| 1 | \input texinfo @c -*-texinfo-*- |
| 2 | @comment %**start of header |
| 3 | @setfilename ../../info/eintr |
| 4 | @c setfilename emacs-lisp-intro.info |
| 5 | @c sethtmlfilename emacs-lisp-intro.html |
| 6 | @settitle Programming in Emacs Lisp |
| 7 | @syncodeindex vr cp |
| 8 | @syncodeindex fn cp |
| 9 | @finalout |
| 10 | |
| 11 | @include emacsver.texi |
| 12 | |
| 13 | @c ================ How to Print a Book in Various Sizes ================ |
| 14 | |
| 15 | @c This book can be printed in any of three different sizes. |
| 16 | @c Set the following @-commands appropriately. |
| 17 | |
| 18 | @c 7 by 9.25 inches: |
| 19 | @c @smallbook |
| 20 | @c @clear largebook |
| 21 | |
| 22 | @c 8.5 by 11 inches: |
| 23 | @c @c smallbook |
| 24 | @c @set largebook |
| 25 | |
| 26 | @c European A4 size paper: |
| 27 | @c @c smallbook |
| 28 | @c @afourpaper |
| 29 | @c @set largebook |
| 30 | |
| 31 | @c (Note: if you edit the book so as to change the length of the |
| 32 | @c table of contents, you may have to change the value of `pageno' below.) |
| 33 | |
| 34 | @c <<<< For hard copy printing, this file is now |
| 35 | @c set for smallbook, which works for all sizes |
| 36 | @c of paper, and with PostScript figures >>>> |
| 37 | |
| 38 | @set smallbook |
| 39 | @ifset smallbook |
| 40 | @smallbook |
| 41 | @clear largebook |
| 42 | @end ifset |
| 43 | |
| 44 | @c ================ Included Figures ================ |
| 45 | |
| 46 | @c If you clear this, the figures will be printed as ASCII diagrams |
| 47 | @c rather than PostScript/PDF. |
| 48 | @c (This is not relevant to Info, since Info only handles ASCII.) |
| 49 | @set print-postscript-figures |
| 50 | @c clear print-postscript-figures |
| 51 | |
| 52 | @comment %**end of header |
| 53 | |
| 54 | @c per rms and peterb, use 10pt fonts for the main text, mostly to |
| 55 | @c save on paper cost. |
| 56 | @c Do this inside @tex for now, so current makeinfo does not complain. |
| 57 | @tex |
| 58 | @ifset smallbook |
| 59 | @fonttextsize 10 |
| 60 | |
| 61 | @end ifset |
| 62 | \global\hbadness=6666 % don't worry about not-too-underfull boxes |
| 63 | @end tex |
| 64 | |
| 65 | @c These refer to the printed book sold by the FSF. |
| 66 | @set edition-number 3.10 |
| 67 | @set update-date 28 October 2009 |
| 68 | |
| 69 | @c For next or subsequent edition: |
| 70 | @c create function using with-output-to-temp-buffer |
| 71 | @c create a major mode, with keymaps |
| 72 | @c run an asynchronous process, like grep or diff |
| 73 | |
| 74 | @c For 8.5 by 11 inch format: do not use such a small amount of |
| 75 | @c whitespace between paragraphs as smallbook format |
| 76 | @ifset largebook |
| 77 | @tex |
| 78 | \global\parskip 6pt plus 1pt |
| 79 | @end tex |
| 80 | @end ifset |
| 81 | |
| 82 | @c For all sized formats: print within-book cross |
| 83 | @c reference with ``...'' rather than [...] |
| 84 | |
| 85 | @c This works with the texinfo.tex file, version 2003-05-04.08, |
| 86 | @c in the Texinfo version 4.6 of the 2003 Jun 13 distribution. |
| 87 | |
| 88 | @tex |
| 89 | \if \xrefprintnodename |
| 90 | \global\def\xrefprintnodename#1{\unskip, ``#1''} |
| 91 | \else |
| 92 | \global\def\xrefprintnodename#1{ ``#1''} |
| 93 | \fi |
| 94 | % \global\def\xrefprintnodename#1{, ``#1''} |
| 95 | @end tex |
| 96 | |
| 97 | @c ---------------------------------------------------- |
| 98 | |
| 99 | @dircategory GNU Emacs Lisp |
| 100 | @direntry |
| 101 | * Emacs Lisp Intro: (eintr). |
| 102 | A simple introduction to Emacs Lisp programming. |
| 103 | @end direntry |
| 104 | |
| 105 | @copying |
| 106 | This is an @cite{Introduction to Programming in Emacs Lisp}, for |
| 107 | people who are not programmers. |
| 108 | @sp 1 |
| 109 | @iftex |
| 110 | Edition @value{edition-number}, @value{update-date} |
| 111 | @end iftex |
| 112 | @ifnottex |
| 113 | Distributed with Emacs version @value{EMACSVER}. |
| 114 | @end ifnottex |
| 115 | @sp 1 |
| 116 | Copyright @copyright{} 1990--1995, 1997, 2001--2013 Free Software |
| 117 | Foundation, Inc. |
| 118 | @sp 1 |
| 119 | |
| 120 | @iftex |
| 121 | Published by the:@* |
| 122 | |
| 123 | GNU Press, @hfill @uref{http://www.fsf.org/licensing/gnu-press/}@* |
| 124 | a division of the @hfill email: @email{sales@@fsf.org}@* |
| 125 | Free Software Foundation, Inc. @hfill Tel: +1 (617) 542-5942@* |
| 126 | 51 Franklin Street, Fifth Floor @hfill Fax: +1 (617) 542-2652@* |
| 127 | Boston, MA 02110-1301 USA |
| 128 | @end iftex |
| 129 | |
| 130 | @ifnottex |
| 131 | Printed copies available from @uref{http://shop.fsf.org/}. Published by: |
| 132 | |
| 133 | @example |
| 134 | GNU Press, http://www.fsf.org/licensing/gnu-press/ |
| 135 | a division of the email: sales@@fsf.org |
| 136 | Free Software Foundation, Inc. Tel: +1 (617) 542-5942 |
| 137 | 51 Franklin Street, Fifth Floor Fax: +1 (617) 542-2652 |
| 138 | Boston, MA 02110-1301 USA |
| 139 | @end example |
| 140 | @end ifnottex |
| 141 | |
| 142 | @sp 1 |
| 143 | ISBN 1-882114-43-4 |
| 144 | |
| 145 | Permission is granted to copy, distribute and/or modify this document |
| 146 | under the terms of the GNU Free Documentation License, Version 1.3 or |
| 147 | any later version published by the Free Software Foundation; there |
| 148 | being no Invariant Section, with the Front-Cover Texts being ``A GNU |
| 149 | Manual'', and with the Back-Cover Texts as in (a) below. A copy of |
| 150 | the license is included in the section entitled ``GNU Free |
| 151 | Documentation License''. |
| 152 | |
| 153 | (a) The FSF's Back-Cover Text is: ``You have the freedom to |
| 154 | copy and modify this GNU manual. Buying copies from the FSF |
| 155 | supports it in developing GNU and promoting software freedom.'' |
| 156 | @end copying |
| 157 | |
| 158 | @c half title; two lines here, so do not use `shorttitlepage' |
| 159 | @tex |
| 160 | {\begingroup% |
| 161 | \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}% |
| 162 | \endgroup}% |
| 163 | {\begingroup\hbox{}\vskip 0.25in \chaprm% |
| 164 | \centerline{Programming in Emacs Lisp}% |
| 165 | \endgroup\page\hbox{}\page} |
| 166 | @end tex |
| 167 | |
| 168 | @titlepage |
| 169 | @sp 6 |
| 170 | @center @titlefont{An Introduction to} |
| 171 | @sp 2 |
| 172 | @center @titlefont{Programming in Emacs Lisp} |
| 173 | @sp 2 |
| 174 | @center Revised Third Edition |
| 175 | @sp 4 |
| 176 | @center by Robert J. Chassell |
| 177 | |
| 178 | @page |
| 179 | @vskip 0pt plus 1filll |
| 180 | @insertcopying |
| 181 | @end titlepage |
| 182 | |
| 183 | @iftex |
| 184 | @headings off |
| 185 | @evenheading @thispage @| @| @thischapter |
| 186 | @oddheading @thissection @| @| @thispage |
| 187 | @end iftex |
| 188 | |
| 189 | @ifnothtml |
| 190 | @c Keep T.O.C. short by tightening up for largebook |
| 191 | @ifset largebook |
| 192 | @tex |
| 193 | \global\parskip 2pt plus 1pt |
| 194 | \global\advance\baselineskip by -1pt |
| 195 | @end tex |
| 196 | @end ifset |
| 197 | @end ifnothtml |
| 198 | |
| 199 | @shortcontents |
| 200 | @contents |
| 201 | |
| 202 | @ifnottex |
| 203 | @node Top |
| 204 | @top An Introduction to Programming in Emacs Lisp |
| 205 | |
| 206 | @ifset WWW_GNU_ORG |
| 207 | @html |
| 208 | <p>The homepage for GNU Emacs is at |
| 209 | <a href="/software/emacs/">http://www.gnu.org/software/emacs/</a>.<br> |
| 210 | To view this manual in other formats, click |
| 211 | <a href="/software/emacs/manual/eintr.html">here</a>. |
| 212 | @end html |
| 213 | @end ifset |
| 214 | |
| 215 | @insertcopying |
| 216 | |
| 217 | This master menu first lists each chapter and index; then it lists |
| 218 | every node in every chapter. |
| 219 | @end ifnottex |
| 220 | |
| 221 | @c >>>> Set pageno appropriately <<<< |
| 222 | |
| 223 | @c The first page of the Preface is a roman numeral; it is the first |
| 224 | @c right handed page after the Table of Contents; hence the following |
| 225 | @c setting must be for an odd negative number. |
| 226 | |
| 227 | @c iftex |
| 228 | @c global@pageno = -11 |
| 229 | @c end iftex |
| 230 | |
| 231 | @set COUNT-WORDS count-words-example |
| 232 | @c Length of variable name chosen so that things still line up when expanded. |
| 233 | |
| 234 | @menu |
| 235 | * Preface:: What to look for. |
| 236 | * List Processing:: What is Lisp? |
| 237 | * Practicing Evaluation:: Running several programs. |
| 238 | * Writing Defuns:: How to write function definitions. |
| 239 | * Buffer Walk Through:: Exploring a few buffer-related functions. |
| 240 | * More Complex:: A few, even more complex functions. |
| 241 | * Narrowing & Widening:: Restricting your and Emacs attention to |
| 242 | a region. |
| 243 | * car cdr & cons:: Fundamental functions in Lisp. |
| 244 | * Cutting & Storing Text:: Removing text and saving it. |
| 245 | * List Implementation:: How lists are implemented in the computer. |
| 246 | * Yanking:: Pasting stored text. |
| 247 | * Loops & Recursion:: How to repeat a process. |
| 248 | * Regexp Search:: Regular expression searches. |
| 249 | * Counting Words:: A review of repetition and regexps. |
| 250 | * Words in a defun:: Counting words in a @code{defun}. |
| 251 | * Readying a Graph:: A prototype graph printing function. |
| 252 | * Emacs Initialization:: How to write a @file{.emacs} file. |
| 253 | * Debugging:: How to run the Emacs Lisp debuggers. |
| 254 | * Conclusion:: Now you have the basics. |
| 255 | * the-the:: An appendix: how to find reduplicated words. |
| 256 | * Kill Ring:: An appendix: how the kill ring works. |
| 257 | * Full Graph:: How to create a graph with labeled axes. |
| 258 | * Free Software and Free Manuals:: |
| 259 | * GNU Free Documentation License:: |
| 260 | * Index:: |
| 261 | * About the Author:: |
| 262 | |
| 263 | @detailmenu |
| 264 | --- The Detailed Node Listing --- |
| 265 | |
| 266 | Preface |
| 267 | |
| 268 | * Why:: Why learn Emacs Lisp? |
| 269 | * On Reading this Text:: Read, gain familiarity, pick up habits.... |
| 270 | * Who You Are:: For whom this is written. |
| 271 | * Lisp History:: |
| 272 | * Note for Novices:: You can read this as a novice. |
| 273 | * Thank You:: |
| 274 | |
| 275 | List Processing |
| 276 | |
| 277 | * Lisp Lists:: What are lists? |
| 278 | * Run a Program:: Any list in Lisp is a program ready to run. |
| 279 | * Making Errors:: Generating an error message. |
| 280 | * Names & Definitions:: Names of symbols and function definitions. |
| 281 | * Lisp Interpreter:: What the Lisp interpreter does. |
| 282 | * Evaluation:: Running a program. |
| 283 | * Variables:: Returning a value from a variable. |
| 284 | * Arguments:: Passing information to a function. |
| 285 | * set & setq:: Setting the value of a variable. |
| 286 | * Summary:: The major points. |
| 287 | * Error Message Exercises:: |
| 288 | |
| 289 | Lisp Lists |
| 290 | |
| 291 | * Numbers Lists:: List have numbers, other lists, in them. |
| 292 | * Lisp Atoms:: Elemental entities. |
| 293 | * Whitespace in Lists:: Formatting lists to be readable. |
| 294 | * Typing Lists:: How GNU Emacs helps you type lists. |
| 295 | |
| 296 | The Lisp Interpreter |
| 297 | |
| 298 | * Complications:: Variables, Special forms, Lists within. |
| 299 | * Byte Compiling:: Specially processing code for speed. |
| 300 | |
| 301 | Evaluation |
| 302 | |
| 303 | * How the Interpreter Acts:: Returns and Side Effects... |
| 304 | * Evaluating Inner Lists:: Lists within lists... |
| 305 | |
| 306 | Variables |
| 307 | |
| 308 | * fill-column Example:: |
| 309 | * Void Function:: The error message for a symbol |
| 310 | without a function. |
| 311 | * Void Variable:: The error message for a symbol without a value. |
| 312 | |
| 313 | Arguments |
| 314 | |
| 315 | * Data types:: Types of data passed to a function. |
| 316 | * Args as Variable or List:: An argument can be the value |
| 317 | of a variable or list. |
| 318 | * Variable Number of Arguments:: Some functions may take a |
| 319 | variable number of arguments. |
| 320 | * Wrong Type of Argument:: Passing an argument of the wrong type |
| 321 | to a function. |
| 322 | * message:: A useful function for sending messages. |
| 323 | |
| 324 | Setting the Value of a Variable |
| 325 | |
| 326 | * Using set:: Setting values. |
| 327 | * Using setq:: Setting a quoted value. |
| 328 | * Counting:: Using @code{setq} to count. |
| 329 | |
| 330 | Practicing Evaluation |
| 331 | |
| 332 | * How to Evaluate:: Typing editing commands or @kbd{C-x C-e} |
| 333 | causes evaluation. |
| 334 | * Buffer Names:: Buffers and files are different. |
| 335 | * Getting Buffers:: Getting a buffer itself, not merely its name. |
| 336 | * Switching Buffers:: How to change to another buffer. |
| 337 | * Buffer Size & Locations:: Where point is located and the size of |
| 338 | the buffer. |
| 339 | * Evaluation Exercise:: |
| 340 | |
| 341 | How To Write Function Definitions |
| 342 | |
| 343 | * Primitive Functions:: |
| 344 | * defun:: The @code{defun} macro. |
| 345 | * Install:: Install a function definition. |
| 346 | * Interactive:: Making a function interactive. |
| 347 | * Interactive Options:: Different options for @code{interactive}. |
| 348 | * Permanent Installation:: Installing code permanently. |
| 349 | * let:: Creating and initializing local variables. |
| 350 | * if:: What if? |
| 351 | * else:: If--then--else expressions. |
| 352 | * Truth & Falsehood:: What Lisp considers false and true. |
| 353 | * save-excursion:: Keeping track of point, mark, and buffer. |
| 354 | * Review:: |
| 355 | * defun Exercises:: |
| 356 | |
| 357 | Install a Function Definition |
| 358 | |
| 359 | * Effect of installation:: |
| 360 | * Change a defun:: How to change a function definition. |
| 361 | |
| 362 | Make a Function Interactive |
| 363 | |
| 364 | * Interactive multiply-by-seven:: An overview. |
| 365 | * multiply-by-seven in detail:: The interactive version. |
| 366 | |
| 367 | @code{let} |
| 368 | |
| 369 | * Prevent confusion:: |
| 370 | * Parts of let Expression:: |
| 371 | * Sample let Expression:: |
| 372 | * Uninitialized let Variables:: |
| 373 | |
| 374 | The @code{if} Special Form |
| 375 | |
| 376 | * if in more detail:: |
| 377 | * type-of-animal in detail:: An example of an @code{if} expression. |
| 378 | |
| 379 | Truth and Falsehood in Emacs Lisp |
| 380 | |
| 381 | * nil explained:: @code{nil} has two meanings. |
| 382 | |
| 383 | @code{save-excursion} |
| 384 | |
| 385 | * Point and mark:: A review of various locations. |
| 386 | * Template for save-excursion:: |
| 387 | |
| 388 | A Few Buffer--Related Functions |
| 389 | |
| 390 | * Finding More:: How to find more information. |
| 391 | * simplified-beginning-of-buffer:: Shows @code{goto-char}, |
| 392 | @code{point-min}, and @code{push-mark}. |
| 393 | * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}. |
| 394 | * append-to-buffer:: Uses @code{save-excursion} and |
| 395 | @code{insert-buffer-substring}. |
| 396 | * Buffer Related Review:: Review. |
| 397 | * Buffer Exercises:: |
| 398 | |
| 399 | The Definition of @code{mark-whole-buffer} |
| 400 | |
| 401 | * mark-whole-buffer overview:: |
| 402 | * Body of mark-whole-buffer:: Only three lines of code. |
| 403 | |
| 404 | The Definition of @code{append-to-buffer} |
| 405 | |
| 406 | * append-to-buffer overview:: |
| 407 | * append interactive:: A two part interactive expression. |
| 408 | * append-to-buffer body:: Incorporates a @code{let} expression. |
| 409 | * append save-excursion:: How the @code{save-excursion} works. |
| 410 | |
| 411 | A Few More Complex Functions |
| 412 | |
| 413 | * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}. |
| 414 | * insert-buffer:: Read-only, and with @code{or}. |
| 415 | * beginning-of-buffer:: Shows @code{goto-char}, |
| 416 | @code{point-min}, and @code{push-mark}. |
| 417 | * Second Buffer Related Review:: |
| 418 | * optional Exercise:: |
| 419 | |
| 420 | The Definition of @code{insert-buffer} |
| 421 | |
| 422 | * insert-buffer code:: |
| 423 | * insert-buffer interactive:: When you can read, but not write. |
| 424 | * insert-buffer body:: The body has an @code{or} and a @code{let}. |
| 425 | * if & or:: Using an @code{if} instead of an @code{or}. |
| 426 | * Insert or:: How the @code{or} expression works. |
| 427 | * Insert let:: Two @code{save-excursion} expressions. |
| 428 | * New insert-buffer:: |
| 429 | |
| 430 | The Interactive Expression in @code{insert-buffer} |
| 431 | |
| 432 | * Read-only buffer:: When a buffer cannot be modified. |
| 433 | * b for interactive:: An existing buffer or else its name. |
| 434 | |
| 435 | Complete Definition of @code{beginning-of-buffer} |
| 436 | |
| 437 | * Optional Arguments:: |
| 438 | * beginning-of-buffer opt arg:: Example with optional argument. |
| 439 | * beginning-of-buffer complete:: |
| 440 | |
| 441 | @code{beginning-of-buffer} with an Argument |
| 442 | |
| 443 | * Disentangle beginning-of-buffer:: |
| 444 | * Large buffer case:: |
| 445 | * Small buffer case:: |
| 446 | |
| 447 | Narrowing and Widening |
| 448 | |
| 449 | * Narrowing advantages:: The advantages of narrowing |
| 450 | * save-restriction:: The @code{save-restriction} special form. |
| 451 | * what-line:: The number of the line that point is on. |
| 452 | * narrow Exercise:: |
| 453 | |
| 454 | @code{car}, @code{cdr}, @code{cons}: Fundamental Functions |
| 455 | |
| 456 | * Strange Names:: An historical aside: why the strange names? |
| 457 | * car & cdr:: Functions for extracting part of a list. |
| 458 | * cons:: Constructing a list. |
| 459 | * nthcdr:: Calling @code{cdr} repeatedly. |
| 460 | * nth:: |
| 461 | * setcar:: Changing the first element of a list. |
| 462 | * setcdr:: Changing the rest of a list. |
| 463 | * cons Exercise:: |
| 464 | |
| 465 | @code{cons} |
| 466 | |
| 467 | * Build a list:: |
| 468 | * length:: How to find the length of a list. |
| 469 | |
| 470 | Cutting and Storing Text |
| 471 | |
| 472 | * Storing Text:: Text is stored in a list. |
| 473 | * zap-to-char:: Cutting out text up to a character. |
| 474 | * kill-region:: Cutting text out of a region. |
| 475 | * copy-region-as-kill:: A definition for copying text. |
| 476 | * Digression into C:: Minor note on C programming language macros. |
| 477 | * defvar:: How to give a variable an initial value. |
| 478 | * cons & search-fwd Review:: |
| 479 | * search Exercises:: |
| 480 | |
| 481 | @code{zap-to-char} |
| 482 | |
| 483 | * Complete zap-to-char:: The complete implementation. |
| 484 | * zap-to-char interactive:: A three part interactive expression. |
| 485 | * zap-to-char body:: A short overview. |
| 486 | * search-forward:: How to search for a string. |
| 487 | * progn:: The @code{progn} special form. |
| 488 | * Summing up zap-to-char:: Using @code{point} and @code{search-forward}. |
| 489 | |
| 490 | @code{kill-region} |
| 491 | |
| 492 | * Complete kill-region:: The function definition. |
| 493 | * condition-case:: Dealing with a problem. |
| 494 | * Lisp macro:: |
| 495 | |
| 496 | @code{copy-region-as-kill} |
| 497 | |
| 498 | * Complete copy-region-as-kill:: The complete function definition. |
| 499 | * copy-region-as-kill body:: The body of @code{copy-region-as-kill}. |
| 500 | |
| 501 | The Body of @code{copy-region-as-kill} |
| 502 | |
| 503 | * last-command & this-command:: |
| 504 | * kill-append function:: |
| 505 | * kill-new function:: |
| 506 | |
| 507 | Initializing a Variable with @code{defvar} |
| 508 | |
| 509 | * See variable current value:: |
| 510 | * defvar and asterisk:: |
| 511 | |
| 512 | How Lists are Implemented |
| 513 | |
| 514 | * Lists diagrammed:: |
| 515 | * Symbols as Chest:: Exploring a powerful metaphor. |
| 516 | * List Exercise:: |
| 517 | |
| 518 | Yanking Text Back |
| 519 | |
| 520 | * Kill Ring Overview:: |
| 521 | * kill-ring-yank-pointer:: The kill ring is a list. |
| 522 | * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable. |
| 523 | |
| 524 | Loops and Recursion |
| 525 | |
| 526 | * while:: Causing a stretch of code to repeat. |
| 527 | * dolist dotimes:: |
| 528 | * Recursion:: Causing a function to call itself. |
| 529 | * Looping exercise:: |
| 530 | |
| 531 | @code{while} |
| 532 | |
| 533 | * Looping with while:: Repeat so long as test returns true. |
| 534 | * Loop Example:: A @code{while} loop that uses a list. |
| 535 | * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}. |
| 536 | * Incrementing Loop:: A loop with an incrementing counter. |
| 537 | * Incrementing Loop Details:: |
| 538 | * Decrementing Loop:: A loop with a decrementing counter. |
| 539 | |
| 540 | Details of an Incrementing Loop |
| 541 | |
| 542 | * Incrementing Example:: Counting pebbles in a triangle. |
| 543 | * Inc Example parts:: The parts of the function definition. |
| 544 | * Inc Example altogether:: Putting the function definition together. |
| 545 | |
| 546 | Loop with a Decrementing Counter |
| 547 | |
| 548 | * Decrementing Example:: More pebbles on the beach. |
| 549 | * Dec Example parts:: The parts of the function definition. |
| 550 | * Dec Example altogether:: Putting the function definition together. |
| 551 | |
| 552 | Save your time: @code{dolist} and @code{dotimes} |
| 553 | |
| 554 | * dolist:: |
| 555 | * dotimes:: |
| 556 | |
| 557 | Recursion |
| 558 | |
| 559 | * Building Robots:: Same model, different serial number ... |
| 560 | * Recursive Definition Parts:: Walk until you stop ... |
| 561 | * Recursion with list:: Using a list as the test whether to recurse. |
| 562 | * Recursive triangle function:: |
| 563 | * Recursion with cond:: |
| 564 | * Recursive Patterns:: Often used templates. |
| 565 | * No Deferment:: Don't store up work ... |
| 566 | * No deferment solution:: |
| 567 | |
| 568 | Recursion in Place of a Counter |
| 569 | |
| 570 | * Recursive Example arg of 1 or 2:: |
| 571 | * Recursive Example arg of 3 or 4:: |
| 572 | |
| 573 | Recursive Patterns |
| 574 | |
| 575 | * Every:: |
| 576 | * Accumulate:: |
| 577 | * Keep:: |
| 578 | |
| 579 | Regular Expression Searches |
| 580 | |
| 581 | * sentence-end:: The regular expression for @code{sentence-end}. |
| 582 | * re-search-forward:: Very similar to @code{search-forward}. |
| 583 | * forward-sentence:: A straightforward example of regexp search. |
| 584 | * forward-paragraph:: A somewhat complex example. |
| 585 | * etags:: How to create your own @file{TAGS} table. |
| 586 | * Regexp Review:: |
| 587 | * re-search Exercises:: |
| 588 | |
| 589 | @code{forward-sentence} |
| 590 | |
| 591 | * Complete forward-sentence:: |
| 592 | * fwd-sentence while loops:: Two @code{while} loops. |
| 593 | * fwd-sentence re-search:: A regular expression search. |
| 594 | |
| 595 | @code{forward-paragraph}: a Goldmine of Functions |
| 596 | |
| 597 | * forward-paragraph in brief:: Key parts of the function definition. |
| 598 | * fwd-para let:: The @code{let*} expression. |
| 599 | * fwd-para while:: The forward motion @code{while} loop. |
| 600 | |
| 601 | Counting: Repetition and Regexps |
| 602 | |
| 603 | * Why Count Words:: |
| 604 | * @value{COUNT-WORDS}:: Use a regexp, but find a problem. |
| 605 | * recursive-count-words:: Start with case of no words in region. |
| 606 | * Counting Exercise:: |
| 607 | |
| 608 | The @code{@value{COUNT-WORDS}} Function |
| 609 | |
| 610 | * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop. |
| 611 | * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}. |
| 612 | |
| 613 | Counting Words in a @code{defun} |
| 614 | |
| 615 | * Divide and Conquer:: |
| 616 | * Words and Symbols:: What to count? |
| 617 | * Syntax:: What constitutes a word or symbol? |
| 618 | * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}. |
| 619 | * Several defuns:: Counting several defuns in a file. |
| 620 | * Find a File:: Do you want to look at a file? |
| 621 | * lengths-list-file:: A list of the lengths of many definitions. |
| 622 | * Several files:: Counting in definitions in different files. |
| 623 | * Several files recursively:: Recursively counting in different files. |
| 624 | * Prepare the data:: Prepare the data for display in a graph. |
| 625 | |
| 626 | Count Words in @code{defuns} in Different Files |
| 627 | |
| 628 | * lengths-list-many-files:: Return a list of the lengths of defuns. |
| 629 | * append:: Attach one list to another. |
| 630 | |
| 631 | Prepare the Data for Display in a Graph |
| 632 | |
| 633 | * Data for Display in Detail:: |
| 634 | * Sorting:: Sorting lists. |
| 635 | * Files List:: Making a list of files. |
| 636 | * Counting function definitions:: |
| 637 | |
| 638 | Readying a Graph |
| 639 | |
| 640 | * Columns of a graph:: |
| 641 | * graph-body-print:: How to print the body of a graph. |
| 642 | * recursive-graph-body-print:: |
| 643 | * Printed Axes:: |
| 644 | * Line Graph Exercise:: |
| 645 | |
| 646 | Your @file{.emacs} File |
| 647 | |
| 648 | * Default Configuration:: |
| 649 | * Site-wide Init:: You can write site-wide init files. |
| 650 | * defcustom:: Emacs will write code for you. |
| 651 | * Beginning init File:: How to write a @file{.emacs} init file. |
| 652 | * Text and Auto-fill:: Automatically wrap lines. |
| 653 | * Mail Aliases:: Use abbreviations for email addresses. |
| 654 | * Indent Tabs Mode:: Don't use tabs with @TeX{} |
| 655 | * Keybindings:: Create some personal keybindings. |
| 656 | * Keymaps:: More about key binding. |
| 657 | * Loading Files:: Load (i.e., evaluate) files automatically. |
| 658 | * Autoload:: Make functions available. |
| 659 | * Simple Extension:: Define a function; bind it to a key. |
| 660 | * X11 Colors:: Colors in X. |
| 661 | * Miscellaneous:: |
| 662 | * Mode Line:: How to customize your mode line. |
| 663 | |
| 664 | Debugging |
| 665 | |
| 666 | * debug:: How to use the built-in debugger. |
| 667 | * debug-on-entry:: Start debugging when you call a function. |
| 668 | * debug-on-quit:: Start debugging when you quit with @kbd{C-g}. |
| 669 | * edebug:: How to use Edebug, a source level debugger. |
| 670 | * Debugging Exercises:: |
| 671 | |
| 672 | Handling the Kill Ring |
| 673 | |
| 674 | * What the Kill Ring Does:: |
| 675 | * current-kill:: |
| 676 | * yank:: Paste a copy of a clipped element. |
| 677 | * yank-pop:: Insert element pointed to. |
| 678 | * ring file:: |
| 679 | |
| 680 | The @code{current-kill} Function |
| 681 | |
| 682 | * Code for current-kill:: |
| 683 | * Understanding current-kill:: |
| 684 | |
| 685 | @code{current-kill} in Outline |
| 686 | |
| 687 | * Body of current-kill:: |
| 688 | * Digression concerning error:: How to mislead humans, but not computers. |
| 689 | * Determining the Element:: |
| 690 | |
| 691 | A Graph with Labeled Axes |
| 692 | |
| 693 | * Labeled Example:: |
| 694 | * print-graph Varlist:: @code{let} expression in @code{print-graph}. |
| 695 | * print-Y-axis:: Print a label for the vertical axis. |
| 696 | * print-X-axis:: Print a horizontal label. |
| 697 | * Print Whole Graph:: The function to print a complete graph. |
| 698 | |
| 699 | The @code{print-Y-axis} Function |
| 700 | |
| 701 | * print-Y-axis in Detail:: |
| 702 | * Height of label:: What height for the Y axis? |
| 703 | * Compute a Remainder:: How to compute the remainder of a division. |
| 704 | * Y Axis Element:: Construct a line for the Y axis. |
| 705 | * Y-axis-column:: Generate a list of Y axis labels. |
| 706 | * print-Y-axis Penultimate:: A not quite final version. |
| 707 | |
| 708 | The @code{print-X-axis} Function |
| 709 | |
| 710 | * Similarities differences:: Much like @code{print-Y-axis}, but not exactly. |
| 711 | * X Axis Tic Marks:: Create tic marks for the horizontal axis. |
| 712 | |
| 713 | Printing the Whole Graph |
| 714 | |
| 715 | * The final version:: A few changes. |
| 716 | * Test print-graph:: Run a short test. |
| 717 | * Graphing words in defuns:: Executing the final code. |
| 718 | * lambda:: How to write an anonymous function. |
| 719 | * mapcar:: Apply a function to elements of a list. |
| 720 | * Another Bug:: Yet another bug @dots{} most insidious. |
| 721 | * Final printed graph:: The graph itself! |
| 722 | |
| 723 | @end detailmenu |
| 724 | @end menu |
| 725 | |
| 726 | @node Preface |
| 727 | @unnumbered Preface |
| 728 | |
| 729 | Most of the GNU Emacs integrated environment is written in the programming |
| 730 | language called Emacs Lisp. The code written in this programming |
| 731 | language is the software---the sets of instructions---that tell the |
| 732 | computer what to do when you give it commands. Emacs is designed so |
| 733 | that you can write new code in Emacs Lisp and easily install it as an |
| 734 | extension to the editor. |
| 735 | |
| 736 | (GNU Emacs is sometimes called an ``extensible editor'', but it does |
| 737 | much more than provide editing capabilities. It is better to refer to |
| 738 | Emacs as an ``extensible computing environment''. However, that |
| 739 | phrase is quite a mouthful. It is easier to refer to Emacs simply as |
| 740 | an editor. Moreover, everything you do in Emacs---find the Mayan date |
| 741 | and phases of the moon, simplify polynomials, debug code, manage |
| 742 | files, read letters, write books---all these activities are kinds of |
| 743 | editing in the most general sense of the word.) |
| 744 | |
| 745 | @menu |
| 746 | * Why:: Why learn Emacs Lisp? |
| 747 | * On Reading this Text:: Read, gain familiarity, pick up habits.... |
| 748 | * Who You Are:: For whom this is written. |
| 749 | * Lisp History:: |
| 750 | * Note for Novices:: You can read this as a novice. |
| 751 | * Thank You:: |
| 752 | @end menu |
| 753 | |
| 754 | @ifnottex |
| 755 | @node Why |
| 756 | @unnumberedsec Why Study Emacs Lisp? |
| 757 | @end ifnottex |
| 758 | |
| 759 | Although Emacs Lisp is usually thought of in association only with Emacs, |
| 760 | it is a full computer programming language. You can use Emacs Lisp as |
| 761 | you would any other programming language. |
| 762 | |
| 763 | Perhaps you want to understand programming; perhaps you want to extend |
| 764 | Emacs; or perhaps you want to become a programmer. This introduction to |
| 765 | Emacs Lisp is designed to get you started: to guide you in learning the |
| 766 | fundamentals of programming, and more importantly, to show you how you |
| 767 | can teach yourself to go further. |
| 768 | |
| 769 | @node On Reading this Text |
| 770 | @unnumberedsec On Reading this Text |
| 771 | |
| 772 | All through this document, you will see little sample programs you can |
| 773 | run inside of Emacs. If you read this document in Info inside of GNU |
| 774 | Emacs, you can run the programs as they appear. (This is easy to do and |
| 775 | is explained when the examples are presented.) Alternatively, you can |
| 776 | read this introduction as a printed book while sitting beside a computer |
| 777 | running Emacs. (This is what I like to do; I like printed books.) If |
| 778 | you don't have a running Emacs beside you, you can still read this book, |
| 779 | but in this case, it is best to treat it as a novel or as a travel guide |
| 780 | to a country not yet visited: interesting, but not the same as being |
| 781 | there. |
| 782 | |
| 783 | Much of this introduction is dedicated to walkthroughs or guided tours |
| 784 | of code used in GNU Emacs. These tours are designed for two purposes: |
| 785 | first, to give you familiarity with real, working code (code you use |
| 786 | every day); and, second, to give you familiarity with the way Emacs |
| 787 | works. It is interesting to see how a working environment is |
| 788 | implemented. |
| 789 | Also, I |
| 790 | hope that you will pick up the habit of browsing through source code. |
| 791 | You can learn from it and mine it for ideas. Having GNU Emacs is like |
| 792 | having a dragon's cave of treasures. |
| 793 | |
| 794 | In addition to learning about Emacs as an editor and Emacs Lisp as a |
| 795 | programming language, the examples and guided tours will give you an |
| 796 | opportunity to get acquainted with Emacs as a Lisp programming |
| 797 | environment. GNU Emacs supports programming and provides tools that |
| 798 | you will want to become comfortable using, such as @kbd{M-.} (the key |
| 799 | which invokes the @code{find-tag} command). You will also learn about |
| 800 | buffers and other objects that are part of the environment. |
| 801 | Learning about these features of Emacs is like learning new routes |
| 802 | around your home town. |
| 803 | |
| 804 | @ignore |
| 805 | In addition, I have written several programs as extended examples. |
| 806 | Although these are examples, the programs are real. I use them. |
| 807 | Other people use them. You may use them. Beyond the fragments of |
| 808 | programs used for illustrations, there is very little in here that is |
| 809 | `just for teaching purposes'; what you see is used. This is a great |
| 810 | advantage of Emacs Lisp: it is easy to learn to use it for work. |
| 811 | @end ignore |
| 812 | |
| 813 | Finally, I hope to convey some of the skills for using Emacs to |
| 814 | learn aspects of programming that you don't know. You can often use |
| 815 | Emacs to help you understand what puzzles you or to find out how to do |
| 816 | something new. This self-reliance is not only a pleasure, but an |
| 817 | advantage. |
| 818 | |
| 819 | @node Who You Are |
| 820 | @unnumberedsec For Whom This is Written |
| 821 | |
| 822 | This text is written as an elementary introduction for people who are |
| 823 | not programmers. If you are a programmer, you may not be satisfied with |
| 824 | this primer. The reason is that you may have become expert at reading |
| 825 | reference manuals and be put off by the way this text is organized. |
| 826 | |
| 827 | An expert programmer who reviewed this text said to me: |
| 828 | |
| 829 | @quotation |
| 830 | @i{I prefer to learn from reference manuals. I ``dive into'' each |
| 831 | paragraph, and ``come up for air'' between paragraphs.} |
| 832 | |
| 833 | @i{When I get to the end of a paragraph, I assume that that subject is |
| 834 | done, finished, that I know everything I need (with the |
| 835 | possible exception of the case when the next paragraph starts talking |
| 836 | about it in more detail). I expect that a well written reference manual |
| 837 | will not have a lot of redundancy, and that it will have excellent |
| 838 | pointers to the (one) place where the information I want is.} |
| 839 | @end quotation |
| 840 | |
| 841 | This introduction is not written for this person! |
| 842 | |
| 843 | Firstly, I try to say everything at least three times: first, to |
| 844 | introduce it; second, to show it in context; and third, to show it in a |
| 845 | different context, or to review it. |
| 846 | |
| 847 | Secondly, I hardly ever put all the information about a subject in one |
| 848 | place, much less in one paragraph. To my way of thinking, that imposes |
| 849 | too heavy a burden on the reader. Instead I try to explain only what |
| 850 | you need to know at the time. (Sometimes I include a little extra |
| 851 | information so you won't be surprised later when the additional |
| 852 | information is formally introduced.) |
| 853 | |
| 854 | When you read this text, you are not expected to learn everything the |
| 855 | first time. Frequently, you need only make, as it were, a `nodding |
| 856 | acquaintance' with some of the items mentioned. My hope is that I have |
| 857 | structured the text and given you enough hints that you will be alert to |
| 858 | what is important, and concentrate on it. |
| 859 | |
| 860 | You will need to ``dive into'' some paragraphs; there is no other way |
| 861 | to read them. But I have tried to keep down the number of such |
| 862 | paragraphs. This book is intended as an approachable hill, rather than |
| 863 | as a daunting mountain. |
| 864 | |
| 865 | This introduction to @cite{Programming in Emacs Lisp} has a companion |
| 866 | document, |
| 867 | @iftex |
| 868 | @cite{The GNU Emacs Lisp Reference Manual}. |
| 869 | @end iftex |
| 870 | @ifnottex |
| 871 | @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU |
| 872 | Emacs Lisp Reference Manual}. |
| 873 | @end ifnottex |
| 874 | The reference manual has more detail than this introduction. In the |
| 875 | reference manual, all the information about one topic is concentrated |
| 876 | in one place. You should turn to it if you are like the programmer |
| 877 | quoted above. And, of course, after you have read this |
| 878 | @cite{Introduction}, you will find the @cite{Reference Manual} useful |
| 879 | when you are writing your own programs. |
| 880 | |
| 881 | @node Lisp History |
| 882 | @unnumberedsec Lisp History |
| 883 | @cindex Lisp history |
| 884 | |
| 885 | Lisp was first developed in the late 1950s at the Massachusetts |
| 886 | Institute of Technology for research in artificial intelligence. The |
| 887 | great power of the Lisp language makes it superior for other purposes as |
| 888 | well, such as writing editor commands and integrated environments. |
| 889 | |
| 890 | @cindex Maclisp |
| 891 | @cindex Common Lisp |
| 892 | GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT |
| 893 | in the 1960s. It is somewhat inspired by Common Lisp, which became a |
| 894 | standard in the 1980s. However, Emacs Lisp is much simpler than Common |
| 895 | Lisp. (The standard Emacs distribution contains an optional extensions |
| 896 | file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.) |
| 897 | |
| 898 | @node Note for Novices |
| 899 | @unnumberedsec A Note for Novices |
| 900 | |
| 901 | If you don't know GNU Emacs, you can still read this document |
| 902 | profitably. However, I recommend you learn Emacs, if only to learn to |
| 903 | move around your computer screen. You can teach yourself how to use |
| 904 | Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This |
| 905 | means you press and release the @key{CTRL} key and the @kbd{h} at the |
| 906 | same time, and then press and release @kbd{t}.) |
| 907 | |
| 908 | Also, I often refer to one of Emacs's standard commands by listing the |
| 909 | keys which you press to invoke the command and then giving the name of |
| 910 | the command in parentheses, like this: @kbd{M-C-\} |
| 911 | (@code{indent-region}). What this means is that the |
| 912 | @code{indent-region} command is customarily invoked by typing |
| 913 | @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to |
| 914 | invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, , |
| 915 | Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your |
| 916 | @key{META} key, @key{CTRL} key and @key{\} key all at the same time. |
| 917 | (On many modern keyboards the @key{META} key is labeled |
| 918 | @key{ALT}.) |
| 919 | Sometimes a combination like this is called a keychord, since it is |
| 920 | similar to the way you play a chord on a piano. If your keyboard does |
| 921 | not have a @key{META} key, the @key{ESC} key prefix is used in place |
| 922 | of it. In this case, @kbd{M-C-\} means that you press and release your |
| 923 | @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at |
| 924 | the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key |
| 925 | along with the key that is labeled @key{ALT} and, at the same time, |
| 926 | press the @key{\} key. |
| 927 | |
| 928 | In addition to typing a lone keychord, you can prefix what you type |
| 929 | with @kbd{C-u}, which is called the `universal argument'. The |
| 930 | @kbd{C-u} keychord passes an argument to the subsequent command. |
| 931 | Thus, to indent a region of plain text by 6 spaces, mark the region, |
| 932 | and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number, |
| 933 | Emacs either passes the number 4 to the command or otherwise runs the |
| 934 | command differently than it would otherwise.) @xref{Arguments, , |
| 935 | Numeric Arguments, emacs, The GNU Emacs Manual}. |
| 936 | |
| 937 | If you are reading this in Info using GNU Emacs, you can read through |
| 938 | this whole document just by pressing the space bar, @key{SPC}. |
| 939 | (To learn about Info, type @kbd{C-h i} and then select Info.) |
| 940 | |
| 941 | A note on terminology: when I use the word Lisp alone, I often am |
| 942 | referring to the various dialects of Lisp in general, but when I speak |
| 943 | of Emacs Lisp, I am referring to GNU Emacs Lisp in particular. |
| 944 | |
| 945 | @node Thank You |
| 946 | @unnumberedsec Thank You |
| 947 | |
| 948 | My thanks to all who helped me with this book. My especial thanks to |
| 949 | @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland |
| 950 | McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M. |
| 951 | Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both |
| 952 | @w{Philip Johnson} and @w{David Stampe} for their patient |
| 953 | encouragement. My mistakes are my own. |
| 954 | |
| 955 | @flushright |
| 956 | Robert J. Chassell |
| 957 | @ifnothtml |
| 958 | @email{bob@@gnu.org} |
| 959 | @end ifnothtml |
| 960 | @ifhtml |
| 961 | bob@@gnu.org |
| 962 | @end ifhtml |
| 963 | @end flushright |
| 964 | |
| 965 | @c ================ Beginning of main text ================ |
| 966 | |
| 967 | @c Start main text on right-hand (verso) page |
| 968 | |
| 969 | @tex |
| 970 | \par\vfill\supereject |
| 971 | \headings off |
| 972 | \ifodd\pageno |
| 973 | \par\vfill\supereject |
| 974 | \else |
| 975 | \par\vfill\supereject |
| 976 | \page\hbox{}\page |
| 977 | \par\vfill\supereject |
| 978 | \fi |
| 979 | @end tex |
| 980 | |
| 981 | @c Note: this resetting of the page number back to 1 causes TeX to gripe |
| 982 | @c about already having seen page numbers 1-4 before (in the preface): |
| 983 | @c pdfTeX warning (ext4): destination with the same identifier (name{1}) |
| 984 | @c has been already used, duplicate ignored |
| 985 | @c I guess that is harmless (what happens if a later part of the text |
| 986 | @c makes a link to something in the first 4 pages though?). |
| 987 | @c E.g., note that the Emacs manual has a preface, but does not bother |
| 988 | @c resetting the page numbers back to 1 after that. |
| 989 | @iftex |
| 990 | @headings off |
| 991 | @evenheading @thispage @| @| @thischapter |
| 992 | @oddheading @thissection @| @| @thispage |
| 993 | @global@pageno = 1 |
| 994 | @end iftex |
| 995 | |
| 996 | @node List Processing |
| 997 | @chapter List Processing |
| 998 | |
| 999 | To the untutored eye, Lisp is a strange programming language. In Lisp |
| 1000 | code there are parentheses everywhere. Some people even claim that |
| 1001 | the name stands for `Lots of Isolated Silly Parentheses'. But the |
| 1002 | claim is unwarranted. Lisp stands for LISt Processing, and the |
| 1003 | programming language handles @emph{lists} (and lists of lists) by |
| 1004 | putting them between parentheses. The parentheses mark the boundaries |
| 1005 | of the list. Sometimes a list is preceded by a single apostrophe or |
| 1006 | quotation mark, @samp{'}@footnote{The single apostrophe or quotation |
| 1007 | mark is an abbreviation for the function @code{quote}; you need not |
| 1008 | think about functions now; functions are defined in @ref{Making |
| 1009 | Errors, , Generate an Error Message}.} Lists are the basis of Lisp. |
| 1010 | |
| 1011 | @menu |
| 1012 | * Lisp Lists:: What are lists? |
| 1013 | * Run a Program:: Any list in Lisp is a program ready to run. |
| 1014 | * Making Errors:: Generating an error message. |
| 1015 | * Names & Definitions:: Names of symbols and function definitions. |
| 1016 | * Lisp Interpreter:: What the Lisp interpreter does. |
| 1017 | * Evaluation:: Running a program. |
| 1018 | * Variables:: Returning a value from a variable. |
| 1019 | * Arguments:: Passing information to a function. |
| 1020 | * set & setq:: Setting the value of a variable. |
| 1021 | * Summary:: The major points. |
| 1022 | * Error Message Exercises:: |
| 1023 | @end menu |
| 1024 | |
| 1025 | @node Lisp Lists |
| 1026 | @section Lisp Lists |
| 1027 | @cindex Lisp Lists |
| 1028 | |
| 1029 | In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}. |
| 1030 | This list is preceded by a single apostrophe. It could just as well be |
| 1031 | written as follows, which looks more like the kind of list you are likely |
| 1032 | to be familiar with: |
| 1033 | |
| 1034 | @smallexample |
| 1035 | @group |
| 1036 | '(rose |
| 1037 | violet |
| 1038 | daisy |
| 1039 | buttercup) |
| 1040 | @end group |
| 1041 | @end smallexample |
| 1042 | |
| 1043 | @noindent |
| 1044 | The elements of this list are the names of the four different flowers, |
| 1045 | separated from each other by whitespace and surrounded by parentheses, |
| 1046 | like flowers in a field with a stone wall around them. |
| 1047 | @cindex Flowers in a field |
| 1048 | |
| 1049 | @menu |
| 1050 | * Numbers Lists:: List have numbers, other lists, in them. |
| 1051 | * Lisp Atoms:: Elemental entities. |
| 1052 | * Whitespace in Lists:: Formatting lists to be readable. |
| 1053 | * Typing Lists:: How GNU Emacs helps you type lists. |
| 1054 | @end menu |
| 1055 | |
| 1056 | @ifnottex |
| 1057 | @node Numbers Lists |
| 1058 | @unnumberedsubsec Numbers, Lists inside of Lists |
| 1059 | @end ifnottex |
| 1060 | |
| 1061 | Lists can also have numbers in them, as in this list: @code{(+ 2 2)}. |
| 1062 | This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each |
| 1063 | separated by whitespace. |
| 1064 | |
| 1065 | In Lisp, both data and programs are represented the same way; that is, |
| 1066 | they are both lists of words, numbers, or other lists, separated by |
| 1067 | whitespace and surrounded by parentheses. (Since a program looks like |
| 1068 | data, one program may easily serve as data for another; this is a very |
| 1069 | powerful feature of Lisp.) (Incidentally, these two parenthetical |
| 1070 | remarks are @emph{not} Lisp lists, because they contain @samp{;} and |
| 1071 | @samp{.} as punctuation marks.) |
| 1072 | |
| 1073 | @need 1200 |
| 1074 | Here is another list, this time with a list inside of it: |
| 1075 | |
| 1076 | @smallexample |
| 1077 | '(this list has (a list inside of it)) |
| 1078 | @end smallexample |
| 1079 | |
| 1080 | The components of this list are the words @samp{this}, @samp{list}, |
| 1081 | @samp{has}, and the list @samp{(a list inside of it)}. The interior |
| 1082 | list is made up of the words @samp{a}, @samp{list}, @samp{inside}, |
| 1083 | @samp{of}, @samp{it}. |
| 1084 | |
| 1085 | @node Lisp Atoms |
| 1086 | @subsection Lisp Atoms |
| 1087 | @cindex Lisp Atoms |
| 1088 | |
| 1089 | In Lisp, what we have been calling words are called @dfn{atoms}. This |
| 1090 | term comes from the historical meaning of the word atom, which means |
| 1091 | `indivisible'. As far as Lisp is concerned, the words we have been |
| 1092 | using in the lists cannot be divided into any smaller parts and still |
| 1093 | mean the same thing as part of a program; likewise with numbers and |
| 1094 | single character symbols like @samp{+}. On the other hand, unlike an |
| 1095 | ancient atom, a list can be split into parts. (@xref{car cdr & cons, |
| 1096 | , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.) |
| 1097 | |
| 1098 | In a list, atoms are separated from each other by whitespace. They can be |
| 1099 | right next to a parenthesis. |
| 1100 | |
| 1101 | @cindex @samp{empty list} defined |
| 1102 | Technically speaking, a list in Lisp consists of parentheses surrounding |
| 1103 | atoms separated by whitespace or surrounding other lists or surrounding |
| 1104 | both atoms and other lists. A list can have just one atom in it or |
| 1105 | have nothing in it at all. A list with nothing in it looks like this: |
| 1106 | @code{()}, and is called the @dfn{empty list}. Unlike anything else, an |
| 1107 | empty list is considered both an atom and a list at the same time. |
| 1108 | |
| 1109 | @cindex Symbolic expressions, introduced |
| 1110 | @cindex @samp{expression} defined |
| 1111 | @cindex @samp{form} defined |
| 1112 | The printed representation of both atoms and lists are called |
| 1113 | @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}. |
| 1114 | The word @dfn{expression} by itself can refer to either the printed |
| 1115 | representation, or to the atom or list as it is held internally in the |
| 1116 | computer. Often, people use the term @dfn{expression} |
| 1117 | indiscriminately. (Also, in many texts, the word @dfn{form} is used |
| 1118 | as a synonym for expression.) |
| 1119 | |
| 1120 | Incidentally, the atoms that make up our universe were named such when |
| 1121 | they were thought to be indivisible; but it has been found that physical |
| 1122 | atoms are not indivisible. Parts can split off an atom or it can |
| 1123 | fission into two parts of roughly equal size. Physical atoms were named |
| 1124 | prematurely, before their truer nature was found. In Lisp, certain |
| 1125 | kinds of atom, such as an array, can be separated into parts; but the |
| 1126 | mechanism for doing this is different from the mechanism for splitting a |
| 1127 | list. As far as list operations are concerned, the atoms of a list are |
| 1128 | unsplittable. |
| 1129 | |
| 1130 | As in English, the meanings of the component letters of a Lisp atom |
| 1131 | are different from the meaning the letters make as a word. For |
| 1132 | example, the word for the South American sloth, the @samp{ai}, is |
| 1133 | completely different from the two words, @samp{a}, and @samp{i}. |
| 1134 | |
| 1135 | There are many kinds of atom in nature but only a few in Lisp: for |
| 1136 | example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such |
| 1137 | as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have |
| 1138 | listed in the examples above are all symbols. In everyday Lisp |
| 1139 | conversation, the word ``atom'' is not often used, because programmers |
| 1140 | usually try to be more specific about what kind of atom they are dealing |
| 1141 | with. Lisp programming is mostly about symbols (and sometimes numbers) |
| 1142 | within lists. (Incidentally, the preceding three word parenthetical |
| 1143 | remark is a proper list in Lisp, since it consists of atoms, which in |
| 1144 | this case are symbols, separated by whitespace and enclosed by |
| 1145 | parentheses, without any non-Lisp punctuation.) |
| 1146 | |
| 1147 | @need 1250 |
| 1148 | Text between double quotation marks---even sentences or |
| 1149 | paragraphs---is also an atom. Here is an example: |
| 1150 | @cindex Text between double quotation marks |
| 1151 | |
| 1152 | @smallexample |
| 1153 | '(this list includes "text between quotation marks.") |
| 1154 | @end smallexample |
| 1155 | |
| 1156 | @cindex @samp{string} defined |
| 1157 | @noindent |
| 1158 | In Lisp, all of the quoted text including the punctuation mark and the |
| 1159 | blank spaces is a single atom. This kind of atom is called a |
| 1160 | @dfn{string} (for `string of characters') and is the sort of thing that |
| 1161 | is used for messages that a computer can print for a human to read. |
| 1162 | Strings are a different kind of atom than numbers or symbols and are |
| 1163 | used differently. |
| 1164 | |
| 1165 | @node Whitespace in Lists |
| 1166 | @subsection Whitespace in Lists |
| 1167 | @cindex Whitespace in lists |
| 1168 | |
| 1169 | @need 1200 |
| 1170 | The amount of whitespace in a list does not matter. From the point of view |
| 1171 | of the Lisp language, |
| 1172 | |
| 1173 | @smallexample |
| 1174 | @group |
| 1175 | '(this list |
| 1176 | looks like this) |
| 1177 | @end group |
| 1178 | @end smallexample |
| 1179 | |
| 1180 | @need 800 |
| 1181 | @noindent |
| 1182 | is exactly the same as this: |
| 1183 | |
| 1184 | @smallexample |
| 1185 | '(this list looks like this) |
| 1186 | @end smallexample |
| 1187 | |
| 1188 | Both examples show what to Lisp is the same list, the list made up of |
| 1189 | the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and |
| 1190 | @samp{this} in that order. |
| 1191 | |
| 1192 | Extra whitespace and newlines are designed to make a list more readable |
| 1193 | by humans. When Lisp reads the expression, it gets rid of all the extra |
| 1194 | whitespace (but it needs to have at least one space between atoms in |
| 1195 | order to tell them apart.) |
| 1196 | |
| 1197 | Odd as it seems, the examples we have seen cover almost all of what Lisp |
| 1198 | lists look like! Every other list in Lisp looks more or less like one |
| 1199 | of these examples, except that the list may be longer and more complex. |
| 1200 | In brief, a list is between parentheses, a string is between quotation |
| 1201 | marks, a symbol looks like a word, and a number looks like a number. |
| 1202 | (For certain situations, square brackets, dots and a few other special |
| 1203 | characters may be used; however, we will go quite far without them.) |
| 1204 | |
| 1205 | @node Typing Lists |
| 1206 | @subsection GNU Emacs Helps You Type Lists |
| 1207 | @cindex Help typing lists |
| 1208 | @cindex Formatting help |
| 1209 | |
| 1210 | When you type a Lisp expression in GNU Emacs using either Lisp |
| 1211 | Interaction mode or Emacs Lisp mode, you have available to you several |
| 1212 | commands to format the Lisp expression so it is easy to read. For |
| 1213 | example, pressing the @key{TAB} key automatically indents the line the |
| 1214 | cursor is on by the right amount. A command to properly indent the |
| 1215 | code in a region is customarily bound to @kbd{M-C-\}. Indentation is |
| 1216 | designed so that you can see which elements of a list belong to which |
| 1217 | list---elements of a sub-list are indented more than the elements of |
| 1218 | the enclosing list. |
| 1219 | |
| 1220 | In addition, when you type a closing parenthesis, Emacs momentarily |
| 1221 | jumps the cursor back to the matching opening parenthesis, so you can |
| 1222 | see which one it is. This is very useful, since every list you type |
| 1223 | in Lisp must have its closing parenthesis match its opening |
| 1224 | parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs |
| 1225 | Manual}, for more information about Emacs's modes.) |
| 1226 | |
| 1227 | @node Run a Program |
| 1228 | @section Run a Program |
| 1229 | @cindex Run a program |
| 1230 | @cindex Program, running one |
| 1231 | |
| 1232 | @cindex @samp{evaluate} defined |
| 1233 | A list in Lisp---any list---is a program ready to run. If you run it |
| 1234 | (for which the Lisp jargon is @dfn{evaluate}), the computer will do one |
| 1235 | of three things: do nothing except return to you the list itself; send |
| 1236 | you an error message; or, treat the first symbol in the list as a |
| 1237 | command to do something. (Usually, of course, it is the last of these |
| 1238 | three things that you really want!) |
| 1239 | |
| 1240 | @c use code for the single apostrophe, not samp. |
| 1241 | The single apostrophe, @code{'}, that I put in front of some of the |
| 1242 | example lists in preceding sections is called a @dfn{quote}; when it |
| 1243 | precedes a list, it tells Lisp to do nothing with the list, other than |
| 1244 | take it as it is written. But if there is no quote preceding a list, |
| 1245 | the first item of the list is special: it is a command for the computer |
| 1246 | to obey. (In Lisp, these commands are called @emph{functions}.) The list |
| 1247 | @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp |
| 1248 | understands that the @code{+} is an instruction to do something with the |
| 1249 | rest of the list: add the numbers that follow. |
| 1250 | |
| 1251 | @need 1250 |
| 1252 | If you are reading this inside of GNU Emacs in Info, here is how you can |
| 1253 | evaluate such a list: place your cursor immediately after the right |
| 1254 | hand parenthesis of the following list and then type @kbd{C-x C-e}: |
| 1255 | |
| 1256 | @smallexample |
| 1257 | (+ 2 2) |
| 1258 | @end smallexample |
| 1259 | |
| 1260 | @c use code for the number four, not samp. |
| 1261 | @noindent |
| 1262 | You will see the number @code{4} appear in the echo area. (In the |
| 1263 | jargon, what you have just done is ``evaluate the list.'' The echo area |
| 1264 | is the line at the bottom of the screen that displays or ``echoes'' |
| 1265 | text.) Now try the same thing with a quoted list: place the cursor |
| 1266 | right after the following list and type @kbd{C-x C-e}: |
| 1267 | |
| 1268 | @smallexample |
| 1269 | '(this is a quoted list) |
| 1270 | @end smallexample |
| 1271 | |
| 1272 | @noindent |
| 1273 | You will see @code{(this is a quoted list)} appear in the echo area. |
| 1274 | |
| 1275 | @cindex Lisp interpreter, explained |
| 1276 | @cindex Interpreter, Lisp, explained |
| 1277 | In both cases, what you are doing is giving a command to the program |
| 1278 | inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the |
| 1279 | interpreter a command to evaluate the expression. The name of the Lisp |
| 1280 | interpreter comes from the word for the task done by a human who comes |
| 1281 | up with the meaning of an expression---who ``interprets'' it. |
| 1282 | |
| 1283 | You can also evaluate an atom that is not part of a list---one that is |
| 1284 | not surrounded by parentheses; again, the Lisp interpreter translates |
| 1285 | from the humanly readable expression to the language of the computer. |
| 1286 | But before discussing this (@pxref{Variables}), we will discuss what the |
| 1287 | Lisp interpreter does when you make an error. |
| 1288 | |
| 1289 | @node Making Errors |
| 1290 | @section Generate an Error Message |
| 1291 | @cindex Generate an error message |
| 1292 | @cindex Error message generation |
| 1293 | |
| 1294 | Partly so you won't worry if you do it accidentally, we will now give |
| 1295 | a command to the Lisp interpreter that generates an error message. |
| 1296 | This is a harmless activity; and indeed, we will often try to generate |
| 1297 | error messages intentionally. Once you understand the jargon, error |
| 1298 | messages can be informative. Instead of being called ``error'' |
| 1299 | messages, they should be called ``help'' messages. They are like |
| 1300 | signposts to a traveler in a strange country; deciphering them can be |
| 1301 | hard, but once understood, they can point the way. |
| 1302 | |
| 1303 | The error message is generated by a built-in GNU Emacs debugger. We |
| 1304 | will `enter the debugger'. You get out of the debugger by typing @code{q}. |
| 1305 | |
| 1306 | What we will do is evaluate a list that is not quoted and does not |
| 1307 | have a meaningful command as its first element. Here is a list almost |
| 1308 | exactly the same as the one we just used, but without the single-quote |
| 1309 | in front of it. Position the cursor right after it and type @kbd{C-x |
| 1310 | C-e}: |
| 1311 | |
| 1312 | @smallexample |
| 1313 | (this is an unquoted list) |
| 1314 | @end smallexample |
| 1315 | |
| 1316 | @ignore |
| 1317 | @noindent |
| 1318 | What you see depends on which version of Emacs you are running. GNU |
| 1319 | Emacs version 22 provides more information than version 20 and before. |
| 1320 | First, the more recent result of generating an error; then the |
| 1321 | earlier, version 20 result. |
| 1322 | |
| 1323 | @need 1250 |
| 1324 | @noindent |
| 1325 | In GNU Emacs version 22, a @file{*Backtrace*} window will open up and |
| 1326 | you will see the following in it: |
| 1327 | @end ignore |
| 1328 | |
| 1329 | A @file{*Backtrace*} window will open up and you should see the |
| 1330 | following in it: |
| 1331 | |
| 1332 | @smallexample |
| 1333 | @group |
| 1334 | ---------- Buffer: *Backtrace* ---------- |
| 1335 | Debugger entered--Lisp error: (void-function this) |
| 1336 | (this is an unquoted list) |
| 1337 | eval((this is an unquoted list)) |
| 1338 | eval-last-sexp-1(nil) |
| 1339 | eval-last-sexp(nil) |
| 1340 | call-interactively(eval-last-sexp) |
| 1341 | ---------- Buffer: *Backtrace* ---------- |
| 1342 | @end group |
| 1343 | @end smallexample |
| 1344 | |
| 1345 | @need 1200 |
| 1346 | @noindent |
| 1347 | Your cursor will be in this window (you may have to wait a few seconds |
| 1348 | before it becomes visible). To quit the debugger and make the |
| 1349 | debugger window go away, type: |
| 1350 | |
| 1351 | @smallexample |
| 1352 | q |
| 1353 | @end smallexample |
| 1354 | |
| 1355 | @noindent |
| 1356 | Please type @kbd{q} right now, so you become confident that you can |
| 1357 | get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter |
| 1358 | it. |
| 1359 | |
| 1360 | @cindex @samp{function} defined |
| 1361 | Based on what we already know, we can almost read this error message. |
| 1362 | |
| 1363 | You read the @file{*Backtrace*} buffer from the bottom up; it tells |
| 1364 | you what Emacs did. When you typed @kbd{C-x C-e}, you made an |
| 1365 | interactive call to the command @code{eval-last-sexp}. @code{eval} is |
| 1366 | an abbreviation for `evaluate' and @code{sexp} is an abbreviation for |
| 1367 | `symbolic expression'. The command means `evaluate last symbolic |
| 1368 | expression', which is the expression just before your cursor. |
| 1369 | |
| 1370 | Each line above tells you what the Lisp interpreter evaluated next. |
| 1371 | The most recent action is at the top. The buffer is called the |
| 1372 | @file{*Backtrace*} buffer because it enables you to track Emacs |
| 1373 | backwards. |
| 1374 | |
| 1375 | @need 800 |
| 1376 | At the top of the @file{*Backtrace*} buffer, you see the line: |
| 1377 | |
| 1378 | @smallexample |
| 1379 | Debugger entered--Lisp error: (void-function this) |
| 1380 | @end smallexample |
| 1381 | |
| 1382 | @noindent |
| 1383 | The Lisp interpreter tried to evaluate the first atom of the list, the |
| 1384 | word @samp{this}. It is this action that generated the error message |
| 1385 | @samp{void-function this}. |
| 1386 | |
| 1387 | The message contains the words @samp{void-function} and @samp{this}. |
| 1388 | |
| 1389 | @cindex @samp{function} defined |
| 1390 | The word @samp{function} was mentioned once before. It is a very |
| 1391 | important word. For our purposes, we can define it by saying that a |
| 1392 | @dfn{function} is a set of instructions to the computer that tell the |
| 1393 | computer to do something. |
| 1394 | |
| 1395 | Now we can begin to understand the error message: @samp{void-function |
| 1396 | this}. The function (that is, the word @samp{this}) does not have a |
| 1397 | definition of any set of instructions for the computer to carry out. |
| 1398 | |
| 1399 | The slightly odd word, @samp{void-function}, is designed to cover the |
| 1400 | way Emacs Lisp is implemented, which is that when a symbol does not |
| 1401 | have a function definition attached to it, the place that should |
| 1402 | contain the instructions is `void'. |
| 1403 | |
| 1404 | On the other hand, since we were able to add 2 plus 2 successfully, by |
| 1405 | evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must |
| 1406 | have a set of instructions for the computer to obey and those |
| 1407 | instructions must be to add the numbers that follow the @code{+}. |
| 1408 | |
| 1409 | It is possible to prevent Emacs entering the debugger in cases like |
| 1410 | this. We do not explain how to do that here, but we will mention what |
| 1411 | the result looks like, because you may encounter a similar situation |
| 1412 | if there is a bug in some Emacs code that you are using. In such |
| 1413 | cases, you will see only one line of error message; it will appear in |
| 1414 | the echo area and look like this: |
| 1415 | |
| 1416 | @smallexample |
| 1417 | Symbol's function definition is void:@: this |
| 1418 | @end smallexample |
| 1419 | |
| 1420 | @noindent |
| 1421 | @ignore |
| 1422 | (Also, your terminal may beep at you---some do, some don't; and others |
| 1423 | blink. This is just a device to get your attention.) |
| 1424 | @end ignore |
| 1425 | The message goes away as soon as you type a key, even just to |
| 1426 | move the cursor. |
| 1427 | |
| 1428 | We know the meaning of the word @samp{Symbol}. It refers to the first |
| 1429 | atom of the list, the word @samp{this}. The word @samp{function} |
| 1430 | refers to the instructions that tell the computer what to do. |
| 1431 | (Technically, the symbol tells the computer where to find the |
| 1432 | instructions, but this is a complication we can ignore for the |
| 1433 | moment.) |
| 1434 | |
| 1435 | The error message can be understood: @samp{Symbol's function |
| 1436 | definition is void:@: this}. The symbol (that is, the word |
| 1437 | @samp{this}) lacks instructions for the computer to carry out. |
| 1438 | |
| 1439 | @node Names & Definitions |
| 1440 | @section Symbol Names and Function Definitions |
| 1441 | @cindex Symbol names |
| 1442 | |
| 1443 | We can articulate another characteristic of Lisp based on what we have |
| 1444 | discussed so far---an important characteristic: a symbol, like |
| 1445 | @code{+}, is not itself the set of instructions for the computer to |
| 1446 | carry out. Instead, the symbol is used, perhaps temporarily, as a way |
| 1447 | of locating the definition or set of instructions. What we see is the |
| 1448 | name through which the instructions can be found. Names of people |
| 1449 | work the same way. I can be referred to as @samp{Bob}; however, I am |
| 1450 | not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the |
| 1451 | consciousness consistently associated with a particular life-form. |
| 1452 | The name is not me, but it can be used to refer to me. |
| 1453 | |
| 1454 | In Lisp, one set of instructions can be attached to several names. |
| 1455 | For example, the computer instructions for adding numbers can be |
| 1456 | linked to the symbol @code{plus} as well as to the symbol @code{+} |
| 1457 | (and are in some dialects of Lisp). Among humans, I can be referred |
| 1458 | to as @samp{Robert} as well as @samp{Bob} and by other words as well. |
| 1459 | |
| 1460 | On the other hand, a symbol can have only one function definition |
| 1461 | attached to it at a time. Otherwise, the computer would be confused as |
| 1462 | to which definition to use. If this were the case among people, only |
| 1463 | one person in the world could be named @samp{Bob}. However, the function |
| 1464 | definition to which the name refers can be changed readily. |
| 1465 | (@xref{Install, , Install a Function Definition}.) |
| 1466 | |
| 1467 | Since Emacs Lisp is large, it is customary to name symbols in a way |
| 1468 | that identifies the part of Emacs to which the function belongs. |
| 1469 | Thus, all the names for functions that deal with Texinfo start with |
| 1470 | @samp{texinfo-} and those for functions that deal with reading mail |
| 1471 | start with @samp{rmail-}. |
| 1472 | |
| 1473 | @node Lisp Interpreter |
| 1474 | @section The Lisp Interpreter |
| 1475 | @cindex Lisp interpreter, what it does |
| 1476 | @cindex Interpreter, what it does |
| 1477 | |
| 1478 | Based on what we have seen, we can now start to figure out what the |
| 1479 | Lisp interpreter does when we command it to evaluate a list. |
| 1480 | First, it looks to see whether there is a quote before the list; if |
| 1481 | there is, the interpreter just gives us the list. On the other |
| 1482 | hand, if there is no quote, the interpreter looks at the first element |
| 1483 | in the list and sees whether it has a function definition. If it does, |
| 1484 | the interpreter carries out the instructions in the function definition. |
| 1485 | Otherwise, the interpreter prints an error message. |
| 1486 | |
| 1487 | This is how Lisp works. Simple. There are added complications which we |
| 1488 | will get to in a minute, but these are the fundamentals. Of course, to |
| 1489 | write Lisp programs, you need to know how to write function definitions |
| 1490 | and attach them to names, and how to do this without confusing either |
| 1491 | yourself or the computer. |
| 1492 | |
| 1493 | @menu |
| 1494 | * Complications:: Variables, Special forms, Lists within. |
| 1495 | * Byte Compiling:: Specially processing code for speed. |
| 1496 | @end menu |
| 1497 | |
| 1498 | @ifnottex |
| 1499 | @node Complications |
| 1500 | @unnumberedsubsec Complications |
| 1501 | @end ifnottex |
| 1502 | |
| 1503 | Now, for the first complication. In addition to lists, the Lisp |
| 1504 | interpreter can evaluate a symbol that is not quoted and does not have |
| 1505 | parentheses around it. The Lisp interpreter will attempt to determine |
| 1506 | the symbol's value as a @dfn{variable}. This situation is described |
| 1507 | in the section on variables. (@xref{Variables}.) |
| 1508 | |
| 1509 | @cindex Special form |
| 1510 | The second complication occurs because some functions are unusual and |
| 1511 | do not work in the usual manner. Those that don't are called |
| 1512 | @dfn{special forms}. They are used for special jobs, like defining a |
| 1513 | function, and there are not many of them. In the next few chapters, |
| 1514 | you will be introduced to several of the more important special forms. |
| 1515 | |
| 1516 | As well as special forms, there are also @dfn{macros}. A macro |
| 1517 | is a construct defined in Lisp, which differs from a function in that it |
| 1518 | translates a Lisp expression into another expression that is to be |
| 1519 | evaluated in place of the original expression. (@xref{Lisp macro}.) |
| 1520 | |
| 1521 | For the purposes of this introduction, you do not need to worry too much |
| 1522 | about whether something is a special form, macro, or ordinary function. |
| 1523 | For example, @code{if} is a special form (@pxref{if}), but @code{when} |
| 1524 | is a macro (@pxref{Lisp macro}). In earlier versions of Emacs, |
| 1525 | @code{defun} was a special form, but now it is a macro (@pxref{defun}). |
| 1526 | It still behaves in the same way. |
| 1527 | |
| 1528 | The final complication is this: if the function that the |
| 1529 | Lisp interpreter is looking at is not a special form, and if it is part |
| 1530 | of a list, the Lisp interpreter looks to see whether the list has a list |
| 1531 | inside of it. If there is an inner list, the Lisp interpreter first |
| 1532 | figures out what it should do with the inside list, and then it works on |
| 1533 | the outside list. If there is yet another list embedded inside the |
| 1534 | inner list, it works on that one first, and so on. It always works on |
| 1535 | the innermost list first. The interpreter works on the innermost list |
| 1536 | first, to evaluate the result of that list. The result may be |
| 1537 | used by the enclosing expression. |
| 1538 | |
| 1539 | Otherwise, the interpreter works left to right, from one expression to |
| 1540 | the next. |
| 1541 | |
| 1542 | @node Byte Compiling |
| 1543 | @subsection Byte Compiling |
| 1544 | @cindex Byte compiling |
| 1545 | |
| 1546 | One other aspect of interpreting: the Lisp interpreter is able to |
| 1547 | interpret two kinds of entity: humanly readable code, on which we will |
| 1548 | focus exclusively, and specially processed code, called @dfn{byte |
| 1549 | compiled} code, which is not humanly readable. Byte compiled code |
| 1550 | runs faster than humanly readable code. |
| 1551 | |
| 1552 | You can transform humanly readable code into byte compiled code by |
| 1553 | running one of the compile commands such as @code{byte-compile-file}. |
| 1554 | Byte compiled code is usually stored in a file that ends with a |
| 1555 | @file{.elc} extension rather than a @file{.el} extension. You will |
| 1556 | see both kinds of file in the @file{emacs/lisp} directory; the files |
| 1557 | to read are those with @file{.el} extensions. |
| 1558 | |
| 1559 | As a practical matter, for most things you might do to customize or |
| 1560 | extend Emacs, you do not need to byte compile; and I will not discuss |
| 1561 | the topic here. @xref{Byte Compilation, , Byte Compilation, elisp, |
| 1562 | The GNU Emacs Lisp Reference Manual}, for a full description of byte |
| 1563 | compilation. |
| 1564 | |
| 1565 | @node Evaluation |
| 1566 | @section Evaluation |
| 1567 | @cindex Evaluation |
| 1568 | |
| 1569 | When the Lisp interpreter works on an expression, the term for the |
| 1570 | activity is called @dfn{evaluation}. We say that the interpreter |
| 1571 | `evaluates the expression'. I've used this term several times before. |
| 1572 | The word comes from its use in everyday language, `to ascertain the |
| 1573 | value or amount of; to appraise', according to @cite{Webster's New |
| 1574 | Collegiate Dictionary}. |
| 1575 | |
| 1576 | @menu |
| 1577 | * How the Interpreter Acts:: Returns and Side Effects... |
| 1578 | * Evaluating Inner Lists:: Lists within lists... |
| 1579 | @end menu |
| 1580 | |
| 1581 | @ifnottex |
| 1582 | @node How the Interpreter Acts |
| 1583 | @unnumberedsubsec How the Lisp Interpreter Acts |
| 1584 | @end ifnottex |
| 1585 | |
| 1586 | @cindex @samp{returned value} explained |
| 1587 | After evaluating an expression, the Lisp interpreter will most likely |
| 1588 | @dfn{return} the value that the computer produces by carrying out the |
| 1589 | instructions it found in the function definition, or perhaps it will |
| 1590 | give up on that function and produce an error message. (The interpreter |
| 1591 | may also find itself tossed, so to speak, to a different function or it |
| 1592 | may attempt to repeat continually what it is doing for ever and ever in |
| 1593 | what is called an `infinite loop'. These actions are less common; and |
| 1594 | we can ignore them.) Most frequently, the interpreter returns a value. |
| 1595 | |
| 1596 | @cindex @samp{side effect} defined |
| 1597 | At the same time the interpreter returns a value, it may do something |
| 1598 | else as well, such as move a cursor or copy a file; this other kind of |
| 1599 | action is called a @dfn{side effect}. Actions that we humans think are |
| 1600 | important, such as printing results, are often ``side effects'' to the |
| 1601 | Lisp interpreter. The jargon can sound peculiar, but it turns out that |
| 1602 | it is fairly easy to learn to use side effects. |
| 1603 | |
| 1604 | In summary, evaluating a symbolic expression most commonly causes the |
| 1605 | Lisp interpreter to return a value and perhaps carry out a side effect; |
| 1606 | or else produce an error. |
| 1607 | |
| 1608 | @node Evaluating Inner Lists |
| 1609 | @subsection Evaluating Inner Lists |
| 1610 | @cindex Inner list evaluation |
| 1611 | @cindex Evaluating inner lists |
| 1612 | |
| 1613 | If evaluation applies to a list that is inside another list, the outer |
| 1614 | list may use the value returned by the first evaluation as information |
| 1615 | when the outer list is evaluated. This explains why inner expressions |
| 1616 | are evaluated first: the values they return are used by the outer |
| 1617 | expressions. |
| 1618 | |
| 1619 | @need 1250 |
| 1620 | We can investigate this process by evaluating another addition example. |
| 1621 | Place your cursor after the following expression and type @kbd{C-x C-e}: |
| 1622 | |
| 1623 | @smallexample |
| 1624 | (+ 2 (+ 3 3)) |
| 1625 | @end smallexample |
| 1626 | |
| 1627 | @noindent |
| 1628 | The number 8 will appear in the echo area. |
| 1629 | |
| 1630 | What happens is that the Lisp interpreter first evaluates the inner |
| 1631 | expression, @code{(+ 3 3)}, for which the value 6 is returned; then it |
| 1632 | evaluates the outer expression as if it were written @code{(+ 2 6)}, which |
| 1633 | returns the value 8. Since there are no more enclosing expressions to |
| 1634 | evaluate, the interpreter prints that value in the echo area. |
| 1635 | |
| 1636 | Now it is easy to understand the name of the command invoked by the |
| 1637 | keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The |
| 1638 | letters @code{sexp} are an abbreviation for `symbolic expression', and |
| 1639 | @code{eval} is an abbreviation for `evaluate'. The command means |
| 1640 | `evaluate last symbolic expression'. |
| 1641 | |
| 1642 | As an experiment, you can try evaluating the expression by putting the |
| 1643 | cursor at the beginning of the next line immediately following the |
| 1644 | expression, or inside the expression. |
| 1645 | |
| 1646 | @need 800 |
| 1647 | Here is another copy of the expression: |
| 1648 | |
| 1649 | @smallexample |
| 1650 | (+ 2 (+ 3 3)) |
| 1651 | @end smallexample |
| 1652 | |
| 1653 | @noindent |
| 1654 | If you place the cursor at the beginning of the blank line that |
| 1655 | immediately follows the expression and type @kbd{C-x C-e}, you will |
| 1656 | still get the value 8 printed in the echo area. Now try putting the |
| 1657 | cursor inside the expression. If you put it right after the next to |
| 1658 | last parenthesis (so it appears to sit on top of the last parenthesis), |
| 1659 | you will get a 6 printed in the echo area! This is because the command |
| 1660 | evaluates the expression @code{(+ 3 3)}. |
| 1661 | |
| 1662 | Now put the cursor immediately after a number. Type @kbd{C-x C-e} and |
| 1663 | you will get the number itself. In Lisp, if you evaluate a number, you |
| 1664 | get the number itself---this is how numbers differ from symbols. If you |
| 1665 | evaluate a list starting with a symbol like @code{+}, you will get a |
| 1666 | value returned that is the result of the computer carrying out the |
| 1667 | instructions in the function definition attached to that name. If a |
| 1668 | symbol by itself is evaluated, something different happens, as we will |
| 1669 | see in the next section. |
| 1670 | |
| 1671 | @node Variables |
| 1672 | @section Variables |
| 1673 | @cindex Variables |
| 1674 | |
| 1675 | In Emacs Lisp, a symbol can have a value attached to it just as it can |
| 1676 | have a function definition attached to it. The two are different. |
| 1677 | The function definition is a set of instructions that a computer will |
| 1678 | obey. A value, on the other hand, is something, such as number or a |
| 1679 | name, that can vary (which is why such a symbol is called a variable). |
| 1680 | The value of a symbol can be any expression in Lisp, such as a symbol, |
| 1681 | number, list, or string. A symbol that has a value is often called a |
| 1682 | @dfn{variable}. |
| 1683 | |
| 1684 | A symbol can have both a function definition and a value attached to |
| 1685 | it at the same time. Or it can have just one or the other. |
| 1686 | The two are separate. This is somewhat similar |
| 1687 | to the way the name Cambridge can refer to the city in Massachusetts |
| 1688 | and have some information attached to the name as well, such as |
| 1689 | ``great programming center''. |
| 1690 | |
| 1691 | @ignore |
| 1692 | (Incidentally, in Emacs Lisp, a symbol can have two |
| 1693 | other things attached to it, too: a property list and a documentation |
| 1694 | string; these are discussed later.) |
| 1695 | @end ignore |
| 1696 | |
| 1697 | Another way to think about this is to imagine a symbol as being a chest |
| 1698 | of drawers. The function definition is put in one drawer, the value in |
| 1699 | another, and so on. What is put in the drawer holding the value can be |
| 1700 | changed without affecting the contents of the drawer holding the |
| 1701 | function definition, and vice-verse. |
| 1702 | |
| 1703 | @menu |
| 1704 | * fill-column Example:: |
| 1705 | * Void Function:: The error message for a symbol |
| 1706 | without a function. |
| 1707 | * Void Variable:: The error message for a symbol without a value. |
| 1708 | @end menu |
| 1709 | |
| 1710 | @ifnottex |
| 1711 | @node fill-column Example |
| 1712 | @unnumberedsubsec @code{fill-column}, an Example Variable |
| 1713 | @end ifnottex |
| 1714 | |
| 1715 | @findex fill-column, @r{an example variable} |
| 1716 | @cindex Example variable, @code{fill-column} |
| 1717 | @cindex Variable, example of, @code{fill-column} |
| 1718 | The variable @code{fill-column} illustrates a symbol with a value |
| 1719 | attached to it: in every GNU Emacs buffer, this symbol is set to some |
| 1720 | value, usually 72 or 70, but sometimes to some other value. To find the |
| 1721 | value of this symbol, evaluate it by itself. If you are reading this in |
| 1722 | Info inside of GNU Emacs, you can do this by putting the cursor after |
| 1723 | the symbol and typing @kbd{C-x C-e}: |
| 1724 | |
| 1725 | @smallexample |
| 1726 | fill-column |
| 1727 | @end smallexample |
| 1728 | |
| 1729 | @noindent |
| 1730 | After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo |
| 1731 | area. This is the value for which @code{fill-column} is set for me as I |
| 1732 | write this. It may be different for you in your Info buffer. Notice |
| 1733 | that the value returned as a variable is printed in exactly the same way |
| 1734 | as the value returned by a function carrying out its instructions. From |
| 1735 | the point of view of the Lisp interpreter, a value returned is a value |
| 1736 | returned. What kind of expression it came from ceases to matter once |
| 1737 | the value is known. |
| 1738 | |
| 1739 | A symbol can have any value attached to it or, to use the jargon, we can |
| 1740 | @dfn{bind} the variable to a value: to a number, such as 72; to a |
| 1741 | string, @code{"such as this"}; to a list, such as @code{(spruce pine |
| 1742 | oak)}; we can even bind a variable to a function definition. |
| 1743 | |
| 1744 | A symbol can be bound to a value in several ways. @xref{set & setq, , |
| 1745 | Setting the Value of a Variable}, for information about one way to do |
| 1746 | this. |
| 1747 | |
| 1748 | @node Void Function |
| 1749 | @subsection Error Message for a Symbol Without a Function |
| 1750 | @cindex Symbol without function error |
| 1751 | @cindex Error for symbol without function |
| 1752 | |
| 1753 | When we evaluated @code{fill-column} to find its value as a variable, |
| 1754 | we did not place parentheses around the word. This is because we did |
| 1755 | not intend to use it as a function name. |
| 1756 | |
| 1757 | If @code{fill-column} were the first or only element of a list, the |
| 1758 | Lisp interpreter would attempt to find the function definition |
| 1759 | attached to it. But @code{fill-column} has no function definition. |
| 1760 | Try evaluating this: |
| 1761 | |
| 1762 | @smallexample |
| 1763 | (fill-column) |
| 1764 | @end smallexample |
| 1765 | |
| 1766 | @need 1250 |
| 1767 | @noindent |
| 1768 | You will create a @file{*Backtrace*} buffer that says: |
| 1769 | |
| 1770 | @smallexample |
| 1771 | @group |
| 1772 | ---------- Buffer: *Backtrace* ---------- |
| 1773 | Debugger entered--Lisp error: (void-function fill-column) |
| 1774 | (fill-column) |
| 1775 | eval((fill-column)) |
| 1776 | eval-last-sexp-1(nil) |
| 1777 | eval-last-sexp(nil) |
| 1778 | call-interactively(eval-last-sexp) |
| 1779 | ---------- Buffer: *Backtrace* ---------- |
| 1780 | @end group |
| 1781 | @end smallexample |
| 1782 | |
| 1783 | @noindent |
| 1784 | (Remember, to quit the debugger and make the debugger window go away, |
| 1785 | type @kbd{q} in the @file{*Backtrace*} buffer.) |
| 1786 | |
| 1787 | @ignore |
| 1788 | @need 800 |
| 1789 | In GNU Emacs 20 and before, you will produce an error message that says: |
| 1790 | |
| 1791 | @smallexample |
| 1792 | Symbol's function definition is void:@: fill-column |
| 1793 | @end smallexample |
| 1794 | |
| 1795 | @noindent |
| 1796 | (The message will go away as soon as you move the cursor or type |
| 1797 | another key.) |
| 1798 | @end ignore |
| 1799 | |
| 1800 | @node Void Variable |
| 1801 | @subsection Error Message for a Symbol Without a Value |
| 1802 | @cindex Symbol without value error |
| 1803 | @cindex Error for symbol without value |
| 1804 | |
| 1805 | If you attempt to evaluate a symbol that does not have a value bound to |
| 1806 | it, you will receive an error message. You can see this by |
| 1807 | experimenting with our 2 plus 2 addition. In the following expression, |
| 1808 | put your cursor right after the @code{+}, before the first number 2, |
| 1809 | type @kbd{C-x C-e}: |
| 1810 | |
| 1811 | @smallexample |
| 1812 | (+ 2 2) |
| 1813 | @end smallexample |
| 1814 | |
| 1815 | @need 1500 |
| 1816 | @noindent |
| 1817 | In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that |
| 1818 | says: |
| 1819 | |
| 1820 | @smallexample |
| 1821 | @group |
| 1822 | ---------- Buffer: *Backtrace* ---------- |
| 1823 | Debugger entered--Lisp error: (void-variable +) |
| 1824 | eval(+) |
| 1825 | eval-last-sexp-1(nil) |
| 1826 | eval-last-sexp(nil) |
| 1827 | call-interactively(eval-last-sexp) |
| 1828 | ---------- Buffer: *Backtrace* ---------- |
| 1829 | @end group |
| 1830 | @end smallexample |
| 1831 | |
| 1832 | @noindent |
| 1833 | (Again, you can quit the debugger by |
| 1834 | typing @kbd{q} in the @file{*Backtrace*} buffer.) |
| 1835 | |
| 1836 | This backtrace is different from the very first error message we saw, |
| 1837 | which said, @samp{Debugger entered--Lisp error: (void-function this)}. |
| 1838 | In this case, the function does not have a value as a variable; while |
| 1839 | in the other error message, the function (the word `this') did not |
| 1840 | have a definition. |
| 1841 | |
| 1842 | In this experiment with the @code{+}, what we did was cause the Lisp |
| 1843 | interpreter to evaluate the @code{+} and look for the value of the |
| 1844 | variable instead of the function definition. We did this by placing the |
| 1845 | cursor right after the symbol rather than after the parenthesis of the |
| 1846 | enclosing list as we did before. As a consequence, the Lisp interpreter |
| 1847 | evaluated the preceding s-expression, which in this case was |
| 1848 | @code{+} by itself. |
| 1849 | |
| 1850 | Since @code{+} does not have a value bound to it, just the function |
| 1851 | definition, the error message reported that the symbol's value as a |
| 1852 | variable was void. |
| 1853 | |
| 1854 | @ignore |
| 1855 | @need 800 |
| 1856 | In GNU Emacs version 20 and before, your error message will say: |
| 1857 | |
| 1858 | @example |
| 1859 | Symbol's value as variable is void:@: + |
| 1860 | @end example |
| 1861 | |
| 1862 | @noindent |
| 1863 | The meaning is the same as in GNU Emacs 22. |
| 1864 | @end ignore |
| 1865 | |
| 1866 | @node Arguments |
| 1867 | @section Arguments |
| 1868 | @cindex Arguments |
| 1869 | @cindex Passing information to functions |
| 1870 | |
| 1871 | To see how information is passed to functions, let's look again at |
| 1872 | our old standby, the addition of two plus two. In Lisp, this is written |
| 1873 | as follows: |
| 1874 | |
| 1875 | @smallexample |
| 1876 | (+ 2 2) |
| 1877 | @end smallexample |
| 1878 | |
| 1879 | If you evaluate this expression, the number 4 will appear in your echo |
| 1880 | area. What the Lisp interpreter does is add the numbers that follow |
| 1881 | the @code{+}. |
| 1882 | |
| 1883 | @cindex @samp{argument} defined |
| 1884 | The numbers added by @code{+} are called the @dfn{arguments} of the |
| 1885 | function @code{+}. These numbers are the information that is given to |
| 1886 | or @dfn{passed} to the function. |
| 1887 | |
| 1888 | The word `argument' comes from the way it is used in mathematics and |
| 1889 | does not refer to a disputation between two people; instead it refers to |
| 1890 | the information presented to the function, in this case, to the |
| 1891 | @code{+}. In Lisp, the arguments to a function are the atoms or lists |
| 1892 | that follow the function. The values returned by the evaluation of |
| 1893 | these atoms or lists are passed to the function. Different functions |
| 1894 | require different numbers of arguments; some functions require none at |
| 1895 | all.@footnote{It is curious to track the path by which the word `argument' |
| 1896 | came to have two different meanings, one in mathematics and the other in |
| 1897 | everyday English. According to the @cite{Oxford English Dictionary}, |
| 1898 | the word derives from the Latin for @samp{to make clear, prove}; thus it |
| 1899 | came to mean, by one thread of derivation, `the evidence offered as |
| 1900 | proof', which is to say, `the information offered', which led to its |
| 1901 | meaning in Lisp. But in the other thread of derivation, it came to mean |
| 1902 | `to assert in a manner against which others may make counter |
| 1903 | assertions', which led to the meaning of the word as a disputation. |
| 1904 | (Note here that the English word has two different definitions attached |
| 1905 | to it at the same time. By contrast, in Emacs Lisp, a symbol cannot |
| 1906 | have two different function definitions at the same time.)} |
| 1907 | |
| 1908 | @menu |
| 1909 | * Data types:: Types of data passed to a function. |
| 1910 | * Args as Variable or List:: An argument can be the value |
| 1911 | of a variable or list. |
| 1912 | * Variable Number of Arguments:: Some functions may take a |
| 1913 | variable number of arguments. |
| 1914 | * Wrong Type of Argument:: Passing an argument of the wrong type |
| 1915 | to a function. |
| 1916 | * message:: A useful function for sending messages. |
| 1917 | @end menu |
| 1918 | |
| 1919 | @node Data types |
| 1920 | @subsection Arguments' Data Types |
| 1921 | @cindex Data types |
| 1922 | @cindex Types of data |
| 1923 | @cindex Arguments' data types |
| 1924 | |
| 1925 | The type of data that should be passed to a function depends on what |
| 1926 | kind of information it uses. The arguments to a function such as |
| 1927 | @code{+} must have values that are numbers, since @code{+} adds numbers. |
| 1928 | Other functions use different kinds of data for their arguments. |
| 1929 | |
| 1930 | @need 1250 |
| 1931 | @findex concat |
| 1932 | For example, the @code{concat} function links together or unites two or |
| 1933 | more strings of text to produce a string. The arguments are strings. |
| 1934 | Concatenating the two character strings @code{abc}, @code{def} produces |
| 1935 | the single string @code{abcdef}. This can be seen by evaluating the |
| 1936 | following: |
| 1937 | |
| 1938 | @smallexample |
| 1939 | (concat "abc" "def") |
| 1940 | @end smallexample |
| 1941 | |
| 1942 | @noindent |
| 1943 | The value produced by evaluating this expression is @code{"abcdef"}. |
| 1944 | |
| 1945 | A function such as @code{substring} uses both a string and numbers as |
| 1946 | arguments. The function returns a part of the string, a substring of |
| 1947 | the first argument. This function takes three arguments. Its first |
| 1948 | argument is the string of characters, the second and third arguments are |
| 1949 | numbers that indicate the beginning and end of the substring. The |
| 1950 | numbers are a count of the number of characters (including spaces and |
| 1951 | punctuation) from the beginning of the string. |
| 1952 | |
| 1953 | @need 800 |
| 1954 | For example, if you evaluate the following: |
| 1955 | |
| 1956 | @smallexample |
| 1957 | (substring "The quick brown fox jumped." 16 19) |
| 1958 | @end smallexample |
| 1959 | |
| 1960 | @noindent |
| 1961 | you will see @code{"fox"} appear in the echo area. The arguments are the |
| 1962 | string and the two numbers. |
| 1963 | |
| 1964 | Note that the string passed to @code{substring} is a single atom even |
| 1965 | though it is made up of several words separated by spaces. Lisp counts |
| 1966 | everything between the two quotation marks as part of the string, |
| 1967 | including the spaces. You can think of the @code{substring} function as |
| 1968 | a kind of `atom smasher' since it takes an otherwise indivisible atom |
| 1969 | and extracts a part. However, @code{substring} is only able to extract |
| 1970 | a substring from an argument that is a string, not from another type of |
| 1971 | atom such as a number or symbol. |
| 1972 | |
| 1973 | @node Args as Variable or List |
| 1974 | @subsection An Argument as the Value of a Variable or List |
| 1975 | |
| 1976 | An argument can be a symbol that returns a value when it is evaluated. |
| 1977 | For example, when the symbol @code{fill-column} by itself is evaluated, |
| 1978 | it returns a number. This number can be used in an addition. |
| 1979 | |
| 1980 | @need 1250 |
| 1981 | Position the cursor after the following expression and type @kbd{C-x |
| 1982 | C-e}: |
| 1983 | |
| 1984 | @smallexample |
| 1985 | (+ 2 fill-column) |
| 1986 | @end smallexample |
| 1987 | |
| 1988 | @noindent |
| 1989 | The value will be a number two more than what you get by evaluating |
| 1990 | @code{fill-column} alone. For me, this is 74, because my value of |
| 1991 | @code{fill-column} is 72. |
| 1992 | |
| 1993 | As we have just seen, an argument can be a symbol that returns a value |
| 1994 | when evaluated. In addition, an argument can be a list that returns a |
| 1995 | value when it is evaluated. For example, in the following expression, |
| 1996 | the arguments to the function @code{concat} are the strings |
| 1997 | @w{@code{"The "}} and @w{@code{" red foxes."}} and the list |
| 1998 | @code{(number-to-string (+ 2 fill-column))}. |
| 1999 | |
| 2000 | @c For GNU Emacs 22, need number-to-string |
| 2001 | @smallexample |
| 2002 | (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.") |
| 2003 | @end smallexample |
| 2004 | |
| 2005 | @noindent |
| 2006 | If you evaluate this expression---and if, as with my Emacs, |
| 2007 | @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will |
| 2008 | appear in the echo area. (Note that you must put spaces after the |
| 2009 | word @samp{The} and before the word @samp{red} so they will appear in |
| 2010 | the final string. The function @code{number-to-string} converts the |
| 2011 | integer that the addition function returns to a string. |
| 2012 | @code{number-to-string} is also known as @code{int-to-string}.) |
| 2013 | |
| 2014 | @node Variable Number of Arguments |
| 2015 | @subsection Variable Number of Arguments |
| 2016 | @cindex Variable number of arguments |
| 2017 | @cindex Arguments, variable number of |
| 2018 | |
| 2019 | Some functions, such as @code{concat}, @code{+} or @code{*}, take any |
| 2020 | number of arguments. (The @code{*} is the symbol for multiplication.) |
| 2021 | This can be seen by evaluating each of the following expressions in |
| 2022 | the usual way. What you will see in the echo area is printed in this |
| 2023 | text after @samp{@result{}}, which you may read as `evaluates to'. |
| 2024 | |
| 2025 | @need 1250 |
| 2026 | In the first set, the functions have no arguments: |
| 2027 | |
| 2028 | @smallexample |
| 2029 | @group |
| 2030 | (+) @result{} 0 |
| 2031 | |
| 2032 | (*) @result{} 1 |
| 2033 | @end group |
| 2034 | @end smallexample |
| 2035 | |
| 2036 | @need 1250 |
| 2037 | In this set, the functions have one argument each: |
| 2038 | |
| 2039 | @smallexample |
| 2040 | @group |
| 2041 | (+ 3) @result{} 3 |
| 2042 | |
| 2043 | (* 3) @result{} 3 |
| 2044 | @end group |
| 2045 | @end smallexample |
| 2046 | |
| 2047 | @need 1250 |
| 2048 | In this set, the functions have three arguments each: |
| 2049 | |
| 2050 | @smallexample |
| 2051 | @group |
| 2052 | (+ 3 4 5) @result{} 12 |
| 2053 | |
| 2054 | (* 3 4 5) @result{} 60 |
| 2055 | @end group |
| 2056 | @end smallexample |
| 2057 | |
| 2058 | @node Wrong Type of Argument |
| 2059 | @subsection Using the Wrong Type Object as an Argument |
| 2060 | @cindex Wrong type of argument |
| 2061 | @cindex Argument, wrong type of |
| 2062 | |
| 2063 | When a function is passed an argument of the wrong type, the Lisp |
| 2064 | interpreter produces an error message. For example, the @code{+} |
| 2065 | function expects the values of its arguments to be numbers. As an |
| 2066 | experiment we can pass it the quoted symbol @code{hello} instead of a |
| 2067 | number. Position the cursor after the following expression and type |
| 2068 | @kbd{C-x C-e}: |
| 2069 | |
| 2070 | @smallexample |
| 2071 | (+ 2 'hello) |
| 2072 | @end smallexample |
| 2073 | |
| 2074 | @noindent |
| 2075 | When you do this you will generate an error message. What has happened |
| 2076 | is that @code{+} has tried to add the 2 to the value returned by |
| 2077 | @code{'hello}, but the value returned by @code{'hello} is the symbol |
| 2078 | @code{hello}, not a number. Only numbers can be added. So @code{+} |
| 2079 | could not carry out its addition. |
| 2080 | |
| 2081 | @need 1250 |
| 2082 | You will create and enter a @file{*Backtrace*} buffer that says: |
| 2083 | |
| 2084 | @noindent |
| 2085 | @smallexample |
| 2086 | @group |
| 2087 | ---------- Buffer: *Backtrace* ---------- |
| 2088 | Debugger entered--Lisp error: |
| 2089 | (wrong-type-argument number-or-marker-p hello) |
| 2090 | +(2 hello) |
| 2091 | eval((+ 2 (quote hello))) |
| 2092 | eval-last-sexp-1(nil) |
| 2093 | eval-last-sexp(nil) |
| 2094 | call-interactively(eval-last-sexp) |
| 2095 | ---------- Buffer: *Backtrace* ---------- |
| 2096 | @end group |
| 2097 | @end smallexample |
| 2098 | |
| 2099 | @need 1250 |
| 2100 | As usual, the error message tries to be helpful and makes sense after you |
| 2101 | learn how to read it.@footnote{@code{(quote hello)} is an expansion of |
| 2102 | the abbreviation @code{'hello}.} |
| 2103 | |
| 2104 | The first part of the error message is straightforward; it says |
| 2105 | @samp{wrong type argument}. Next comes the mysterious jargon word |
| 2106 | @w{@samp{number-or-marker-p}}. This word is trying to tell you what |
| 2107 | kind of argument the @code{+} expected. |
| 2108 | |
| 2109 | The symbol @code{number-or-marker-p} says that the Lisp interpreter is |
| 2110 | trying to determine whether the information presented it (the value of |
| 2111 | the argument) is a number or a marker (a special object representing a |
| 2112 | buffer position). What it does is test to see whether the @code{+} is |
| 2113 | being given numbers to add. It also tests to see whether the |
| 2114 | argument is something called a marker, which is a specific feature of |
| 2115 | Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers. |
| 2116 | When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command, |
| 2117 | its position is kept as a marker. The mark can be considered a |
| 2118 | number---the number of characters the location is from the beginning |
| 2119 | of the buffer.) In Emacs Lisp, @code{+} can be used to add the |
| 2120 | numeric value of marker positions as numbers. |
| 2121 | |
| 2122 | The @samp{p} of @code{number-or-marker-p} is the embodiment of a |
| 2123 | practice started in the early days of Lisp programming. The @samp{p} |
| 2124 | stands for `predicate'. In the jargon used by the early Lisp |
| 2125 | researchers, a predicate refers to a function to determine whether some |
| 2126 | property is true or false. So the @samp{p} tells us that |
| 2127 | @code{number-or-marker-p} is the name of a function that determines |
| 2128 | whether it is true or false that the argument supplied is a number or |
| 2129 | a marker. Other Lisp symbols that end in @samp{p} include @code{zerop}, |
| 2130 | a function that tests whether its argument has the value of zero, and |
| 2131 | @code{listp}, a function that tests whether its argument is a list. |
| 2132 | |
| 2133 | Finally, the last part of the error message is the symbol @code{hello}. |
| 2134 | This is the value of the argument that was passed to @code{+}. If the |
| 2135 | addition had been passed the correct type of object, the value passed |
| 2136 | would have been a number, such as 37, rather than a symbol like |
| 2137 | @code{hello}. But then you would not have got the error message. |
| 2138 | |
| 2139 | @ignore |
| 2140 | @need 1250 |
| 2141 | In GNU Emacs version 20 and before, the echo area displays an error |
| 2142 | message that says: |
| 2143 | |
| 2144 | @smallexample |
| 2145 | Wrong type argument:@: number-or-marker-p, hello |
| 2146 | @end smallexample |
| 2147 | |
| 2148 | This says, in different words, the same as the top line of the |
| 2149 | @file{*Backtrace*} buffer. |
| 2150 | @end ignore |
| 2151 | |
| 2152 | @node message |
| 2153 | @subsection The @code{message} Function |
| 2154 | @findex message |
| 2155 | |
| 2156 | Like @code{+}, the @code{message} function takes a variable number of |
| 2157 | arguments. It is used to send messages to the user and is so useful |
| 2158 | that we will describe it here. |
| 2159 | |
| 2160 | @need 1250 |
| 2161 | A message is printed in the echo area. For example, you can print a |
| 2162 | message in your echo area by evaluating the following list: |
| 2163 | |
| 2164 | @smallexample |
| 2165 | (message "This message appears in the echo area!") |
| 2166 | @end smallexample |
| 2167 | |
| 2168 | The whole string between double quotation marks is a single argument |
| 2169 | and is printed @i{in toto}. (Note that in this example, the message |
| 2170 | itself will appear in the echo area within double quotes; that is |
| 2171 | because you see the value returned by the @code{message} function. In |
| 2172 | most uses of @code{message} in programs that you write, the text will |
| 2173 | be printed in the echo area as a side-effect, without the quotes. |
| 2174 | @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in |
| 2175 | detail}, for an example of this.) |
| 2176 | |
| 2177 | However, if there is a @samp{%s} in the quoted string of characters, the |
| 2178 | @code{message} function does not print the @samp{%s} as such, but looks |
| 2179 | to the argument that follows the string. It evaluates the second |
| 2180 | argument and prints the value at the location in the string where the |
| 2181 | @samp{%s} is. |
| 2182 | |
| 2183 | @need 1250 |
| 2184 | You can see this by positioning the cursor after the following |
| 2185 | expression and typing @kbd{C-x C-e}: |
| 2186 | |
| 2187 | @smallexample |
| 2188 | (message "The name of this buffer is: %s." (buffer-name)) |
| 2189 | @end smallexample |
| 2190 | |
| 2191 | @noindent |
| 2192 | In Info, @code{"The name of this buffer is: *info*."} will appear in the |
| 2193 | echo area. The function @code{buffer-name} returns the name of the |
| 2194 | buffer as a string, which the @code{message} function inserts in place |
| 2195 | of @code{%s}. |
| 2196 | |
| 2197 | To print a value as an integer, use @samp{%d} in the same way as |
| 2198 | @samp{%s}. For example, to print a message in the echo area that |
| 2199 | states the value of the @code{fill-column}, evaluate the following: |
| 2200 | |
| 2201 | @smallexample |
| 2202 | (message "The value of fill-column is %d." fill-column) |
| 2203 | @end smallexample |
| 2204 | |
| 2205 | @noindent |
| 2206 | On my system, when I evaluate this list, @code{"The value of |
| 2207 | fill-column is 72."} appears in my echo area@footnote{Actually, you |
| 2208 | can use @code{%s} to print a number. It is non-specific. @code{%d} |
| 2209 | prints only the part of a number left of a decimal point, and not |
| 2210 | anything that is not a number.}. |
| 2211 | |
| 2212 | If there is more than one @samp{%s} in the quoted string, the value of |
| 2213 | the first argument following the quoted string is printed at the |
| 2214 | location of the first @samp{%s} and the value of the second argument is |
| 2215 | printed at the location of the second @samp{%s}, and so on. |
| 2216 | |
| 2217 | @need 1250 |
| 2218 | For example, if you evaluate the following, |
| 2219 | |
| 2220 | @smallexample |
| 2221 | @group |
| 2222 | (message "There are %d %s in the office!" |
| 2223 | (- fill-column 14) "pink elephants") |
| 2224 | @end group |
| 2225 | @end smallexample |
| 2226 | |
| 2227 | @noindent |
| 2228 | a rather whimsical message will appear in your echo area. On my system |
| 2229 | it says, @code{"There are 58 pink elephants in the office!"}. |
| 2230 | |
| 2231 | The expression @code{(- fill-column 14)} is evaluated and the resulting |
| 2232 | number is inserted in place of the @samp{%d}; and the string in double |
| 2233 | quotes, @code{"pink elephants"}, is treated as a single argument and |
| 2234 | inserted in place of the @samp{%s}. (That is to say, a string between |
| 2235 | double quotes evaluates to itself, like a number.) |
| 2236 | |
| 2237 | Finally, here is a somewhat complex example that not only illustrates |
| 2238 | the computation of a number, but also shows how you can use an |
| 2239 | expression within an expression to generate the text that is substituted |
| 2240 | for @samp{%s}: |
| 2241 | |
| 2242 | @smallexample |
| 2243 | @group |
| 2244 | (message "He saw %d %s" |
| 2245 | (- fill-column 32) |
| 2246 | (concat "red " |
| 2247 | (substring |
| 2248 | "The quick brown foxes jumped." 16 21) |
| 2249 | " leaping.")) |
| 2250 | @end group |
| 2251 | @end smallexample |
| 2252 | |
| 2253 | In this example, @code{message} has three arguments: the string, |
| 2254 | @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and |
| 2255 | the expression beginning with the function @code{concat}. The value |
| 2256 | resulting from the evaluation of @code{(- fill-column 32)} is inserted |
| 2257 | in place of the @samp{%d}; and the value returned by the expression |
| 2258 | beginning with @code{concat} is inserted in place of the @samp{%s}. |
| 2259 | |
| 2260 | When your fill column is 70 and you evaluate the expression, the |
| 2261 | message @code{"He saw 38 red foxes leaping."} appears in your echo |
| 2262 | area. |
| 2263 | |
| 2264 | @node set & setq |
| 2265 | @section Setting the Value of a Variable |
| 2266 | @cindex Variable, setting value |
| 2267 | @cindex Setting value of variable |
| 2268 | |
| 2269 | @cindex @samp{bind} defined |
| 2270 | There are several ways by which a variable can be given a value. One of |
| 2271 | the ways is to use either the function @code{set} or the function |
| 2272 | @code{setq}. Another way is to use @code{let} (@pxref{let}). (The |
| 2273 | jargon for this process is to @dfn{bind} a variable to a value.) |
| 2274 | |
| 2275 | The following sections not only describe how @code{set} and @code{setq} |
| 2276 | work but also illustrate how arguments are passed. |
| 2277 | |
| 2278 | @menu |
| 2279 | * Using set:: Setting values. |
| 2280 | * Using setq:: Setting a quoted value. |
| 2281 | * Counting:: Using @code{setq} to count. |
| 2282 | @end menu |
| 2283 | |
| 2284 | @node Using set |
| 2285 | @subsection Using @code{set} |
| 2286 | @findex set |
| 2287 | |
| 2288 | To set the value of the symbol @code{flowers} to the list @code{'(rose |
| 2289 | violet daisy buttercup)}, evaluate the following expression by |
| 2290 | positioning the cursor after the expression and typing @kbd{C-x C-e}. |
| 2291 | |
| 2292 | @smallexample |
| 2293 | (set 'flowers '(rose violet daisy buttercup)) |
| 2294 | @end smallexample |
| 2295 | |
| 2296 | @noindent |
| 2297 | The list @code{(rose violet daisy buttercup)} will appear in the echo |
| 2298 | area. This is what is @emph{returned} by the @code{set} function. As a |
| 2299 | side effect, the symbol @code{flowers} is bound to the list; that is, |
| 2300 | the symbol @code{flowers}, which can be viewed as a variable, is given |
| 2301 | the list as its value. (This process, by the way, illustrates how a |
| 2302 | side effect to the Lisp interpreter, setting the value, can be the |
| 2303 | primary effect that we humans are interested in. This is because every |
| 2304 | Lisp function must return a value if it does not get an error, but it |
| 2305 | will only have a side effect if it is designed to have one.) |
| 2306 | |
| 2307 | After evaluating the @code{set} expression, you can evaluate the symbol |
| 2308 | @code{flowers} and it will return the value you just set. Here is the |
| 2309 | symbol. Place your cursor after it and type @kbd{C-x C-e}. |
| 2310 | |
| 2311 | @smallexample |
| 2312 | flowers |
| 2313 | @end smallexample |
| 2314 | |
| 2315 | @noindent |
| 2316 | When you evaluate @code{flowers}, the list |
| 2317 | @code{(rose violet daisy buttercup)} appears in the echo area. |
| 2318 | |
| 2319 | Incidentally, if you evaluate @code{'flowers}, the variable with a quote |
| 2320 | in front of it, what you will see in the echo area is the symbol itself, |
| 2321 | @code{flowers}. Here is the quoted symbol, so you can try this: |
| 2322 | |
| 2323 | @smallexample |
| 2324 | 'flowers |
| 2325 | @end smallexample |
| 2326 | |
| 2327 | Note also, that when you use @code{set}, you need to quote both |
| 2328 | arguments to @code{set}, unless you want them evaluated. Since we do |
| 2329 | not want either argument evaluated, neither the variable |
| 2330 | @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both |
| 2331 | are quoted. (When you use @code{set} without quoting its first |
| 2332 | argument, the first argument is evaluated before anything else is |
| 2333 | done. If you did this and @code{flowers} did not have a value |
| 2334 | already, you would get an error message that the @samp{Symbol's value |
| 2335 | as variable is void}; on the other hand, if @code{flowers} did return |
| 2336 | a value after it was evaluated, the @code{set} would attempt to set |
| 2337 | the value that was returned. There are situations where this is the |
| 2338 | right thing for the function to do; but such situations are rare.) |
| 2339 | |
| 2340 | @node Using setq |
| 2341 | @subsection Using @code{setq} |
| 2342 | @findex setq |
| 2343 | |
| 2344 | As a practical matter, you almost always quote the first argument to |
| 2345 | @code{set}. The combination of @code{set} and a quoted first argument |
| 2346 | is so common that it has its own name: the special form @code{setq}. |
| 2347 | This special form is just like @code{set} except that the first argument |
| 2348 | is quoted automatically, so you don't need to type the quote mark |
| 2349 | yourself. Also, as an added convenience, @code{setq} permits you to set |
| 2350 | several different variables to different values, all in one expression. |
| 2351 | |
| 2352 | To set the value of the variable @code{carnivores} to the list |
| 2353 | @code{'(lion tiger leopard)} using @code{setq}, the following expression |
| 2354 | is used: |
| 2355 | |
| 2356 | @smallexample |
| 2357 | (setq carnivores '(lion tiger leopard)) |
| 2358 | @end smallexample |
| 2359 | |
| 2360 | @noindent |
| 2361 | This is exactly the same as using @code{set} except the first argument |
| 2362 | is automatically quoted by @code{setq}. (The @samp{q} in @code{setq} |
| 2363 | means @code{quote}.) |
| 2364 | |
| 2365 | @need 1250 |
| 2366 | With @code{set}, the expression would look like this: |
| 2367 | |
| 2368 | @smallexample |
| 2369 | (set 'carnivores '(lion tiger leopard)) |
| 2370 | @end smallexample |
| 2371 | |
| 2372 | Also, @code{setq} can be used to assign different values to |
| 2373 | different variables. The first argument is bound to the value |
| 2374 | of the second argument, the third argument is bound to the value of the |
| 2375 | fourth argument, and so on. For example, you could use the following to |
| 2376 | assign a list of trees to the symbol @code{trees} and a list of herbivores |
| 2377 | to the symbol @code{herbivores}: |
| 2378 | |
| 2379 | @smallexample |
| 2380 | @group |
| 2381 | (setq trees '(pine fir oak maple) |
| 2382 | herbivores '(gazelle antelope zebra)) |
| 2383 | @end group |
| 2384 | @end smallexample |
| 2385 | |
| 2386 | @noindent |
| 2387 | (The expression could just as well have been on one line, but it might |
| 2388 | not have fit on a page; and humans find it easier to read nicely |
| 2389 | formatted lists.) |
| 2390 | |
| 2391 | Although I have been using the term `assign', there is another way of |
| 2392 | thinking about the workings of @code{set} and @code{setq}; and that is to |
| 2393 | say that @code{set} and @code{setq} make the symbol @emph{point} to the |
| 2394 | list. This latter way of thinking is very common and in forthcoming |
| 2395 | chapters we shall come upon at least one symbol that has `pointer' as |
| 2396 | part of its name. The name is chosen because the symbol has a value, |
| 2397 | specifically a list, attached to it; or, expressed another way, |
| 2398 | the symbol is set to ``point'' to the list. |
| 2399 | |
| 2400 | @node Counting |
| 2401 | @subsection Counting |
| 2402 | @cindex Counting |
| 2403 | |
| 2404 | Here is an example that shows how to use @code{setq} in a counter. You |
| 2405 | might use this to count how many times a part of your program repeats |
| 2406 | itself. First set a variable to zero; then add one to the number each |
| 2407 | time the program repeats itself. To do this, you need a variable that |
| 2408 | serves as a counter, and two expressions: an initial @code{setq} |
| 2409 | expression that sets the counter variable to zero; and a second |
| 2410 | @code{setq} expression that increments the counter each time it is |
| 2411 | evaluated. |
| 2412 | |
| 2413 | @smallexample |
| 2414 | @group |
| 2415 | (setq counter 0) ; @r{Let's call this the initializer.} |
| 2416 | |
| 2417 | (setq counter (+ counter 1)) ; @r{This is the incrementer.} |
| 2418 | |
| 2419 | counter ; @r{This is the counter.} |
| 2420 | @end group |
| 2421 | @end smallexample |
| 2422 | |
| 2423 | @noindent |
| 2424 | (The text following the @samp{;} are comments. @xref{Change a |
| 2425 | defun, , Change a Function Definition}.) |
| 2426 | |
| 2427 | If you evaluate the first of these expressions, the initializer, |
| 2428 | @code{(setq counter 0)}, and then evaluate the third expression, |
| 2429 | @code{counter}, the number @code{0} will appear in the echo area. If |
| 2430 | you then evaluate the second expression, the incrementer, @code{(setq |
| 2431 | counter (+ counter 1))}, the counter will get the value 1. So if you |
| 2432 | again evaluate @code{counter}, the number @code{1} will appear in the |
| 2433 | echo area. Each time you evaluate the second expression, the value of |
| 2434 | the counter will be incremented. |
| 2435 | |
| 2436 | When you evaluate the incrementer, @code{(setq counter (+ counter 1))}, |
| 2437 | the Lisp interpreter first evaluates the innermost list; this is the |
| 2438 | addition. In order to evaluate this list, it must evaluate the variable |
| 2439 | @code{counter} and the number @code{1}. When it evaluates the variable |
| 2440 | @code{counter}, it receives its current value. It passes this value and |
| 2441 | the number @code{1} to the @code{+} which adds them together. The sum |
| 2442 | is then returned as the value of the inner list and passed to the |
| 2443 | @code{setq} which sets the variable @code{counter} to this new value. |
| 2444 | Thus, the value of the variable, @code{counter}, is changed. |
| 2445 | |
| 2446 | @node Summary |
| 2447 | @section Summary |
| 2448 | |
| 2449 | Learning Lisp is like climbing a hill in which the first part is the |
| 2450 | steepest. You have now climbed the most difficult part; what remains |
| 2451 | becomes easier as you progress onwards. |
| 2452 | |
| 2453 | @need 1000 |
| 2454 | In summary, |
| 2455 | |
| 2456 | @itemize @bullet |
| 2457 | |
| 2458 | @item |
| 2459 | Lisp programs are made up of expressions, which are lists or single atoms. |
| 2460 | |
| 2461 | @item |
| 2462 | Lists are made up of zero or more atoms or inner lists, separated by whitespace and |
| 2463 | surrounded by parentheses. A list can be empty. |
| 2464 | |
| 2465 | @item |
| 2466 | Atoms are multi-character symbols, like @code{forward-paragraph}, single |
| 2467 | character symbols like @code{+}, strings of characters between double |
| 2468 | quotation marks, or numbers. |
| 2469 | |
| 2470 | @item |
| 2471 | A number evaluates to itself. |
| 2472 | |
| 2473 | @item |
| 2474 | A string between double quotes also evaluates to itself. |
| 2475 | |
| 2476 | @item |
| 2477 | When you evaluate a symbol by itself, its value is returned. |
| 2478 | |
| 2479 | @item |
| 2480 | When you evaluate a list, the Lisp interpreter looks at the first symbol |
| 2481 | in the list and then at the function definition bound to that symbol. |
| 2482 | Then the instructions in the function definition are carried out. |
| 2483 | |
| 2484 | @item |
| 2485 | A single quotation mark, |
| 2486 | @ifinfo |
| 2487 | ' |
| 2488 | @end ifinfo |
| 2489 | @ifnotinfo |
| 2490 | @code{'} |
| 2491 | @end ifnotinfo |
| 2492 | , tells the Lisp interpreter that it should |
| 2493 | return the following expression as written, and not evaluate it as it |
| 2494 | would if the quote were not there. |
| 2495 | |
| 2496 | @item |
| 2497 | Arguments are the information passed to a function. The arguments to a |
| 2498 | function are computed by evaluating the rest of the elements of the list |
| 2499 | of which the function is the first element. |
| 2500 | |
| 2501 | @item |
| 2502 | A function always returns a value when it is evaluated (unless it gets |
| 2503 | an error); in addition, it may also carry out some action called a |
| 2504 | ``side effect''. In many cases, a function's primary purpose is to |
| 2505 | create a side effect. |
| 2506 | @end itemize |
| 2507 | |
| 2508 | @node Error Message Exercises |
| 2509 | @section Exercises |
| 2510 | |
| 2511 | A few simple exercises: |
| 2512 | |
| 2513 | @itemize @bullet |
| 2514 | @item |
| 2515 | Generate an error message by evaluating an appropriate symbol that is |
| 2516 | not within parentheses. |
| 2517 | |
| 2518 | @item |
| 2519 | Generate an error message by evaluating an appropriate symbol that is |
| 2520 | between parentheses. |
| 2521 | |
| 2522 | @item |
| 2523 | Create a counter that increments by two rather than one. |
| 2524 | |
| 2525 | @item |
| 2526 | Write an expression that prints a message in the echo area when |
| 2527 | evaluated. |
| 2528 | @end itemize |
| 2529 | |
| 2530 | @node Practicing Evaluation |
| 2531 | @chapter Practicing Evaluation |
| 2532 | @cindex Practicing evaluation |
| 2533 | @cindex Evaluation practice |
| 2534 | |
| 2535 | Before learning how to write a function definition in Emacs Lisp, it is |
| 2536 | useful to spend a little time evaluating various expressions that have |
| 2537 | already been written. These expressions will be lists with the |
| 2538 | functions as their first (and often only) element. Since some of the |
| 2539 | functions associated with buffers are both simple and interesting, we |
| 2540 | will start with those. In this section, we will evaluate a few of |
| 2541 | these. In another section, we will study the code of several other |
| 2542 | buffer-related functions, to see how they were written. |
| 2543 | |
| 2544 | @menu |
| 2545 | * How to Evaluate:: Typing editing commands or @kbd{C-x C-e} |
| 2546 | causes evaluation. |
| 2547 | * Buffer Names:: Buffers and files are different. |
| 2548 | * Getting Buffers:: Getting a buffer itself, not merely its name. |
| 2549 | * Switching Buffers:: How to change to another buffer. |
| 2550 | * Buffer Size & Locations:: Where point is located and the size of |
| 2551 | the buffer. |
| 2552 | * Evaluation Exercise:: |
| 2553 | @end menu |
| 2554 | |
| 2555 | @ifnottex |
| 2556 | @node How to Evaluate |
| 2557 | @unnumberedsec How to Evaluate |
| 2558 | @end ifnottex |
| 2559 | |
| 2560 | @i{Whenever you give an editing command} to Emacs Lisp, such as the |
| 2561 | command to move the cursor or to scroll the screen, @i{you are evaluating |
| 2562 | an expression,} the first element of which is a function. @i{This is |
| 2563 | how Emacs works.} |
| 2564 | |
| 2565 | @cindex @samp{interactive function} defined |
| 2566 | @cindex @samp{command} defined |
| 2567 | When you type keys, you cause the Lisp interpreter to evaluate an |
| 2568 | expression and that is how you get your results. Even typing plain text |
| 2569 | involves evaluating an Emacs Lisp function, in this case, one that uses |
| 2570 | @code{self-insert-command}, which simply inserts the character you |
| 2571 | typed. The functions you evaluate by typing keystrokes are called |
| 2572 | @dfn{interactive} functions, or @dfn{commands}; how you make a function |
| 2573 | interactive will be illustrated in the chapter on how to write function |
| 2574 | definitions. @xref{Interactive, , Making a Function Interactive}. |
| 2575 | |
| 2576 | In addition to typing keyboard commands, we have seen a second way to |
| 2577 | evaluate an expression: by positioning the cursor after a list and |
| 2578 | typing @kbd{C-x C-e}. This is what we will do in the rest of this |
| 2579 | section. There are other ways to evaluate an expression as well; these |
| 2580 | will be described as we come to them. |
| 2581 | |
| 2582 | Besides being used for practicing evaluation, the functions shown in the |
| 2583 | next few sections are important in their own right. A study of these |
| 2584 | functions makes clear the distinction between buffers and files, how to |
| 2585 | switch to a buffer, and how to determine a location within it. |
| 2586 | |
| 2587 | @node Buffer Names |
| 2588 | @section Buffer Names |
| 2589 | @findex buffer-name |
| 2590 | @findex buffer-file-name |
| 2591 | |
| 2592 | The two functions, @code{buffer-name} and @code{buffer-file-name}, show |
| 2593 | the difference between a file and a buffer. When you evaluate the |
| 2594 | following expression, @code{(buffer-name)}, the name of the buffer |
| 2595 | appears in the echo area. When you evaluate @code{(buffer-file-name)}, |
| 2596 | the name of the file to which the buffer refers appears in the echo |
| 2597 | area. Usually, the name returned by @code{(buffer-name)} is the same as |
| 2598 | the name of the file to which it refers, and the name returned by |
| 2599 | @code{(buffer-file-name)} is the full path-name of the file. |
| 2600 | |
| 2601 | A file and a buffer are two different entities. A file is information |
| 2602 | recorded permanently in the computer (unless you delete it). A buffer, |
| 2603 | on the other hand, is information inside of Emacs that will vanish at |
| 2604 | the end of the editing session (or when you kill the buffer). Usually, |
| 2605 | a buffer contains information that you have copied from a file; we say |
| 2606 | the buffer is @dfn{visiting} that file. This copy is what you work on |
| 2607 | and modify. Changes to the buffer do not change the file, until you |
| 2608 | save the buffer. When you save the buffer, the buffer is copied to the file |
| 2609 | and is thus saved permanently. |
| 2610 | |
| 2611 | @need 1250 |
| 2612 | If you are reading this in Info inside of GNU Emacs, you can evaluate |
| 2613 | each of the following expressions by positioning the cursor after it and |
| 2614 | typing @kbd{C-x C-e}. |
| 2615 | |
| 2616 | @example |
| 2617 | @group |
| 2618 | (buffer-name) |
| 2619 | |
| 2620 | (buffer-file-name) |
| 2621 | @end group |
| 2622 | @end example |
| 2623 | |
| 2624 | @noindent |
| 2625 | When I do this in Info, the value returned by evaluating |
| 2626 | @code{(buffer-name)} is @file{"*info*"}, and the value returned by |
| 2627 | evaluating @code{(buffer-file-name)} is @file{nil}. |
| 2628 | |
| 2629 | On the other hand, while I am writing this document, the value |
| 2630 | returned by evaluating @code{(buffer-name)} is |
| 2631 | @file{"introduction.texinfo"}, and the value returned by evaluating |
| 2632 | @code{(buffer-file-name)} is |
| 2633 | @file{"/gnu/work/intro/introduction.texinfo"}. |
| 2634 | |
| 2635 | @cindex @code{nil}, history of word |
| 2636 | The former is the name of the buffer and the latter is the name of the |
| 2637 | file. In Info, the buffer name is @file{"*info*"}. Info does not |
| 2638 | point to any file, so the result of evaluating |
| 2639 | @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is |
| 2640 | from the Latin word for `nothing'; in this case, it means that the |
| 2641 | buffer is not associated with any file. (In Lisp, @code{nil} is also |
| 2642 | used to mean `false' and is a synonym for the empty list, @code{()}.) |
| 2643 | |
| 2644 | When I am writing, the name of my buffer is |
| 2645 | @file{"introduction.texinfo"}. The name of the file to which it |
| 2646 | points is @file{"/gnu/work/intro/introduction.texinfo"}. |
| 2647 | |
| 2648 | (In the expressions, the parentheses tell the Lisp interpreter to |
| 2649 | treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as |
| 2650 | functions; without the parentheses, the interpreter would attempt to |
| 2651 | evaluate the symbols as variables. @xref{Variables}.) |
| 2652 | |
| 2653 | In spite of the distinction between files and buffers, you will often |
| 2654 | find that people refer to a file when they mean a buffer and vice-verse. |
| 2655 | Indeed, most people say, ``I am editing a file,'' rather than saying, |
| 2656 | ``I am editing a buffer which I will soon save to a file.'' It is |
| 2657 | almost always clear from context what people mean. When dealing with |
| 2658 | computer programs, however, it is important to keep the distinction in mind, |
| 2659 | since the computer is not as smart as a person. |
| 2660 | |
| 2661 | @cindex Buffer, history of word |
| 2662 | The word `buffer', by the way, comes from the meaning of the word as a |
| 2663 | cushion that deadens the force of a collision. In early computers, a |
| 2664 | buffer cushioned the interaction between files and the computer's |
| 2665 | central processing unit. The drums or tapes that held a file and the |
| 2666 | central processing unit were pieces of equipment that were very |
| 2667 | different from each other, working at their own speeds, in spurts. The |
| 2668 | buffer made it possible for them to work together effectively. |
| 2669 | Eventually, the buffer grew from being an intermediary, a temporary |
| 2670 | holding place, to being the place where work is done. This |
| 2671 | transformation is rather like that of a small seaport that grew into a |
| 2672 | great city: once it was merely the place where cargo was warehoused |
| 2673 | temporarily before being loaded onto ships; then it became a business |
| 2674 | and cultural center in its own right. |
| 2675 | |
| 2676 | Not all buffers are associated with files. For example, a |
| 2677 | @file{*scratch*} buffer does not visit any file. Similarly, a |
| 2678 | @file{*Help*} buffer is not associated with any file. |
| 2679 | |
| 2680 | In the old days, when you lacked a @file{~/.emacs} file and started an |
| 2681 | Emacs session by typing the command @code{emacs} alone, without naming |
| 2682 | any files, Emacs started with the @file{*scratch*} buffer visible. |
| 2683 | Nowadays, you will see a splash screen. You can follow one of the |
| 2684 | commands suggested on the splash screen, visit a file, or press the |
| 2685 | spacebar to reach the @file{*scratch*} buffer. |
| 2686 | |
| 2687 | If you switch to the @file{*scratch*} buffer, type |
| 2688 | @code{(buffer-name)}, position the cursor after it, and then type |
| 2689 | @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"} |
| 2690 | will be returned and will appear in the echo area. @code{"*scratch*"} |
| 2691 | is the name of the buffer. When you type @code{(buffer-file-name)} in |
| 2692 | the @file{*scratch*} buffer and evaluate that, @code{nil} will appear |
| 2693 | in the echo area, just as it does when you evaluate |
| 2694 | @code{(buffer-file-name)} in Info. |
| 2695 | |
| 2696 | Incidentally, if you are in the @file{*scratch*} buffer and want the |
| 2697 | value returned by an expression to appear in the @file{*scratch*} |
| 2698 | buffer itself rather than in the echo area, type @kbd{C-u C-x C-e} |
| 2699 | instead of @kbd{C-x C-e}. This causes the value returned to appear |
| 2700 | after the expression. The buffer will look like this: |
| 2701 | |
| 2702 | @smallexample |
| 2703 | (buffer-name)"*scratch*" |
| 2704 | @end smallexample |
| 2705 | |
| 2706 | @noindent |
| 2707 | You cannot do this in Info since Info is read-only and it will not allow |
| 2708 | you to change the contents of the buffer. But you can do this in any |
| 2709 | buffer you can edit; and when you write code or documentation (such as |
| 2710 | this book), this feature is very useful. |
| 2711 | |
| 2712 | @node Getting Buffers |
| 2713 | @section Getting Buffers |
| 2714 | @findex current-buffer |
| 2715 | @findex other-buffer |
| 2716 | @cindex Getting a buffer |
| 2717 | |
| 2718 | The @code{buffer-name} function returns the @emph{name} of the buffer; |
| 2719 | to get the buffer @emph{itself}, a different function is needed: the |
| 2720 | @code{current-buffer} function. If you use this function in code, what |
| 2721 | you get is the buffer itself. |
| 2722 | |
| 2723 | A name and the object or entity to which the name refers are different |
| 2724 | from each other. You are not your name. You are a person to whom |
| 2725 | others refer by name. If you ask to speak to George and someone hands you |
| 2726 | a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r}, |
| 2727 | @samp{g}, and @samp{e} written on it, you might be amused, but you would |
| 2728 | not be satisfied. You do not want to speak to the name, but to the |
| 2729 | person to whom the name refers. A buffer is similar: the name of the |
| 2730 | scratch buffer is @file{*scratch*}, but the name is not the buffer. To |
| 2731 | get a buffer itself, you need to use a function such as |
| 2732 | @code{current-buffer}. |
| 2733 | |
| 2734 | However, there is a slight complication: if you evaluate |
| 2735 | @code{current-buffer} in an expression on its own, as we will do here, |
| 2736 | what you see is a printed representation of the name of the buffer |
| 2737 | without the contents of the buffer. Emacs works this way for two |
| 2738 | reasons: the buffer may be thousands of lines long---too long to be |
| 2739 | conveniently displayed; and, another buffer may have the same contents |
| 2740 | but a different name, and it is important to distinguish between them. |
| 2741 | |
| 2742 | @need 800 |
| 2743 | Here is an expression containing the function: |
| 2744 | |
| 2745 | @smallexample |
| 2746 | (current-buffer) |
| 2747 | @end smallexample |
| 2748 | |
| 2749 | @noindent |
| 2750 | If you evaluate this expression in Info in Emacs in the usual way, |
| 2751 | @file{#<buffer *info*>} will appear in the echo area. The special |
| 2752 | format indicates that the buffer itself is being returned, rather than |
| 2753 | just its name. |
| 2754 | |
| 2755 | Incidentally, while you can type a number or symbol into a program, you |
| 2756 | cannot do that with the printed representation of a buffer: the only way |
| 2757 | to get a buffer itself is with a function such as @code{current-buffer}. |
| 2758 | |
| 2759 | A related function is @code{other-buffer}. This returns the most |
| 2760 | recently selected buffer other than the one you are in currently, not |
| 2761 | a printed representation of its name. If you have recently switched |
| 2762 | back and forth from the @file{*scratch*} buffer, @code{other-buffer} |
| 2763 | will return that buffer. |
| 2764 | |
| 2765 | @need 800 |
| 2766 | You can see this by evaluating the expression: |
| 2767 | |
| 2768 | @smallexample |
| 2769 | (other-buffer) |
| 2770 | @end smallexample |
| 2771 | |
| 2772 | @noindent |
| 2773 | You should see @file{#<buffer *scratch*>} appear in the echo area, or |
| 2774 | the name of whatever other buffer you switched back from most |
| 2775 | recently@footnote{Actually, by default, if the buffer from which you |
| 2776 | just switched is visible to you in another window, @code{other-buffer} |
| 2777 | will choose the most recent buffer that you cannot see; this is a |
| 2778 | subtlety that I often forget.}. |
| 2779 | |
| 2780 | @node Switching Buffers |
| 2781 | @section Switching Buffers |
| 2782 | @findex switch-to-buffer |
| 2783 | @findex set-buffer |
| 2784 | @cindex Switching to a buffer |
| 2785 | |
| 2786 | The @code{other-buffer} function actually provides a buffer when it is |
| 2787 | used as an argument to a function that requires one. We can see this |
| 2788 | by using @code{other-buffer} and @code{switch-to-buffer} to switch to a |
| 2789 | different buffer. |
| 2790 | |
| 2791 | But first, a brief introduction to the @code{switch-to-buffer} |
| 2792 | function. When you switched back and forth from Info to the |
| 2793 | @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most |
| 2794 | likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or |
| 2795 | rather, to save typing, you probably only typed @kbd{RET} if the |
| 2796 | default buffer was @file{*scratch*}, or if it was different, then you |
| 2797 | typed just part of the name, such as @code{*sc}, pressed your |
| 2798 | @kbd{TAB} key to cause it to expand to the full name, and then typed |
| 2799 | @kbd{RET}.} when prompted in the minibuffer for the name of |
| 2800 | the buffer to which you wanted to switch. The keystrokes, @kbd{C-x |
| 2801 | b}, cause the Lisp interpreter to evaluate the interactive function |
| 2802 | @code{switch-to-buffer}. As we said before, this is how Emacs works: |
| 2803 | different keystrokes call or run different functions. For example, |
| 2804 | @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls |
| 2805 | @code{forward-sentence}, and so on. |
| 2806 | |
| 2807 | By writing @code{switch-to-buffer} in an expression, and giving it a |
| 2808 | buffer to switch to, we can switch buffers just the way @kbd{C-x b} |
| 2809 | does: |
| 2810 | |
| 2811 | @smallexample |
| 2812 | (switch-to-buffer (other-buffer)) |
| 2813 | @end smallexample |
| 2814 | |
| 2815 | @noindent |
| 2816 | The symbol @code{switch-to-buffer} is the first element of the list, |
| 2817 | so the Lisp interpreter will treat it as a function and carry out the |
| 2818 | instructions that are attached to it. But before doing that, the |
| 2819 | interpreter will note that @code{other-buffer} is inside parentheses |
| 2820 | and work on that symbol first. @code{other-buffer} is the first (and |
| 2821 | in this case, the only) element of this list, so the Lisp interpreter |
| 2822 | calls or runs the function. It returns another buffer. Next, the |
| 2823 | interpreter runs @code{switch-to-buffer}, passing to it, as an |
| 2824 | argument, the other buffer, which is what Emacs will switch to. If |
| 2825 | you are reading this in Info, try this now. Evaluate the expression. |
| 2826 | (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this |
| 2827 | expression will move you to your most recent other buffer that you |
| 2828 | cannot see. If you really want to go to your most recently selected |
| 2829 | buffer, even if you can still see it, you need to evaluate the |
| 2830 | following more complex expression: |
| 2831 | |
| 2832 | @smallexample |
| 2833 | (switch-to-buffer (other-buffer (current-buffer) t)) |
| 2834 | @end smallexample |
| 2835 | |
| 2836 | @c noindent |
| 2837 | In this case, the first argument to @code{other-buffer} tells it which |
| 2838 | buffer to skip---the current one---and the second argument tells |
| 2839 | @code{other-buffer} it is OK to switch to a visible buffer. |
| 2840 | In regular use, @code{switch-to-buffer} takes you to an invisible |
| 2841 | window since you would most likely use @kbd{C-x o} (@code{other-window}) |
| 2842 | to go to another visible buffer.} |
| 2843 | |
| 2844 | In the programming examples in later sections of this document, you will |
| 2845 | see the function @code{set-buffer} more often than |
| 2846 | @code{switch-to-buffer}. This is because of a difference between |
| 2847 | computer programs and humans: humans have eyes and expect to see the |
| 2848 | buffer on which they are working on their computer terminals. This is |
| 2849 | so obvious, it almost goes without saying. However, programs do not |
| 2850 | have eyes. When a computer program works on a buffer, that buffer does |
| 2851 | not need to be visible on the screen. |
| 2852 | |
| 2853 | @code{switch-to-buffer} is designed for humans and does two different |
| 2854 | things: it switches the buffer to which Emacs's attention is directed; and |
| 2855 | it switches the buffer displayed in the window to the new buffer. |
| 2856 | @code{set-buffer}, on the other hand, does only one thing: it switches |
| 2857 | the attention of the computer program to a different buffer. The buffer |
| 2858 | on the screen remains unchanged (of course, normally nothing happens |
| 2859 | there until the command finishes running). |
| 2860 | |
| 2861 | @cindex @samp{call} defined |
| 2862 | Also, we have just introduced another jargon term, the word @dfn{call}. |
| 2863 | When you evaluate a list in which the first symbol is a function, you |
| 2864 | are calling that function. The use of the term comes from the notion of |
| 2865 | the function as an entity that can do something for you if you `call' |
| 2866 | it---just as a plumber is an entity who can fix a leak if you call him |
| 2867 | or her. |
| 2868 | |
| 2869 | @node Buffer Size & Locations |
| 2870 | @section Buffer Size and the Location of Point |
| 2871 | @cindex Size of buffer |
| 2872 | @cindex Buffer size |
| 2873 | @cindex Point location |
| 2874 | @cindex Location of point |
| 2875 | |
| 2876 | Finally, let's look at several rather simple functions, |
| 2877 | @code{buffer-size}, @code{point}, @code{point-min}, and |
| 2878 | @code{point-max}. These give information about the size of a buffer and |
| 2879 | the location of point within it. |
| 2880 | |
| 2881 | The function @code{buffer-size} tells you the size of the current |
| 2882 | buffer; that is, the function returns a count of the number of |
| 2883 | characters in the buffer. |
| 2884 | |
| 2885 | @smallexample |
| 2886 | (buffer-size) |
| 2887 | @end smallexample |
| 2888 | |
| 2889 | @noindent |
| 2890 | You can evaluate this in the usual way, by positioning the |
| 2891 | cursor after the expression and typing @kbd{C-x C-e}. |
| 2892 | |
| 2893 | @cindex @samp{point} defined |
| 2894 | In Emacs, the current position of the cursor is called @dfn{point}. |
| 2895 | The expression @code{(point)} returns a number that tells you where the |
| 2896 | cursor is located as a count of the number of characters from the |
| 2897 | beginning of the buffer up to point. |
| 2898 | |
| 2899 | @need 1250 |
| 2900 | You can see the character count for point in this buffer by evaluating |
| 2901 | the following expression in the usual way: |
| 2902 | |
| 2903 | @smallexample |
| 2904 | (point) |
| 2905 | @end smallexample |
| 2906 | |
| 2907 | @noindent |
| 2908 | As I write this, the value of @code{point} is 65724. The @code{point} |
| 2909 | function is frequently used in some of the examples later in this |
| 2910 | book. |
| 2911 | |
| 2912 | @need 1250 |
| 2913 | The value of point depends, of course, on its location within the |
| 2914 | buffer. If you evaluate point in this spot, the number will be larger: |
| 2915 | |
| 2916 | @smallexample |
| 2917 | (point) |
| 2918 | @end smallexample |
| 2919 | |
| 2920 | @noindent |
| 2921 | For me, the value of point in this location is 66043, which means that |
| 2922 | there are 319 characters (including spaces) between the two |
| 2923 | expressions. (Doubtless, you will see different numbers, since I will |
| 2924 | have edited this since I first evaluated point.) |
| 2925 | |
| 2926 | @cindex @samp{narrowing} defined |
| 2927 | The function @code{point-min} is somewhat similar to @code{point}, but |
| 2928 | it returns the value of the minimum permissible value of point in the |
| 2929 | current buffer. This is the number 1 unless @dfn{narrowing} is in |
| 2930 | effect. (Narrowing is a mechanism whereby you can restrict yourself, |
| 2931 | or a program, to operations on just a part of a buffer. |
| 2932 | @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the |
| 2933 | function @code{point-max} returns the value of the maximum permissible |
| 2934 | value of point in the current buffer. |
| 2935 | |
| 2936 | @node Evaluation Exercise |
| 2937 | @section Exercise |
| 2938 | |
| 2939 | Find a file with which you are working and move towards its middle. |
| 2940 | Find its buffer name, file name, length, and your position in the file. |
| 2941 | |
| 2942 | @node Writing Defuns |
| 2943 | @chapter How To Write Function Definitions |
| 2944 | @cindex Definition writing |
| 2945 | @cindex Function definition writing |
| 2946 | @cindex Writing a function definition |
| 2947 | |
| 2948 | When the Lisp interpreter evaluates a list, it looks to see whether the |
| 2949 | first symbol on the list has a function definition attached to it; or, |
| 2950 | put another way, whether the symbol points to a function definition. If |
| 2951 | it does, the computer carries out the instructions in the definition. A |
| 2952 | symbol that has a function definition is called, simply, a function |
| 2953 | (although, properly speaking, the definition is the function and the |
| 2954 | symbol refers to it.) |
| 2955 | |
| 2956 | @menu |
| 2957 | * Primitive Functions:: |
| 2958 | * defun:: The @code{defun} macro. |
| 2959 | * Install:: Install a function definition. |
| 2960 | * Interactive:: Making a function interactive. |
| 2961 | * Interactive Options:: Different options for @code{interactive}. |
| 2962 | * Permanent Installation:: Installing code permanently. |
| 2963 | * let:: Creating and initializing local variables. |
| 2964 | * if:: What if? |
| 2965 | * else:: If--then--else expressions. |
| 2966 | * Truth & Falsehood:: What Lisp considers false and true. |
| 2967 | * save-excursion:: Keeping track of point, mark, and buffer. |
| 2968 | * Review:: |
| 2969 | * defun Exercises:: |
| 2970 | @end menu |
| 2971 | |
| 2972 | @ifnottex |
| 2973 | @node Primitive Functions |
| 2974 | @unnumberedsec An Aside about Primitive Functions |
| 2975 | @end ifnottex |
| 2976 | @cindex Primitive functions |
| 2977 | @cindex Functions, primitive |
| 2978 | |
| 2979 | @cindex C language primitives |
| 2980 | @cindex Primitives written in C |
| 2981 | All functions are defined in terms of other functions, except for a few |
| 2982 | @dfn{primitive} functions that are written in the C programming |
| 2983 | language. When you write functions' definitions, you will write them in |
| 2984 | Emacs Lisp and use other functions as your building blocks. Some of the |
| 2985 | functions you will use will themselves be written in Emacs Lisp (perhaps |
| 2986 | by you) and some will be primitives written in C@. The primitive |
| 2987 | functions are used exactly like those written in Emacs Lisp and behave |
| 2988 | like them. They are written in C so we can easily run GNU Emacs on any |
| 2989 | computer that has sufficient power and can run C. |
| 2990 | |
| 2991 | Let me re-emphasize this: when you write code in Emacs Lisp, you do not |
| 2992 | distinguish between the use of functions written in C and the use of |
| 2993 | functions written in Emacs Lisp. The difference is irrelevant. I |
| 2994 | mention the distinction only because it is interesting to know. Indeed, |
| 2995 | unless you investigate, you won't know whether an already-written |
| 2996 | function is written in Emacs Lisp or C. |
| 2997 | |
| 2998 | @node defun |
| 2999 | @section The @code{defun} Macro |
| 3000 | @findex defun |
| 3001 | |
| 3002 | @cindex @samp{function definition} defined |
| 3003 | In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to |
| 3004 | it that tells the computer what to do when the function is called. |
| 3005 | This code is called the @dfn{function definition} and is created by |
| 3006 | evaluating a Lisp expression that starts with the symbol @code{defun} |
| 3007 | (which is an abbreviation for @emph{define function}). |
| 3008 | |
| 3009 | In subsequent sections, we will look at function definitions from the |
| 3010 | Emacs source code, such as @code{mark-whole-buffer}. In this section, |
| 3011 | we will describe a simple function definition so you can see how it |
| 3012 | looks. This function definition uses arithmetic because it makes for a |
| 3013 | simple example. Some people dislike examples using arithmetic; however, |
| 3014 | if you are such a person, do not despair. Hardly any of the code we |
| 3015 | will study in the remainder of this introduction involves arithmetic or |
| 3016 | mathematics. The examples mostly involve text in one way or another. |
| 3017 | |
| 3018 | A function definition has up to five parts following the word |
| 3019 | @code{defun}: |
| 3020 | |
| 3021 | @enumerate |
| 3022 | @item |
| 3023 | The name of the symbol to which the function definition should be |
| 3024 | attached. |
| 3025 | |
| 3026 | @item |
| 3027 | A list of the arguments that will be passed to the function. If no |
| 3028 | arguments will be passed to the function, this is an empty list, |
| 3029 | @code{()}. |
| 3030 | |
| 3031 | @item |
| 3032 | Documentation describing the function. (Technically optional, but |
| 3033 | strongly recommended.) |
| 3034 | |
| 3035 | @item |
| 3036 | Optionally, an expression to make the function interactive so you can |
| 3037 | use it by typing @kbd{M-x} and then the name of the function; or by |
| 3038 | typing an appropriate key or keychord. |
| 3039 | |
| 3040 | @cindex @samp{body} defined |
| 3041 | @item |
| 3042 | The code that instructs the computer what to do: the @dfn{body} of the |
| 3043 | function definition. |
| 3044 | @end enumerate |
| 3045 | |
| 3046 | It is helpful to think of the five parts of a function definition as |
| 3047 | being organized in a template, with slots for each part: |
| 3048 | |
| 3049 | @smallexample |
| 3050 | @group |
| 3051 | (defun @var{function-name} (@var{arguments}@dots{}) |
| 3052 | "@var{optional-documentation}@dots{}" |
| 3053 | (interactive @var{argument-passing-info}) ; @r{optional} |
| 3054 | @var{body}@dots{}) |
| 3055 | @end group |
| 3056 | @end smallexample |
| 3057 | |
| 3058 | As an example, here is the code for a function that multiplies its |
| 3059 | argument by 7. (This example is not interactive. @xref{Interactive, |
| 3060 | , Making a Function Interactive}, for that information.) |
| 3061 | |
| 3062 | @smallexample |
| 3063 | @group |
| 3064 | (defun multiply-by-seven (number) |
| 3065 | "Multiply NUMBER by seven." |
| 3066 | (* 7 number)) |
| 3067 | @end group |
| 3068 | @end smallexample |
| 3069 | |
| 3070 | This definition begins with a parenthesis and the symbol @code{defun}, |
| 3071 | followed by the name of the function. |
| 3072 | |
| 3073 | @cindex @samp{argument list} defined |
| 3074 | The name of the function is followed by a list that contains the |
| 3075 | arguments that will be passed to the function. This list is called |
| 3076 | the @dfn{argument list}. In this example, the list has only one |
| 3077 | element, the symbol, @code{number}. When the function is used, the |
| 3078 | symbol will be bound to the value that is used as the argument to the |
| 3079 | function. |
| 3080 | |
| 3081 | Instead of choosing the word @code{number} for the name of the argument, |
| 3082 | I could have picked any other name. For example, I could have chosen |
| 3083 | the word @code{multiplicand}. I picked the word `number' because it |
| 3084 | tells what kind of value is intended for this slot; but I could just as |
| 3085 | well have chosen the word `multiplicand' to indicate the role that the |
| 3086 | value placed in this slot will play in the workings of the function. I |
| 3087 | could have called it @code{foogle}, but that would have been a bad |
| 3088 | choice because it would not tell humans what it means. The choice of |
| 3089 | name is up to the programmer and should be chosen to make the meaning of |
| 3090 | the function clear. |
| 3091 | |
| 3092 | Indeed, you can choose any name you wish for a symbol in an argument |
| 3093 | list, even the name of a symbol used in some other function: the name |
| 3094 | you use in an argument list is private to that particular definition. |
| 3095 | In that definition, the name refers to a different entity than any use |
| 3096 | of the same name outside the function definition. Suppose you have a |
| 3097 | nick-name `Shorty' in your family; when your family members refer to |
| 3098 | `Shorty', they mean you. But outside your family, in a movie, for |
| 3099 | example, the name `Shorty' refers to someone else. Because a name in an |
| 3100 | argument list is private to the function definition, you can change the |
| 3101 | value of such a symbol inside the body of a function without changing |
| 3102 | its value outside the function. The effect is similar to that produced |
| 3103 | by a @code{let} expression. (@xref{let, , @code{let}}.) |
| 3104 | |
| 3105 | @ignore |
| 3106 | Note also that we discuss the word `number' in two different ways: as a |
| 3107 | symbol that appears in the code, and as the name of something that will |
| 3108 | be replaced by a something else during the evaluation of the function. |
| 3109 | In the first case, @code{number} is a symbol, not a number; it happens |
| 3110 | that within the function, it is a variable who value is the number in |
| 3111 | question, but our primary interest in it is as a symbol. On the other |
| 3112 | hand, when we are talking about the function, our interest is that we |
| 3113 | will substitute a number for the word @var{number}. To keep this |
| 3114 | distinction clear, we use different typography for the two |
| 3115 | circumstances. When we talk about this function, or about how it works, |
| 3116 | we refer to this number by writing @var{number}. In the function |
| 3117 | itself, we refer to it by writing @code{number}. |
| 3118 | @end ignore |
| 3119 | |
| 3120 | The argument list is followed by the documentation string that |
| 3121 | describes the function. This is what you see when you type |
| 3122 | @w{@kbd{C-h f}} and the name of a function. Incidentally, when you |
| 3123 | write a documentation string like this, you should make the first line |
| 3124 | a complete sentence since some commands, such as @code{apropos}, print |
| 3125 | only the first line of a multi-line documentation string. Also, you |
| 3126 | should not indent the second line of a documentation string, if you |
| 3127 | have one, because that looks odd when you use @kbd{C-h f} |
| 3128 | (@code{describe-function}). The documentation string is optional, but |
| 3129 | it is so useful, it should be included in almost every function you |
| 3130 | write. |
| 3131 | |
| 3132 | @findex * @r{(multiplication)} |
| 3133 | The third line of the example consists of the body of the function |
| 3134 | definition. (Most functions' definitions, of course, are longer than |
| 3135 | this.) In this function, the body is the list, @code{(* 7 number)}, which |
| 3136 | says to multiply the value of @var{number} by 7. (In Emacs Lisp, |
| 3137 | @code{*} is the function for multiplication, just as @code{+} is the |
| 3138 | function for addition.) |
| 3139 | |
| 3140 | When you use the @code{multiply-by-seven} function, the argument |
| 3141 | @code{number} evaluates to the actual number you want used. Here is an |
| 3142 | example that shows how @code{multiply-by-seven} is used; but don't try |
| 3143 | to evaluate this yet! |
| 3144 | |
| 3145 | @smallexample |
| 3146 | (multiply-by-seven 3) |
| 3147 | @end smallexample |
| 3148 | |
| 3149 | @noindent |
| 3150 | The symbol @code{number}, specified in the function definition in the |
| 3151 | next section, is given or ``bound to'' the value 3 in the actual use of |
| 3152 | the function. Note that although @code{number} was inside parentheses |
| 3153 | in the function definition, the argument passed to the |
| 3154 | @code{multiply-by-seven} function is not in parentheses. The |
| 3155 | parentheses are written in the function definition so the computer can |
| 3156 | figure out where the argument list ends and the rest of the function |
| 3157 | definition begins. |
| 3158 | |
| 3159 | If you evaluate this example, you are likely to get an error message. |
| 3160 | (Go ahead, try it!) This is because we have written the function |
| 3161 | definition, but not yet told the computer about the definition---we have |
| 3162 | not yet installed (or `loaded') the function definition in Emacs. |
| 3163 | Installing a function is the process that tells the Lisp interpreter the |
| 3164 | definition of the function. Installation is described in the next |
| 3165 | section. |
| 3166 | |
| 3167 | @node Install |
| 3168 | @section Install a Function Definition |
| 3169 | @cindex Install a Function Definition |
| 3170 | @cindex Definition installation |
| 3171 | @cindex Function definition installation |
| 3172 | |
| 3173 | If you are reading this inside of Info in Emacs, you can try out the |
| 3174 | @code{multiply-by-seven} function by first evaluating the function |
| 3175 | definition and then evaluating @code{(multiply-by-seven 3)}. A copy of |
| 3176 | the function definition follows. Place the cursor after the last |
| 3177 | parenthesis of the function definition and type @kbd{C-x C-e}. When you |
| 3178 | do this, @code{multiply-by-seven} will appear in the echo area. (What |
| 3179 | this means is that when a function definition is evaluated, the value it |
| 3180 | returns is the name of the defined function.) At the same time, this |
| 3181 | action installs the function definition. |
| 3182 | |
| 3183 | @smallexample |
| 3184 | @group |
| 3185 | (defun multiply-by-seven (number) |
| 3186 | "Multiply NUMBER by seven." |
| 3187 | (* 7 number)) |
| 3188 | @end group |
| 3189 | @end smallexample |
| 3190 | |
| 3191 | @noindent |
| 3192 | By evaluating this @code{defun}, you have just installed |
| 3193 | @code{multiply-by-seven} in Emacs. The function is now just as much a |
| 3194 | part of Emacs as @code{forward-word} or any other editing function you |
| 3195 | use. (@code{multiply-by-seven} will stay installed until you quit |
| 3196 | Emacs. To reload code automatically whenever you start Emacs, see |
| 3197 | @ref{Permanent Installation, , Installing Code Permanently}.) |
| 3198 | |
| 3199 | @menu |
| 3200 | * Effect of installation:: |
| 3201 | * Change a defun:: How to change a function definition. |
| 3202 | @end menu |
| 3203 | |
| 3204 | @ifnottex |
| 3205 | @node Effect of installation |
| 3206 | @unnumberedsubsec The effect of installation |
| 3207 | @end ifnottex |
| 3208 | |
| 3209 | You can see the effect of installing @code{multiply-by-seven} by |
| 3210 | evaluating the following sample. Place the cursor after the following |
| 3211 | expression and type @kbd{C-x C-e}. The number 21 will appear in the |
| 3212 | echo area. |
| 3213 | |
| 3214 | @smallexample |
| 3215 | (multiply-by-seven 3) |
| 3216 | @end smallexample |
| 3217 | |
| 3218 | If you wish, you can read the documentation for the function by typing |
| 3219 | @kbd{C-h f} (@code{describe-function}) and then the name of the |
| 3220 | function, @code{multiply-by-seven}. When you do this, a |
| 3221 | @file{*Help*} window will appear on your screen that says: |
| 3222 | |
| 3223 | @smallexample |
| 3224 | @group |
| 3225 | multiply-by-seven is a Lisp function. |
| 3226 | (multiply-by-seven NUMBER) |
| 3227 | |
| 3228 | Multiply NUMBER by seven. |
| 3229 | @end group |
| 3230 | @end smallexample |
| 3231 | |
| 3232 | @noindent |
| 3233 | (To return to a single window on your screen, type @kbd{C-x 1}.) |
| 3234 | |
| 3235 | @node Change a defun |
| 3236 | @subsection Change a Function Definition |
| 3237 | @cindex Changing a function definition |
| 3238 | @cindex Function definition, how to change |
| 3239 | @cindex Definition, how to change |
| 3240 | |
| 3241 | If you want to change the code in @code{multiply-by-seven}, just rewrite |
| 3242 | it. To install the new version in place of the old one, evaluate the |
| 3243 | function definition again. This is how you modify code in Emacs. It is |
| 3244 | very simple. |
| 3245 | |
| 3246 | As an example, you can change the @code{multiply-by-seven} function to |
| 3247 | add the number to itself seven times instead of multiplying the number |
| 3248 | by seven. It produces the same answer, but by a different path. At |
| 3249 | the same time, we will add a comment to the code; a comment is text |
| 3250 | that the Lisp interpreter ignores, but that a human reader may find |
| 3251 | useful or enlightening. The comment is that this is the ``second |
| 3252 | version''. |
| 3253 | |
| 3254 | @smallexample |
| 3255 | @group |
| 3256 | (defun multiply-by-seven (number) ; @r{Second version.} |
| 3257 | "Multiply NUMBER by seven." |
| 3258 | (+ number number number number number number number)) |
| 3259 | @end group |
| 3260 | @end smallexample |
| 3261 | |
| 3262 | @cindex Comments in Lisp code |
| 3263 | The comment follows a semicolon, @samp{;}. In Lisp, everything on a |
| 3264 | line that follows a semicolon is a comment. The end of the line is the |
| 3265 | end of the comment. To stretch a comment over two or more lines, begin |
| 3266 | each line with a semicolon. |
| 3267 | |
| 3268 | @xref{Beginning init File, , Beginning a @file{.emacs} |
| 3269 | File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp |
| 3270 | Reference Manual}, for more about comments. |
| 3271 | |
| 3272 | You can install this version of the @code{multiply-by-seven} function by |
| 3273 | evaluating it in the same way you evaluated the first function: place |
| 3274 | the cursor after the last parenthesis and type @kbd{C-x C-e}. |
| 3275 | |
| 3276 | In summary, this is how you write code in Emacs Lisp: you write a |
| 3277 | function; install it; test it; and then make fixes or enhancements and |
| 3278 | install it again. |
| 3279 | |
| 3280 | @node Interactive |
| 3281 | @section Make a Function Interactive |
| 3282 | @cindex Interactive functions |
| 3283 | @findex interactive |
| 3284 | |
| 3285 | You make a function interactive by placing a list that begins with |
| 3286 | the special form @code{interactive} immediately after the |
| 3287 | documentation. A user can invoke an interactive function by typing |
| 3288 | @kbd{M-x} and then the name of the function; or by typing the keys to |
| 3289 | which it is bound, for example, by typing @kbd{C-n} for |
| 3290 | @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}. |
| 3291 | |
| 3292 | Interestingly, when you call an interactive function interactively, |
| 3293 | the value returned is not automatically displayed in the echo area. |
| 3294 | This is because you often call an interactive function for its side |
| 3295 | effects, such as moving forward by a word or line, and not for the |
| 3296 | value returned. If the returned value were displayed in the echo area |
| 3297 | each time you typed a key, it would be very distracting. |
| 3298 | |
| 3299 | @menu |
| 3300 | * Interactive multiply-by-seven:: An overview. |
| 3301 | * multiply-by-seven in detail:: The interactive version. |
| 3302 | @end menu |
| 3303 | |
| 3304 | @ifnottex |
| 3305 | @node Interactive multiply-by-seven |
| 3306 | @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview |
| 3307 | @end ifnottex |
| 3308 | |
| 3309 | Both the use of the special form @code{interactive} and one way to |
| 3310 | display a value in the echo area can be illustrated by creating an |
| 3311 | interactive version of @code{multiply-by-seven}. |
| 3312 | |
| 3313 | @need 1250 |
| 3314 | Here is the code: |
| 3315 | |
| 3316 | @smallexample |
| 3317 | @group |
| 3318 | (defun multiply-by-seven (number) ; @r{Interactive version.} |
| 3319 | "Multiply NUMBER by seven." |
| 3320 | (interactive "p") |
| 3321 | (message "The result is %d" (* 7 number))) |
| 3322 | @end group |
| 3323 | @end smallexample |
| 3324 | |
| 3325 | @noindent |
| 3326 | You can install this code by placing your cursor after it and typing |
| 3327 | @kbd{C-x C-e}. The name of the function will appear in your echo area. |
| 3328 | Then, you can use this code by typing @kbd{C-u} and a number and then |
| 3329 | typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase |
| 3330 | @samp{The result is @dots{}} followed by the product will appear in the |
| 3331 | echo area. |
| 3332 | |
| 3333 | Speaking more generally, you invoke a function like this in either of two |
| 3334 | ways: |
| 3335 | |
| 3336 | @enumerate |
| 3337 | @item |
| 3338 | By typing a prefix argument that contains the number to be passed, and |
| 3339 | then typing @kbd{M-x} and the name of the function, as with |
| 3340 | @kbd{C-u 3 M-x forward-sentence}; or, |
| 3341 | |
| 3342 | @item |
| 3343 | By typing whatever key or keychord the function is bound to, as with |
| 3344 | @kbd{C-u 3 M-e}. |
| 3345 | @end enumerate |
| 3346 | |
| 3347 | @noindent |
| 3348 | Both the examples just mentioned work identically to move point forward |
| 3349 | three sentences. (Since @code{multiply-by-seven} is not bound to a key, |
| 3350 | it could not be used as an example of key binding.) |
| 3351 | |
| 3352 | (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command |
| 3353 | to a key.) |
| 3354 | |
| 3355 | A prefix argument is passed to an interactive function by typing the |
| 3356 | @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by |
| 3357 | typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you |
| 3358 | type @kbd{C-u} without a number, it defaults to 4). |
| 3359 | |
| 3360 | @node multiply-by-seven in detail |
| 3361 | @subsection An Interactive @code{multiply-by-seven} |
| 3362 | |
| 3363 | Let's look at the use of the special form @code{interactive} and then at |
| 3364 | the function @code{message} in the interactive version of |
| 3365 | @code{multiply-by-seven}. You will recall that the function definition |
| 3366 | looks like this: |
| 3367 | |
| 3368 | @smallexample |
| 3369 | @group |
| 3370 | (defun multiply-by-seven (number) ; @r{Interactive version.} |
| 3371 | "Multiply NUMBER by seven." |
| 3372 | (interactive "p") |
| 3373 | (message "The result is %d" (* 7 number))) |
| 3374 | @end group |
| 3375 | @end smallexample |
| 3376 | |
| 3377 | In this function, the expression, @code{(interactive "p")}, is a list of |
| 3378 | two elements. The @code{"p"} tells Emacs to pass the prefix argument to |
| 3379 | the function and use its value for the argument of the function. |
| 3380 | |
| 3381 | @need 1000 |
| 3382 | The argument will be a number. This means that the symbol |
| 3383 | @code{number} will be bound to a number in the line: |
| 3384 | |
| 3385 | @smallexample |
| 3386 | (message "The result is %d" (* 7 number)) |
| 3387 | @end smallexample |
| 3388 | |
| 3389 | @need 1250 |
| 3390 | @noindent |
| 3391 | For example, if your prefix argument is 5, the Lisp interpreter will |
| 3392 | evaluate the line as if it were: |
| 3393 | |
| 3394 | @smallexample |
| 3395 | (message "The result is %d" (* 7 5)) |
| 3396 | @end smallexample |
| 3397 | |
| 3398 | @noindent |
| 3399 | (If you are reading this in GNU Emacs, you can evaluate this expression |
| 3400 | yourself.) First, the interpreter will evaluate the inner list, which |
| 3401 | is @code{(* 7 5)}. This returns a value of 35. Next, it |
| 3402 | will evaluate the outer list, passing the values of the second and |
| 3403 | subsequent elements of the list to the function @code{message}. |
| 3404 | |
| 3405 | As we have seen, @code{message} is an Emacs Lisp function especially |
| 3406 | designed for sending a one line message to a user. (@xref{message, , |
| 3407 | The @code{message} function}.) In summary, the @code{message} |
| 3408 | function prints its first argument in the echo area as is, except for |
| 3409 | occurrences of @samp{%d} or @samp{%s} (and various other %-sequences |
| 3410 | which we have not mentioned). When it sees a control sequence, the |
| 3411 | function looks to the second or subsequent arguments and prints the |
| 3412 | value of the argument in the location in the string where the control |
| 3413 | sequence is located. |
| 3414 | |
| 3415 | In the interactive @code{multiply-by-seven} function, the control string |
| 3416 | is @samp{%d}, which requires a number, and the value returned by |
| 3417 | evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35 |
| 3418 | is printed in place of the @samp{%d} and the message is @samp{The result |
| 3419 | is 35}. |
| 3420 | |
| 3421 | (Note that when you call the function @code{multiply-by-seven}, the |
| 3422 | message is printed without quotes, but when you call @code{message}, the |
| 3423 | text is printed in double quotes. This is because the value returned by |
| 3424 | @code{message} is what appears in the echo area when you evaluate an |
| 3425 | expression whose first element is @code{message}; but when embedded in a |
| 3426 | function, @code{message} prints the text as a side effect without |
| 3427 | quotes.) |
| 3428 | |
| 3429 | @node Interactive Options |
| 3430 | @section Different Options for @code{interactive} |
| 3431 | @cindex Options for @code{interactive} |
| 3432 | @cindex Interactive options |
| 3433 | |
| 3434 | In the example, @code{multiply-by-seven} used @code{"p"} as the |
| 3435 | argument to @code{interactive}. This argument told Emacs to interpret |
| 3436 | your typing either @kbd{C-u} followed by a number or @key{META} |
| 3437 | followed by a number as a command to pass that number to the function |
| 3438 | as its argument. Emacs has more than twenty characters predefined for |
| 3439 | use with @code{interactive}. In almost every case, one of these |
| 3440 | options will enable you to pass the right information interactively to |
| 3441 | a function. (@xref{Interactive Codes, , Code Characters for |
| 3442 | @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.) |
| 3443 | |
| 3444 | @need 1250 |
| 3445 | Consider the function @code{zap-to-char}. Its interactive expression |
| 3446 | is |
| 3447 | |
| 3448 | @smallexample |
| 3449 | (interactive "p\ncZap to char: ") |
| 3450 | @end smallexample |
| 3451 | |
| 3452 | The first part of the argument to @code{interactive} is @samp{p}, with |
| 3453 | which you are already familiar. This argument tells Emacs to |
| 3454 | interpret a `prefix', as a number to be passed to the function. You |
| 3455 | can specify a prefix either by typing @kbd{C-u} followed by a number |
| 3456 | or by typing @key{META} followed by a number. The prefix is the |
| 3457 | number of specified characters. Thus, if your prefix is three and the |
| 3458 | specified character is @samp{x}, then you will delete all the text up |
| 3459 | to and including the third next @samp{x}. If you do not set a prefix, |
| 3460 | then you delete all the text up to and including the specified |
| 3461 | character, but no more. |
| 3462 | |
| 3463 | The @samp{c} tells the function the name of the character to which to delete. |
| 3464 | |
| 3465 | More formally, a function with two or more arguments can have |
| 3466 | information passed to each argument by adding parts to the string that |
| 3467 | follows @code{interactive}. When you do this, the information is |
| 3468 | passed to each argument in the same order it is specified in the |
| 3469 | @code{interactive} list. In the string, each part is separated from |
| 3470 | the next part by a @samp{\n}, which is a newline. For example, you |
| 3471 | can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }. |
| 3472 | This causes Emacs to pass the value of the prefix argument (if there |
| 3473 | is one) and the character. |
| 3474 | |
| 3475 | In this case, the function definition looks like the following, where |
| 3476 | @code{arg} and @code{char} are the symbols to which @code{interactive} |
| 3477 | binds the prefix argument and the specified character: |
| 3478 | |
| 3479 | @smallexample |
| 3480 | @group |
| 3481 | (defun @var{name-of-function} (arg char) |
| 3482 | "@var{documentation}@dots{}" |
| 3483 | (interactive "p\ncZap to char: ") |
| 3484 | @var{body-of-function}@dots{}) |
| 3485 | @end group |
| 3486 | @end smallexample |
| 3487 | |
| 3488 | @noindent |
| 3489 | (The space after the colon in the prompt makes it look better when you |
| 3490 | are prompted. @xref{copy-to-buffer, , The Definition of |
| 3491 | @code{copy-to-buffer}}, for an example.) |
| 3492 | |
| 3493 | When a function does not take arguments, @code{interactive} does not |
| 3494 | require any. Such a function contains the simple expression |
| 3495 | @code{(interactive)}. The @code{mark-whole-buffer} function is like |
| 3496 | this. |
| 3497 | |
| 3498 | Alternatively, if the special letter-codes are not right for your |
| 3499 | application, you can pass your own arguments to @code{interactive} as |
| 3500 | a list. |
| 3501 | |
| 3502 | @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}, |
| 3503 | for an example. @xref{Using Interactive, , Using @code{Interactive}, |
| 3504 | elisp, The GNU Emacs Lisp Reference Manual}, for a more complete |
| 3505 | explanation about this technique. |
| 3506 | |
| 3507 | @node Permanent Installation |
| 3508 | @section Install Code Permanently |
| 3509 | @cindex Install code permanently |
| 3510 | @cindex Permanent code installation |
| 3511 | @cindex Code installation |
| 3512 | |
| 3513 | When you install a function definition by evaluating it, it will stay |
| 3514 | installed until you quit Emacs. The next time you start a new session |
| 3515 | of Emacs, the function will not be installed unless you evaluate the |
| 3516 | function definition again. |
| 3517 | |
| 3518 | At some point, you may want to have code installed automatically |
| 3519 | whenever you start a new session of Emacs. There are several ways of |
| 3520 | doing this: |
| 3521 | |
| 3522 | @itemize @bullet |
| 3523 | @item |
| 3524 | If you have code that is just for yourself, you can put the code for the |
| 3525 | function definition in your @file{.emacs} initialization file. When you |
| 3526 | start Emacs, your @file{.emacs} file is automatically evaluated and all |
| 3527 | the function definitions within it are installed. |
| 3528 | @xref{Emacs Initialization, , Your @file{.emacs} File}. |
| 3529 | |
| 3530 | @item |
| 3531 | Alternatively, you can put the function definitions that you want |
| 3532 | installed in one or more files of their own and use the @code{load} |
| 3533 | function to cause Emacs to evaluate and thereby install each of the |
| 3534 | functions in the files. |
| 3535 | @xref{Loading Files, , Loading Files}. |
| 3536 | |
| 3537 | @item |
| 3538 | Thirdly, if you have code that your whole site will use, it is usual |
| 3539 | to put it in a file called @file{site-init.el} that is loaded when |
| 3540 | Emacs is built. This makes the code available to everyone who uses |
| 3541 | your machine. (See the @file{INSTALL} file that is part of the Emacs |
| 3542 | distribution.) |
| 3543 | @end itemize |
| 3544 | |
| 3545 | Finally, if you have code that everyone who uses Emacs may want, you |
| 3546 | can post it on a computer network or send a copy to the Free Software |
| 3547 | Foundation. (When you do this, please license the code and its |
| 3548 | documentation under a license that permits other people to run, copy, |
| 3549 | study, modify, and redistribute the code and which protects you from |
| 3550 | having your work taken from you.) If you send a copy of your code to |
| 3551 | the Free Software Foundation, and properly protect yourself and |
| 3552 | others, it may be included in the next release of Emacs. In large |
| 3553 | part, this is how Emacs has grown over the past years, by donations. |
| 3554 | |
| 3555 | @node let |
| 3556 | @section @code{let} |
| 3557 | @findex let |
| 3558 | |
| 3559 | The @code{let} expression is a special form in Lisp that you will need |
| 3560 | to use in most function definitions. |
| 3561 | |
| 3562 | @code{let} is used to attach or bind a symbol to a value in such a way |
| 3563 | that the Lisp interpreter will not confuse the variable with a |
| 3564 | variable of the same name that is not part of the function. |
| 3565 | |
| 3566 | To understand why the @code{let} special form is necessary, consider |
| 3567 | the situation in which you own a home that you generally refer to as |
| 3568 | `the house', as in the sentence, ``The house needs painting.'' If you |
| 3569 | are visiting a friend and your host refers to `the house', he is |
| 3570 | likely to be referring to @emph{his} house, not yours, that is, to a |
| 3571 | different house. |
| 3572 | |
| 3573 | If your friend is referring to his house and you think he is referring |
| 3574 | to your house, you may be in for some confusion. The same thing could |
| 3575 | happen in Lisp if a variable that is used inside of one function has |
| 3576 | the same name as a variable that is used inside of another function, |
| 3577 | and the two are not intended to refer to the same value. The |
| 3578 | @code{let} special form prevents this kind of confusion. |
| 3579 | |
| 3580 | @menu |
| 3581 | * Prevent confusion:: |
| 3582 | * Parts of let Expression:: |
| 3583 | * Sample let Expression:: |
| 3584 | * Uninitialized let Variables:: |
| 3585 | @end menu |
| 3586 | |
| 3587 | @ifnottex |
| 3588 | @node Prevent confusion |
| 3589 | @unnumberedsubsec @code{let} Prevents Confusion |
| 3590 | @end ifnottex |
| 3591 | |
| 3592 | @cindex @samp{local variable} defined |
| 3593 | @cindex @samp{variable, local}, defined |
| 3594 | The @code{let} special form prevents confusion. @code{let} creates a |
| 3595 | name for a @dfn{local variable} that overshadows any use of the same |
| 3596 | name outside the @code{let} expression. This is like understanding |
| 3597 | that whenever your host refers to `the house', he means his house, not |
| 3598 | yours. (Symbols used in argument lists work the same way. |
| 3599 | @xref{defun, , The @code{defun} Macro}.) |
| 3600 | |
| 3601 | Local variables created by a @code{let} expression retain their value |
| 3602 | @emph{only} within the @code{let} expression itself (and within |
| 3603 | expressions called within the @code{let} expression); the local |
| 3604 | variables have no effect outside the @code{let} expression. |
| 3605 | |
| 3606 | Another way to think about @code{let} is that it is like a @code{setq} |
| 3607 | that is temporary and local. The values set by @code{let} are |
| 3608 | automatically undone when the @code{let} is finished. The setting |
| 3609 | only affects expressions that are inside the bounds of the @code{let} |
| 3610 | expression. In computer science jargon, we would say ``the binding of |
| 3611 | a symbol is visible only in functions called in the @code{let} form; |
| 3612 | in Emacs Lisp, scoping is dynamic, not lexical.'' |
| 3613 | |
| 3614 | @code{let} can create more than one variable at once. Also, |
| 3615 | @code{let} gives each variable it creates an initial value, either a |
| 3616 | value specified by you, or @code{nil}. (In the jargon, this is called |
| 3617 | `binding the variable to the value'.) After @code{let} has created |
| 3618 | and bound the variables, it executes the code in the body of the |
| 3619 | @code{let}, and returns the value of the last expression in the body, |
| 3620 | as the value of the whole @code{let} expression. (`Execute' is a jargon |
| 3621 | term that means to evaluate a list; it comes from the use of the word |
| 3622 | meaning `to give practical effect to' (@cite{Oxford English |
| 3623 | Dictionary}). Since you evaluate an expression to perform an action, |
| 3624 | `execute' has evolved as a synonym to `evaluate'.) |
| 3625 | |
| 3626 | @node Parts of let Expression |
| 3627 | @subsection The Parts of a @code{let} Expression |
| 3628 | @cindex @code{let} expression, parts of |
| 3629 | @cindex Parts of @code{let} expression |
| 3630 | |
| 3631 | @cindex @samp{varlist} defined |
| 3632 | A @code{let} expression is a list of three parts. The first part is |
| 3633 | the symbol @code{let}. The second part is a list, called a |
| 3634 | @dfn{varlist}, each element of which is either a symbol by itself or a |
| 3635 | two-element list, the first element of which is a symbol. The third |
| 3636 | part of the @code{let} expression is the body of the @code{let}. The |
| 3637 | body usually consists of one or more lists. |
| 3638 | |
| 3639 | @need 800 |
| 3640 | A template for a @code{let} expression looks like this: |
| 3641 | |
| 3642 | @smallexample |
| 3643 | (let @var{varlist} @var{body}@dots{}) |
| 3644 | @end smallexample |
| 3645 | |
| 3646 | @noindent |
| 3647 | The symbols in the varlist are the variables that are given initial |
| 3648 | values by the @code{let} special form. Symbols by themselves are given |
| 3649 | the initial value of @code{nil}; and each symbol that is the first |
| 3650 | element of a two-element list is bound to the value that is returned |
| 3651 | when the Lisp interpreter evaluates the second element. |
| 3652 | |
| 3653 | Thus, a varlist might look like this: @code{(thread (needles 3))}. In |
| 3654 | this case, in a @code{let} expression, Emacs binds the symbol |
| 3655 | @code{thread} to an initial value of @code{nil}, and binds the symbol |
| 3656 | @code{needles} to an initial value of 3. |
| 3657 | |
| 3658 | When you write a @code{let} expression, what you do is put the |
| 3659 | appropriate expressions in the slots of the @code{let} expression |
| 3660 | template. |
| 3661 | |
| 3662 | If the varlist is composed of two-element lists, as is often the case, |
| 3663 | the template for the @code{let} expression looks like this: |
| 3664 | |
| 3665 | @smallexample |
| 3666 | @group |
| 3667 | (let ((@var{variable} @var{value}) |
| 3668 | (@var{variable} @var{value}) |
| 3669 | @dots{}) |
| 3670 | @var{body}@dots{}) |
| 3671 | @end group |
| 3672 | @end smallexample |
| 3673 | |
| 3674 | @node Sample let Expression |
| 3675 | @subsection Sample @code{let} Expression |
| 3676 | @cindex Sample @code{let} expression |
| 3677 | @cindex @code{let} expression sample |
| 3678 | |
| 3679 | The following expression creates and gives initial values |
| 3680 | to the two variables @code{zebra} and @code{tiger}. The body of the |
| 3681 | @code{let} expression is a list which calls the @code{message} function. |
| 3682 | |
| 3683 | @smallexample |
| 3684 | @group |
| 3685 | (let ((zebra 'stripes) |
| 3686 | (tiger 'fierce)) |
| 3687 | (message "One kind of animal has %s and another is %s." |
| 3688 | zebra tiger)) |
| 3689 | @end group |
| 3690 | @end smallexample |
| 3691 | |
| 3692 | Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}. |
| 3693 | |
| 3694 | The two variables are @code{zebra} and @code{tiger}. Each variable is |
| 3695 | the first element of a two-element list and each value is the second |
| 3696 | element of its two-element list. In the varlist, Emacs binds the |
| 3697 | variable @code{zebra} to the value @code{stripes}@footnote{According |
| 3698 | to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras |
| 3699 | become impossibly dangerous as they grow older'' but the claim here is |
| 3700 | that they do not become fierce like a tiger. (1997, W. W. Norton and |
| 3701 | Co., ISBN 0-393-03894-2, page 171)}, and binds the |
| 3702 | variable @code{tiger} to the value @code{fierce}. In this example, |
| 3703 | both values are symbols preceded by a quote. The values could just as |
| 3704 | well have been another list or a string. The body of the @code{let} |
| 3705 | follows after the list holding the variables. In this example, the |
| 3706 | body is a list that uses the @code{message} function to print a string |
| 3707 | in the echo area. |
| 3708 | |
| 3709 | @need 1500 |
| 3710 | You may evaluate the example in the usual fashion, by placing the |
| 3711 | cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do |
| 3712 | this, the following will appear in the echo area: |
| 3713 | |
| 3714 | @smallexample |
| 3715 | "One kind of animal has stripes and another is fierce." |
| 3716 | @end smallexample |
| 3717 | |
| 3718 | As we have seen before, the @code{message} function prints its first |
| 3719 | argument, except for @samp{%s}. In this example, the value of the variable |
| 3720 | @code{zebra} is printed at the location of the first @samp{%s} and the |
| 3721 | value of the variable @code{tiger} is printed at the location of the |
| 3722 | second @samp{%s}. |
| 3723 | |
| 3724 | @node Uninitialized let Variables |
| 3725 | @subsection Uninitialized Variables in a @code{let} Statement |
| 3726 | @cindex Uninitialized @code{let} variables |
| 3727 | @cindex @code{let} variables uninitialized |
| 3728 | |
| 3729 | If you do not bind the variables in a @code{let} statement to specific |
| 3730 | initial values, they will automatically be bound to an initial value of |
| 3731 | @code{nil}, as in the following expression: |
| 3732 | |
| 3733 | @smallexample |
| 3734 | @group |
| 3735 | (let ((birch 3) |
| 3736 | pine |
| 3737 | fir |
| 3738 | (oak 'some)) |
| 3739 | (message |
| 3740 | "Here are %d variables with %s, %s, and %s value." |
| 3741 | birch pine fir oak)) |
| 3742 | @end group |
| 3743 | @end smallexample |
| 3744 | |
| 3745 | @noindent |
| 3746 | Here, the varlist is @code{((birch 3) pine fir (oak 'some))}. |
| 3747 | |
| 3748 | @need 1250 |
| 3749 | If you evaluate this expression in the usual way, the following will |
| 3750 | appear in your echo area: |
| 3751 | |
| 3752 | @smallexample |
| 3753 | "Here are 3 variables with nil, nil, and some value." |
| 3754 | @end smallexample |
| 3755 | |
| 3756 | @noindent |
| 3757 | In this example, Emacs binds the symbol @code{birch} to the number 3, |
| 3758 | binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds |
| 3759 | the symbol @code{oak} to the value @code{some}. |
| 3760 | |
| 3761 | Note that in the first part of the @code{let}, the variables @code{pine} |
| 3762 | and @code{fir} stand alone as atoms that are not surrounded by |
| 3763 | parentheses; this is because they are being bound to @code{nil}, the |
| 3764 | empty list. But @code{oak} is bound to @code{some} and so is a part of |
| 3765 | the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the |
| 3766 | number 3 and so is in a list with that number. (Since a number |
| 3767 | evaluates to itself, the number does not need to be quoted. Also, the |
| 3768 | number is printed in the message using a @samp{%d} rather than a |
| 3769 | @samp{%s}.) The four variables as a group are put into a list to |
| 3770 | delimit them from the body of the @code{let}. |
| 3771 | |
| 3772 | @node if |
| 3773 | @section The @code{if} Special Form |
| 3774 | @findex if |
| 3775 | @cindex Conditional with @code{if} |
| 3776 | |
| 3777 | A third special form, in addition to @code{defun} and @code{let}, is the |
| 3778 | conditional @code{if}. This form is used to instruct the computer to |
| 3779 | make decisions. You can write function definitions without using |
| 3780 | @code{if}, but it is used often enough, and is important enough, to be |
| 3781 | included here. It is used, for example, in the code for the |
| 3782 | function @code{beginning-of-buffer}. |
| 3783 | |
| 3784 | The basic idea behind an @code{if}, is that ``@emph{if} a test is true, |
| 3785 | @emph{then} an expression is evaluated.'' If the test is not true, the |
| 3786 | expression is not evaluated. For example, you might make a decision |
| 3787 | such as, ``if it is warm and sunny, then go to the beach!'' |
| 3788 | |
| 3789 | @menu |
| 3790 | * if in more detail:: |
| 3791 | * type-of-animal in detail:: An example of an @code{if} expression. |
| 3792 | @end menu |
| 3793 | |
| 3794 | @ifnottex |
| 3795 | @node if in more detail |
| 3796 | @unnumberedsubsec @code{if} in more detail |
| 3797 | @end ifnottex |
| 3798 | |
| 3799 | @cindex @samp{if-part} defined |
| 3800 | @cindex @samp{then-part} defined |
| 3801 | An @code{if} expression written in Lisp does not use the word `then'; |
| 3802 | the test and the action are the second and third elements of the list |
| 3803 | whose first element is @code{if}. Nonetheless, the test part of an |
| 3804 | @code{if} expression is often called the @dfn{if-part} and the second |
| 3805 | argument is often called the @dfn{then-part}. |
| 3806 | |
| 3807 | Also, when an @code{if} expression is written, the true-or-false-test |
| 3808 | is usually written on the same line as the symbol @code{if}, but the |
| 3809 | action to carry out if the test is true, the ``then-part'', is written |
| 3810 | on the second and subsequent lines. This makes the @code{if} |
| 3811 | expression easier to read. |
| 3812 | |
| 3813 | @smallexample |
| 3814 | @group |
| 3815 | (if @var{true-or-false-test} |
| 3816 | @var{action-to-carry-out-if-test-is-true}) |
| 3817 | @end group |
| 3818 | @end smallexample |
| 3819 | |
| 3820 | @noindent |
| 3821 | The true-or-false-test will be an expression that |
| 3822 | is evaluated by the Lisp interpreter. |
| 3823 | |
| 3824 | Here is an example that you can evaluate in the usual manner. The test |
| 3825 | is whether the number 5 is greater than the number 4. Since it is, the |
| 3826 | message @samp{5 is greater than 4!} will be printed. |
| 3827 | |
| 3828 | @smallexample |
| 3829 | @group |
| 3830 | (if (> 5 4) ; @r{if-part} |
| 3831 | (message "5 is greater than 4!")) ; @r{then-part} |
| 3832 | @end group |
| 3833 | @end smallexample |
| 3834 | |
| 3835 | @noindent |
| 3836 | (The function @code{>} tests whether its first argument is greater than |
| 3837 | its second argument and returns true if it is.) |
| 3838 | @findex > (greater than) |
| 3839 | |
| 3840 | Of course, in actual use, the test in an @code{if} expression will not |
| 3841 | be fixed for all time as it is by the expression @code{(> 5 4)}. |
| 3842 | Instead, at least one of the variables used in the test will be bound to |
| 3843 | a value that is not known ahead of time. (If the value were known ahead |
| 3844 | of time, we would not need to run the test!) |
| 3845 | |
| 3846 | For example, the value may be bound to an argument of a function |
| 3847 | definition. In the following function definition, the character of the |
| 3848 | animal is a value that is passed to the function. If the value bound to |
| 3849 | @code{characteristic} is @code{fierce}, then the message, @samp{It's a |
| 3850 | tiger!} will be printed; otherwise, @code{nil} will be returned. |
| 3851 | |
| 3852 | @smallexample |
| 3853 | @group |
| 3854 | (defun type-of-animal (characteristic) |
| 3855 | "Print message in echo area depending on CHARACTERISTIC. |
| 3856 | If the CHARACTERISTIC is the symbol `fierce', |
| 3857 | then warn of a tiger." |
| 3858 | (if (equal characteristic 'fierce) |
| 3859 | (message "It's a tiger!"))) |
| 3860 | @end group |
| 3861 | @end smallexample |
| 3862 | |
| 3863 | @need 1500 |
| 3864 | @noindent |
| 3865 | If you are reading this inside of GNU Emacs, you can evaluate the |
| 3866 | function definition in the usual way to install it in Emacs, and then you |
| 3867 | can evaluate the following two expressions to see the results: |
| 3868 | |
| 3869 | @smallexample |
| 3870 | @group |
| 3871 | (type-of-animal 'fierce) |
| 3872 | |
| 3873 | (type-of-animal 'zebra) |
| 3874 | |
| 3875 | @end group |
| 3876 | @end smallexample |
| 3877 | |
| 3878 | @c Following sentences rewritten to prevent overfull hbox. |
| 3879 | @noindent |
| 3880 | When you evaluate @code{(type-of-animal 'fierce)}, you will see the |
| 3881 | following message printed in the echo area: @code{"It's a tiger!"}; and |
| 3882 | when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil} |
| 3883 | printed in the echo area. |
| 3884 | |
| 3885 | @node type-of-animal in detail |
| 3886 | @subsection The @code{type-of-animal} Function in Detail |
| 3887 | |
| 3888 | Let's look at the @code{type-of-animal} function in detail. |
| 3889 | |
| 3890 | The function definition for @code{type-of-animal} was written by filling |
| 3891 | the slots of two templates, one for a function definition as a whole, and |
| 3892 | a second for an @code{if} expression. |
| 3893 | |
| 3894 | @need 1250 |
| 3895 | The template for every function that is not interactive is: |
| 3896 | |
| 3897 | @smallexample |
| 3898 | @group |
| 3899 | (defun @var{name-of-function} (@var{argument-list}) |
| 3900 | "@var{documentation}@dots{}" |
| 3901 | @var{body}@dots{}) |
| 3902 | @end group |
| 3903 | @end smallexample |
| 3904 | |
| 3905 | @need 800 |
| 3906 | The parts of the function that match this template look like this: |
| 3907 | |
| 3908 | @smallexample |
| 3909 | @group |
| 3910 | (defun type-of-animal (characteristic) |
| 3911 | "Print message in echo area depending on CHARACTERISTIC. |
| 3912 | If the CHARACTERISTIC is the symbol `fierce', |
| 3913 | then warn of a tiger." |
| 3914 | @var{body: the} @code{if} @var{expression}) |
| 3915 | @end group |
| 3916 | @end smallexample |
| 3917 | |
| 3918 | The name of function is @code{type-of-animal}; it is passed the value |
| 3919 | of one argument. The argument list is followed by a multi-line |
| 3920 | documentation string. The documentation string is included in the |
| 3921 | example because it is a good habit to write documentation string for |
| 3922 | every function definition. The body of the function definition |
| 3923 | consists of the @code{if} expression. |
| 3924 | |
| 3925 | @need 800 |
| 3926 | The template for an @code{if} expression looks like this: |
| 3927 | |
| 3928 | @smallexample |
| 3929 | @group |
| 3930 | (if @var{true-or-false-test} |
| 3931 | @var{action-to-carry-out-if-the-test-returns-true}) |
| 3932 | @end group |
| 3933 | @end smallexample |
| 3934 | |
| 3935 | @need 1250 |
| 3936 | In the @code{type-of-animal} function, the code for the @code{if} |
| 3937 | looks like this: |
| 3938 | |
| 3939 | @smallexample |
| 3940 | @group |
| 3941 | (if (equal characteristic 'fierce) |
| 3942 | (message "It's a tiger!"))) |
| 3943 | @end group |
| 3944 | @end smallexample |
| 3945 | |
| 3946 | @need 800 |
| 3947 | Here, the true-or-false-test is the expression: |
| 3948 | |
| 3949 | @smallexample |
| 3950 | (equal characteristic 'fierce) |
| 3951 | @end smallexample |
| 3952 | |
| 3953 | @noindent |
| 3954 | In Lisp, @code{equal} is a function that determines whether its first |
| 3955 | argument is equal to its second argument. The second argument is the |
| 3956 | quoted symbol @code{'fierce} and the first argument is the value of the |
| 3957 | symbol @code{characteristic}---in other words, the argument passed to |
| 3958 | this function. |
| 3959 | |
| 3960 | In the first exercise of @code{type-of-animal}, the argument |
| 3961 | @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce} |
| 3962 | is equal to @code{fierce}, the expression, @code{(equal characteristic |
| 3963 | 'fierce)}, returns a value of true. When this happens, the @code{if} |
| 3964 | evaluates the second argument or then-part of the @code{if}: |
| 3965 | @code{(message "It's tiger!")}. |
| 3966 | |
| 3967 | On the other hand, in the second exercise of @code{type-of-animal}, the |
| 3968 | argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra} |
| 3969 | is not equal to @code{fierce}, so the then-part is not evaluated and |
| 3970 | @code{nil} is returned by the @code{if} expression. |
| 3971 | |
| 3972 | @node else |
| 3973 | @section If--then--else Expressions |
| 3974 | @cindex Else |
| 3975 | |
| 3976 | An @code{if} expression may have an optional third argument, called |
| 3977 | the @dfn{else-part}, for the case when the true-or-false-test returns |
| 3978 | false. When this happens, the second argument or then-part of the |
| 3979 | overall @code{if} expression is @emph{not} evaluated, but the third or |
| 3980 | else-part @emph{is} evaluated. You might think of this as the cloudy |
| 3981 | day alternative for the decision ``if it is warm and sunny, then go to |
| 3982 | the beach, else read a book!''. |
| 3983 | |
| 3984 | The word ``else'' is not written in the Lisp code; the else-part of an |
| 3985 | @code{if} expression comes after the then-part. In the written Lisp, the |
| 3986 | else-part is usually written to start on a line of its own and is |
| 3987 | indented less than the then-part: |
| 3988 | |
| 3989 | @smallexample |
| 3990 | @group |
| 3991 | (if @var{true-or-false-test} |
| 3992 | @var{action-to-carry-out-if-the-test-returns-true} |
| 3993 | @var{action-to-carry-out-if-the-test-returns-false}) |
| 3994 | @end group |
| 3995 | @end smallexample |
| 3996 | |
| 3997 | For example, the following @code{if} expression prints the message @samp{4 |
| 3998 | is not greater than 5!} when you evaluate it in the usual way: |
| 3999 | |
| 4000 | @smallexample |
| 4001 | @group |
| 4002 | (if (> 4 5) ; @r{if-part} |
| 4003 | (message "4 falsely greater than 5!") ; @r{then-part} |
| 4004 | (message "4 is not greater than 5!")) ; @r{else-part} |
| 4005 | @end group |
| 4006 | @end smallexample |
| 4007 | |
| 4008 | @noindent |
| 4009 | Note that the different levels of indentation make it easy to |
| 4010 | distinguish the then-part from the else-part. (GNU Emacs has several |
| 4011 | commands that automatically indent @code{if} expressions correctly. |
| 4012 | @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.) |
| 4013 | |
| 4014 | We can extend the @code{type-of-animal} function to include an |
| 4015 | else-part by simply incorporating an additional part to the @code{if} |
| 4016 | expression. |
| 4017 | |
| 4018 | @need 1500 |
| 4019 | You can see the consequences of doing this if you evaluate the following |
| 4020 | version of the @code{type-of-animal} function definition to install it |
| 4021 | and then evaluate the two subsequent expressions to pass different |
| 4022 | arguments to the function. |
| 4023 | |
| 4024 | @smallexample |
| 4025 | @group |
| 4026 | (defun type-of-animal (characteristic) ; @r{Second version.} |
| 4027 | "Print message in echo area depending on CHARACTERISTIC. |
| 4028 | If the CHARACTERISTIC is the symbol `fierce', |
| 4029 | then warn of a tiger; |
| 4030 | else say it's not fierce." |
| 4031 | (if (equal characteristic 'fierce) |
| 4032 | (message "It's a tiger!") |
| 4033 | (message "It's not fierce!"))) |
| 4034 | @end group |
| 4035 | @end smallexample |
| 4036 | @sp 1 |
| 4037 | |
| 4038 | @smallexample |
| 4039 | @group |
| 4040 | (type-of-animal 'fierce) |
| 4041 | |
| 4042 | (type-of-animal 'zebra) |
| 4043 | |
| 4044 | @end group |
| 4045 | @end smallexample |
| 4046 | |
| 4047 | @c Following sentence rewritten to prevent overfull hbox. |
| 4048 | @noindent |
| 4049 | When you evaluate @code{(type-of-animal 'fierce)}, you will see the |
| 4050 | following message printed in the echo area: @code{"It's a tiger!"}; but |
| 4051 | when you evaluate @code{(type-of-animal 'zebra)}, you will see |
| 4052 | @code{"It's not fierce!"}. |
| 4053 | |
| 4054 | (Of course, if the @var{characteristic} were @code{ferocious}, the |
| 4055 | message @code{"It's not fierce!"} would be printed; and it would be |
| 4056 | misleading! When you write code, you need to take into account the |
| 4057 | possibility that some such argument will be tested by the @code{if} |
| 4058 | and write your program accordingly.) |
| 4059 | |
| 4060 | @node Truth & Falsehood |
| 4061 | @section Truth and Falsehood in Emacs Lisp |
| 4062 | @cindex Truth and falsehood in Emacs Lisp |
| 4063 | @cindex Falsehood and truth in Emacs Lisp |
| 4064 | @findex nil |
| 4065 | |
| 4066 | There is an important aspect to the truth test in an @code{if} |
| 4067 | expression. So far, we have spoken of `true' and `false' as values of |
| 4068 | predicates as if they were new kinds of Emacs Lisp objects. In fact, |
| 4069 | `false' is just our old friend @code{nil}. Anything else---anything |
| 4070 | at all---is `true'. |
| 4071 | |
| 4072 | The expression that tests for truth is interpreted as @dfn{true} |
| 4073 | if the result of evaluating it is a value that is not @code{nil}. In |
| 4074 | other words, the result of the test is considered true if the value |
| 4075 | returned is a number such as 47, a string such as @code{"hello"}, or a |
| 4076 | symbol (other than @code{nil}) such as @code{flowers}, or a list (so |
| 4077 | long as it is not empty), or even a buffer! |
| 4078 | |
| 4079 | @menu |
| 4080 | * nil explained:: @code{nil} has two meanings. |
| 4081 | @end menu |
| 4082 | |
| 4083 | @ifnottex |
| 4084 | @node nil explained |
| 4085 | @unnumberedsubsec An explanation of @code{nil} |
| 4086 | @end ifnottex |
| 4087 | |
| 4088 | Before illustrating a test for truth, we need an explanation of @code{nil}. |
| 4089 | |
| 4090 | In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the |
| 4091 | empty list. Second, it means false and is the value returned when a |
| 4092 | true-or-false-test tests false. @code{nil} can be written as an empty |
| 4093 | list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is |
| 4094 | concerned, @code{()} and @code{nil} are the same. Humans, however, tend |
| 4095 | to use @code{nil} for false and @code{()} for the empty list. |
| 4096 | |
| 4097 | In Emacs Lisp, any value that is not @code{nil}---is not the empty |
| 4098 | list---is considered true. This means that if an evaluation returns |
| 4099 | something that is not an empty list, an @code{if} expression will test |
| 4100 | true. For example, if a number is put in the slot for the test, it |
| 4101 | will be evaluated and will return itself, since that is what numbers |
| 4102 | do when evaluated. In this conditional, the @code{if} expression will |
| 4103 | test true. The expression tests false only when @code{nil}, an empty |
| 4104 | list, is returned by evaluating the expression. |
| 4105 | |
| 4106 | You can see this by evaluating the two expressions in the following examples. |
| 4107 | |
| 4108 | In the first example, the number 4 is evaluated as the test in the |
| 4109 | @code{if} expression and returns itself; consequently, the then-part |
| 4110 | of the expression is evaluated and returned: @samp{true} appears in |
| 4111 | the echo area. In the second example, the @code{nil} indicates false; |
| 4112 | consequently, the else-part of the expression is evaluated and |
| 4113 | returned: @samp{false} appears in the echo area. |
| 4114 | |
| 4115 | @smallexample |
| 4116 | @group |
| 4117 | (if 4 |
| 4118 | 'true |
| 4119 | 'false) |
| 4120 | @end group |
| 4121 | |
| 4122 | @group |
| 4123 | (if nil |
| 4124 | 'true |
| 4125 | 'false) |
| 4126 | @end group |
| 4127 | @end smallexample |
| 4128 | |
| 4129 | @need 1250 |
| 4130 | Incidentally, if some other useful value is not available for a test that |
| 4131 | returns true, then the Lisp interpreter will return the symbol @code{t} |
| 4132 | for true. For example, the expression @code{(> 5 4)} returns @code{t} |
| 4133 | when evaluated, as you can see by evaluating it in the usual way: |
| 4134 | |
| 4135 | @smallexample |
| 4136 | (> 5 4) |
| 4137 | @end smallexample |
| 4138 | |
| 4139 | @need 1250 |
| 4140 | @noindent |
| 4141 | On the other hand, this function returns @code{nil} if the test is false. |
| 4142 | |
| 4143 | @smallexample |
| 4144 | (> 4 5) |
| 4145 | @end smallexample |
| 4146 | |
| 4147 | @node save-excursion |
| 4148 | @section @code{save-excursion} |
| 4149 | @findex save-excursion |
| 4150 | @cindex Region, what it is |
| 4151 | @cindex Preserving point, mark, and buffer |
| 4152 | @cindex Point, mark, buffer preservation |
| 4153 | @findex point |
| 4154 | @findex mark |
| 4155 | |
| 4156 | The @code{save-excursion} function is the third and final special form |
| 4157 | that we will discuss in this chapter. |
| 4158 | |
| 4159 | In Emacs Lisp programs used for editing, the @code{save-excursion} |
| 4160 | function is very common. It saves the location of point and mark, |
| 4161 | executes the body of the function, and then restores point and mark to |
| 4162 | their previous positions if their locations were changed. Its primary |
| 4163 | purpose is to keep the user from being surprised and disturbed by |
| 4164 | unexpected movement of point or mark. |
| 4165 | |
| 4166 | @menu |
| 4167 | * Point and mark:: A review of various locations. |
| 4168 | * Template for save-excursion:: |
| 4169 | @end menu |
| 4170 | |
| 4171 | @ifnottex |
| 4172 | @node Point and mark |
| 4173 | @unnumberedsubsec Point and Mark |
| 4174 | @end ifnottex |
| 4175 | |
| 4176 | Before discussing @code{save-excursion}, however, it may be useful |
| 4177 | first to review what point and mark are in GNU Emacs. @dfn{Point} is |
| 4178 | the current location of the cursor. Wherever the cursor |
| 4179 | is, that is point. More precisely, on terminals where the cursor |
| 4180 | appears to be on top of a character, point is immediately before the |
| 4181 | character. In Emacs Lisp, point is an integer. The first character in |
| 4182 | a buffer is number one, the second is number two, and so on. The |
| 4183 | function @code{point} returns the current position of the cursor as a |
| 4184 | number. Each buffer has its own value for point. |
| 4185 | |
| 4186 | The @dfn{mark} is another position in the buffer; its value can be set |
| 4187 | with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If |
| 4188 | a mark has been set, you can use the command @kbd{C-x C-x} |
| 4189 | (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark |
| 4190 | and set the mark to be the previous position of point. In addition, if |
| 4191 | you set another mark, the position of the previous mark is saved in the |
| 4192 | mark ring. Many mark positions can be saved this way. You can jump the |
| 4193 | cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more |
| 4194 | times. |
| 4195 | |
| 4196 | The part of the buffer between point and mark is called @dfn{the |
| 4197 | region}. Numerous commands work on the region, including |
| 4198 | @code{center-region}, @code{count-lines-region}, @code{kill-region}, and |
| 4199 | @code{print-region}. |
| 4200 | |
| 4201 | The @code{save-excursion} special form saves the locations of point and |
| 4202 | mark and restores those positions after the code within the body of the |
| 4203 | special form is evaluated by the Lisp interpreter. Thus, if point were |
| 4204 | in the beginning of a piece of text and some code moved point to the end |
| 4205 | of the buffer, the @code{save-excursion} would put point back to where |
| 4206 | it was before, after the expressions in the body of the function were |
| 4207 | evaluated. |
| 4208 | |
| 4209 | In Emacs, a function frequently moves point as part of its internal |
| 4210 | workings even though a user would not expect this. For example, |
| 4211 | @code{count-lines-region} moves point. To prevent the user from being |
| 4212 | bothered by jumps that are both unexpected and (from the user's point of |
| 4213 | view) unnecessary, @code{save-excursion} is often used to keep point and |
| 4214 | mark in the location expected by the user. The use of |
| 4215 | @code{save-excursion} is good housekeeping. |
| 4216 | |
| 4217 | To make sure the house stays clean, @code{save-excursion} restores the |
| 4218 | values of point and mark even if something goes wrong in the code inside |
| 4219 | of it (or, to be more precise and to use the proper jargon, ``in case of |
| 4220 | abnormal exit''). This feature is very helpful. |
| 4221 | |
| 4222 | In addition to recording the values of point and mark, |
| 4223 | @code{save-excursion} keeps track of the current buffer, and restores |
| 4224 | it, too. This means you can write code that will change the buffer and |
| 4225 | have @code{save-excursion} switch you back to the original buffer. |
| 4226 | This is how @code{save-excursion} is used in @code{append-to-buffer}. |
| 4227 | (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.) |
| 4228 | |
| 4229 | @node Template for save-excursion |
| 4230 | @subsection Template for a @code{save-excursion} Expression |
| 4231 | |
| 4232 | @need 800 |
| 4233 | The template for code using @code{save-excursion} is simple: |
| 4234 | |
| 4235 | @smallexample |
| 4236 | @group |
| 4237 | (save-excursion |
| 4238 | @var{body}@dots{}) |
| 4239 | @end group |
| 4240 | @end smallexample |
| 4241 | |
| 4242 | @noindent |
| 4243 | The body of the function is one or more expressions that will be |
| 4244 | evaluated in sequence by the Lisp interpreter. If there is more than |
| 4245 | one expression in the body, the value of the last one will be returned |
| 4246 | as the value of the @code{save-excursion} function. The other |
| 4247 | expressions in the body are evaluated only for their side effects; and |
| 4248 | @code{save-excursion} itself is used only for its side effect (which |
| 4249 | is restoring the positions of point and mark). |
| 4250 | |
| 4251 | @need 1250 |
| 4252 | In more detail, the template for a @code{save-excursion} expression |
| 4253 | looks like this: |
| 4254 | |
| 4255 | @smallexample |
| 4256 | @group |
| 4257 | (save-excursion |
| 4258 | @var{first-expression-in-body} |
| 4259 | @var{second-expression-in-body} |
| 4260 | @var{third-expression-in-body} |
| 4261 | @dots{} |
| 4262 | @var{last-expression-in-body}) |
| 4263 | @end group |
| 4264 | @end smallexample |
| 4265 | |
| 4266 | @noindent |
| 4267 | An expression, of course, may be a symbol on its own or a list. |
| 4268 | |
| 4269 | In Emacs Lisp code, a @code{save-excursion} expression often occurs |
| 4270 | within the body of a @code{let} expression. It looks like this: |
| 4271 | |
| 4272 | @smallexample |
| 4273 | @group |
| 4274 | (let @var{varlist} |
| 4275 | (save-excursion |
| 4276 | @var{body}@dots{})) |
| 4277 | @end group |
| 4278 | @end smallexample |
| 4279 | |
| 4280 | @node Review |
| 4281 | @section Review |
| 4282 | |
| 4283 | In the last few chapters we have introduced a macro and a fair number |
| 4284 | of functions and special forms. Here they are described in brief, |
| 4285 | along with a few similar functions that have not been mentioned yet. |
| 4286 | |
| 4287 | @table @code |
| 4288 | @item eval-last-sexp |
| 4289 | Evaluate the last symbolic expression before the current location of |
| 4290 | point. The value is printed in the echo area unless the function is |
| 4291 | invoked with an argument; in that case, the output is printed in the |
| 4292 | current buffer. This command is normally bound to @kbd{C-x C-e}. |
| 4293 | |
| 4294 | @item defun |
| 4295 | Define function. This macro has up to five parts: the name, a |
| 4296 | template for the arguments that will be passed to the function, |
| 4297 | documentation, an optional interactive declaration, and the body of |
| 4298 | the definition. |
| 4299 | |
| 4300 | @need 1250 |
| 4301 | For example, in an early version of Emacs, the function definition was |
| 4302 | as follows. (It is slightly more complex now that it seeks the first |
| 4303 | non-whitespace character rather than the first visible character.) |
| 4304 | |
| 4305 | @smallexample |
| 4306 | @group |
| 4307 | (defun back-to-indentation () |
| 4308 | "Move point to first visible character on line." |
| 4309 | (interactive) |
| 4310 | (beginning-of-line 1) |
| 4311 | (skip-chars-forward " \t")) |
| 4312 | @end group |
| 4313 | @end smallexample |
| 4314 | |
| 4315 | @ignore |
| 4316 | In GNU Emacs 22, |
| 4317 | |
| 4318 | (defun backward-to-indentation (&optional arg) |
| 4319 | "Move backward ARG lines and position at first nonblank character." |
| 4320 | (interactive "p") |
| 4321 | (forward-line (- (or arg 1))) |
| 4322 | (skip-chars-forward " \t")) |
| 4323 | |
| 4324 | (defun back-to-indentation () |
| 4325 | "Move point to the first non-whitespace character on this line." |
| 4326 | (interactive) |
| 4327 | (beginning-of-line 1) |
| 4328 | (skip-syntax-forward " " (line-end-position)) |
| 4329 | ;; Move back over chars that have whitespace syntax but have the p flag. |
| 4330 | (backward-prefix-chars)) |
| 4331 | @end ignore |
| 4332 | |
| 4333 | @item interactive |
| 4334 | Declare to the interpreter that the function can be used |
| 4335 | interactively. This special form may be followed by a string with one |
| 4336 | or more parts that pass the information to the arguments of the |
| 4337 | function, in sequence. These parts may also tell the interpreter to |
| 4338 | prompt for information. Parts of the string are separated by |
| 4339 | newlines, @samp{\n}. |
| 4340 | |
| 4341 | @need 1000 |
| 4342 | Common code characters are: |
| 4343 | |
| 4344 | @table @code |
| 4345 | @item b |
| 4346 | The name of an existing buffer. |
| 4347 | |
| 4348 | @item f |
| 4349 | The name of an existing file. |
| 4350 | |
| 4351 | @item p |
| 4352 | The numeric prefix argument. (Note that this `p' is lower case.) |
| 4353 | |
| 4354 | @item r |
| 4355 | Point and the mark, as two numeric arguments, smallest first. This |
| 4356 | is the only code letter that specifies two successive arguments |
| 4357 | rather than one. |
| 4358 | @end table |
| 4359 | |
| 4360 | @xref{Interactive Codes, , Code Characters for @samp{interactive}, |
| 4361 | elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of |
| 4362 | code characters. |
| 4363 | |
| 4364 | @item let |
| 4365 | Declare that a list of variables is for use within the body of the |
| 4366 | @code{let} and give them an initial value, either @code{nil} or a |
| 4367 | specified value; then evaluate the rest of the expressions in the body |
| 4368 | of the @code{let} and return the value of the last one. Inside the |
| 4369 | body of the @code{let}, the Lisp interpreter does not see the values of |
| 4370 | the variables of the same names that are bound outside of the |
| 4371 | @code{let}. |
| 4372 | |
| 4373 | @need 1250 |
| 4374 | For example, |
| 4375 | |
| 4376 | @smallexample |
| 4377 | @group |
| 4378 | (let ((foo (buffer-name)) |
| 4379 | (bar (buffer-size))) |
| 4380 | (message |
| 4381 | "This buffer is %s and has %d characters." |
| 4382 | foo bar)) |
| 4383 | @end group |
| 4384 | @end smallexample |
| 4385 | |
| 4386 | @item save-excursion |
| 4387 | Record the values of point and mark and the current buffer before |
| 4388 | evaluating the body of this special form. Restore the values of point |
| 4389 | and mark and buffer afterward. |
| 4390 | |
| 4391 | @need 1250 |
| 4392 | For example, |
| 4393 | |
| 4394 | @smallexample |
| 4395 | @group |
| 4396 | (message "We are %d characters into this buffer." |
| 4397 | (- (point) |
| 4398 | (save-excursion |
| 4399 | (goto-char (point-min)) (point)))) |
| 4400 | @end group |
| 4401 | @end smallexample |
| 4402 | |
| 4403 | @item if |
| 4404 | Evaluate the first argument to the function; if it is true, evaluate |
| 4405 | the second argument; else evaluate the third argument, if there is one. |
| 4406 | |
| 4407 | The @code{if} special form is called a @dfn{conditional}. There are |
| 4408 | other conditionals in Emacs Lisp, but @code{if} is perhaps the most |
| 4409 | commonly used. |
| 4410 | |
| 4411 | @need 1250 |
| 4412 | For example, |
| 4413 | |
| 4414 | @smallexample |
| 4415 | @group |
| 4416 | (if (= 22 emacs-major-version) |
| 4417 | (message "This is version 22 Emacs") |
| 4418 | (message "This is not version 22 Emacs")) |
| 4419 | @end group |
| 4420 | @end smallexample |
| 4421 | |
| 4422 | @need 1250 |
| 4423 | @item < |
| 4424 | @itemx > |
| 4425 | @itemx <= |
| 4426 | @itemx >= |
| 4427 | The @code{<} function tests whether its first argument is smaller than |
| 4428 | its second argument. A corresponding function, @code{>}, tests whether |
| 4429 | the first argument is greater than the second. Likewise, @code{<=} |
| 4430 | tests whether the first argument is less than or equal to the second and |
| 4431 | @code{>=} tests whether the first argument is greater than or equal to |
| 4432 | the second. In all cases, both arguments must be numbers or markers |
| 4433 | (markers indicate positions in buffers). |
| 4434 | |
| 4435 | @need 800 |
| 4436 | @item = |
| 4437 | The @code{=} function tests whether two arguments, both numbers or |
| 4438 | markers, are equal. |
| 4439 | |
| 4440 | @need 1250 |
| 4441 | @item equal |
| 4442 | @itemx eq |
| 4443 | Test whether two objects are the same. @code{equal} uses one meaning |
| 4444 | of the word `same' and @code{eq} uses another: @code{equal} returns |
| 4445 | true if the two objects have a similar structure and contents, such as |
| 4446 | two copies of the same book. On the other hand, @code{eq}, returns |
| 4447 | true if both arguments are actually the same object. |
| 4448 | @findex equal |
| 4449 | @findex eq |
| 4450 | |
| 4451 | @need 1250 |
| 4452 | @item string< |
| 4453 | @itemx string-lessp |
| 4454 | @itemx string= |
| 4455 | @itemx string-equal |
| 4456 | The @code{string-lessp} function tests whether its first argument is |
| 4457 | smaller than the second argument. A shorter, alternative name for the |
| 4458 | same function (a @code{defalias}) is @code{string<}. |
| 4459 | |
| 4460 | The arguments to @code{string-lessp} must be strings or symbols; the |
| 4461 | ordering is lexicographic, so case is significant. The print names of |
| 4462 | symbols are used instead of the symbols themselves. |
| 4463 | |
| 4464 | @cindex @samp{empty string} defined |
| 4465 | An empty string, @samp{""}, a string with no characters in it, is |
| 4466 | smaller than any string of characters. |
| 4467 | |
| 4468 | @code{string-equal} provides the corresponding test for equality. Its |
| 4469 | shorter, alternative name is @code{string=}. There are no string test |
| 4470 | functions that correspond to @var{>}, @code{>=}, or @code{<=}. |
| 4471 | |
| 4472 | @item message |
| 4473 | Print a message in the echo area. The first argument is a string that |
| 4474 | can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of |
| 4475 | arguments that follow the string. The argument used by @samp{%s} must |
| 4476 | be a string or a symbol; the argument used by @samp{%d} must be a |
| 4477 | number. The argument used by @samp{%c} must be an @sc{ascii} code |
| 4478 | number; it will be printed as the character with that @sc{ascii} code. |
| 4479 | (Various other %-sequences have not been mentioned.) |
| 4480 | |
| 4481 | @item setq |
| 4482 | @itemx set |
| 4483 | The @code{setq} function sets the value of its first argument to the |
| 4484 | value of the second argument. The first argument is automatically |
| 4485 | quoted by @code{setq}. It does the same for succeeding pairs of |
| 4486 | arguments. Another function, @code{set}, takes only two arguments and |
| 4487 | evaluates both of them before setting the value returned by its first |
| 4488 | argument to the value returned by its second argument. |
| 4489 | |
| 4490 | @item buffer-name |
| 4491 | Without an argument, return the name of the buffer, as a string. |
| 4492 | |
| 4493 | @item buffer-file-name |
| 4494 | Without an argument, return the name of the file the buffer is |
| 4495 | visiting. |
| 4496 | |
| 4497 | @item current-buffer |
| 4498 | Return the buffer in which Emacs is active; it may not be |
| 4499 | the buffer that is visible on the screen. |
| 4500 | |
| 4501 | @item other-buffer |
| 4502 | Return the most recently selected buffer (other than the buffer passed |
| 4503 | to @code{other-buffer} as an argument and other than the current |
| 4504 | buffer). |
| 4505 | |
| 4506 | @item switch-to-buffer |
| 4507 | Select a buffer for Emacs to be active in and display it in the current |
| 4508 | window so users can look at it. Usually bound to @kbd{C-x b}. |
| 4509 | |
| 4510 | @item set-buffer |
| 4511 | Switch Emacs's attention to a buffer on which programs will run. Don't |
| 4512 | alter what the window is showing. |
| 4513 | |
| 4514 | @item buffer-size |
| 4515 | Return the number of characters in the current buffer. |
| 4516 | |
| 4517 | @item point |
| 4518 | Return the value of the current position of the cursor, as an |
| 4519 | integer counting the number of characters from the beginning of the |
| 4520 | buffer. |
| 4521 | |
| 4522 | @item point-min |
| 4523 | Return the minimum permissible value of point in |
| 4524 | the current buffer. This is 1, unless narrowing is in effect. |
| 4525 | |
| 4526 | @item point-max |
| 4527 | Return the value of the maximum permissible value of point in the |
| 4528 | current buffer. This is the end of the buffer, unless narrowing is in |
| 4529 | effect. |
| 4530 | @end table |
| 4531 | |
| 4532 | @need 1500 |
| 4533 | @node defun Exercises |
| 4534 | @section Exercises |
| 4535 | |
| 4536 | @itemize @bullet |
| 4537 | @item |
| 4538 | Write a non-interactive function that doubles the value of its |
| 4539 | argument, a number. Make that function interactive. |
| 4540 | |
| 4541 | @item |
| 4542 | Write a function that tests whether the current value of |
| 4543 | @code{fill-column} is greater than the argument passed to the function, |
| 4544 | and if so, prints an appropriate message. |
| 4545 | @end itemize |
| 4546 | |
| 4547 | @node Buffer Walk Through |
| 4548 | @chapter A Few Buffer--Related Functions |
| 4549 | |
| 4550 | In this chapter we study in detail several of the functions used in GNU |
| 4551 | Emacs. This is called a ``walk-through''. These functions are used as |
| 4552 | examples of Lisp code, but are not imaginary examples; with the |
| 4553 | exception of the first, simplified function definition, these functions |
| 4554 | show the actual code used in GNU Emacs. You can learn a great deal from |
| 4555 | these definitions. The functions described here are all related to |
| 4556 | buffers. Later, we will study other functions. |
| 4557 | |
| 4558 | @menu |
| 4559 | * Finding More:: How to find more information. |
| 4560 | * simplified-beginning-of-buffer:: Shows @code{goto-char}, |
| 4561 | @code{point-min}, and @code{push-mark}. |
| 4562 | * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}. |
| 4563 | * append-to-buffer:: Uses @code{save-excursion} and |
| 4564 | @code{insert-buffer-substring}. |
| 4565 | * Buffer Related Review:: Review. |
| 4566 | * Buffer Exercises:: |
| 4567 | @end menu |
| 4568 | |
| 4569 | @node Finding More |
| 4570 | @section Finding More Information |
| 4571 | |
| 4572 | @findex describe-function, @r{introduced} |
| 4573 | @cindex Find function documentation |
| 4574 | In this walk-through, I will describe each new function as we come to |
| 4575 | it, sometimes in detail and sometimes briefly. If you are interested, |
| 4576 | you can get the full documentation of any Emacs Lisp function at any |
| 4577 | time by typing @kbd{C-h f} and then the name of the function (and then |
| 4578 | @key{RET}). Similarly, you can get the full documentation for a |
| 4579 | variable by typing @kbd{C-h v} and then the name of the variable (and |
| 4580 | then @key{RET}). |
| 4581 | |
| 4582 | @cindex Find source of function |
| 4583 | @c In version 22, tells location both of C and of Emacs Lisp |
| 4584 | Also, @code{describe-function} will tell you the location of the |
| 4585 | function definition. |
| 4586 | |
| 4587 | Put point into the name of the file that contains the function and |
| 4588 | press the @key{RET} key. In this case, @key{RET} means |
| 4589 | @code{push-button} rather than `return' or `enter'. Emacs will take |
| 4590 | you directly to the function definition. |
| 4591 | |
| 4592 | @ignore |
| 4593 | Not In version 22 |
| 4594 | |
| 4595 | If you move point over the file name and press |
| 4596 | the @key{RET} key, which in this case means @code{help-follow} rather |
| 4597 | than `return' or `enter', Emacs will take you directly to the function |
| 4598 | definition. |
| 4599 | @end ignore |
| 4600 | |
| 4601 | More generally, if you want to see a function in its original source |
| 4602 | file, you can use the @code{find-tag} function to jump to it. |
| 4603 | @code{find-tag} works with a wide variety of languages, not just |
| 4604 | Lisp, and C, and it works with non-programming text as well. For |
| 4605 | example, @code{find-tag} will jump to the various nodes in the |
| 4606 | Texinfo source file of this document. |
| 4607 | The @code{find-tag} function depends on `tags tables' that record |
| 4608 | the locations of the functions, variables, and other items to which |
| 4609 | @code{find-tag} jumps. |
| 4610 | |
| 4611 | To use the @code{find-tag} command, type @kbd{M-.} (i.e., press the |
| 4612 | period key while holding down the @key{META} key, or else type the |
| 4613 | @key{ESC} key and then type the period key), and then, at the prompt, |
| 4614 | type in the name of the function whose source code you want to see, |
| 4615 | such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will |
| 4616 | switch buffers and display the source code for the function on your |
| 4617 | screen. To switch back to your current buffer, type @kbd{C-x b |
| 4618 | @key{RET}}. (On some keyboards, the @key{META} key is labeled |
| 4619 | @key{ALT}.) |
| 4620 | |
| 4621 | @c !!! 22.1.1 tags table location in this paragraph |
| 4622 | @cindex TAGS table, specifying |
| 4623 | @findex find-tag |
| 4624 | Depending on how the initial default values of your copy of Emacs are |
| 4625 | set, you may also need to specify the location of your `tags table', |
| 4626 | which is a file called @file{TAGS}. For example, if you are |
| 4627 | interested in Emacs sources, the tags table you will most likely want, |
| 4628 | if it has already been created for you, will be in a subdirectory of |
| 4629 | the @file{/usr/local/share/emacs/} directory; thus you would use the |
| 4630 | @code{M-x visit-tags-table} command and specify a pathname such as |
| 4631 | @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table |
| 4632 | has not already been created, you will have to create it yourself. It |
| 4633 | will be in a file such as @file{/usr/local/src/emacs/src/TAGS}. |
| 4634 | |
| 4635 | @need 1250 |
| 4636 | To create a @file{TAGS} file in a specific directory, switch to that |
| 4637 | directory in Emacs using @kbd{M-x cd} command, or list the directory |
| 4638 | with @kbd{C-x d} (@code{dired}). Then run the compile command, with |
| 4639 | @w{@code{etags *.el}} as the command to execute: |
| 4640 | |
| 4641 | @smallexample |
| 4642 | M-x compile RET etags *.el RET |
| 4643 | @end smallexample |
| 4644 | |
| 4645 | For more information, see @ref{etags, , Create Your Own @file{TAGS} File}. |
| 4646 | |
| 4647 | After you become more familiar with Emacs Lisp, you will find that you will |
| 4648 | frequently use @code{find-tag} to navigate your way around source code; |
| 4649 | and you will create your own @file{TAGS} tables. |
| 4650 | |
| 4651 | @cindex Library, as term for `file' |
| 4652 | Incidentally, the files that contain Lisp code are conventionally |
| 4653 | called @dfn{libraries}. The metaphor is derived from that of a |
| 4654 | specialized library, such as a law library or an engineering library, |
| 4655 | rather than a general library. Each library, or file, contains |
| 4656 | functions that relate to a particular topic or activity, such as |
| 4657 | @file{abbrev.el} for handling abbreviations and other typing |
| 4658 | shortcuts, and @file{help.el} for on-line help. (Sometimes several |
| 4659 | libraries provide code for a single activity, as the various |
| 4660 | @file{rmail@dots{}} files provide code for reading electronic mail.) |
| 4661 | In @cite{The GNU Emacs Manual}, you will see sentences such as ``The |
| 4662 | @kbd{C-h p} command lets you search the standard Emacs Lisp libraries |
| 4663 | by topic keywords.'' |
| 4664 | |
| 4665 | @node simplified-beginning-of-buffer |
| 4666 | @section A Simplified @code{beginning-of-buffer} Definition |
| 4667 | @findex simplified-beginning-of-buffer |
| 4668 | |
| 4669 | The @code{beginning-of-buffer} command is a good function to start with |
| 4670 | since you are likely to be familiar with it and it is easy to |
| 4671 | understand. Used as an interactive command, @code{beginning-of-buffer} |
| 4672 | moves the cursor to the beginning of the buffer, leaving the mark at the |
| 4673 | previous position. It is generally bound to @kbd{M-<}. |
| 4674 | |
| 4675 | In this section, we will discuss a shortened version of the function |
| 4676 | that shows how it is most frequently used. This shortened function |
| 4677 | works as written, but it does not contain the code for a complex option. |
| 4678 | In another section, we will describe the entire function. |
| 4679 | (@xref{beginning-of-buffer, , Complete Definition of |
| 4680 | @code{beginning-of-buffer}}.) |
| 4681 | |
| 4682 | Before looking at the code, let's consider what the function |
| 4683 | definition has to contain: it must include an expression that makes |
| 4684 | the function interactive so it can be called by typing @kbd{M-x |
| 4685 | beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it |
| 4686 | must include code to leave a mark at the original position in the |
| 4687 | buffer; and it must include code to move the cursor to the beginning |
| 4688 | of the buffer. |
| 4689 | |
| 4690 | @need 1250 |
| 4691 | Here is the complete text of the shortened version of the function: |
| 4692 | |
| 4693 | @smallexample |
| 4694 | @group |
| 4695 | (defun simplified-beginning-of-buffer () |
| 4696 | "Move point to the beginning of the buffer; |
| 4697 | leave mark at previous position." |
| 4698 | (interactive) |
| 4699 | (push-mark) |
| 4700 | (goto-char (point-min))) |
| 4701 | @end group |
| 4702 | @end smallexample |
| 4703 | |
| 4704 | Like all function definitions, this definition has five parts following |
| 4705 | the macro @code{defun}: |
| 4706 | |
| 4707 | @enumerate |
| 4708 | @item |
| 4709 | The name: in this example, @code{simplified-beginning-of-buffer}. |
| 4710 | |
| 4711 | @item |
| 4712 | A list of the arguments: in this example, an empty list, @code{()}, |
| 4713 | |
| 4714 | @item |
| 4715 | The documentation string. |
| 4716 | |
| 4717 | @item |
| 4718 | The interactive expression. |
| 4719 | |
| 4720 | @item |
| 4721 | The body. |
| 4722 | @end enumerate |
| 4723 | |
| 4724 | @noindent |
| 4725 | In this function definition, the argument list is empty; this means that |
| 4726 | this function does not require any arguments. (When we look at the |
| 4727 | definition for the complete function, we will see that it may be passed |
| 4728 | an optional argument.) |
| 4729 | |
| 4730 | The interactive expression tells Emacs that the function is intended to |
| 4731 | be used interactively. In this example, @code{interactive} does not have |
| 4732 | an argument because @code{simplified-beginning-of-buffer} does not |
| 4733 | require one. |
| 4734 | |
| 4735 | @need 800 |
| 4736 | The body of the function consists of the two lines: |
| 4737 | |
| 4738 | @smallexample |
| 4739 | @group |
| 4740 | (push-mark) |
| 4741 | (goto-char (point-min)) |
| 4742 | @end group |
| 4743 | @end smallexample |
| 4744 | |
| 4745 | The first of these lines is the expression, @code{(push-mark)}. When |
| 4746 | this expression is evaluated by the Lisp interpreter, it sets a mark at |
| 4747 | the current position of the cursor, wherever that may be. The position |
| 4748 | of this mark is saved in the mark ring. |
| 4749 | |
| 4750 | The next line is @code{(goto-char (point-min))}. This expression |
| 4751 | jumps the cursor to the minimum point in the buffer, that is, to the |
| 4752 | beginning of the buffer (or to the beginning of the accessible portion |
| 4753 | of the buffer if it is narrowed. @xref{Narrowing & Widening, , |
| 4754 | Narrowing and Widening}.) |
| 4755 | |
| 4756 | The @code{push-mark} command sets a mark at the place where the cursor |
| 4757 | was located before it was moved to the beginning of the buffer by the |
| 4758 | @code{(goto-char (point-min))} expression. Consequently, you can, if |
| 4759 | you wish, go back to where you were originally by typing @kbd{C-x C-x}. |
| 4760 | |
| 4761 | That is all there is to the function definition! |
| 4762 | |
| 4763 | @findex describe-function |
| 4764 | When you are reading code such as this and come upon an unfamiliar |
| 4765 | function, such as @code{goto-char}, you can find out what it does by |
| 4766 | using the @code{describe-function} command. To use this command, type |
| 4767 | @kbd{C-h f} and then type in the name of the function and press |
| 4768 | @key{RET}. The @code{describe-function} command will print the |
| 4769 | function's documentation string in a @file{*Help*} window. For |
| 4770 | example, the documentation for @code{goto-char} is: |
| 4771 | |
| 4772 | @smallexample |
| 4773 | @group |
| 4774 | Set point to POSITION, a number or marker. |
| 4775 | Beginning of buffer is position (point-min), end is (point-max). |
| 4776 | @end group |
| 4777 | @end smallexample |
| 4778 | |
| 4779 | @noindent |
| 4780 | The function's one argument is the desired position. |
| 4781 | |
| 4782 | @noindent |
| 4783 | (The prompt for @code{describe-function} will offer you the symbol |
| 4784 | under or preceding the cursor, so you can save typing by positioning |
| 4785 | the cursor right over or after the function and then typing @kbd{C-h f |
| 4786 | @key{RET}}.) |
| 4787 | |
| 4788 | The @code{end-of-buffer} function definition is written in the same way as |
| 4789 | the @code{beginning-of-buffer} definition except that the body of the |
| 4790 | function contains the expression @code{(goto-char (point-max))} in place |
| 4791 | of @code{(goto-char (point-min))}. |
| 4792 | |
| 4793 | @node mark-whole-buffer |
| 4794 | @section The Definition of @code{mark-whole-buffer} |
| 4795 | @findex mark-whole-buffer |
| 4796 | |
| 4797 | The @code{mark-whole-buffer} function is no harder to understand than the |
| 4798 | @code{simplified-beginning-of-buffer} function. In this case, however, |
| 4799 | we will look at the complete function, not a shortened version. |
| 4800 | |
| 4801 | The @code{mark-whole-buffer} function is not as commonly used as the |
| 4802 | @code{beginning-of-buffer} function, but is useful nonetheless: it |
| 4803 | marks a whole buffer as a region by putting point at the beginning and |
| 4804 | a mark at the end of the buffer. It is generally bound to @kbd{C-x |
| 4805 | h}. |
| 4806 | |
| 4807 | @menu |
| 4808 | * mark-whole-buffer overview:: |
| 4809 | * Body of mark-whole-buffer:: Only three lines of code. |
| 4810 | @end menu |
| 4811 | |
| 4812 | @ifnottex |
| 4813 | @node mark-whole-buffer overview |
| 4814 | @unnumberedsubsec An overview of @code{mark-whole-buffer} |
| 4815 | @end ifnottex |
| 4816 | |
| 4817 | @need 1250 |
| 4818 | In GNU Emacs 22, the code for the complete function looks like this: |
| 4819 | |
| 4820 | @smallexample |
| 4821 | @group |
| 4822 | (defun mark-whole-buffer () |
| 4823 | "Put point at beginning and mark at end of buffer. |
| 4824 | You probably should not use this function in Lisp programs; |
| 4825 | it is usually a mistake for a Lisp function to use any subroutine |
| 4826 | that uses or sets the mark." |
| 4827 | (interactive) |
| 4828 | (push-mark (point)) |
| 4829 | (push-mark (point-max) nil t) |
| 4830 | (goto-char (point-min))) |
| 4831 | @end group |
| 4832 | @end smallexample |
| 4833 | |
| 4834 | @need 1250 |
| 4835 | Like all other functions, the @code{mark-whole-buffer} function fits |
| 4836 | into the template for a function definition. The template looks like |
| 4837 | this: |
| 4838 | |
| 4839 | @smallexample |
| 4840 | @group |
| 4841 | (defun @var{name-of-function} (@var{argument-list}) |
| 4842 | "@var{documentation}@dots{}" |
| 4843 | (@var{interactive-expression}@dots{}) |
| 4844 | @var{body}@dots{}) |
| 4845 | @end group |
| 4846 | @end smallexample |
| 4847 | |
| 4848 | Here is how the function works: the name of the function is |
| 4849 | @code{mark-whole-buffer}; it is followed by an empty argument list, |
| 4850 | @samp{()}, which means that the function does not require arguments. |
| 4851 | The documentation comes next. |
| 4852 | |
| 4853 | The next line is an @code{(interactive)} expression that tells Emacs |
| 4854 | that the function will be used interactively. These details are similar |
| 4855 | to the @code{simplified-beginning-of-buffer} function described in the |
| 4856 | previous section. |
| 4857 | |
| 4858 | @need 1250 |
| 4859 | @node Body of mark-whole-buffer |
| 4860 | @subsection Body of @code{mark-whole-buffer} |
| 4861 | |
| 4862 | The body of the @code{mark-whole-buffer} function consists of three |
| 4863 | lines of code: |
| 4864 | |
| 4865 | @c GNU Emacs 22 |
| 4866 | @smallexample |
| 4867 | @group |
| 4868 | (push-mark (point)) |
| 4869 | (push-mark (point-max) nil t) |
| 4870 | (goto-char (point-min)) |
| 4871 | @end group |
| 4872 | @end smallexample |
| 4873 | |
| 4874 | The first of these lines is the expression, @code{(push-mark (point))}. |
| 4875 | |
| 4876 | This line does exactly the same job as the first line of the body of |
| 4877 | the @code{simplified-beginning-of-buffer} function, which is written |
| 4878 | @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark |
| 4879 | at the current position of the cursor. |
| 4880 | |
| 4881 | I don't know why the expression in @code{mark-whole-buffer} is written |
| 4882 | @code{(push-mark (point))} and the expression in |
| 4883 | @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps |
| 4884 | whoever wrote the code did not know that the arguments for |
| 4885 | @code{push-mark} are optional and that if @code{push-mark} is not |
| 4886 | passed an argument, the function automatically sets mark at the |
| 4887 | location of point by default. Or perhaps the expression was written |
| 4888 | so as to parallel the structure of the next line. In any case, the |
| 4889 | line causes Emacs to determine the position of point and set a mark |
| 4890 | there. |
| 4891 | |
| 4892 | In earlier versions of GNU Emacs, the next line of |
| 4893 | @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This |
| 4894 | expression sets a mark at the point in the buffer that has the highest |
| 4895 | number. This will be the end of the buffer (or, if the buffer is |
| 4896 | narrowed, the end of the accessible portion of the buffer. |
| 4897 | @xref{Narrowing & Widening, , Narrowing and Widening}, for more about |
| 4898 | narrowing.) After this mark has been set, the previous mark, the one |
| 4899 | set at point, is no longer set, but Emacs remembers its position, just |
| 4900 | as all other recent marks are always remembered. This means that you |
| 4901 | can, if you wish, go back to that position by typing @kbd{C-u |
| 4902 | C-@key{SPC}} twice. |
| 4903 | |
| 4904 | @need 1250 |
| 4905 | In GNU Emacs 22, the @code{(point-max)} is slightly more complicated. |
| 4906 | The line reads |
| 4907 | |
| 4908 | @smallexample |
| 4909 | (push-mark (point-max) nil t) |
| 4910 | @end smallexample |
| 4911 | |
| 4912 | @noindent |
| 4913 | The expression works nearly the same as before. It sets a mark at the |
| 4914 | highest numbered place in the buffer that it can. However, in this |
| 4915 | version, @code{push-mark} has two additional arguments. The second |
| 4916 | argument to @code{push-mark} is @code{nil}. This tells the function |
| 4917 | it @emph{should} display a message that says `Mark set' when it pushes |
| 4918 | the mark. The third argument is @code{t}. This tells |
| 4919 | @code{push-mark} to activate the mark when Transient Mark mode is |
| 4920 | turned on. Transient Mark mode highlights the currently active |
| 4921 | region. It is often turned off. |
| 4922 | |
| 4923 | Finally, the last line of the function is @code{(goto-char |
| 4924 | (point-min)))}. This is written exactly the same way as it is written |
| 4925 | in @code{beginning-of-buffer}. The expression moves the cursor to |
| 4926 | the minimum point in the buffer, that is, to the beginning of the buffer |
| 4927 | (or to the beginning of the accessible portion of the buffer). As a |
| 4928 | result of this, point is placed at the beginning of the buffer and mark |
| 4929 | is set at the end of the buffer. The whole buffer is, therefore, the |
| 4930 | region. |
| 4931 | |
| 4932 | @node append-to-buffer |
| 4933 | @section The Definition of @code{append-to-buffer} |
| 4934 | @findex append-to-buffer |
| 4935 | |
| 4936 | The @code{append-to-buffer} command is more complex than the |
| 4937 | @code{mark-whole-buffer} command. What it does is copy the region |
| 4938 | (that is, the part of the buffer between point and mark) from the |
| 4939 | current buffer to a specified buffer. |
| 4940 | |
| 4941 | @menu |
| 4942 | * append-to-buffer overview:: |
| 4943 | * append interactive:: A two part interactive expression. |
| 4944 | * append-to-buffer body:: Incorporates a @code{let} expression. |
| 4945 | * append save-excursion:: How the @code{save-excursion} works. |
| 4946 | @end menu |
| 4947 | |
| 4948 | @ifnottex |
| 4949 | @node append-to-buffer overview |
| 4950 | @unnumberedsubsec An Overview of @code{append-to-buffer} |
| 4951 | @end ifnottex |
| 4952 | |
| 4953 | @findex insert-buffer-substring |
| 4954 | The @code{append-to-buffer} command uses the |
| 4955 | @code{insert-buffer-substring} function to copy the region. |
| 4956 | @code{insert-buffer-substring} is described by its name: it takes a |
| 4957 | string of characters from part of a buffer, a ``substring'', and |
| 4958 | inserts them into another buffer. |
| 4959 | |
| 4960 | Most of @code{append-to-buffer} is |
| 4961 | concerned with setting up the conditions for |
| 4962 | @code{insert-buffer-substring} to work: the code must specify both the |
| 4963 | buffer to which the text will go, the window it comes from and goes |
| 4964 | to, and the region that will be copied. |
| 4965 | |
| 4966 | @need 1250 |
| 4967 | Here is the complete text of the function: |
| 4968 | |
| 4969 | @smallexample |
| 4970 | @group |
| 4971 | (defun append-to-buffer (buffer start end) |
| 4972 | "Append to specified buffer the text of the region. |
| 4973 | It is inserted into that buffer before its point. |
| 4974 | @end group |
| 4975 | |
| 4976 | @group |
| 4977 | When calling from a program, give three arguments: |
| 4978 | BUFFER (or buffer name), START and END. |
| 4979 | START and END specify the portion of the current buffer to be copied." |
| 4980 | (interactive |
| 4981 | (list (read-buffer "Append to buffer: " (other-buffer |
| 4982 | (current-buffer) t)) |
| 4983 | (region-beginning) (region-end))) |
| 4984 | @end group |
| 4985 | @group |
| 4986 | (let ((oldbuf (current-buffer))) |
| 4987 | (save-excursion |
| 4988 | (let* ((append-to (get-buffer-create buffer)) |
| 4989 | (windows (get-buffer-window-list append-to t t)) |
| 4990 | point) |
| 4991 | (set-buffer append-to) |
| 4992 | (setq point (point)) |
| 4993 | (barf-if-buffer-read-only) |
| 4994 | (insert-buffer-substring oldbuf start end) |
| 4995 | (dolist (window windows) |
| 4996 | (when (= (window-point window) point) |
| 4997 | (set-window-point window (point)))))))) |
| 4998 | @end group |
| 4999 | @end smallexample |
| 5000 | |
| 5001 | The function can be understood by looking at it as a series of |
| 5002 | filled-in templates. |
| 5003 | |
| 5004 | The outermost template is for the function definition. In this |
| 5005 | function, it looks like this (with several slots filled in): |
| 5006 | |
| 5007 | @smallexample |
| 5008 | @group |
| 5009 | (defun append-to-buffer (buffer start end) |
| 5010 | "@var{documentation}@dots{}" |
| 5011 | (interactive @dots{}) |
| 5012 | @var{body}@dots{}) |
| 5013 | @end group |
| 5014 | @end smallexample |
| 5015 | |
| 5016 | The first line of the function includes its name and three arguments. |
| 5017 | The arguments are the @code{buffer} to which the text will be copied, and |
| 5018 | the @code{start} and @code{end} of the region in the current buffer that |
| 5019 | will be copied. |
| 5020 | |
| 5021 | The next part of the function is the documentation, which is clear and |
| 5022 | complete. As is conventional, the three arguments are written in |
| 5023 | upper case so you will notice them easily. Even better, they are |
| 5024 | described in the same order as in the argument list. |
| 5025 | |
| 5026 | Note that the documentation distinguishes between a buffer and its |
| 5027 | name. (The function can handle either.) |
| 5028 | |
| 5029 | @node append interactive |
| 5030 | @subsection The @code{append-to-buffer} Interactive Expression |
| 5031 | |
| 5032 | Since the @code{append-to-buffer} function will be used interactively, |
| 5033 | the function must have an @code{interactive} expression. (For a |
| 5034 | review of @code{interactive}, see @ref{Interactive, , Making a |
| 5035 | Function Interactive}.) The expression reads as follows: |
| 5036 | |
| 5037 | @smallexample |
| 5038 | @group |
| 5039 | (interactive |
| 5040 | (list (read-buffer |
| 5041 | "Append to buffer: " |
| 5042 | (other-buffer (current-buffer) t)) |
| 5043 | (region-beginning) |
| 5044 | (region-end))) |
| 5045 | @end group |
| 5046 | @end smallexample |
| 5047 | |
| 5048 | @noindent |
| 5049 | This expression is not one with letters standing for parts, as |
| 5050 | described earlier. Instead, it starts a list with these parts: |
| 5051 | |
| 5052 | The first part of the list is an expression to read the name of a |
| 5053 | buffer and return it as a string. That is @code{read-buffer}. The |
| 5054 | function requires a prompt as its first argument, @samp{"Append to |
| 5055 | buffer: "}. Its second argument tells the command what value to |
| 5056 | provide if you don't specify anything. |
| 5057 | |
| 5058 | In this case that second argument is an expression containing the |
| 5059 | function @code{other-buffer}, an exception, and a @samp{t}, standing |
| 5060 | for true. |
| 5061 | |
| 5062 | The first argument to @code{other-buffer}, the exception, is yet |
| 5063 | another function, @code{current-buffer}. That is not going to be |
| 5064 | returned. The second argument is the symbol for true, @code{t}. that |
| 5065 | tells @code{other-buffer} that it may show visible buffers (except in |
| 5066 | this case, it will not show the current buffer, which makes sense). |
| 5067 | |
| 5068 | @need 1250 |
| 5069 | The expression looks like this: |
| 5070 | |
| 5071 | @smallexample |
| 5072 | (other-buffer (current-buffer) t) |
| 5073 | @end smallexample |
| 5074 | |
| 5075 | The second and third arguments to the @code{list} expression are |
| 5076 | @code{(region-beginning)} and @code{(region-end)}. These two |
| 5077 | functions specify the beginning and end of the text to be appended. |
| 5078 | |
| 5079 | @need 1250 |
| 5080 | Originally, the command used the letters @samp{B} and @samp{r}. |
| 5081 | The whole @code{interactive} expression looked like this: |
| 5082 | |
| 5083 | @smallexample |
| 5084 | (interactive "BAppend to buffer:@: \nr") |
| 5085 | @end smallexample |
| 5086 | |
| 5087 | @noindent |
| 5088 | But when that was done, the default value of the buffer switched to |
| 5089 | was invisible. That was not wanted. |
| 5090 | |
| 5091 | (The prompt was separated from the second argument with a newline, |
| 5092 | @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the |
| 5093 | two arguments that follow the symbol @code{buffer} in the function's |
| 5094 | argument list (that is, @code{start} and @code{end}) to the values of |
| 5095 | point and mark. That argument worked fine.) |
| 5096 | |
| 5097 | @node append-to-buffer body |
| 5098 | @subsection The Body of @code{append-to-buffer} |
| 5099 | |
| 5100 | @ignore |
| 5101 | in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el |
| 5102 | |
| 5103 | (defun append-to-buffer (buffer start end) |
| 5104 | "Append to specified buffer the text of the region. |
| 5105 | It is inserted into that buffer before its point. |
| 5106 | |
| 5107 | When calling from a program, give three arguments: |
| 5108 | BUFFER (or buffer name), START and END. |
| 5109 | START and END specify the portion of the current buffer to be copied." |
| 5110 | (interactive |
| 5111 | (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t)) |
| 5112 | (region-beginning) (region-end))) |
| 5113 | (let ((oldbuf (current-buffer))) |
| 5114 | (save-excursion |
| 5115 | (let* ((append-to (get-buffer-create buffer)) |
| 5116 | (windows (get-buffer-window-list append-to t t)) |
| 5117 | point) |
| 5118 | (set-buffer append-to) |
| 5119 | (setq point (point)) |
| 5120 | (barf-if-buffer-read-only) |
| 5121 | (insert-buffer-substring oldbuf start end) |
| 5122 | (dolist (window windows) |
| 5123 | (when (= (window-point window) point) |
| 5124 | (set-window-point window (point)))))))) |
| 5125 | @end ignore |
| 5126 | |
| 5127 | The body of the @code{append-to-buffer} function begins with @code{let}. |
| 5128 | |
| 5129 | As we have seen before (@pxref{let, , @code{let}}), the purpose of a |
| 5130 | @code{let} expression is to create and give initial values to one or |
| 5131 | more variables that will only be used within the body of the |
| 5132 | @code{let}. This means that such a variable will not be confused with |
| 5133 | any variable of the same name outside the @code{let} expression. |
| 5134 | |
| 5135 | We can see how the @code{let} expression fits into the function as a |
| 5136 | whole by showing a template for @code{append-to-buffer} with the |
| 5137 | @code{let} expression in outline: |
| 5138 | |
| 5139 | @smallexample |
| 5140 | @group |
| 5141 | (defun append-to-buffer (buffer start end) |
| 5142 | "@var{documentation}@dots{}" |
| 5143 | (interactive @dots{}) |
| 5144 | (let ((@var{variable} @var{value})) |
| 5145 | @var{body}@dots{}) |
| 5146 | @end group |
| 5147 | @end smallexample |
| 5148 | |
| 5149 | The @code{let} expression has three elements: |
| 5150 | |
| 5151 | @enumerate |
| 5152 | @item |
| 5153 | The symbol @code{let}; |
| 5154 | |
| 5155 | @item |
| 5156 | A varlist containing, in this case, a single two-element list, |
| 5157 | @code{(@var{variable} @var{value})}; |
| 5158 | |
| 5159 | @item |
| 5160 | The body of the @code{let} expression. |
| 5161 | @end enumerate |
| 5162 | |
| 5163 | @need 800 |
| 5164 | In the @code{append-to-buffer} function, the varlist looks like this: |
| 5165 | |
| 5166 | @smallexample |
| 5167 | (oldbuf (current-buffer)) |
| 5168 | @end smallexample |
| 5169 | |
| 5170 | @noindent |
| 5171 | In this part of the @code{let} expression, the one variable, |
| 5172 | @code{oldbuf}, is bound to the value returned by the |
| 5173 | @code{(current-buffer)} expression. The variable, @code{oldbuf}, is |
| 5174 | used to keep track of the buffer in which you are working and from |
| 5175 | which you will copy. |
| 5176 | |
| 5177 | The element or elements of a varlist are surrounded by a set of |
| 5178 | parentheses so the Lisp interpreter can distinguish the varlist from |
| 5179 | the body of the @code{let}. As a consequence, the two-element list |
| 5180 | within the varlist is surrounded by a circumscribing set of parentheses. |
| 5181 | The line looks like this: |
| 5182 | |
| 5183 | @smallexample |
| 5184 | @group |
| 5185 | (let ((oldbuf (current-buffer))) |
| 5186 | @dots{} ) |
| 5187 | @end group |
| 5188 | @end smallexample |
| 5189 | |
| 5190 | @noindent |
| 5191 | The two parentheses before @code{oldbuf} might surprise you if you did |
| 5192 | not realize that the first parenthesis before @code{oldbuf} marks the |
| 5193 | boundary of the varlist and the second parenthesis marks the beginning |
| 5194 | of the two-element list, @code{(oldbuf (current-buffer))}. |
| 5195 | |
| 5196 | @node append save-excursion |
| 5197 | @subsection @code{save-excursion} in @code{append-to-buffer} |
| 5198 | |
| 5199 | The body of the @code{let} expression in @code{append-to-buffer} |
| 5200 | consists of a @code{save-excursion} expression. |
| 5201 | |
| 5202 | The @code{save-excursion} function saves the locations of point and |
| 5203 | mark, and restores them to those positions after the expressions in the |
| 5204 | body of the @code{save-excursion} complete execution. In addition, |
| 5205 | @code{save-excursion} keeps track of the original buffer, and |
| 5206 | restores it. This is how @code{save-excursion} is used in |
| 5207 | @code{append-to-buffer}. |
| 5208 | |
| 5209 | @need 1500 |
| 5210 | @cindex Indentation for formatting |
| 5211 | @cindex Formatting convention |
| 5212 | Incidentally, it is worth noting here that a Lisp function is normally |
| 5213 | formatted so that everything that is enclosed in a multi-line spread is |
| 5214 | indented more to the right than the first symbol. In this function |
| 5215 | definition, the @code{let} is indented more than the @code{defun}, and |
| 5216 | the @code{save-excursion} is indented more than the @code{let}, like |
| 5217 | this: |
| 5218 | |
| 5219 | @smallexample |
| 5220 | @group |
| 5221 | (defun @dots{} |
| 5222 | @dots{} |
| 5223 | @dots{} |
| 5224 | (let@dots{} |
| 5225 | (save-excursion |
| 5226 | @dots{} |
| 5227 | @end group |
| 5228 | @end smallexample |
| 5229 | |
| 5230 | @need 1500 |
| 5231 | @noindent |
| 5232 | This formatting convention makes it easy to see that the lines in |
| 5233 | the body of the @code{save-excursion} are enclosed by the parentheses |
| 5234 | associated with @code{save-excursion}, just as the |
| 5235 | @code{save-excursion} itself is enclosed by the parentheses associated |
| 5236 | with the @code{let}: |
| 5237 | |
| 5238 | @smallexample |
| 5239 | @group |
| 5240 | (let ((oldbuf (current-buffer))) |
| 5241 | (save-excursion |
| 5242 | @dots{} |
| 5243 | (set-buffer @dots{}) |
| 5244 | (insert-buffer-substring oldbuf start end) |
| 5245 | @dots{})) |
| 5246 | @end group |
| 5247 | @end smallexample |
| 5248 | |
| 5249 | @need 1200 |
| 5250 | The use of the @code{save-excursion} function can be viewed as a process |
| 5251 | of filling in the slots of a template: |
| 5252 | |
| 5253 | @smallexample |
| 5254 | @group |
| 5255 | (save-excursion |
| 5256 | @var{first-expression-in-body} |
| 5257 | @var{second-expression-in-body} |
| 5258 | @dots{} |
| 5259 | @var{last-expression-in-body}) |
| 5260 | @end group |
| 5261 | @end smallexample |
| 5262 | |
| 5263 | @need 1200 |
| 5264 | @noindent |
| 5265 | In this function, the body of the @code{save-excursion} contains only |
| 5266 | one expression, the @code{let*} expression. You know about a |
| 5267 | @code{let} function. The @code{let*} function is different. It has a |
| 5268 | @samp{*} in its name. It enables Emacs to set each variable in its |
| 5269 | varlist in sequence, one after another. |
| 5270 | |
| 5271 | Its critical feature is that variables later in the varlist can make |
| 5272 | use of the values to which Emacs set variables earlier in the varlist. |
| 5273 | @xref{fwd-para let, , The @code{let*} expression}. |
| 5274 | |
| 5275 | We will skip functions like @code{let*} and focus on two: the |
| 5276 | @code{set-buffer} function and the @code{insert-buffer-substring} |
| 5277 | function. |
| 5278 | |
| 5279 | @need 1250 |
| 5280 | In the old days, the @code{set-buffer} expression was simply |
| 5281 | |
| 5282 | @smallexample |
| 5283 | (set-buffer (get-buffer-create buffer)) |
| 5284 | @end smallexample |
| 5285 | |
| 5286 | @need 1250 |
| 5287 | @noindent |
| 5288 | but now it is |
| 5289 | |
| 5290 | @smallexample |
| 5291 | (set-buffer append-to) |
| 5292 | @end smallexample |
| 5293 | |
| 5294 | @noindent |
| 5295 | @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier |
| 5296 | on in the @code{let*} expression. That extra binding would not be |
| 5297 | necessary except for that @code{append-to} is used later in the |
| 5298 | varlist as an argument to @code{get-buffer-window-list}. |
| 5299 | |
| 5300 | @ignore |
| 5301 | in GNU Emacs 22 |
| 5302 | |
| 5303 | (let ((oldbuf (current-buffer))) |
| 5304 | (save-excursion |
| 5305 | (let* ((append-to (get-buffer-create buffer)) |
| 5306 | (windows (get-buffer-window-list append-to t t)) |
| 5307 | point) |
| 5308 | (set-buffer append-to) |
| 5309 | (setq point (point)) |
| 5310 | (barf-if-buffer-read-only) |
| 5311 | (insert-buffer-substring oldbuf start end) |
| 5312 | (dolist (window windows) |
| 5313 | (when (= (window-point window) point) |
| 5314 | (set-window-point window (point)))))))) |
| 5315 | @end ignore |
| 5316 | |
| 5317 | The @code{append-to-buffer} function definition inserts text from the |
| 5318 | buffer in which you are currently to a named buffer. It happens that |
| 5319 | @code{insert-buffer-substring} copies text from another buffer to the |
| 5320 | current buffer, just the reverse---that is why the |
| 5321 | @code{append-to-buffer} definition starts out with a @code{let} that |
| 5322 | binds the local symbol @code{oldbuf} to the value returned by |
| 5323 | @code{current-buffer}. |
| 5324 | |
| 5325 | @need 1250 |
| 5326 | The @code{insert-buffer-substring} expression looks like this: |
| 5327 | |
| 5328 | @smallexample |
| 5329 | (insert-buffer-substring oldbuf start end) |
| 5330 | @end smallexample |
| 5331 | |
| 5332 | @noindent |
| 5333 | The @code{insert-buffer-substring} function copies a string |
| 5334 | @emph{from} the buffer specified as its first argument and inserts the |
| 5335 | string into the present buffer. In this case, the argument to |
| 5336 | @code{insert-buffer-substring} is the value of the variable created |
| 5337 | and bound by the @code{let}, namely the value of @code{oldbuf}, which |
| 5338 | was the current buffer when you gave the @code{append-to-buffer} |
| 5339 | command. |
| 5340 | |
| 5341 | After @code{insert-buffer-substring} has done its work, |
| 5342 | @code{save-excursion} will restore the action to the original buffer |
| 5343 | and @code{append-to-buffer} will have done its job. |
| 5344 | |
| 5345 | @need 800 |
| 5346 | Written in skeletal form, the workings of the body look like this: |
| 5347 | |
| 5348 | @smallexample |
| 5349 | @group |
| 5350 | (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer}) |
| 5351 | (save-excursion ; @r{Keep track of buffer.} |
| 5352 | @var{change-buffer} |
| 5353 | @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer}) |
| 5354 | |
| 5355 | @var{change-back-to-original-buffer-when-finished} |
| 5356 | @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished} |
| 5357 | @end group |
| 5358 | @end smallexample |
| 5359 | |
| 5360 | In summary, @code{append-to-buffer} works as follows: it saves the |
| 5361 | value of the current buffer in the variable called @code{oldbuf}. It |
| 5362 | gets the new buffer (creating one if need be) and switches Emacs's |
| 5363 | attention to it. Using the value of @code{oldbuf}, it inserts the |
| 5364 | region of text from the old buffer into the new buffer; and then using |
| 5365 | @code{save-excursion}, it brings you back to your original buffer. |
| 5366 | |
| 5367 | In looking at @code{append-to-buffer}, you have explored a fairly |
| 5368 | complex function. It shows how to use @code{let} and |
| 5369 | @code{save-excursion}, and how to change to and come back from another |
| 5370 | buffer. Many function definitions use @code{let}, |
| 5371 | @code{save-excursion}, and @code{set-buffer} this way. |
| 5372 | |
| 5373 | @node Buffer Related Review |
| 5374 | @section Review |
| 5375 | |
| 5376 | Here is a brief summary of the various functions discussed in this chapter. |
| 5377 | |
| 5378 | @table @code |
| 5379 | @item describe-function |
| 5380 | @itemx describe-variable |
| 5381 | Print the documentation for a function or variable. |
| 5382 | Conventionally bound to @kbd{C-h f} and @kbd{C-h v}. |
| 5383 | |
| 5384 | @item find-tag |
| 5385 | Find the file containing the source for a function or variable and |
| 5386 | switch buffers to it, positioning point at the beginning of the item. |
| 5387 | Conventionally bound to @kbd{M-.} (that's a period following the |
| 5388 | @key{META} key). |
| 5389 | |
| 5390 | @item save-excursion |
| 5391 | Save the location of point and mark and restore their values after the |
| 5392 | arguments to @code{save-excursion} have been evaluated. Also, remember |
| 5393 | the current buffer and return to it. |
| 5394 | |
| 5395 | @item push-mark |
| 5396 | Set mark at a location and record the value of the previous mark on the |
| 5397 | mark ring. The mark is a location in the buffer that will keep its |
| 5398 | relative position even if text is added to or removed from the buffer. |
| 5399 | |
| 5400 | @item goto-char |
| 5401 | Set point to the location specified by the value of the argument, which |
| 5402 | can be a number, a marker, or an expression that returns the number of |
| 5403 | a position, such as @code{(point-min)}. |
| 5404 | |
| 5405 | @item insert-buffer-substring |
| 5406 | Copy a region of text from a buffer that is passed to the function as |
| 5407 | an argument and insert the region into the current buffer. |
| 5408 | |
| 5409 | @item mark-whole-buffer |
| 5410 | Mark the whole buffer as a region. Normally bound to @kbd{C-x h}. |
| 5411 | |
| 5412 | @item set-buffer |
| 5413 | Switch the attention of Emacs to another buffer, but do not change the |
| 5414 | window being displayed. Used when the program rather than a human is |
| 5415 | to work on a different buffer. |
| 5416 | |
| 5417 | @item get-buffer-create |
| 5418 | @itemx get-buffer |
| 5419 | Find a named buffer or create one if a buffer of that name does not |
| 5420 | exist. The @code{get-buffer} function returns @code{nil} if the named |
| 5421 | buffer does not exist. |
| 5422 | @end table |
| 5423 | |
| 5424 | @need 1500 |
| 5425 | @node Buffer Exercises |
| 5426 | @section Exercises |
| 5427 | |
| 5428 | @itemize @bullet |
| 5429 | @item |
| 5430 | Write your own @code{simplified-end-of-buffer} function definition; |
| 5431 | then test it to see whether it works. |
| 5432 | |
| 5433 | @item |
| 5434 | Use @code{if} and @code{get-buffer} to write a function that prints a |
| 5435 | message telling you whether a buffer exists. |
| 5436 | |
| 5437 | @item |
| 5438 | Using @code{find-tag}, find the source for the @code{copy-to-buffer} |
| 5439 | function. |
| 5440 | @end itemize |
| 5441 | |
| 5442 | @node More Complex |
| 5443 | @chapter A Few More Complex Functions |
| 5444 | |
| 5445 | In this chapter, we build on what we have learned in previous chapters |
| 5446 | by looking at more complex functions. The @code{copy-to-buffer} |
| 5447 | function illustrates use of two @code{save-excursion} expressions in |
| 5448 | one definition, while the @code{insert-buffer} function illustrates |
| 5449 | use of an asterisk in an @code{interactive} expression, use of |
| 5450 | @code{or}, and the important distinction between a name and the object |
| 5451 | to which the name refers. |
| 5452 | |
| 5453 | @menu |
| 5454 | * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}. |
| 5455 | * insert-buffer:: Read-only, and with @code{or}. |
| 5456 | * beginning-of-buffer:: Shows @code{goto-char}, |
| 5457 | @code{point-min}, and @code{push-mark}. |
| 5458 | * Second Buffer Related Review:: |
| 5459 | * optional Exercise:: |
| 5460 | @end menu |
| 5461 | |
| 5462 | @node copy-to-buffer |
| 5463 | @section The Definition of @code{copy-to-buffer} |
| 5464 | @findex copy-to-buffer |
| 5465 | |
| 5466 | After understanding how @code{append-to-buffer} works, it is easy to |
| 5467 | understand @code{copy-to-buffer}. This function copies text into a |
| 5468 | buffer, but instead of adding to the second buffer, it replaces all the |
| 5469 | previous text in the second buffer. |
| 5470 | |
| 5471 | @need 800 |
| 5472 | The body of @code{copy-to-buffer} looks like this, |
| 5473 | |
| 5474 | @smallexample |
| 5475 | @group |
| 5476 | @dots{} |
| 5477 | (interactive "BCopy to buffer: \nr") |
| 5478 | (let ((oldbuf (current-buffer))) |
| 5479 | (with-current-buffer (get-buffer-create buffer) |
| 5480 | (barf-if-buffer-read-only) |
| 5481 | (erase-buffer) |
| 5482 | (save-excursion |
| 5483 | (insert-buffer-substring oldbuf start end))))) |
| 5484 | @end group |
| 5485 | @end smallexample |
| 5486 | |
| 5487 | The @code{copy-to-buffer} function has a simpler @code{interactive} |
| 5488 | expression than @code{append-to-buffer}. |
| 5489 | |
| 5490 | @need 800 |
| 5491 | The definition then says |
| 5492 | |
| 5493 | @smallexample |
| 5494 | (with-current-buffer (get-buffer-create buffer) @dots{} |
| 5495 | @end smallexample |
| 5496 | |
| 5497 | First, look at the earliest inner expression; that is evaluated first. |
| 5498 | That expression starts with @code{get-buffer-create buffer}. The |
| 5499 | function tells the computer to use the buffer with the name specified |
| 5500 | as the one to which you are copying, or if such a buffer does not |
| 5501 | exist, to create it. Then, the @code{with-current-buffer} function |
| 5502 | evaluates its body with that buffer temporarily current. |
| 5503 | |
| 5504 | (This demonstrates another way to shift the computer's attention but |
| 5505 | not the user's. The @code{append-to-buffer} function showed how to do |
| 5506 | the same with @code{save-excursion} and @code{set-buffer}. |
| 5507 | @code{with-current-buffer} is a newer, and arguably easier, |
| 5508 | mechanism.) |
| 5509 | |
| 5510 | The @code{barf-if-buffer-read-only} function sends you an error |
| 5511 | message saying the buffer is read-only if you cannot modify it. |
| 5512 | |
| 5513 | The next line has the @code{erase-buffer} function as its sole |
| 5514 | contents. That function erases the buffer. |
| 5515 | |
| 5516 | Finally, the last two lines contain the @code{save-excursion} |
| 5517 | expression with @code{insert-buffer-substring} as its body. |
| 5518 | The @code{insert-buffer-substring} expression copies the text from |
| 5519 | the buffer you are in (and you have not seen the computer shift its |
| 5520 | attention, so you don't know that that buffer is now called |
| 5521 | @code{oldbuf}). |
| 5522 | |
| 5523 | Incidentally, this is what is meant by `replacement'. To replace text, |
| 5524 | Emacs erases the previous text and then inserts new text. |
| 5525 | |
| 5526 | @need 1250 |
| 5527 | In outline, the body of @code{copy-to-buffer} looks like this: |
| 5528 | |
| 5529 | @smallexample |
| 5530 | @group |
| 5531 | (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer}) |
| 5532 | (@var{with-the-buffer-you-are-copying-to} |
| 5533 | (@var{but-do-not-erase-or-copy-to-a-read-only-buffer}) |
| 5534 | (erase-buffer) |
| 5535 | (save-excursion |
| 5536 | @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer}))) |
| 5537 | @end group |
| 5538 | @end smallexample |
| 5539 | |
| 5540 | @node insert-buffer |
| 5541 | @section The Definition of @code{insert-buffer} |
| 5542 | @findex insert-buffer |
| 5543 | |
| 5544 | @code{insert-buffer} is yet another buffer-related function. This |
| 5545 | command copies another buffer @emph{into} the current buffer. It is the |
| 5546 | reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they |
| 5547 | copy a region of text @emph{from} the current buffer to another buffer. |
| 5548 | |
| 5549 | Here is a discussion based on the original code. The code was |
| 5550 | simplified in 2003 and is harder to understand. |
| 5551 | |
| 5552 | (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see |
| 5553 | a discussion of the new body.) |
| 5554 | |
| 5555 | In addition, this code illustrates the use of @code{interactive} with a |
| 5556 | buffer that might be @dfn{read-only} and the important distinction |
| 5557 | between the name of an object and the object actually referred to. |
| 5558 | |
| 5559 | @menu |
| 5560 | * insert-buffer code:: |
| 5561 | * insert-buffer interactive:: When you can read, but not write. |
| 5562 | * insert-buffer body:: The body has an @code{or} and a @code{let}. |
| 5563 | * if & or:: Using an @code{if} instead of an @code{or}. |
| 5564 | * Insert or:: How the @code{or} expression works. |
| 5565 | * Insert let:: Two @code{save-excursion} expressions. |
| 5566 | * New insert-buffer:: |
| 5567 | @end menu |
| 5568 | |
| 5569 | @ifnottex |
| 5570 | @node insert-buffer code |
| 5571 | @unnumberedsubsec The Code for @code{insert-buffer} |
| 5572 | @end ifnottex |
| 5573 | |
| 5574 | @need 800 |
| 5575 | Here is the earlier code: |
| 5576 | |
| 5577 | @smallexample |
| 5578 | @group |
| 5579 | (defun insert-buffer (buffer) |
| 5580 | "Insert after point the contents of BUFFER. |
| 5581 | Puts mark after the inserted text. |
| 5582 | BUFFER may be a buffer or a buffer name." |
| 5583 | (interactive "*bInsert buffer:@: ") |
| 5584 | @end group |
| 5585 | @group |
| 5586 | (or (bufferp buffer) |
| 5587 | (setq buffer (get-buffer buffer))) |
| 5588 | (let (start end newmark) |
| 5589 | (save-excursion |
| 5590 | (save-excursion |
| 5591 | (set-buffer buffer) |
| 5592 | (setq start (point-min) end (point-max))) |
| 5593 | @end group |
| 5594 | @group |
| 5595 | (insert-buffer-substring buffer start end) |
| 5596 | (setq newmark (point))) |
| 5597 | (push-mark newmark))) |
| 5598 | @end group |
| 5599 | @end smallexample |
| 5600 | |
| 5601 | @need 1200 |
| 5602 | As with other function definitions, you can use a template to see an |
| 5603 | outline of the function: |
| 5604 | |
| 5605 | @smallexample |
| 5606 | @group |
| 5607 | (defun insert-buffer (buffer) |
| 5608 | "@var{documentation}@dots{}" |
| 5609 | (interactive "*bInsert buffer:@: ") |
| 5610 | @var{body}@dots{}) |
| 5611 | @end group |
| 5612 | @end smallexample |
| 5613 | |
| 5614 | @node insert-buffer interactive |
| 5615 | @subsection The Interactive Expression in @code{insert-buffer} |
| 5616 | @findex interactive, @r{example use of} |
| 5617 | |
| 5618 | In @code{insert-buffer}, the argument to the @code{interactive} |
| 5619 | declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert |
| 5620 | buffer:@: }. |
| 5621 | |
| 5622 | @menu |
| 5623 | * Read-only buffer:: When a buffer cannot be modified. |
| 5624 | * b for interactive:: An existing buffer or else its name. |
| 5625 | @end menu |
| 5626 | |
| 5627 | @node Read-only buffer |
| 5628 | @unnumberedsubsubsec A Read-only Buffer |
| 5629 | @cindex Read-only buffer |
| 5630 | @cindex Asterisk for read-only buffer |
| 5631 | @findex * @r{for read-only buffer} |
| 5632 | |
| 5633 | The asterisk is for the situation when the current buffer is a |
| 5634 | read-only buffer---a buffer that cannot be modified. If |
| 5635 | @code{insert-buffer} is called when the current buffer is read-only, a |
| 5636 | message to this effect is printed in the echo area and the terminal |
| 5637 | may beep or blink at you; you will not be permitted to insert anything |
| 5638 | into current buffer. The asterisk does not need to be followed by a |
| 5639 | newline to separate it from the next argument. |
| 5640 | |
| 5641 | @node b for interactive |
| 5642 | @unnumberedsubsubsec @samp{b} in an Interactive Expression |
| 5643 | |
| 5644 | The next argument in the interactive expression starts with a lower |
| 5645 | case @samp{b}. (This is different from the code for |
| 5646 | @code{append-to-buffer}, which uses an upper-case @samp{B}. |
| 5647 | @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.) |
| 5648 | The lower-case @samp{b} tells the Lisp interpreter that the argument |
| 5649 | for @code{insert-buffer} should be an existing buffer or else its |
| 5650 | name. (The upper-case @samp{B} option provides for the possibility |
| 5651 | that the buffer does not exist.) Emacs will prompt you for the name |
| 5652 | of the buffer, offering you a default buffer, with name completion |
| 5653 | enabled. If the buffer does not exist, you receive a message that |
| 5654 | says ``No match''; your terminal may beep at you as well. |
| 5655 | |
| 5656 | The new and simplified code generates a list for @code{interactive}. |
| 5657 | It uses the @code{barf-if-buffer-read-only} and @code{read-buffer} |
| 5658 | functions with which we are already familiar and the @code{progn} |
| 5659 | special form with which we are not. (It will be described later.) |
| 5660 | |
| 5661 | @node insert-buffer body |
| 5662 | @subsection The Body of the @code{insert-buffer} Function |
| 5663 | |
| 5664 | The body of the @code{insert-buffer} function has two major parts: an |
| 5665 | @code{or} expression and a @code{let} expression. The purpose of the |
| 5666 | @code{or} expression is to ensure that the argument @code{buffer} is |
| 5667 | bound to a buffer and not just the name of a buffer. The body of the |
| 5668 | @code{let} expression contains the code which copies the other buffer |
| 5669 | into the current buffer. |
| 5670 | |
| 5671 | @need 1250 |
| 5672 | In outline, the two expressions fit into the @code{insert-buffer} |
| 5673 | function like this: |
| 5674 | |
| 5675 | @smallexample |
| 5676 | @group |
| 5677 | (defun insert-buffer (buffer) |
| 5678 | "@var{documentation}@dots{}" |
| 5679 | (interactive "*bInsert buffer:@: ") |
| 5680 | (or @dots{} |
| 5681 | @dots{} |
| 5682 | @end group |
| 5683 | @group |
| 5684 | (let (@var{varlist}) |
| 5685 | @var{body-of-}@code{let}@dots{} ) |
| 5686 | @end group |
| 5687 | @end smallexample |
| 5688 | |
| 5689 | To understand how the @code{or} expression ensures that the argument |
| 5690 | @code{buffer} is bound to a buffer and not to the name of a buffer, it |
| 5691 | is first necessary to understand the @code{or} function. |
| 5692 | |
| 5693 | Before doing this, let me rewrite this part of the function using |
| 5694 | @code{if} so that you can see what is done in a manner that will be familiar. |
| 5695 | |
| 5696 | @node if & or |
| 5697 | @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or} |
| 5698 | |
| 5699 | The job to be done is to make sure the value of @code{buffer} is a |
| 5700 | buffer itself and not the name of a buffer. If the value is the name, |
| 5701 | then the buffer itself must be got. |
| 5702 | |
| 5703 | You can imagine yourself at a conference where an usher is wandering |
| 5704 | around holding a list with your name on it and looking for you: the |
| 5705 | usher is ``bound'' to your name, not to you; but when the usher finds |
| 5706 | you and takes your arm, the usher becomes ``bound'' to you. |
| 5707 | |
| 5708 | @need 800 |
| 5709 | In Lisp, you might describe this situation like this: |
| 5710 | |
| 5711 | @smallexample |
| 5712 | @group |
| 5713 | (if (not (holding-on-to-guest)) |
| 5714 | (find-and-take-arm-of-guest)) |
| 5715 | @end group |
| 5716 | @end smallexample |
| 5717 | |
| 5718 | We want to do the same thing with a buffer---if we do not have the |
| 5719 | buffer itself, we want to get it. |
| 5720 | |
| 5721 | @need 1200 |
| 5722 | Using a predicate called @code{bufferp} that tells us whether we have a |
| 5723 | buffer (rather than its name), we can write the code like this: |
| 5724 | |
| 5725 | @smallexample |
| 5726 | @group |
| 5727 | (if (not (bufferp buffer)) ; @r{if-part} |
| 5728 | (setq buffer (get-buffer buffer))) ; @r{then-part} |
| 5729 | @end group |
| 5730 | @end smallexample |
| 5731 | |
| 5732 | @noindent |
| 5733 | Here, the true-or-false-test of the @code{if} expression is |
| 5734 | @w{@code{(not (bufferp buffer))}}; and the then-part is the expression |
| 5735 | @w{@code{(setq buffer (get-buffer buffer))}}. |
| 5736 | |
| 5737 | In the test, the function @code{bufferp} returns true if its argument is |
| 5738 | a buffer---but false if its argument is the name of the buffer. (The |
| 5739 | last character of the function name @code{bufferp} is the character |
| 5740 | @samp{p}; as we saw earlier, such use of @samp{p} is a convention that |
| 5741 | indicates that the function is a predicate, which is a term that means |
| 5742 | that the function will determine whether some property is true or false. |
| 5743 | @xref{Wrong Type of Argument, , Using the Wrong Type Object as an |
| 5744 | Argument}.) |
| 5745 | |
| 5746 | @need 1200 |
| 5747 | The function @code{not} precedes the expression @code{(bufferp buffer)}, |
| 5748 | so the true-or-false-test looks like this: |
| 5749 | |
| 5750 | @smallexample |
| 5751 | (not (bufferp buffer)) |
| 5752 | @end smallexample |
| 5753 | |
| 5754 | @noindent |
| 5755 | @code{not} is a function that returns true if its argument is false |
| 5756 | and false if its argument is true. So if @code{(bufferp buffer)} |
| 5757 | returns true, the @code{not} expression returns false and vice-verse: |
| 5758 | what is ``not true'' is false and what is ``not false'' is true. |
| 5759 | |
| 5760 | Using this test, the @code{if} expression works as follows: when the |
| 5761 | value of the variable @code{buffer} is actually a buffer rather than |
| 5762 | its name, the true-or-false-test returns false and the @code{if} |
| 5763 | expression does not evaluate the then-part. This is fine, since we do |
| 5764 | not need to do anything to the variable @code{buffer} if it really is |
| 5765 | a buffer. |
| 5766 | |
| 5767 | On the other hand, when the value of @code{buffer} is not a buffer |
| 5768 | itself, but the name of a buffer, the true-or-false-test returns true |
| 5769 | and the then-part of the expression is evaluated. In this case, the |
| 5770 | then-part is @code{(setq buffer (get-buffer buffer))}. This |
| 5771 | expression uses the @code{get-buffer} function to return an actual |
| 5772 | buffer itself, given its name. The @code{setq} then sets the variable |
| 5773 | @code{buffer} to the value of the buffer itself, replacing its previous |
| 5774 | value (which was the name of the buffer). |
| 5775 | |
| 5776 | @node Insert or |
| 5777 | @subsection The @code{or} in the Body |
| 5778 | |
| 5779 | The purpose of the @code{or} expression in the @code{insert-buffer} |
| 5780 | function is to ensure that the argument @code{buffer} is bound to a |
| 5781 | buffer and not just to the name of a buffer. The previous section shows |
| 5782 | how the job could have been done using an @code{if} expression. |
| 5783 | However, the @code{insert-buffer} function actually uses @code{or}. |
| 5784 | To understand this, it is necessary to understand how @code{or} works. |
| 5785 | |
| 5786 | @findex or |
| 5787 | An @code{or} function can have any number of arguments. It evaluates |
| 5788 | each argument in turn and returns the value of the first of its |
| 5789 | arguments that is not @code{nil}. Also, and this is a crucial feature |
| 5790 | of @code{or}, it does not evaluate any subsequent arguments after |
| 5791 | returning the first non-@code{nil} value. |
| 5792 | |
| 5793 | @need 800 |
| 5794 | The @code{or} expression looks like this: |
| 5795 | |
| 5796 | @smallexample |
| 5797 | @group |
| 5798 | (or (bufferp buffer) |
| 5799 | (setq buffer (get-buffer buffer))) |
| 5800 | @end group |
| 5801 | @end smallexample |
| 5802 | |
| 5803 | @noindent |
| 5804 | The first argument to @code{or} is the expression @code{(bufferp buffer)}. |
| 5805 | This expression returns true (a non-@code{nil} value) if the buffer is |
| 5806 | actually a buffer, and not just the name of a buffer. In the @code{or} |
| 5807 | expression, if this is the case, the @code{or} expression returns this |
| 5808 | true value and does not evaluate the next expression---and this is fine |
| 5809 | with us, since we do not want to do anything to the value of |
| 5810 | @code{buffer} if it really is a buffer. |
| 5811 | |
| 5812 | On the other hand, if the value of @code{(bufferp buffer)} is @code{nil}, |
| 5813 | which it will be if the value of @code{buffer} is the name of a buffer, |
| 5814 | the Lisp interpreter evaluates the next element of the @code{or} |
| 5815 | expression. This is the expression @code{(setq buffer (get-buffer |
| 5816 | buffer))}. This expression returns a non-@code{nil} value, which |
| 5817 | is the value to which it sets the variable @code{buffer}---and this |
| 5818 | value is a buffer itself, not the name of a buffer. |
| 5819 | |
| 5820 | The result of all this is that the symbol @code{buffer} is always |
| 5821 | bound to a buffer itself rather than to the name of a buffer. All |
| 5822 | this is necessary because the @code{set-buffer} function in a |
| 5823 | following line only works with a buffer itself, not with the name to a |
| 5824 | buffer. |
| 5825 | |
| 5826 | @need 1250 |
| 5827 | Incidentally, using @code{or}, the situation with the usher would be |
| 5828 | written like this: |
| 5829 | |
| 5830 | @smallexample |
| 5831 | (or (holding-on-to-guest) (find-and-take-arm-of-guest)) |
| 5832 | @end smallexample |
| 5833 | |
| 5834 | @node Insert let |
| 5835 | @subsection The @code{let} Expression in @code{insert-buffer} |
| 5836 | |
| 5837 | After ensuring that the variable @code{buffer} refers to a buffer itself |
| 5838 | and not just to the name of a buffer, the @code{insert-buffer function} |
| 5839 | continues with a @code{let} expression. This specifies three local |
| 5840 | variables, @code{start}, @code{end}, and @code{newmark} and binds them |
| 5841 | to the initial value @code{nil}. These variables are used inside the |
| 5842 | remainder of the @code{let} and temporarily hide any other occurrence of |
| 5843 | variables of the same name in Emacs until the end of the @code{let}. |
| 5844 | |
| 5845 | @need 1200 |
| 5846 | The body of the @code{let} contains two @code{save-excursion} |
| 5847 | expressions. First, we will look at the inner @code{save-excursion} |
| 5848 | expression in detail. The expression looks like this: |
| 5849 | |
| 5850 | @smallexample |
| 5851 | @group |
| 5852 | (save-excursion |
| 5853 | (set-buffer buffer) |
| 5854 | (setq start (point-min) end (point-max))) |
| 5855 | @end group |
| 5856 | @end smallexample |
| 5857 | |
| 5858 | @noindent |
| 5859 | The expression @code{(set-buffer buffer)} changes Emacs's attention |
| 5860 | from the current buffer to the one from which the text will copied. |
| 5861 | In that buffer, the variables @code{start} and @code{end} are set to |
| 5862 | the beginning and end of the buffer, using the commands |
| 5863 | @code{point-min} and @code{point-max}. Note that we have here an |
| 5864 | illustration of how @code{setq} is able to set two variables in the |
| 5865 | same expression. The first argument of @code{setq} is set to the |
| 5866 | value of its second, and its third argument is set to the value of its |
| 5867 | fourth. |
| 5868 | |
| 5869 | After the body of the inner @code{save-excursion} is evaluated, the |
| 5870 | @code{save-excursion} restores the original buffer, but @code{start} and |
| 5871 | @code{end} remain set to the values of the beginning and end of the |
| 5872 | buffer from which the text will be copied. |
| 5873 | |
| 5874 | @need 1250 |
| 5875 | The outer @code{save-excursion} expression looks like this: |
| 5876 | |
| 5877 | @smallexample |
| 5878 | @group |
| 5879 | (save-excursion |
| 5880 | (@var{inner-}@code{save-excursion}@var{-expression} |
| 5881 | (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end}) |
| 5882 | (insert-buffer-substring buffer start end) |
| 5883 | (setq newmark (point))) |
| 5884 | @end group |
| 5885 | @end smallexample |
| 5886 | |
| 5887 | @noindent |
| 5888 | The @code{insert-buffer-substring} function copies the text |
| 5889 | @emph{into} the current buffer @emph{from} the region indicated by |
| 5890 | @code{start} and @code{end} in @code{buffer}. Since the whole of the |
| 5891 | second buffer lies between @code{start} and @code{end}, the whole of |
| 5892 | the second buffer is copied into the buffer you are editing. Next, |
| 5893 | the value of point, which will be at the end of the inserted text, is |
| 5894 | recorded in the variable @code{newmark}. |
| 5895 | |
| 5896 | After the body of the outer @code{save-excursion} is evaluated, point |
| 5897 | and mark are relocated to their original places. |
| 5898 | |
| 5899 | However, it is convenient to locate a mark at the end of the newly |
| 5900 | inserted text and locate point at its beginning. The @code{newmark} |
| 5901 | variable records the end of the inserted text. In the last line of |
| 5902 | the @code{let} expression, the @code{(push-mark newmark)} expression |
| 5903 | function sets a mark to this location. (The previous location of the |
| 5904 | mark is still accessible; it is recorded on the mark ring and you can |
| 5905 | go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is |
| 5906 | located at the beginning of the inserted text, which is where it was |
| 5907 | before you called the insert function, the position of which was saved |
| 5908 | by the first @code{save-excursion}. |
| 5909 | |
| 5910 | @need 1250 |
| 5911 | The whole @code{let} expression looks like this: |
| 5912 | |
| 5913 | @smallexample |
| 5914 | @group |
| 5915 | (let (start end newmark) |
| 5916 | (save-excursion |
| 5917 | (save-excursion |
| 5918 | (set-buffer buffer) |
| 5919 | (setq start (point-min) end (point-max))) |
| 5920 | (insert-buffer-substring buffer start end) |
| 5921 | (setq newmark (point))) |
| 5922 | (push-mark newmark)) |
| 5923 | @end group |
| 5924 | @end smallexample |
| 5925 | |
| 5926 | Like the @code{append-to-buffer} function, the @code{insert-buffer} |
| 5927 | function uses @code{let}, @code{save-excursion}, and |
| 5928 | @code{set-buffer}. In addition, the function illustrates one way to |
| 5929 | use @code{or}. All these functions are building blocks that we will |
| 5930 | find and use again and again. |
| 5931 | |
| 5932 | @node New insert-buffer |
| 5933 | @subsection New Body for @code{insert-buffer} |
| 5934 | @findex insert-buffer, new version body |
| 5935 | @findex new version body for insert-buffer |
| 5936 | |
| 5937 | The body in the GNU Emacs 22 version is more confusing than the original. |
| 5938 | |
| 5939 | @need 1250 |
| 5940 | It consists of two expressions, |
| 5941 | |
| 5942 | @smallexample |
| 5943 | @group |
| 5944 | (push-mark |
| 5945 | (save-excursion |
| 5946 | (insert-buffer-substring (get-buffer buffer)) |
| 5947 | (point))) |
| 5948 | |
| 5949 | nil |
| 5950 | @end group |
| 5951 | @end smallexample |
| 5952 | |
| 5953 | @noindent |
| 5954 | except, and this is what confuses novices, very important work is done |
| 5955 | inside the @code{push-mark} expression. |
| 5956 | |
| 5957 | The @code{get-buffer} function returns a buffer with the name |
| 5958 | provided. You will note that the function is @emph{not} called |
| 5959 | @code{get-buffer-create}; it does not create a buffer if one does not |
| 5960 | already exist. The buffer returned by @code{get-buffer}, an existing |
| 5961 | buffer, is passed to @code{insert-buffer-substring}, which inserts the |
| 5962 | whole of the buffer (since you did not specify anything else). |
| 5963 | |
| 5964 | The location into which the buffer is inserted is recorded by |
| 5965 | @code{push-mark}. Then the function returns @code{nil}, the value of |
| 5966 | its last command. Put another way, the @code{insert-buffer} function |
| 5967 | exists only to produce a side effect, inserting another buffer, not to |
| 5968 | return any value. |
| 5969 | |
| 5970 | @node beginning-of-buffer |
| 5971 | @section Complete Definition of @code{beginning-of-buffer} |
| 5972 | @findex beginning-of-buffer |
| 5973 | |
| 5974 | The basic structure of the @code{beginning-of-buffer} function has |
| 5975 | already been discussed. (@xref{simplified-beginning-of-buffer, , A |
| 5976 | Simplified @code{beginning-of-buffer} Definition}.) |
| 5977 | This section describes the complex part of the definition. |
| 5978 | |
| 5979 | As previously described, when invoked without an argument, |
| 5980 | @code{beginning-of-buffer} moves the cursor to the beginning of the |
| 5981 | buffer (in truth, the beginning of the accessible portion of the |
| 5982 | buffer), leaving the mark at the previous position. However, when the |
| 5983 | command is invoked with a number between one and ten, the function |
| 5984 | considers that number to be a fraction of the length of the buffer, |
| 5985 | measured in tenths, and Emacs moves the cursor that fraction of the |
| 5986 | way from the beginning of the buffer. Thus, you can either call this |
| 5987 | function with the key command @kbd{M-<}, which will move the cursor to |
| 5988 | the beginning of the buffer, or with a key command such as @kbd{C-u 7 |
| 5989 | M-<} which will move the cursor to a point 70% of the way through the |
| 5990 | buffer. If a number bigger than ten is used for the argument, it |
| 5991 | moves to the end of the buffer. |
| 5992 | |
| 5993 | The @code{beginning-of-buffer} function can be called with or without an |
| 5994 | argument. The use of the argument is optional. |
| 5995 | |
| 5996 | @menu |
| 5997 | * Optional Arguments:: |
| 5998 | * beginning-of-buffer opt arg:: Example with optional argument. |
| 5999 | * beginning-of-buffer complete:: |
| 6000 | @end menu |
| 6001 | |
| 6002 | @node Optional Arguments |
| 6003 | @subsection Optional Arguments |
| 6004 | |
| 6005 | Unless told otherwise, Lisp expects that a function with an argument in |
| 6006 | its function definition will be called with a value for that argument. |
| 6007 | If that does not happen, you get an error and a message that says |
| 6008 | @samp{Wrong number of arguments}. |
| 6009 | |
| 6010 | @cindex Optional arguments |
| 6011 | @cindex Keyword |
| 6012 | @findex optional |
| 6013 | However, optional arguments are a feature of Lisp: a particular |
| 6014 | @dfn{keyword} is used to tell the Lisp interpreter that an argument is |
| 6015 | optional. The keyword is @code{&optional}. (The @samp{&} in front of |
| 6016 | @samp{optional} is part of the keyword.) In a function definition, if |
| 6017 | an argument follows the keyword @code{&optional}, no value need be |
| 6018 | passed to that argument when the function is called. |
| 6019 | |
| 6020 | @need 1200 |
| 6021 | The first line of the function definition of @code{beginning-of-buffer} |
| 6022 | therefore looks like this: |
| 6023 | |
| 6024 | @smallexample |
| 6025 | (defun beginning-of-buffer (&optional arg) |
| 6026 | @end smallexample |
| 6027 | |
| 6028 | @need 1250 |
| 6029 | In outline, the whole function looks like this: |
| 6030 | |
| 6031 | @smallexample |
| 6032 | @group |
| 6033 | (defun beginning-of-buffer (&optional arg) |
| 6034 | "@var{documentation}@dots{}" |
| 6035 | (interactive "P") |
| 6036 | (or (@var{is-the-argument-a-cons-cell} arg) |
| 6037 | (and @var{are-both-transient-mark-mode-and-mark-active-true}) |
| 6038 | (push-mark)) |
| 6039 | (let (@var{determine-size-and-set-it}) |
| 6040 | (goto-char |
| 6041 | (@var{if-there-is-an-argument} |
| 6042 | @var{figure-out-where-to-go} |
| 6043 | @var{else-go-to} |
| 6044 | (point-min)))) |
| 6045 | @var{do-nicety} |
| 6046 | @end group |
| 6047 | @end smallexample |
| 6048 | |
| 6049 | The function is similar to the @code{simplified-beginning-of-buffer} |
| 6050 | function except that the @code{interactive} expression has @code{"P"} |
| 6051 | as an argument and the @code{goto-char} function is followed by an |
| 6052 | if-then-else expression that figures out where to put the cursor if |
| 6053 | there is an argument that is not a cons cell. |
| 6054 | |
| 6055 | (Since I do not explain a cons cell for many more chapters, please |
| 6056 | consider ignoring the function @code{consp}. @xref{List |
| 6057 | Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type, |
| 6058 | , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference |
| 6059 | Manual}.) |
| 6060 | |
| 6061 | The @code{"P"} in the @code{interactive} expression tells Emacs to |
| 6062 | pass a prefix argument, if there is one, to the function in raw form. |
| 6063 | A prefix argument is made by typing the @key{META} key followed by a |
| 6064 | number, or by typing @kbd{C-u} and then a number. (If you don't type |
| 6065 | a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase |
| 6066 | @code{"p"} in the @code{interactive} expression causes the function to |
| 6067 | convert a prefix arg to a number.) |
| 6068 | |
| 6069 | The true-or-false-test of the @code{if} expression looks complex, but |
| 6070 | it is not: it checks whether @code{arg} has a value that is not |
| 6071 | @code{nil} and whether it is a cons cell. (That is what @code{consp} |
| 6072 | does; it checks whether its argument is a cons cell.) If @code{arg} |
| 6073 | has a value that is not @code{nil} (and is not a cons cell), which |
| 6074 | will be the case if @code{beginning-of-buffer} is called with a |
| 6075 | numeric argument, then this true-or-false-test will return true and |
| 6076 | the then-part of the @code{if} expression will be evaluated. On the |
| 6077 | other hand, if @code{beginning-of-buffer} is not called with an |
| 6078 | argument, the value of @code{arg} will be @code{nil} and the else-part |
| 6079 | of the @code{if} expression will be evaluated. The else-part is |
| 6080 | simply @code{point-min}, and when this is the outcome, the whole |
| 6081 | @code{goto-char} expression is @code{(goto-char (point-min))}, which |
| 6082 | is how we saw the @code{beginning-of-buffer} function in its |
| 6083 | simplified form. |
| 6084 | |
| 6085 | @node beginning-of-buffer opt arg |
| 6086 | @subsection @code{beginning-of-buffer} with an Argument |
| 6087 | |
| 6088 | When @code{beginning-of-buffer} is called with an argument, an |
| 6089 | expression is evaluated which calculates what value to pass to |
| 6090 | @code{goto-char}. This expression is rather complicated at first sight. |
| 6091 | It includes an inner @code{if} expression and much arithmetic. It looks |
| 6092 | like this: |
| 6093 | |
| 6094 | @smallexample |
| 6095 | @group |
| 6096 | (if (> (buffer-size) 10000) |
| 6097 | ;; @r{Avoid overflow for large buffer sizes!} |
| 6098 | (* (prefix-numeric-value arg) |
| 6099 | (/ size 10)) |
| 6100 | (/ |
| 6101 | (+ 10 |
| 6102 | (* |
| 6103 | size (prefix-numeric-value arg))) 10))) |
| 6104 | @end group |
| 6105 | @end smallexample |
| 6106 | |
| 6107 | @menu |
| 6108 | * Disentangle beginning-of-buffer:: |
| 6109 | * Large buffer case:: |
| 6110 | * Small buffer case:: |
| 6111 | @end menu |
| 6112 | |
| 6113 | @ifnottex |
| 6114 | @node Disentangle beginning-of-buffer |
| 6115 | @unnumberedsubsubsec Disentangle @code{beginning-of-buffer} |
| 6116 | @end ifnottex |
| 6117 | |
| 6118 | Like other complex-looking expressions, the conditional expression |
| 6119 | within @code{beginning-of-buffer} can be disentangled by looking at it |
| 6120 | as parts of a template, in this case, the template for an if-then-else |
| 6121 | expression. In skeletal form, the expression looks like this: |
| 6122 | |
| 6123 | @smallexample |
| 6124 | @group |
| 6125 | (if (@var{buffer-is-large} |
| 6126 | @var{divide-buffer-size-by-10-and-multiply-by-arg} |
| 6127 | @var{else-use-alternate-calculation} |
| 6128 | @end group |
| 6129 | @end smallexample |
| 6130 | |
| 6131 | The true-or-false-test of this inner @code{if} expression checks the |
| 6132 | size of the buffer. The reason for this is that the old version 18 |
| 6133 | Emacs used numbers that are no bigger than eight million or so and in |
| 6134 | the computation that followed, the programmer feared that Emacs might |
| 6135 | try to use over-large numbers if the buffer were large. The term |
| 6136 | `overflow', mentioned in the comment, means numbers that are over |
| 6137 | large. More recent versions of Emacs use larger numbers, but this |
| 6138 | code has not been touched, if only because people now look at buffers |
| 6139 | that are far, far larger than ever before. |
| 6140 | |
| 6141 | There are two cases: if the buffer is large and if it is not. |
| 6142 | |
| 6143 | @node Large buffer case |
| 6144 | @unnumberedsubsubsec What happens in a large buffer |
| 6145 | |
| 6146 | In @code{beginning-of-buffer}, the inner @code{if} expression tests |
| 6147 | whether the size of the buffer is greater than 10,000 characters. To do |
| 6148 | this, it uses the @code{>} function and the computation of @code{size} |
| 6149 | that comes from the let expression. |
| 6150 | |
| 6151 | In the old days, the function @code{buffer-size} was used. Not only |
| 6152 | was that function called several times, it gave the size of the whole |
| 6153 | buffer, not the accessible part. The computation makes much more |
| 6154 | sense when it handles just the accessible part. (@xref{Narrowing & |
| 6155 | Widening, , Narrowing and Widening}, for more information on focusing |
| 6156 | attention to an `accessible' part.) |
| 6157 | |
| 6158 | @need 800 |
| 6159 | The line looks like this: |
| 6160 | |
| 6161 | @smallexample |
| 6162 | (if (> size 10000) |
| 6163 | @end smallexample |
| 6164 | |
| 6165 | @need 1200 |
| 6166 | @noindent |
| 6167 | When the buffer is large, the then-part of the @code{if} expression is |
| 6168 | evaluated. It reads like this (after formatting for easy reading): |
| 6169 | |
| 6170 | @smallexample |
| 6171 | @group |
| 6172 | (* |
| 6173 | (prefix-numeric-value arg) |
| 6174 | (/ size 10)) |
| 6175 | @end group |
| 6176 | @end smallexample |
| 6177 | |
| 6178 | @noindent |
| 6179 | This expression is a multiplication, with two arguments to the function |
| 6180 | @code{*}. |
| 6181 | |
| 6182 | The first argument is @code{(prefix-numeric-value arg)}. When |
| 6183 | @code{"P"} is used as the argument for @code{interactive}, the value |
| 6184 | passed to the function as its argument is passed a ``raw prefix |
| 6185 | argument'', and not a number. (It is a number in a list.) To perform |
| 6186 | the arithmetic, a conversion is necessary, and |
| 6187 | @code{prefix-numeric-value} does the job. |
| 6188 | |
| 6189 | @findex / @r{(division)} |
| 6190 | @cindex Division |
| 6191 | The second argument is @code{(/ size 10)}. This expression divides |
| 6192 | the numeric value by ten---the numeric value of the size of the |
| 6193 | accessible portion of the buffer. This produces a number that tells |
| 6194 | how many characters make up one tenth of the buffer size. (In Lisp, |
| 6195 | @code{/} is used for division, just as @code{*} is used for |
| 6196 | multiplication.) |
| 6197 | |
| 6198 | @need 1200 |
| 6199 | In the multiplication expression as a whole, this amount is multiplied |
| 6200 | by the value of the prefix argument---the multiplication looks like this: |
| 6201 | |
| 6202 | @smallexample |
| 6203 | @group |
| 6204 | (* @var{numeric-value-of-prefix-arg} |
| 6205 | @var{number-of-characters-in-one-tenth-of-the-accessible-buffer}) |
| 6206 | @end group |
| 6207 | @end smallexample |
| 6208 | |
| 6209 | @noindent |
| 6210 | If, for example, the prefix argument is @samp{7}, the one-tenth value |
| 6211 | will be multiplied by 7 to give a position 70% of the way through. |
| 6212 | |
| 6213 | @need 1200 |
| 6214 | The result of all this is that if the accessible portion of the buffer |
| 6215 | is large, the @code{goto-char} expression reads like this: |
| 6216 | |
| 6217 | @smallexample |
| 6218 | @group |
| 6219 | (goto-char (* (prefix-numeric-value arg) |
| 6220 | (/ size 10))) |
| 6221 | @end group |
| 6222 | @end smallexample |
| 6223 | |
| 6224 | This puts the cursor where we want it. |
| 6225 | |
| 6226 | @node Small buffer case |
| 6227 | @unnumberedsubsubsec What happens in a small buffer |
| 6228 | |
| 6229 | If the buffer contains fewer than 10,000 characters, a slightly |
| 6230 | different computation is performed. You might think this is not |
| 6231 | necessary, since the first computation could do the job. However, in |
| 6232 | a small buffer, the first method may not put the cursor on exactly the |
| 6233 | desired line; the second method does a better job. |
| 6234 | |
| 6235 | @need 800 |
| 6236 | The code looks like this: |
| 6237 | |
| 6238 | @c Keep this on one line. |
| 6239 | @smallexample |
| 6240 | (/ (+ 10 (* size (prefix-numeric-value arg))) 10)) |
| 6241 | @end smallexample |
| 6242 | |
| 6243 | @need 1200 |
| 6244 | @noindent |
| 6245 | This is code in which you figure out what happens by discovering how the |
| 6246 | functions are embedded in parentheses. It is easier to read if you |
| 6247 | reformat it with each expression indented more deeply than its |
| 6248 | enclosing expression: |
| 6249 | |
| 6250 | @smallexample |
| 6251 | @group |
| 6252 | (/ |
| 6253 | (+ 10 |
| 6254 | (* |
| 6255 | size |
| 6256 | (prefix-numeric-value arg))) |
| 6257 | 10)) |
| 6258 | @end group |
| 6259 | @end smallexample |
| 6260 | |
| 6261 | @need 1200 |
| 6262 | @noindent |
| 6263 | Looking at parentheses, we see that the innermost operation is |
| 6264 | @code{(prefix-numeric-value arg)}, which converts the raw argument to |
| 6265 | a number. In the following expression, this number is multiplied by |
| 6266 | the size of the accessible portion of the buffer: |
| 6267 | |
| 6268 | @smallexample |
| 6269 | (* size (prefix-numeric-value arg)) |
| 6270 | @end smallexample |
| 6271 | |
| 6272 | @noindent |
| 6273 | This multiplication creates a number that may be larger than the size of |
| 6274 | the buffer---seven times larger if the argument is 7, for example. Ten |
| 6275 | is then added to this number and finally the large number is divided by |
| 6276 | ten to provide a value that is one character larger than the percentage |
| 6277 | position in the buffer. |
| 6278 | |
| 6279 | The number that results from all this is passed to @code{goto-char} and |
| 6280 | the cursor is moved to that point. |
| 6281 | |
| 6282 | @need 1500 |
| 6283 | @node beginning-of-buffer complete |
| 6284 | @subsection The Complete @code{beginning-of-buffer} |
| 6285 | |
| 6286 | @need 1000 |
| 6287 | Here is the complete text of the @code{beginning-of-buffer} function: |
| 6288 | @sp 1 |
| 6289 | |
| 6290 | @c In GNU Emacs 22 |
| 6291 | @smallexample |
| 6292 | @group |
| 6293 | (defun beginning-of-buffer (&optional arg) |
| 6294 | "Move point to the beginning of the buffer; |
| 6295 | leave mark at previous position. |
| 6296 | With \\[universal-argument] prefix, |
| 6297 | do not set mark at previous position. |
| 6298 | With numeric arg N, |
| 6299 | put point N/10 of the way from the beginning. |
| 6300 | |
| 6301 | If the buffer is narrowed, |
| 6302 | this command uses the beginning and size |
| 6303 | of the accessible part of the buffer. |
| 6304 | @end group |
| 6305 | |
| 6306 | @group |
| 6307 | Don't use this command in Lisp programs! |
| 6308 | \(goto-char (point-min)) is faster |
| 6309 | and avoids clobbering the mark." |
| 6310 | (interactive "P") |
| 6311 | (or (consp arg) |
| 6312 | (and transient-mark-mode mark-active) |
| 6313 | (push-mark)) |
| 6314 | @end group |
| 6315 | @group |
| 6316 | (let ((size (- (point-max) (point-min)))) |
| 6317 | (goto-char (if (and arg (not (consp arg))) |
| 6318 | (+ (point-min) |
| 6319 | (if (> size 10000) |
| 6320 | ;; Avoid overflow for large buffer sizes! |
| 6321 | (* (prefix-numeric-value arg) |
| 6322 | (/ size 10)) |
| 6323 | (/ (+ 10 (* size (prefix-numeric-value arg))) |
| 6324 | 10))) |
| 6325 | (point-min)))) |
| 6326 | (if arg (forward-line 1))) |
| 6327 | @end group |
| 6328 | @end smallexample |
| 6329 | |
| 6330 | @ignore |
| 6331 | From before GNU Emacs 22 |
| 6332 | @smallexample |
| 6333 | @group |
| 6334 | (defun beginning-of-buffer (&optional arg) |
| 6335 | "Move point to the beginning of the buffer; |
| 6336 | leave mark at previous position. |
| 6337 | With arg N, put point N/10 of the way |
| 6338 | from the true beginning. |
| 6339 | @end group |
| 6340 | @group |
| 6341 | Don't use this in Lisp programs! |
| 6342 | \(goto-char (point-min)) is faster |
| 6343 | and does not set the mark." |
| 6344 | (interactive "P") |
| 6345 | (push-mark) |
| 6346 | @end group |
| 6347 | @group |
| 6348 | (goto-char |
| 6349 | (if arg |
| 6350 | (if (> (buffer-size) 10000) |
| 6351 | ;; @r{Avoid overflow for large buffer sizes!} |
| 6352 | (* (prefix-numeric-value arg) |
| 6353 | (/ (buffer-size) 10)) |
| 6354 | @end group |
| 6355 | @group |
| 6356 | (/ (+ 10 (* (buffer-size) |
| 6357 | (prefix-numeric-value arg))) |
| 6358 | 10)) |
| 6359 | (point-min))) |
| 6360 | (if arg (forward-line 1))) |
| 6361 | @end group |
| 6362 | @end smallexample |
| 6363 | @end ignore |
| 6364 | |
| 6365 | @noindent |
| 6366 | Except for two small points, the previous discussion shows how this |
| 6367 | function works. The first point deals with a detail in the |
| 6368 | documentation string, and the second point concerns the last line of |
| 6369 | the function. |
| 6370 | |
| 6371 | @need 800 |
| 6372 | In the documentation string, there is reference to an expression: |
| 6373 | |
| 6374 | @smallexample |
| 6375 | \\[universal-argument] |
| 6376 | @end smallexample |
| 6377 | |
| 6378 | @noindent |
| 6379 | A @samp{\\} is used before the first square bracket of this |
| 6380 | expression. This @samp{\\} tells the Lisp interpreter to substitute |
| 6381 | whatever key is currently bound to the @samp{[@dots{}]}. In the case |
| 6382 | of @code{universal-argument}, that is usually @kbd{C-u}, but it might |
| 6383 | be different. (@xref{Documentation Tips, , Tips for Documentation |
| 6384 | Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more |
| 6385 | information.) |
| 6386 | |
| 6387 | @need 1200 |
| 6388 | Finally, the last line of the @code{beginning-of-buffer} command says |
| 6389 | to move point to the beginning of the next line if the command is |
| 6390 | invoked with an argument: |
| 6391 | |
| 6392 | @smallexample |
| 6393 | (if arg (forward-line 1))) |
| 6394 | @end smallexample |
| 6395 | |
| 6396 | @noindent |
| 6397 | This puts the cursor at the beginning of the first line after the |
| 6398 | appropriate tenths position in the buffer. This is a flourish that |
| 6399 | means that the cursor is always located @emph{at least} the requested |
| 6400 | tenths of the way through the buffer, which is a nicety that is, |
| 6401 | perhaps, not necessary, but which, if it did not occur, would be sure |
| 6402 | to draw complaints. |
| 6403 | |
| 6404 | On the other hand, it also means that if you specify the command with |
| 6405 | a @kbd{C-u}, but without a number, that is to say, if the `raw prefix |
| 6406 | argument' is simply a cons cell, then the command puts you at the |
| 6407 | beginning of the second line @dots{} I don't know whether this is |
| 6408 | intended or whether no one has dealt with the code to avoid this |
| 6409 | happening. |
| 6410 | |
| 6411 | @node Second Buffer Related Review |
| 6412 | @section Review |
| 6413 | |
| 6414 | Here is a brief summary of some of the topics covered in this chapter. |
| 6415 | |
| 6416 | @table @code |
| 6417 | @item or |
| 6418 | Evaluate each argument in sequence, and return the value of the first |
| 6419 | argument that is not @code{nil}; if none return a value that is not |
| 6420 | @code{nil}, return @code{nil}. In brief, return the first true value |
| 6421 | of the arguments; return a true value if one @emph{or} any of the |
| 6422 | others are true. |
| 6423 | |
| 6424 | @item and |
| 6425 | Evaluate each argument in sequence, and if any are @code{nil}, return |
| 6426 | @code{nil}; if none are @code{nil}, return the value of the last |
| 6427 | argument. In brief, return a true value only if all the arguments are |
| 6428 | true; return a true value if one @emph{and} each of the others is |
| 6429 | true. |
| 6430 | |
| 6431 | @item &optional |
| 6432 | A keyword used to indicate that an argument to a function definition |
| 6433 | is optional; this means that the function can be evaluated without the |
| 6434 | argument, if desired. |
| 6435 | |
| 6436 | @item prefix-numeric-value |
| 6437 | Convert the `raw prefix argument' produced by @code{(interactive |
| 6438 | "P")} to a numeric value. |
| 6439 | |
| 6440 | @item forward-line |
| 6441 | Move point forward to the beginning of the next line, or if the argument |
| 6442 | is greater than one, forward that many lines. If it can't move as far |
| 6443 | forward as it is supposed to, @code{forward-line} goes forward as far as |
| 6444 | it can and then returns a count of the number of additional lines it was |
| 6445 | supposed to move but couldn't. |
| 6446 | |
| 6447 | @item erase-buffer |
| 6448 | Delete the entire contents of the current buffer. |
| 6449 | |
| 6450 | @item bufferp |
| 6451 | Return @code{t} if its argument is a buffer; otherwise return @code{nil}. |
| 6452 | @end table |
| 6453 | |
| 6454 | @node optional Exercise |
| 6455 | @section @code{optional} Argument Exercise |
| 6456 | |
| 6457 | Write an interactive function with an optional argument that tests |
| 6458 | whether its argument, a number, is greater than or equal to, or else, |
| 6459 | less than the value of @code{fill-column}, and tells you which, in a |
| 6460 | message. However, if you do not pass an argument to the function, use |
| 6461 | 56 as a default value. |
| 6462 | |
| 6463 | @node Narrowing & Widening |
| 6464 | @chapter Narrowing and Widening |
| 6465 | @cindex Focusing attention (narrowing) |
| 6466 | @cindex Narrowing |
| 6467 | @cindex Widening |
| 6468 | |
| 6469 | Narrowing is a feature of Emacs that makes it possible for you to focus |
| 6470 | on a specific part of a buffer, and work without accidentally changing |
| 6471 | other parts. Narrowing is normally disabled since it can confuse |
| 6472 | novices. |
| 6473 | |
| 6474 | @menu |
| 6475 | * Narrowing advantages:: The advantages of narrowing |
| 6476 | * save-restriction:: The @code{save-restriction} special form. |
| 6477 | * what-line:: The number of the line that point is on. |
| 6478 | * narrow Exercise:: |
| 6479 | @end menu |
| 6480 | |
| 6481 | @ifnottex |
| 6482 | @node Narrowing advantages |
| 6483 | @unnumberedsec The Advantages of Narrowing |
| 6484 | @end ifnottex |
| 6485 | |
| 6486 | With narrowing, the rest of a buffer is made invisible, as if it weren't |
| 6487 | there. This is an advantage if, for example, you want to replace a word |
| 6488 | in one part of a buffer but not in another: you narrow to the part you want |
| 6489 | and the replacement is carried out only in that section, not in the rest |
| 6490 | of the buffer. Searches will only work within a narrowed region, not |
| 6491 | outside of one, so if you are fixing a part of a document, you can keep |
| 6492 | yourself from accidentally finding parts you do not need to fix by |
| 6493 | narrowing just to the region you want. |
| 6494 | (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.) |
| 6495 | |
| 6496 | However, narrowing does make the rest of the buffer invisible, which |
| 6497 | can scare people who inadvertently invoke narrowing and think they |
| 6498 | have deleted a part of their file. Moreover, the @code{undo} command |
| 6499 | (which is usually bound to @kbd{C-x u}) does not turn off narrowing |
| 6500 | (nor should it), so people can become quite desperate if they do not |
| 6501 | know that they can return the rest of a buffer to visibility with the |
| 6502 | @code{widen} command. |
| 6503 | (The key binding for @code{widen} is @kbd{C-x n w}.) |
| 6504 | |
| 6505 | Narrowing is just as useful to the Lisp interpreter as to a human. |
| 6506 | Often, an Emacs Lisp function is designed to work on just part of a |
| 6507 | buffer; or conversely, an Emacs Lisp function needs to work on all of a |
| 6508 | buffer that has been narrowed. The @code{what-line} function, for |
| 6509 | example, removes the narrowing from a buffer, if it has any narrowing |
| 6510 | and when it has finished its job, restores the narrowing to what it was. |
| 6511 | On the other hand, the @code{count-lines} function |
| 6512 | uses narrowing to restrict itself to just that portion |
| 6513 | of the buffer in which it is interested and then restores the previous |
| 6514 | situation. |
| 6515 | |
| 6516 | @node save-restriction |
| 6517 | @section The @code{save-restriction} Special Form |
| 6518 | @findex save-restriction |
| 6519 | |
| 6520 | In Emacs Lisp, you can use the @code{save-restriction} special form to |
| 6521 | keep track of whatever narrowing is in effect, if any. When the Lisp |
| 6522 | interpreter meets with @code{save-restriction}, it executes the code |
| 6523 | in the body of the @code{save-restriction} expression, and then undoes |
| 6524 | any changes to narrowing that the code caused. If, for example, the |
| 6525 | buffer is narrowed and the code that follows @code{save-restriction} |
| 6526 | gets rid of the narrowing, @code{save-restriction} returns the buffer |
| 6527 | to its narrowed region afterwards. In the @code{what-line} command, |
| 6528 | any narrowing the buffer may have is undone by the @code{widen} |
| 6529 | command that immediately follows the @code{save-restriction} command. |
| 6530 | Any original narrowing is restored just before the completion of the |
| 6531 | function. |
| 6532 | |
| 6533 | @need 1250 |
| 6534 | The template for a @code{save-restriction} expression is simple: |
| 6535 | |
| 6536 | @smallexample |
| 6537 | @group |
| 6538 | (save-restriction |
| 6539 | @var{body}@dots{} ) |
| 6540 | @end group |
| 6541 | @end smallexample |
| 6542 | |
| 6543 | @noindent |
| 6544 | The body of the @code{save-restriction} is one or more expressions that |
| 6545 | will be evaluated in sequence by the Lisp interpreter. |
| 6546 | |
| 6547 | Finally, a point to note: when you use both @code{save-excursion} and |
| 6548 | @code{save-restriction}, one right after the other, you should use |
| 6549 | @code{save-excursion} outermost. If you write them in reverse order, |
| 6550 | you may fail to record narrowing in the buffer to which Emacs switches |
| 6551 | after calling @code{save-excursion}. Thus, when written together, |
| 6552 | @code{save-excursion} and @code{save-restriction} should be written |
| 6553 | like this: |
| 6554 | |
| 6555 | @smallexample |
| 6556 | @group |
| 6557 | (save-excursion |
| 6558 | (save-restriction |
| 6559 | @var{body}@dots{})) |
| 6560 | @end group |
| 6561 | @end smallexample |
| 6562 | |
| 6563 | In other circumstances, when not written together, the |
| 6564 | @code{save-excursion} and @code{save-restriction} special forms must |
| 6565 | be written in the order appropriate to the function. |
| 6566 | |
| 6567 | @need 1250 |
| 6568 | For example, |
| 6569 | |
| 6570 | @smallexample |
| 6571 | @group |
| 6572 | (save-restriction |
| 6573 | (widen) |
| 6574 | (save-excursion |
| 6575 | @var{body}@dots{})) |
| 6576 | @end group |
| 6577 | @end smallexample |
| 6578 | |
| 6579 | @ignore |
| 6580 | Emacs 22 |
| 6581 | /usr/local/src/emacs/lisp/simple.el |
| 6582 | |
| 6583 | (defun what-line () |
| 6584 | "Print the current buffer line number and narrowed line number of point." |
| 6585 | (interactive) |
| 6586 | (let ((start (point-min)) |
| 6587 | (n (line-number-at-pos))) |
| 6588 | (if (= start 1) |
| 6589 | (message "Line %d" n) |
| 6590 | (save-excursion |
| 6591 | (save-restriction |
| 6592 | (widen) |
| 6593 | (message "line %d (narrowed line %d)" |
| 6594 | (+ n (line-number-at-pos start) -1) n)))))) |
| 6595 | |
| 6596 | (defun line-number-at-pos (&optional pos) |
| 6597 | "Return (narrowed) buffer line number at position POS. |
| 6598 | If POS is nil, use current buffer location. |
| 6599 | Counting starts at (point-min), so the value refers |
| 6600 | to the contents of the accessible portion of the buffer." |
| 6601 | (let ((opoint (or pos (point))) start) |
| 6602 | (save-excursion |
| 6603 | (goto-char (point-min)) |
| 6604 | (setq start (point)) |
| 6605 | (goto-char opoint) |
| 6606 | (forward-line 0) |
| 6607 | (1+ (count-lines start (point)))))) |
| 6608 | |
| 6609 | (defun count-lines (start end) |
| 6610 | "Return number of lines between START and END. |
| 6611 | This is usually the number of newlines between them, |
| 6612 | but can be one more if START is not equal to END |
| 6613 | and the greater of them is not at the start of a line." |
| 6614 | (save-excursion |
| 6615 | (save-restriction |
| 6616 | (narrow-to-region start end) |
| 6617 | (goto-char (point-min)) |
| 6618 | (if (eq selective-display t) |
| 6619 | (save-match-data |
| 6620 | (let ((done 0)) |
| 6621 | (while (re-search-forward "[\n\C-m]" nil t 40) |
| 6622 | (setq done (+ 40 done))) |
| 6623 | (while (re-search-forward "[\n\C-m]" nil t 1) |
| 6624 | (setq done (+ 1 done))) |
| 6625 | (goto-char (point-max)) |
| 6626 | (if (and (/= start end) |
| 6627 | (not (bolp))) |
| 6628 | (1+ done) |
| 6629 | done))) |
| 6630 | (- (buffer-size) (forward-line (buffer-size))))))) |
| 6631 | @end ignore |
| 6632 | |
| 6633 | @node what-line |
| 6634 | @section @code{what-line} |
| 6635 | @findex what-line |
| 6636 | @cindex Widening, example of |
| 6637 | |
| 6638 | The @code{what-line} command tells you the number of the line in which |
| 6639 | the cursor is located. The function illustrates the use of the |
| 6640 | @code{save-restriction} and @code{save-excursion} commands. Here is the |
| 6641 | original text of the function: |
| 6642 | |
| 6643 | @smallexample |
| 6644 | @group |
| 6645 | (defun what-line () |
| 6646 | "Print the current line number (in the buffer) of point." |
| 6647 | (interactive) |
| 6648 | (save-restriction |
| 6649 | (widen) |
| 6650 | (save-excursion |
| 6651 | (beginning-of-line) |
| 6652 | (message "Line %d" |
| 6653 | (1+ (count-lines 1 (point))))))) |
| 6654 | @end group |
| 6655 | @end smallexample |
| 6656 | |
| 6657 | (In recent versions of GNU Emacs, the @code{what-line} function has |
| 6658 | been expanded to tell you your line number in a narrowed buffer as |
| 6659 | well as your line number in a widened buffer. The recent version is |
| 6660 | more complex than the version shown here. If you feel adventurous, |
| 6661 | you might want to look at it after figuring out how this version |
| 6662 | works. You will probably need to use @kbd{C-h f} |
| 6663 | (@code{describe-function}). The newer version uses a conditional to |
| 6664 | determine whether the buffer has been narrowed. |
| 6665 | |
| 6666 | (Also, it uses @code{line-number-at-pos}, which among other simple |
| 6667 | expressions, such as @code{(goto-char (point-min))}, moves point to |
| 6668 | the beginning of the current line with @code{(forward-line 0)} rather |
| 6669 | than @code{beginning-of-line}.) |
| 6670 | |
| 6671 | The @code{what-line} function as shown here has a documentation line |
| 6672 | and is interactive, as you would expect. The next two lines use the |
| 6673 | functions @code{save-restriction} and @code{widen}. |
| 6674 | |
| 6675 | The @code{save-restriction} special form notes whatever narrowing is in |
| 6676 | effect, if any, in the current buffer and restores that narrowing after |
| 6677 | the code in the body of the @code{save-restriction} has been evaluated. |
| 6678 | |
| 6679 | The @code{save-restriction} special form is followed by @code{widen}. |
| 6680 | This function undoes any narrowing the current buffer may have had |
| 6681 | when @code{what-line} was called. (The narrowing that was there is |
| 6682 | the narrowing that @code{save-restriction} remembers.) This widening |
| 6683 | makes it possible for the line counting commands to count from the |
| 6684 | beginning of the buffer. Otherwise, they would have been limited to |
| 6685 | counting within the accessible region. Any original narrowing is |
| 6686 | restored just before the completion of the function by the |
| 6687 | @code{save-restriction} special form. |
| 6688 | |
| 6689 | The call to @code{widen} is followed by @code{save-excursion}, which |
| 6690 | saves the location of the cursor (i.e., of point) and of the mark, and |
| 6691 | restores them after the code in the body of the @code{save-excursion} |
| 6692 | uses the @code{beginning-of-line} function to move point. |
| 6693 | |
| 6694 | (Note that the @code{(widen)} expression comes between the |
| 6695 | @code{save-restriction} and @code{save-excursion} special forms. When |
| 6696 | you write the two @code{save- @dots{}} expressions in sequence, write |
| 6697 | @code{save-excursion} outermost.) |
| 6698 | |
| 6699 | @need 1200 |
| 6700 | The last two lines of the @code{what-line} function are functions to |
| 6701 | count the number of lines in the buffer and then print the number in the |
| 6702 | echo area. |
| 6703 | |
| 6704 | @smallexample |
| 6705 | @group |
| 6706 | (message "Line %d" |
| 6707 | (1+ (count-lines 1 (point))))))) |
| 6708 | @end group |
| 6709 | @end smallexample |
| 6710 | |
| 6711 | The @code{message} function prints a one-line message at the bottom of |
| 6712 | the Emacs screen. The first argument is inside of quotation marks and |
| 6713 | is printed as a string of characters. However, it may contain a |
| 6714 | @samp{%d} expression to print a following argument. @samp{%d} prints |
| 6715 | the argument as a decimal, so the message will say something such as |
| 6716 | @samp{Line 243}. |
| 6717 | |
| 6718 | @need 1200 |
| 6719 | The number that is printed in place of the @samp{%d} is computed by the |
| 6720 | last line of the function: |
| 6721 | |
| 6722 | @smallexample |
| 6723 | (1+ (count-lines 1 (point))) |
| 6724 | @end smallexample |
| 6725 | |
| 6726 | @ignore |
| 6727 | GNU Emacs 22 |
| 6728 | |
| 6729 | (defun count-lines (start end) |
| 6730 | "Return number of lines between START and END. |
| 6731 | This is usually the number of newlines between them, |
| 6732 | but can be one more if START is not equal to END |
| 6733 | and the greater of them is not at the start of a line." |
| 6734 | (save-excursion |
| 6735 | (save-restriction |
| 6736 | (narrow-to-region start end) |
| 6737 | (goto-char (point-min)) |
| 6738 | (if (eq selective-display t) |
| 6739 | (save-match-data |
| 6740 | (let ((done 0)) |
| 6741 | (while (re-search-forward "[\n\C-m]" nil t 40) |
| 6742 | (setq done (+ 40 done))) |
| 6743 | (while (re-search-forward "[\n\C-m]" nil t 1) |
| 6744 | (setq done (+ 1 done))) |
| 6745 | (goto-char (point-max)) |
| 6746 | (if (and (/= start end) |
| 6747 | (not (bolp))) |
| 6748 | (1+ done) |
| 6749 | done))) |
| 6750 | (- (buffer-size) (forward-line (buffer-size))))))) |
| 6751 | @end ignore |
| 6752 | |
| 6753 | @noindent |
| 6754 | What this does is count the lines from the first position of the |
| 6755 | buffer, indicated by the @code{1}, up to @code{(point)}, and then add |
| 6756 | one to that number. (The @code{1+} function adds one to its |
| 6757 | argument.) We add one to it because line 2 has only one line before |
| 6758 | it, and @code{count-lines} counts only the lines @emph{before} the |
| 6759 | current line. |
| 6760 | |
| 6761 | After @code{count-lines} has done its job, and the message has been |
| 6762 | printed in the echo area, the @code{save-excursion} restores point and |
| 6763 | mark to their original positions; and @code{save-restriction} restores |
| 6764 | the original narrowing, if any. |
| 6765 | |
| 6766 | @node narrow Exercise |
| 6767 | @section Exercise with Narrowing |
| 6768 | |
| 6769 | Write a function that will display the first 60 characters of the |
| 6770 | current buffer, even if you have narrowed the buffer to its latter |
| 6771 | half so that the first line is inaccessible. Restore point, mark, and |
| 6772 | narrowing. For this exercise, you need to use a whole potpourri of |
| 6773 | functions, including @code{save-restriction}, @code{widen}, |
| 6774 | @code{goto-char}, @code{point-min}, @code{message}, and |
| 6775 | @code{buffer-substring}. |
| 6776 | |
| 6777 | @cindex Properties, mention of @code{buffer-substring-no-properties} |
| 6778 | (@code{buffer-substring} is a previously unmentioned function you will |
| 6779 | have to investigate yourself; or perhaps you will have to use |
| 6780 | @code{buffer-substring-no-properties} or |
| 6781 | @code{filter-buffer-substring} @dots{}, yet other functions. Text |
| 6782 | properties are a feature otherwise not discussed here. @xref{Text |
| 6783 | Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference |
| 6784 | Manual}.) |
| 6785 | |
| 6786 | Additionally, do you really need @code{goto-char} or @code{point-min}? |
| 6787 | Or can you write the function without them? |
| 6788 | |
| 6789 | @node car cdr & cons |
| 6790 | @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions |
| 6791 | @findex car, @r{introduced} |
| 6792 | @findex cdr, @r{introduced} |
| 6793 | |
| 6794 | In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental |
| 6795 | functions. The @code{cons} function is used to construct lists, and |
| 6796 | the @code{car} and @code{cdr} functions are used to take them apart. |
| 6797 | |
| 6798 | In the walk through of the @code{copy-region-as-kill} function, we |
| 6799 | will see @code{cons} as well as two variants on @code{cdr}, |
| 6800 | namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.) |
| 6801 | |
| 6802 | @menu |
| 6803 | * Strange Names:: An historical aside: why the strange names? |
| 6804 | * car & cdr:: Functions for extracting part of a list. |
| 6805 | * cons:: Constructing a list. |
| 6806 | * nthcdr:: Calling @code{cdr} repeatedly. |
| 6807 | * nth:: |
| 6808 | * setcar:: Changing the first element of a list. |
| 6809 | * setcdr:: Changing the rest of a list. |
| 6810 | * cons Exercise:: |
| 6811 | @end menu |
| 6812 | |
| 6813 | @ifnottex |
| 6814 | @node Strange Names |
| 6815 | @unnumberedsec Strange Names |
| 6816 | @end ifnottex |
| 6817 | |
| 6818 | The name of the @code{cons} function is not unreasonable: it is an |
| 6819 | abbreviation of the word `construct'. The origins of the names for |
| 6820 | @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car} |
| 6821 | is an acronym from the phrase `Contents of the Address part of the |
| 6822 | Register'; and @code{cdr} (pronounced `could-er') is an acronym from |
| 6823 | the phrase `Contents of the Decrement part of the Register'. These |
| 6824 | phrases refer to specific pieces of hardware on the very early |
| 6825 | computer on which the original Lisp was developed. Besides being |
| 6826 | obsolete, the phrases have been completely irrelevant for more than 25 |
| 6827 | years to anyone thinking about Lisp. Nonetheless, although a few |
| 6828 | brave scholars have begun to use more reasonable names for these |
| 6829 | functions, the old terms are still in use. In particular, since the |
| 6830 | terms are used in the Emacs Lisp source code, we will use them in this |
| 6831 | introduction. |
| 6832 | |
| 6833 | @node car & cdr |
| 6834 | @section @code{car} and @code{cdr} |
| 6835 | |
| 6836 | The @sc{car} of a list is, quite simply, the first item in the list. |
| 6837 | Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is |
| 6838 | @code{rose}. |
| 6839 | |
| 6840 | @need 1200 |
| 6841 | If you are reading this in Info in GNU Emacs, you can see this by |
| 6842 | evaluating the following: |
| 6843 | |
| 6844 | @smallexample |
| 6845 | (car '(rose violet daisy buttercup)) |
| 6846 | @end smallexample |
| 6847 | |
| 6848 | @noindent |
| 6849 | After evaluating the expression, @code{rose} will appear in the echo |
| 6850 | area. |
| 6851 | |
| 6852 | Clearly, a more reasonable name for the @code{car} function would be |
| 6853 | @code{first} and this is often suggested. |
| 6854 | |
| 6855 | @code{car} does not remove the first item from the list; it only reports |
| 6856 | what it is. After @code{car} has been applied to a list, the list is |
| 6857 | still the same as it was. In the jargon, @code{car} is |
| 6858 | `non-destructive'. This feature turns out to be important. |
| 6859 | |
| 6860 | The @sc{cdr} of a list is the rest of the list, that is, the |
| 6861 | @code{cdr} function returns the part of the list that follows the |
| 6862 | first item. Thus, while the @sc{car} of the list @code{'(rose violet |
| 6863 | daisy buttercup)} is @code{rose}, the rest of the list, the value |
| 6864 | returned by the @code{cdr} function, is @code{(violet daisy |
| 6865 | buttercup)}. |
| 6866 | |
| 6867 | @need 800 |
| 6868 | You can see this by evaluating the following in the usual way: |
| 6869 | |
| 6870 | @smallexample |
| 6871 | (cdr '(rose violet daisy buttercup)) |
| 6872 | @end smallexample |
| 6873 | |
| 6874 | @noindent |
| 6875 | When you evaluate this, @code{(violet daisy buttercup)} will appear in |
| 6876 | the echo area. |
| 6877 | |
| 6878 | Like @code{car}, @code{cdr} does not remove any elements from the |
| 6879 | list---it just returns a report of what the second and subsequent |
| 6880 | elements are. |
| 6881 | |
| 6882 | Incidentally, in the example, the list of flowers is quoted. If it were |
| 6883 | not, the Lisp interpreter would try to evaluate the list by calling |
| 6884 | @code{rose} as a function. In this example, we do not want to do that. |
| 6885 | |
| 6886 | Clearly, a more reasonable name for @code{cdr} would be @code{rest}. |
| 6887 | |
| 6888 | (There is a lesson here: when you name new functions, consider very |
| 6889 | carefully what you are doing, since you may be stuck with the names |
| 6890 | for far longer than you expect. The reason this document perpetuates |
| 6891 | these names is that the Emacs Lisp source code uses them, and if I did |
| 6892 | not use them, you would have a hard time reading the code; but do, |
| 6893 | please, try to avoid using these terms yourself. The people who come |
| 6894 | after you will be grateful to you.) |
| 6895 | |
| 6896 | When @code{car} and @code{cdr} are applied to a list made up of symbols, |
| 6897 | such as the list @code{(pine fir oak maple)}, the element of the list |
| 6898 | returned by the function @code{car} is the symbol @code{pine} without |
| 6899 | any parentheses around it. @code{pine} is the first element in the |
| 6900 | list. However, the @sc{cdr} of the list is a list itself, @code{(fir |
| 6901 | oak maple)}, as you can see by evaluating the following expressions in |
| 6902 | the usual way: |
| 6903 | |
| 6904 | @smallexample |
| 6905 | @group |
| 6906 | (car '(pine fir oak maple)) |
| 6907 | |
| 6908 | (cdr '(pine fir oak maple)) |
| 6909 | @end group |
| 6910 | @end smallexample |
| 6911 | |
| 6912 | On the other hand, in a list of lists, the first element is itself a |
| 6913 | list. @code{car} returns this first element as a list. For example, |
| 6914 | the following list contains three sub-lists, a list of carnivores, a |
| 6915 | list of herbivores and a list of sea mammals: |
| 6916 | |
| 6917 | @smallexample |
| 6918 | @group |
| 6919 | (car '((lion tiger cheetah) |
| 6920 | (gazelle antelope zebra) |
| 6921 | (whale dolphin seal))) |
| 6922 | @end group |
| 6923 | @end smallexample |
| 6924 | |
| 6925 | @noindent |
| 6926 | In this example, the first element or @sc{car} of the list is the list of |
| 6927 | carnivores, @code{(lion tiger cheetah)}, and the rest of the list is |
| 6928 | @code{((gazelle antelope zebra) (whale dolphin seal))}. |
| 6929 | |
| 6930 | @smallexample |
| 6931 | @group |
| 6932 | (cdr '((lion tiger cheetah) |
| 6933 | (gazelle antelope zebra) |
| 6934 | (whale dolphin seal))) |
| 6935 | @end group |
| 6936 | @end smallexample |
| 6937 | |
| 6938 | It is worth saying again that @code{car} and @code{cdr} are |
| 6939 | non-destructive---that is, they do not modify or change lists to which |
| 6940 | they are applied. This is very important for how they are used. |
| 6941 | |
| 6942 | Also, in the first chapter, in the discussion about atoms, I said that |
| 6943 | in Lisp, ``certain kinds of atom, such as an array, can be separated |
| 6944 | into parts; but the mechanism for doing this is different from the |
| 6945 | mechanism for splitting a list. As far as Lisp is concerned, the |
| 6946 | atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The |
| 6947 | @code{car} and @code{cdr} functions are used for splitting lists and |
| 6948 | are considered fundamental to Lisp. Since they cannot split or gain |
| 6949 | access to the parts of an array, an array is considered an atom. |
| 6950 | Conversely, the other fundamental function, @code{cons}, can put |
| 6951 | together or construct a list, but not an array. (Arrays are handled |
| 6952 | by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU |
| 6953 | Emacs Lisp Reference Manual}.) |
| 6954 | |
| 6955 | @node cons |
| 6956 | @section @code{cons} |
| 6957 | @findex cons, @r{introduced} |
| 6958 | |
| 6959 | The @code{cons} function constructs lists; it is the inverse of |
| 6960 | @code{car} and @code{cdr}. For example, @code{cons} can be used to make |
| 6961 | a four element list from the three element list, @code{(fir oak maple)}: |
| 6962 | |
| 6963 | @smallexample |
| 6964 | (cons 'pine '(fir oak maple)) |
| 6965 | @end smallexample |
| 6966 | |
| 6967 | @need 800 |
| 6968 | @noindent |
| 6969 | After evaluating this list, you will see |
| 6970 | |
| 6971 | @smallexample |
| 6972 | (pine fir oak maple) |
| 6973 | @end smallexample |
| 6974 | |
| 6975 | @noindent |
| 6976 | appear in the echo area. @code{cons} causes the creation of a new |
| 6977 | list in which the element is followed by the elements of the original |
| 6978 | list. |
| 6979 | |
| 6980 | We often say that `@code{cons} puts a new element at the beginning of |
| 6981 | a list; it attaches or pushes elements onto the list', but this |
| 6982 | phrasing can be misleading, since @code{cons} does not change an |
| 6983 | existing list, but creates a new one. |
| 6984 | |
| 6985 | Like @code{car} and @code{cdr}, @code{cons} is non-destructive. |
| 6986 | |
| 6987 | @menu |
| 6988 | * Build a list:: |
| 6989 | * length:: How to find the length of a list. |
| 6990 | @end menu |
| 6991 | |
| 6992 | @ifnottex |
| 6993 | @node Build a list |
| 6994 | @unnumberedsubsec Build a list |
| 6995 | @end ifnottex |
| 6996 | |
| 6997 | @code{cons} must have a list to attach to.@footnote{Actually, you can |
| 6998 | @code{cons} an element to an atom to produce a dotted pair. Dotted |
| 6999 | pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted |
| 7000 | Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You |
| 7001 | cannot start from absolutely nothing. If you are building a list, you |
| 7002 | need to provide at least an empty list at the beginning. Here is a |
| 7003 | series of @code{cons} expressions that build up a list of flowers. If |
| 7004 | you are reading this in Info in GNU Emacs, you can evaluate each of |
| 7005 | the expressions in the usual way; the value is printed in this text |
| 7006 | after @samp{@result{}}, which you may read as `evaluates to'. |
| 7007 | |
| 7008 | @smallexample |
| 7009 | @group |
| 7010 | (cons 'buttercup ()) |
| 7011 | @result{} (buttercup) |
| 7012 | @end group |
| 7013 | |
| 7014 | @group |
| 7015 | (cons 'daisy '(buttercup)) |
| 7016 | @result{} (daisy buttercup) |
| 7017 | @end group |
| 7018 | |
| 7019 | @group |
| 7020 | (cons 'violet '(daisy buttercup)) |
| 7021 | @result{} (violet daisy buttercup) |
| 7022 | @end group |
| 7023 | |
| 7024 | @group |
| 7025 | (cons 'rose '(violet daisy buttercup)) |
| 7026 | @result{} (rose violet daisy buttercup) |
| 7027 | @end group |
| 7028 | @end smallexample |
| 7029 | |
| 7030 | @noindent |
| 7031 | In the first example, the empty list is shown as @code{()} and a list |
| 7032 | made up of @code{buttercup} followed by the empty list is constructed. |
| 7033 | As you can see, the empty list is not shown in the list that was |
| 7034 | constructed. All that you see is @code{(buttercup)}. The empty list is |
| 7035 | not counted as an element of a list because there is nothing in an empty |
| 7036 | list. Generally speaking, an empty list is invisible. |
| 7037 | |
| 7038 | The second example, @code{(cons 'daisy '(buttercup))} constructs a new, |
| 7039 | two element list by putting @code{daisy} in front of @code{buttercup}; |
| 7040 | and the third example constructs a three element list by putting |
| 7041 | @code{violet} in front of @code{daisy} and @code{buttercup}. |
| 7042 | |
| 7043 | @node length |
| 7044 | @subsection Find the Length of a List: @code{length} |
| 7045 | @findex length |
| 7046 | |
| 7047 | You can find out how many elements there are in a list by using the Lisp |
| 7048 | function @code{length}, as in the following examples: |
| 7049 | |
| 7050 | @smallexample |
| 7051 | @group |
| 7052 | (length '(buttercup)) |
| 7053 | @result{} 1 |
| 7054 | @end group |
| 7055 | |
| 7056 | @group |
| 7057 | (length '(daisy buttercup)) |
| 7058 | @result{} 2 |
| 7059 | @end group |
| 7060 | |
| 7061 | @group |
| 7062 | (length (cons 'violet '(daisy buttercup))) |
| 7063 | @result{} 3 |
| 7064 | @end group |
| 7065 | @end smallexample |
| 7066 | |
| 7067 | @noindent |
| 7068 | In the third example, the @code{cons} function is used to construct a |
| 7069 | three element list which is then passed to the @code{length} function as |
| 7070 | its argument. |
| 7071 | |
| 7072 | @need 1200 |
| 7073 | We can also use @code{length} to count the number of elements in an |
| 7074 | empty list: |
| 7075 | |
| 7076 | @smallexample |
| 7077 | @group |
| 7078 | (length ()) |
| 7079 | @result{} 0 |
| 7080 | @end group |
| 7081 | @end smallexample |
| 7082 | |
| 7083 | @noindent |
| 7084 | As you would expect, the number of elements in an empty list is zero. |
| 7085 | |
| 7086 | An interesting experiment is to find out what happens if you try to find |
| 7087 | the length of no list at all; that is, if you try to call @code{length} |
| 7088 | without giving it an argument, not even an empty list: |
| 7089 | |
| 7090 | @smallexample |
| 7091 | (length ) |
| 7092 | @end smallexample |
| 7093 | |
| 7094 | @need 800 |
| 7095 | @noindent |
| 7096 | What you see, if you evaluate this, is the error message |
| 7097 | |
| 7098 | @smallexample |
| 7099 | Lisp error: (wrong-number-of-arguments length 0) |
| 7100 | @end smallexample |
| 7101 | |
| 7102 | @noindent |
| 7103 | This means that the function receives the wrong number of |
| 7104 | arguments, zero, when it expects some other number of arguments. In |
| 7105 | this case, one argument is expected, the argument being a list whose |
| 7106 | length the function is measuring. (Note that @emph{one} list is |
| 7107 | @emph{one} argument, even if the list has many elements inside it.) |
| 7108 | |
| 7109 | The part of the error message that says @samp{length} is the name of |
| 7110 | the function. |
| 7111 | |
| 7112 | @ignore |
| 7113 | @code{length} is still a subroutine, but you need C-h f to discover that. |
| 7114 | |
| 7115 | In an earlier version: |
| 7116 | This is written with a special notation, @samp{#<subr}, |
| 7117 | that indicates that the function @code{length} is one of the primitive |
| 7118 | functions written in C rather than in Emacs Lisp. (@samp{subr} is an |
| 7119 | abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a |
| 7120 | Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more |
| 7121 | about subroutines. |
| 7122 | @end ignore |
| 7123 | |
| 7124 | @node nthcdr |
| 7125 | @section @code{nthcdr} |
| 7126 | @findex nthcdr |
| 7127 | |
| 7128 | The @code{nthcdr} function is associated with the @code{cdr} function. |
| 7129 | What it does is take the @sc{cdr} of a list repeatedly. |
| 7130 | |
| 7131 | If you take the @sc{cdr} of the list @code{(pine fir |
| 7132 | oak maple)}, you will be returned the list @code{(fir oak maple)}. If you |
| 7133 | repeat this on what was returned, you will be returned the list |
| 7134 | @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original |
| 7135 | list will just give you the original @sc{cdr} since the function does |
| 7136 | not change the list. You need to evaluate the @sc{cdr} of the |
| 7137 | @sc{cdr} and so on.) If you continue this, eventually you will be |
| 7138 | returned an empty list, which in this case, instead of being shown as |
| 7139 | @code{()} is shown as @code{nil}. |
| 7140 | |
| 7141 | @need 1200 |
| 7142 | For review, here is a series of repeated @sc{cdr}s, the text following |
| 7143 | the @samp{@result{}} shows what is returned. |
| 7144 | |
| 7145 | @smallexample |
| 7146 | @group |
| 7147 | (cdr '(pine fir oak maple)) |
| 7148 | @result{}(fir oak maple) |
| 7149 | @end group |
| 7150 | |
| 7151 | @group |
| 7152 | (cdr '(fir oak maple)) |
| 7153 | @result{} (oak maple) |
| 7154 | @end group |
| 7155 | |
| 7156 | @group |
| 7157 | (cdr '(oak maple)) |
| 7158 | @result{}(maple) |
| 7159 | @end group |
| 7160 | |
| 7161 | @group |
| 7162 | (cdr '(maple)) |
| 7163 | @result{} nil |
| 7164 | @end group |
| 7165 | |
| 7166 | @group |
| 7167 | (cdr 'nil) |
| 7168 | @result{} nil |
| 7169 | @end group |
| 7170 | |
| 7171 | @group |
| 7172 | (cdr ()) |
| 7173 | @result{} nil |
| 7174 | @end group |
| 7175 | @end smallexample |
| 7176 | |
| 7177 | @need 1200 |
| 7178 | You can also do several @sc{cdr}s without printing the values in |
| 7179 | between, like this: |
| 7180 | |
| 7181 | @smallexample |
| 7182 | @group |
| 7183 | (cdr (cdr '(pine fir oak maple))) |
| 7184 | @result{} (oak maple) |
| 7185 | @end group |
| 7186 | @end smallexample |
| 7187 | |
| 7188 | @noindent |
| 7189 | In this example, the Lisp interpreter evaluates the innermost list first. |
| 7190 | The innermost list is quoted, so it just passes the list as it is to the |
| 7191 | innermost @code{cdr}. This @code{cdr} passes a list made up of the |
| 7192 | second and subsequent elements of the list to the outermost @code{cdr}, |
| 7193 | which produces a list composed of the third and subsequent elements of |
| 7194 | the original list. In this example, the @code{cdr} function is repeated |
| 7195 | and returns a list that consists of the original list without its |
| 7196 | first two elements. |
| 7197 | |
| 7198 | The @code{nthcdr} function does the same as repeating the call to |
| 7199 | @code{cdr}. In the following example, the argument 2 is passed to the |
| 7200 | function @code{nthcdr}, along with the list, and the value returned is |
| 7201 | the list without its first two items, which is exactly the same |
| 7202 | as repeating @code{cdr} twice on the list: |
| 7203 | |
| 7204 | @smallexample |
| 7205 | @group |
| 7206 | (nthcdr 2 '(pine fir oak maple)) |
| 7207 | @result{} (oak maple) |
| 7208 | @end group |
| 7209 | @end smallexample |
| 7210 | |
| 7211 | @need 1200 |
| 7212 | Using the original four element list, we can see what happens when |
| 7213 | various numeric arguments are passed to @code{nthcdr}, including 0, 1, |
| 7214 | and 5: |
| 7215 | |
| 7216 | @smallexample |
| 7217 | @group |
| 7218 | ;; @r{Leave the list as it was.} |
| 7219 | (nthcdr 0 '(pine fir oak maple)) |
| 7220 | @result{} (pine fir oak maple) |
| 7221 | @end group |
| 7222 | |
| 7223 | @group |
| 7224 | ;; @r{Return a copy without the first element.} |
| 7225 | (nthcdr 1 '(pine fir oak maple)) |
| 7226 | @result{} (fir oak maple) |
| 7227 | @end group |
| 7228 | |
| 7229 | @group |
| 7230 | ;; @r{Return a copy of the list without three elements.} |
| 7231 | (nthcdr 3 '(pine fir oak maple)) |
| 7232 | @result{} (maple) |
| 7233 | @end group |
| 7234 | |
| 7235 | @group |
| 7236 | ;; @r{Return a copy lacking all four elements.} |
| 7237 | (nthcdr 4 '(pine fir oak maple)) |
| 7238 | @result{} nil |
| 7239 | @end group |
| 7240 | |
| 7241 | @group |
| 7242 | ;; @r{Return a copy lacking all elements.} |
| 7243 | (nthcdr 5 '(pine fir oak maple)) |
| 7244 | @result{} nil |
| 7245 | @end group |
| 7246 | @end smallexample |
| 7247 | |
| 7248 | @node nth |
| 7249 | @section @code{nth} |
| 7250 | @findex nth |
| 7251 | |
| 7252 | The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly. |
| 7253 | The @code{nth} function takes the @sc{car} of the result returned by |
| 7254 | @code{nthcdr}. It returns the Nth element of the list. |
| 7255 | |
| 7256 | @need 1500 |
| 7257 | Thus, if it were not defined in C for speed, the definition of |
| 7258 | @code{nth} would be: |
| 7259 | |
| 7260 | @smallexample |
| 7261 | @group |
| 7262 | (defun nth (n list) |
| 7263 | "Returns the Nth element of LIST. |
| 7264 | N counts from zero. If LIST is not that long, nil is returned." |
| 7265 | (car (nthcdr n list))) |
| 7266 | @end group |
| 7267 | @end smallexample |
| 7268 | |
| 7269 | @noindent |
| 7270 | (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el}, |
| 7271 | but its definition was redone in C in the 1980s.) |
| 7272 | |
| 7273 | The @code{nth} function returns a single element of a list. |
| 7274 | This can be very convenient. |
| 7275 | |
| 7276 | Note that the elements are numbered from zero, not one. That is to |
| 7277 | say, the first element of a list, its @sc{car} is the zeroth element. |
| 7278 | This is called `zero-based' counting and often bothers people who |
| 7279 | are accustomed to the first element in a list being number one, which |
| 7280 | is `one-based'. |
| 7281 | |
| 7282 | @need 1250 |
| 7283 | For example: |
| 7284 | |
| 7285 | @smallexample |
| 7286 | @group |
| 7287 | (nth 0 '("one" "two" "three")) |
| 7288 | @result{} "one" |
| 7289 | |
| 7290 | (nth 1 '("one" "two" "three")) |
| 7291 | @result{} "two" |
| 7292 | @end group |
| 7293 | @end smallexample |
| 7294 | |
| 7295 | It is worth mentioning that @code{nth}, like @code{nthcdr} and |
| 7296 | @code{cdr}, does not change the original list---the function is |
| 7297 | non-destructive. This is in sharp contrast to the @code{setcar} and |
| 7298 | @code{setcdr} functions. |
| 7299 | |
| 7300 | @node setcar |
| 7301 | @section @code{setcar} |
| 7302 | @findex setcar |
| 7303 | |
| 7304 | As you might guess from their names, the @code{setcar} and @code{setcdr} |
| 7305 | functions set the @sc{car} or the @sc{cdr} of a list to a new value. |
| 7306 | They actually change the original list, unlike @code{car} and @code{cdr} |
| 7307 | which leave the original list as it was. One way to find out how this |
| 7308 | works is to experiment. We will start with the @code{setcar} function. |
| 7309 | |
| 7310 | @need 1200 |
| 7311 | First, we can make a list and then set the value of a variable to the |
| 7312 | list, using the @code{setq} function. Here is a list of animals: |
| 7313 | |
| 7314 | @smallexample |
| 7315 | (setq animals '(antelope giraffe lion tiger)) |
| 7316 | @end smallexample |
| 7317 | |
| 7318 | @noindent |
| 7319 | If you are reading this in Info inside of GNU Emacs, you can evaluate |
| 7320 | this expression in the usual fashion, by positioning the cursor after |
| 7321 | the expression and typing @kbd{C-x C-e}. (I'm doing this right here |
| 7322 | as I write this. This is one of the advantages of having the |
| 7323 | interpreter built into the computing environment. Incidentally, when |
| 7324 | there is nothing on the line after the final parentheses, such as a |
| 7325 | comment, point can be on the next line. Thus, if your cursor is in |
| 7326 | the first column of the next line, you do not need to move it. |
| 7327 | Indeed, Emacs permits any amount of white space after the final |
| 7328 | parenthesis.) |
| 7329 | |
| 7330 | @need 1200 |
| 7331 | When we evaluate the variable @code{animals}, we see that it is bound to |
| 7332 | the list @code{(antelope giraffe lion tiger)}: |
| 7333 | |
| 7334 | @smallexample |
| 7335 | @group |
| 7336 | animals |
| 7337 | @result{} (antelope giraffe lion tiger) |
| 7338 | @end group |
| 7339 | @end smallexample |
| 7340 | |
| 7341 | @noindent |
| 7342 | Put another way, the variable @code{animals} points to the list |
| 7343 | @code{(antelope giraffe lion tiger)}. |
| 7344 | |
| 7345 | Next, evaluate the function @code{setcar} while passing it two |
| 7346 | arguments, the variable @code{animals} and the quoted symbol |
| 7347 | @code{hippopotamus}; this is done by writing the three element list |
| 7348 | @code{(setcar animals 'hippopotamus)} and then evaluating it in the |
| 7349 | usual fashion: |
| 7350 | |
| 7351 | @smallexample |
| 7352 | (setcar animals 'hippopotamus) |
| 7353 | @end smallexample |
| 7354 | |
| 7355 | @need 1200 |
| 7356 | @noindent |
| 7357 | After evaluating this expression, evaluate the variable @code{animals} |
| 7358 | again. You will see that the list of animals has changed: |
| 7359 | |
| 7360 | @smallexample |
| 7361 | @group |
| 7362 | animals |
| 7363 | @result{} (hippopotamus giraffe lion tiger) |
| 7364 | @end group |
| 7365 | @end smallexample |
| 7366 | |
| 7367 | @noindent |
| 7368 | The first element on the list, @code{antelope} is replaced by |
| 7369 | @code{hippopotamus}. |
| 7370 | |
| 7371 | So we can see that @code{setcar} did not add a new element to the list |
| 7372 | as @code{cons} would have; it replaced @code{antelope} with |
| 7373 | @code{hippopotamus}; it @emph{changed} the list. |
| 7374 | |
| 7375 | @node setcdr |
| 7376 | @section @code{setcdr} |
| 7377 | @findex setcdr |
| 7378 | |
| 7379 | The @code{setcdr} function is similar to the @code{setcar} function, |
| 7380 | except that the function replaces the second and subsequent elements of |
| 7381 | a list rather than the first element. |
| 7382 | |
| 7383 | (To see how to change the last element of a list, look ahead to |
| 7384 | @ref{kill-new function, , The @code{kill-new} function}, which uses |
| 7385 | the @code{nthcdr} and @code{setcdr} functions.) |
| 7386 | |
| 7387 | @need 1200 |
| 7388 | To see how this works, set the value of the variable to a list of |
| 7389 | domesticated animals by evaluating the following expression: |
| 7390 | |
| 7391 | @smallexample |
| 7392 | (setq domesticated-animals '(horse cow sheep goat)) |
| 7393 | @end smallexample |
| 7394 | |
| 7395 | @need 1200 |
| 7396 | @noindent |
| 7397 | If you now evaluate the list, you will be returned the list |
| 7398 | @code{(horse cow sheep goat)}: |
| 7399 | |
| 7400 | @smallexample |
| 7401 | @group |
| 7402 | domesticated-animals |
| 7403 | @result{} (horse cow sheep goat) |
| 7404 | @end group |
| 7405 | @end smallexample |
| 7406 | |
| 7407 | @need 1200 |
| 7408 | Next, evaluate @code{setcdr} with two arguments, the name of the |
| 7409 | variable which has a list as its value, and the list to which the |
| 7410 | @sc{cdr} of the first list will be set; |
| 7411 | |
| 7412 | @smallexample |
| 7413 | (setcdr domesticated-animals '(cat dog)) |
| 7414 | @end smallexample |
| 7415 | |
| 7416 | @noindent |
| 7417 | If you evaluate this expression, the list @code{(cat dog)} will appear |
| 7418 | in the echo area. This is the value returned by the function. The |
| 7419 | result we are interested in is the ``side effect'', which we can see by |
| 7420 | evaluating the variable @code{domesticated-animals}: |
| 7421 | |
| 7422 | @smallexample |
| 7423 | @group |
| 7424 | domesticated-animals |
| 7425 | @result{} (horse cat dog) |
| 7426 | @end group |
| 7427 | @end smallexample |
| 7428 | |
| 7429 | @noindent |
| 7430 | Indeed, the list is changed from @code{(horse cow sheep goat)} to |
| 7431 | @code{(horse cat dog)}. The @sc{cdr} of the list is changed from |
| 7432 | @code{(cow sheep goat)} to @code{(cat dog)}. |
| 7433 | |
| 7434 | @node cons Exercise |
| 7435 | @section Exercise |
| 7436 | |
| 7437 | Construct a list of four birds by evaluating several expressions with |
| 7438 | @code{cons}. Find out what happens when you @code{cons} a list onto |
| 7439 | itself. Replace the first element of the list of four birds with a |
| 7440 | fish. Replace the rest of that list with a list of other fish. |
| 7441 | |
| 7442 | @node Cutting & Storing Text |
| 7443 | @chapter Cutting and Storing Text |
| 7444 | @cindex Cutting and storing text |
| 7445 | @cindex Storing and cutting text |
| 7446 | @cindex Killing text |
| 7447 | @cindex Clipping text |
| 7448 | @cindex Erasing text |
| 7449 | @cindex Deleting text |
| 7450 | |
| 7451 | Whenever you cut or clip text out of a buffer with a `kill' command in |
| 7452 | GNU Emacs, it is stored in a list and you can bring it back with a |
| 7453 | `yank' command. |
| 7454 | |
| 7455 | (The use of the word `kill' in Emacs for processes which specifically |
| 7456 | @emph{do not} destroy the values of the entities is an unfortunate |
| 7457 | historical accident. A much more appropriate word would be `clip' since |
| 7458 | that is what the kill commands do; they clip text out of a buffer and |
| 7459 | put it into storage from which it can be brought back. I have often |
| 7460 | been tempted to replace globally all occurrences of `kill' in the Emacs |
| 7461 | sources with `clip' and all occurrences of `killed' with `clipped'.) |
| 7462 | |
| 7463 | @menu |
| 7464 | * Storing Text:: Text is stored in a list. |
| 7465 | * zap-to-char:: Cutting out text up to a character. |
| 7466 | * kill-region:: Cutting text out of a region. |
| 7467 | * copy-region-as-kill:: A definition for copying text. |
| 7468 | * Digression into C:: Minor note on C programming language macros. |
| 7469 | * defvar:: How to give a variable an initial value. |
| 7470 | * cons & search-fwd Review:: |
| 7471 | * search Exercises:: |
| 7472 | @end menu |
| 7473 | |
| 7474 | @ifnottex |
| 7475 | @node Storing Text |
| 7476 | @unnumberedsec Storing Text in a List |
| 7477 | @end ifnottex |
| 7478 | |
| 7479 | When text is cut out of a buffer, it is stored on a list. Successive |
| 7480 | pieces of text are stored on the list successively, so the list might |
| 7481 | look like this: |
| 7482 | |
| 7483 | @smallexample |
| 7484 | ("a piece of text" "previous piece") |
| 7485 | @end smallexample |
| 7486 | |
| 7487 | @need 1200 |
| 7488 | @noindent |
| 7489 | The function @code{cons} can be used to create a new list from a piece |
| 7490 | of text (an `atom', to use the jargon) and an existing list, like |
| 7491 | this: |
| 7492 | |
| 7493 | @smallexample |
| 7494 | @group |
| 7495 | (cons "another piece" |
| 7496 | '("a piece of text" "previous piece")) |
| 7497 | @end group |
| 7498 | @end smallexample |
| 7499 | |
| 7500 | @need 1200 |
| 7501 | @noindent |
| 7502 | If you evaluate this expression, a list of three elements will appear in |
| 7503 | the echo area: |
| 7504 | |
| 7505 | @smallexample |
| 7506 | ("another piece" "a piece of text" "previous piece") |
| 7507 | @end smallexample |
| 7508 | |
| 7509 | With the @code{car} and @code{nthcdr} functions, you can retrieve |
| 7510 | whichever piece of text you want. For example, in the following code, |
| 7511 | @code{nthcdr 1 @dots{}} returns the list with the first item removed; |
| 7512 | and the @code{car} returns the first element of that remainder---the |
| 7513 | second element of the original list: |
| 7514 | |
| 7515 | @smallexample |
| 7516 | @group |
| 7517 | (car (nthcdr 1 '("another piece" |
| 7518 | "a piece of text" |
| 7519 | "previous piece"))) |
| 7520 | @result{} "a piece of text" |
| 7521 | @end group |
| 7522 | @end smallexample |
| 7523 | |
| 7524 | The actual functions in Emacs are more complex than this, of course. |
| 7525 | The code for cutting and retrieving text has to be written so that |
| 7526 | Emacs can figure out which element in the list you want---the first, |
| 7527 | second, third, or whatever. In addition, when you get to the end of |
| 7528 | the list, Emacs should give you the first element of the list, rather |
| 7529 | than nothing at all. |
| 7530 | |
| 7531 | The list that holds the pieces of text is called the @dfn{kill ring}. |
| 7532 | This chapter leads up to a description of the kill ring and how it is |
| 7533 | used by first tracing how the @code{zap-to-char} function works. This |
| 7534 | function uses (or `calls') a function that invokes a function that |
| 7535 | manipulates the kill ring. Thus, before reaching the mountains, we |
| 7536 | climb the foothills. |
| 7537 | |
| 7538 | A subsequent chapter describes how text that is cut from the buffer is |
| 7539 | retrieved. @xref{Yanking, , Yanking Text Back}. |
| 7540 | |
| 7541 | @node zap-to-char |
| 7542 | @section @code{zap-to-char} |
| 7543 | @findex zap-to-char |
| 7544 | |
| 7545 | Let us look at the interactive @code{zap-to-char} function. |
| 7546 | |
| 7547 | @menu |
| 7548 | * Complete zap-to-char:: The complete implementation. |
| 7549 | * zap-to-char interactive:: A three part interactive expression. |
| 7550 | * zap-to-char body:: A short overview. |
| 7551 | * search-forward:: How to search for a string. |
| 7552 | * progn:: The @code{progn} special form. |
| 7553 | * Summing up zap-to-char:: Using @code{point} and @code{search-forward}. |
| 7554 | @end menu |
| 7555 | |
| 7556 | @ifnottex |
| 7557 | @node Complete zap-to-char |
| 7558 | @unnumberedsubsec The Complete @code{zap-to-char} Implementation |
| 7559 | @end ifnottex |
| 7560 | |
| 7561 | The @code{zap-to-char} function removes the text in the region between |
| 7562 | the location of the cursor (i.e., of point) up to and including the |
| 7563 | next occurrence of a specified character. The text that |
| 7564 | @code{zap-to-char} removes is put in the kill ring; and it can be |
| 7565 | retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If |
| 7566 | the command is given an argument, it removes text through that number |
| 7567 | of occurrences. Thus, if the cursor were at the beginning of this |
| 7568 | sentence and the character were @samp{s}, @samp{Thus} would be |
| 7569 | removed. If the argument were two, @samp{Thus, if the curs} would be |
| 7570 | removed, up to and including the @samp{s} in @samp{cursor}. |
| 7571 | |
| 7572 | If the specified character is not found, @code{zap-to-char} will say |
| 7573 | ``Search failed'', tell you the character you typed, and not remove |
| 7574 | any text. |
| 7575 | |
| 7576 | In order to determine how much text to remove, @code{zap-to-char} uses |
| 7577 | a search function. Searches are used extensively in code that |
| 7578 | manipulates text, and we will focus attention on them as well as on the |
| 7579 | deletion command. |
| 7580 | |
| 7581 | @ignore |
| 7582 | @c GNU Emacs version 19 |
| 7583 | (defun zap-to-char (arg char) ; version 19 implementation |
| 7584 | "Kill up to and including ARG'th occurrence of CHAR. |
| 7585 | Goes backward if ARG is negative; error if CHAR not found." |
| 7586 | (interactive "*p\ncZap to char: ") |
| 7587 | (kill-region (point) |
| 7588 | (progn |
| 7589 | (search-forward |
| 7590 | (char-to-string char) nil nil arg) |
| 7591 | (point)))) |
| 7592 | @end ignore |
| 7593 | |
| 7594 | @need 1250 |
| 7595 | Here is the complete text of the version 22 implementation of the function: |
| 7596 | |
| 7597 | @c GNU Emacs 22 |
| 7598 | @smallexample |
| 7599 | @group |
| 7600 | (defun zap-to-char (arg char) |
| 7601 | "Kill up to and including ARG'th occurrence of CHAR. |
| 7602 | Case is ignored if `case-fold-search' is non-nil in the current buffer. |
| 7603 | Goes backward if ARG is negative; error if CHAR not found." |
| 7604 | (interactive "p\ncZap to char: ") |
| 7605 | (if (char-table-p translation-table-for-input) |
| 7606 | (setq char (or (aref translation-table-for-input char) char))) |
| 7607 | (kill-region (point) (progn |
| 7608 | (search-forward (char-to-string char) |
| 7609 | nil nil arg) |
| 7610 | (point)))) |
| 7611 | @end group |
| 7612 | @end smallexample |
| 7613 | |
| 7614 | The documentation is thorough. You do need to know the jargon meaning |
| 7615 | of the word `kill'. |
| 7616 | |
| 7617 | @node zap-to-char interactive |
| 7618 | @subsection The @code{interactive} Expression |
| 7619 | |
| 7620 | @need 800 |
| 7621 | The interactive expression in the @code{zap-to-char} command looks like |
| 7622 | this: |
| 7623 | |
| 7624 | @smallexample |
| 7625 | (interactive "p\ncZap to char: ") |
| 7626 | @end smallexample |
| 7627 | |
| 7628 | The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies |
| 7629 | two different things. First, and most simply, is the @samp{p}. |
| 7630 | This part is separated from the next part by a newline, @samp{\n}. |
| 7631 | The @samp{p} means that the first argument to the function will be |
| 7632 | passed the value of a `processed prefix'. The prefix argument is |
| 7633 | passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If |
| 7634 | the function is called interactively without a prefix, 1 is passed to |
| 7635 | this argument. |
| 7636 | |
| 7637 | The second part of @code{"p\ncZap to char:@: "} is |
| 7638 | @samp{cZap to char:@: }. In this part, the lower case @samp{c} |
| 7639 | indicates that @code{interactive} expects a prompt and that the |
| 7640 | argument will be a character. The prompt follows the @samp{c} and is |
| 7641 | the string @samp{Zap to char:@: } (with a space after the colon to |
| 7642 | make it look good). |
| 7643 | |
| 7644 | What all this does is prepare the arguments to @code{zap-to-char} so they |
| 7645 | are of the right type, and give the user a prompt. |
| 7646 | |
| 7647 | In a read-only buffer, the @code{zap-to-char} function copies the text |
| 7648 | to the kill ring, but does not remove it. The echo area displays a |
| 7649 | message saying that the buffer is read-only. Also, the terminal may |
| 7650 | beep or blink at you. |
| 7651 | |
| 7652 | @node zap-to-char body |
| 7653 | @subsection The Body of @code{zap-to-char} |
| 7654 | |
| 7655 | The body of the @code{zap-to-char} function contains the code that |
| 7656 | kills (that is, removes) the text in the region from the current |
| 7657 | position of the cursor up to and including the specified character. |
| 7658 | |
| 7659 | The first part of the code looks like this: |
| 7660 | |
| 7661 | @smallexample |
| 7662 | (if (char-table-p translation-table-for-input) |
| 7663 | (setq char (or (aref translation-table-for-input char) char))) |
| 7664 | (kill-region (point) (progn |
| 7665 | (search-forward (char-to-string char) nil nil arg) |
| 7666 | (point))) |
| 7667 | @end smallexample |
| 7668 | |
| 7669 | @noindent |
| 7670 | @code{char-table-p} is an hitherto unseen function. It determines |
| 7671 | whether its argument is a character table. When it is, it sets the |
| 7672 | character passed to @code{zap-to-char} to one of them, if that |
| 7673 | character exists, or to the character itself. (This becomes important |
| 7674 | for certain characters in non-European languages. The @code{aref} |
| 7675 | function extracts an element from an array. It is an array-specific |
| 7676 | function that is not described in this document. @xref{Arrays, , |
| 7677 | Arrays, elisp, The GNU Emacs Lisp Reference Manual}.) |
| 7678 | |
| 7679 | @noindent |
| 7680 | @code{(point)} is the current position of the cursor. |
| 7681 | |
| 7682 | The next part of the code is an expression using @code{progn}. The body |
| 7683 | of the @code{progn} consists of calls to @code{search-forward} and |
| 7684 | @code{point}. |
| 7685 | |
| 7686 | It is easier to understand how @code{progn} works after learning about |
| 7687 | @code{search-forward}, so we will look at @code{search-forward} and |
| 7688 | then at @code{progn}. |
| 7689 | |
| 7690 | @node search-forward |
| 7691 | @subsection The @code{search-forward} Function |
| 7692 | @findex search-forward |
| 7693 | |
| 7694 | The @code{search-forward} function is used to locate the |
| 7695 | zapped-for-character in @code{zap-to-char}. If the search is |
| 7696 | successful, @code{search-forward} leaves point immediately after the |
| 7697 | last character in the target string. (In @code{zap-to-char}, the |
| 7698 | target string is just one character long. @code{zap-to-char} uses the |
| 7699 | function @code{char-to-string} to ensure that the computer treats that |
| 7700 | character as a string.) If the search is backwards, |
| 7701 | @code{search-forward} leaves point just before the first character in |
| 7702 | the target. Also, @code{search-forward} returns @code{t} for true. |
| 7703 | (Moving point is therefore a `side effect'.) |
| 7704 | |
| 7705 | @need 1250 |
| 7706 | In @code{zap-to-char}, the @code{search-forward} function looks like this: |
| 7707 | |
| 7708 | @smallexample |
| 7709 | (search-forward (char-to-string char) nil nil arg) |
| 7710 | @end smallexample |
| 7711 | |
| 7712 | The @code{search-forward} function takes four arguments: |
| 7713 | |
| 7714 | @enumerate |
| 7715 | @item |
| 7716 | The first argument is the target, what is searched for. This must be a |
| 7717 | string, such as @samp{"z"}. |
| 7718 | |
| 7719 | As it happens, the argument passed to @code{zap-to-char} is a single |
| 7720 | character. Because of the way computers are built, the Lisp |
| 7721 | interpreter may treat a single character as being different from a |
| 7722 | string of characters. Inside the computer, a single character has a |
| 7723 | different electronic format than a string of one character. (A single |
| 7724 | character can often be recorded in the computer using exactly one |
| 7725 | byte; but a string may be longer, and the computer needs to be ready |
| 7726 | for this.) Since the @code{search-forward} function searches for a |
| 7727 | string, the character that the @code{zap-to-char} function receives as |
| 7728 | its argument must be converted inside the computer from one format to |
| 7729 | the other; otherwise the @code{search-forward} function will fail. |
| 7730 | The @code{char-to-string} function is used to make this conversion. |
| 7731 | |
| 7732 | @item |
| 7733 | The second argument bounds the search; it is specified as a position in |
| 7734 | the buffer. In this case, the search can go to the end of the buffer, |
| 7735 | so no bound is set and the second argument is @code{nil}. |
| 7736 | |
| 7737 | @item |
| 7738 | The third argument tells the function what it should do if the search |
| 7739 | fails---it can signal an error (and print a message) or it can return |
| 7740 | @code{nil}. A @code{nil} as the third argument causes the function to |
| 7741 | signal an error when the search fails. |
| 7742 | |
| 7743 | @item |
| 7744 | The fourth argument to @code{search-forward} is the repeat count---how |
| 7745 | many occurrences of the string to look for. This argument is optional |
| 7746 | and if the function is called without a repeat count, this argument is |
| 7747 | passed the value 1. If this argument is negative, the search goes |
| 7748 | backwards. |
| 7749 | @end enumerate |
| 7750 | |
| 7751 | @need 800 |
| 7752 | In template form, a @code{search-forward} expression looks like this: |
| 7753 | |
| 7754 | @smallexample |
| 7755 | @group |
| 7756 | (search-forward "@var{target-string}" |
| 7757 | @var{limit-of-search} |
| 7758 | @var{what-to-do-if-search-fails} |
| 7759 | @var{repeat-count}) |
| 7760 | @end group |
| 7761 | @end smallexample |
| 7762 | |
| 7763 | We will look at @code{progn} next. |
| 7764 | |
| 7765 | @node progn |
| 7766 | @subsection The @code{progn} Special Form |
| 7767 | @findex progn |
| 7768 | |
| 7769 | @code{progn} is a special form that causes each of its arguments to be |
| 7770 | evaluated in sequence and then returns the value of the last one. The |
| 7771 | preceding expressions are evaluated only for the side effects they |
| 7772 | perform. The values produced by them are discarded. |
| 7773 | |
| 7774 | @need 800 |
| 7775 | The template for a @code{progn} expression is very simple: |
| 7776 | |
| 7777 | @smallexample |
| 7778 | @group |
| 7779 | (progn |
| 7780 | @var{body}@dots{}) |
| 7781 | @end group |
| 7782 | @end smallexample |
| 7783 | |
| 7784 | In @code{zap-to-char}, the @code{progn} expression has to do two things: |
| 7785 | put point in exactly the right position; and return the location of |
| 7786 | point so that @code{kill-region} will know how far to kill to. |
| 7787 | |
| 7788 | The first argument to the @code{progn} is @code{search-forward}. When |
| 7789 | @code{search-forward} finds the string, the function leaves point |
| 7790 | immediately after the last character in the target string. (In this |
| 7791 | case the target string is just one character long.) If the search is |
| 7792 | backwards, @code{search-forward} leaves point just before the first |
| 7793 | character in the target. The movement of point is a side effect. |
| 7794 | |
| 7795 | The second and last argument to @code{progn} is the expression |
| 7796 | @code{(point)}. This expression returns the value of point, which in |
| 7797 | this case will be the location to which it has been moved by |
| 7798 | @code{search-forward}. (In the source, a line that tells the function |
| 7799 | to go to the previous character, if it is going forward, was commented |
| 7800 | out in 1999; I don't remember whether that feature or mis-feature was |
| 7801 | ever a part of the distributed source.) The value of @code{point} is |
| 7802 | returned by the @code{progn} expression and is passed to |
| 7803 | @code{kill-region} as @code{kill-region}'s second argument. |
| 7804 | |
| 7805 | @node Summing up zap-to-char |
| 7806 | @subsection Summing up @code{zap-to-char} |
| 7807 | |
| 7808 | Now that we have seen how @code{search-forward} and @code{progn} work, |
| 7809 | we can see how the @code{zap-to-char} function works as a whole. |
| 7810 | |
| 7811 | The first argument to @code{kill-region} is the position of the cursor |
| 7812 | when the @code{zap-to-char} command is given---the value of point at |
| 7813 | that time. Within the @code{progn}, the search function then moves |
| 7814 | point to just after the zapped-to-character and @code{point} returns the |
| 7815 | value of this location. The @code{kill-region} function puts together |
| 7816 | these two values of point, the first one as the beginning of the region |
| 7817 | and the second one as the end of the region, and removes the region. |
| 7818 | |
| 7819 | The @code{progn} special form is necessary because the |
| 7820 | @code{kill-region} command takes two arguments; and it would fail if |
| 7821 | @code{search-forward} and @code{point} expressions were written in |
| 7822 | sequence as two additional arguments. The @code{progn} expression is |
| 7823 | a single argument to @code{kill-region} and returns the one value that |
| 7824 | @code{kill-region} needs for its second argument. |
| 7825 | |
| 7826 | @node kill-region |
| 7827 | @section @code{kill-region} |
| 7828 | @findex kill-region |
| 7829 | |
| 7830 | The @code{zap-to-char} function uses the @code{kill-region} function. |
| 7831 | This function clips text from a region and copies that text to |
| 7832 | the kill ring, from which it may be retrieved. |
| 7833 | |
| 7834 | @ignore |
| 7835 | GNU Emacs 22: |
| 7836 | |
| 7837 | (defun kill-region (beg end &optional yank-handler) |
| 7838 | "Kill (\"cut\") text between point and mark. |
| 7839 | This deletes the text from the buffer and saves it in the kill ring. |
| 7840 | The command \\[yank] can retrieve it from there. |
| 7841 | \(If you want to kill and then yank immediately, use \\[kill-ring-save].) |
| 7842 | |
| 7843 | If you want to append the killed region to the last killed text, |
| 7844 | use \\[append-next-kill] before \\[kill-region]. |
| 7845 | |
| 7846 | If the buffer is read-only, Emacs will beep and refrain from deleting |
| 7847 | the text, but put the text in the kill ring anyway. This means that |
| 7848 | you can use the killing commands to copy text from a read-only buffer. |
| 7849 | |
| 7850 | This is the primitive for programs to kill text (as opposed to deleting it). |
| 7851 | Supply two arguments, character positions indicating the stretch of text |
| 7852 | to be killed. |
| 7853 | Any command that calls this function is a \"kill command\". |
| 7854 | If the previous command was also a kill command, |
| 7855 | the text killed this time appends to the text killed last time |
| 7856 | to make one entry in the kill ring. |
| 7857 | |
| 7858 | In Lisp code, optional third arg YANK-HANDLER, if non-nil, |
| 7859 | specifies the yank-handler text property to be set on the killed |
| 7860 | text. See `insert-for-yank'." |
| 7861 | ;; Pass point first, then mark, because the order matters |
| 7862 | ;; when calling kill-append. |
| 7863 | (interactive (list (point) (mark))) |
| 7864 | (unless (and beg end) |
| 7865 | (error "The mark is not set now, so there is no region")) |
| 7866 | (condition-case nil |
| 7867 | (let ((string (filter-buffer-substring beg end t))) |
| 7868 | (when string ;STRING is nil if BEG = END |
| 7869 | ;; Add that string to the kill ring, one way or another. |
| 7870 | (if (eq last-command 'kill-region) |
| 7871 | (kill-append string (< end beg) yank-handler) |
| 7872 | (kill-new string nil yank-handler))) |
| 7873 | (when (or string (eq last-command 'kill-region)) |
| 7874 | (setq this-command 'kill-region)) |
| 7875 | nil) |
| 7876 | ((buffer-read-only text-read-only) |
| 7877 | ;; The code above failed because the buffer, or some of the characters |
| 7878 | ;; in the region, are read-only. |
| 7879 | ;; We should beep, in case the user just isn't aware of this. |
| 7880 | ;; However, there's no harm in putting |
| 7881 | ;; the region's text in the kill ring, anyway. |
| 7882 | (copy-region-as-kill beg end) |
| 7883 | ;; Set this-command now, so it will be set even if we get an error. |
| 7884 | (setq this-command 'kill-region) |
| 7885 | ;; This should barf, if appropriate, and give us the correct error. |
| 7886 | (if kill-read-only-ok |
| 7887 | (progn (message "Read only text copied to kill ring") nil) |
| 7888 | ;; Signal an error if the buffer is read-only. |
| 7889 | (barf-if-buffer-read-only) |
| 7890 | ;; If the buffer isn't read-only, the text is. |
| 7891 | (signal 'text-read-only (list (current-buffer))))))) |
| 7892 | @end ignore |
| 7893 | |
| 7894 | The Emacs 22 version of that function uses @code{condition-case} and |
| 7895 | @code{copy-region-as-kill}, both of which we will explain. |
| 7896 | @code{condition-case} is an important special form. |
| 7897 | |
| 7898 | In essence, the @code{kill-region} function calls |
| 7899 | @code{condition-case}, which takes three arguments. In this function, |
| 7900 | the first argument does nothing. The second argument contains the |
| 7901 | code that does the work when all goes well. The third argument |
| 7902 | contains the code that is called in the event of an error. |
| 7903 | |
| 7904 | @menu |
| 7905 | * Complete kill-region:: The function definition. |
| 7906 | * condition-case:: Dealing with a problem. |
| 7907 | * Lisp macro:: |
| 7908 | @end menu |
| 7909 | |
| 7910 | @ifnottex |
| 7911 | @node Complete kill-region |
| 7912 | @unnumberedsubsec The Complete @code{kill-region} Definition |
| 7913 | @end ifnottex |
| 7914 | |
| 7915 | @need 1200 |
| 7916 | We will go through the @code{condition-case} code in a moment. First, |
| 7917 | let us look at the definition of @code{kill-region}, with comments |
| 7918 | added: |
| 7919 | |
| 7920 | @c GNU Emacs 22: |
| 7921 | @smallexample |
| 7922 | @group |
| 7923 | (defun kill-region (beg end) |
| 7924 | "Kill (\"cut\") text between point and mark. |
| 7925 | This deletes the text from the buffer and saves it in the kill ring. |
| 7926 | The command \\[yank] can retrieve it from there. @dots{} " |
| 7927 | @end group |
| 7928 | |
| 7929 | @group |
| 7930 | ;; @bullet{} Since order matters, pass point first. |
| 7931 | (interactive (list (point) (mark))) |
| 7932 | ;; @bullet{} And tell us if we cannot cut the text. |
| 7933 | ;; `unless' is an `if' without a then-part. |
| 7934 | (unless (and beg end) |
| 7935 | (error "The mark is not set now, so there is no region")) |
| 7936 | @end group |
| 7937 | |
| 7938 | @group |
| 7939 | ;; @bullet{} `condition-case' takes three arguments. |
| 7940 | ;; If the first argument is nil, as it is here, |
| 7941 | ;; information about the error signal is not |
| 7942 | ;; stored for use by another function. |
| 7943 | (condition-case nil |
| 7944 | @end group |
| 7945 | |
| 7946 | @group |
| 7947 | ;; @bullet{} The second argument to `condition-case' tells the |
| 7948 | ;; Lisp interpreter what to do when all goes well. |
| 7949 | @end group |
| 7950 | |
| 7951 | @group |
| 7952 | ;; It starts with a `let' function that extracts the string |
| 7953 | ;; and tests whether it exists. If so (that is what the |
| 7954 | ;; `when' checks), it calls an `if' function that determines |
| 7955 | ;; whether the previous command was another call to |
| 7956 | ;; `kill-region'; if it was, then the new text is appended to |
| 7957 | ;; the previous text; if not, then a different function, |
| 7958 | ;; `kill-new', is called. |
| 7959 | @end group |
| 7960 | |
| 7961 | @group |
| 7962 | ;; The `kill-append' function concatenates the new string and |
| 7963 | ;; the old. The `kill-new' function inserts text into a new |
| 7964 | ;; item in the kill ring. |
| 7965 | @end group |
| 7966 | |
| 7967 | @group |
| 7968 | ;; `when' is an `if' without an else-part. The second `when' |
| 7969 | ;; again checks whether the current string exists; in |
| 7970 | ;; addition, it checks whether the previous command was |
| 7971 | ;; another call to `kill-region'. If one or the other |
| 7972 | ;; condition is true, then it sets the current command to |
| 7973 | ;; be `kill-region'. |
| 7974 | @end group |
| 7975 | @group |
| 7976 | (let ((string (filter-buffer-substring beg end t))) |
| 7977 | (when string ;STRING is nil if BEG = END |
| 7978 | ;; Add that string to the kill ring, one way or another. |
| 7979 | (if (eq last-command 'kill-region) |
| 7980 | @end group |
| 7981 | @group |
| 7982 | ;; @minus{} `yank-handler' is an optional argument to |
| 7983 | ;; `kill-region' that tells the `kill-append' and |
| 7984 | ;; `kill-new' functions how deal with properties |
| 7985 | ;; added to the text, such as `bold' or `italics'. |
| 7986 | (kill-append string (< end beg) yank-handler) |
| 7987 | (kill-new string nil yank-handler))) |
| 7988 | (when (or string (eq last-command 'kill-region)) |
| 7989 | (setq this-command 'kill-region)) |
| 7990 | nil) |
| 7991 | @end group |
| 7992 | |
| 7993 | @group |
| 7994 | ;; @bullet{} The third argument to `condition-case' tells the interpreter |
| 7995 | ;; what to do with an error. |
| 7996 | @end group |
| 7997 | @group |
| 7998 | ;; The third argument has a conditions part and a body part. |
| 7999 | ;; If the conditions are met (in this case, |
| 8000 | ;; if text or buffer are read-only) |
| 8001 | ;; then the body is executed. |
| 8002 | @end group |
| 8003 | @group |
| 8004 | ;; The first part of the third argument is the following: |
| 8005 | ((buffer-read-only text-read-only) ;; the if-part |
| 8006 | ;; @dots{} the then-part |
| 8007 | (copy-region-as-kill beg end) |
| 8008 | @end group |
| 8009 | @group |
| 8010 | ;; Next, also as part of the then-part, set this-command, so |
| 8011 | ;; it will be set in an error |
| 8012 | (setq this-command 'kill-region) |
| 8013 | ;; Finally, in the then-part, send a message if you may copy |
| 8014 | ;; the text to the kill ring without signaling an error, but |
| 8015 | ;; don't if you may not. |
| 8016 | @end group |
| 8017 | @group |
| 8018 | (if kill-read-only-ok |
| 8019 | (progn (message "Read only text copied to kill ring") nil) |
| 8020 | (barf-if-buffer-read-only) |
| 8021 | ;; If the buffer isn't read-only, the text is. |
| 8022 | (signal 'text-read-only (list (current-buffer))))) |
| 8023 | @end group |
| 8024 | @end smallexample |
| 8025 | |
| 8026 | @ignore |
| 8027 | @c v 21 |
| 8028 | @smallexample |
| 8029 | @group |
| 8030 | (defun kill-region (beg end) |
| 8031 | "Kill between point and mark. |
| 8032 | The text is deleted but saved in the kill ring." |
| 8033 | (interactive "r") |
| 8034 | @end group |
| 8035 | |
| 8036 | @group |
| 8037 | ;; 1. `condition-case' takes three arguments. |
| 8038 | ;; If the first argument is nil, as it is here, |
| 8039 | ;; information about the error signal is not |
| 8040 | ;; stored for use by another function. |
| 8041 | (condition-case nil |
| 8042 | @end group |
| 8043 | |
| 8044 | @group |
| 8045 | ;; 2. The second argument to `condition-case' |
| 8046 | ;; tells the Lisp interpreter what to do when all goes well. |
| 8047 | @end group |
| 8048 | |
| 8049 | @group |
| 8050 | ;; The `delete-and-extract-region' function usually does the |
| 8051 | ;; work. If the beginning and ending of the region are both |
| 8052 | ;; the same, then the variable `string' will be empty, or nil |
| 8053 | (let ((string (delete-and-extract-region beg end))) |
| 8054 | @end group |
| 8055 | |
| 8056 | @group |
| 8057 | ;; `when' is an `if' clause that cannot take an `else-part'. |
| 8058 | ;; Emacs normally sets the value of `last-command' to the |
| 8059 | ;; previous command. |
| 8060 | @end group |
| 8061 | @group |
| 8062 | ;; `kill-append' concatenates the new string and the old. |
| 8063 | ;; `kill-new' inserts text into a new item in the kill ring. |
| 8064 | (when string |
| 8065 | (if (eq last-command 'kill-region) |
| 8066 | ;; if true, prepend string |
| 8067 | (kill-append string (< end beg)) |
| 8068 | (kill-new string))) |
| 8069 | (setq this-command 'kill-region)) |
| 8070 | @end group |
| 8071 | |
| 8072 | @group |
| 8073 | ;; 3. The third argument to `condition-case' tells the interpreter |
| 8074 | ;; what to do with an error. |
| 8075 | @end group |
| 8076 | @group |
| 8077 | ;; The third argument has a conditions part and a body part. |
| 8078 | ;; If the conditions are met (in this case, |
| 8079 | ;; if text or buffer are read-only) |
| 8080 | ;; then the body is executed. |
| 8081 | @end group |
| 8082 | @group |
| 8083 | ((buffer-read-only text-read-only) ;; this is the if-part |
| 8084 | ;; then... |
| 8085 | (copy-region-as-kill beg end) |
| 8086 | @end group |
| 8087 | @group |
| 8088 | (if kill-read-only-ok ;; usually this variable is nil |
| 8089 | (message "Read only text copied to kill ring") |
| 8090 | ;; or else, signal an error if the buffer is read-only; |
| 8091 | (barf-if-buffer-read-only) |
| 8092 | ;; and, in any case, signal that the text is read-only. |
| 8093 | (signal 'text-read-only (list (current-buffer))))))) |
| 8094 | @end group |
| 8095 | @end smallexample |
| 8096 | @end ignore |
| 8097 | |
| 8098 | @node condition-case |
| 8099 | @subsection @code{condition-case} |
| 8100 | @findex condition-case |
| 8101 | |
| 8102 | As we have seen earlier (@pxref{Making Errors, , Generate an Error |
| 8103 | Message}), when the Emacs Lisp interpreter has trouble evaluating an |
| 8104 | expression, it provides you with help; in the jargon, this is called |
| 8105 | ``signaling an error''. Usually, the computer stops the program and |
| 8106 | shows you a message. |
| 8107 | |
| 8108 | However, some programs undertake complicated actions. They should not |
| 8109 | simply stop on an error. In the @code{kill-region} function, the most |
| 8110 | likely error is that you will try to kill text that is read-only and |
| 8111 | cannot be removed. So the @code{kill-region} function contains code |
| 8112 | to handle this circumstance. This code, which makes up the body of |
| 8113 | the @code{kill-region} function, is inside of a @code{condition-case} |
| 8114 | special form. |
| 8115 | |
| 8116 | @need 800 |
| 8117 | The template for @code{condition-case} looks like this: |
| 8118 | |
| 8119 | @smallexample |
| 8120 | @group |
| 8121 | (condition-case |
| 8122 | @var{var} |
| 8123 | @var{bodyform} |
| 8124 | @var{error-handler}@dots{}) |
| 8125 | @end group |
| 8126 | @end smallexample |
| 8127 | |
| 8128 | The second argument, @var{bodyform}, is straightforward. The |
| 8129 | @code{condition-case} special form causes the Lisp interpreter to |
| 8130 | evaluate the code in @var{bodyform}. If no error occurs, the special |
| 8131 | form returns the code's value and produces the side-effects, if any. |
| 8132 | |
| 8133 | In short, the @var{bodyform} part of a @code{condition-case} |
| 8134 | expression determines what should happen when everything works |
| 8135 | correctly. |
| 8136 | |
| 8137 | However, if an error occurs, among its other actions, the function |
| 8138 | generating the error signal will define one or more error condition |
| 8139 | names. |
| 8140 | |
| 8141 | An error handler is the third argument to @code{condition case}. |
| 8142 | An error handler has two parts, a @var{condition-name} and a |
| 8143 | @var{body}. If the @var{condition-name} part of an error handler |
| 8144 | matches a condition name generated by an error, then the @var{body} |
| 8145 | part of the error handler is run. |
| 8146 | |
| 8147 | As you will expect, the @var{condition-name} part of an error handler |
| 8148 | may be either a single condition name or a list of condition names. |
| 8149 | |
| 8150 | Also, a complete @code{condition-case} expression may contain more |
| 8151 | than one error handler. When an error occurs, the first applicable |
| 8152 | handler is run. |
| 8153 | |
| 8154 | Lastly, the first argument to the @code{condition-case} expression, |
| 8155 | the @var{var} argument, is sometimes bound to a variable that |
| 8156 | contains information about the error. However, if that argument is |
| 8157 | nil, as is the case in @code{kill-region}, that information is |
| 8158 | discarded. |
| 8159 | |
| 8160 | @need 1200 |
| 8161 | In brief, in the @code{kill-region} function, the code |
| 8162 | @code{condition-case} works like this: |
| 8163 | |
| 8164 | @smallexample |
| 8165 | @group |
| 8166 | @var{If no errors}, @var{run only this code} |
| 8167 | @var{but}, @var{if errors}, @var{run this other code}. |
| 8168 | @end group |
| 8169 | @end smallexample |
| 8170 | |
| 8171 | @ignore |
| 8172 | 2006 Oct 24 |
| 8173 | In Emacs 22, |
| 8174 | copy-region-as-kill is short, 12 lines, and uses |
| 8175 | filter-buffer-substring, which is longer, 39 lines |
| 8176 | and has delete-and-extract-region in it. |
| 8177 | delete-and-extract-region is written in C. |
| 8178 | |
| 8179 | see Initializing a Variable with @code{defvar} |
| 8180 | this is line 8054 |
| 8181 | Initializing a Variable with @code{defvar} includes line 8350 |
| 8182 | @end ignore |
| 8183 | |
| 8184 | @node Lisp macro |
| 8185 | @subsection Lisp macro |
| 8186 | @cindex Macro, lisp |
| 8187 | @cindex Lisp macro |
| 8188 | |
| 8189 | The part of the @code{condition-case} expression that is evaluated in |
| 8190 | the expectation that all goes well has a @code{when}. The code uses |
| 8191 | @code{when} to determine whether the @code{string} variable points to |
| 8192 | text that exists. |
| 8193 | |
| 8194 | A @code{when} expression is simply a programmers' convenience. It is |
| 8195 | an @code{if} without the possibility of an else clause. In your mind, |
| 8196 | you can replace @code{when} with @code{if} and understand what goes |
| 8197 | on. That is what the Lisp interpreter does. |
| 8198 | |
| 8199 | Technically speaking, @code{when} is a Lisp macro. A Lisp macro |
| 8200 | enables you to define new control constructs and other language |
| 8201 | features. It tells the interpreter how to compute another Lisp |
| 8202 | expression which will in turn compute the value. In this case, the |
| 8203 | `other expression' is an @code{if} expression. |
| 8204 | |
| 8205 | The @code{kill-region} function definition also has an @code{unless} |
| 8206 | macro; it is the converse of @code{when}. The @code{unless} macro is |
| 8207 | an @code{if} without a then clause |
| 8208 | |
| 8209 | For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU |
| 8210 | Emacs Lisp Reference Manual}. The C programming language also |
| 8211 | provides macros. These are different, but also useful. |
| 8212 | |
| 8213 | @ignore |
| 8214 | We will briefly look at C macros in |
| 8215 | @ref{Digression into C}. |
| 8216 | @end ignore |
| 8217 | |
| 8218 | @need 1200 |
| 8219 | Regarding the @code{when} macro, in the @code{condition-case} |
| 8220 | expression, when the string has content, then another conditional |
| 8221 | expression is executed. This is an @code{if} with both a then-part |
| 8222 | and an else-part. |
| 8223 | |
| 8224 | @smallexample |
| 8225 | @group |
| 8226 | (if (eq last-command 'kill-region) |
| 8227 | (kill-append string (< end beg) yank-handler) |
| 8228 | (kill-new string nil yank-handler)) |
| 8229 | @end group |
| 8230 | @end smallexample |
| 8231 | |
| 8232 | The then-part is evaluated if the previous command was another call to |
| 8233 | @code{kill-region}; if not, the else-part is evaluated. |
| 8234 | |
| 8235 | @code{yank-handler} is an optional argument to @code{kill-region} that |
| 8236 | tells the @code{kill-append} and @code{kill-new} functions how deal |
| 8237 | with properties added to the text, such as `bold' or `italics'. |
| 8238 | |
| 8239 | @code{last-command} is a variable that comes with Emacs that we have |
| 8240 | not seen before. Normally, whenever a function is executed, Emacs |
| 8241 | sets the value of @code{last-command} to the previous command. |
| 8242 | |
| 8243 | @need 1200 |
| 8244 | In this segment of the definition, the @code{if} expression checks |
| 8245 | whether the previous command was @code{kill-region}. If it was, |
| 8246 | |
| 8247 | @smallexample |
| 8248 | (kill-append string (< end beg) yank-handler) |
| 8249 | @end smallexample |
| 8250 | |
| 8251 | @noindent |
| 8252 | concatenates a copy of the newly clipped text to the just previously |
| 8253 | clipped text in the kill ring. |
| 8254 | |
| 8255 | @node copy-region-as-kill |
| 8256 | @section @code{copy-region-as-kill} |
| 8257 | @findex copy-region-as-kill |
| 8258 | @findex nthcdr |
| 8259 | |
| 8260 | The @code{copy-region-as-kill} function copies a region of text from a |
| 8261 | buffer and (via either @code{kill-append} or @code{kill-new}) saves it |
| 8262 | in the @code{kill-ring}. |
| 8263 | |
| 8264 | If you call @code{copy-region-as-kill} immediately after a |
| 8265 | @code{kill-region} command, Emacs appends the newly copied text to the |
| 8266 | previously copied text. This means that if you yank back the text, you |
| 8267 | get it all, from both this and the previous operation. On the other |
| 8268 | hand, if some other command precedes the @code{copy-region-as-kill}, |
| 8269 | the function copies the text into a separate entry in the kill ring. |
| 8270 | |
| 8271 | @menu |
| 8272 | * Complete copy-region-as-kill:: The complete function definition. |
| 8273 | * copy-region-as-kill body:: The body of @code{copy-region-as-kill}. |
| 8274 | @end menu |
| 8275 | |
| 8276 | @ifnottex |
| 8277 | @node Complete copy-region-as-kill |
| 8278 | @unnumberedsubsec The complete @code{copy-region-as-kill} function definition |
| 8279 | @end ifnottex |
| 8280 | |
| 8281 | @need 1200 |
| 8282 | Here is the complete text of the version 22 @code{copy-region-as-kill} |
| 8283 | function: |
| 8284 | |
| 8285 | @smallexample |
| 8286 | @group |
| 8287 | (defun copy-region-as-kill (beg end) |
| 8288 | "Save the region as if killed, but don't kill it. |
| 8289 | In Transient Mark mode, deactivate the mark. |
| 8290 | If `interprogram-cut-function' is non-nil, also save the text for a window |
| 8291 | system cut and paste." |
| 8292 | (interactive "r") |
| 8293 | @end group |
| 8294 | @group |
| 8295 | (if (eq last-command 'kill-region) |
| 8296 | (kill-append (filter-buffer-substring beg end) (< end beg)) |
| 8297 | (kill-new (filter-buffer-substring beg end))) |
| 8298 | @end group |
| 8299 | @group |
| 8300 | (if transient-mark-mode |
| 8301 | (setq deactivate-mark t)) |
| 8302 | nil) |
| 8303 | @end group |
| 8304 | @end smallexample |
| 8305 | |
| 8306 | @need 800 |
| 8307 | As usual, this function can be divided into its component parts: |
| 8308 | |
| 8309 | @smallexample |
| 8310 | @group |
| 8311 | (defun copy-region-as-kill (@var{argument-list}) |
| 8312 | "@var{documentation}@dots{}" |
| 8313 | (interactive "r") |
| 8314 | @var{body}@dots{}) |
| 8315 | @end group |
| 8316 | @end smallexample |
| 8317 | |
| 8318 | The arguments are @code{beg} and @code{end} and the function is |
| 8319 | interactive with @code{"r"}, so the two arguments must refer to the |
| 8320 | beginning and end of the region. If you have been reading through this |
| 8321 | document from the beginning, understanding these parts of a function is |
| 8322 | almost becoming routine. |
| 8323 | |
| 8324 | The documentation is somewhat confusing unless you remember that the |
| 8325 | word `kill' has a meaning different from usual. The `Transient Mark' |
| 8326 | and @code{interprogram-cut-function} comments explain certain |
| 8327 | side-effects. |
| 8328 | |
| 8329 | After you once set a mark, a buffer always contains a region. If you |
| 8330 | wish, you can use Transient Mark mode to highlight the region |
| 8331 | temporarily. (No one wants to highlight the region all the time, so |
| 8332 | Transient Mark mode highlights it only at appropriate times. Many |
| 8333 | people turn off Transient Mark mode, so the region is never |
| 8334 | highlighted.) |
| 8335 | |
| 8336 | Also, a windowing system allows you to copy, cut, and paste among |
| 8337 | different programs. In the X windowing system, for example, the |
| 8338 | @code{interprogram-cut-function} function is @code{x-select-text}, |
| 8339 | which works with the windowing system's equivalent of the Emacs kill |
| 8340 | ring. |
| 8341 | |
| 8342 | The body of the @code{copy-region-as-kill} function starts with an |
| 8343 | @code{if} clause. What this clause does is distinguish between two |
| 8344 | different situations: whether or not this command is executed |
| 8345 | immediately after a previous @code{kill-region} command. In the first |
| 8346 | case, the new region is appended to the previously copied text. |
| 8347 | Otherwise, it is inserted into the beginning of the kill ring as a |
| 8348 | separate piece of text from the previous piece. |
| 8349 | |
| 8350 | The last two lines of the function prevent the region from lighting up |
| 8351 | if Transient Mark mode is turned on. |
| 8352 | |
| 8353 | The body of @code{copy-region-as-kill} merits discussion in detail. |
| 8354 | |
| 8355 | @node copy-region-as-kill body |
| 8356 | @subsection The Body of @code{copy-region-as-kill} |
| 8357 | |
| 8358 | The @code{copy-region-as-kill} function works in much the same way as |
| 8359 | the @code{kill-region} function. Both are written so that two or more |
| 8360 | kills in a row combine their text into a single entry. If you yank |
| 8361 | back the text from the kill ring, you get it all in one piece. |
| 8362 | Moreover, kills that kill forward from the current position of the |
| 8363 | cursor are added to the end of the previously copied text and commands |
| 8364 | that copy text backwards add it to the beginning of the previously |
| 8365 | copied text. This way, the words in the text stay in the proper |
| 8366 | order. |
| 8367 | |
| 8368 | Like @code{kill-region}, the @code{copy-region-as-kill} function makes |
| 8369 | use of the @code{last-command} variable that keeps track of the |
| 8370 | previous Emacs command. |
| 8371 | |
| 8372 | @menu |
| 8373 | * last-command & this-command:: |
| 8374 | * kill-append function:: |
| 8375 | * kill-new function:: |
| 8376 | @end menu |
| 8377 | |
| 8378 | @ifnottex |
| 8379 | @node last-command & this-command |
| 8380 | @unnumberedsubsubsec @code{last-command} and @code{this-command} |
| 8381 | @end ifnottex |
| 8382 | |
| 8383 | Normally, whenever a function is executed, Emacs sets the value of |
| 8384 | @code{this-command} to the function being executed (which in this case |
| 8385 | would be @code{copy-region-as-kill}). At the same time, Emacs sets |
| 8386 | the value of @code{last-command} to the previous value of |
| 8387 | @code{this-command}. |
| 8388 | |
| 8389 | In the first part of the body of the @code{copy-region-as-kill} |
| 8390 | function, an @code{if} expression determines whether the value of |
| 8391 | @code{last-command} is @code{kill-region}. If so, the then-part of |
| 8392 | the @code{if} expression is evaluated; it uses the @code{kill-append} |
| 8393 | function to concatenate the text copied at this call to the function |
| 8394 | with the text already in the first element (the @sc{car}) of the kill |
| 8395 | ring. On the other hand, if the value of @code{last-command} is not |
| 8396 | @code{kill-region}, then the @code{copy-region-as-kill} function |
| 8397 | attaches a new element to the kill ring using the @code{kill-new} |
| 8398 | function. |
| 8399 | |
| 8400 | @need 1250 |
| 8401 | The @code{if} expression reads as follows; it uses @code{eq}: |
| 8402 | |
| 8403 | @smallexample |
| 8404 | @group |
| 8405 | (if (eq last-command 'kill-region) |
| 8406 | ;; @r{then-part} |
| 8407 | (kill-append (filter-buffer-substring beg end) (< end beg)) |
| 8408 | ;; @r{else-part} |
| 8409 | (kill-new (filter-buffer-substring beg end))) |
| 8410 | @end group |
| 8411 | @end smallexample |
| 8412 | |
| 8413 | @findex filter-buffer-substring |
| 8414 | (The @code{filter-buffer-substring} function returns a filtered |
| 8415 | substring of the buffer, if any. Optionally---the arguments are not |
| 8416 | here, so neither is done---the function may delete the initial text or |
| 8417 | return the text without its properties; this function is a replacement |
| 8418 | for the older @code{buffer-substring} function, which came before text |
| 8419 | properties were implemented.) |
| 8420 | |
| 8421 | @findex eq @r{(example of use)} |
| 8422 | @noindent |
| 8423 | The @code{eq} function tests whether its first argument is the same Lisp |
| 8424 | object as its second argument. The @code{eq} function is similar to the |
| 8425 | @code{equal} function in that it is used to test for equality, but |
| 8426 | differs in that it determines whether two representations are actually |
| 8427 | the same object inside the computer, but with different names. |
| 8428 | @code{equal} determines whether the structure and contents of two |
| 8429 | expressions are the same. |
| 8430 | |
| 8431 | If the previous command was @code{kill-region}, then the Emacs Lisp |
| 8432 | interpreter calls the @code{kill-append} function |
| 8433 | |
| 8434 | @node kill-append function |
| 8435 | @unnumberedsubsubsec The @code{kill-append} function |
| 8436 | @findex kill-append |
| 8437 | |
| 8438 | @need 800 |
| 8439 | The @code{kill-append} function looks like this: |
| 8440 | |
| 8441 | @c in GNU Emacs 22 |
| 8442 | @smallexample |
| 8443 | @group |
| 8444 | (defun kill-append (string before-p &optional yank-handler) |
| 8445 | "Append STRING to the end of the latest kill in the kill ring. |
| 8446 | If BEFORE-P is non-nil, prepend STRING to the kill. |
| 8447 | @dots{} " |
| 8448 | (let* ((cur (car kill-ring))) |
| 8449 | (kill-new (if before-p (concat string cur) (concat cur string)) |
| 8450 | (or (= (length cur) 0) |
| 8451 | (equal yank-handler |
| 8452 | (get-text-property 0 'yank-handler cur))) |
| 8453 | yank-handler))) |
| 8454 | @end group |
| 8455 | @end smallexample |
| 8456 | |
| 8457 | @ignore |
| 8458 | was: |
| 8459 | (defun kill-append (string before-p) |
| 8460 | "Append STRING to the end of the latest kill in the kill ring. |
| 8461 | If BEFORE-P is non-nil, prepend STRING to the kill. |
| 8462 | If `interprogram-cut-function' is set, pass the resulting kill to |
| 8463 | it." |
| 8464 | (kill-new (if before-p |
| 8465 | (concat string (car kill-ring)) |
| 8466 | (concat (car kill-ring) string)) |
| 8467 | t)) |
| 8468 | @end ignore |
| 8469 | |
| 8470 | @noindent |
| 8471 | The @code{kill-append} function is fairly straightforward. It uses |
| 8472 | the @code{kill-new} function, which we will discuss in more detail in |
| 8473 | a moment. |
| 8474 | |
| 8475 | (Also, the function provides an optional argument called |
| 8476 | @code{yank-handler}; when invoked, this argument tells the function |
| 8477 | how to deal with properties added to the text, such as `bold' or |
| 8478 | `italics'.) |
| 8479 | |
| 8480 | @c !!! bug in GNU Emacs 22 version of kill-append ? |
| 8481 | It has a @code{let*} function to set the value of the first element of |
| 8482 | the kill ring to @code{cur}. (I do not know why the function does not |
| 8483 | use @code{let} instead; only one value is set in the expression. |
| 8484 | Perhaps this is a bug that produces no problems?) |
| 8485 | |
| 8486 | Consider the conditional that is one of the two arguments to |
| 8487 | @code{kill-new}. It uses @code{concat} to concatenate the new text to |
| 8488 | the @sc{car} of the kill ring. Whether it prepends or appends the |
| 8489 | text depends on the results of an @code{if} expression: |
| 8490 | |
| 8491 | @smallexample |
| 8492 | @group |
| 8493 | (if before-p ; @r{if-part} |
| 8494 | (concat string cur) ; @r{then-part} |
| 8495 | (concat cur string)) ; @r{else-part} |
| 8496 | @end group |
| 8497 | @end smallexample |
| 8498 | |
| 8499 | @noindent |
| 8500 | If the region being killed is before the region that was killed in the |
| 8501 | last command, then it should be prepended before the material that was |
| 8502 | saved in the previous kill; and conversely, if the killed text follows |
| 8503 | what was just killed, it should be appended after the previous text. |
| 8504 | The @code{if} expression depends on the predicate @code{before-p} to |
| 8505 | decide whether the newly saved text should be put before or after the |
| 8506 | previously saved text. |
| 8507 | |
| 8508 | The symbol @code{before-p} is the name of one of the arguments to |
| 8509 | @code{kill-append}. When the @code{kill-append} function is |
| 8510 | evaluated, it is bound to the value returned by evaluating the actual |
| 8511 | argument. In this case, this is the expression @code{(< end beg)}. |
| 8512 | This expression does not directly determine whether the killed text in |
| 8513 | this command is located before or after the kill text of the last |
| 8514 | command; what it does is determine whether the value of the variable |
| 8515 | @code{end} is less than the value of the variable @code{beg}. If it |
| 8516 | is, it means that the user is most likely heading towards the |
| 8517 | beginning of the buffer. Also, the result of evaluating the predicate |
| 8518 | expression, @code{(< end beg)}, will be true and the text will be |
| 8519 | prepended before the previous text. On the other hand, if the value of |
| 8520 | the variable @code{end} is greater than the value of the variable |
| 8521 | @code{beg}, the text will be appended after the previous text. |
| 8522 | |
| 8523 | @need 800 |
| 8524 | When the newly saved text will be prepended, then the string with the new |
| 8525 | text will be concatenated before the old text: |
| 8526 | |
| 8527 | @smallexample |
| 8528 | (concat string cur) |
| 8529 | @end smallexample |
| 8530 | |
| 8531 | @need 1200 |
| 8532 | @noindent |
| 8533 | But if the text will be appended, it will be concatenated |
| 8534 | after the old text: |
| 8535 | |
| 8536 | @smallexample |
| 8537 | (concat cur string)) |
| 8538 | @end smallexample |
| 8539 | |
| 8540 | To understand how this works, we first need to review the |
| 8541 | @code{concat} function. The @code{concat} function links together or |
| 8542 | unites two strings of text. The result is a string. For example: |
| 8543 | |
| 8544 | @smallexample |
| 8545 | @group |
| 8546 | (concat "abc" "def") |
| 8547 | @result{} "abcdef" |
| 8548 | @end group |
| 8549 | |
| 8550 | @group |
| 8551 | (concat "new " |
| 8552 | (car '("first element" "second element"))) |
| 8553 | @result{} "new first element" |
| 8554 | |
| 8555 | (concat (car |
| 8556 | '("first element" "second element")) " modified") |
| 8557 | @result{} "first element modified" |
| 8558 | @end group |
| 8559 | @end smallexample |
| 8560 | |
| 8561 | We can now make sense of @code{kill-append}: it modifies the contents |
| 8562 | of the kill ring. The kill ring is a list, each element of which is |
| 8563 | saved text. The @code{kill-append} function uses the @code{kill-new} |
| 8564 | function which in turn uses the @code{setcar} function. |
| 8565 | |
| 8566 | @node kill-new function |
| 8567 | @unnumberedsubsubsec The @code{kill-new} function |
| 8568 | @findex kill-new |
| 8569 | |
| 8570 | @c in GNU Emacs 22, additional documentation to kill-new: |
| 8571 | @ignore |
| 8572 | Optional third arguments YANK-HANDLER controls how the STRING is later |
| 8573 | inserted into a buffer; see `insert-for-yank' for details. |
| 8574 | When a yank handler is specified, STRING must be non-empty (the yank |
| 8575 | handler, if non-nil, is stored as a `yank-handler' text property on STRING). |
| 8576 | |
| 8577 | When the yank handler has a non-nil PARAM element, the original STRING |
| 8578 | argument is not used by `insert-for-yank'. However, since Lisp code |
| 8579 | may access and use elements from the kill ring directly, the STRING |
| 8580 | argument should still be a \"useful\" string for such uses." |
| 8581 | @end ignore |
| 8582 | @need 1200 |
| 8583 | The @code{kill-new} function looks like this: |
| 8584 | |
| 8585 | @smallexample |
| 8586 | @group |
| 8587 | (defun kill-new (string &optional replace yank-handler) |
| 8588 | "Make STRING the latest kill in the kill ring. |
| 8589 | Set `kill-ring-yank-pointer' to point to it. |
| 8590 | |
| 8591 | If `interprogram-cut-function' is non-nil, apply it to STRING. |
| 8592 | Optional second argument REPLACE non-nil means that STRING will replace |
| 8593 | the front of the kill ring, rather than being added to the list. |
| 8594 | @dots{}" |
| 8595 | @end group |
| 8596 | @group |
| 8597 | (if (> (length string) 0) |
| 8598 | (if yank-handler |
| 8599 | (put-text-property 0 (length string) |
| 8600 | 'yank-handler yank-handler string)) |
| 8601 | (if yank-handler |
| 8602 | (signal 'args-out-of-range |
| 8603 | (list string "yank-handler specified for empty string")))) |
| 8604 | @end group |
| 8605 | @group |
| 8606 | (if (fboundp 'menu-bar-update-yank-menu) |
| 8607 | (menu-bar-update-yank-menu string (and replace (car kill-ring)))) |
| 8608 | @end group |
| 8609 | @group |
| 8610 | (if (and replace kill-ring) |
| 8611 | (setcar kill-ring string) |
| 8612 | (push string kill-ring) |
| 8613 | (if (> (length kill-ring) kill-ring-max) |
| 8614 | (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))) |
| 8615 | @end group |
| 8616 | @group |
| 8617 | (setq kill-ring-yank-pointer kill-ring) |
| 8618 | (if interprogram-cut-function |
| 8619 | (funcall interprogram-cut-function string (not replace)))) |
| 8620 | @end group |
| 8621 | @end smallexample |
| 8622 | @ignore |
| 8623 | was: |
| 8624 | (defun kill-new (string &optional replace) |
| 8625 | "Make STRING the latest kill in the kill ring. |
| 8626 | Set the kill-ring-yank pointer to point to it. |
| 8627 | If `interprogram-cut-function' is non-nil, apply it to STRING. |
| 8628 | Optional second argument REPLACE non-nil means that STRING will replace |
| 8629 | the front of the kill ring, rather than being added to the list." |
| 8630 | (and (fboundp 'menu-bar-update-yank-menu) |
| 8631 | (menu-bar-update-yank-menu string (and replace (car kill-ring)))) |
| 8632 | (if (and replace kill-ring) |
| 8633 | (setcar kill-ring string) |
| 8634 | (setq kill-ring (cons string kill-ring)) |
| 8635 | (if (> (length kill-ring) kill-ring-max) |
| 8636 | (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))) |
| 8637 | (setq kill-ring-yank-pointer kill-ring) |
| 8638 | (if interprogram-cut-function |
| 8639 | (funcall interprogram-cut-function string (not replace)))) |
| 8640 | @end ignore |
| 8641 | |
| 8642 | (Notice that the function is not interactive.) |
| 8643 | |
| 8644 | As usual, we can look at this function in parts. |
| 8645 | |
| 8646 | The function definition has an optional @code{yank-handler} argument, |
| 8647 | which when invoked tells the function how to deal with properties |
| 8648 | added to the text, such as `bold' or `italics'. We will skip that. |
| 8649 | |
| 8650 | @need 1200 |
| 8651 | The first line of the documentation makes sense: |
| 8652 | |
| 8653 | @smallexample |
| 8654 | Make STRING the latest kill in the kill ring. |
| 8655 | @end smallexample |
| 8656 | |
| 8657 | @noindent |
| 8658 | Let's skip over the rest of the documentation for the moment. |
| 8659 | |
| 8660 | @noindent |
| 8661 | Also, let's skip over the initial @code{if} expression and those lines |
| 8662 | of code involving @code{menu-bar-update-yank-menu}. We will explain |
| 8663 | them below. |
| 8664 | |
| 8665 | @need 1200 |
| 8666 | The critical lines are these: |
| 8667 | |
| 8668 | @smallexample |
| 8669 | @group |
| 8670 | (if (and replace kill-ring) |
| 8671 | ;; @r{then} |
| 8672 | (setcar kill-ring string) |
| 8673 | @end group |
| 8674 | @group |
| 8675 | ;; @r{else} |
| 8676 | (push string kill-ring) |
| 8677 | @end group |
| 8678 | @group |
| 8679 | (setq kill-ring (cons string kill-ring)) |
| 8680 | (if (> (length kill-ring) kill-ring-max) |
| 8681 | ;; @r{avoid overly long kill ring} |
| 8682 | (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))) |
| 8683 | @end group |
| 8684 | @group |
| 8685 | (setq kill-ring-yank-pointer kill-ring) |
| 8686 | (if interprogram-cut-function |
| 8687 | (funcall interprogram-cut-function string (not replace)))) |
| 8688 | @end group |
| 8689 | @end smallexample |
| 8690 | |
| 8691 | The conditional test is @w{@code{(and replace kill-ring)}}. |
| 8692 | This will be true when two conditions are met: the kill ring has |
| 8693 | something in it, and the @code{replace} variable is true. |
| 8694 | |
| 8695 | @need 1250 |
| 8696 | When the @code{kill-append} function sets @code{replace} to be true |
| 8697 | and when the kill ring has at least one item in it, the @code{setcar} |
| 8698 | expression is executed: |
| 8699 | |
| 8700 | @smallexample |
| 8701 | (setcar kill-ring string) |
| 8702 | @end smallexample |
| 8703 | |
| 8704 | The @code{setcar} function actually changes the first element of the |
| 8705 | @code{kill-ring} list to the value of @code{string}. It replaces the |
| 8706 | first element. |
| 8707 | |
| 8708 | @need 1250 |
| 8709 | On the other hand, if the kill ring is empty, or replace is false, the |
| 8710 | else-part of the condition is executed: |
| 8711 | |
| 8712 | @smallexample |
| 8713 | (push string kill-ring) |
| 8714 | @end smallexample |
| 8715 | |
| 8716 | @noindent |
| 8717 | @need 1250 |
| 8718 | @code{push} puts its first argument onto the second. It is similar to |
| 8719 | the older |
| 8720 | |
| 8721 | @smallexample |
| 8722 | (setq kill-ring (cons string kill-ring)) |
| 8723 | @end smallexample |
| 8724 | |
| 8725 | @noindent |
| 8726 | @need 1250 |
| 8727 | or the newer |
| 8728 | |
| 8729 | @smallexample |
| 8730 | (add-to-list kill-ring string) |
| 8731 | @end smallexample |
| 8732 | |
| 8733 | @noindent |
| 8734 | When it is false, the expression first constructs a new version of the |
| 8735 | kill ring by prepending @code{string} to the existing kill ring as a |
| 8736 | new element (that is what the @code{push} does). Then it executes a |
| 8737 | second @code{if} clause. This second @code{if} clause keeps the kill |
| 8738 | ring from growing too long. |
| 8739 | |
| 8740 | Let's look at these two expressions in order. |
| 8741 | |
| 8742 | The @code{push} line of the else-part sets the new value of the kill |
| 8743 | ring to what results from adding the string being killed to the old |
| 8744 | kill ring. |
| 8745 | |
| 8746 | We can see how this works with an example. |
| 8747 | |
| 8748 | @need 800 |
| 8749 | First, |
| 8750 | |
| 8751 | @smallexample |
| 8752 | (setq example-list '("here is a clause" "another clause")) |
| 8753 | @end smallexample |
| 8754 | |
| 8755 | @need 1200 |
| 8756 | @noindent |
| 8757 | After evaluating this expression with @kbd{C-x C-e}, you can evaluate |
| 8758 | @code{example-list} and see what it returns: |
| 8759 | |
| 8760 | @smallexample |
| 8761 | @group |
| 8762 | example-list |
| 8763 | @result{} ("here is a clause" "another clause") |
| 8764 | @end group |
| 8765 | @end smallexample |
| 8766 | |
| 8767 | @need 1200 |
| 8768 | @noindent |
| 8769 | Now, we can add a new element on to this list by evaluating the |
| 8770 | following expression: |
| 8771 | @findex push, @r{example} |
| 8772 | |
| 8773 | @smallexample |
| 8774 | (push "a third clause" example-list) |
| 8775 | @end smallexample |
| 8776 | |
| 8777 | @need 800 |
| 8778 | @noindent |
| 8779 | When we evaluate @code{example-list}, we find its value is: |
| 8780 | |
| 8781 | @smallexample |
| 8782 | @group |
| 8783 | example-list |
| 8784 | @result{} ("a third clause" "here is a clause" "another clause") |
| 8785 | @end group |
| 8786 | @end smallexample |
| 8787 | |
| 8788 | @noindent |
| 8789 | Thus, the third clause is added to the list by @code{push}. |
| 8790 | |
| 8791 | @need 1200 |
| 8792 | Now for the second part of the @code{if} clause. This expression |
| 8793 | keeps the kill ring from growing too long. It looks like this: |
| 8794 | |
| 8795 | @smallexample |
| 8796 | @group |
| 8797 | (if (> (length kill-ring) kill-ring-max) |
| 8798 | (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)) |
| 8799 | @end group |
| 8800 | @end smallexample |
| 8801 | |
| 8802 | The code checks whether the length of the kill ring is greater than |
| 8803 | the maximum permitted length. This is the value of |
| 8804 | @code{kill-ring-max} (which is 60, by default). If the length of the |
| 8805 | kill ring is too long, then this code sets the last element of the |
| 8806 | kill ring to @code{nil}. It does this by using two functions, |
| 8807 | @code{nthcdr} and @code{setcdr}. |
| 8808 | |
| 8809 | We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}). |
| 8810 | It sets the @sc{cdr} of a list, just as @code{setcar} sets the |
| 8811 | @sc{car} of a list. In this case, however, @code{setcdr} will not be |
| 8812 | setting the @sc{cdr} of the whole kill ring; the @code{nthcdr} |
| 8813 | function is used to cause it to set the @sc{cdr} of the next to last |
| 8814 | element of the kill ring---this means that since the @sc{cdr} of the |
| 8815 | next to last element is the last element of the kill ring, it will set |
| 8816 | the last element of the kill ring. |
| 8817 | |
| 8818 | @findex nthcdr, @r{example} |
| 8819 | The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a |
| 8820 | list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr} |
| 8821 | @dots{} It does this @var{N} times and returns the results. |
| 8822 | (@xref{nthcdr, , @code{nthcdr}}.) |
| 8823 | |
| 8824 | @findex setcdr, @r{example} |
| 8825 | Thus, if we had a four element list that was supposed to be three |
| 8826 | elements long, we could set the @sc{cdr} of the next to last element |
| 8827 | to @code{nil}, and thereby shorten the list. (If you set the last |
| 8828 | element to some other value than @code{nil}, which you could do, then |
| 8829 | you would not have shortened the list. @xref{setcdr, , |
| 8830 | @code{setcdr}}.) |
| 8831 | |
| 8832 | You can see shortening by evaluating the following three expressions |
| 8833 | in turn. First set the value of @code{trees} to @code{(maple oak pine |
| 8834 | birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil} |
| 8835 | and then find the value of @code{trees}: |
| 8836 | |
| 8837 | @smallexample |
| 8838 | @group |
| 8839 | (setq trees '(maple oak pine birch)) |
| 8840 | @result{} (maple oak pine birch) |
| 8841 | @end group |
| 8842 | |
| 8843 | @group |
| 8844 | (setcdr (nthcdr 2 trees) nil) |
| 8845 | @result{} nil |
| 8846 | |
| 8847 | trees |
| 8848 | @result{} (maple oak pine) |
| 8849 | @end group |
| 8850 | @end smallexample |
| 8851 | |
| 8852 | @noindent |
| 8853 | (The value returned by the @code{setcdr} expression is @code{nil} since |
| 8854 | that is what the @sc{cdr} is set to.) |
| 8855 | |
| 8856 | To repeat, in @code{kill-new}, the @code{nthcdr} function takes the |
| 8857 | @sc{cdr} a number of times that is one less than the maximum permitted |
| 8858 | size of the kill ring and @code{setcdr} sets the @sc{cdr} of that |
| 8859 | element (which will be the rest of the elements in the kill ring) to |
| 8860 | @code{nil}. This prevents the kill ring from growing too long. |
| 8861 | |
| 8862 | @need 800 |
| 8863 | The next to last expression in the @code{kill-new} function is |
| 8864 | |
| 8865 | @smallexample |
| 8866 | (setq kill-ring-yank-pointer kill-ring) |
| 8867 | @end smallexample |
| 8868 | |
| 8869 | The @code{kill-ring-yank-pointer} is a global variable that is set to be |
| 8870 | the @code{kill-ring}. |
| 8871 | |
| 8872 | Even though the @code{kill-ring-yank-pointer} is called a |
| 8873 | @samp{pointer}, it is a variable just like the kill ring. However, the |
| 8874 | name has been chosen to help humans understand how the variable is used. |
| 8875 | |
| 8876 | @need 1200 |
| 8877 | Now, to return to an early expression in the body of the function: |
| 8878 | |
| 8879 | @smallexample |
| 8880 | @group |
| 8881 | (if (fboundp 'menu-bar-update-yank-menu) |
| 8882 | (menu-bar-update-yank-menu string (and replace (car kill-ring)))) |
| 8883 | @end group |
| 8884 | @end smallexample |
| 8885 | |
| 8886 | @noindent |
| 8887 | It starts with an @code{if} expression |
| 8888 | |
| 8889 | In this case, the expression tests first to see whether |
| 8890 | @code{menu-bar-update-yank-menu} exists as a function, and if so, |
| 8891 | calls it. The @code{fboundp} function returns true if the symbol it |
| 8892 | is testing has a function definition that `is not void'. If the |
| 8893 | symbol's function definition were void, we would receive an error |
| 8894 | message, as we did when we created errors intentionally (@pxref{Making |
| 8895 | Errors, , Generate an Error Message}). |
| 8896 | |
| 8897 | @noindent |
| 8898 | The then-part contains an expression whose first element is the |
| 8899 | function @code{and}. |
| 8900 | |
| 8901 | @findex and |
| 8902 | The @code{and} special form evaluates each of its arguments until one |
| 8903 | of the arguments returns a value of @code{nil}, in which case the |
| 8904 | @code{and} expression returns @code{nil}; however, if none of the |
| 8905 | arguments returns a value of @code{nil}, the value resulting from |
| 8906 | evaluating the last argument is returned. (Since such a value is not |
| 8907 | @code{nil}, it is considered true in Emacs Lisp.) In other words, an |
| 8908 | @code{and} expression returns a true value only if all its arguments |
| 8909 | are true. (@xref{Second Buffer Related Review}.) |
| 8910 | |
| 8911 | The expression determines whether the second argument to |
| 8912 | @code{menu-bar-update-yank-menu} is true or not. |
| 8913 | @ignore |
| 8914 | ;; If we're supposed to be extending an existing string, and that |
| 8915 | ;; string really is at the front of the menu, then update it in place. |
| 8916 | @end ignore |
| 8917 | |
| 8918 | @code{menu-bar-update-yank-menu} is one of the functions that make it |
| 8919 | possible to use the `Select and Paste' menu in the Edit item of a menu |
| 8920 | bar; using a mouse, you can look at the various pieces of text you |
| 8921 | have saved and select one piece to paste. |
| 8922 | |
| 8923 | The last expression in the @code{kill-new} function adds the newly |
| 8924 | copied string to whatever facility exists for copying and pasting |
| 8925 | among different programs running in a windowing system. In the X |
| 8926 | Windowing system, for example, the @code{x-select-text} function takes |
| 8927 | the string and stores it in memory operated by X@. You can paste the |
| 8928 | string in another program, such as an Xterm. |
| 8929 | |
| 8930 | @need 1200 |
| 8931 | The expression looks like this: |
| 8932 | |
| 8933 | @smallexample |
| 8934 | @group |
| 8935 | (if interprogram-cut-function |
| 8936 | (funcall interprogram-cut-function string (not replace)))) |
| 8937 | @end group |
| 8938 | @end smallexample |
| 8939 | |
| 8940 | If an @code{interprogram-cut-function} exists, then Emacs executes |
| 8941 | @code{funcall}, which in turn calls its first argument as a function |
| 8942 | and passes the remaining arguments to it. (Incidentally, as far as I |
| 8943 | can see, this @code{if} expression could be replaced by an @code{and} |
| 8944 | expression similar to the one in the first part of the function.) |
| 8945 | |
| 8946 | We are not going to discuss windowing systems and other programs |
| 8947 | further, but merely note that this is a mechanism that enables GNU |
| 8948 | Emacs to work easily and well with other programs. |
| 8949 | |
| 8950 | This code for placing text in the kill ring, either concatenated with |
| 8951 | an existing element or as a new element, leads us to the code for |
| 8952 | bringing back text that has been cut out of the buffer---the yank |
| 8953 | commands. However, before discussing the yank commands, it is better |
| 8954 | to learn how lists are implemented in a computer. This will make |
| 8955 | clear such mysteries as the use of the term `pointer'. But before |
| 8956 | that, we will digress into C. |
| 8957 | |
| 8958 | @ignore |
| 8959 | @c is this true in Emacs 22? Does not seems to be |
| 8960 | |
| 8961 | (If the @w{@code{(< end beg))}} |
| 8962 | expression is true, @code{kill-append} prepends the string to the just |
| 8963 | previously clipped text. For a detailed discussion, see |
| 8964 | @ref{kill-append function, , The @code{kill-append} function}.) |
| 8965 | |
| 8966 | If you then yank back the text, i.e., `paste' it, you get both |
| 8967 | pieces of text at once. That way, if you delete two words in a row, |
| 8968 | and then yank them back, you get both words, in their proper order, |
| 8969 | with one yank. (The @w{@code{(< end beg))}} expression makes sure the |
| 8970 | order is correct.) |
| 8971 | |
| 8972 | On the other hand, if the previous command is not @code{kill-region}, |
| 8973 | then the @code{kill-new} function is called, which adds the text to |
| 8974 | the kill ring as the latest item, and sets the |
| 8975 | @code{kill-ring-yank-pointer} variable to point to it. |
| 8976 | @end ignore |
| 8977 | @ignore |
| 8978 | |
| 8979 | @c Evidently, changed for Emacs 22. The zap-to-char command does not |
| 8980 | @c use the delete-and-extract-region function |
| 8981 | |
| 8982 | 2006 Oct 26, the Digression into C is now OK but should come after |
| 8983 | copy-region-as-kill and filter-buffer-substring |
| 8984 | |
| 8985 | 2006 Oct 24 |
| 8986 | In Emacs 22, |
| 8987 | copy-region-as-kill is short, 12 lines, and uses |
| 8988 | filter-buffer-substring, which is longer, 39 lines |
| 8989 | and has delete-and-extract-region in it. |
| 8990 | delete-and-extract-region is written in C. |
| 8991 | |
| 8992 | see Initializing a Variable with @code{defvar} |
| 8993 | @end ignore |
| 8994 | |
| 8995 | @node Digression into C |
| 8996 | @section Digression into C |
| 8997 | @findex delete-and-extract-region |
| 8998 | @cindex C, a digression into |
| 8999 | @cindex Digression into C |
| 9000 | |
| 9001 | The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, , |
| 9002 | @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring} |
| 9003 | function, which in turn uses the @code{delete-and-extract-region} |
| 9004 | function. It removes the contents of a region and you cannot get them |
| 9005 | back. |
| 9006 | |
| 9007 | Unlike the other code discussed here, the |
| 9008 | @code{delete-and-extract-region} function is not written in Emacs |
| 9009 | Lisp; it is written in C and is one of the primitives of the GNU Emacs |
| 9010 | system. Since it is very simple, I will digress briefly from Lisp and |
| 9011 | describe it here. |
| 9012 | |
| 9013 | @c GNU Emacs 24 in src/editfns.c |
| 9014 | @c the DEFUN for delete-and-extract-region |
| 9015 | |
| 9016 | @need 1500 |
| 9017 | Like many of the other Emacs primitives, |
| 9018 | @code{delete-and-extract-region} is written as an instance of a C |
| 9019 | macro, a macro being a template for code. The complete macro looks |
| 9020 | like this: |
| 9021 | |
| 9022 | @smallexample |
| 9023 | @group |
| 9024 | DEFUN ("delete-and-extract-region", Fdelete_and_extract_region, |
| 9025 | Sdelete_and_extract_region, 2, 2, 0, |
| 9026 | doc: /* Delete the text between START and END and return it. */) |
| 9027 | (Lisp_Object start, Lisp_Object end) |
| 9028 | @{ |
| 9029 | validate_region (&start, &end); |
| 9030 | if (XINT (start) == XINT (end)) |
| 9031 | return empty_unibyte_string; |
| 9032 | return del_range_1 (XINT (start), XINT (end), 1, 1); |
| 9033 | @} |
| 9034 | @end group |
| 9035 | @end smallexample |
| 9036 | |
| 9037 | Without going into the details of the macro writing process, let me |
| 9038 | point out that this macro starts with the word @code{DEFUN}. The word |
| 9039 | @code{DEFUN} was chosen since the code serves the same purpose as |
| 9040 | @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in |
| 9041 | @file{emacs/src/lisp.h}.) |
| 9042 | |
| 9043 | The word @code{DEFUN} is followed by seven parts inside of |
| 9044 | parentheses: |
| 9045 | |
| 9046 | @itemize @bullet |
| 9047 | @item |
| 9048 | The first part is the name given to the function in Lisp, |
| 9049 | @code{delete-and-extract-region}. |
| 9050 | |
| 9051 | @item |
| 9052 | The second part is the name of the function in C, |
| 9053 | @code{Fdelete_and_extract_region}. By convention, it starts with |
| 9054 | @samp{F}. Since C does not use hyphens in names, underscores are used |
| 9055 | instead. |
| 9056 | |
| 9057 | @item |
| 9058 | The third part is the name for the C constant structure that records |
| 9059 | information on this function for internal use. It is the name of the |
| 9060 | function in C but begins with an @samp{S} instead of an @samp{F}. |
| 9061 | |
| 9062 | @item |
| 9063 | The fourth and fifth parts specify the minimum and maximum number of |
| 9064 | arguments the function can have. This function demands exactly 2 |
| 9065 | arguments. |
| 9066 | |
| 9067 | @item |
| 9068 | The sixth part is nearly like the argument that follows the |
| 9069 | @code{interactive} declaration in a function written in Lisp: a letter |
| 9070 | followed, perhaps, by a prompt. The only difference from the Lisp is |
| 9071 | when the macro is called with no arguments. Then you write a @code{0} |
| 9072 | (which is a `null string'), as in this macro. |
| 9073 | |
| 9074 | If you were to specify arguments, you would place them between |
| 9075 | quotation marks. The C macro for @code{goto-char} includes |
| 9076 | @code{"NGoto char: "} in this position to indicate that the function |
| 9077 | expects a raw prefix, in this case, a numerical location in a buffer, |
| 9078 | and provides a prompt. |
| 9079 | |
| 9080 | @item |
| 9081 | The seventh part is a documentation string, just like the one for a |
| 9082 | function written in Emacs Lisp. This is written as a C comment. (When |
| 9083 | you build Emacs, the program @command{lib-src/make-docfile} extracts |
| 9084 | these comments and uses them to make the ``real'' documentation.) |
| 9085 | @end itemize |
| 9086 | |
| 9087 | @need 1200 |
| 9088 | In a C macro, the formal parameters come next, with a statement of |
| 9089 | what kind of object they are, followed by what might be called the `body' |
| 9090 | of the macro. For @code{delete-and-extract-region} the `body' |
| 9091 | consists of the following four lines: |
| 9092 | |
| 9093 | @smallexample |
| 9094 | @group |
| 9095 | validate_region (&start, &end); |
| 9096 | if (XINT (start) == XINT (end)) |
| 9097 | return empty_unibyte_string; |
| 9098 | return del_range_1 (XINT (start), XINT (end), 1, 1); |
| 9099 | @end group |
| 9100 | @end smallexample |
| 9101 | |
| 9102 | The @code{validate_region} function checks whether the values |
| 9103 | passed as the beginning and end of the region are the proper type and |
| 9104 | are within range. If the beginning and end positions are the same, |
| 9105 | then return an empty string. |
| 9106 | |
| 9107 | The @code{del_range_1} function actually deletes the text. It is a |
| 9108 | complex function we will not look into. It updates the buffer and |
| 9109 | does other things. However, it is worth looking at the two arguments |
| 9110 | passed to @code{del_range}. These are @w{@code{XINT (start)}} and |
| 9111 | @w{@code{XINT (end)}}. |
| 9112 | |
| 9113 | As far as the C language is concerned, @code{start} and @code{end} are |
| 9114 | two integers that mark the beginning and end of the region to be |
| 9115 | deleted@footnote{More precisely, and requiring more expert knowledge |
| 9116 | to understand, the two integers are of type `Lisp_Object', which can |
| 9117 | also be a C union instead of an integer type.}. |
| 9118 | |
| 9119 | In early versions of Emacs, these two numbers were thirty-two bits |
| 9120 | long, but the code is slowly being generalized to handle other |
| 9121 | lengths. Three of the available bits are used to specify the type of |
| 9122 | information; the remaining bits are used as `content'. |
| 9123 | |
| 9124 | @samp{XINT} is a C macro that extracts the relevant number from the |
| 9125 | longer collection of bits; the three other bits are discarded. |
| 9126 | |
| 9127 | @need 800 |
| 9128 | The command in @code{delete-and-extract-region} looks like this: |
| 9129 | |
| 9130 | @smallexample |
| 9131 | del_range_1 (XINT (start), XINT (end), 1, 1); |
| 9132 | @end smallexample |
| 9133 | |
| 9134 | @noindent |
| 9135 | It deletes the region between the beginning position, @code{start}, |
| 9136 | and the ending position, @code{end}. |
| 9137 | |
| 9138 | From the point of view of the person writing Lisp, Emacs is all very |
| 9139 | simple; but hidden underneath is a great deal of complexity to make it |
| 9140 | all work. |
| 9141 | |
| 9142 | @node defvar |
| 9143 | @section Initializing a Variable with @code{defvar} |
| 9144 | @findex defvar |
| 9145 | @cindex Initializing a variable |
| 9146 | @cindex Variable initialization |
| 9147 | |
| 9148 | @ignore |
| 9149 | 2006 Oct 24 |
| 9150 | In Emacs 22, |
| 9151 | copy-region-as-kill is short, 12 lines, and uses |
| 9152 | filter-buffer-substring, which is longer, 39 lines |
| 9153 | and has delete-and-extract-region in it. |
| 9154 | delete-and-extract-region is written in C. |
| 9155 | |
| 9156 | see Initializing a Variable with @code{defvar} |
| 9157 | |
| 9158 | @end ignore |
| 9159 | |
| 9160 | The @code{copy-region-as-kill} function is written in Emacs Lisp. Two |
| 9161 | functions within it, @code{kill-append} and @code{kill-new}, copy a |
| 9162 | region in a buffer and save it in a variable called the |
| 9163 | @code{kill-ring}. This section describes how the @code{kill-ring} |
| 9164 | variable is created and initialized using the @code{defvar} special |
| 9165 | form. |
| 9166 | |
| 9167 | (Again we note that the term @code{kill-ring} is a misnomer. The text |
| 9168 | that is clipped out of the buffer can be brought back; it is not a ring |
| 9169 | of corpses, but a ring of resurrectable text.) |
| 9170 | |
| 9171 | In Emacs Lisp, a variable such as the @code{kill-ring} is created and |
| 9172 | given an initial value by using the @code{defvar} special form. The |
| 9173 | name comes from ``define variable''. |
| 9174 | |
| 9175 | The @code{defvar} special form is similar to @code{setq} in that it sets |
| 9176 | the value of a variable. It is unlike @code{setq} in two ways: first, |
| 9177 | it only sets the value of the variable if the variable does not already |
| 9178 | have a value. If the variable already has a value, @code{defvar} does |
| 9179 | not override the existing value. Second, @code{defvar} has a |
| 9180 | documentation string. |
| 9181 | |
| 9182 | (There is a related macro, @code{defcustom}, designed for variables |
| 9183 | that people customize. It has more features than @code{defvar}. |
| 9184 | (@xref{defcustom, , Setting Variables with @code{defcustom}}.) |
| 9185 | |
| 9186 | @menu |
| 9187 | * See variable current value:: |
| 9188 | * defvar and asterisk:: |
| 9189 | @end menu |
| 9190 | |
| 9191 | @ifnottex |
| 9192 | @node See variable current value |
| 9193 | @unnumberedsubsec Seeing the Current Value of a Variable |
| 9194 | @end ifnottex |
| 9195 | |
| 9196 | You can see the current value of a variable, any variable, by using |
| 9197 | the @code{describe-variable} function, which is usually invoked by |
| 9198 | typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring} |
| 9199 | (followed by @key{RET}) when prompted, you will see what is in your |
| 9200 | current kill ring---this may be quite a lot! Conversely, if you have |
| 9201 | been doing nothing this Emacs session except read this document, you |
| 9202 | may have nothing in it. Also, you will see the documentation for |
| 9203 | @code{kill-ring}: |
| 9204 | |
| 9205 | @smallexample |
| 9206 | @group |
| 9207 | Documentation: |
| 9208 | List of killed text sequences. |
| 9209 | Since the kill ring is supposed to interact nicely with cut-and-paste |
| 9210 | facilities offered by window systems, use of this variable should |
| 9211 | @end group |
| 9212 | @group |
| 9213 | interact nicely with `interprogram-cut-function' and |
| 9214 | `interprogram-paste-function'. The functions `kill-new', |
| 9215 | `kill-append', and `current-kill' are supposed to implement this |
| 9216 | interaction; you may want to use them instead of manipulating the kill |
| 9217 | ring directly. |
| 9218 | @end group |
| 9219 | @end smallexample |
| 9220 | |
| 9221 | @need 800 |
| 9222 | The kill ring is defined by a @code{defvar} in the following way: |
| 9223 | |
| 9224 | @smallexample |
| 9225 | @group |
| 9226 | (defvar kill-ring nil |
| 9227 | "List of killed text sequences. |
| 9228 | @dots{}") |
| 9229 | @end group |
| 9230 | @end smallexample |
| 9231 | |
| 9232 | @noindent |
| 9233 | In this variable definition, the variable is given an initial value of |
| 9234 | @code{nil}, which makes sense, since if you have saved nothing, you want |
| 9235 | nothing back if you give a @code{yank} command. The documentation |
| 9236 | string is written just like the documentation string of a @code{defun}. |
| 9237 | As with the documentation string of the @code{defun}, the first line of |
| 9238 | the documentation should be a complete sentence, since some commands, |
| 9239 | like @code{apropos}, print only the first line of documentation. |
| 9240 | Succeeding lines should not be indented; otherwise they look odd when |
| 9241 | you use @kbd{C-h v} (@code{describe-variable}). |
| 9242 | |
| 9243 | @node defvar and asterisk |
| 9244 | @subsection @code{defvar} and an asterisk |
| 9245 | @findex defvar @r{for a user customizable variable} |
| 9246 | @findex defvar @r{with an asterisk} |
| 9247 | |
| 9248 | In the past, Emacs used the @code{defvar} special form both for |
| 9249 | internal variables that you would not expect a user to change and for |
| 9250 | variables that you do expect a user to change. Although you can still |
| 9251 | use @code{defvar} for user customizable variables, please use |
| 9252 | @code{defcustom} instead, since it provides a path into |
| 9253 | the Customization commands. (@xref{defcustom, , Specifying Variables |
| 9254 | using @code{defcustom}}.) |
| 9255 | |
| 9256 | When you specified a variable using the @code{defvar} special form, |
| 9257 | you could distinguish a variable that a user might want to change from |
| 9258 | others by typing an asterisk, @samp{*}, in the first column of its |
| 9259 | documentation string. For example: |
| 9260 | |
| 9261 | @smallexample |
| 9262 | @group |
| 9263 | (defvar shell-command-default-error-buffer nil |
| 9264 | "*Buffer name for `shell-command' @dots{} error output. |
| 9265 | @dots{} ") |
| 9266 | @end group |
| 9267 | @end smallexample |
| 9268 | |
| 9269 | @findex set-variable |
| 9270 | @noindent |
| 9271 | You could (and still can) use the @code{set-variable} command to |
| 9272 | change the value of @code{shell-command-default-error-buffer} |
| 9273 | temporarily. However, options set using @code{set-variable} are set |
| 9274 | only for the duration of your editing session. The new values are not |
| 9275 | saved between sessions. Each time Emacs starts, it reads the original |
| 9276 | value, unless you change the value within your @file{.emacs} file, |
| 9277 | either by setting it manually or by using @code{customize}. |
| 9278 | @xref{Emacs Initialization, , Your @file{.emacs} File}. |
| 9279 | |
| 9280 | For me, the major use of the @code{set-variable} command is to suggest |
| 9281 | variables that I might want to set in my @file{.emacs} file. There |
| 9282 | are now more than 700 such variables, far too many to remember |
| 9283 | readily. Fortunately, you can press @key{TAB} after calling the |
| 9284 | @code{M-x set-variable} command to see the list of variables. |
| 9285 | (@xref{Examining, , Examining and Setting Variables, emacs, |
| 9286 | The GNU Emacs Manual}.) |
| 9287 | |
| 9288 | @need 1250 |
| 9289 | @node cons & search-fwd Review |
| 9290 | @section Review |
| 9291 | |
| 9292 | Here is a brief summary of some recently introduced functions. |
| 9293 | |
| 9294 | @table @code |
| 9295 | @item car |
| 9296 | @itemx cdr |
| 9297 | @code{car} returns the first element of a list; @code{cdr} returns the |
| 9298 | second and subsequent elements of a list. |
| 9299 | |
| 9300 | @need 1250 |
| 9301 | For example: |
| 9302 | |
| 9303 | @smallexample |
| 9304 | @group |
| 9305 | (car '(1 2 3 4 5 6 7)) |
| 9306 | @result{} 1 |
| 9307 | (cdr '(1 2 3 4 5 6 7)) |
| 9308 | @result{} (2 3 4 5 6 7) |
| 9309 | @end group |
| 9310 | @end smallexample |
| 9311 | |
| 9312 | @item cons |
| 9313 | @code{cons} constructs a list by prepending its first argument to its |
| 9314 | second argument. |
| 9315 | |
| 9316 | @need 1250 |
| 9317 | For example: |
| 9318 | |
| 9319 | @smallexample |
| 9320 | @group |
| 9321 | (cons 1 '(2 3 4)) |
| 9322 | @result{} (1 2 3 4) |
| 9323 | @end group |
| 9324 | @end smallexample |
| 9325 | |
| 9326 | @item funcall |
| 9327 | @code{funcall} evaluates its first argument as a function. It passes |
| 9328 | its remaining arguments to its first argument. |
| 9329 | |
| 9330 | @item nthcdr |
| 9331 | Return the result of taking @sc{cdr} `n' times on a list. |
| 9332 | @iftex |
| 9333 | The |
| 9334 | @tex |
| 9335 | $n^{th}$ |
| 9336 | @end tex |
| 9337 | @code{cdr}. |
| 9338 | @end iftex |
| 9339 | The `rest of the rest', as it were. |
| 9340 | |
| 9341 | @need 1250 |
| 9342 | For example: |
| 9343 | |
| 9344 | @smallexample |
| 9345 | @group |
| 9346 | (nthcdr 3 '(1 2 3 4 5 6 7)) |
| 9347 | @result{} (4 5 6 7) |
| 9348 | @end group |
| 9349 | @end smallexample |
| 9350 | |
| 9351 | @item setcar |
| 9352 | @itemx setcdr |
| 9353 | @code{setcar} changes the first element of a list; @code{setcdr} |
| 9354 | changes the second and subsequent elements of a list. |
| 9355 | |
| 9356 | @need 1250 |
| 9357 | For example: |
| 9358 | |
| 9359 | @smallexample |
| 9360 | @group |
| 9361 | (setq triple '(1 2 3)) |
| 9362 | |
| 9363 | (setcar triple '37) |
| 9364 | |
| 9365 | triple |
| 9366 | @result{} (37 2 3) |
| 9367 | |
| 9368 | (setcdr triple '("foo" "bar")) |
| 9369 | |
| 9370 | triple |
| 9371 | @result{} (37 "foo" "bar") |
| 9372 | @end group |
| 9373 | @end smallexample |
| 9374 | |
| 9375 | @item progn |
| 9376 | Evaluate each argument in sequence and then return the value of the |
| 9377 | last. |
| 9378 | |
| 9379 | @need 1250 |
| 9380 | For example: |
| 9381 | |
| 9382 | @smallexample |
| 9383 | @group |
| 9384 | (progn 1 2 3 4) |
| 9385 | @result{} 4 |
| 9386 | @end group |
| 9387 | @end smallexample |
| 9388 | |
| 9389 | @item save-restriction |
| 9390 | Record whatever narrowing is in effect in the current buffer, if any, |
| 9391 | and restore that narrowing after evaluating the arguments. |
| 9392 | |
| 9393 | @item search-forward |
| 9394 | Search for a string, and if the string is found, move point. With a |
| 9395 | regular expression, use the similar @code{re-search-forward}. |
| 9396 | (@xref{Regexp Search, , Regular Expression Searches}, for an |
| 9397 | explanation of regular expression patterns and searches.) |
| 9398 | |
| 9399 | @need 1250 |
| 9400 | @noindent |
| 9401 | @code{search-forward} and @code{re-search-forward} take four |
| 9402 | arguments: |
| 9403 | |
| 9404 | @enumerate |
| 9405 | @item |
| 9406 | The string or regular expression to search for. |
| 9407 | |
| 9408 | @item |
| 9409 | Optionally, the limit of the search. |
| 9410 | |
| 9411 | @item |
| 9412 | Optionally, what to do if the search fails, return @code{nil} or an |
| 9413 | error message. |
| 9414 | |
| 9415 | @item |
| 9416 | Optionally, how many times to repeat the search; if negative, the |
| 9417 | search goes backwards. |
| 9418 | @end enumerate |
| 9419 | |
| 9420 | @item kill-region |
| 9421 | @itemx delete-and-extract-region |
| 9422 | @itemx copy-region-as-kill |
| 9423 | |
| 9424 | @code{kill-region} cuts the text between point and mark from the |
| 9425 | buffer and stores that text in the kill ring, so you can get it back |
| 9426 | by yanking. |
| 9427 | |
| 9428 | @code{copy-region-as-kill} copies the text between point and mark into |
| 9429 | the kill ring, from which you can get it by yanking. The function |
| 9430 | does not cut or remove the text from the buffer. |
| 9431 | @end table |
| 9432 | |
| 9433 | @code{delete-and-extract-region} removes the text between point and |
| 9434 | mark from the buffer and throws it away. You cannot get it back. |
| 9435 | (This is not an interactive command.) |
| 9436 | |
| 9437 | @need 1500 |
| 9438 | @node search Exercises |
| 9439 | @section Searching Exercises |
| 9440 | |
| 9441 | @itemize @bullet |
| 9442 | @item |
| 9443 | Write an interactive function that searches for a string. If the |
| 9444 | search finds the string, leave point after it and display a message |
| 9445 | that says ``Found!''. (Do not use @code{search-forward} for the name |
| 9446 | of this function; if you do, you will overwrite the existing version of |
| 9447 | @code{search-forward} that comes with Emacs. Use a name such as |
| 9448 | @code{test-search} instead.) |
| 9449 | |
| 9450 | @item |
| 9451 | Write a function that prints the third element of the kill ring in the |
| 9452 | echo area, if any; if the kill ring does not contain a third element, |
| 9453 | print an appropriate message. |
| 9454 | @end itemize |
| 9455 | |
| 9456 | @node List Implementation |
| 9457 | @chapter How Lists are Implemented |
| 9458 | @cindex Lists in a computer |
| 9459 | |
| 9460 | In Lisp, atoms are recorded in a straightforward fashion; if the |
| 9461 | implementation is not straightforward in practice, it is, nonetheless, |
| 9462 | straightforward in theory. The atom @samp{rose}, for example, is |
| 9463 | recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s}, |
| 9464 | @samp{e}. A list, on the other hand, is kept differently. The mechanism |
| 9465 | is equally simple, but it takes a moment to get used to the idea. A |
| 9466 | list is kept using a series of pairs of pointers. In the series, the |
| 9467 | first pointer in each pair points to an atom or to another list, and the |
| 9468 | second pointer in each pair points to the next pair, or to the symbol |
| 9469 | @code{nil}, which marks the end of the list. |
| 9470 | |
| 9471 | A pointer itself is quite simply the electronic address of what is |
| 9472 | pointed to. Hence, a list is kept as a series of electronic addresses. |
| 9473 | |
| 9474 | @menu |
| 9475 | * Lists diagrammed:: |
| 9476 | * Symbols as Chest:: Exploring a powerful metaphor. |
| 9477 | * List Exercise:: |
| 9478 | @end menu |
| 9479 | |
| 9480 | @ifnottex |
| 9481 | @node Lists diagrammed |
| 9482 | @unnumberedsec Lists diagrammed |
| 9483 | @end ifnottex |
| 9484 | |
| 9485 | For example, the list @code{(rose violet buttercup)} has three elements, |
| 9486 | @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the |
| 9487 | electronic address of @samp{rose} is recorded in a segment of computer |
| 9488 | memory along with the address that gives the electronic address of where |
| 9489 | the atom @samp{violet} is located; and that address (the one that tells |
| 9490 | where @samp{violet} is located) is kept along with an address that tells |
| 9491 | where the address for the atom @samp{buttercup} is located. |
| 9492 | |
| 9493 | @need 1200 |
| 9494 | This sounds more complicated than it is and is easier seen in a diagram: |
| 9495 | |
| 9496 | @c clear print-postscript-figures |
| 9497 | @c !!! cons-cell-diagram #1 |
| 9498 | @ifnottex |
| 9499 | @smallexample |
| 9500 | @group |
| 9501 | ___ ___ ___ ___ ___ ___ |
| 9502 | |___|___|--> |___|___|--> |___|___|--> nil |
| 9503 | | | | |
| 9504 | | | | |
| 9505 | --> rose --> violet --> buttercup |
| 9506 | @end group |
| 9507 | @end smallexample |
| 9508 | @end ifnottex |
| 9509 | @ifset print-postscript-figures |
| 9510 | @sp 1 |
| 9511 | @tex |
| 9512 | @center @image{cons-1} |
| 9513 | @end tex |
| 9514 | @sp 1 |
| 9515 | @end ifset |
| 9516 | @ifclear print-postscript-figures |
| 9517 | @iftex |
| 9518 | @smallexample |
| 9519 | @group |
| 9520 | ___ ___ ___ ___ ___ ___ |
| 9521 | |___|___|--> |___|___|--> |___|___|--> nil |
| 9522 | | | | |
| 9523 | | | | |
| 9524 | --> rose --> violet --> buttercup |
| 9525 | @end group |
| 9526 | @end smallexample |
| 9527 | @end iftex |
| 9528 | @end ifclear |
| 9529 | |
| 9530 | @noindent |
| 9531 | In the diagram, each box represents a word of computer memory that |
| 9532 | holds a Lisp object, usually in the form of a memory address. The boxes, |
| 9533 | i.e., the addresses, are in pairs. Each arrow points to what the address |
| 9534 | is the address of, either an atom or another pair of addresses. The |
| 9535 | first box is the electronic address of @samp{rose} and the arrow points |
| 9536 | to @samp{rose}; the second box is the address of the next pair of boxes, |
| 9537 | the first part of which is the address of @samp{violet} and the second |
| 9538 | part of which is the address of the next pair. The very last box |
| 9539 | points to the symbol @code{nil}, which marks the end of the list. |
| 9540 | |
| 9541 | @need 1200 |
| 9542 | When a variable is set to a list with a function such as @code{setq}, |
| 9543 | it stores the address of the first box in the variable. Thus, |
| 9544 | evaluation of the expression |
| 9545 | |
| 9546 | @smallexample |
| 9547 | (setq bouquet '(rose violet buttercup)) |
| 9548 | @end smallexample |
| 9549 | |
| 9550 | @need 1250 |
| 9551 | @noindent |
| 9552 | creates a situation like this: |
| 9553 | |
| 9554 | @c cons-cell-diagram #2 |
| 9555 | @ifnottex |
| 9556 | @smallexample |
| 9557 | @group |
| 9558 | bouquet |
| 9559 | | |
| 9560 | | ___ ___ ___ ___ ___ ___ |
| 9561 | --> |___|___|--> |___|___|--> |___|___|--> nil |
| 9562 | | | | |
| 9563 | | | | |
| 9564 | --> rose --> violet --> buttercup |
| 9565 | @end group |
| 9566 | @end smallexample |
| 9567 | @end ifnottex |
| 9568 | @ifset print-postscript-figures |
| 9569 | @sp 1 |
| 9570 | @tex |
| 9571 | @center @image{cons-2} |
| 9572 | @end tex |
| 9573 | @sp 1 |
| 9574 | @end ifset |
| 9575 | @ifclear print-postscript-figures |
| 9576 | @iftex |
| 9577 | @smallexample |
| 9578 | @group |
| 9579 | bouquet |
| 9580 | | |
| 9581 | | ___ ___ ___ ___ ___ ___ |
| 9582 | --> |___|___|--> |___|___|--> |___|___|--> nil |
| 9583 | | | | |
| 9584 | | | | |
| 9585 | --> rose --> violet --> buttercup |
| 9586 | @end group |
| 9587 | @end smallexample |
| 9588 | @end iftex |
| 9589 | @end ifclear |
| 9590 | |
| 9591 | @noindent |
| 9592 | In this example, the symbol @code{bouquet} holds the address of the first |
| 9593 | pair of boxes. |
| 9594 | |
| 9595 | @need 1200 |
| 9596 | This same list can be illustrated in a different sort of box notation |
| 9597 | like this: |
| 9598 | |
| 9599 | @c cons-cell-diagram #2a |
| 9600 | @ifnottex |
| 9601 | @smallexample |
| 9602 | @group |
| 9603 | bouquet |
| 9604 | | |
| 9605 | | -------------- --------------- ---------------- |
| 9606 | | | car | cdr | | car | cdr | | car | cdr | |
| 9607 | -->| rose | o------->| violet | o------->| butter- | nil | |
| 9608 | | | | | | | | cup | | |
| 9609 | -------------- --------------- ---------------- |
| 9610 | @end group |
| 9611 | @end smallexample |
| 9612 | @end ifnottex |
| 9613 | @ifset print-postscript-figures |
| 9614 | @sp 1 |
| 9615 | @tex |
| 9616 | @center @image{cons-2a} |
| 9617 | @end tex |
| 9618 | @sp 1 |
| 9619 | @end ifset |
| 9620 | @ifclear print-postscript-figures |
| 9621 | @iftex |
| 9622 | @smallexample |
| 9623 | @group |
| 9624 | bouquet |
| 9625 | | |
| 9626 | | -------------- --------------- ---------------- |
| 9627 | | | car | cdr | | car | cdr | | car | cdr | |
| 9628 | -->| rose | o------->| violet | o------->| butter- | nil | |
| 9629 | | | | | | | | cup | | |
| 9630 | -------------- --------------- ---------------- |
| 9631 | @end group |
| 9632 | @end smallexample |
| 9633 | @end iftex |
| 9634 | @end ifclear |
| 9635 | |
| 9636 | (Symbols consist of more than pairs of addresses, but the structure of |
| 9637 | a symbol is made up of addresses. Indeed, the symbol @code{bouquet} |
| 9638 | consists of a group of address-boxes, one of which is the address of |
| 9639 | the printed word @samp{bouquet}, a second of which is the address of a |
| 9640 | function definition attached to the symbol, if any, a third of which |
| 9641 | is the address of the first pair of address-boxes for the list |
| 9642 | @code{(rose violet buttercup)}, and so on. Here we are showing that |
| 9643 | the symbol's third address-box points to the first pair of |
| 9644 | address-boxes for the list.) |
| 9645 | |
| 9646 | If a symbol is set to the @sc{cdr} of a list, the list itself is not |
| 9647 | changed; the symbol simply has an address further down the list. (In |
| 9648 | the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus, |
| 9649 | evaluation of the following expression |
| 9650 | |
| 9651 | @smallexample |
| 9652 | (setq flowers (cdr bouquet)) |
| 9653 | @end smallexample |
| 9654 | |
| 9655 | @need 800 |
| 9656 | @noindent |
| 9657 | produces this: |
| 9658 | |
| 9659 | @c cons-cell-diagram #3 |
| 9660 | @ifnottex |
| 9661 | @sp 1 |
| 9662 | @smallexample |
| 9663 | @group |
| 9664 | bouquet flowers |
| 9665 | | | |
| 9666 | | ___ ___ | ___ ___ ___ ___ |
| 9667 | --> | | | --> | | | | | | |
| 9668 | |___|___|----> |___|___|--> |___|___|--> nil |
| 9669 | | | | |
| 9670 | | | | |
| 9671 | --> rose --> violet --> buttercup |
| 9672 | @end group |
| 9673 | @end smallexample |
| 9674 | @sp 1 |
| 9675 | @end ifnottex |
| 9676 | @ifset print-postscript-figures |
| 9677 | @sp 1 |
| 9678 | @tex |
| 9679 | @center @image{cons-3} |
| 9680 | @end tex |
| 9681 | @sp 1 |
| 9682 | @end ifset |
| 9683 | @ifclear print-postscript-figures |
| 9684 | @iftex |
| 9685 | @sp 1 |
| 9686 | @smallexample |
| 9687 | @group |
| 9688 | bouquet flowers |
| 9689 | | | |
| 9690 | | ___ ___ | ___ ___ ___ ___ |
| 9691 | --> | | | --> | | | | | | |
| 9692 | |___|___|----> |___|___|--> |___|___|--> nil |
| 9693 | | | | |
| 9694 | | | | |
| 9695 | --> rose --> violet --> buttercup |
| 9696 | @end group |
| 9697 | @end smallexample |
| 9698 | @sp 1 |
| 9699 | @end iftex |
| 9700 | @end ifclear |
| 9701 | |
| 9702 | @noindent |
| 9703 | The value of @code{flowers} is @code{(violet buttercup)}, which is |
| 9704 | to say, the symbol @code{flowers} holds the address of the pair of |
| 9705 | address-boxes, the first of which holds the address of @code{violet}, |
| 9706 | and the second of which holds the address of @code{buttercup}. |
| 9707 | |
| 9708 | A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted |
| 9709 | pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp |
| 9710 | Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair |
| 9711 | Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more |
| 9712 | information about cons cells and dotted pairs. |
| 9713 | |
| 9714 | @need 1200 |
| 9715 | The function @code{cons} adds a new pair of addresses to the front of |
| 9716 | a series of addresses like that shown above. For example, evaluating |
| 9717 | the expression |
| 9718 | |
| 9719 | @smallexample |
| 9720 | (setq bouquet (cons 'lily bouquet)) |
| 9721 | @end smallexample |
| 9722 | |
| 9723 | @need 1500 |
| 9724 | @noindent |
| 9725 | produces: |
| 9726 | |
| 9727 | @c cons-cell-diagram #4 |
| 9728 | @ifnottex |
| 9729 | @sp 1 |
| 9730 | @smallexample |
| 9731 | @group |
| 9732 | bouquet flowers |
| 9733 | | | |
| 9734 | | ___ ___ ___ ___ | ___ ___ ___ ___ |
| 9735 | --> | | | | | | --> | | | | | | |
| 9736 | |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil |
| 9737 | | | | | |
| 9738 | | | | | |
| 9739 | --> lily --> rose --> violet --> buttercup |
| 9740 | @end group |
| 9741 | @end smallexample |
| 9742 | @sp 1 |
| 9743 | @end ifnottex |
| 9744 | @ifset print-postscript-figures |
| 9745 | @sp 1 |
| 9746 | @tex |
| 9747 | @center @image{cons-4} |
| 9748 | @end tex |
| 9749 | @sp 1 |
| 9750 | @end ifset |
| 9751 | @ifclear print-postscript-figures |
| 9752 | @iftex |
| 9753 | @sp 1 |
| 9754 | @smallexample |
| 9755 | @group |
| 9756 | bouquet flowers |
| 9757 | | | |
| 9758 | | ___ ___ ___ ___ | ___ ___ ___ ___ |
| 9759 | --> | | | | | | --> | | | | | | |
| 9760 | |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil |
| 9761 | | | | | |
| 9762 | | | | | |
| 9763 | --> lily --> rose --> violet --> buttercup |
| 9764 | @end group |
| 9765 | @end smallexample |
| 9766 | @sp 1 |
| 9767 | @end iftex |
| 9768 | @end ifclear |
| 9769 | |
| 9770 | @need 1200 |
| 9771 | @noindent |
| 9772 | However, this does not change the value of the symbol |
| 9773 | @code{flowers}, as you can see by evaluating the following, |
| 9774 | |
| 9775 | @smallexample |
| 9776 | (eq (cdr (cdr bouquet)) flowers) |
| 9777 | @end smallexample |
| 9778 | |
| 9779 | @noindent |
| 9780 | which returns @code{t} for true. |
| 9781 | |
| 9782 | Until it is reset, @code{flowers} still has the value |
| 9783 | @code{(violet buttercup)}; that is, it has the address of the cons |
| 9784 | cell whose first address is of @code{violet}. Also, this does not |
| 9785 | alter any of the pre-existing cons cells; they are all still there. |
| 9786 | |
| 9787 | Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address |
| 9788 | of the next cons cell in the series; to get the @sc{car} of a list, |
| 9789 | you get the address of the first element of the list; to @code{cons} a |
| 9790 | new element on a list, you add a new cons cell to the front of the list. |
| 9791 | That is all there is to it! The underlying structure of Lisp is |
| 9792 | brilliantly simple! |
| 9793 | |
| 9794 | And what does the last address in a series of cons cells refer to? It |
| 9795 | is the address of the empty list, of @code{nil}. |
| 9796 | |
| 9797 | In summary, when a Lisp variable is set to a value, it is provided with |
| 9798 | the address of the list to which the variable refers. |
| 9799 | |
| 9800 | @node Symbols as Chest |
| 9801 | @section Symbols as a Chest of Drawers |
| 9802 | @cindex Symbols as a Chest of Drawers |
| 9803 | @cindex Chest of Drawers, metaphor for a symbol |
| 9804 | @cindex Drawers, Chest of, metaphor for a symbol |
| 9805 | |
| 9806 | In an earlier section, I suggested that you might imagine a symbol as |
| 9807 | being a chest of drawers. The function definition is put in one |
| 9808 | drawer, the value in another, and so on. What is put in the drawer |
| 9809 | holding the value can be changed without affecting the contents of the |
| 9810 | drawer holding the function definition, and vice-verse. |
| 9811 | |
| 9812 | Actually, what is put in each drawer is the address of the value or |
| 9813 | function definition. It is as if you found an old chest in the attic, |
| 9814 | and in one of its drawers you found a map giving you directions to |
| 9815 | where the buried treasure lies. |
| 9816 | |
| 9817 | (In addition to its name, symbol definition, and variable value, a |
| 9818 | symbol has a `drawer' for a @dfn{property list} which can be used to |
| 9819 | record other information. Property lists are not discussed here; see |
| 9820 | @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp |
| 9821 | Reference Manual}.) |
| 9822 | |
| 9823 | @need 1500 |
| 9824 | Here is a fanciful representation: |
| 9825 | |
| 9826 | @c chest-of-drawers diagram |
| 9827 | @ifnottex |
| 9828 | @sp 1 |
| 9829 | @smallexample |
| 9830 | @group |
| 9831 | Chest of Drawers Contents of Drawers |
| 9832 | |
| 9833 | __ o0O0o __ |
| 9834 | / \ |
| 9835 | --------------------- |
| 9836 | | directions to | [map to] |
| 9837 | | symbol name | bouquet |
| 9838 | | | |
| 9839 | +---------------------+ |
| 9840 | | directions to | |
| 9841 | | symbol definition | [none] |
| 9842 | | | |
| 9843 | +---------------------+ |
| 9844 | | directions to | [map to] |
| 9845 | | variable value | (rose violet buttercup) |
| 9846 | | | |
| 9847 | +---------------------+ |
| 9848 | | directions to | |
| 9849 | | property list | [not described here] |
| 9850 | | | |
| 9851 | +---------------------+ |
| 9852 | |/ \| |
| 9853 | @end group |
| 9854 | @end smallexample |
| 9855 | @sp 1 |
| 9856 | @end ifnottex |
| 9857 | @ifset print-postscript-figures |
| 9858 | @sp 1 |
| 9859 | @tex |
| 9860 | @center @image{drawers} |
| 9861 | @end tex |
| 9862 | @sp 1 |
| 9863 | @end ifset |
| 9864 | @ifclear print-postscript-figures |
| 9865 | @iftex |
| 9866 | @sp 1 |
| 9867 | @smallexample |
| 9868 | @group |
| 9869 | Chest of Drawers Contents of Drawers |
| 9870 | |
| 9871 | __ o0O0o __ |
| 9872 | / \ |
| 9873 | --------------------- |
| 9874 | | directions to | [map to] |
| 9875 | | symbol name | bouquet |
| 9876 | | | |
| 9877 | +---------------------+ |
| 9878 | | directions to | |
| 9879 | | symbol definition | [none] |
| 9880 | | | |
| 9881 | +---------------------+ |
| 9882 | | directions to | [map to] |
| 9883 | | variable value | (rose violet buttercup) |
| 9884 | | | |
| 9885 | +---------------------+ |
| 9886 | | directions to | |
| 9887 | | property list | [not described here] |
| 9888 | | | |
| 9889 | +---------------------+ |
| 9890 | |/ \| |
| 9891 | @end group |
| 9892 | @end smallexample |
| 9893 | @sp 1 |
| 9894 | @end iftex |
| 9895 | @end ifclear |
| 9896 | |
| 9897 | @node List Exercise |
| 9898 | @section Exercise |
| 9899 | |
| 9900 | Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two |
| 9901 | more flowers on to this list and set this new list to |
| 9902 | @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish. |
| 9903 | What does the @code{more-flowers} list now contain? |
| 9904 | |
| 9905 | @node Yanking |
| 9906 | @chapter Yanking Text Back |
| 9907 | @findex yank |
| 9908 | @cindex Text retrieval |
| 9909 | @cindex Retrieving text |
| 9910 | @cindex Pasting text |
| 9911 | |
| 9912 | Whenever you cut text out of a buffer with a `kill' command in GNU Emacs, |
| 9913 | you can bring it back with a `yank' command. The text that is cut out of |
| 9914 | the buffer is put in the kill ring and the yank commands insert the |
| 9915 | appropriate contents of the kill ring back into a buffer (not necessarily |
| 9916 | the original buffer). |
| 9917 | |
| 9918 | A simple @kbd{C-y} (@code{yank}) command inserts the first item from |
| 9919 | the kill ring into the current buffer. If the @kbd{C-y} command is |
| 9920 | followed immediately by @kbd{M-y}, the first element is replaced by |
| 9921 | the second element. Successive @kbd{M-y} commands replace the second |
| 9922 | element with the third, fourth, or fifth element, and so on. When the |
| 9923 | last element in the kill ring is reached, it is replaced by the first |
| 9924 | element and the cycle is repeated. (Thus the kill ring is called a |
| 9925 | `ring' rather than just a `list'. However, the actual data structure |
| 9926 | that holds the text is a list. |
| 9927 | @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the |
| 9928 | list is handled as a ring.) |
| 9929 | |
| 9930 | @menu |
| 9931 | * Kill Ring Overview:: |
| 9932 | * kill-ring-yank-pointer:: The kill ring is a list. |
| 9933 | * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable. |
| 9934 | @end menu |
| 9935 | |
| 9936 | @node Kill Ring Overview |
| 9937 | @section Kill Ring Overview |
| 9938 | @cindex Kill ring overview |
| 9939 | |
| 9940 | The kill ring is a list of textual strings. This is what it looks like: |
| 9941 | |
| 9942 | @smallexample |
| 9943 | ("some text" "a different piece of text" "yet more text") |
| 9944 | @end smallexample |
| 9945 | |
| 9946 | If this were the contents of my kill ring and I pressed @kbd{C-y}, the |
| 9947 | string of characters saying @samp{some text} would be inserted in this |
| 9948 | buffer where my cursor is located. |
| 9949 | |
| 9950 | The @code{yank} command is also used for duplicating text by copying it. |
| 9951 | The copied text is not cut from the buffer, but a copy of it is put on the |
| 9952 | kill ring and is inserted by yanking it back. |
| 9953 | |
| 9954 | Three functions are used for bringing text back from the kill ring: |
| 9955 | @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop}, |
| 9956 | which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer}, |
| 9957 | which is used by the two other functions. |
| 9958 | |
| 9959 | These functions refer to the kill ring through a variable called the |
| 9960 | @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the |
| 9961 | @code{yank} and @code{yank-pop} functions is: |
| 9962 | |
| 9963 | @smallexample |
| 9964 | (insert (car kill-ring-yank-pointer)) |
| 9965 | @end smallexample |
| 9966 | |
| 9967 | @noindent |
| 9968 | (Well, no more. In GNU Emacs 22, the function has been replaced by |
| 9969 | @code{insert-for-yank} which calls @code{insert-for-yank-1} |
| 9970 | repetitively for each @code{yank-handler} segment. In turn, |
| 9971 | @code{insert-for-yank-1} strips text properties from the inserted text |
| 9972 | according to @code{yank-excluded-properties}. Otherwise, it is just |
| 9973 | like @code{insert}. We will stick with plain @code{insert} since it |
| 9974 | is easier to understand.) |
| 9975 | |
| 9976 | To begin to understand how @code{yank} and @code{yank-pop} work, it is |
| 9977 | first necessary to look at the @code{kill-ring-yank-pointer} variable. |
| 9978 | |
| 9979 | @node kill-ring-yank-pointer |
| 9980 | @section The @code{kill-ring-yank-pointer} Variable |
| 9981 | |
| 9982 | @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is |
| 9983 | a variable. It points to something by being bound to the value of what |
| 9984 | it points to, like any other Lisp variable. |
| 9985 | |
| 9986 | @need 1000 |
| 9987 | Thus, if the value of the kill ring is: |
| 9988 | |
| 9989 | @smallexample |
| 9990 | ("some text" "a different piece of text" "yet more text") |
| 9991 | @end smallexample |
| 9992 | |
| 9993 | @need 1250 |
| 9994 | @noindent |
| 9995 | and the @code{kill-ring-yank-pointer} points to the second clause, the |
| 9996 | value of @code{kill-ring-yank-pointer} is: |
| 9997 | |
| 9998 | @smallexample |
| 9999 | ("a different piece of text" "yet more text") |
| 10000 | @end smallexample |
| 10001 | |
| 10002 | As explained in the previous chapter (@pxref{List Implementation}), the |
| 10003 | computer does not keep two different copies of the text being pointed to |
| 10004 | by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The |
| 10005 | words ``a different piece of text'' and ``yet more text'' are not |
| 10006 | duplicated. Instead, the two Lisp variables point to the same pieces of |
| 10007 | text. Here is a diagram: |
| 10008 | |
| 10009 | @c cons-cell-diagram #5 |
| 10010 | @ifnottex |
| 10011 | @smallexample |
| 10012 | @group |
| 10013 | kill-ring kill-ring-yank-pointer |
| 10014 | | | |
| 10015 | | ___ ___ | ___ ___ ___ ___ |
| 10016 | ---> | | | --> | | | | | | |
| 10017 | |___|___|----> |___|___|--> |___|___|--> nil |
| 10018 | | | | |
| 10019 | | | | |
| 10020 | | | --> "yet more text" |
| 10021 | | | |
| 10022 | | --> "a different piece of text" |
| 10023 | | |
| 10024 | --> "some text" |
| 10025 | @end group |
| 10026 | @end smallexample |
| 10027 | @sp 1 |
| 10028 | @end ifnottex |
| 10029 | @ifset print-postscript-figures |
| 10030 | @sp 1 |
| 10031 | @tex |
| 10032 | @center @image{cons-5} |
| 10033 | @end tex |
| 10034 | @sp 1 |
| 10035 | @end ifset |
| 10036 | @ifclear print-postscript-figures |
| 10037 | @iftex |
| 10038 | @smallexample |
| 10039 | @group |
| 10040 | kill-ring kill-ring-yank-pointer |
| 10041 | | | |
| 10042 | | ___ ___ | ___ ___ ___ ___ |
| 10043 | ---> | | | --> | | | | | | |
| 10044 | |___|___|----> |___|___|--> |___|___|--> nil |
| 10045 | | | | |
| 10046 | | | | |
| 10047 | | | --> "yet more text" |
| 10048 | | | |
| 10049 | | --> "a different piece of text |
| 10050 | | |
| 10051 | --> "some text" |
| 10052 | @end group |
| 10053 | @end smallexample |
| 10054 | @sp 1 |
| 10055 | @end iftex |
| 10056 | @end ifclear |
| 10057 | |
| 10058 | Both the variable @code{kill-ring} and the variable |
| 10059 | @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is |
| 10060 | usually described as if it were actually what it is composed of. The |
| 10061 | @code{kill-ring} is spoken of as if it were the list rather than that it |
| 10062 | points to the list. Conversely, the @code{kill-ring-yank-pointer} is |
| 10063 | spoken of as pointing to a list. |
| 10064 | |
| 10065 | These two ways of talking about the same thing sound confusing at first but |
| 10066 | make sense on reflection. The kill ring is generally thought of as the |
| 10067 | complete structure of data that holds the information of what has recently |
| 10068 | been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer} |
| 10069 | on the other hand, serves to indicate---that is, to `point to'---that part |
| 10070 | of the kill ring of which the first element (the @sc{car}) will be |
| 10071 | inserted. |
| 10072 | |
| 10073 | @ignore |
| 10074 | In GNU Emacs 22, the @code{kill-new} function calls |
| 10075 | |
| 10076 | @code{(setq kill-ring-yank-pointer kill-ring)} |
| 10077 | |
| 10078 | (defun rotate-yank-pointer (arg) |
| 10079 | "Rotate the yanking point in the kill ring. |
| 10080 | With argument, rotate that many kills forward (or backward, if negative)." |
| 10081 | (interactive "p") |
| 10082 | (current-kill arg)) |
| 10083 | |
| 10084 | (defun current-kill (n &optional do-not-move) |
| 10085 | "Rotate the yanking point by N places, and then return that kill. |
| 10086 | If N is zero, `interprogram-paste-function' is set, and calling it |
| 10087 | returns a string, then that string is added to the front of the |
| 10088 | kill ring and returned as the latest kill. |
| 10089 | If optional arg DO-NOT-MOVE is non-nil, then don't actually move the |
| 10090 | yanking point; just return the Nth kill forward." |
| 10091 | (let ((interprogram-paste (and (= n 0) |
| 10092 | interprogram-paste-function |
| 10093 | (funcall interprogram-paste-function)))) |
| 10094 | (if interprogram-paste |
| 10095 | (progn |
| 10096 | ;; Disable the interprogram cut function when we add the new |
| 10097 | ;; text to the kill ring, so Emacs doesn't try to own the |
| 10098 | ;; selection, with identical text. |
| 10099 | (let ((interprogram-cut-function nil)) |
| 10100 | (kill-new interprogram-paste)) |
| 10101 | interprogram-paste) |
| 10102 | (or kill-ring (error "Kill ring is empty")) |
| 10103 | (let ((ARGth-kill-element |
| 10104 | (nthcdr (mod (- n (length kill-ring-yank-pointer)) |
| 10105 | (length kill-ring)) |
| 10106 | kill-ring))) |
| 10107 | (or do-not-move |
| 10108 | (setq kill-ring-yank-pointer ARGth-kill-element)) |
| 10109 | (car ARGth-kill-element))))) |
| 10110 | |
| 10111 | @end ignore |
| 10112 | |
| 10113 | @need 1500 |
| 10114 | @node yank nthcdr Exercises |
| 10115 | @section Exercises with @code{yank} and @code{nthcdr} |
| 10116 | |
| 10117 | @itemize @bullet |
| 10118 | @item |
| 10119 | Using @kbd{C-h v} (@code{describe-variable}), look at the value of |
| 10120 | your kill ring. Add several items to your kill ring; look at its |
| 10121 | value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way |
| 10122 | around the kill ring. How many items were in your kill ring? Find |
| 10123 | the value of @code{kill-ring-max}. Was your kill ring full, or could |
| 10124 | you have kept more blocks of text within it? |
| 10125 | |
| 10126 | @item |
| 10127 | Using @code{nthcdr} and @code{car}, construct a series of expressions |
| 10128 | to return the first, second, third, and fourth elements of a list. |
| 10129 | @end itemize |
| 10130 | |
| 10131 | @node Loops & Recursion |
| 10132 | @chapter Loops and Recursion |
| 10133 | @cindex Loops and recursion |
| 10134 | @cindex Recursion and loops |
| 10135 | @cindex Repetition (loops) |
| 10136 | |
| 10137 | Emacs Lisp has two primary ways to cause an expression, or a series of |
| 10138 | expressions, to be evaluated repeatedly: one uses a @code{while} |
| 10139 | loop, and the other uses @dfn{recursion}. |
| 10140 | |
| 10141 | Repetition can be very valuable. For example, to move forward four |
| 10142 | sentences, you need only write a program that will move forward one |
| 10143 | sentence and then repeat the process four times. Since a computer does |
| 10144 | not get bored or tired, such repetitive action does not have the |
| 10145 | deleterious effects that excessive or the wrong kinds of repetition can |
| 10146 | have on humans. |
| 10147 | |
| 10148 | People mostly write Emacs Lisp functions using @code{while} loops and |
| 10149 | their kin; but you can use recursion, which provides a very powerful |
| 10150 | way to think about and then to solve problems@footnote{You can write |
| 10151 | recursive functions to be frugal or wasteful of mental or computer |
| 10152 | resources; as it happens, methods that people find easy---that are |
| 10153 | frugal of `mental resources'---sometimes use considerable computer |
| 10154 | resources. Emacs was designed to run on machines that we now consider |
| 10155 | limited and its default settings are conservative. You may want to |
| 10156 | increase the values of @code{max-specpdl-size} and |
| 10157 | @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to |
| 10158 | 15 and 30 times their default value.}. |
| 10159 | |
| 10160 | @menu |
| 10161 | * while:: Causing a stretch of code to repeat. |
| 10162 | * dolist dotimes:: |
| 10163 | * Recursion:: Causing a function to call itself. |
| 10164 | * Looping exercise:: |
| 10165 | @end menu |
| 10166 | |
| 10167 | @node while |
| 10168 | @section @code{while} |
| 10169 | @cindex Loops |
| 10170 | @findex while |
| 10171 | |
| 10172 | The @code{while} special form tests whether the value returned by |
| 10173 | evaluating its first argument is true or false. This is similar to what |
| 10174 | the Lisp interpreter does with an @code{if}; what the interpreter does |
| 10175 | next, however, is different. |
| 10176 | |
| 10177 | In a @code{while} expression, if the value returned by evaluating the |
| 10178 | first argument is false, the Lisp interpreter skips the rest of the |
| 10179 | expression (the @dfn{body} of the expression) and does not evaluate it. |
| 10180 | However, if the value is true, the Lisp interpreter evaluates the body |
| 10181 | of the expression and then again tests whether the first argument to |
| 10182 | @code{while} is true or false. If the value returned by evaluating the |
| 10183 | first argument is again true, the Lisp interpreter again evaluates the |
| 10184 | body of the expression. |
| 10185 | |
| 10186 | @need 1200 |
| 10187 | The template for a @code{while} expression looks like this: |
| 10188 | |
| 10189 | @smallexample |
| 10190 | @group |
| 10191 | (while @var{true-or-false-test} |
| 10192 | @var{body}@dots{}) |
| 10193 | @end group |
| 10194 | @end smallexample |
| 10195 | |
| 10196 | @menu |
| 10197 | * Looping with while:: Repeat so long as test returns true. |
| 10198 | * Loop Example:: A @code{while} loop that uses a list. |
| 10199 | * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}. |
| 10200 | * Incrementing Loop:: A loop with an incrementing counter. |
| 10201 | * Incrementing Loop Details:: |
| 10202 | * Decrementing Loop:: A loop with a decrementing counter. |
| 10203 | @end menu |
| 10204 | |
| 10205 | @ifnottex |
| 10206 | @node Looping with while |
| 10207 | @unnumberedsubsec Looping with @code{while} |
| 10208 | @end ifnottex |
| 10209 | |
| 10210 | So long as the true-or-false-test of the @code{while} expression |
| 10211 | returns a true value when it is evaluated, the body is repeatedly |
| 10212 | evaluated. This process is called a loop since the Lisp interpreter |
| 10213 | repeats the same thing again and again, like an airplane doing a loop. |
| 10214 | When the result of evaluating the true-or-false-test is false, the |
| 10215 | Lisp interpreter does not evaluate the rest of the @code{while} |
| 10216 | expression and `exits the loop'. |
| 10217 | |
| 10218 | Clearly, if the value returned by evaluating the first argument to |
| 10219 | @code{while} is always true, the body following will be evaluated |
| 10220 | again and again @dots{} and again @dots{} forever. Conversely, if the |
| 10221 | value returned is never true, the expressions in the body will never |
| 10222 | be evaluated. The craft of writing a @code{while} loop consists of |
| 10223 | choosing a mechanism such that the true-or-false-test returns true |
| 10224 | just the number of times that you want the subsequent expressions to |
| 10225 | be evaluated, and then have the test return false. |
| 10226 | |
| 10227 | The value returned by evaluating a @code{while} is the value of the |
| 10228 | true-or-false-test. An interesting consequence of this is that a |
| 10229 | @code{while} loop that evaluates without error will return @code{nil} |
| 10230 | or false regardless of whether it has looped 1 or 100 times or none at |
| 10231 | all. A @code{while} expression that evaluates successfully never |
| 10232 | returns a true value! What this means is that @code{while} is always |
| 10233 | evaluated for its side effects, which is to say, the consequences of |
| 10234 | evaluating the expressions within the body of the @code{while} loop. |
| 10235 | This makes sense. It is not the mere act of looping that is desired, |
| 10236 | but the consequences of what happens when the expressions in the loop |
| 10237 | are repeatedly evaluated. |
| 10238 | |
| 10239 | @node Loop Example |
| 10240 | @subsection A @code{while} Loop and a List |
| 10241 | |
| 10242 | A common way to control a @code{while} loop is to test whether a list |
| 10243 | has any elements. If it does, the loop is repeated; but if it does not, |
| 10244 | the repetition is ended. Since this is an important technique, we will |
| 10245 | create a short example to illustrate it. |
| 10246 | |
| 10247 | A simple way to test whether a list has elements is to evaluate the |
| 10248 | list: if it has no elements, it is an empty list and will return the |
| 10249 | empty list, @code{()}, which is a synonym for @code{nil} or false. On |
| 10250 | the other hand, a list with elements will return those elements when it |
| 10251 | is evaluated. Since Emacs Lisp considers as true any value that is not |
| 10252 | @code{nil}, a list that returns elements will test true in a |
| 10253 | @code{while} loop. |
| 10254 | |
| 10255 | @need 1200 |
| 10256 | For example, you can set the variable @code{empty-list} to @code{nil} by |
| 10257 | evaluating the following @code{setq} expression: |
| 10258 | |
| 10259 | @smallexample |
| 10260 | (setq empty-list ()) |
| 10261 | @end smallexample |
| 10262 | |
| 10263 | @noindent |
| 10264 | After evaluating the @code{setq} expression, you can evaluate the |
| 10265 | variable @code{empty-list} in the usual way, by placing the cursor after |
| 10266 | the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your |
| 10267 | echo area: |
| 10268 | |
| 10269 | @smallexample |
| 10270 | empty-list |
| 10271 | @end smallexample |
| 10272 | |
| 10273 | On the other hand, if you set a variable to be a list with elements, the |
| 10274 | list will appear when you evaluate the variable, as you can see by |
| 10275 | evaluating the following two expressions: |
| 10276 | |
| 10277 | @smallexample |
| 10278 | @group |
| 10279 | (setq animals '(gazelle giraffe lion tiger)) |
| 10280 | |
| 10281 | animals |
| 10282 | @end group |
| 10283 | @end smallexample |
| 10284 | |
| 10285 | Thus, to create a @code{while} loop that tests whether there are any |
| 10286 | items in the list @code{animals}, the first part of the loop will be |
| 10287 | written like this: |
| 10288 | |
| 10289 | @smallexample |
| 10290 | @group |
| 10291 | (while animals |
| 10292 | @dots{} |
| 10293 | @end group |
| 10294 | @end smallexample |
| 10295 | |
| 10296 | @noindent |
| 10297 | When the @code{while} tests its first argument, the variable |
| 10298 | @code{animals} is evaluated. It returns a list. So long as the list |
| 10299 | has elements, the @code{while} considers the results of the test to be |
| 10300 | true; but when the list is empty, it considers the results of the test |
| 10301 | to be false. |
| 10302 | |
| 10303 | To prevent the @code{while} loop from running forever, some mechanism |
| 10304 | needs to be provided to empty the list eventually. An oft-used |
| 10305 | technique is to have one of the subsequent forms in the @code{while} |
| 10306 | expression set the value of the list to be the @sc{cdr} of the list. |
| 10307 | Each time the @code{cdr} function is evaluated, the list will be made |
| 10308 | shorter, until eventually only the empty list will be left. At this |
| 10309 | point, the test of the @code{while} loop will return false, and the |
| 10310 | arguments to the @code{while} will no longer be evaluated. |
| 10311 | |
| 10312 | For example, the list of animals bound to the variable @code{animals} |
| 10313 | can be set to be the @sc{cdr} of the original list with the |
| 10314 | following expression: |
| 10315 | |
| 10316 | @smallexample |
| 10317 | (setq animals (cdr animals)) |
| 10318 | @end smallexample |
| 10319 | |
| 10320 | @noindent |
| 10321 | If you have evaluated the previous expressions and then evaluate this |
| 10322 | expression, you will see @code{(giraffe lion tiger)} appear in the echo |
| 10323 | area. If you evaluate the expression again, @code{(lion tiger)} will |
| 10324 | appear in the echo area. If you evaluate it again and yet again, |
| 10325 | @code{(tiger)} appears and then the empty list, shown by @code{nil}. |
| 10326 | |
| 10327 | A template for a @code{while} loop that uses the @code{cdr} function |
| 10328 | repeatedly to cause the true-or-false-test eventually to test false |
| 10329 | looks like this: |
| 10330 | |
| 10331 | @smallexample |
| 10332 | @group |
| 10333 | (while @var{test-whether-list-is-empty} |
| 10334 | @var{body}@dots{} |
| 10335 | @var{set-list-to-cdr-of-list}) |
| 10336 | @end group |
| 10337 | @end smallexample |
| 10338 | |
| 10339 | This test and use of @code{cdr} can be put together in a function that |
| 10340 | goes through a list and prints each element of the list on a line of its |
| 10341 | own. |
| 10342 | |
| 10343 | @node print-elements-of-list |
| 10344 | @subsection An Example: @code{print-elements-of-list} |
| 10345 | @findex print-elements-of-list |
| 10346 | |
| 10347 | The @code{print-elements-of-list} function illustrates a @code{while} |
| 10348 | loop with a list. |
| 10349 | |
| 10350 | @cindex @file{*scratch*} buffer |
| 10351 | The function requires several lines for its output. If you are |
| 10352 | reading this in a recent instance of GNU Emacs, |
| 10353 | @c GNU Emacs 21, GNU Emacs 22, or a later version, |
| 10354 | you can evaluate the following expression inside of Info, as usual. |
| 10355 | |
| 10356 | If you are using an earlier version of Emacs, you need to copy the |
| 10357 | necessary expressions to your @file{*scratch*} buffer and evaluate |
| 10358 | them there. This is because the echo area had only one line in the |
| 10359 | earlier versions. |
| 10360 | |
| 10361 | You can copy the expressions by marking the beginning of the region |
| 10362 | with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to |
| 10363 | the end of the region and then copying the region using @kbd{M-w} |
| 10364 | (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and |
| 10365 | then provides visual feedback). In the @file{*scratch*} |
| 10366 | buffer, you can yank the expressions back by typing @kbd{C-y} |
| 10367 | (@code{yank}). |
| 10368 | |
| 10369 | After you have copied the expressions to the @file{*scratch*} buffer, |
| 10370 | evaluate each expression in turn. Be sure to evaluate the last |
| 10371 | expression, @code{(print-elements-of-list animals)}, by typing |
| 10372 | @kbd{C-u C-x C-e}, that is, by giving an argument to |
| 10373 | @code{eval-last-sexp}. This will cause the result of the evaluation |
| 10374 | to be printed in the @file{*scratch*} buffer instead of being printed |
| 10375 | in the echo area. (Otherwise you will see something like this in your |
| 10376 | echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which |
| 10377 | each @samp{^J} stands for a `newline'.) |
| 10378 | |
| 10379 | @need 1500 |
| 10380 | In a recent instance of GNU Emacs, you can evaluate these expressions |
| 10381 | directly in the Info buffer, and the echo area will grow to show the |
| 10382 | results. |
| 10383 | |
| 10384 | @smallexample |
| 10385 | @group |
| 10386 | (setq animals '(gazelle giraffe lion tiger)) |
| 10387 | |
| 10388 | (defun print-elements-of-list (list) |
| 10389 | "Print each element of LIST on a line of its own." |
| 10390 | (while list |
| 10391 | (print (car list)) |
| 10392 | (setq list (cdr list)))) |
| 10393 | |
| 10394 | (print-elements-of-list animals) |
| 10395 | @end group |
| 10396 | @end smallexample |
| 10397 | |
| 10398 | @need 1200 |
| 10399 | @noindent |
| 10400 | When you evaluate the three expressions in sequence, you will see |
| 10401 | this: |
| 10402 | |
| 10403 | @smallexample |
| 10404 | @group |
| 10405 | gazelle |
| 10406 | |
| 10407 | giraffe |
| 10408 | |
| 10409 | lion |
| 10410 | |
| 10411 | tiger |
| 10412 | nil |
| 10413 | @end group |
| 10414 | @end smallexample |
| 10415 | |
| 10416 | Each element of the list is printed on a line of its own (that is what |
| 10417 | the function @code{print} does) and then the value returned by the |
| 10418 | function is printed. Since the last expression in the function is the |
| 10419 | @code{while} loop, and since @code{while} loops always return |
| 10420 | @code{nil}, a @code{nil} is printed after the last element of the list. |
| 10421 | |
| 10422 | @node Incrementing Loop |
| 10423 | @subsection A Loop with an Incrementing Counter |
| 10424 | |
| 10425 | A loop is not useful unless it stops when it ought. Besides |
| 10426 | controlling a loop with a list, a common way of stopping a loop is to |
| 10427 | write the first argument as a test that returns false when the correct |
| 10428 | number of repetitions are complete. This means that the loop must |
| 10429 | have a counter---an expression that counts how many times the loop |
| 10430 | repeats itself. |
| 10431 | |
| 10432 | @ifnottex |
| 10433 | @node Incrementing Loop Details |
| 10434 | @unnumberedsubsec Details of an Incrementing Loop |
| 10435 | @end ifnottex |
| 10436 | |
| 10437 | The test for a loop with an incrementing counter can be an expression |
| 10438 | such as @code{(< count desired-number)} which returns @code{t} for |
| 10439 | true if the value of @code{count} is less than the |
| 10440 | @code{desired-number} of repetitions and @code{nil} for false if the |
| 10441 | value of @code{count} is equal to or is greater than the |
| 10442 | @code{desired-number}. The expression that increments the count can |
| 10443 | be a simple @code{setq} such as @code{(setq count (1+ count))}, where |
| 10444 | @code{1+} is a built-in function in Emacs Lisp that adds 1 to its |
| 10445 | argument. (The expression @w{@code{(1+ count)}} has the same result |
| 10446 | as @w{@code{(+ count 1)}}, but is easier for a human to read.) |
| 10447 | |
| 10448 | @need 1250 |
| 10449 | The template for a @code{while} loop controlled by an incrementing |
| 10450 | counter looks like this: |
| 10451 | |
| 10452 | @smallexample |
| 10453 | @group |
| 10454 | @var{set-count-to-initial-value} |
| 10455 | (while (< count desired-number) ; @r{true-or-false-test} |
| 10456 | @var{body}@dots{} |
| 10457 | (setq count (1+ count))) ; @r{incrementer} |
| 10458 | @end group |
| 10459 | @end smallexample |
| 10460 | |
| 10461 | @noindent |
| 10462 | Note that you need to set the initial value of @code{count}; usually it |
| 10463 | is set to 1. |
| 10464 | |
| 10465 | @menu |
| 10466 | * Incrementing Example:: Counting pebbles in a triangle. |
| 10467 | * Inc Example parts:: The parts of the function definition. |
| 10468 | * Inc Example altogether:: Putting the function definition together. |
| 10469 | @end menu |
| 10470 | |
| 10471 | @node Incrementing Example |
| 10472 | @unnumberedsubsubsec Example with incrementing counter |
| 10473 | |
| 10474 | Suppose you are playing on the beach and decide to make a triangle of |
| 10475 | pebbles, putting one pebble in the first row, two in the second row, |
| 10476 | three in the third row and so on, like this: |
| 10477 | |
| 10478 | @sp 1 |
| 10479 | @c pebble diagram |
| 10480 | @ifnottex |
| 10481 | @smallexample |
| 10482 | @group |
| 10483 | * |
| 10484 | * * |
| 10485 | * * * |
| 10486 | * * * * |
| 10487 | @end group |
| 10488 | @end smallexample |
| 10489 | @end ifnottex |
| 10490 | @iftex |
| 10491 | @smallexample |
| 10492 | @group |
| 10493 | @bullet{} |
| 10494 | @bullet{} @bullet{} |
| 10495 | @bullet{} @bullet{} @bullet{} |
| 10496 | @bullet{} @bullet{} @bullet{} @bullet{} |
| 10497 | @end group |
| 10498 | @end smallexample |
| 10499 | @end iftex |
| 10500 | @sp 1 |
| 10501 | |
| 10502 | @noindent |
| 10503 | (About 2500 years ago, Pythagoras and others developed the beginnings of |
| 10504 | number theory by considering questions such as this.) |
| 10505 | |
| 10506 | Suppose you want to know how many pebbles you will need to make a |
| 10507 | triangle with 7 rows? |
| 10508 | |
| 10509 | Clearly, what you need to do is add up the numbers from 1 to 7. There |
| 10510 | are two ways to do this; start with the smallest number, one, and add up |
| 10511 | the list in sequence, 1, 2, 3, 4 and so on; or start with the largest |
| 10512 | number and add the list going down: 7, 6, 5, 4 and so on. Because both |
| 10513 | mechanisms illustrate common ways of writing @code{while} loops, we will |
| 10514 | create two examples, one counting up and the other counting down. In |
| 10515 | this first example, we will start with 1 and add 2, 3, 4 and so on. |
| 10516 | |
| 10517 | If you are just adding up a short list of numbers, the easiest way to do |
| 10518 | it is to add up all the numbers at once. However, if you do not know |
| 10519 | ahead of time how many numbers your list will have, or if you want to be |
| 10520 | prepared for a very long list, then you need to design your addition so |
| 10521 | that what you do is repeat a simple process many times instead of doing |
| 10522 | a more complex process once. |
| 10523 | |
| 10524 | For example, instead of adding up all the pebbles all at once, what you |
| 10525 | can do is add the number of pebbles in the first row, 1, to the number |
| 10526 | in the second row, 2, and then add the total of those two rows to the |
| 10527 | third row, 3. Then you can add the number in the fourth row, 4, to the |
| 10528 | total of the first three rows; and so on. |
| 10529 | |
| 10530 | The critical characteristic of the process is that each repetitive |
| 10531 | action is simple. In this case, at each step we add only two numbers, |
| 10532 | the number of pebbles in the row and the total already found. This |
| 10533 | process of adding two numbers is repeated again and again until the last |
| 10534 | row has been added to the total of all the preceding rows. In a more |
| 10535 | complex loop the repetitive action might not be so simple, but it will |
| 10536 | be simpler than doing everything all at once. |
| 10537 | |
| 10538 | @node Inc Example parts |
| 10539 | @unnumberedsubsubsec The parts of the function definition |
| 10540 | |
| 10541 | The preceding analysis gives us the bones of our function definition: |
| 10542 | first, we will need a variable that we can call @code{total} that will |
| 10543 | be the total number of pebbles. This will be the value returned by |
| 10544 | the function. |
| 10545 | |
| 10546 | Second, we know that the function will require an argument: this |
| 10547 | argument will be the total number of rows in the triangle. It can be |
| 10548 | called @code{number-of-rows}. |
| 10549 | |
| 10550 | Finally, we need a variable to use as a counter. We could call this |
| 10551 | variable @code{counter}, but a better name is @code{row-number}. That |
| 10552 | is because what the counter does in this function is count rows, and a |
| 10553 | program should be written to be as understandable as possible. |
| 10554 | |
| 10555 | When the Lisp interpreter first starts evaluating the expressions in the |
| 10556 | function, the value of @code{total} should be set to zero, since we have |
| 10557 | not added anything to it. Then the function should add the number of |
| 10558 | pebbles in the first row to the total, and then add the number of |
| 10559 | pebbles in the second to the total, and then add the number of |
| 10560 | pebbles in the third row to the total, and so on, until there are no |
| 10561 | more rows left to add. |
| 10562 | |
| 10563 | Both @code{total} and @code{row-number} are used only inside the |
| 10564 | function, so they can be declared as local variables with @code{let} |
| 10565 | and given initial values. Clearly, the initial value for @code{total} |
| 10566 | should be 0. The initial value of @code{row-number} should be 1, |
| 10567 | since we start with the first row. This means that the @code{let} |
| 10568 | statement will look like this: |
| 10569 | |
| 10570 | @smallexample |
| 10571 | @group |
| 10572 | (let ((total 0) |
| 10573 | (row-number 1)) |
| 10574 | @var{body}@dots{}) |
| 10575 | @end group |
| 10576 | @end smallexample |
| 10577 | |
| 10578 | After the internal variables are declared and bound to their initial |
| 10579 | values, we can begin the @code{while} loop. The expression that serves |
| 10580 | as the test should return a value of @code{t} for true so long as the |
| 10581 | @code{row-number} is less than or equal to the @code{number-of-rows}. |
| 10582 | (If the expression tests true only so long as the row number is less |
| 10583 | than the number of rows in the triangle, the last row will never be |
| 10584 | added to the total; hence the row number has to be either less than or |
| 10585 | equal to the number of rows.) |
| 10586 | |
| 10587 | @need 1500 |
| 10588 | @findex <= @r{(less than or equal)} |
| 10589 | Lisp provides the @code{<=} function that returns true if the value of |
| 10590 | its first argument is less than or equal to the value of its second |
| 10591 | argument and false otherwise. So the expression that the @code{while} |
| 10592 | will evaluate as its test should look like this: |
| 10593 | |
| 10594 | @smallexample |
| 10595 | (<= row-number number-of-rows) |
| 10596 | @end smallexample |
| 10597 | |
| 10598 | The total number of pebbles can be found by repeatedly adding the number |
| 10599 | of pebbles in a row to the total already found. Since the number of |
| 10600 | pebbles in the row is equal to the row number, the total can be found by |
| 10601 | adding the row number to the total. (Clearly, in a more complex |
| 10602 | situation, the number of pebbles in the row might be related to the row |
| 10603 | number in a more complicated way; if this were the case, the row number |
| 10604 | would be replaced by the appropriate expression.) |
| 10605 | |
| 10606 | @smallexample |
| 10607 | (setq total (+ total row-number)) |
| 10608 | @end smallexample |
| 10609 | |
| 10610 | @noindent |
| 10611 | What this does is set the new value of @code{total} to be equal to the |
| 10612 | sum of adding the number of pebbles in the row to the previous total. |
| 10613 | |
| 10614 | After setting the value of @code{total}, the conditions need to be |
| 10615 | established for the next repetition of the loop, if there is one. This |
| 10616 | is done by incrementing the value of the @code{row-number} variable, |
| 10617 | which serves as a counter. After the @code{row-number} variable has |
| 10618 | been incremented, the true-or-false-test at the beginning of the |
| 10619 | @code{while} loop tests whether its value is still less than or equal to |
| 10620 | the value of the @code{number-of-rows} and if it is, adds the new value |
| 10621 | of the @code{row-number} variable to the @code{total} of the previous |
| 10622 | repetition of the loop. |
| 10623 | |
| 10624 | @need 1200 |
| 10625 | The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the |
| 10626 | @code{row-number} variable can be incremented with this expression: |
| 10627 | |
| 10628 | @smallexample |
| 10629 | (setq row-number (1+ row-number)) |
| 10630 | @end smallexample |
| 10631 | |
| 10632 | @node Inc Example altogether |
| 10633 | @unnumberedsubsubsec Putting the function definition together |
| 10634 | |
| 10635 | We have created the parts for the function definition; now we need to |
| 10636 | put them together. |
| 10637 | |
| 10638 | @need 800 |
| 10639 | First, the contents of the @code{while} expression: |
| 10640 | |
| 10641 | @smallexample |
| 10642 | @group |
| 10643 | (while (<= row-number number-of-rows) ; @r{true-or-false-test} |
| 10644 | (setq total (+ total row-number)) |
| 10645 | (setq row-number (1+ row-number))) ; @r{incrementer} |
| 10646 | @end group |
| 10647 | @end smallexample |
| 10648 | |
| 10649 | Along with the @code{let} expression varlist, this very nearly |
| 10650 | completes the body of the function definition. However, it requires |
| 10651 | one final element, the need for which is somewhat subtle. |
| 10652 | |
| 10653 | The final touch is to place the variable @code{total} on a line by |
| 10654 | itself after the @code{while} expression. Otherwise, the value returned |
| 10655 | by the whole function is the value of the last expression that is |
| 10656 | evaluated in the body of the @code{let}, and this is the value |
| 10657 | returned by the @code{while}, which is always @code{nil}. |
| 10658 | |
| 10659 | This may not be evident at first sight. It almost looks as if the |
| 10660 | incrementing expression is the last expression of the whole function. |
| 10661 | But that expression is part of the body of the @code{while}; it is the |
| 10662 | last element of the list that starts with the symbol @code{while}. |
| 10663 | Moreover, the whole of the @code{while} loop is a list within the body |
| 10664 | of the @code{let}. |
| 10665 | |
| 10666 | @need 1250 |
| 10667 | In outline, the function will look like this: |
| 10668 | |
| 10669 | @smallexample |
| 10670 | @group |
| 10671 | (defun @var{name-of-function} (@var{argument-list}) |
| 10672 | "@var{documentation}@dots{}" |
| 10673 | (let (@var{varlist}) |
| 10674 | (while (@var{true-or-false-test}) |
| 10675 | @var{body-of-while}@dots{} ) |
| 10676 | @dots{} )) ; @r{Need final expression here.} |
| 10677 | @end group |
| 10678 | @end smallexample |
| 10679 | |
| 10680 | The result of evaluating the @code{let} is what is going to be returned |
| 10681 | by the @code{defun} since the @code{let} is not embedded within any |
| 10682 | containing list, except for the @code{defun} as a whole. However, if |
| 10683 | the @code{while} is the last element of the @code{let} expression, the |
| 10684 | function will always return @code{nil}. This is not what we want! |
| 10685 | Instead, what we want is the value of the variable @code{total}. This |
| 10686 | is returned by simply placing the symbol as the last element of the list |
| 10687 | starting with @code{let}. It gets evaluated after the preceding |
| 10688 | elements of the list are evaluated, which means it gets evaluated after |
| 10689 | it has been assigned the correct value for the total. |
| 10690 | |
| 10691 | It may be easier to see this by printing the list starting with |
| 10692 | @code{let} all on one line. This format makes it evident that the |
| 10693 | @var{varlist} and @code{while} expressions are the second and third |
| 10694 | elements of the list starting with @code{let}, and the @code{total} is |
| 10695 | the last element: |
| 10696 | |
| 10697 | @smallexample |
| 10698 | @group |
| 10699 | (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total) |
| 10700 | @end group |
| 10701 | @end smallexample |
| 10702 | |
| 10703 | @need 1200 |
| 10704 | Putting everything together, the @code{triangle} function definition |
| 10705 | looks like this: |
| 10706 | |
| 10707 | @smallexample |
| 10708 | @group |
| 10709 | (defun triangle (number-of-rows) ; @r{Version with} |
| 10710 | ; @r{ incrementing counter.} |
| 10711 | "Add up the number of pebbles in a triangle. |
| 10712 | The first row has one pebble, the second row two pebbles, |
| 10713 | the third row three pebbles, and so on. |
| 10714 | The argument is NUMBER-OF-ROWS." |
| 10715 | @end group |
| 10716 | @group |
| 10717 | (let ((total 0) |
| 10718 | (row-number 1)) |
| 10719 | (while (<= row-number number-of-rows) |
| 10720 | (setq total (+ total row-number)) |
| 10721 | (setq row-number (1+ row-number))) |
| 10722 | total)) |
| 10723 | @end group |
| 10724 | @end smallexample |
| 10725 | |
| 10726 | @need 1200 |
| 10727 | After you have installed @code{triangle} by evaluating the function, you |
| 10728 | can try it out. Here are two examples: |
| 10729 | |
| 10730 | @smallexample |
| 10731 | @group |
| 10732 | (triangle 4) |
| 10733 | |
| 10734 | (triangle 7) |
| 10735 | @end group |
| 10736 | @end smallexample |
| 10737 | |
| 10738 | @noindent |
| 10739 | The sum of the first four numbers is 10 and the sum of the first seven |
| 10740 | numbers is 28. |
| 10741 | |
| 10742 | @node Decrementing Loop |
| 10743 | @subsection Loop with a Decrementing Counter |
| 10744 | |
| 10745 | Another common way to write a @code{while} loop is to write the test |
| 10746 | so that it determines whether a counter is greater than zero. So long |
| 10747 | as the counter is greater than zero, the loop is repeated. But when |
| 10748 | the counter is equal to or less than zero, the loop is stopped. For |
| 10749 | this to work, the counter has to start out greater than zero and then |
| 10750 | be made smaller and smaller by a form that is evaluated |
| 10751 | repeatedly. |
| 10752 | |
| 10753 | The test will be an expression such as @code{(> counter 0)} which |
| 10754 | returns @code{t} for true if the value of @code{counter} is greater |
| 10755 | than zero, and @code{nil} for false if the value of @code{counter} is |
| 10756 | equal to or less than zero. The expression that makes the number |
| 10757 | smaller and smaller can be a simple @code{setq} such as @code{(setq |
| 10758 | counter (1- counter))}, where @code{1-} is a built-in function in |
| 10759 | Emacs Lisp that subtracts 1 from its argument. |
| 10760 | |
| 10761 | @need 1250 |
| 10762 | The template for a decrementing @code{while} loop looks like this: |
| 10763 | |
| 10764 | @smallexample |
| 10765 | @group |
| 10766 | (while (> counter 0) ; @r{true-or-false-test} |
| 10767 | @var{body}@dots{} |
| 10768 | (setq counter (1- counter))) ; @r{decrementer} |
| 10769 | @end group |
| 10770 | @end smallexample |
| 10771 | |
| 10772 | @menu |
| 10773 | * Decrementing Example:: More pebbles on the beach. |
| 10774 | * Dec Example parts:: The parts of the function definition. |
| 10775 | * Dec Example altogether:: Putting the function definition together. |
| 10776 | @end menu |
| 10777 | |
| 10778 | @node Decrementing Example |
| 10779 | @unnumberedsubsubsec Example with decrementing counter |
| 10780 | |
| 10781 | To illustrate a loop with a decrementing counter, we will rewrite the |
| 10782 | @code{triangle} function so the counter decreases to zero. |
| 10783 | |
| 10784 | This is the reverse of the earlier version of the function. In this |
| 10785 | case, to find out how many pebbles are needed to make a triangle with |
| 10786 | 3 rows, add the number of pebbles in the third row, 3, to the number |
| 10787 | in the preceding row, 2, and then add the total of those two rows to |
| 10788 | the row that precedes them, which is 1. |
| 10789 | |
| 10790 | Likewise, to find the number of pebbles in a triangle with 7 rows, add |
| 10791 | the number of pebbles in the seventh row, 7, to the number in the |
| 10792 | preceding row, which is 6, and then add the total of those two rows to |
| 10793 | the row that precedes them, which is 5, and so on. As in the previous |
| 10794 | example, each addition only involves adding two numbers, the total of |
| 10795 | the rows already added up and the number of pebbles in the row that is |
| 10796 | being added to the total. This process of adding two numbers is |
| 10797 | repeated again and again until there are no more pebbles to add. |
| 10798 | |
| 10799 | We know how many pebbles to start with: the number of pebbles in the |
| 10800 | last row is equal to the number of rows. If the triangle has seven |
| 10801 | rows, the number of pebbles in the last row is 7. Likewise, we know how |
| 10802 | many pebbles are in the preceding row: it is one less than the number in |
| 10803 | the row. |
| 10804 | |
| 10805 | @node Dec Example parts |
| 10806 | @unnumberedsubsubsec The parts of the function definition |
| 10807 | |
| 10808 | We start with three variables: the total number of rows in the |
| 10809 | triangle; the number of pebbles in a row; and the total number of |
| 10810 | pebbles, which is what we want to calculate. These variables can be |
| 10811 | named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and |
| 10812 | @code{total}, respectively. |
| 10813 | |
| 10814 | Both @code{total} and @code{number-of-pebbles-in-row} are used only |
| 10815 | inside the function and are declared with @code{let}. The initial |
| 10816 | value of @code{total} should, of course, be zero. However, the |
| 10817 | initial value of @code{number-of-pebbles-in-row} should be equal to |
| 10818 | the number of rows in the triangle, since the addition will start with |
| 10819 | the longest row. |
| 10820 | |
| 10821 | @need 1250 |
| 10822 | This means that the beginning of the @code{let} expression will look |
| 10823 | like this: |
| 10824 | |
| 10825 | @smallexample |
| 10826 | @group |
| 10827 | (let ((total 0) |
| 10828 | (number-of-pebbles-in-row number-of-rows)) |
| 10829 | @var{body}@dots{}) |
| 10830 | @end group |
| 10831 | @end smallexample |
| 10832 | |
| 10833 | The total number of pebbles can be found by repeatedly adding the number |
| 10834 | of pebbles in a row to the total already found, that is, by repeatedly |
| 10835 | evaluating the following expression: |
| 10836 | |
| 10837 | @smallexample |
| 10838 | (setq total (+ total number-of-pebbles-in-row)) |
| 10839 | @end smallexample |
| 10840 | |
| 10841 | @noindent |
| 10842 | After the @code{number-of-pebbles-in-row} is added to the @code{total}, |
| 10843 | the @code{number-of-pebbles-in-row} should be decremented by one, since |
| 10844 | the next time the loop repeats, the preceding row will be |
| 10845 | added to the total. |
| 10846 | |
| 10847 | The number of pebbles in a preceding row is one less than the number of |
| 10848 | pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be |
| 10849 | used to compute the number of pebbles in the preceding row. This can be |
| 10850 | done with the following expression: |
| 10851 | |
| 10852 | @smallexample |
| 10853 | @group |
| 10854 | (setq number-of-pebbles-in-row |
| 10855 | (1- number-of-pebbles-in-row)) |
| 10856 | @end group |
| 10857 | @end smallexample |
| 10858 | |
| 10859 | Finally, we know that the @code{while} loop should stop making repeated |
| 10860 | additions when there are no pebbles in a row. So the test for |
| 10861 | the @code{while} loop is simply: |
| 10862 | |
| 10863 | @smallexample |
| 10864 | (while (> number-of-pebbles-in-row 0) |
| 10865 | @end smallexample |
| 10866 | |
| 10867 | @node Dec Example altogether |
| 10868 | @unnumberedsubsubsec Putting the function definition together |
| 10869 | |
| 10870 | We can put these expressions together to create a function definition |
| 10871 | that works. However, on examination, we find that one of the local |
| 10872 | variables is unneeded! |
| 10873 | |
| 10874 | @need 1250 |
| 10875 | The function definition looks like this: |
| 10876 | |
| 10877 | @smallexample |
| 10878 | @group |
| 10879 | ;;; @r{First subtractive version.} |
| 10880 | (defun triangle (number-of-rows) |
| 10881 | "Add up the number of pebbles in a triangle." |
| 10882 | (let ((total 0) |
| 10883 | (number-of-pebbles-in-row number-of-rows)) |
| 10884 | (while (> number-of-pebbles-in-row 0) |
| 10885 | (setq total (+ total number-of-pebbles-in-row)) |
| 10886 | (setq number-of-pebbles-in-row |
| 10887 | (1- number-of-pebbles-in-row))) |
| 10888 | total)) |
| 10889 | @end group |
| 10890 | @end smallexample |
| 10891 | |
| 10892 | As written, this function works. |
| 10893 | |
| 10894 | However, we do not need @code{number-of-pebbles-in-row}. |
| 10895 | |
| 10896 | @cindex Argument as local variable |
| 10897 | When the @code{triangle} function is evaluated, the symbol |
| 10898 | @code{number-of-rows} will be bound to a number, giving it an initial |
| 10899 | value. That number can be changed in the body of the function as if |
| 10900 | it were a local variable, without any fear that such a change will |
| 10901 | effect the value of the variable outside of the function. This is a |
| 10902 | very useful characteristic of Lisp; it means that the variable |
| 10903 | @code{number-of-rows} can be used anywhere in the function where |
| 10904 | @code{number-of-pebbles-in-row} is used. |
| 10905 | |
| 10906 | @need 800 |
| 10907 | Here is a second version of the function written a bit more cleanly: |
| 10908 | |
| 10909 | @smallexample |
| 10910 | @group |
| 10911 | (defun triangle (number) ; @r{Second version.} |
| 10912 | "Return sum of numbers 1 through NUMBER inclusive." |
| 10913 | (let ((total 0)) |
| 10914 | (while (> number 0) |
| 10915 | (setq total (+ total number)) |
| 10916 | (setq number (1- number))) |
| 10917 | total)) |
| 10918 | @end group |
| 10919 | @end smallexample |
| 10920 | |
| 10921 | In brief, a properly written @code{while} loop will consist of three parts: |
| 10922 | |
| 10923 | @enumerate |
| 10924 | @item |
| 10925 | A test that will return false after the loop has repeated itself the |
| 10926 | correct number of times. |
| 10927 | |
| 10928 | @item |
| 10929 | An expression the evaluation of which will return the value desired |
| 10930 | after being repeatedly evaluated. |
| 10931 | |
| 10932 | @item |
| 10933 | An expression to change the value passed to the true-or-false-test so |
| 10934 | that the test returns false after the loop has repeated itself the right |
| 10935 | number of times. |
| 10936 | @end enumerate |
| 10937 | |
| 10938 | @node dolist dotimes |
| 10939 | @section Save your time: @code{dolist} and @code{dotimes} |
| 10940 | |
| 10941 | In addition to @code{while}, both @code{dolist} and @code{dotimes} |
| 10942 | provide for looping. Sometimes these are quicker to write than the |
| 10943 | equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, , |
| 10944 | Macros, elisp, The GNU Emacs Lisp Reference Manual}. ) |
| 10945 | |
| 10946 | @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a |
| 10947 | list': @code{dolist} automatically shortens the list each time it |
| 10948 | loops---takes the @sc{cdr} of the list---and binds the @sc{car} of |
| 10949 | each shorter version of the list to the first of its arguments. |
| 10950 | |
| 10951 | @code{dotimes} loops a specific number of times: you specify the number. |
| 10952 | |
| 10953 | @menu |
| 10954 | * dolist:: |
| 10955 | * dotimes:: |
| 10956 | @end menu |
| 10957 | |
| 10958 | @node dolist |
| 10959 | @unnumberedsubsec The @code{dolist} Macro |
| 10960 | @findex dolist |
| 10961 | |
| 10962 | Suppose, for example, you want to reverse a list, so that |
| 10963 | ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''. |
| 10964 | |
| 10965 | @need 1250 |
| 10966 | In practice, you would use the @code{reverse} function, like this: |
| 10967 | |
| 10968 | @smallexample |
| 10969 | @group |
| 10970 | (setq animals '(gazelle giraffe lion tiger)) |
| 10971 | |
| 10972 | (reverse animals) |
| 10973 | @end group |
| 10974 | @end smallexample |
| 10975 | |
| 10976 | @need 800 |
| 10977 | @noindent |
| 10978 | Here is how you could reverse the list using a @code{while} loop: |
| 10979 | |
| 10980 | @smallexample |
| 10981 | @group |
| 10982 | (setq animals '(gazelle giraffe lion tiger)) |
| 10983 | |
| 10984 | (defun reverse-list-with-while (list) |
| 10985 | "Using while, reverse the order of LIST." |
| 10986 | (let (value) ; make sure list starts empty |
| 10987 | (while list |
| 10988 | (setq value (cons (car list) value)) |
| 10989 | (setq list (cdr list))) |
| 10990 | value)) |
| 10991 | |
| 10992 | (reverse-list-with-while animals) |
| 10993 | @end group |
| 10994 | @end smallexample |
| 10995 | |
| 10996 | @need 800 |
| 10997 | @noindent |
| 10998 | And here is how you could use the @code{dolist} macro: |
| 10999 | |
| 11000 | @smallexample |
| 11001 | @group |
| 11002 | (setq animals '(gazelle giraffe lion tiger)) |
| 11003 | |
| 11004 | (defun reverse-list-with-dolist (list) |
| 11005 | "Using dolist, reverse the order of LIST." |
| 11006 | (let (value) ; make sure list starts empty |
| 11007 | (dolist (element list value) |
| 11008 | (setq value (cons element value))))) |
| 11009 | |
| 11010 | (reverse-list-with-dolist animals) |
| 11011 | @end group |
| 11012 | @end smallexample |
| 11013 | |
| 11014 | @need 1250 |
| 11015 | @noindent |
| 11016 | In Info, you can place your cursor after the closing parenthesis of |
| 11017 | each expression and type @kbd{C-x C-e}; in each case, you should see |
| 11018 | |
| 11019 | @smallexample |
| 11020 | (tiger lion giraffe gazelle) |
| 11021 | @end smallexample |
| 11022 | |
| 11023 | @noindent |
| 11024 | in the echo area. |
| 11025 | |
| 11026 | For this example, the existing @code{reverse} function is obviously best. |
| 11027 | The @code{while} loop is just like our first example (@pxref{Loop |
| 11028 | Example, , A @code{while} Loop and a List}). The @code{while} first |
| 11029 | checks whether the list has elements; if so, it constructs a new list |
| 11030 | by adding the first element of the list to the existing list (which in |
| 11031 | the first iteration of the loop is @code{nil}). Since the second |
| 11032 | element is prepended in front of the first element, and the third |
| 11033 | element is prepended in front of the second element, the list is reversed. |
| 11034 | |
| 11035 | In the expression using a @code{while} loop, |
| 11036 | the @w{@code{(setq list (cdr list))}} |
| 11037 | expression shortens the list, so the @code{while} loop eventually |
| 11038 | stops. In addition, it provides the @code{cons} expression with a new |
| 11039 | first element by creating a new and shorter list at each repetition of |
| 11040 | the loop. |
| 11041 | |
| 11042 | The @code{dolist} expression does very much the same as the |
| 11043 | @code{while} expression, except that the @code{dolist} macro does some |
| 11044 | of the work you have to do when writing a @code{while} expression. |
| 11045 | |
| 11046 | Like a @code{while} loop, a @code{dolist} loops. What is different is |
| 11047 | that it automatically shortens the list each time it loops---it |
| 11048 | `@sc{cdr}s down the list' on its own---and it automatically binds |
| 11049 | the @sc{car} of each shorter version of the list to the first of its |
| 11050 | arguments. |
| 11051 | |
| 11052 | In the example, the @sc{car} of each shorter version of the list is |
| 11053 | referred to using the symbol @samp{element}, the list itself is called |
| 11054 | @samp{list}, and the value returned is called @samp{value}. The |
| 11055 | remainder of the @code{dolist} expression is the body. |
| 11056 | |
| 11057 | The @code{dolist} expression binds the @sc{car} of each shorter |
| 11058 | version of the list to @code{element} and then evaluates the body of |
| 11059 | the expression; and repeats the loop. The result is returned in |
| 11060 | @code{value}. |
| 11061 | |
| 11062 | @node dotimes |
| 11063 | @unnumberedsubsec The @code{dotimes} Macro |
| 11064 | @findex dotimes |
| 11065 | |
| 11066 | The @code{dotimes} macro is similar to @code{dolist}, except that it |
| 11067 | loops a specific number of times. |
| 11068 | |
| 11069 | The first argument to @code{dotimes} is assigned the numbers 0, 1, 2 |
| 11070 | and so forth each time around the loop, and the value of the third |
| 11071 | argument is returned. You need to provide the value of the second |
| 11072 | argument, which is how many times the macro loops. |
| 11073 | |
| 11074 | @need 1250 |
| 11075 | For example, the following binds the numbers from 0 up to, but not |
| 11076 | including, the number 3 to the first argument, @var{number}, and then |
| 11077 | constructs a list of the three numbers. (The first number is 0, the |
| 11078 | second number is 1, and the third number is 2; this makes a total of |
| 11079 | three numbers in all, starting with zero as the first number.) |
| 11080 | |
| 11081 | @smallexample |
| 11082 | @group |
| 11083 | (let (value) ; otherwise a value is a void variable |
| 11084 | (dotimes (number 3 value) |
| 11085 | (setq value (cons number value)))) |
| 11086 | |
| 11087 | @result{} (2 1 0) |
| 11088 | @end group |
| 11089 | @end smallexample |
| 11090 | |
| 11091 | @noindent |
| 11092 | @code{dotimes} returns @code{value}, so the way to use |
| 11093 | @code{dotimes} is to operate on some expression @var{number} number of |
| 11094 | times and then return the result, either as a list or an atom. |
| 11095 | |
| 11096 | @need 1250 |
| 11097 | Here is an example of a @code{defun} that uses @code{dotimes} to add |
| 11098 | up the number of pebbles in a triangle. |
| 11099 | |
| 11100 | @smallexample |
| 11101 | @group |
| 11102 | (defun triangle-using-dotimes (number-of-rows) |
| 11103 | "Using dotimes, add up the number of pebbles in a triangle." |
| 11104 | (let ((total 0)) ; otherwise a total is a void variable |
| 11105 | (dotimes (number number-of-rows total) |
| 11106 | (setq total (+ total (1+ number)))))) |
| 11107 | |
| 11108 | (triangle-using-dotimes 4) |
| 11109 | @end group |
| 11110 | @end smallexample |
| 11111 | |
| 11112 | @node Recursion |
| 11113 | @section Recursion |
| 11114 | @cindex Recursion |
| 11115 | |
| 11116 | A recursive function contains code that tells the Lisp interpreter to |
| 11117 | call a program that runs exactly like itself, but with slightly |
| 11118 | different arguments. The code runs exactly the same because it has |
| 11119 | the same name. However, even though the program has the same name, it |
| 11120 | is not the same entity. It is different. In the jargon, it is a |
| 11121 | different `instance'. |
| 11122 | |
| 11123 | Eventually, if the program is written correctly, the `slightly |
| 11124 | different arguments' will become sufficiently different from the first |
| 11125 | arguments that the final instance will stop. |
| 11126 | |
| 11127 | @menu |
| 11128 | * Building Robots:: Same model, different serial number ... |
| 11129 | * Recursive Definition Parts:: Walk until you stop ... |
| 11130 | * Recursion with list:: Using a list as the test whether to recurse. |
| 11131 | * Recursive triangle function:: |
| 11132 | * Recursion with cond:: |
| 11133 | * Recursive Patterns:: Often used templates. |
| 11134 | * No Deferment:: Don't store up work ... |
| 11135 | * No deferment solution:: |
| 11136 | @end menu |
| 11137 | |
| 11138 | @node Building Robots |
| 11139 | @subsection Building Robots: Extending the Metaphor |
| 11140 | @cindex Building robots |
| 11141 | @cindex Robots, building |
| 11142 | |
| 11143 | It is sometimes helpful to think of a running program as a robot that |
| 11144 | does a job. In doing its job, a recursive function calls on a second |
| 11145 | robot to help it. The second robot is identical to the first in every |
| 11146 | way, except that the second robot helps the first and has been |
| 11147 | passed different arguments than the first. |
| 11148 | |
| 11149 | In a recursive function, the second robot may call a third; and the |
| 11150 | third may call a fourth, and so on. Each of these is a different |
| 11151 | entity; but all are clones. |
| 11152 | |
| 11153 | Since each robot has slightly different instructions---the arguments |
| 11154 | will differ from one robot to the next---the last robot should know |
| 11155 | when to stop. |
| 11156 | |
| 11157 | Let's expand on the metaphor in which a computer program is a robot. |
| 11158 | |
| 11159 | A function definition provides the blueprints for a robot. When you |
| 11160 | install a function definition, that is, when you evaluate a |
| 11161 | @code{defun} macro, you install the necessary equipment to build |
| 11162 | robots. It is as if you were in a factory, setting up an assembly |
| 11163 | line. Robots with the same name are built according to the same |
| 11164 | blueprints. So they have, as it were, the same `model number', but a |
| 11165 | different `serial number'. |
| 11166 | |
| 11167 | We often say that a recursive function `calls itself'. What we mean |
| 11168 | is that the instructions in a recursive function cause the Lisp |
| 11169 | interpreter to run a different function that has the same name and |
| 11170 | does the same job as the first, but with different arguments. |
| 11171 | |
| 11172 | It is important that the arguments differ from one instance to the |
| 11173 | next; otherwise, the process will never stop. |
| 11174 | |
| 11175 | @node Recursive Definition Parts |
| 11176 | @subsection The Parts of a Recursive Definition |
| 11177 | @cindex Parts of a Recursive Definition |
| 11178 | @cindex Recursive Definition Parts |
| 11179 | |
| 11180 | A recursive function typically contains a conditional expression which |
| 11181 | has three parts: |
| 11182 | |
| 11183 | @enumerate |
| 11184 | @item |
| 11185 | A true-or-false-test that determines whether the function is called |
| 11186 | again, here called the @dfn{do-again-test}. |
| 11187 | |
| 11188 | @item |
| 11189 | The name of the function. When this name is called, a new instance of |
| 11190 | the function---a new robot, as it were---is created and told what to do. |
| 11191 | |
| 11192 | @item |
| 11193 | An expression that returns a different value each time the function is |
| 11194 | called, here called the @dfn{next-step-expression}. Consequently, the |
| 11195 | argument (or arguments) passed to the new instance of the function |
| 11196 | will be different from that passed to the previous instance. This |
| 11197 | causes the conditional expression, the @dfn{do-again-test}, to test |
| 11198 | false after the correct number of repetitions. |
| 11199 | @end enumerate |
| 11200 | |
| 11201 | Recursive functions can be much simpler than any other kind of |
| 11202 | function. Indeed, when people first start to use them, they often look |
| 11203 | so mysteriously simple as to be incomprehensible. Like riding a |
| 11204 | bicycle, reading a recursive function definition takes a certain knack |
| 11205 | which is hard at first but then seems simple. |
| 11206 | |
| 11207 | @need 1200 |
| 11208 | There are several different common recursive patterns. A very simple |
| 11209 | pattern looks like this: |
| 11210 | |
| 11211 | @smallexample |
| 11212 | @group |
| 11213 | (defun @var{name-of-recursive-function} (@var{argument-list}) |
| 11214 | "@var{documentation}@dots{}" |
| 11215 | (if @var{do-again-test} |
| 11216 | @var{body}@dots{} |
| 11217 | (@var{name-of-recursive-function} |
| 11218 | @var{next-step-expression}))) |
| 11219 | @end group |
| 11220 | @end smallexample |
| 11221 | |
| 11222 | Each time a recursive function is evaluated, a new instance of it is |
| 11223 | created and told what to do. The arguments tell the instance what to do. |
| 11224 | |
| 11225 | An argument is bound to the value of the next-step-expression. Each |
| 11226 | instance runs with a different value of the next-step-expression. |
| 11227 | |
| 11228 | The value in the next-step-expression is used in the do-again-test. |
| 11229 | |
| 11230 | The value returned by the next-step-expression is passed to the new |
| 11231 | instance of the function, which evaluates it (or some |
| 11232 | transmogrification of it) to determine whether to continue or stop. |
| 11233 | The next-step-expression is designed so that the do-again-test returns |
| 11234 | false when the function should no longer be repeated. |
| 11235 | |
| 11236 | The do-again-test is sometimes called the @dfn{stop condition}, |
| 11237 | since it stops the repetitions when it tests false. |
| 11238 | |
| 11239 | @node Recursion with list |
| 11240 | @subsection Recursion with a List |
| 11241 | |
| 11242 | The example of a @code{while} loop that printed the elements of a list |
| 11243 | of numbers can be written recursively. Here is the code, including |
| 11244 | an expression to set the value of the variable @code{animals} to a list. |
| 11245 | |
| 11246 | If you are reading this in Info in Emacs, you can evaluate this |
| 11247 | expression directly in Info. Otherwise, you must copy the example |
| 11248 | to the @file{*scratch*} buffer and evaluate each expression there. |
| 11249 | Use @kbd{C-u C-x C-e} to evaluate the |
| 11250 | @code{(print-elements-recursively animals)} expression so that the |
| 11251 | results are printed in the buffer; otherwise the Lisp interpreter will |
| 11252 | try to squeeze the results into the one line of the echo area. |
| 11253 | |
| 11254 | Also, place your cursor immediately after the last closing parenthesis |
| 11255 | of the @code{print-elements-recursively} function, before the comment. |
| 11256 | Otherwise, the Lisp interpreter will try to evaluate the comment. |
| 11257 | |
| 11258 | @findex print-elements-recursively |
| 11259 | @smallexample |
| 11260 | @group |
| 11261 | (setq animals '(gazelle giraffe lion tiger)) |
| 11262 | |
| 11263 | (defun print-elements-recursively (list) |
| 11264 | "Print each element of LIST on a line of its own. |
| 11265 | Uses recursion." |
| 11266 | (when list ; @r{do-again-test} |
| 11267 | (print (car list)) ; @r{body} |
| 11268 | (print-elements-recursively ; @r{recursive call} |
| 11269 | (cdr list)))) ; @r{next-step-expression} |
| 11270 | |
| 11271 | (print-elements-recursively animals) |
| 11272 | @end group |
| 11273 | @end smallexample |
| 11274 | |
| 11275 | The @code{print-elements-recursively} function first tests whether |
| 11276 | there is any content in the list; if there is, the function prints the |
| 11277 | first element of the list, the @sc{car} of the list. Then the |
| 11278 | function `invokes itself', but gives itself as its argument, not the |
| 11279 | whole list, but the second and subsequent elements of the list, the |
| 11280 | @sc{cdr} of the list. |
| 11281 | |
| 11282 | Put another way, if the list is not empty, the function invokes |
| 11283 | another instance of code that is similar to the initial code, but is a |
| 11284 | different thread of execution, with different arguments than the first |
| 11285 | instance. |
| 11286 | |
| 11287 | Put in yet another way, if the list is not empty, the first robot |
| 11288 | assembles a second robot and tells it what to do; the second robot is |
| 11289 | a different individual from the first, but is the same model. |
| 11290 | |
| 11291 | When the second evaluation occurs, the @code{when} expression is |
| 11292 | evaluated and if true, prints the first element of the list it |
| 11293 | receives as its argument (which is the second element of the original |
| 11294 | list). Then the function `calls itself' with the @sc{cdr} of the list |
| 11295 | it is invoked with, which (the second time around) is the @sc{cdr} of |
| 11296 | the @sc{cdr} of the original list. |
| 11297 | |
| 11298 | Note that although we say that the function `calls itself', what we |
| 11299 | mean is that the Lisp interpreter assembles and instructs a new |
| 11300 | instance of the program. The new instance is a clone of the first, |
| 11301 | but is a separate individual. |
| 11302 | |
| 11303 | Each time the function `invokes itself', it invokes itself on a |
| 11304 | shorter version of the original list. It creates a new instance that |
| 11305 | works on a shorter list. |
| 11306 | |
| 11307 | Eventually, the function invokes itself on an empty list. It creates |
| 11308 | a new instance whose argument is @code{nil}. The conditional expression |
| 11309 | tests the value of @code{list}. Since the value of @code{list} is |
| 11310 | @code{nil}, the @code{when} expression tests false so the then-part is |
| 11311 | not evaluated. The function as a whole then returns @code{nil}. |
| 11312 | |
| 11313 | @need 1200 |
| 11314 | When you evaluate the expression @code{(print-elements-recursively |
| 11315 | animals)} in the @file{*scratch*} buffer, you see this result: |
| 11316 | |
| 11317 | @smallexample |
| 11318 | @group |
| 11319 | gazelle |
| 11320 | |
| 11321 | giraffe |
| 11322 | |
| 11323 | lion |
| 11324 | |
| 11325 | tiger |
| 11326 | nil |
| 11327 | @end group |
| 11328 | @end smallexample |
| 11329 | |
| 11330 | @need 2000 |
| 11331 | @node Recursive triangle function |
| 11332 | @subsection Recursion in Place of a Counter |
| 11333 | @findex triangle-recursively |
| 11334 | |
| 11335 | @need 1200 |
| 11336 | The @code{triangle} function described in a previous section can also |
| 11337 | be written recursively. It looks like this: |
| 11338 | |
| 11339 | @smallexample |
| 11340 | @group |
| 11341 | (defun triangle-recursively (number) |
| 11342 | "Return the sum of the numbers 1 through NUMBER inclusive. |
| 11343 | Uses recursion." |
| 11344 | (if (= number 1) ; @r{do-again-test} |
| 11345 | 1 ; @r{then-part} |
| 11346 | (+ number ; @r{else-part} |
| 11347 | (triangle-recursively ; @r{recursive call} |
| 11348 | (1- number))))) ; @r{next-step-expression} |
| 11349 | |
| 11350 | (triangle-recursively 7) |
| 11351 | @end group |
| 11352 | @end smallexample |
| 11353 | |
| 11354 | @noindent |
| 11355 | You can install this function by evaluating it and then try it by |
| 11356 | evaluating @code{(triangle-recursively 7)}. (Remember to put your |
| 11357 | cursor immediately after the last parenthesis of the function |
| 11358 | definition, before the comment.) The function evaluates to 28. |
| 11359 | |
| 11360 | To understand how this function works, let's consider what happens in the |
| 11361 | various cases when the function is passed 1, 2, 3, or 4 as the value of |
| 11362 | its argument. |
| 11363 | |
| 11364 | @menu |
| 11365 | * Recursive Example arg of 1 or 2:: |
| 11366 | * Recursive Example arg of 3 or 4:: |
| 11367 | @end menu |
| 11368 | |
| 11369 | @ifnottex |
| 11370 | @node Recursive Example arg of 1 or 2 |
| 11371 | @unnumberedsubsubsec An argument of 1 or 2 |
| 11372 | @end ifnottex |
| 11373 | |
| 11374 | First, what happens if the value of the argument is 1? |
| 11375 | |
| 11376 | The function has an @code{if} expression after the documentation |
| 11377 | string. It tests whether the value of @code{number} is equal to 1; if |
| 11378 | so, Emacs evaluates the then-part of the @code{if} expression, which |
| 11379 | returns the number 1 as the value of the function. (A triangle with |
| 11380 | one row has one pebble in it.) |
| 11381 | |
| 11382 | Suppose, however, that the value of the argument is 2. In this case, |
| 11383 | Emacs evaluates the else-part of the @code{if} expression. |
| 11384 | |
| 11385 | @need 1200 |
| 11386 | The else-part consists of an addition, the recursive call to |
| 11387 | @code{triangle-recursively} and a decrementing action; and it looks like |
| 11388 | this: |
| 11389 | |
| 11390 | @smallexample |
| 11391 | (+ number (triangle-recursively (1- number))) |
| 11392 | @end smallexample |
| 11393 | |
| 11394 | When Emacs evaluates this expression, the innermost expression is |
| 11395 | evaluated first; then the other parts in sequence. Here are the steps |
| 11396 | in detail: |
| 11397 | |
| 11398 | @table @i |
| 11399 | @item Step 1 @w{ } Evaluate the innermost expression. |
| 11400 | |
| 11401 | The innermost expression is @code{(1- number)} so Emacs decrements the |
| 11402 | value of @code{number} from 2 to 1. |
| 11403 | |
| 11404 | @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function. |
| 11405 | |
| 11406 | The Lisp interpreter creates an individual instance of |
| 11407 | @code{triangle-recursively}. It does not matter that this function is |
| 11408 | contained within itself. Emacs passes the result Step 1 as the |
| 11409 | argument used by this instance of the @code{triangle-recursively} |
| 11410 | function |
| 11411 | |
| 11412 | In this case, Emacs evaluates @code{triangle-recursively} with an |
| 11413 | argument of 1. This means that this evaluation of |
| 11414 | @code{triangle-recursively} returns 1. |
| 11415 | |
| 11416 | @item Step 3 @w{ } Evaluate the value of @code{number}. |
| 11417 | |
| 11418 | The variable @code{number} is the second element of the list that |
| 11419 | starts with @code{+}; its value is 2. |
| 11420 | |
| 11421 | @item Step 4 @w{ } Evaluate the @code{+} expression. |
| 11422 | |
| 11423 | The @code{+} expression receives two arguments, the first |
| 11424 | from the evaluation of @code{number} (Step 3) and the second from the |
| 11425 | evaluation of @code{triangle-recursively} (Step 2). |
| 11426 | |
| 11427 | The result of the addition is the sum of 2 plus 1, and the number 3 is |
| 11428 | returned, which is correct. A triangle with two rows has three |
| 11429 | pebbles in it. |
| 11430 | @end table |
| 11431 | |
| 11432 | @node Recursive Example arg of 3 or 4 |
| 11433 | @unnumberedsubsubsec An argument of 3 or 4 |
| 11434 | |
| 11435 | Suppose that @code{triangle-recursively} is called with an argument of |
| 11436 | 3. |
| 11437 | |
| 11438 | @table @i |
| 11439 | @item Step 1 @w{ } Evaluate the do-again-test. |
| 11440 | |
| 11441 | The @code{if} expression is evaluated first. This is the do-again |
| 11442 | test and returns false, so the else-part of the @code{if} expression |
| 11443 | is evaluated. (Note that in this example, the do-again-test causes |
| 11444 | the function to call itself when it tests false, not when it tests |
| 11445 | true.) |
| 11446 | |
| 11447 | @item Step 2 @w{ } Evaluate the innermost expression of the else-part. |
| 11448 | |
| 11449 | The innermost expression of the else-part is evaluated, which decrements |
| 11450 | 3 to 2. This is the next-step-expression. |
| 11451 | |
| 11452 | @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function. |
| 11453 | |
| 11454 | The number 2 is passed to the @code{triangle-recursively} function. |
| 11455 | |
| 11456 | We already know what happens when Emacs evaluates @code{triangle-recursively} with |
| 11457 | an argument of 2. After going through the sequence of actions described |
| 11458 | earlier, it returns a value of 3. So that is what will happen here. |
| 11459 | |
| 11460 | @item Step 4 @w{ } Evaluate the addition. |
| 11461 | |
| 11462 | 3 will be passed as an argument to the addition and will be added to the |
| 11463 | number with which the function was called, which is 3. |
| 11464 | @end table |
| 11465 | |
| 11466 | @noindent |
| 11467 | The value returned by the function as a whole will be 6. |
| 11468 | |
| 11469 | Now that we know what will happen when @code{triangle-recursively} is |
| 11470 | called with an argument of 3, it is evident what will happen if it is |
| 11471 | called with an argument of 4: |
| 11472 | |
| 11473 | @quotation |
| 11474 | @need 800 |
| 11475 | In the recursive call, the evaluation of |
| 11476 | |
| 11477 | @smallexample |
| 11478 | (triangle-recursively (1- 4)) |
| 11479 | @end smallexample |
| 11480 | |
| 11481 | @need 800 |
| 11482 | @noindent |
| 11483 | will return the value of evaluating |
| 11484 | |
| 11485 | @smallexample |
| 11486 | (triangle-recursively 3) |
| 11487 | @end smallexample |
| 11488 | |
| 11489 | @noindent |
| 11490 | which is 6 and this value will be added to 4 by the addition in the |
| 11491 | third line. |
| 11492 | @end quotation |
| 11493 | |
| 11494 | @noindent |
| 11495 | The value returned by the function as a whole will be 10. |
| 11496 | |
| 11497 | Each time @code{triangle-recursively} is evaluated, it evaluates a |
| 11498 | version of itself---a different instance of itself---with a smaller |
| 11499 | argument, until the argument is small enough so that it does not |
| 11500 | evaluate itself. |
| 11501 | |
| 11502 | Note that this particular design for a recursive function |
| 11503 | requires that operations be deferred. |
| 11504 | |
| 11505 | Before @code{(triangle-recursively 7)} can calculate its answer, it |
| 11506 | must call @code{(triangle-recursively 6)}; and before |
| 11507 | @code{(triangle-recursively 6)} can calculate its answer, it must call |
| 11508 | @code{(triangle-recursively 5)}; and so on. That is to say, the |
| 11509 | calculation that @code{(triangle-recursively 7)} makes must be |
| 11510 | deferred until @code{(triangle-recursively 6)} makes its calculation; |
| 11511 | and @code{(triangle-recursively 6)} must defer until |
| 11512 | @code{(triangle-recursively 5)} completes; and so on. |
| 11513 | |
| 11514 | If each of these instances of @code{triangle-recursively} are thought |
| 11515 | of as different robots, the first robot must wait for the second to |
| 11516 | complete its job, which must wait until the third completes, and so |
| 11517 | on. |
| 11518 | |
| 11519 | There is a way around this kind of waiting, which we will discuss in |
| 11520 | @ref{No Deferment, , Recursion without Deferments}. |
| 11521 | |
| 11522 | @node Recursion with cond |
| 11523 | @subsection Recursion Example Using @code{cond} |
| 11524 | @findex cond |
| 11525 | |
| 11526 | The version of @code{triangle-recursively} described earlier is written |
| 11527 | with the @code{if} special form. It can also be written using another |
| 11528 | special form called @code{cond}. The name of the special form |
| 11529 | @code{cond} is an abbreviation of the word @samp{conditional}. |
| 11530 | |
| 11531 | Although the @code{cond} special form is not used as often in the |
| 11532 | Emacs Lisp sources as @code{if}, it is used often enough to justify |
| 11533 | explaining it. |
| 11534 | |
| 11535 | @need 800 |
| 11536 | The template for a @code{cond} expression looks like this: |
| 11537 | |
| 11538 | @smallexample |
| 11539 | @group |
| 11540 | (cond |
| 11541 | @var{body}@dots{}) |
| 11542 | @end group |
| 11543 | @end smallexample |
| 11544 | |
| 11545 | @noindent |
| 11546 | where the @var{body} is a series of lists. |
| 11547 | |
| 11548 | @need 800 |
| 11549 | Written out more fully, the template looks like this: |
| 11550 | |
| 11551 | @smallexample |
| 11552 | @group |
| 11553 | (cond |
| 11554 | (@var{first-true-or-false-test} @var{first-consequent}) |
| 11555 | (@var{second-true-or-false-test} @var{second-consequent}) |
| 11556 | (@var{third-true-or-false-test} @var{third-consequent}) |
| 11557 | @dots{}) |
| 11558 | @end group |
| 11559 | @end smallexample |
| 11560 | |
| 11561 | When the Lisp interpreter evaluates the @code{cond} expression, it |
| 11562 | evaluates the first element (the @sc{car} or true-or-false-test) of |
| 11563 | the first expression in a series of expressions within the body of the |
| 11564 | @code{cond}. |
| 11565 | |
| 11566 | If the true-or-false-test returns @code{nil} the rest of that |
| 11567 | expression, the consequent, is skipped and the true-or-false-test of the |
| 11568 | next expression is evaluated. When an expression is found whose |
| 11569 | true-or-false-test returns a value that is not @code{nil}, the |
| 11570 | consequent of that expression is evaluated. The consequent can be one |
| 11571 | or more expressions. If the consequent consists of more than one |
| 11572 | expression, the expressions are evaluated in sequence and the value of |
| 11573 | the last one is returned. If the expression does not have a consequent, |
| 11574 | the value of the true-or-false-test is returned. |
| 11575 | |
| 11576 | If none of the true-or-false-tests test true, the @code{cond} expression |
| 11577 | returns @code{nil}. |
| 11578 | |
| 11579 | @need 1250 |
| 11580 | Written using @code{cond}, the @code{triangle} function looks like this: |
| 11581 | |
| 11582 | @smallexample |
| 11583 | @group |
| 11584 | (defun triangle-using-cond (number) |
| 11585 | (cond ((<= number 0) 0) |
| 11586 | ((= number 1) 1) |
| 11587 | ((> number 1) |
| 11588 | (+ number (triangle-using-cond (1- number)))))) |
| 11589 | @end group |
| 11590 | @end smallexample |
| 11591 | |
| 11592 | @noindent |
| 11593 | In this example, the @code{cond} returns 0 if the number is less than or |
| 11594 | equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+ |
| 11595 | number (triangle-using-cond (1- number)))} if the number is greater than |
| 11596 | 1. |
| 11597 | |
| 11598 | @node Recursive Patterns |
| 11599 | @subsection Recursive Patterns |
| 11600 | @cindex Recursive Patterns |
| 11601 | |
| 11602 | Here are three common recursive patterns. Each involves a list. |
| 11603 | Recursion does not need to involve lists, but Lisp is designed for lists |
| 11604 | and this provides a sense of its primal capabilities. |
| 11605 | |
| 11606 | @menu |
| 11607 | * Every:: |
| 11608 | * Accumulate:: |
| 11609 | * Keep:: |
| 11610 | @end menu |
| 11611 | |
| 11612 | @node Every |
| 11613 | @unnumberedsubsubsec Recursive Pattern: @emph{every} |
| 11614 | @cindex Every, type of recursive pattern |
| 11615 | @cindex Recursive pattern: every |
| 11616 | |
| 11617 | In the @code{every} recursive pattern, an action is performed on every |
| 11618 | element of a list. |
| 11619 | |
| 11620 | @need 1500 |
| 11621 | The basic pattern is: |
| 11622 | |
| 11623 | @itemize @bullet |
| 11624 | @item |
| 11625 | If a list be empty, return @code{nil}. |
| 11626 | @item |
| 11627 | Else, act on the beginning of the list (the @sc{car} of the list) |
| 11628 | @itemize @minus |
| 11629 | @item |
| 11630 | through a recursive call by the function on the rest (the |
| 11631 | @sc{cdr}) of the list, |
| 11632 | @item |
| 11633 | and, optionally, combine the acted-on element, using @code{cons}, |
| 11634 | with the results of acting on the rest. |
| 11635 | @end itemize |
| 11636 | @end itemize |
| 11637 | |
| 11638 | @need 1500 |
| 11639 | Here is example: |
| 11640 | |
| 11641 | @smallexample |
| 11642 | @group |
| 11643 | (defun square-each (numbers-list) |
| 11644 | "Square each of a NUMBERS LIST, recursively." |
| 11645 | (if (not numbers-list) ; do-again-test |
| 11646 | nil |
| 11647 | (cons |
| 11648 | (* (car numbers-list) (car numbers-list)) |
| 11649 | (square-each (cdr numbers-list))))) ; next-step-expression |
| 11650 | @end group |
| 11651 | |
| 11652 | @group |
| 11653 | (square-each '(1 2 3)) |
| 11654 | @result{} (1 4 9) |
| 11655 | @end group |
| 11656 | @end smallexample |
| 11657 | |
| 11658 | @need 1200 |
| 11659 | @noindent |
| 11660 | If @code{numbers-list} is empty, do nothing. But if it has content, |
| 11661 | construct a list combining the square of the first number in the list |
| 11662 | with the result of the recursive call. |
| 11663 | |
| 11664 | (The example follows the pattern exactly: @code{nil} is returned if |
| 11665 | the numbers' list is empty. In practice, you would write the |
| 11666 | conditional so it carries out the action when the numbers' list is not |
| 11667 | empty.) |
| 11668 | |
| 11669 | The @code{print-elements-recursively} function (@pxref{Recursion with |
| 11670 | list, , Recursion with a List}) is another example of an @code{every} |
| 11671 | pattern, except in this case, rather than bring the results together |
| 11672 | using @code{cons}, we print each element of output. |
| 11673 | |
| 11674 | @need 1250 |
| 11675 | The @code{print-elements-recursively} function looks like this: |
| 11676 | |
| 11677 | @smallexample |
| 11678 | @group |
| 11679 | (setq animals '(gazelle giraffe lion tiger)) |
| 11680 | @end group |
| 11681 | |
| 11682 | @group |
| 11683 | (defun print-elements-recursively (list) |
| 11684 | "Print each element of LIST on a line of its own. |
| 11685 | Uses recursion." |
| 11686 | (when list ; @r{do-again-test} |
| 11687 | (print (car list)) ; @r{body} |
| 11688 | (print-elements-recursively ; @r{recursive call} |
| 11689 | (cdr list)))) ; @r{next-step-expression} |
| 11690 | |
| 11691 | (print-elements-recursively animals) |
| 11692 | @end group |
| 11693 | @end smallexample |
| 11694 | |
| 11695 | @need 1500 |
| 11696 | The pattern for @code{print-elements-recursively} is: |
| 11697 | |
| 11698 | @itemize @bullet |
| 11699 | @item |
| 11700 | When the list is empty, do nothing. |
| 11701 | @item |
| 11702 | But when the list has at least one element, |
| 11703 | @itemize @minus |
| 11704 | @item |
| 11705 | act on the beginning of the list (the @sc{car} of the list), |
| 11706 | @item |
| 11707 | and make a recursive call on the rest (the @sc{cdr}) of the list. |
| 11708 | @end itemize |
| 11709 | @end itemize |
| 11710 | |
| 11711 | @node Accumulate |
| 11712 | @unnumberedsubsubsec Recursive Pattern: @emph{accumulate} |
| 11713 | @cindex Accumulate, type of recursive pattern |
| 11714 | @cindex Recursive pattern: accumulate |
| 11715 | |
| 11716 | Another recursive pattern is called the @code{accumulate} pattern. In |
| 11717 | the @code{accumulate} recursive pattern, an action is performed on |
| 11718 | every element of a list and the result of that action is accumulated |
| 11719 | with the results of performing the action on the other elements. |
| 11720 | |
| 11721 | This is very like the `every' pattern using @code{cons}, except that |
| 11722 | @code{cons} is not used, but some other combiner. |
| 11723 | |
| 11724 | @need 1500 |
| 11725 | The pattern is: |
| 11726 | |
| 11727 | @itemize @bullet |
| 11728 | @item |
| 11729 | If a list be empty, return zero or some other constant. |
| 11730 | @item |
| 11731 | Else, act on the beginning of the list (the @sc{car} of the list), |
| 11732 | @itemize @minus |
| 11733 | @item |
| 11734 | and combine that acted-on element, using @code{+} or |
| 11735 | some other combining function, with |
| 11736 | @item |
| 11737 | a recursive call by the function on the rest (the @sc{cdr}) of the list. |
| 11738 | @end itemize |
| 11739 | @end itemize |
| 11740 | |
| 11741 | @need 1500 |
| 11742 | Here is an example: |
| 11743 | |
| 11744 | @smallexample |
| 11745 | @group |
| 11746 | (defun add-elements (numbers-list) |
| 11747 | "Add the elements of NUMBERS-LIST together." |
| 11748 | (if (not numbers-list) |
| 11749 | 0 |
| 11750 | (+ (car numbers-list) (add-elements (cdr numbers-list))))) |
| 11751 | @end group |
| 11752 | |
| 11753 | @group |
| 11754 | (add-elements '(1 2 3 4)) |
| 11755 | @result{} 10 |
| 11756 | @end group |
| 11757 | @end smallexample |
| 11758 | |
| 11759 | @xref{Files List, , Making a List of Files}, for an example of the |
| 11760 | accumulate pattern. |
| 11761 | |
| 11762 | @node Keep |
| 11763 | @unnumberedsubsubsec Recursive Pattern: @emph{keep} |
| 11764 | @cindex Keep, type of recursive pattern |
| 11765 | @cindex Recursive pattern: keep |
| 11766 | |
| 11767 | A third recursive pattern is called the @code{keep} pattern. |
| 11768 | In the @code{keep} recursive pattern, each element of a list is tested; |
| 11769 | the element is acted on and the results are kept only if the element |
| 11770 | meets a criterion. |
| 11771 | |
| 11772 | Again, this is very like the `every' pattern, except the element is |
| 11773 | skipped unless it meets a criterion. |
| 11774 | |
| 11775 | @need 1500 |
| 11776 | The pattern has three parts: |
| 11777 | |
| 11778 | @itemize @bullet |
| 11779 | @item |
| 11780 | If a list be empty, return @code{nil}. |
| 11781 | @item |
| 11782 | Else, if the beginning of the list (the @sc{car} of the list) passes |
| 11783 | a test |
| 11784 | @itemize @minus |
| 11785 | @item |
| 11786 | act on that element and combine it, using @code{cons} with |
| 11787 | @item |
| 11788 | a recursive call by the function on the rest (the @sc{cdr}) of the list. |
| 11789 | @end itemize |
| 11790 | @item |
| 11791 | Otherwise, if the beginning of the list (the @sc{car} of the list) fails |
| 11792 | the test |
| 11793 | @itemize @minus |
| 11794 | @item |
| 11795 | skip on that element, |
| 11796 | @item |
| 11797 | and, recursively call the function on the rest (the @sc{cdr}) of the list. |
| 11798 | @end itemize |
| 11799 | @end itemize |
| 11800 | |
| 11801 | @need 1500 |
| 11802 | Here is an example that uses @code{cond}: |
| 11803 | |
| 11804 | @smallexample |
| 11805 | @group |
| 11806 | (defun keep-three-letter-words (word-list) |
| 11807 | "Keep three letter words in WORD-LIST." |
| 11808 | (cond |
| 11809 | ;; First do-again-test: stop-condition |
| 11810 | ((not word-list) nil) |
| 11811 | |
| 11812 | ;; Second do-again-test: when to act |
| 11813 | ((eq 3 (length (symbol-name (car word-list)))) |
| 11814 | ;; combine acted-on element with recursive call on shorter list |
| 11815 | (cons (car word-list) (keep-three-letter-words (cdr word-list)))) |
| 11816 | |
| 11817 | ;; Third do-again-test: when to skip element; |
| 11818 | ;; recursively call shorter list with next-step expression |
| 11819 | (t (keep-three-letter-words (cdr word-list))))) |
| 11820 | @end group |
| 11821 | |
| 11822 | @group |
| 11823 | (keep-three-letter-words '(one two three four five six)) |
| 11824 | @result{} (one two six) |
| 11825 | @end group |
| 11826 | @end smallexample |
| 11827 | |
| 11828 | It goes without saying that you need not use @code{nil} as the test for |
| 11829 | when to stop; and you can, of course, combine these patterns. |
| 11830 | |
| 11831 | @node No Deferment |
| 11832 | @subsection Recursion without Deferments |
| 11833 | @cindex Deferment in recursion |
| 11834 | @cindex Recursion without Deferments |
| 11835 | |
| 11836 | Let's consider again what happens with the @code{triangle-recursively} |
| 11837 | function. We will find that the intermediate calculations are |
| 11838 | deferred until all can be done. |
| 11839 | |
| 11840 | @need 800 |
| 11841 | Here is the function definition: |
| 11842 | |
| 11843 | @smallexample |
| 11844 | @group |
| 11845 | (defun triangle-recursively (number) |
| 11846 | "Return the sum of the numbers 1 through NUMBER inclusive. |
| 11847 | Uses recursion." |
| 11848 | (if (= number 1) ; @r{do-again-test} |
| 11849 | 1 ; @r{then-part} |
| 11850 | (+ number ; @r{else-part} |
| 11851 | (triangle-recursively ; @r{recursive call} |
| 11852 | (1- number))))) ; @r{next-step-expression} |
| 11853 | @end group |
| 11854 | @end smallexample |
| 11855 | |
| 11856 | What happens when we call this function with a argument of 7? |
| 11857 | |
| 11858 | The first instance of the @code{triangle-recursively} function adds |
| 11859 | the number 7 to the value returned by a second instance of |
| 11860 | @code{triangle-recursively}, an instance that has been passed an |
| 11861 | argument of 6. That is to say, the first calculation is: |
| 11862 | |
| 11863 | @smallexample |
| 11864 | (+ 7 (triangle-recursively 6)) |
| 11865 | @end smallexample |
| 11866 | |
| 11867 | @noindent |
| 11868 | The first instance of @code{triangle-recursively}---you may want to |
| 11869 | think of it as a little robot---cannot complete its job. It must hand |
| 11870 | off the calculation for @code{(triangle-recursively 6)} to a second |
| 11871 | instance of the program, to a second robot. This second individual is |
| 11872 | completely different from the first one; it is, in the jargon, a |
| 11873 | `different instantiation'. Or, put another way, it is a different |
| 11874 | robot. It is the same model as the first; it calculates triangle |
| 11875 | numbers recursively; but it has a different serial number. |
| 11876 | |
| 11877 | And what does @code{(triangle-recursively 6)} return? It returns the |
| 11878 | number 6 added to the value returned by evaluating |
| 11879 | @code{triangle-recursively} with an argument of 5. Using the robot |
| 11880 | metaphor, it asks yet another robot to help it. |
| 11881 | |
| 11882 | @need 800 |
| 11883 | Now the total is: |
| 11884 | |
| 11885 | @smallexample |
| 11886 | (+ 7 6 (triangle-recursively 5)) |
| 11887 | @end smallexample |
| 11888 | |
| 11889 | @need 800 |
| 11890 | And what happens next? |
| 11891 | |
| 11892 | @smallexample |
| 11893 | (+ 7 6 5 (triangle-recursively 4)) |
| 11894 | @end smallexample |
| 11895 | |
| 11896 | Each time @code{triangle-recursively} is called, except for the last |
| 11897 | time, it creates another instance of the program---another robot---and |
| 11898 | asks it to make a calculation. |
| 11899 | |
| 11900 | @need 800 |
| 11901 | Eventually, the full addition is set up and performed: |
| 11902 | |
| 11903 | @smallexample |
| 11904 | (+ 7 6 5 4 3 2 1) |
| 11905 | @end smallexample |
| 11906 | |
| 11907 | This design for the function defers the calculation of the first step |
| 11908 | until the second can be done, and defers that until the third can be |
| 11909 | done, and so on. Each deferment means the computer must remember what |
| 11910 | is being waited on. This is not a problem when there are only a few |
| 11911 | steps, as in this example. But it can be a problem when there are |
| 11912 | more steps. |
| 11913 | |
| 11914 | @node No deferment solution |
| 11915 | @subsection No Deferment Solution |
| 11916 | @cindex No deferment solution |
| 11917 | @cindex Defermentless solution |
| 11918 | @cindex Solution without deferment |
| 11919 | |
| 11920 | The solution to the problem of deferred operations is to write in a |
| 11921 | manner that does not defer operations@footnote{The phrase @dfn{tail |
| 11922 | recursive} is used to describe such a process, one that uses |
| 11923 | `constant space'.}. This requires |
| 11924 | writing to a different pattern, often one that involves writing two |
| 11925 | function definitions, an `initialization' function and a `helper' |
| 11926 | function. |
| 11927 | |
| 11928 | The `initialization' function sets up the job; the `helper' function |
| 11929 | does the work. |
| 11930 | |
| 11931 | @need 1200 |
| 11932 | Here are the two function definitions for adding up numbers. They are |
| 11933 | so simple, I find them hard to understand. |
| 11934 | |
| 11935 | @smallexample |
| 11936 | @group |
| 11937 | (defun triangle-initialization (number) |
| 11938 | "Return the sum of the numbers 1 through NUMBER inclusive. |
| 11939 | This is the `initialization' component of a two function |
| 11940 | duo that uses recursion." |
| 11941 | (triangle-recursive-helper 0 0 number)) |
| 11942 | @end group |
| 11943 | @end smallexample |
| 11944 | |
| 11945 | @smallexample |
| 11946 | @group |
| 11947 | (defun triangle-recursive-helper (sum counter number) |
| 11948 | "Return SUM, using COUNTER, through NUMBER inclusive. |
| 11949 | This is the `helper' component of a two function duo |
| 11950 | that uses recursion." |
| 11951 | (if (> counter number) |
| 11952 | sum |
| 11953 | (triangle-recursive-helper (+ sum counter) ; @r{sum} |
| 11954 | (1+ counter) ; @r{counter} |
| 11955 | number))) ; @r{number} |
| 11956 | @end group |
| 11957 | @end smallexample |
| 11958 | |
| 11959 | @need 1250 |
| 11960 | Install both function definitions by evaluating them, then call |
| 11961 | @code{triangle-initialization} with 2 rows: |
| 11962 | |
| 11963 | @smallexample |
| 11964 | @group |
| 11965 | (triangle-initialization 2) |
| 11966 | @result{} 3 |
| 11967 | @end group |
| 11968 | @end smallexample |
| 11969 | |
| 11970 | The `initialization' function calls the first instance of the `helper' |
| 11971 | function with three arguments: zero, zero, and a number which is the |
| 11972 | number of rows in the triangle. |
| 11973 | |
| 11974 | The first two arguments passed to the `helper' function are |
| 11975 | initialization values. These values are changed when |
| 11976 | @code{triangle-recursive-helper} invokes new instances.@footnote{The |
| 11977 | jargon is mildly confusing: @code{triangle-recursive-helper} uses a |
| 11978 | process that is iterative in a procedure that is recursive. The |
| 11979 | process is called iterative because the computer need only record the |
| 11980 | three values, @code{sum}, @code{counter}, and @code{number}; the |
| 11981 | procedure is recursive because the function `calls itself'. On the |
| 11982 | other hand, both the process and the procedure used by |
| 11983 | @code{triangle-recursively} are called recursive. The word |
| 11984 | `recursive' has different meanings in the two contexts.} |
| 11985 | |
| 11986 | Let's see what happens when we have a triangle that has one row. (This |
| 11987 | triangle will have one pebble in it!) |
| 11988 | |
| 11989 | @need 1200 |
| 11990 | @code{triangle-initialization} will call its helper with |
| 11991 | the arguments @w{@code{0 0 1}}. That function will run the conditional |
| 11992 | test whether @code{(> counter number)}: |
| 11993 | |
| 11994 | @smallexample |
| 11995 | (> 0 1) |
| 11996 | @end smallexample |
| 11997 | |
| 11998 | @need 1200 |
| 11999 | @noindent |
| 12000 | and find that the result is false, so it will invoke |
| 12001 | the else-part of the @code{if} clause: |
| 12002 | |
| 12003 | @smallexample |
| 12004 | @group |
| 12005 | (triangle-recursive-helper |
| 12006 | (+ sum counter) ; @r{sum plus counter} @result{} @r{sum} |
| 12007 | (1+ counter) ; @r{increment counter} @result{} @r{counter} |
| 12008 | number) ; @r{number stays the same} |
| 12009 | @end group |
| 12010 | @end smallexample |
| 12011 | |
| 12012 | @need 800 |
| 12013 | @noindent |
| 12014 | which will first compute: |
| 12015 | |
| 12016 | @smallexample |
| 12017 | @group |
| 12018 | (triangle-recursive-helper (+ 0 0) ; @r{sum} |
| 12019 | (1+ 0) ; @r{counter} |
| 12020 | 1) ; @r{number} |
| 12021 | @exdent which is: |
| 12022 | |
| 12023 | (triangle-recursive-helper 0 1 1) |
| 12024 | @end group |
| 12025 | @end smallexample |
| 12026 | |
| 12027 | Again, @code{(> counter number)} will be false, so again, the Lisp |
| 12028 | interpreter will evaluate @code{triangle-recursive-helper}, creating a |
| 12029 | new instance with new arguments. |
| 12030 | |
| 12031 | @need 800 |
| 12032 | This new instance will be; |
| 12033 | |
| 12034 | @smallexample |
| 12035 | @group |
| 12036 | (triangle-recursive-helper |
| 12037 | (+ sum counter) ; @r{sum plus counter} @result{} @r{sum} |
| 12038 | (1+ counter) ; @r{increment counter} @result{} @r{counter} |
| 12039 | number) ; @r{number stays the same} |
| 12040 | |
| 12041 | @exdent which is: |
| 12042 | |
| 12043 | (triangle-recursive-helper 1 2 1) |
| 12044 | @end group |
| 12045 | @end smallexample |
| 12046 | |
| 12047 | In this case, the @code{(> counter number)} test will be true! So the |
| 12048 | instance will return the value of the sum, which will be 1, as |
| 12049 | expected. |
| 12050 | |
| 12051 | Now, let's pass @code{triangle-initialization} an argument |
| 12052 | of 2, to find out how many pebbles there are in a triangle with two rows. |
| 12053 | |
| 12054 | That function calls @code{(triangle-recursive-helper 0 0 2)}. |
| 12055 | |
| 12056 | @need 800 |
| 12057 | In stages, the instances called will be: |
| 12058 | |
| 12059 | @smallexample |
| 12060 | @group |
| 12061 | @r{sum counter number} |
| 12062 | (triangle-recursive-helper 0 1 2) |
| 12063 | |
| 12064 | (triangle-recursive-helper 1 2 2) |
| 12065 | |
| 12066 | (triangle-recursive-helper 3 3 2) |
| 12067 | @end group |
| 12068 | @end smallexample |
| 12069 | |
| 12070 | When the last instance is called, the @code{(> counter number)} test |
| 12071 | will be true, so the instance will return the value of @code{sum}, |
| 12072 | which will be 3. |
| 12073 | |
| 12074 | This kind of pattern helps when you are writing functions that can use |
| 12075 | many resources in a computer. |
| 12076 | |
| 12077 | @need 1500 |
| 12078 | @node Looping exercise |
| 12079 | @section Looping Exercise |
| 12080 | |
| 12081 | @itemize @bullet |
| 12082 | @item |
| 12083 | Write a function similar to @code{triangle} in which each row has a |
| 12084 | value which is the square of the row number. Use a @code{while} loop. |
| 12085 | |
| 12086 | @item |
| 12087 | Write a function similar to @code{triangle} that multiplies instead of |
| 12088 | adds the values. |
| 12089 | |
| 12090 | @item |
| 12091 | Rewrite these two functions recursively. Rewrite these functions |
| 12092 | using @code{cond}. |
| 12093 | |
| 12094 | @c comma in printed title causes problem in Info cross reference |
| 12095 | @item |
| 12096 | Write a function for Texinfo mode that creates an index entry at the |
| 12097 | beginning of a paragraph for every @samp{@@dfn} within the paragraph. |
| 12098 | (In a Texinfo file, @samp{@@dfn} marks a definition. This book is |
| 12099 | written in Texinfo.) |
| 12100 | |
| 12101 | Many of the functions you will need are described in two of the |
| 12102 | previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing |
| 12103 | Text}, and @ref{Yanking, , Yanking Text Back}. If you use |
| 12104 | @code{forward-paragraph} to put the index entry at the beginning of |
| 12105 | the paragraph, you will have to use @w{@kbd{C-h f}} |
| 12106 | (@code{describe-function}) to find out how to make the command go |
| 12107 | backwards. |
| 12108 | |
| 12109 | For more information, see |
| 12110 | @ifinfo |
| 12111 | @ref{Indicating, , Indicating Definitions, texinfo}. |
| 12112 | @end ifinfo |
| 12113 | @ifhtml |
| 12114 | @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to |
| 12115 | a Texinfo manual in the current directory. Or, if you are on the |
| 12116 | Internet, see |
| 12117 | @uref{http://www.gnu.org/software/texinfo/manual/texinfo/} |
| 12118 | @end ifhtml |
| 12119 | @iftex |
| 12120 | ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU |
| 12121 | Documentation Format}. |
| 12122 | @end iftex |
| 12123 | @end itemize |
| 12124 | |
| 12125 | @node Regexp Search |
| 12126 | @chapter Regular Expression Searches |
| 12127 | @cindex Searches, illustrating |
| 12128 | @cindex Regular expression searches |
| 12129 | @cindex Patterns, searching for |
| 12130 | @cindex Motion by sentence and paragraph |
| 12131 | @cindex Sentences, movement by |
| 12132 | @cindex Paragraphs, movement by |
| 12133 | |
| 12134 | Regular expression searches are used extensively in GNU Emacs. The |
| 12135 | two functions, @code{forward-sentence} and @code{forward-paragraph}, |
| 12136 | illustrate these searches well. They use regular expressions to find |
| 12137 | where to move point. The phrase `regular expression' is often written |
| 12138 | as `regexp'. |
| 12139 | |
| 12140 | Regular expression searches are described in @ref{Regexp Search, , |
| 12141 | Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in |
| 12142 | @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference |
| 12143 | Manual}. In writing this chapter, I am presuming that you have at |
| 12144 | least a mild acquaintance with them. The major point to remember is |
| 12145 | that regular expressions permit you to search for patterns as well as |
| 12146 | for literal strings of characters. For example, the code in |
| 12147 | @code{forward-sentence} searches for the pattern of possible |
| 12148 | characters that could mark the end of a sentence, and moves point to |
| 12149 | that spot. |
| 12150 | |
| 12151 | Before looking at the code for the @code{forward-sentence} function, it |
| 12152 | is worth considering what the pattern that marks the end of a sentence |
| 12153 | must be. The pattern is discussed in the next section; following that |
| 12154 | is a description of the regular expression search function, |
| 12155 | @code{re-search-forward}. The @code{forward-sentence} function |
| 12156 | is described in the section following. Finally, the |
| 12157 | @code{forward-paragraph} function is described in the last section of |
| 12158 | this chapter. @code{forward-paragraph} is a complex function that |
| 12159 | introduces several new features. |
| 12160 | |
| 12161 | @menu |
| 12162 | * sentence-end:: The regular expression for @code{sentence-end}. |
| 12163 | * re-search-forward:: Very similar to @code{search-forward}. |
| 12164 | * forward-sentence:: A straightforward example of regexp search. |
| 12165 | * forward-paragraph:: A somewhat complex example. |
| 12166 | * etags:: How to create your own @file{TAGS} table. |
| 12167 | * Regexp Review:: |
| 12168 | * re-search Exercises:: |
| 12169 | @end menu |
| 12170 | |
| 12171 | @node sentence-end |
| 12172 | @section The Regular Expression for @code{sentence-end} |
| 12173 | @findex sentence-end |
| 12174 | |
| 12175 | The symbol @code{sentence-end} is bound to the pattern that marks the |
| 12176 | end of a sentence. What should this regular expression be? |
| 12177 | |
| 12178 | Clearly, a sentence may be ended by a period, a question mark, or an |
| 12179 | exclamation mark. Indeed, in English, only clauses that end with one |
| 12180 | of those three characters should be considered the end of a sentence. |
| 12181 | This means that the pattern should include the character set: |
| 12182 | |
| 12183 | @smallexample |
| 12184 | [.?!] |
| 12185 | @end smallexample |
| 12186 | |
| 12187 | However, we do not want @code{forward-sentence} merely to jump to a |
| 12188 | period, a question mark, or an exclamation mark, because such a character |
| 12189 | might be used in the middle of a sentence. A period, for example, is |
| 12190 | used after abbreviations. So other information is needed. |
| 12191 | |
| 12192 | According to convention, you type two spaces after every sentence, but |
| 12193 | only one space after a period, a question mark, or an exclamation mark in |
| 12194 | the body of a sentence. So a period, a question mark, or an exclamation |
| 12195 | mark followed by two spaces is a good indicator of an end of sentence. |
| 12196 | However, in a file, the two spaces may instead be a tab or the end of a |
| 12197 | line. This means that the regular expression should include these three |
| 12198 | items as alternatives. |
| 12199 | |
| 12200 | @need 800 |
| 12201 | This group of alternatives will look like this: |
| 12202 | |
| 12203 | @smallexample |
| 12204 | @group |
| 12205 | \\($\\| \\| \\) |
| 12206 | ^ ^^ |
| 12207 | TAB SPC |
| 12208 | @end group |
| 12209 | @end smallexample |
| 12210 | |
| 12211 | @noindent |
| 12212 | Here, @samp{$} indicates the end of the line, and I have pointed out |
| 12213 | where the tab and two spaces are inserted in the expression. Both are |
| 12214 | inserted by putting the actual characters into the expression. |
| 12215 | |
| 12216 | Two backslashes, @samp{\\}, are required before the parentheses and |
| 12217 | vertical bars: the first backslash quotes the following backslash in |
| 12218 | Emacs; and the second indicates that the following character, the |
| 12219 | parenthesis or the vertical bar, is special. |
| 12220 | |
| 12221 | @need 1000 |
| 12222 | Also, a sentence may be followed by one or more carriage returns, like |
| 12223 | this: |
| 12224 | |
| 12225 | @smallexample |
| 12226 | @group |
| 12227 | [ |
| 12228 | ]* |
| 12229 | @end group |
| 12230 | @end smallexample |
| 12231 | |
| 12232 | @noindent |
| 12233 | Like tabs and spaces, a carriage return is inserted into a regular |
| 12234 | expression by inserting it literally. The asterisk indicates that the |
| 12235 | @key{RET} is repeated zero or more times. |
| 12236 | |
| 12237 | But a sentence end does not consist only of a period, a question mark or |
| 12238 | an exclamation mark followed by appropriate space: a closing quotation |
| 12239 | mark or a closing brace of some kind may precede the space. Indeed more |
| 12240 | than one such mark or brace may precede the space. These require a |
| 12241 | expression that looks like this: |
| 12242 | |
| 12243 | @smallexample |
| 12244 | []\"')@}]* |
| 12245 | @end smallexample |
| 12246 | |
| 12247 | In this expression, the first @samp{]} is the first character in the |
| 12248 | expression; the second character is @samp{"}, which is preceded by a |
| 12249 | @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last |
| 12250 | three characters are @samp{'}, @samp{)}, and @samp{@}}. |
| 12251 | |
| 12252 | All this suggests what the regular expression pattern for matching the |
| 12253 | end of a sentence should be; and, indeed, if we evaluate |
| 12254 | @code{sentence-end} we find that it returns the following value: |
| 12255 | |
| 12256 | @smallexample |
| 12257 | @group |
| 12258 | sentence-end |
| 12259 | @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[ |
| 12260 | ]*" |
| 12261 | @end group |
| 12262 | @end smallexample |
| 12263 | |
| 12264 | @noindent |
| 12265 | (Well, not in GNU Emacs 22; that is because of an effort to make the |
| 12266 | process simpler and to handle more glyphs and languages. When the |
| 12267 | value of @code{sentence-end} is @code{nil}, then use the value defined |
| 12268 | by the function @code{sentence-end}. (Here is a use of the difference |
| 12269 | between a value and a function in Emacs Lisp.) The function returns a |
| 12270 | value constructed from the variables @code{sentence-end-base}, |
| 12271 | @code{sentence-end-double-space}, @code{sentence-end-without-period}, |
| 12272 | and @code{sentence-end-without-space}. The critical variable is |
| 12273 | @code{sentence-end-base}; its global value is similar to the one |
| 12274 | described above but it also contains two additional quotation marks. |
| 12275 | These have differing degrees of curliness. The |
| 12276 | @code{sentence-end-without-period} variable, when true, tells Emacs |
| 12277 | that a sentence may end without a period, such as text in Thai.) |
| 12278 | |
| 12279 | @ignore |
| 12280 | @noindent |
| 12281 | (Note that here the @key{TAB}, two spaces, and @key{RET} are shown |
| 12282 | literally in the pattern.) |
| 12283 | |
| 12284 | This regular expression can be deciphered as follows: |
| 12285 | |
| 12286 | @table @code |
| 12287 | @item [.?!] |
| 12288 | The first part of the pattern is the three characters, a period, a question |
| 12289 | mark and an exclamation mark, within square brackets. The pattern must |
| 12290 | begin with one or other of these characters. |
| 12291 | |
| 12292 | @item []\"')@}]* |
| 12293 | The second part of the pattern is the group of closing braces and |
| 12294 | quotation marks, which can appear zero or more times. These may follow |
| 12295 | the period, question mark or exclamation mark. In a regular expression, |
| 12296 | the backslash, @samp{\}, followed by the double quotation mark, |
| 12297 | @samp{"}, indicates the class of string-quote characters. Usually, the |
| 12298 | double quotation mark is the only character in this class. The |
| 12299 | asterisk, @samp{*}, indicates that the items in the previous group (the |
| 12300 | group surrounded by square brackets, @samp{[]}) may be repeated zero or |
| 12301 | more times. |
| 12302 | |
| 12303 | @item \\($\\| \\| \\) |
| 12304 | The third part of the pattern is one or other of: either the end of a |
| 12305 | line, or two blank spaces, or a tab. The double back-slashes are used |
| 12306 | to prevent Emacs from reading the parentheses and vertical bars as part |
| 12307 | of the search pattern; the parentheses are used to mark the group and |
| 12308 | the vertical bars are used to indicated that the patterns to either side |
| 12309 | of them are alternatives. The dollar sign is used to indicate the end |
| 12310 | of a line and both the two spaces and the tab are each inserted as is to |
| 12311 | indicate what they are. |
| 12312 | |
| 12313 | @item [@key{RET}]* |
| 12314 | Finally, the last part of the pattern indicates that the end of the line |
| 12315 | or the whitespace following the period, question mark or exclamation |
| 12316 | mark may, but need not, be followed by one or more carriage returns. In |
| 12317 | the pattern, the carriage return is inserted as an actual carriage |
| 12318 | return between square brackets but here it is shown as @key{RET}. |
| 12319 | @end table |
| 12320 | @end ignore |
| 12321 | |
| 12322 | @node re-search-forward |
| 12323 | @section The @code{re-search-forward} Function |
| 12324 | @findex re-search-forward |
| 12325 | |
| 12326 | The @code{re-search-forward} function is very like the |
| 12327 | @code{search-forward} function. (@xref{search-forward, , The |
| 12328 | @code{search-forward} Function}.) |
| 12329 | |
| 12330 | @code{re-search-forward} searches for a regular expression. If the |
| 12331 | search is successful, it leaves point immediately after the last |
| 12332 | character in the target. If the search is backwards, it leaves point |
| 12333 | just before the first character in the target. You may tell |
| 12334 | @code{re-search-forward} to return @code{t} for true. (Moving point |
| 12335 | is therefore a `side effect'.) |
| 12336 | |
| 12337 | Like @code{search-forward}, the @code{re-search-forward} function takes |
| 12338 | four arguments: |
| 12339 | |
| 12340 | @enumerate |
| 12341 | @item |
| 12342 | The first argument is the regular expression that the function searches |
| 12343 | for. The regular expression will be a string between quotation marks. |
| 12344 | |
| 12345 | @item |
| 12346 | The optional second argument limits how far the function will search; it is a |
| 12347 | bound, which is specified as a position in the buffer. |
| 12348 | |
| 12349 | @item |
| 12350 | The optional third argument specifies how the function responds to |
| 12351 | failure: @code{nil} as the third argument causes the function to |
| 12352 | signal an error (and print a message) when the search fails; any other |
| 12353 | value causes it to return @code{nil} if the search fails and @code{t} |
| 12354 | if the search succeeds. |
| 12355 | |
| 12356 | @item |
| 12357 | The optional fourth argument is the repeat count. A negative repeat |
| 12358 | count causes @code{re-search-forward} to search backwards. |
| 12359 | @end enumerate |
| 12360 | |
| 12361 | @need 800 |
| 12362 | The template for @code{re-search-forward} looks like this: |
| 12363 | |
| 12364 | @smallexample |
| 12365 | @group |
| 12366 | (re-search-forward "@var{regular-expression}" |
| 12367 | @var{limit-of-search} |
| 12368 | @var{what-to-do-if-search-fails} |
| 12369 | @var{repeat-count}) |
| 12370 | @end group |
| 12371 | @end smallexample |
| 12372 | |
| 12373 | The second, third, and fourth arguments are optional. However, if you |
| 12374 | want to pass a value to either or both of the last two arguments, you |
| 12375 | must also pass a value to all the preceding arguments. Otherwise, the |
| 12376 | Lisp interpreter will mistake which argument you are passing the value |
| 12377 | to. |
| 12378 | |
| 12379 | @need 1200 |
| 12380 | In the @code{forward-sentence} function, the regular expression will be |
| 12381 | the value of the variable @code{sentence-end}. In simple form, that is: |
| 12382 | |
| 12383 | @smallexample |
| 12384 | @group |
| 12385 | "[.?!][]\"')@}]*\\($\\| \\| \\)[ |
| 12386 | ]*" |
| 12387 | @end group |
| 12388 | @end smallexample |
| 12389 | |
| 12390 | @noindent |
| 12391 | The limit of the search will be the end of the paragraph (since a |
| 12392 | sentence cannot go beyond a paragraph). If the search fails, the |
| 12393 | function will return @code{nil}; and the repeat count will be provided |
| 12394 | by the argument to the @code{forward-sentence} function. |
| 12395 | |
| 12396 | @node forward-sentence |
| 12397 | @section @code{forward-sentence} |
| 12398 | @findex forward-sentence |
| 12399 | |
| 12400 | The command to move the cursor forward a sentence is a straightforward |
| 12401 | illustration of how to use regular expression searches in Emacs Lisp. |
| 12402 | Indeed, the function looks longer and more complicated than it is; this |
| 12403 | is because the function is designed to go backwards as well as forwards; |
| 12404 | and, optionally, over more than one sentence. The function is usually |
| 12405 | bound to the key command @kbd{M-e}. |
| 12406 | |
| 12407 | @menu |
| 12408 | * Complete forward-sentence:: |
| 12409 | * fwd-sentence while loops:: Two @code{while} loops. |
| 12410 | * fwd-sentence re-search:: A regular expression search. |
| 12411 | @end menu |
| 12412 | |
| 12413 | @ifnottex |
| 12414 | @node Complete forward-sentence |
| 12415 | @unnumberedsubsec Complete @code{forward-sentence} function definition |
| 12416 | @end ifnottex |
| 12417 | |
| 12418 | @need 1250 |
| 12419 | Here is the code for @code{forward-sentence}: |
| 12420 | |
| 12421 | @c in GNU Emacs 22 |
| 12422 | @smallexample |
| 12423 | @group |
| 12424 | (defun forward-sentence (&optional arg) |
| 12425 | "Move forward to next `sentence-end'. With argument, repeat. |
| 12426 | With negative argument, move backward repeatedly to `sentence-beginning'. |
| 12427 | |
| 12428 | The variable `sentence-end' is a regular expression that matches ends of |
| 12429 | sentences. Also, every paragraph boundary terminates sentences as well." |
| 12430 | @end group |
| 12431 | @group |
| 12432 | (interactive "p") |
| 12433 | (or arg (setq arg 1)) |
| 12434 | (let ((opoint (point)) |
| 12435 | (sentence-end (sentence-end))) |
| 12436 | (while (< arg 0) |
| 12437 | (let ((pos (point)) |
| 12438 | (par-beg (save-excursion (start-of-paragraph-text) (point)))) |
| 12439 | (if (and (re-search-backward sentence-end par-beg t) |
| 12440 | (or (< (match-end 0) pos) |
| 12441 | (re-search-backward sentence-end par-beg t))) |
| 12442 | (goto-char (match-end 0)) |
| 12443 | (goto-char par-beg))) |
| 12444 | (setq arg (1+ arg))) |
| 12445 | @end group |
| 12446 | @group |
| 12447 | (while (> arg 0) |
| 12448 | (let ((par-end (save-excursion (end-of-paragraph-text) (point)))) |
| 12449 | (if (re-search-forward sentence-end par-end t) |
| 12450 | (skip-chars-backward " \t\n") |
| 12451 | (goto-char par-end))) |
| 12452 | (setq arg (1- arg))) |
| 12453 | (constrain-to-field nil opoint t))) |
| 12454 | @end group |
| 12455 | @end smallexample |
| 12456 | |
| 12457 | @ignore |
| 12458 | GNU Emacs 21 |
| 12459 | @smallexample |
| 12460 | @group |
| 12461 | (defun forward-sentence (&optional arg) |
| 12462 | "Move forward to next sentence-end. With argument, repeat. |
| 12463 | With negative argument, move backward repeatedly to sentence-beginning. |
| 12464 | Sentence ends are identified by the value of sentence-end |
| 12465 | treated as a regular expression. Also, every paragraph boundary |
| 12466 | terminates sentences as well." |
| 12467 | @end group |
| 12468 | @group |
| 12469 | (interactive "p") |
| 12470 | (or arg (setq arg 1)) |
| 12471 | (while (< arg 0) |
| 12472 | (let ((par-beg |
| 12473 | (save-excursion (start-of-paragraph-text) (point)))) |
| 12474 | (if (re-search-backward |
| 12475 | (concat sentence-end "[^ \t\n]") par-beg t) |
| 12476 | (goto-char (1- (match-end 0))) |
| 12477 | (goto-char par-beg))) |
| 12478 | (setq arg (1+ arg))) |
| 12479 | (while (> arg 0) |
| 12480 | (let ((par-end |
| 12481 | (save-excursion (end-of-paragraph-text) (point)))) |
| 12482 | (if (re-search-forward sentence-end par-end t) |
| 12483 | (skip-chars-backward " \t\n") |
| 12484 | (goto-char par-end))) |
| 12485 | (setq arg (1- arg)))) |
| 12486 | @end group |
| 12487 | @end smallexample |
| 12488 | @end ignore |
| 12489 | |
| 12490 | The function looks long at first sight and it is best to look at its |
| 12491 | skeleton first, and then its muscle. The way to see the skeleton is to |
| 12492 | look at the expressions that start in the left-most columns: |
| 12493 | |
| 12494 | @smallexample |
| 12495 | @group |
| 12496 | (defun forward-sentence (&optional arg) |
| 12497 | "@var{documentation}@dots{}" |
| 12498 | (interactive "p") |
| 12499 | (or arg (setq arg 1)) |
| 12500 | (let ((opoint (point)) (sentence-end (sentence-end))) |
| 12501 | (while (< arg 0) |
| 12502 | (let ((pos (point)) |
| 12503 | (par-beg (save-excursion (start-of-paragraph-text) (point)))) |
| 12504 | @var{rest-of-body-of-while-loop-when-going-backwards} |
| 12505 | (while (> arg 0) |
| 12506 | (let ((par-end (save-excursion (end-of-paragraph-text) (point)))) |
| 12507 | @var{rest-of-body-of-while-loop-when-going-forwards} |
| 12508 | @var{handle-forms-and-equivalent} |
| 12509 | @end group |
| 12510 | @end smallexample |
| 12511 | |
| 12512 | This looks much simpler! The function definition consists of |
| 12513 | documentation, an @code{interactive} expression, an @code{or} |
| 12514 | expression, a @code{let} expression, and @code{while} loops. |
| 12515 | |
| 12516 | Let's look at each of these parts in turn. |
| 12517 | |
| 12518 | We note that the documentation is thorough and understandable. |
| 12519 | |
| 12520 | The function has an @code{interactive "p"} declaration. This means |
| 12521 | that the processed prefix argument, if any, is passed to the |
| 12522 | function as its argument. (This will be a number.) If the function |
| 12523 | is not passed an argument (it is optional) then the argument |
| 12524 | @code{arg} will be bound to 1. |
| 12525 | |
| 12526 | When @code{forward-sentence} is called non-interactively without an |
| 12527 | argument, @code{arg} is bound to @code{nil}. The @code{or} expression |
| 12528 | handles this. What it does is either leave the value of @code{arg} as |
| 12529 | it is, but only if @code{arg} is bound to a value; or it sets the |
| 12530 | value of @code{arg} to 1, in the case when @code{arg} is bound to |
| 12531 | @code{nil}. |
| 12532 | |
| 12533 | Next is a @code{let}. That specifies the values of two local |
| 12534 | variables, @code{point} and @code{sentence-end}. The local value of |
| 12535 | point, from before the search, is used in the |
| 12536 | @code{constrain-to-field} function which handles forms and |
| 12537 | equivalents. The @code{sentence-end} variable is set by the |
| 12538 | @code{sentence-end} function. |
| 12539 | |
| 12540 | @node fwd-sentence while loops |
| 12541 | @unnumberedsubsec The @code{while} loops |
| 12542 | |
| 12543 | Two @code{while} loops follow. The first @code{while} has a |
| 12544 | true-or-false-test that tests true if the prefix argument for |
| 12545 | @code{forward-sentence} is a negative number. This is for going |
| 12546 | backwards. The body of this loop is similar to the body of the second |
| 12547 | @code{while} clause, but it is not exactly the same. We will skip |
| 12548 | this @code{while} loop and concentrate on the second @code{while} |
| 12549 | loop. |
| 12550 | |
| 12551 | @need 1500 |
| 12552 | The second @code{while} loop is for moving point forward. Its skeleton |
| 12553 | looks like this: |
| 12554 | |
| 12555 | @smallexample |
| 12556 | @group |
| 12557 | (while (> arg 0) ; @r{true-or-false-test} |
| 12558 | (let @var{varlist} |
| 12559 | (if (@var{true-or-false-test}) |
| 12560 | @var{then-part} |
| 12561 | @var{else-part} |
| 12562 | (setq arg (1- arg)))) ; @code{while} @r{loop decrementer} |
| 12563 | @end group |
| 12564 | @end smallexample |
| 12565 | |
| 12566 | The @code{while} loop is of the decrementing kind. |
| 12567 | (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It |
| 12568 | has a true-or-false-test that tests true so long as the counter (in |
| 12569 | this case, the variable @code{arg}) is greater than zero; and it has a |
| 12570 | decrementer that subtracts 1 from the value of the counter every time |
| 12571 | the loop repeats. |
| 12572 | |
| 12573 | If no prefix argument is given to @code{forward-sentence}, which is |
| 12574 | the most common way the command is used, this @code{while} loop will |
| 12575 | run once, since the value of @code{arg} will be 1. |
| 12576 | |
| 12577 | The body of the @code{while} loop consists of a @code{let} expression, |
| 12578 | which creates and binds a local variable, and has, as its body, an |
| 12579 | @code{if} expression. |
| 12580 | |
| 12581 | @need 1250 |
| 12582 | The body of the @code{while} loop looks like this: |
| 12583 | |
| 12584 | @smallexample |
| 12585 | @group |
| 12586 | (let ((par-end |
| 12587 | (save-excursion (end-of-paragraph-text) (point)))) |
| 12588 | (if (re-search-forward sentence-end par-end t) |
| 12589 | (skip-chars-backward " \t\n") |
| 12590 | (goto-char par-end))) |
| 12591 | @end group |
| 12592 | @end smallexample |
| 12593 | |
| 12594 | The @code{let} expression creates and binds the local variable |
| 12595 | @code{par-end}. As we shall see, this local variable is designed to |
| 12596 | provide a bound or limit to the regular expression search. If the |
| 12597 | search fails to find a proper sentence ending in the paragraph, it will |
| 12598 | stop on reaching the end of the paragraph. |
| 12599 | |
| 12600 | But first, let us examine how @code{par-end} is bound to the value of |
| 12601 | the end of the paragraph. What happens is that the @code{let} sets the |
| 12602 | value of @code{par-end} to the value returned when the Lisp interpreter |
| 12603 | evaluates the expression |
| 12604 | |
| 12605 | @smallexample |
| 12606 | @group |
| 12607 | (save-excursion (end-of-paragraph-text) (point)) |
| 12608 | @end group |
| 12609 | @end smallexample |
| 12610 | |
| 12611 | @noindent |
| 12612 | In this expression, @code{(end-of-paragraph-text)} moves point to the |
| 12613 | end of the paragraph, @code{(point)} returns the value of point, and then |
| 12614 | @code{save-excursion} restores point to its original position. Thus, |
| 12615 | the @code{let} binds @code{par-end} to the value returned by the |
| 12616 | @code{save-excursion} expression, which is the position of the end of |
| 12617 | the paragraph. (The @code{end-of-paragraph-text} function uses |
| 12618 | @code{forward-paragraph}, which we will discuss shortly.) |
| 12619 | |
| 12620 | @need 1200 |
| 12621 | Emacs next evaluates the body of the @code{let}, which is an @code{if} |
| 12622 | expression that looks like this: |
| 12623 | |
| 12624 | @smallexample |
| 12625 | @group |
| 12626 | (if (re-search-forward sentence-end par-end t) ; @r{if-part} |
| 12627 | (skip-chars-backward " \t\n") ; @r{then-part} |
| 12628 | (goto-char par-end))) ; @r{else-part} |
| 12629 | @end group |
| 12630 | @end smallexample |
| 12631 | |
| 12632 | The @code{if} tests whether its first argument is true and if so, |
| 12633 | evaluates its then-part; otherwise, the Emacs Lisp interpreter |
| 12634 | evaluates the else-part. The true-or-false-test of the @code{if} |
| 12635 | expression is the regular expression search. |
| 12636 | |
| 12637 | It may seem odd to have what looks like the `real work' of |
| 12638 | the @code{forward-sentence} function buried here, but this is a common |
| 12639 | way this kind of operation is carried out in Lisp. |
| 12640 | |
| 12641 | @node fwd-sentence re-search |
| 12642 | @unnumberedsubsec The regular expression search |
| 12643 | |
| 12644 | The @code{re-search-forward} function searches for the end of the |
| 12645 | sentence, that is, for the pattern defined by the @code{sentence-end} |
| 12646 | regular expression. If the pattern is found---if the end of the sentence is |
| 12647 | found---then the @code{re-search-forward} function does two things: |
| 12648 | |
| 12649 | @enumerate |
| 12650 | @item |
| 12651 | The @code{re-search-forward} function carries out a side effect, which |
| 12652 | is to move point to the end of the occurrence found. |
| 12653 | |
| 12654 | @item |
| 12655 | The @code{re-search-forward} function returns a value of true. This is |
| 12656 | the value received by the @code{if}, and means that the search was |
| 12657 | successful. |
| 12658 | @end enumerate |
| 12659 | |
| 12660 | @noindent |
| 12661 | The side effect, the movement of point, is completed before the |
| 12662 | @code{if} function is handed the value returned by the successful |
| 12663 | conclusion of the search. |
| 12664 | |
| 12665 | When the @code{if} function receives the value of true from a successful |
| 12666 | call to @code{re-search-forward}, the @code{if} evaluates the then-part, |
| 12667 | which is the expression @code{(skip-chars-backward " \t\n")}. This |
| 12668 | expression moves backwards over any blank spaces, tabs or carriage |
| 12669 | returns until a printed character is found and then leaves point after |
| 12670 | the character. Since point has already been moved to the end of the |
| 12671 | pattern that marks the end of the sentence, this action leaves point |
| 12672 | right after the closing printed character of the sentence, which is |
| 12673 | usually a period. |
| 12674 | |
| 12675 | On the other hand, if the @code{re-search-forward} function fails to |
| 12676 | find a pattern marking the end of the sentence, the function returns |
| 12677 | false. The false then causes the @code{if} to evaluate its third |
| 12678 | argument, which is @code{(goto-char par-end)}: it moves point to the |
| 12679 | end of the paragraph. |
| 12680 | |
| 12681 | (And if the text is in a form or equivalent, and point may not move |
| 12682 | fully, then the @code{constrain-to-field} function comes into play.) |
| 12683 | |
| 12684 | Regular expression searches are exceptionally useful and the pattern |
| 12685 | illustrated by @code{re-search-forward}, in which the search is the |
| 12686 | test of an @code{if} expression, is handy. You will see or write code |
| 12687 | incorporating this pattern often. |
| 12688 | |
| 12689 | @node forward-paragraph |
| 12690 | @section @code{forward-paragraph}: a Goldmine of Functions |
| 12691 | @findex forward-paragraph |
| 12692 | |
| 12693 | @ignore |
| 12694 | @c in GNU Emacs 22 |
| 12695 | (defun forward-paragraph (&optional arg) |
| 12696 | "Move forward to end of paragraph. |
| 12697 | With argument ARG, do it ARG times; |
| 12698 | a negative argument ARG = -N means move backward N paragraphs. |
| 12699 | |
| 12700 | A line which `paragraph-start' matches either separates paragraphs |
| 12701 | \(if `paragraph-separate' matches it also) or is the first line of a paragraph. |
| 12702 | A paragraph end is the beginning of a line which is not part of the paragraph |
| 12703 | to which the end of the previous line belongs, or the end of the buffer. |
| 12704 | Returns the count of paragraphs left to move." |
| 12705 | (interactive "p") |
| 12706 | (or arg (setq arg 1)) |
| 12707 | (let* ((opoint (point)) |
| 12708 | (fill-prefix-regexp |
| 12709 | (and fill-prefix (not (equal fill-prefix "")) |
| 12710 | (not paragraph-ignore-fill-prefix) |
| 12711 | (regexp-quote fill-prefix))) |
| 12712 | ;; Remove ^ from paragraph-start and paragraph-sep if they are there. |
| 12713 | ;; These regexps shouldn't be anchored, because we look for them |
| 12714 | ;; starting at the left-margin. This allows paragraph commands to |
| 12715 | ;; work normally with indented text. |
| 12716 | ;; This hack will not find problem cases like "whatever\\|^something". |
| 12717 | (parstart (if (and (not (equal "" paragraph-start)) |
| 12718 | (equal ?^ (aref paragraph-start 0))) |
| 12719 | (substring paragraph-start 1) |
| 12720 | paragraph-start)) |
| 12721 | (parsep (if (and (not (equal "" paragraph-separate)) |
| 12722 | (equal ?^ (aref paragraph-separate 0))) |
| 12723 | (substring paragraph-separate 1) |
| 12724 | paragraph-separate)) |
| 12725 | (parsep |
| 12726 | (if fill-prefix-regexp |
| 12727 | (concat parsep "\\|" |
| 12728 | fill-prefix-regexp "[ \t]*$") |
| 12729 | parsep)) |
| 12730 | ;; This is used for searching. |
| 12731 | (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)")) |
| 12732 | start found-start) |
| 12733 | (while (and (< arg 0) (not (bobp))) |
| 12734 | (if (and (not (looking-at parsep)) |
| 12735 | (re-search-backward "^\n" (max (1- (point)) (point-min)) t) |
| 12736 | (looking-at parsep)) |
| 12737 | (setq arg (1+ arg)) |
| 12738 | (setq start (point)) |
| 12739 | ;; Move back over paragraph-separating lines. |
| 12740 | (forward-char -1) (beginning-of-line) |
| 12741 | (while (and (not (bobp)) |
| 12742 | (progn (move-to-left-margin) |
| 12743 | (looking-at parsep))) |
| 12744 | (forward-line -1)) |
| 12745 | (if (bobp) |
| 12746 | nil |
| 12747 | (setq arg (1+ arg)) |
| 12748 | ;; Go to end of the previous (non-separating) line. |
| 12749 | (end-of-line) |
| 12750 | ;; Search back for line that starts or separates paragraphs. |
| 12751 | (if (if fill-prefix-regexp |
| 12752 | ;; There is a fill prefix; it overrides parstart. |
| 12753 | (let (multiple-lines) |
| 12754 | (while (and (progn (beginning-of-line) (not (bobp))) |
| 12755 | (progn (move-to-left-margin) |
| 12756 | (not (looking-at parsep))) |
| 12757 | (looking-at fill-prefix-regexp)) |
| 12758 | (unless (= (point) start) |
| 12759 | (setq multiple-lines t)) |
| 12760 | (forward-line -1)) |
| 12761 | (move-to-left-margin) |
| 12762 | ;; This deleted code caused a long hanging-indent line |
| 12763 | ;; not to be filled together with the following lines. |
| 12764 | ;; ;; Don't move back over a line before the paragraph |
| 12765 | ;; ;; which doesn't start with fill-prefix |
| 12766 | ;; ;; unless that is the only line we've moved over. |
| 12767 | ;; (and (not (looking-at fill-prefix-regexp)) |
| 12768 | ;; multiple-lines |
| 12769 | ;; (forward-line 1)) |
| 12770 | (not (bobp))) |
| 12771 | (while (and (re-search-backward sp-parstart nil 1) |
| 12772 | (setq found-start t) |
| 12773 | ;; Found a candidate, but need to check if it is a |
| 12774 | ;; REAL parstart. |
| 12775 | (progn (setq start (point)) |
| 12776 | (move-to-left-margin) |
| 12777 | (not (looking-at parsep))) |
| 12778 | (not (and (looking-at parstart) |
| 12779 | (or (not use-hard-newlines) |
| 12780 | (bobp) |
| 12781 | (get-text-property |
| 12782 | (1- start) 'hard))))) |
| 12783 | (setq found-start nil) |
| 12784 | (goto-char start)) |
| 12785 | found-start) |
| 12786 | ;; Found one. |
| 12787 | (progn |
| 12788 | ;; Move forward over paragraph separators. |
| 12789 | ;; We know this cannot reach the place we started |
| 12790 | ;; because we know we moved back over a non-separator. |
| 12791 | (while (and (not (eobp)) |
| 12792 | (progn (move-to-left-margin) |
| 12793 | (looking-at parsep))) |
| 12794 | (forward-line 1)) |
| 12795 | ;; If line before paragraph is just margin, back up to there. |
| 12796 | (end-of-line 0) |
| 12797 | (if (> (current-column) (current-left-margin)) |
| 12798 | (forward-char 1) |
| 12799 | (skip-chars-backward " \t") |
| 12800 | (if (not (bolp)) |
| 12801 | (forward-line 1)))) |
| 12802 | ;; No starter or separator line => use buffer beg. |
| 12803 | (goto-char (point-min)))))) |
| 12804 | |
| 12805 | (while (and (> arg 0) (not (eobp))) |
| 12806 | ;; Move forward over separator lines... |
| 12807 | (while (and (not (eobp)) |
| 12808 | (progn (move-to-left-margin) (not (eobp))) |
| 12809 | (looking-at parsep)) |
| 12810 | (forward-line 1)) |
| 12811 | (unless (eobp) (setq arg (1- arg))) |
| 12812 | ;; ... and one more line. |
| 12813 | (forward-line 1) |
| 12814 | (if fill-prefix-regexp |
| 12815 | ;; There is a fill prefix; it overrides parstart. |
| 12816 | (while (and (not (eobp)) |
| 12817 | (progn (move-to-left-margin) (not (eobp))) |
| 12818 | (not (looking-at parsep)) |
| 12819 | (looking-at fill-prefix-regexp)) |
| 12820 | (forward-line 1)) |
| 12821 | (while (and (re-search-forward sp-parstart nil 1) |
| 12822 | (progn (setq start (match-beginning 0)) |
| 12823 | (goto-char start) |
| 12824 | (not (eobp))) |
| 12825 | (progn (move-to-left-margin) |
| 12826 | (not (looking-at parsep))) |
| 12827 | (or (not (looking-at parstart)) |
| 12828 | (and use-hard-newlines |
| 12829 | (not (get-text-property (1- start) 'hard))))) |
| 12830 | (forward-char 1)) |
| 12831 | (if (< (point) (point-max)) |
| 12832 | (goto-char start)))) |
| 12833 | (constrain-to-field nil opoint t) |
| 12834 | ;; Return the number of steps that could not be done. |
| 12835 | arg)) |
| 12836 | @end ignore |
| 12837 | |
| 12838 | The @code{forward-paragraph} function moves point forward to the end |
| 12839 | of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a |
| 12840 | number of functions that are important in themselves, including |
| 12841 | @code{let*}, @code{match-beginning}, and @code{looking-at}. |
| 12842 | |
| 12843 | The function definition for @code{forward-paragraph} is considerably |
| 12844 | longer than the function definition for @code{forward-sentence} |
| 12845 | because it works with a paragraph, each line of which may begin with a |
| 12846 | fill prefix. |
| 12847 | |
| 12848 | A fill prefix consists of a string of characters that are repeated at |
| 12849 | the beginning of each line. For example, in Lisp code, it is a |
| 12850 | convention to start each line of a paragraph-long comment with |
| 12851 | @samp{;;; }. In Text mode, four blank spaces make up another common |
| 12852 | fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , , |
| 12853 | emacs, The GNU Emacs Manual}, for more information about fill |
| 12854 | prefixes.) |
| 12855 | |
| 12856 | The existence of a fill prefix means that in addition to being able to |
| 12857 | find the end of a paragraph whose lines begin on the left-most |
| 12858 | column, the @code{forward-paragraph} function must be able to find the |
| 12859 | end of a paragraph when all or many of the lines in the buffer begin |
| 12860 | with the fill prefix. |
| 12861 | |
| 12862 | Moreover, it is sometimes practical to ignore a fill prefix that |
| 12863 | exists, especially when blank lines separate paragraphs. |
| 12864 | This is an added complication. |
| 12865 | |
| 12866 | @menu |
| 12867 | * forward-paragraph in brief:: Key parts of the function definition. |
| 12868 | * fwd-para let:: The @code{let*} expression. |
| 12869 | * fwd-para while:: The forward motion @code{while} loop. |
| 12870 | @end menu |
| 12871 | |
| 12872 | @ifnottex |
| 12873 | @node forward-paragraph in brief |
| 12874 | @unnumberedsubsec Shortened @code{forward-paragraph} function definition |
| 12875 | @end ifnottex |
| 12876 | |
| 12877 | Rather than print all of the @code{forward-paragraph} function, we |
| 12878 | will only print parts of it. Read without preparation, the function |
| 12879 | can be daunting! |
| 12880 | |
| 12881 | @need 800 |
| 12882 | In outline, the function looks like this: |
| 12883 | |
| 12884 | @smallexample |
| 12885 | @group |
| 12886 | (defun forward-paragraph (&optional arg) |
| 12887 | "@var{documentation}@dots{}" |
| 12888 | (interactive "p") |
| 12889 | (or arg (setq arg 1)) |
| 12890 | (let* |
| 12891 | @var{varlist} |
| 12892 | (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code} |
| 12893 | @dots{} |
| 12894 | (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code} |
| 12895 | @dots{} |
| 12896 | @end group |
| 12897 | @end smallexample |
| 12898 | |
| 12899 | The first parts of the function are routine: the function's argument |
| 12900 | list consists of one optional argument. Documentation follows. |
| 12901 | |
| 12902 | The lower case @samp{p} in the @code{interactive} declaration means |
| 12903 | that the processed prefix argument, if any, is passed to the function. |
| 12904 | This will be a number, and is the repeat count of how many paragraphs |
| 12905 | point will move. The @code{or} expression in the next line handles |
| 12906 | the common case when no argument is passed to the function, which occurs |
| 12907 | if the function is called from other code rather than interactively. |
| 12908 | This case was described earlier. (@xref{forward-sentence, The |
| 12909 | @code{forward-sentence} function}.) Now we reach the end of the |
| 12910 | familiar part of this function. |
| 12911 | |
| 12912 | @node fwd-para let |
| 12913 | @unnumberedsubsec The @code{let*} expression |
| 12914 | |
| 12915 | The next line of the @code{forward-paragraph} function begins a |
| 12916 | @code{let*} expression. This is a different than @code{let}. The |
| 12917 | symbol is @code{let*} not @code{let}. |
| 12918 | |
| 12919 | The @code{let*} special form is like @code{let} except that Emacs sets |
| 12920 | each variable in sequence, one after another, and variables in the |
| 12921 | latter part of the varlist can make use of the values to which Emacs |
| 12922 | set variables in the earlier part of the varlist. |
| 12923 | |
| 12924 | @ignore |
| 12925 | ( refappend save-excursion, , code save-excursion in code append-to-buffer .) |
| 12926 | @end ignore |
| 12927 | |
| 12928 | (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.) |
| 12929 | |
| 12930 | In the @code{let*} expression in this function, Emacs binds a total of |
| 12931 | seven variables: @code{opoint}, @code{fill-prefix-regexp}, |
| 12932 | @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and |
| 12933 | @code{found-start}. |
| 12934 | |
| 12935 | The variable @code{parsep} appears twice, first, to remove instances |
| 12936 | of @samp{^}, and second, to handle fill prefixes. |
| 12937 | |
| 12938 | The variable @code{opoint} is just the value of @code{point}. As you |
| 12939 | can guess, it is used in a @code{constrain-to-field} expression, just |
| 12940 | as in @code{forward-sentence}. |
| 12941 | |
| 12942 | The variable @code{fill-prefix-regexp} is set to the value returned by |
| 12943 | evaluating the following list: |
| 12944 | |
| 12945 | @smallexample |
| 12946 | @group |
| 12947 | (and fill-prefix |
| 12948 | (not (equal fill-prefix "")) |
| 12949 | (not paragraph-ignore-fill-prefix) |
| 12950 | (regexp-quote fill-prefix)) |
| 12951 | @end group |
| 12952 | @end smallexample |
| 12953 | |
| 12954 | @noindent |
| 12955 | This is an expression whose first element is the @code{and} special form. |
| 12956 | |
| 12957 | As we learned earlier (@pxref{kill-new function, , The @code{kill-new} |
| 12958 | function}), the @code{and} special form evaluates each of its |
| 12959 | arguments until one of the arguments returns a value of @code{nil}, in |
| 12960 | which case the @code{and} expression returns @code{nil}; however, if |
| 12961 | none of the arguments returns a value of @code{nil}, the value |
| 12962 | resulting from evaluating the last argument is returned. (Since such |
| 12963 | a value is not @code{nil}, it is considered true in Lisp.) In other |
| 12964 | words, an @code{and} expression returns a true value only if all its |
| 12965 | arguments are true. |
| 12966 | @findex and |
| 12967 | |
| 12968 | In this case, the variable @code{fill-prefix-regexp} is bound to a |
| 12969 | non-@code{nil} value only if the following four expressions produce a |
| 12970 | true (i.e., a non-@code{nil}) value when they are evaluated; otherwise, |
| 12971 | @code{fill-prefix-regexp} is bound to @code{nil}. |
| 12972 | |
| 12973 | @table @code |
| 12974 | @item fill-prefix |
| 12975 | When this variable is evaluated, the value of the fill prefix, if any, |
| 12976 | is returned. If there is no fill prefix, this variable returns |
| 12977 | @code{nil}. |
| 12978 | |
| 12979 | @item (not (equal fill-prefix "") |
| 12980 | This expression checks whether an existing fill prefix is an empty |
| 12981 | string, that is, a string with no characters in it. An empty string is |
| 12982 | not a useful fill prefix. |
| 12983 | |
| 12984 | @item (not paragraph-ignore-fill-prefix) |
| 12985 | This expression returns @code{nil} if the variable |
| 12986 | @code{paragraph-ignore-fill-prefix} has been turned on by being set to a |
| 12987 | true value such as @code{t}. |
| 12988 | |
| 12989 | @item (regexp-quote fill-prefix) |
| 12990 | This is the last argument to the @code{and} special form. If all the |
| 12991 | arguments to the @code{and} are true, the value resulting from |
| 12992 | evaluating this expression will be returned by the @code{and} expression |
| 12993 | and bound to the variable @code{fill-prefix-regexp}, |
| 12994 | @end table |
| 12995 | |
| 12996 | @findex regexp-quote |
| 12997 | @noindent |
| 12998 | The result of evaluating this @code{and} expression successfully is that |
| 12999 | @code{fill-prefix-regexp} will be bound to the value of |
| 13000 | @code{fill-prefix} as modified by the @code{regexp-quote} function. |
| 13001 | What @code{regexp-quote} does is read a string and return a regular |
| 13002 | expression that will exactly match the string and match nothing else. |
| 13003 | This means that @code{fill-prefix-regexp} will be set to a value that |
| 13004 | will exactly match the fill prefix if the fill prefix exists. |
| 13005 | Otherwise, the variable will be set to @code{nil}. |
| 13006 | |
| 13007 | The next two local variables in the @code{let*} expression are |
| 13008 | designed to remove instances of @samp{^} from @code{parstart} and |
| 13009 | @code{parsep}, the local variables which indicate the paragraph start |
| 13010 | and the paragraph separator. The next expression sets @code{parsep} |
| 13011 | again. That is to handle fill prefixes. |
| 13012 | |
| 13013 | This is the setting that requires the definition call @code{let*} |
| 13014 | rather than @code{let}. The true-or-false-test for the @code{if} |
| 13015 | depends on whether the variable @code{fill-prefix-regexp} evaluates to |
| 13016 | @code{nil} or some other value. |
| 13017 | |
| 13018 | If @code{fill-prefix-regexp} does not have a value, Emacs evaluates |
| 13019 | the else-part of the @code{if} expression and binds @code{parsep} to |
| 13020 | its local value. (@code{parsep} is a regular expression that matches |
| 13021 | what separates paragraphs.) |
| 13022 | |
| 13023 | But if @code{fill-prefix-regexp} does have a value, Emacs evaluates |
| 13024 | the then-part of the @code{if} expression and binds @code{parsep} to a |
| 13025 | regular expression that includes the @code{fill-prefix-regexp} as part |
| 13026 | of the pattern. |
| 13027 | |
| 13028 | Specifically, @code{parsep} is set to the original value of the |
| 13029 | paragraph separate regular expression concatenated with an alternative |
| 13030 | expression that consists of the @code{fill-prefix-regexp} followed by |
| 13031 | optional whitespace to the end of the line. The whitespace is defined |
| 13032 | by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the |
| 13033 | regexp as an alternative to @code{parsep}. |
| 13034 | |
| 13035 | According to a comment in the code, the next local variable, |
| 13036 | @code{sp-parstart}, is used for searching, and then the final two, |
| 13037 | @code{start} and @code{found-start}, are set to @code{nil}. |
| 13038 | |
| 13039 | Now we get into the body of the @code{let*}. The first part of the body |
| 13040 | of the @code{let*} deals with the case when the function is given a |
| 13041 | negative argument and is therefore moving backwards. We will skip this |
| 13042 | section. |
| 13043 | |
| 13044 | @node fwd-para while |
| 13045 | @unnumberedsubsec The forward motion @code{while} loop |
| 13046 | |
| 13047 | The second part of the body of the @code{let*} deals with forward |
| 13048 | motion. It is a @code{while} loop that repeats itself so long as the |
| 13049 | value of @code{arg} is greater than zero. In the most common use of |
| 13050 | the function, the value of the argument is 1, so the body of the |
| 13051 | @code{while} loop is evaluated exactly once, and the cursor moves |
| 13052 | forward one paragraph. |
| 13053 | |
| 13054 | @ignore |
| 13055 | (while (and (> arg 0) (not (eobp))) |
| 13056 | |
| 13057 | ;; Move forward over separator lines... |
| 13058 | (while (and (not (eobp)) |
| 13059 | (progn (move-to-left-margin) (not (eobp))) |
| 13060 | (looking-at parsep)) |
| 13061 | (forward-line 1)) |
| 13062 | (unless (eobp) (setq arg (1- arg))) |
| 13063 | ;; ... and one more line. |
| 13064 | (forward-line 1) |
| 13065 | |
| 13066 | (if fill-prefix-regexp |
| 13067 | ;; There is a fill prefix; it overrides parstart. |
| 13068 | (while (and (not (eobp)) |
| 13069 | (progn (move-to-left-margin) (not (eobp))) |
| 13070 | (not (looking-at parsep)) |
| 13071 | (looking-at fill-prefix-regexp)) |
| 13072 | (forward-line 1)) |
| 13073 | |
| 13074 | (while (and (re-search-forward sp-parstart nil 1) |
| 13075 | (progn (setq start (match-beginning 0)) |
| 13076 | (goto-char start) |
| 13077 | (not (eobp))) |
| 13078 | (progn (move-to-left-margin) |
| 13079 | (not (looking-at parsep))) |
| 13080 | (or (not (looking-at parstart)) |
| 13081 | (and use-hard-newlines |
| 13082 | (not (get-text-property (1- start) 'hard))))) |
| 13083 | (forward-char 1)) |
| 13084 | |
| 13085 | (if (< (point) (point-max)) |
| 13086 | (goto-char start)))) |
| 13087 | @end ignore |
| 13088 | |
| 13089 | This part handles three situations: when point is between paragraphs, |
| 13090 | when there is a fill prefix and when there is no fill prefix. |
| 13091 | |
| 13092 | @need 800 |
| 13093 | The @code{while} loop looks like this: |
| 13094 | |
| 13095 | @smallexample |
| 13096 | @group |
| 13097 | ;; @r{going forwards and not at the end of the buffer} |
| 13098 | (while (and (> arg 0) (not (eobp))) |
| 13099 | |
| 13100 | ;; @r{between paragraphs} |
| 13101 | ;; Move forward over separator lines... |
| 13102 | (while (and (not (eobp)) |
| 13103 | (progn (move-to-left-margin) (not (eobp))) |
| 13104 | (looking-at parsep)) |
| 13105 | (forward-line 1)) |
| 13106 | ;; @r{This decrements the loop} |
| 13107 | (unless (eobp) (setq arg (1- arg))) |
| 13108 | ;; ... and one more line. |
| 13109 | (forward-line 1) |
| 13110 | @end group |
| 13111 | |
| 13112 | @group |
| 13113 | (if fill-prefix-regexp |
| 13114 | ;; There is a fill prefix; it overrides parstart; |
| 13115 | ;; we go forward line by line |
| 13116 | (while (and (not (eobp)) |
| 13117 | (progn (move-to-left-margin) (not (eobp))) |
| 13118 | (not (looking-at parsep)) |
| 13119 | (looking-at fill-prefix-regexp)) |
| 13120 | (forward-line 1)) |
| 13121 | @end group |
| 13122 | |
| 13123 | @group |
| 13124 | ;; There is no fill prefix; |
| 13125 | ;; we go forward character by character |
| 13126 | (while (and (re-search-forward sp-parstart nil 1) |
| 13127 | (progn (setq start (match-beginning 0)) |
| 13128 | (goto-char start) |
| 13129 | (not (eobp))) |
| 13130 | (progn (move-to-left-margin) |
| 13131 | (not (looking-at parsep))) |
| 13132 | (or (not (looking-at parstart)) |
| 13133 | (and use-hard-newlines |
| 13134 | (not (get-text-property (1- start) 'hard))))) |
| 13135 | (forward-char 1)) |
| 13136 | @end group |
| 13137 | |
| 13138 | @group |
| 13139 | ;; and if there is no fill prefix and if we are not at the end, |
| 13140 | ;; go to whatever was found in the regular expression search |
| 13141 | ;; for sp-parstart |
| 13142 | (if (< (point) (point-max)) |
| 13143 | (goto-char start)))) |
| 13144 | @end group |
| 13145 | @end smallexample |
| 13146 | |
| 13147 | @findex eobp |
| 13148 | We can see that this is a decrementing counter @code{while} loop, |
| 13149 | using the expression @code{(setq arg (1- arg))} as the decrementer. |
| 13150 | That expression is not far from the @code{while}, but is hidden in |
| 13151 | another Lisp macro, an @code{unless} macro. Unless we are at the end |
| 13152 | of the buffer---that is what the @code{eobp} function determines; it |
| 13153 | is an abbreviation of @samp{End Of Buffer P}---we decrease the value |
| 13154 | of @code{arg} by one. |
| 13155 | |
| 13156 | (If we are at the end of the buffer, we cannot go forward any more and |
| 13157 | the next loop of the @code{while} expression will test false since the |
| 13158 | test is an @code{and} with @code{(not (eobp))}. The @code{not} |
| 13159 | function means exactly as you expect; it is another name for |
| 13160 | @code{null}, a function that returns true when its argument is false.) |
| 13161 | |
| 13162 | Interestingly, the loop count is not decremented until we leave the |
| 13163 | space between paragraphs, unless we come to the end of buffer or stop |
| 13164 | seeing the local value of the paragraph separator. |
| 13165 | |
| 13166 | That second @code{while} also has a @code{(move-to-left-margin)} |
| 13167 | expression. The function is self-explanatory. It is inside a |
| 13168 | @code{progn} expression and not the last element of its body, so it is |
| 13169 | only invoked for its side effect, which is to move point to the left |
| 13170 | margin of the current line. |
| 13171 | |
| 13172 | @findex looking-at |
| 13173 | The @code{looking-at} function is also self-explanatory; it returns |
| 13174 | true if the text after point matches the regular expression given as |
| 13175 | its argument. |
| 13176 | |
| 13177 | The rest of the body of the loop looks difficult at first, but makes |
| 13178 | sense as you come to understand it. |
| 13179 | |
| 13180 | @need 800 |
| 13181 | First consider what happens if there is a fill prefix: |
| 13182 | |
| 13183 | @smallexample |
| 13184 | @group |
| 13185 | (if fill-prefix-regexp |
| 13186 | ;; There is a fill prefix; it overrides parstart; |
| 13187 | ;; we go forward line by line |
| 13188 | (while (and (not (eobp)) |
| 13189 | (progn (move-to-left-margin) (not (eobp))) |
| 13190 | (not (looking-at parsep)) |
| 13191 | (looking-at fill-prefix-regexp)) |
| 13192 | (forward-line 1)) |
| 13193 | @end group |
| 13194 | @end smallexample |
| 13195 | |
| 13196 | @noindent |
| 13197 | This expression moves point forward line by line so long |
| 13198 | as four conditions are true: |
| 13199 | |
| 13200 | @enumerate |
| 13201 | @item |
| 13202 | Point is not at the end of the buffer. |
| 13203 | |
| 13204 | @item |
| 13205 | We can move to the left margin of the text and are |
| 13206 | not at the end of the buffer. |
| 13207 | |
| 13208 | @item |
| 13209 | The text following point does not separate paragraphs. |
| 13210 | |
| 13211 | @item |
| 13212 | The pattern following point is the fill prefix regular expression. |
| 13213 | @end enumerate |
| 13214 | |
| 13215 | The last condition may be puzzling, until you remember that point was |
| 13216 | moved to the beginning of the line early in the @code{forward-paragraph} |
| 13217 | function. This means that if the text has a fill prefix, the |
| 13218 | @code{looking-at} function will see it. |
| 13219 | |
| 13220 | @need 1250 |
| 13221 | Consider what happens when there is no fill prefix. |
| 13222 | |
| 13223 | @smallexample |
| 13224 | @group |
| 13225 | (while (and (re-search-forward sp-parstart nil 1) |
| 13226 | (progn (setq start (match-beginning 0)) |
| 13227 | (goto-char start) |
| 13228 | (not (eobp))) |
| 13229 | (progn (move-to-left-margin) |
| 13230 | (not (looking-at parsep))) |
| 13231 | (or (not (looking-at parstart)) |
| 13232 | (and use-hard-newlines |
| 13233 | (not (get-text-property (1- start) 'hard))))) |
| 13234 | (forward-char 1)) |
| 13235 | @end group |
| 13236 | @end smallexample |
| 13237 | |
| 13238 | @noindent |
| 13239 | This @code{while} loop has us searching forward for |
| 13240 | @code{sp-parstart}, which is the combination of possible whitespace |
| 13241 | with a the local value of the start of a paragraph or of a paragraph |
| 13242 | separator. (The latter two are within an expression starting |
| 13243 | @code{\(?:} so that they are not referenced by the |
| 13244 | @code{match-beginning} function.) |
| 13245 | |
| 13246 | @need 800 |
| 13247 | The two expressions, |
| 13248 | |
| 13249 | @smallexample |
| 13250 | @group |
| 13251 | (setq start (match-beginning 0)) |
| 13252 | (goto-char start) |
| 13253 | @end group |
| 13254 | @end smallexample |
| 13255 | |
| 13256 | @noindent |
| 13257 | mean go to the start of the text matched by the regular expression |
| 13258 | search. |
| 13259 | |
| 13260 | The @code{(match-beginning 0)} expression is new. It returns a number |
| 13261 | specifying the location of the start of the text that was matched by |
| 13262 | the last search. |
| 13263 | |
| 13264 | The @code{match-beginning} function is used here because of a |
| 13265 | characteristic of a forward search: a successful forward search, |
| 13266 | regardless of whether it is a plain search or a regular expression |
| 13267 | search, moves point to the end of the text that is found. In this |
| 13268 | case, a successful search moves point to the end of the pattern for |
| 13269 | @code{sp-parstart}. |
| 13270 | |
| 13271 | However, we want to put point at the end of the current paragraph, not |
| 13272 | somewhere else. Indeed, since the search possibly includes the |
| 13273 | paragraph separator, point may end up at the beginning of the next one |
| 13274 | unless we use an expression that includes @code{match-beginning}. |
| 13275 | |
| 13276 | @findex match-beginning |
| 13277 | When given an argument of 0, @code{match-beginning} returns the |
| 13278 | position that is the start of the text matched by the most recent |
| 13279 | search. In this case, the most recent search looks for |
| 13280 | @code{sp-parstart}. The @code{(match-beginning 0)} expression returns |
| 13281 | the beginning position of that pattern, rather than the end position |
| 13282 | of that pattern. |
| 13283 | |
| 13284 | (Incidentally, when passed a positive number as an argument, the |
| 13285 | @code{match-beginning} function returns the location of point at that |
| 13286 | parenthesized expression in the last search unless that parenthesized |
| 13287 | expression begins with @code{\(?:}. I don't know why @code{\(?:} |
| 13288 | appears here since the argument is 0.) |
| 13289 | |
| 13290 | @need 1250 |
| 13291 | The last expression when there is no fill prefix is |
| 13292 | |
| 13293 | @smallexample |
| 13294 | @group |
| 13295 | (if (< (point) (point-max)) |
| 13296 | (goto-char start)))) |
| 13297 | @end group |
| 13298 | @end smallexample |
| 13299 | |
| 13300 | @noindent |
| 13301 | This says that if there is no fill prefix and if we are not at the |
| 13302 | end, point should move to the beginning of whatever was found by the |
| 13303 | regular expression search for @code{sp-parstart}. |
| 13304 | |
| 13305 | The full definition for the @code{forward-paragraph} function not only |
| 13306 | includes code for going forwards, but also code for going backwards. |
| 13307 | |
| 13308 | If you are reading this inside of GNU Emacs and you want to see the |
| 13309 | whole function, you can type @kbd{C-h f} (@code{describe-function}) |
| 13310 | and the name of the function. This gives you the function |
| 13311 | documentation and the name of the library containing the function's |
| 13312 | source. Place point over the name of the library and press the RET |
| 13313 | key; you will be taken directly to the source. (Be sure to install |
| 13314 | your sources! Without them, you are like a person who tries to drive |
| 13315 | a car with his eyes shut!) |
| 13316 | |
| 13317 | @node etags |
| 13318 | @section Create Your Own @file{TAGS} File |
| 13319 | @findex etags |
| 13320 | @cindex @file{TAGS} file, create own |
| 13321 | |
| 13322 | Besides @kbd{C-h f} (@code{describe-function}), another way to see the |
| 13323 | source of a function is to type @kbd{M-.} (@code{find-tag}) and the |
| 13324 | name of the function when prompted for it. This is a good habit to |
| 13325 | get into. The @kbd{M-.} (@code{find-tag}) command takes you directly |
| 13326 | to the source for a function, variable, or node. The function depends |
| 13327 | on tags tables to tell it where to go. |
| 13328 | |
| 13329 | If the @code{find-tag} function first asks you for the name of a |
| 13330 | @file{TAGS} table, give it the name of a @file{TAGS} file such as |
| 13331 | @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your |
| 13332 | @file{TAGS} file depends on how your copy of Emacs was installed. I |
| 13333 | just told you the location that provides both my C and my Emacs Lisp |
| 13334 | sources.) |
| 13335 | |
| 13336 | You can also create your own @file{TAGS} file for directories that |
| 13337 | lack one. |
| 13338 | |
| 13339 | You often need to build and install tags tables yourself. They are |
| 13340 | not built automatically. A tags table is called a @file{TAGS} file; |
| 13341 | the name is in upper case letters. |
| 13342 | |
| 13343 | You can create a @file{TAGS} file by calling the @code{etags} program |
| 13344 | that comes as a part of the Emacs distribution. Usually, @code{etags} |
| 13345 | is compiled and installed when Emacs is built. (@code{etags} is not |
| 13346 | an Emacs Lisp function or a part of Emacs; it is a C program.) |
| 13347 | |
| 13348 | @need 1250 |
| 13349 | To create a @file{TAGS} file, first switch to the directory in which |
| 13350 | you want to create the file. In Emacs you can do this with the |
| 13351 | @kbd{M-x cd} command, or by visiting a file in the directory, or by |
| 13352 | listing the directory with @kbd{C-x d} (@code{dired}). Then run the |
| 13353 | compile command, with @w{@code{etags *.el}} as the command to execute |
| 13354 | |
| 13355 | @smallexample |
| 13356 | M-x compile RET etags *.el RET |
| 13357 | @end smallexample |
| 13358 | |
| 13359 | @noindent |
| 13360 | to create a @file{TAGS} file for Emacs Lisp. |
| 13361 | |
| 13362 | For example, if you have a large number of files in your |
| 13363 | @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it, |
| 13364 | of which I load 12---you can create a @file{TAGS} file for the Emacs |
| 13365 | Lisp files in that directory. |
| 13366 | |
| 13367 | @need 1250 |
| 13368 | The @code{etags} program takes all the usual shell `wildcards'. For |
| 13369 | example, if you have two directories for which you want a single |
| 13370 | @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where |
| 13371 | @file{../elisp/} is the second directory: |
| 13372 | |
| 13373 | @smallexample |
| 13374 | M-x compile RET etags *.el ../elisp/*.el RET |
| 13375 | @end smallexample |
| 13376 | |
| 13377 | @need 1250 |
| 13378 | Type |
| 13379 | |
| 13380 | @smallexample |
| 13381 | M-x compile RET etags --help RET |
| 13382 | @end smallexample |
| 13383 | |
| 13384 | @noindent |
| 13385 | to see a list of the options accepted by @code{etags} as well as a |
| 13386 | list of supported languages. |
| 13387 | |
| 13388 | The @code{etags} program handles more than 20 languages, including |
| 13389 | Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java, |
| 13390 | LaTeX, Pascal, Perl, PostScript, Python, TeX, Texinfo, makefiles, and |
| 13391 | most assemblers. The program has no switches for specifying the |
| 13392 | language; it recognizes the language in an input file according to its |
| 13393 | file name and contents. |
| 13394 | |
| 13395 | @file{etags} is very helpful when you are writing code yourself and |
| 13396 | want to refer back to functions you have already written. Just run |
| 13397 | @code{etags} again at intervals as you write new functions, so they |
| 13398 | become part of the @file{TAGS} file. |
| 13399 | |
| 13400 | If you think an appropriate @file{TAGS} file already exists for what |
| 13401 | you want, but do not know where it is, you can use the @code{locate} |
| 13402 | program to attempt to find it. |
| 13403 | |
| 13404 | Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list |
| 13405 | for you the full path names of all your @file{TAGS} files. On my |
| 13406 | system, this command lists 34 @file{TAGS} files. On the other hand, a |
| 13407 | `plain vanilla' system I recently installed did not contain any |
| 13408 | @file{TAGS} files. |
| 13409 | |
| 13410 | If the tags table you want has been created, you can use the @code{M-x |
| 13411 | visit-tags-table} command to specify it. Otherwise, you will need to |
| 13412 | create the tag table yourself and then use @code{M-x |
| 13413 | visit-tags-table}. |
| 13414 | |
| 13415 | @subsubheading Building Tags in the Emacs sources |
| 13416 | @cindex Building Tags in the Emacs sources |
| 13417 | @cindex Tags in the Emacs sources |
| 13418 | @findex make tags |
| 13419 | |
| 13420 | The GNU Emacs sources come with a @file{Makefile} that contains a |
| 13421 | sophisticated @code{etags} command that creates, collects, and merges |
| 13422 | tags tables from all over the Emacs sources and puts the information |
| 13423 | into one @file{TAGS} file in the @file{src/} directory. (The |
| 13424 | @file{src/} directory is below the top level of your Emacs directory.) |
| 13425 | |
| 13426 | @need 1250 |
| 13427 | To build this @file{TAGS} file, go to the top level of your Emacs |
| 13428 | source directory and run the compile command @code{make tags}: |
| 13429 | |
| 13430 | @smallexample |
| 13431 | M-x compile RET make tags RET |
| 13432 | @end smallexample |
| 13433 | |
| 13434 | @noindent |
| 13435 | (The @code{make tags} command works well with the GNU Emacs sources, |
| 13436 | as well as with some other source packages.) |
| 13437 | |
| 13438 | For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs |
| 13439 | Manual}. |
| 13440 | |
| 13441 | @node Regexp Review |
| 13442 | @section Review |
| 13443 | |
| 13444 | Here is a brief summary of some recently introduced functions. |
| 13445 | |
| 13446 | @table @code |
| 13447 | @item while |
| 13448 | Repeatedly evaluate the body of the expression so long as the first |
| 13449 | element of the body tests true. Then return @code{nil}. (The |
| 13450 | expression is evaluated only for its side effects.) |
| 13451 | |
| 13452 | @need 1250 |
| 13453 | For example: |
| 13454 | |
| 13455 | @smallexample |
| 13456 | @group |
| 13457 | (let ((foo 2)) |
| 13458 | (while (> foo 0) |
| 13459 | (insert (format "foo is %d.\n" foo)) |
| 13460 | (setq foo (1- foo)))) |
| 13461 | |
| 13462 | @result{} foo is 2. |
| 13463 | foo is 1. |
| 13464 | nil |
| 13465 | @end group |
| 13466 | @end smallexample |
| 13467 | |
| 13468 | @noindent |
| 13469 | (The @code{insert} function inserts its arguments at point; the |
| 13470 | @code{format} function returns a string formatted from its arguments |
| 13471 | the way @code{message} formats its arguments; @code{\n} produces a new |
| 13472 | line.) |
| 13473 | |
| 13474 | @item re-search-forward |
| 13475 | Search for a pattern, and if the pattern is found, move point to rest |
| 13476 | just after it. |
| 13477 | |
| 13478 | @noindent |
| 13479 | Takes four arguments, like @code{search-forward}: |
| 13480 | |
| 13481 | @enumerate |
| 13482 | @item |
| 13483 | A regular expression that specifies the pattern to search for. |
| 13484 | (Remember to put quotation marks around this argument!) |
| 13485 | |
| 13486 | @item |
| 13487 | Optionally, the limit of the search. |
| 13488 | |
| 13489 | @item |
| 13490 | Optionally, what to do if the search fails, return @code{nil} or an |
| 13491 | error message. |
| 13492 | |
| 13493 | @item |
| 13494 | Optionally, how many times to repeat the search; if negative, the |
| 13495 | search goes backwards. |
| 13496 | @end enumerate |
| 13497 | |
| 13498 | @item let* |
| 13499 | Bind some variables locally to particular values, |
| 13500 | and then evaluate the remaining arguments, returning the value of the |
| 13501 | last one. While binding the local variables, use the local values of |
| 13502 | variables bound earlier, if any. |
| 13503 | |
| 13504 | @need 1250 |
| 13505 | For example: |
| 13506 | |
| 13507 | @smallexample |
| 13508 | @group |
| 13509 | (let* ((foo 7) |
| 13510 | (bar (* 3 foo))) |
| 13511 | (message "`bar' is %d." bar)) |
| 13512 | @result{} `bar' is 21. |
| 13513 | @end group |
| 13514 | @end smallexample |
| 13515 | |
| 13516 | @item match-beginning |
| 13517 | Return the position of the start of the text found by the last regular |
| 13518 | expression search. |
| 13519 | |
| 13520 | @item looking-at |
| 13521 | Return @code{t} for true if the text after point matches the argument, |
| 13522 | which should be a regular expression. |
| 13523 | |
| 13524 | @item eobp |
| 13525 | Return @code{t} for true if point is at the end of the accessible part |
| 13526 | of a buffer. The end of the accessible part is the end of the buffer |
| 13527 | if the buffer is not narrowed; it is the end of the narrowed part if |
| 13528 | the buffer is narrowed. |
| 13529 | @end table |
| 13530 | |
| 13531 | @need 1500 |
| 13532 | @node re-search Exercises |
| 13533 | @section Exercises with @code{re-search-forward} |
| 13534 | |
| 13535 | @itemize @bullet |
| 13536 | @item |
| 13537 | Write a function to search for a regular expression that matches two |
| 13538 | or more blank lines in sequence. |
| 13539 | |
| 13540 | @item |
| 13541 | Write a function to search for duplicated words, such as `the the'. |
| 13542 | @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs |
| 13543 | Manual}, for information on how to write a regexp (a regular |
| 13544 | expression) to match a string that is composed of two identical |
| 13545 | halves. You can devise several regexps; some are better than others. |
| 13546 | The function I use is described in an appendix, along with several |
| 13547 | regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}. |
| 13548 | @end itemize |
| 13549 | |
| 13550 | @node Counting Words |
| 13551 | @chapter Counting: Repetition and Regexps |
| 13552 | @cindex Repetition for word counting |
| 13553 | @cindex Regular expressions for word counting |
| 13554 | |
| 13555 | Repetition and regular expression searches are powerful tools that you |
| 13556 | often use when you write code in Emacs Lisp. This chapter illustrates |
| 13557 | the use of regular expression searches through the construction of |
| 13558 | word count commands using @code{while} loops and recursion. |
| 13559 | |
| 13560 | @menu |
| 13561 | * Why Count Words:: |
| 13562 | * @value{COUNT-WORDS}:: Use a regexp, but find a problem. |
| 13563 | * recursive-count-words:: Start with case of no words in region. |
| 13564 | * Counting Exercise:: |
| 13565 | @end menu |
| 13566 | |
| 13567 | @ifnottex |
| 13568 | @node Why Count Words |
| 13569 | @unnumberedsec Counting words |
| 13570 | @end ifnottex |
| 13571 | |
| 13572 | The standard Emacs distribution contains functions for counting the |
| 13573 | number of lines and words within a region. |
| 13574 | |
| 13575 | Certain types of writing ask you to count words. Thus, if you write |
| 13576 | an essay, you may be limited to 800 words; if you write a novel, you |
| 13577 | may discipline yourself to write 1000 words a day. It seems odd, but |
| 13578 | for a long time, Emacs lacked a word count command. Perhaps people used |
| 13579 | Emacs mostly for code or types of documentation that did not require |
| 13580 | word counts; or perhaps they restricted themselves to the operating |
| 13581 | system word count command, @code{wc}. Alternatively, people may have |
| 13582 | followed the publishers' convention and computed a word count by |
| 13583 | dividing the number of characters in a document by five. |
| 13584 | |
| 13585 | There are many ways to implement a command to count words. Here are |
| 13586 | some examples, which you may wish to compare with the standard Emacs |
| 13587 | command, @code{count-words-region}. |
| 13588 | |
| 13589 | @node @value{COUNT-WORDS} |
| 13590 | @section The @code{@value{COUNT-WORDS}} Function |
| 13591 | @findex @value{COUNT-WORDS} |
| 13592 | |
| 13593 | A word count command could count words in a line, paragraph, region, |
| 13594 | or buffer. What should the command cover? You could design the |
| 13595 | command to count the number of words in a complete buffer. However, |
| 13596 | the Emacs tradition encourages flexibility---you may want to count |
| 13597 | words in just a section, rather than all of a buffer. So it makes |
| 13598 | more sense to design the command to count the number of words in a |
| 13599 | region. Once you have a command to count words in a region, you can, |
| 13600 | if you wish, count words in a whole buffer by marking it with |
| 13601 | @w{@kbd{C-x h}} (@code{mark-whole-buffer}). |
| 13602 | |
| 13603 | Clearly, counting words is a repetitive act: starting from the |
| 13604 | beginning of the region, you count the first word, then the second |
| 13605 | word, then the third word, and so on, until you reach the end of the |
| 13606 | region. This means that word counting is ideally suited to recursion |
| 13607 | or to a @code{while} loop. |
| 13608 | |
| 13609 | @menu |
| 13610 | * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop. |
| 13611 | * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}. |
| 13612 | @end menu |
| 13613 | |
| 13614 | @ifnottex |
| 13615 | @node Design @value{COUNT-WORDS} |
| 13616 | @unnumberedsubsec Designing @code{@value{COUNT-WORDS}} |
| 13617 | @end ifnottex |
| 13618 | |
| 13619 | First, we will implement the word count command with a @code{while} |
| 13620 | loop, then with recursion. The command will, of course, be |
| 13621 | interactive. |
| 13622 | |
| 13623 | @need 800 |
| 13624 | The template for an interactive function definition is, as always: |
| 13625 | |
| 13626 | @smallexample |
| 13627 | @group |
| 13628 | (defun @var{name-of-function} (@var{argument-list}) |
| 13629 | "@var{documentation}@dots{}" |
| 13630 | (@var{interactive-expression}@dots{}) |
| 13631 | @var{body}@dots{}) |
| 13632 | @end group |
| 13633 | @end smallexample |
| 13634 | |
| 13635 | What we need to do is fill in the slots. |
| 13636 | |
| 13637 | The name of the function should be self-explanatory and similar to the |
| 13638 | existing @code{count-lines-region} name. This makes the name easier |
| 13639 | to remember. @code{count-words-region} is the obvious choice. Since |
| 13640 | that name is now used for the standard Emacs command to count words, we |
| 13641 | will name our implementation @code{@value{COUNT-WORDS}}. |
| 13642 | |
| 13643 | The function counts words within a region. This means that the |
| 13644 | argument list must contain symbols that are bound to the two |
| 13645 | positions, the beginning and end of the region. These two positions |
| 13646 | can be called @samp{beginning} and @samp{end} respectively. The first |
| 13647 | line of the documentation should be a single sentence, since that is |
| 13648 | all that is printed as documentation by a command such as |
| 13649 | @code{apropos}. The interactive expression will be of the form |
| 13650 | @samp{(interactive "r")}, since that will cause Emacs to pass the |
| 13651 | beginning and end of the region to the function's argument list. All |
| 13652 | this is routine. |
| 13653 | |
| 13654 | The body of the function needs to be written to do three tasks: |
| 13655 | first, to set up conditions under which the @code{while} loop can |
| 13656 | count words, second, to run the @code{while} loop, and third, to send |
| 13657 | a message to the user. |
| 13658 | |
| 13659 | When a user calls @code{@value{COUNT-WORDS}}, point may be at the |
| 13660 | beginning or the end of the region. However, the counting process |
| 13661 | must start at the beginning of the region. This means we will want |
| 13662 | to put point there if it is not already there. Executing |
| 13663 | @code{(goto-char beginning)} ensures this. Of course, we will want to |
| 13664 | return point to its expected position when the function finishes its |
| 13665 | work. For this reason, the body must be enclosed in a |
| 13666 | @code{save-excursion} expression. |
| 13667 | |
| 13668 | The central part of the body of the function consists of a |
| 13669 | @code{while} loop in which one expression jumps point forward word by |
| 13670 | word, and another expression counts those jumps. The true-or-false-test |
| 13671 | of the @code{while} loop should test true so long as point should jump |
| 13672 | forward, and false when point is at the end of the region. |
| 13673 | |
| 13674 | We could use @code{(forward-word 1)} as the expression for moving point |
| 13675 | forward word by word, but it is easier to see what Emacs identifies as a |
| 13676 | `word' if we use a regular expression search. |
| 13677 | |
| 13678 | A regular expression search that finds the pattern for which it is |
| 13679 | searching leaves point after the last character matched. This means |
| 13680 | that a succession of successful word searches will move point forward |
| 13681 | word by word. |
| 13682 | |
| 13683 | As a practical matter, we want the regular expression search to jump |
| 13684 | over whitespace and punctuation between words as well as over the |
| 13685 | words themselves. A regexp that refuses to jump over interword |
| 13686 | whitespace would never jump more than one word! This means that |
| 13687 | the regexp should include the whitespace and punctuation that follows |
| 13688 | a word, if any, as well as the word itself. (A word may end a buffer |
| 13689 | and not have any following whitespace or punctuation, so that part of |
| 13690 | the regexp must be optional.) |
| 13691 | |
| 13692 | Thus, what we want for the regexp is a pattern defining one or more |
| 13693 | word constituent characters followed, optionally, by one or more |
| 13694 | characters that are not word constituents. The regular expression for |
| 13695 | this is: |
| 13696 | |
| 13697 | @smallexample |
| 13698 | \w+\W* |
| 13699 | @end smallexample |
| 13700 | |
| 13701 | @noindent |
| 13702 | The buffer's syntax table determines which characters are and are not |
| 13703 | word constituents. For more information about syntax, |
| 13704 | @pxref{Syntax Tables, , Syntax Tables, elisp, The GNU Emacs Lisp |
| 13705 | Reference Manual}. |
| 13706 | |
| 13707 | @need 800 |
| 13708 | The search expression looks like this: |
| 13709 | |
| 13710 | @smallexample |
| 13711 | (re-search-forward "\\w+\\W*") |
| 13712 | @end smallexample |
| 13713 | |
| 13714 | @noindent |
| 13715 | (Note that paired backslashes precede the @samp{w} and @samp{W}. A |
| 13716 | single backslash has special meaning to the Emacs Lisp interpreter. |
| 13717 | It indicates that the following character is interpreted differently |
| 13718 | than usual. For example, the two characters, @samp{\n}, stand for |
| 13719 | @samp{newline}, rather than for a backslash followed by @samp{n}. Two |
| 13720 | backslashes in a row stand for an ordinary, `unspecial' backslash, so |
| 13721 | Emacs Lisp interpreter ends of seeing a single backslash followed by a |
| 13722 | letter. So it discovers the letter is special.) |
| 13723 | |
| 13724 | We need a counter to count how many words there are; this variable |
| 13725 | must first be set to 0 and then incremented each time Emacs goes |
| 13726 | around the @code{while} loop. The incrementing expression is simply: |
| 13727 | |
| 13728 | @smallexample |
| 13729 | (setq count (1+ count)) |
| 13730 | @end smallexample |
| 13731 | |
| 13732 | Finally, we want to tell the user how many words there are in the |
| 13733 | region. The @code{message} function is intended for presenting this |
| 13734 | kind of information to the user. The message has to be phrased so |
| 13735 | that it reads properly regardless of how many words there are in the |
| 13736 | region: we don't want to say that ``there are 1 words in the region''. |
| 13737 | The conflict between singular and plural is ungrammatical. We can |
| 13738 | solve this problem by using a conditional expression that evaluates |
| 13739 | different messages depending on the number of words in the region. |
| 13740 | There are three possibilities: no words in the region, one word in the |
| 13741 | region, and more than one word. This means that the @code{cond} |
| 13742 | special form is appropriate. |
| 13743 | |
| 13744 | @need 1500 |
| 13745 | All this leads to the following function definition: |
| 13746 | |
| 13747 | @smallexample |
| 13748 | @group |
| 13749 | ;;; @r{First version; has bugs!} |
| 13750 | (defun @value{COUNT-WORDS} (beginning end) |
| 13751 | "Print number of words in the region. |
| 13752 | Words are defined as at least one word-constituent |
| 13753 | character followed by at least one character that |
| 13754 | is not a word-constituent. The buffer's syntax |
| 13755 | table determines which characters these are." |
| 13756 | (interactive "r") |
| 13757 | (message "Counting words in region ... ") |
| 13758 | @end group |
| 13759 | |
| 13760 | @group |
| 13761 | ;;; @r{1. Set up appropriate conditions.} |
| 13762 | (save-excursion |
| 13763 | (goto-char beginning) |
| 13764 | (let ((count 0)) |
| 13765 | @end group |
| 13766 | |
| 13767 | @group |
| 13768 | ;;; @r{2. Run the} while @r{loop.} |
| 13769 | (while (< (point) end) |
| 13770 | (re-search-forward "\\w+\\W*") |
| 13771 | (setq count (1+ count))) |
| 13772 | @end group |
| 13773 | |
| 13774 | @group |
| 13775 | ;;; @r{3. Send a message to the user.} |
| 13776 | (cond ((zerop count) |
| 13777 | (message |
| 13778 | "The region does NOT have any words.")) |
| 13779 | ((= 1 count) |
| 13780 | (message |
| 13781 | "The region has 1 word.")) |
| 13782 | (t |
| 13783 | (message |
| 13784 | "The region has %d words." count)))))) |
| 13785 | @end group |
| 13786 | @end smallexample |
| 13787 | |
| 13788 | @noindent |
| 13789 | As written, the function works, but not in all circumstances. |
| 13790 | |
| 13791 | @node Whitespace Bug |
| 13792 | @subsection The Whitespace Bug in @code{@value{COUNT-WORDS}} |
| 13793 | |
| 13794 | The @code{@value{COUNT-WORDS}} command described in the preceding |
| 13795 | section has two bugs, or rather, one bug with two manifestations. |
| 13796 | First, if you mark a region containing only whitespace in the middle |
| 13797 | of some text, the @code{@value{COUNT-WORDS}} command tells you that the |
| 13798 | region contains one word! Second, if you mark a region containing |
| 13799 | only whitespace at the end of the buffer or the accessible portion of |
| 13800 | a narrowed buffer, the command displays an error message that looks |
| 13801 | like this: |
| 13802 | |
| 13803 | @smallexample |
| 13804 | Search failed: "\\w+\\W*" |
| 13805 | @end smallexample |
| 13806 | |
| 13807 | If you are reading this in Info in GNU Emacs, you can test for these |
| 13808 | bugs yourself. |
| 13809 | |
| 13810 | First, evaluate the function in the usual manner to install it. |
| 13811 | @ifinfo |
| 13812 | Here is a copy of the definition. Place your cursor after the closing |
| 13813 | parenthesis and type @kbd{C-x C-e} to install it. |
| 13814 | |
| 13815 | @smallexample |
| 13816 | @group |
| 13817 | ;; @r{First version; has bugs!} |
| 13818 | (defun @value{COUNT-WORDS} (beginning end) |
| 13819 | "Print number of words in the region. |
| 13820 | Words are defined as at least one word-constituent character followed |
| 13821 | by at least one character that is not a word-constituent. The buffer's |
| 13822 | syntax table determines which characters these are." |
| 13823 | @end group |
| 13824 | @group |
| 13825 | (interactive "r") |
| 13826 | (message "Counting words in region ... ") |
| 13827 | @end group |
| 13828 | |
| 13829 | @group |
| 13830 | ;;; @r{1. Set up appropriate conditions.} |
| 13831 | (save-excursion |
| 13832 | (goto-char beginning) |
| 13833 | (let ((count 0)) |
| 13834 | @end group |
| 13835 | |
| 13836 | @group |
| 13837 | ;;; @r{2. Run the} while @r{loop.} |
| 13838 | (while (< (point) end) |
| 13839 | (re-search-forward "\\w+\\W*") |
| 13840 | (setq count (1+ count))) |
| 13841 | @end group |
| 13842 | |
| 13843 | @group |
| 13844 | ;;; @r{3. Send a message to the user.} |
| 13845 | (cond ((zerop count) |
| 13846 | (message "The region does NOT have any words.")) |
| 13847 | ((= 1 count) (message "The region has 1 word.")) |
| 13848 | (t (message "The region has %d words." count)))))) |
| 13849 | @end group |
| 13850 | @end smallexample |
| 13851 | @end ifinfo |
| 13852 | |
| 13853 | @need 1000 |
| 13854 | If you wish, you can also install this keybinding by evaluating it: |
| 13855 | |
| 13856 | @smallexample |
| 13857 | (global-set-key "\C-c=" '@value{COUNT-WORDS}) |
| 13858 | @end smallexample |
| 13859 | |
| 13860 | To conduct the first test, set mark and point to the beginning and end |
| 13861 | of the following line and then type @kbd{C-c =} (or @kbd{M-x |
| 13862 | @value{COUNT-WORDS}} if you have not bound @kbd{C-c =}): |
| 13863 | |
| 13864 | @smallexample |
| 13865 | one two three |
| 13866 | @end smallexample |
| 13867 | |
| 13868 | @noindent |
| 13869 | Emacs will tell you, correctly, that the region has three words. |
| 13870 | |
| 13871 | Repeat the test, but place mark at the beginning of the line and place |
| 13872 | point just @emph{before} the word @samp{one}. Again type the command |
| 13873 | @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}). Emacs should tell you |
| 13874 | that the region has no words, since it is composed only of the |
| 13875 | whitespace at the beginning of the line. But instead Emacs tells you |
| 13876 | that the region has one word! |
| 13877 | |
| 13878 | For the third test, copy the sample line to the end of the |
| 13879 | @file{*scratch*} buffer and then type several spaces at the end of the |
| 13880 | line. Place mark right after the word @samp{three} and point at the |
| 13881 | end of line. (The end of the line will be the end of the buffer.) |
| 13882 | Type @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}) as you did before. |
| 13883 | Again, Emacs should tell you that the region has no words, since it is |
| 13884 | composed only of the whitespace at the end of the line. Instead, |
| 13885 | Emacs displays an error message saying @samp{Search failed}. |
| 13886 | |
| 13887 | The two bugs stem from the same problem. |
| 13888 | |
| 13889 | Consider the first manifestation of the bug, in which the command |
| 13890 | tells you that the whitespace at the beginning of the line contains |
| 13891 | one word. What happens is this: The @code{M-x @value{COUNT-WORDS}} |
| 13892 | command moves point to the beginning of the region. The @code{while} |
| 13893 | tests whether the value of point is smaller than the value of |
| 13894 | @code{end}, which it is. Consequently, the regular expression search |
| 13895 | looks for and finds the first word. It leaves point after the word. |
| 13896 | @code{count} is set to one. The @code{while} loop repeats; but this |
| 13897 | time the value of point is larger than the value of @code{end}, the |
| 13898 | loop is exited; and the function displays a message saying the number |
| 13899 | of words in the region is one. In brief, the regular expression |
| 13900 | search looks for and finds the word even though it is outside |
| 13901 | the marked region. |
| 13902 | |
| 13903 | In the second manifestation of the bug, the region is whitespace at |
| 13904 | the end of the buffer. Emacs says @samp{Search failed}. What happens |
| 13905 | is that the true-or-false-test in the @code{while} loop tests true, so |
| 13906 | the search expression is executed. But since there are no more words |
| 13907 | in the buffer, the search fails. |
| 13908 | |
| 13909 | In both manifestations of the bug, the search extends or attempts to |
| 13910 | extend outside of the region. |
| 13911 | |
| 13912 | The solution is to limit the search to the region---this is a fairly |
| 13913 | simple action, but as you may have come to expect, it is not quite as |
| 13914 | simple as you might think. |
| 13915 | |
| 13916 | As we have seen, the @code{re-search-forward} function takes a search |
| 13917 | pattern as its first argument. But in addition to this first, |
| 13918 | mandatory argument, it accepts three optional arguments. The optional |
| 13919 | second argument bounds the search. The optional third argument, if |
| 13920 | @code{t}, causes the function to return @code{nil} rather than signal |
| 13921 | an error if the search fails. The optional fourth argument is a |
| 13922 | repeat count. (In Emacs, you can see a function's documentation by |
| 13923 | typing @kbd{C-h f}, the name of the function, and then @key{RET}.) |
| 13924 | |
| 13925 | In the @code{@value{COUNT-WORDS}} definition, the value of the end of |
| 13926 | the region is held by the variable @code{end} which is passed as an |
| 13927 | argument to the function. Thus, we can add @code{end} as an argument |
| 13928 | to the regular expression search expression: |
| 13929 | |
| 13930 | @smallexample |
| 13931 | (re-search-forward "\\w+\\W*" end) |
| 13932 | @end smallexample |
| 13933 | |
| 13934 | However, if you make only this change to the @code{@value{COUNT-WORDS}} |
| 13935 | definition and then test the new version of the definition on a |
| 13936 | stretch of whitespace, you will receive an error message saying |
| 13937 | @samp{Search failed}. |
| 13938 | |
| 13939 | What happens is this: the search is limited to the region, and fails |
| 13940 | as you expect because there are no word-constituent characters in the |
| 13941 | region. Since it fails, we receive an error message. But we do not |
| 13942 | want to receive an error message in this case; we want to receive the |
| 13943 | message that "The region does NOT have any words." |
| 13944 | |
| 13945 | The solution to this problem is to provide @code{re-search-forward} |
| 13946 | with a third argument of @code{t}, which causes the function to return |
| 13947 | @code{nil} rather than signal an error if the search fails. |
| 13948 | |
| 13949 | However, if you make this change and try it, you will see the message |
| 13950 | ``Counting words in region ... '' and @dots{} you will keep on seeing |
| 13951 | that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}). |
| 13952 | |
| 13953 | Here is what happens: the search is limited to the region, as before, |
| 13954 | and it fails because there are no word-constituent characters in the |
| 13955 | region, as expected. Consequently, the @code{re-search-forward} |
| 13956 | expression returns @code{nil}. It does nothing else. In particular, |
| 13957 | it does not move point, which it does as a side effect if it finds the |
| 13958 | search target. After the @code{re-search-forward} expression returns |
| 13959 | @code{nil}, the next expression in the @code{while} loop is evaluated. |
| 13960 | This expression increments the count. Then the loop repeats. The |
| 13961 | true-or-false-test tests true because the value of point is still less |
| 13962 | than the value of end, since the @code{re-search-forward} expression |
| 13963 | did not move point. @dots{} and the cycle repeats @dots{} |
| 13964 | |
| 13965 | The @code{@value{COUNT-WORDS}} definition requires yet another |
| 13966 | modification, to cause the true-or-false-test of the @code{while} loop |
| 13967 | to test false if the search fails. Put another way, there are two |
| 13968 | conditions that must be satisfied in the true-or-false-test before the |
| 13969 | word count variable is incremented: point must still be within the |
| 13970 | region and the search expression must have found a word to count. |
| 13971 | |
| 13972 | Since both the first condition and the second condition must be true |
| 13973 | together, the two expressions, the region test and the search |
| 13974 | expression, can be joined with an @code{and} special form and embedded in |
| 13975 | the @code{while} loop as the true-or-false-test, like this: |
| 13976 | |
| 13977 | @smallexample |
| 13978 | (and (< (point) end) (re-search-forward "\\w+\\W*" end t)) |
| 13979 | @end smallexample |
| 13980 | |
| 13981 | @c colon in printed section title causes problem in Info cross reference |
| 13982 | @c also trouble with an overfull hbox |
| 13983 | @iftex |
| 13984 | @noindent |
| 13985 | (For information about @code{and}, see |
| 13986 | @ref{kill-new function, , The @code{kill-new} function}.) |
| 13987 | @end iftex |
| 13988 | @ifinfo |
| 13989 | @noindent |
| 13990 | (@xref{kill-new function, , The @code{kill-new} function}, for |
| 13991 | information about @code{and}.) |
| 13992 | @end ifinfo |
| 13993 | |
| 13994 | The @code{re-search-forward} expression returns @code{t} if the search |
| 13995 | succeeds and as a side effect moves point. Consequently, as words are |
| 13996 | found, point is moved through the region. When the search expression |
| 13997 | fails to find another word, or when point reaches the end of the |
| 13998 | region, the true-or-false-test tests false, the @code{while} loop |
| 13999 | exits, and the @code{@value{COUNT-WORDS}} function displays one or |
| 14000 | other of its messages. |
| 14001 | |
| 14002 | After incorporating these final changes, the @code{@value{COUNT-WORDS}} |
| 14003 | works without bugs (or at least, without bugs that I have found!). |
| 14004 | Here is what it looks like: |
| 14005 | |
| 14006 | @smallexample |
| 14007 | @group |
| 14008 | ;;; @r{Final version:} @code{while} |
| 14009 | (defun @value{COUNT-WORDS} (beginning end) |
| 14010 | "Print number of words in the region." |
| 14011 | (interactive "r") |
| 14012 | (message "Counting words in region ... ") |
| 14013 | @end group |
| 14014 | |
| 14015 | @group |
| 14016 | ;;; @r{1. Set up appropriate conditions.} |
| 14017 | (save-excursion |
| 14018 | (let ((count 0)) |
| 14019 | (goto-char beginning) |
| 14020 | @end group |
| 14021 | |
| 14022 | @group |
| 14023 | ;;; @r{2. Run the} while @r{loop.} |
| 14024 | (while (and (< (point) end) |
| 14025 | (re-search-forward "\\w+\\W*" end t)) |
| 14026 | (setq count (1+ count))) |
| 14027 | @end group |
| 14028 | |
| 14029 | @group |
| 14030 | ;;; @r{3. Send a message to the user.} |
| 14031 | (cond ((zerop count) |
| 14032 | (message |
| 14033 | "The region does NOT have any words.")) |
| 14034 | ((= 1 count) |
| 14035 | (message |
| 14036 | "The region has 1 word.")) |
| 14037 | (t |
| 14038 | (message |
| 14039 | "The region has %d words." count)))))) |
| 14040 | @end group |
| 14041 | @end smallexample |
| 14042 | |
| 14043 | @node recursive-count-words |
| 14044 | @section Count Words Recursively |
| 14045 | @cindex Count words recursively |
| 14046 | @cindex Recursively counting words |
| 14047 | @cindex Words, counted recursively |
| 14048 | |
| 14049 | You can write the function for counting words recursively as well as |
| 14050 | with a @code{while} loop. Let's see how this is done. |
| 14051 | |
| 14052 | First, we need to recognize that the @code{@value{COUNT-WORDS}} |
| 14053 | function has three jobs: it sets up the appropriate conditions for |
| 14054 | counting to occur; it counts the words in the region; and it sends a |
| 14055 | message to the user telling how many words there are. |
| 14056 | |
| 14057 | If we write a single recursive function to do everything, we will |
| 14058 | receive a message for every recursive call. If the region contains 13 |
| 14059 | words, we will receive thirteen messages, one right after the other. |
| 14060 | We don't want this! Instead, we must write two functions to do the |
| 14061 | job, one of which (the recursive function) will be used inside of the |
| 14062 | other. One function will set up the conditions and display the |
| 14063 | message; the other will return the word count. |
| 14064 | |
| 14065 | Let us start with the function that causes the message to be displayed. |
| 14066 | We can continue to call this @code{@value{COUNT-WORDS}}. |
| 14067 | |
| 14068 | This is the function that the user will call. It will be interactive. |
| 14069 | Indeed, it will be similar to our previous versions of this |
| 14070 | function, except that it will call @code{recursive-count-words} to |
| 14071 | determine how many words are in the region. |
| 14072 | |
| 14073 | @need 1250 |
| 14074 | We can readily construct a template for this function, based on our |
| 14075 | previous versions: |
| 14076 | |
| 14077 | @smallexample |
| 14078 | @group |
| 14079 | ;; @r{Recursive version; uses regular expression search} |
| 14080 | (defun @value{COUNT-WORDS} (beginning end) |
| 14081 | "@var{documentation}@dots{}" |
| 14082 | (@var{interactive-expression}@dots{}) |
| 14083 | @end group |
| 14084 | @group |
| 14085 | |
| 14086 | ;;; @r{1. Set up appropriate conditions.} |
| 14087 | (@var{explanatory message}) |
| 14088 | (@var{set-up functions}@dots{} |
| 14089 | @end group |
| 14090 | @group |
| 14091 | |
| 14092 | ;;; @r{2. Count the words.} |
| 14093 | @var{recursive call} |
| 14094 | @end group |
| 14095 | @group |
| 14096 | |
| 14097 | ;;; @r{3. Send a message to the user.} |
| 14098 | @var{message providing word count})) |
| 14099 | @end group |
| 14100 | @end smallexample |
| 14101 | |
| 14102 | The definition looks straightforward, except that somehow the count |
| 14103 | returned by the recursive call must be passed to the message |
| 14104 | displaying the word count. A little thought suggests that this can be |
| 14105 | done by making use of a @code{let} expression: we can bind a variable |
| 14106 | in the varlist of a @code{let} expression to the number of words in |
| 14107 | the region, as returned by the recursive call; and then the |
| 14108 | @code{cond} expression, using binding, can display the value to the |
| 14109 | user. |
| 14110 | |
| 14111 | Often, one thinks of the binding within a @code{let} expression as |
| 14112 | somehow secondary to the `primary' work of a function. But in this |
| 14113 | case, what you might consider the `primary' job of the function, |
| 14114 | counting words, is done within the @code{let} expression. |
| 14115 | |
| 14116 | @need 1250 |
| 14117 | Using @code{let}, the function definition looks like this: |
| 14118 | |
| 14119 | @smallexample |
| 14120 | @group |
| 14121 | (defun @value{COUNT-WORDS} (beginning end) |
| 14122 | "Print number of words in the region." |
| 14123 | (interactive "r") |
| 14124 | @end group |
| 14125 | |
| 14126 | @group |
| 14127 | ;;; @r{1. Set up appropriate conditions.} |
| 14128 | (message "Counting words in region ... ") |
| 14129 | (save-excursion |
| 14130 | (goto-char beginning) |
| 14131 | @end group |
| 14132 | |
| 14133 | @group |
| 14134 | ;;; @r{2. Count the words.} |
| 14135 | (let ((count (recursive-count-words end))) |
| 14136 | @end group |
| 14137 | |
| 14138 | @group |
| 14139 | ;;; @r{3. Send a message to the user.} |
| 14140 | (cond ((zerop count) |
| 14141 | (message |
| 14142 | "The region does NOT have any words.")) |
| 14143 | ((= 1 count) |
| 14144 | (message |
| 14145 | "The region has 1 word.")) |
| 14146 | (t |
| 14147 | (message |
| 14148 | "The region has %d words." count)))))) |
| 14149 | @end group |
| 14150 | @end smallexample |
| 14151 | |
| 14152 | Next, we need to write the recursive counting function. |
| 14153 | |
| 14154 | A recursive function has at least three parts: the `do-again-test', the |
| 14155 | `next-step-expression', and the recursive call. |
| 14156 | |
| 14157 | The do-again-test determines whether the function will or will not be |
| 14158 | called again. Since we are counting words in a region and can use a |
| 14159 | function that moves point forward for every word, the do-again-test |
| 14160 | can check whether point is still within the region. The do-again-test |
| 14161 | should find the value of point and determine whether point is before, |
| 14162 | at, or after the value of the end of the region. We can use the |
| 14163 | @code{point} function to locate point. Clearly, we must pass the |
| 14164 | value of the end of the region to the recursive counting function as an |
| 14165 | argument. |
| 14166 | |
| 14167 | In addition, the do-again-test should also test whether the search finds a |
| 14168 | word. If it does not, the function should not call itself again. |
| 14169 | |
| 14170 | The next-step-expression changes a value so that when the recursive |
| 14171 | function is supposed to stop calling itself, it stops. More |
| 14172 | precisely, the next-step-expression changes a value so that at the |
| 14173 | right time, the do-again-test stops the recursive function from |
| 14174 | calling itself again. In this case, the next-step-expression can be |
| 14175 | the expression that moves point forward, word by word. |
| 14176 | |
| 14177 | The third part of a recursive function is the recursive call. |
| 14178 | |
| 14179 | Somewhere, also, we also need a part that does the `work' of the |
| 14180 | function, a part that does the counting. A vital part! |
| 14181 | |
| 14182 | @need 1250 |
| 14183 | But already, we have an outline of the recursive counting function: |
| 14184 | |
| 14185 | @smallexample |
| 14186 | @group |
| 14187 | (defun recursive-count-words (region-end) |
| 14188 | "@var{documentation}@dots{}" |
| 14189 | @var{do-again-test} |
| 14190 | @var{next-step-expression} |
| 14191 | @var{recursive call}) |
| 14192 | @end group |
| 14193 | @end smallexample |
| 14194 | |
| 14195 | Now we need to fill in the slots. Let's start with the simplest cases |
| 14196 | first: if point is at or beyond the end of the region, there cannot |
| 14197 | be any words in the region, so the function should return zero. |
| 14198 | Likewise, if the search fails, there are no words to count, so the |
| 14199 | function should return zero. |
| 14200 | |
| 14201 | On the other hand, if point is within the region and the search |
| 14202 | succeeds, the function should call itself again. |
| 14203 | |
| 14204 | @need 800 |
| 14205 | Thus, the do-again-test should look like this: |
| 14206 | |
| 14207 | @smallexample |
| 14208 | @group |
| 14209 | (and (< (point) region-end) |
| 14210 | (re-search-forward "\\w+\\W*" region-end t)) |
| 14211 | @end group |
| 14212 | @end smallexample |
| 14213 | |
| 14214 | Note that the search expression is part of the do-again-test---the |
| 14215 | function returns @code{t} if its search succeeds and @code{nil} if it |
| 14216 | fails. (@xref{Whitespace Bug, , The Whitespace Bug in |
| 14217 | @code{@value{COUNT-WORDS}}}, for an explanation of how |
| 14218 | @code{re-search-forward} works.) |
| 14219 | |
| 14220 | The do-again-test is the true-or-false test of an @code{if} clause. |
| 14221 | Clearly, if the do-again-test succeeds, the then-part of the @code{if} |
| 14222 | clause should call the function again; but if it fails, the else-part |
| 14223 | should return zero since either point is outside the region or the |
| 14224 | search failed because there were no words to find. |
| 14225 | |
| 14226 | But before considering the recursive call, we need to consider the |
| 14227 | next-step-expression. What is it? Interestingly, it is the search |
| 14228 | part of the do-again-test. |
| 14229 | |
| 14230 | In addition to returning @code{t} or @code{nil} for the |
| 14231 | do-again-test, @code{re-search-forward} moves point forward as a side |
| 14232 | effect of a successful search. This is the action that changes the |
| 14233 | value of point so that the recursive function stops calling itself |
| 14234 | when point completes its movement through the region. Consequently, |
| 14235 | the @code{re-search-forward} expression is the next-step-expression. |
| 14236 | |
| 14237 | @need 1200 |
| 14238 | In outline, then, the body of the @code{recursive-count-words} |
| 14239 | function looks like this: |
| 14240 | |
| 14241 | @smallexample |
| 14242 | @group |
| 14243 | (if @var{do-again-test-and-next-step-combined} |
| 14244 | ;; @r{then} |
| 14245 | @var{recursive-call-returning-count} |
| 14246 | ;; @r{else} |
| 14247 | @var{return-zero}) |
| 14248 | @end group |
| 14249 | @end smallexample |
| 14250 | |
| 14251 | How to incorporate the mechanism that counts? |
| 14252 | |
| 14253 | If you are not used to writing recursive functions, a question like |
| 14254 | this can be troublesome. But it can and should be approached |
| 14255 | systematically. |
| 14256 | |
| 14257 | We know that the counting mechanism should be associated in some way |
| 14258 | with the recursive call. Indeed, since the next-step-expression moves |
| 14259 | point forward by one word, and since a recursive call is made for |
| 14260 | each word, the counting mechanism must be an expression that adds one |
| 14261 | to the value returned by a call to @code{recursive-count-words}. |
| 14262 | |
| 14263 | @need 800 |
| 14264 | Consider several cases: |
| 14265 | |
| 14266 | @itemize @bullet |
| 14267 | @item |
| 14268 | If there are two words in the region, the function should return |
| 14269 | a value resulting from adding one to the value returned when it counts |
| 14270 | the first word, plus the number returned when it counts the remaining |
| 14271 | words in the region, which in this case is one. |
| 14272 | |
| 14273 | @item |
| 14274 | If there is one word in the region, the function should return |
| 14275 | a value resulting from adding one to the value returned when it counts |
| 14276 | that word, plus the number returned when it counts the remaining |
| 14277 | words in the region, which in this case is zero. |
| 14278 | |
| 14279 | @item |
| 14280 | If there are no words in the region, the function should return zero. |
| 14281 | @end itemize |
| 14282 | |
| 14283 | From the sketch we can see that the else-part of the @code{if} returns |
| 14284 | zero for the case of no words. This means that the then-part of the |
| 14285 | @code{if} must return a value resulting from adding one to the value |
| 14286 | returned from a count of the remaining words. |
| 14287 | |
| 14288 | @need 1200 |
| 14289 | The expression will look like this, where @code{1+} is a function that |
| 14290 | adds one to its argument. |
| 14291 | |
| 14292 | @smallexample |
| 14293 | (1+ (recursive-count-words region-end)) |
| 14294 | @end smallexample |
| 14295 | |
| 14296 | @need 1200 |
| 14297 | The whole @code{recursive-count-words} function will then look like |
| 14298 | this: |
| 14299 | |
| 14300 | @smallexample |
| 14301 | @group |
| 14302 | (defun recursive-count-words (region-end) |
| 14303 | "@var{documentation}@dots{}" |
| 14304 | |
| 14305 | ;;; @r{1. do-again-test} |
| 14306 | (if (and (< (point) region-end) |
| 14307 | (re-search-forward "\\w+\\W*" region-end t)) |
| 14308 | @end group |
| 14309 | |
| 14310 | @group |
| 14311 | ;;; @r{2. then-part: the recursive call} |
| 14312 | (1+ (recursive-count-words region-end)) |
| 14313 | |
| 14314 | ;;; @r{3. else-part} |
| 14315 | 0)) |
| 14316 | @end group |
| 14317 | @end smallexample |
| 14318 | |
| 14319 | @need 1250 |
| 14320 | Let's examine how this works: |
| 14321 | |
| 14322 | If there are no words in the region, the else part of the @code{if} |
| 14323 | expression is evaluated and consequently the function returns zero. |
| 14324 | |
| 14325 | If there is one word in the region, the value of point is less than |
| 14326 | the value of @code{region-end} and the search succeeds. In this case, |
| 14327 | the true-or-false-test of the @code{if} expression tests true, and the |
| 14328 | then-part of the @code{if} expression is evaluated. The counting |
| 14329 | expression is evaluated. This expression returns a value (which will |
| 14330 | be the value returned by the whole function) that is the sum of one |
| 14331 | added to the value returned by a recursive call. |
| 14332 | |
| 14333 | Meanwhile, the next-step-expression has caused point to jump over the |
| 14334 | first (and in this case only) word in the region. This means that |
| 14335 | when @code{(recursive-count-words region-end)} is evaluated a second |
| 14336 | time, as a result of the recursive call, the value of point will be |
| 14337 | equal to or greater than the value of region end. So this time, |
| 14338 | @code{recursive-count-words} will return zero. The zero will be added |
| 14339 | to one, and the original evaluation of @code{recursive-count-words} |
| 14340 | will return one plus zero, which is one, which is the correct amount. |
| 14341 | |
| 14342 | Clearly, if there are two words in the region, the first call to |
| 14343 | @code{recursive-count-words} returns one added to the value returned |
| 14344 | by calling @code{recursive-count-words} on a region containing the |
| 14345 | remaining word---that is, it adds one to one, producing two, which is |
| 14346 | the correct amount. |
| 14347 | |
| 14348 | Similarly, if there are three words in the region, the first call to |
| 14349 | @code{recursive-count-words} returns one added to the value returned |
| 14350 | by calling @code{recursive-count-words} on a region containing the |
| 14351 | remaining two words---and so on and so on. |
| 14352 | |
| 14353 | @need 1250 |
| 14354 | @noindent |
| 14355 | With full documentation the two functions look like this: |
| 14356 | |
| 14357 | @need 1250 |
| 14358 | @noindent |
| 14359 | The recursive function: |
| 14360 | |
| 14361 | @findex recursive-count-words |
| 14362 | @smallexample |
| 14363 | @group |
| 14364 | (defun recursive-count-words (region-end) |
| 14365 | "Number of words between point and REGION-END." |
| 14366 | @end group |
| 14367 | |
| 14368 | @group |
| 14369 | ;;; @r{1. do-again-test} |
| 14370 | (if (and (< (point) region-end) |
| 14371 | (re-search-forward "\\w+\\W*" region-end t)) |
| 14372 | @end group |
| 14373 | |
| 14374 | @group |
| 14375 | ;;; @r{2. then-part: the recursive call} |
| 14376 | (1+ (recursive-count-words region-end)) |
| 14377 | |
| 14378 | ;;; @r{3. else-part} |
| 14379 | 0)) |
| 14380 | @end group |
| 14381 | @end smallexample |
| 14382 | |
| 14383 | @need 800 |
| 14384 | @noindent |
| 14385 | The wrapper: |
| 14386 | |
| 14387 | @smallexample |
| 14388 | @group |
| 14389 | ;;; @r{Recursive version} |
| 14390 | (defun @value{COUNT-WORDS} (beginning end) |
| 14391 | "Print number of words in the region. |
| 14392 | @end group |
| 14393 | |
| 14394 | @group |
| 14395 | Words are defined as at least one word-constituent |
| 14396 | character followed by at least one character that is |
| 14397 | not a word-constituent. The buffer's syntax table |
| 14398 | determines which characters these are." |
| 14399 | @end group |
| 14400 | @group |
| 14401 | (interactive "r") |
| 14402 | (message "Counting words in region ... ") |
| 14403 | (save-excursion |
| 14404 | (goto-char beginning) |
| 14405 | (let ((count (recursive-count-words end))) |
| 14406 | @end group |
| 14407 | @group |
| 14408 | (cond ((zerop count) |
| 14409 | (message |
| 14410 | "The region does NOT have any words.")) |
| 14411 | @end group |
| 14412 | @group |
| 14413 | ((= 1 count) |
| 14414 | (message "The region has 1 word.")) |
| 14415 | (t |
| 14416 | (message |
| 14417 | "The region has %d words." count)))))) |
| 14418 | @end group |
| 14419 | @end smallexample |
| 14420 | |
| 14421 | @node Counting Exercise |
| 14422 | @section Exercise: Counting Punctuation |
| 14423 | |
| 14424 | Using a @code{while} loop, write a function to count the number of |
| 14425 | punctuation marks in a region---period, comma, semicolon, colon, |
| 14426 | exclamation mark, and question mark. Do the same using recursion. |
| 14427 | |
| 14428 | @node Words in a defun |
| 14429 | @chapter Counting Words in a @code{defun} |
| 14430 | @cindex Counting words in a @code{defun} |
| 14431 | @cindex Word counting in a @code{defun} |
| 14432 | |
| 14433 | Our next project is to count the number of words in a function |
| 14434 | definition. Clearly, this can be done using some variant of |
| 14435 | @code{@value{COUNT-WORDS}}. @xref{Counting Words, , Counting Words: |
| 14436 | Repetition and Regexps}. If we are just going to count the words in |
| 14437 | one definition, it is easy enough to mark the definition with the |
| 14438 | @kbd{C-M-h} (@code{mark-defun}) command, and then call |
| 14439 | @code{@value{COUNT-WORDS}}. |
| 14440 | |
| 14441 | However, I am more ambitious: I want to count the words and symbols in |
| 14442 | every definition in the Emacs sources and then print a graph that |
| 14443 | shows how many functions there are of each length: how many contain 40 |
| 14444 | to 49 words or symbols, how many contain 50 to 59 words or symbols, |
| 14445 | and so on. I have often been curious how long a typical function is, |
| 14446 | and this will tell. |
| 14447 | |
| 14448 | @menu |
| 14449 | * Divide and Conquer:: |
| 14450 | * Words and Symbols:: What to count? |
| 14451 | * Syntax:: What constitutes a word or symbol? |
| 14452 | * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}. |
| 14453 | * Several defuns:: Counting several defuns in a file. |
| 14454 | * Find a File:: Do you want to look at a file? |
| 14455 | * lengths-list-file:: A list of the lengths of many definitions. |
| 14456 | * Several files:: Counting in definitions in different files. |
| 14457 | * Several files recursively:: Recursively counting in different files. |
| 14458 | * Prepare the data:: Prepare the data for display in a graph. |
| 14459 | @end menu |
| 14460 | |
| 14461 | @ifnottex |
| 14462 | @node Divide and Conquer |
| 14463 | @unnumberedsec Divide and Conquer |
| 14464 | @end ifnottex |
| 14465 | |
| 14466 | Described in one phrase, the histogram project is daunting; but |
| 14467 | divided into numerous small steps, each of which we can take one at a |
| 14468 | time, the project becomes less fearsome. Let us consider what the |
| 14469 | steps must be: |
| 14470 | |
| 14471 | @itemize @bullet |
| 14472 | @item |
| 14473 | First, write a function to count the words in one definition. This |
| 14474 | includes the problem of handling symbols as well as words. |
| 14475 | |
| 14476 | @item |
| 14477 | Second, write a function to list the numbers of words in each function |
| 14478 | in a file. This function can use the @code{count-words-in-defun} |
| 14479 | function. |
| 14480 | |
| 14481 | @item |
| 14482 | Third, write a function to list the numbers of words in each function |
| 14483 | in each of several files. This entails automatically finding the |
| 14484 | various files, switching to them, and counting the words in the |
| 14485 | definitions within them. |
| 14486 | |
| 14487 | @item |
| 14488 | Fourth, write a function to convert the list of numbers that we |
| 14489 | created in step three to a form that will be suitable for printing as |
| 14490 | a graph. |
| 14491 | |
| 14492 | @item |
| 14493 | Fifth, write a function to print the results as a graph. |
| 14494 | @end itemize |
| 14495 | |
| 14496 | This is quite a project! But if we take each step slowly, it will not |
| 14497 | be difficult. |
| 14498 | |
| 14499 | @node Words and Symbols |
| 14500 | @section What to Count? |
| 14501 | @cindex Words and symbols in defun |
| 14502 | |
| 14503 | When we first start thinking about how to count the words in a |
| 14504 | function definition, the first question is (or ought to be) what are |
| 14505 | we going to count? When we speak of `words' with respect to a Lisp |
| 14506 | function definition, we are actually speaking, in large part, of |
| 14507 | `symbols'. For example, the following @code{multiply-by-seven} |
| 14508 | function contains the five symbols @code{defun}, |
| 14509 | @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In |
| 14510 | addition, in the documentation string, it contains the four words |
| 14511 | @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The |
| 14512 | symbol @samp{number} is repeated, so the definition contains a total |
| 14513 | of ten words and symbols. |
| 14514 | |
| 14515 | @smallexample |
| 14516 | @group |
| 14517 | (defun multiply-by-seven (number) |
| 14518 | "Multiply NUMBER by seven." |
| 14519 | (* 7 number)) |
| 14520 | @end group |
| 14521 | @end smallexample |
| 14522 | |
| 14523 | @noindent |
| 14524 | However, if we mark the @code{multiply-by-seven} definition with |
| 14525 | @kbd{C-M-h} (@code{mark-defun}), and then call |
| 14526 | @code{@value{COUNT-WORDS}} on it, we will find that |
| 14527 | @code{@value{COUNT-WORDS}} claims the definition has eleven words, not |
| 14528 | ten! Something is wrong! |
| 14529 | |
| 14530 | The problem is twofold: @code{@value{COUNT-WORDS}} does not count the |
| 14531 | @samp{*} as a word, and it counts the single symbol, |
| 14532 | @code{multiply-by-seven}, as containing three words. The hyphens are |
| 14533 | treated as if they were interword spaces rather than intraword |
| 14534 | connectors: @samp{multiply-by-seven} is counted as if it were written |
| 14535 | @samp{multiply by seven}. |
| 14536 | |
| 14537 | The cause of this confusion is the regular expression search within |
| 14538 | the @code{@value{COUNT-WORDS}} definition that moves point forward word |
| 14539 | by word. In the canonical version of @code{@value{COUNT-WORDS}}, the |
| 14540 | regexp is: |
| 14541 | |
| 14542 | @smallexample |
| 14543 | "\\w+\\W*" |
| 14544 | @end smallexample |
| 14545 | |
| 14546 | @noindent |
| 14547 | This regular expression is a pattern defining one or more word |
| 14548 | constituent characters possibly followed by one or more characters |
| 14549 | that are not word constituents. What is meant by `word constituent |
| 14550 | characters' brings us to the issue of syntax, which is worth a section |
| 14551 | of its own. |
| 14552 | |
| 14553 | @node Syntax |
| 14554 | @section What Constitutes a Word or Symbol? |
| 14555 | @cindex Syntax categories and tables |
| 14556 | |
| 14557 | Emacs treats different characters as belonging to different |
| 14558 | @dfn{syntax categories}. For example, the regular expression, |
| 14559 | @samp{\\w+}, is a pattern specifying one or more @emph{word |
| 14560 | constituent} characters. Word constituent characters are members of |
| 14561 | one syntax category. Other syntax categories include the class of |
| 14562 | punctuation characters, such as the period and the comma, and the |
| 14563 | class of whitespace characters, such as the blank space and the tab |
| 14564 | character. (For more information, @pxref{Syntax Tables, , Syntax |
| 14565 | Tables, elisp, The GNU Emacs Lisp Reference Manual}.) |
| 14566 | |
| 14567 | Syntax tables specify which characters belong to which categories. |
| 14568 | Usually, a hyphen is not specified as a `word constituent character'. |
| 14569 | Instead, it is specified as being in the `class of characters that are |
| 14570 | part of symbol names but not words.' This means that the |
| 14571 | @code{@value{COUNT-WORDS}} function treats it in the same way it treats |
| 14572 | an interword white space, which is why @code{@value{COUNT-WORDS}} |
| 14573 | counts @samp{multiply-by-seven} as three words. |
| 14574 | |
| 14575 | There are two ways to cause Emacs to count @samp{multiply-by-seven} as |
| 14576 | one symbol: modify the syntax table or modify the regular expression. |
| 14577 | |
| 14578 | We could redefine a hyphen as a word constituent character by |
| 14579 | modifying the syntax table that Emacs keeps for each mode. This |
| 14580 | action would serve our purpose, except that a hyphen is merely the |
| 14581 | most common character within symbols that is not typically a word |
| 14582 | constituent character; there are others, too. |
| 14583 | |
| 14584 | Alternatively, we can redefine the regexp used in the |
| 14585 | @code{@value{COUNT-WORDS}} definition so as to include symbols. This |
| 14586 | procedure has the merit of clarity, but the task is a little tricky. |
| 14587 | |
| 14588 | @need 1200 |
| 14589 | The first part is simple enough: the pattern must match ``at least one |
| 14590 | character that is a word or symbol constituent''. Thus: |
| 14591 | |
| 14592 | @smallexample |
| 14593 | "\\(\\w\\|\\s_\\)+" |
| 14594 | @end smallexample |
| 14595 | |
| 14596 | @noindent |
| 14597 | The @samp{\\(} is the first part of the grouping construct that |
| 14598 | includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated |
| 14599 | by the @samp{\\|}. The @samp{\\w} matches any word-constituent |
| 14600 | character and the @samp{\\s_} matches any character that is part of a |
| 14601 | symbol name but not a word-constituent character. The @samp{+} |
| 14602 | following the group indicates that the word or symbol constituent |
| 14603 | characters must be matched at least once. |
| 14604 | |
| 14605 | However, the second part of the regexp is more difficult to design. |
| 14606 | What we want is to follow the first part with ``optionally one or more |
| 14607 | characters that are not constituents of a word or symbol''. At first, |
| 14608 | I thought I could define this with the following: |
| 14609 | |
| 14610 | @smallexample |
| 14611 | "\\(\\W\\|\\S_\\)*" |
| 14612 | @end smallexample |
| 14613 | |
| 14614 | @noindent |
| 14615 | The upper case @samp{W} and @samp{S} match characters that are |
| 14616 | @emph{not} word or symbol constituents. Unfortunately, this |
| 14617 | expression matches any character that is either not a word constituent |
| 14618 | or not a symbol constituent. This matches any character! |
| 14619 | |
| 14620 | I then noticed that every word or symbol in my test region was |
| 14621 | followed by white space (blank space, tab, or newline). So I tried |
| 14622 | placing a pattern to match one or more blank spaces after the pattern |
| 14623 | for one or more word or symbol constituents. This failed, too. Words |
| 14624 | and symbols are often separated by whitespace, but in actual code |
| 14625 | parentheses may follow symbols and punctuation may follow words. So |
| 14626 | finally, I designed a pattern in which the word or symbol constituents |
| 14627 | are followed optionally by characters that are not white space and |
| 14628 | then followed optionally by white space. |
| 14629 | |
| 14630 | @need 800 |
| 14631 | Here is the full regular expression: |
| 14632 | |
| 14633 | @smallexample |
| 14634 | "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" |
| 14635 | @end smallexample |
| 14636 | |
| 14637 | @node count-words-in-defun |
| 14638 | @section The @code{count-words-in-defun} Function |
| 14639 | @cindex Counting words in a @code{defun} |
| 14640 | |
| 14641 | We have seen that there are several ways to write a |
| 14642 | @code{count-words-region} function. To write a |
| 14643 | @code{count-words-in-defun}, we need merely adapt one of these |
| 14644 | versions. |
| 14645 | |
| 14646 | The version that uses a @code{while} loop is easy to understand, so I |
| 14647 | am going to adapt that. Because @code{count-words-in-defun} will be |
| 14648 | part of a more complex program, it need not be interactive and it need |
| 14649 | not display a message but just return the count. These considerations |
| 14650 | simplify the definition a little. |
| 14651 | |
| 14652 | On the other hand, @code{count-words-in-defun} will be used within a |
| 14653 | buffer that contains function definitions. Consequently, it is |
| 14654 | reasonable to ask that the function determine whether it is called |
| 14655 | when point is within a function definition, and if it is, to return |
| 14656 | the count for that definition. This adds complexity to the |
| 14657 | definition, but saves us from needing to pass arguments to the |
| 14658 | function. |
| 14659 | |
| 14660 | @need 1250 |
| 14661 | These considerations lead us to prepare the following template: |
| 14662 | |
| 14663 | @smallexample |
| 14664 | @group |
| 14665 | (defun count-words-in-defun () |
| 14666 | "@var{documentation}@dots{}" |
| 14667 | (@var{set up}@dots{} |
| 14668 | (@var{while loop}@dots{}) |
| 14669 | @var{return count}) |
| 14670 | @end group |
| 14671 | @end smallexample |
| 14672 | |
| 14673 | @noindent |
| 14674 | As usual, our job is to fill in the slots. |
| 14675 | |
| 14676 | First, the set up. |
| 14677 | |
| 14678 | We are presuming that this function will be called within a buffer |
| 14679 | containing function definitions. Point will either be within a |
| 14680 | function definition or not. For @code{count-words-in-defun} to work, |
| 14681 | point must move to the beginning of the definition, a counter must |
| 14682 | start at zero, and the counting loop must stop when point reaches the |
| 14683 | end of the definition. |
| 14684 | |
| 14685 | The @code{beginning-of-defun} function searches backwards for an |
| 14686 | opening delimiter such as a @samp{(} at the beginning of a line, and |
| 14687 | moves point to that position, or else to the limit of the search. In |
| 14688 | practice, this means that @code{beginning-of-defun} moves point to the |
| 14689 | beginning of an enclosing or preceding function definition, or else to |
| 14690 | the beginning of the buffer. We can use @code{beginning-of-defun} to |
| 14691 | place point where we wish to start. |
| 14692 | |
| 14693 | The @code{while} loop requires a counter to keep track of the words or |
| 14694 | symbols being counted. A @code{let} expression can be used to create |
| 14695 | a local variable for this purpose, and bind it to an initial value of zero. |
| 14696 | |
| 14697 | The @code{end-of-defun} function works like @code{beginning-of-defun} |
| 14698 | except that it moves point to the end of the definition. |
| 14699 | @code{end-of-defun} can be used as part of an expression that |
| 14700 | determines the position of the end of the definition. |
| 14701 | |
| 14702 | The set up for @code{count-words-in-defun} takes shape rapidly: first |
| 14703 | we move point to the beginning of the definition, then we create a |
| 14704 | local variable to hold the count, and finally, we record the position |
| 14705 | of the end of the definition so the @code{while} loop will know when to stop |
| 14706 | looping. |
| 14707 | |
| 14708 | @need 1250 |
| 14709 | The code looks like this: |
| 14710 | |
| 14711 | @smallexample |
| 14712 | @group |
| 14713 | (beginning-of-defun) |
| 14714 | (let ((count 0) |
| 14715 | (end (save-excursion (end-of-defun) (point)))) |
| 14716 | @end group |
| 14717 | @end smallexample |
| 14718 | |
| 14719 | @noindent |
| 14720 | The code is simple. The only slight complication is likely to concern |
| 14721 | @code{end}: it is bound to the position of the end of the definition |
| 14722 | by a @code{save-excursion} expression that returns the value of point |
| 14723 | after @code{end-of-defun} temporarily moves it to the end of the |
| 14724 | definition. |
| 14725 | |
| 14726 | The second part of the @code{count-words-in-defun}, after the set up, |
| 14727 | is the @code{while} loop. |
| 14728 | |
| 14729 | The loop must contain an expression that jumps point forward word by |
| 14730 | word and symbol by symbol, and another expression that counts the |
| 14731 | jumps. The true-or-false-test for the @code{while} loop should test |
| 14732 | true so long as point should jump forward, and false when point is at |
| 14733 | the end of the definition. We have already redefined the regular |
| 14734 | expression for this, so the loop is straightforward: |
| 14735 | |
| 14736 | @smallexample |
| 14737 | @group |
| 14738 | (while (and (< (point) end) |
| 14739 | (re-search-forward |
| 14740 | "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t)) |
| 14741 | (setq count (1+ count))) |
| 14742 | @end group |
| 14743 | @end smallexample |
| 14744 | |
| 14745 | The third part of the function definition returns the count of words |
| 14746 | and symbols. This part is the last expression within the body of the |
| 14747 | @code{let} expression, and can be, very simply, the local variable |
| 14748 | @code{count}, which when evaluated returns the count. |
| 14749 | |
| 14750 | @need 1250 |
| 14751 | Put together, the @code{count-words-in-defun} definition looks like this: |
| 14752 | |
| 14753 | @findex count-words-in-defun |
| 14754 | @smallexample |
| 14755 | @group |
| 14756 | (defun count-words-in-defun () |
| 14757 | "Return the number of words and symbols in a defun." |
| 14758 | (beginning-of-defun) |
| 14759 | (let ((count 0) |
| 14760 | (end (save-excursion (end-of-defun) (point)))) |
| 14761 | @end group |
| 14762 | @group |
| 14763 | (while |
| 14764 | (and (< (point) end) |
| 14765 | (re-search-forward |
| 14766 | "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" |
| 14767 | end t)) |
| 14768 | (setq count (1+ count))) |
| 14769 | count)) |
| 14770 | @end group |
| 14771 | @end smallexample |
| 14772 | |
| 14773 | How to test this? The function is not interactive, but it is easy to |
| 14774 | put a wrapper around the function to make it interactive; we can use |
| 14775 | almost the same code as for the recursive version of |
| 14776 | @code{@value{COUNT-WORDS}}: |
| 14777 | |
| 14778 | @smallexample |
| 14779 | @group |
| 14780 | ;;; @r{Interactive version.} |
| 14781 | (defun count-words-defun () |
| 14782 | "Number of words and symbols in a function definition." |
| 14783 | (interactive) |
| 14784 | (message |
| 14785 | "Counting words and symbols in function definition ... ") |
| 14786 | @end group |
| 14787 | @group |
| 14788 | (let ((count (count-words-in-defun))) |
| 14789 | (cond |
| 14790 | ((zerop count) |
| 14791 | (message |
| 14792 | "The definition does NOT have any words or symbols.")) |
| 14793 | @end group |
| 14794 | @group |
| 14795 | ((= 1 count) |
| 14796 | (message |
| 14797 | "The definition has 1 word or symbol.")) |
| 14798 | (t |
| 14799 | (message |
| 14800 | "The definition has %d words or symbols." count))))) |
| 14801 | @end group |
| 14802 | @end smallexample |
| 14803 | |
| 14804 | @need 800 |
| 14805 | @noindent |
| 14806 | Let's re-use @kbd{C-c =} as a convenient keybinding: |
| 14807 | |
| 14808 | @smallexample |
| 14809 | (global-set-key "\C-c=" 'count-words-defun) |
| 14810 | @end smallexample |
| 14811 | |
| 14812 | Now we can try out @code{count-words-defun}: install both |
| 14813 | @code{count-words-in-defun} and @code{count-words-defun}, and set the |
| 14814 | keybinding, and then place the cursor within the following definition: |
| 14815 | |
| 14816 | @smallexample |
| 14817 | @group |
| 14818 | (defun multiply-by-seven (number) |
| 14819 | "Multiply NUMBER by seven." |
| 14820 | (* 7 number)) |
| 14821 | @result{} 10 |
| 14822 | @end group |
| 14823 | @end smallexample |
| 14824 | |
| 14825 | @noindent |
| 14826 | Success! The definition has 10 words and symbols. |
| 14827 | |
| 14828 | The next problem is to count the numbers of words and symbols in |
| 14829 | several definitions within a single file. |
| 14830 | |
| 14831 | @node Several defuns |
| 14832 | @section Count Several @code{defuns} Within a File |
| 14833 | |
| 14834 | A file such as @file{simple.el} may have a hundred or more function |
| 14835 | definitions within it. Our long term goal is to collect statistics on |
| 14836 | many files, but as a first step, our immediate goal is to collect |
| 14837 | statistics on one file. |
| 14838 | |
| 14839 | The information will be a series of numbers, each number being the |
| 14840 | length of a function definition. We can store the numbers in a list. |
| 14841 | |
| 14842 | We know that we will want to incorporate the information regarding one |
| 14843 | file with information about many other files; this means that the |
| 14844 | function for counting definition lengths within one file need only |
| 14845 | return the list of lengths. It need not and should not display any |
| 14846 | messages. |
| 14847 | |
| 14848 | The word count commands contain one expression to jump point forward |
| 14849 | word by word and another expression to count the jumps. The function |
| 14850 | to return the lengths of definitions can be designed to work the same |
| 14851 | way, with one expression to jump point forward definition by |
| 14852 | definition and another expression to construct the lengths' list. |
| 14853 | |
| 14854 | This statement of the problem makes it elementary to write the |
| 14855 | function definition. Clearly, we will start the count at the |
| 14856 | beginning of the file, so the first command will be @code{(goto-char |
| 14857 | (point-min))}. Next, we start the @code{while} loop; and the |
| 14858 | true-or-false test of the loop can be a regular expression search for |
| 14859 | the next function definition---so long as the search succeeds, point |
| 14860 | is moved forward and then the body of the loop is evaluated. The body |
| 14861 | needs an expression that constructs the lengths' list. @code{cons}, |
| 14862 | the list construction command, can be used to create the list. That |
| 14863 | is almost all there is to it. |
| 14864 | |
| 14865 | @need 800 |
| 14866 | Here is what this fragment of code looks like: |
| 14867 | |
| 14868 | @smallexample |
| 14869 | @group |
| 14870 | (goto-char (point-min)) |
| 14871 | (while (re-search-forward "^(defun" nil t) |
| 14872 | (setq lengths-list |
| 14873 | (cons (count-words-in-defun) lengths-list))) |
| 14874 | @end group |
| 14875 | @end smallexample |
| 14876 | |
| 14877 | What we have left out is the mechanism for finding the file that |
| 14878 | contains the function definitions. |
| 14879 | |
| 14880 | In previous examples, we either used this, the Info file, or we |
| 14881 | switched back and forth to some other buffer, such as the |
| 14882 | @file{*scratch*} buffer. |
| 14883 | |
| 14884 | Finding a file is a new process that we have not yet discussed. |
| 14885 | |
| 14886 | @node Find a File |
| 14887 | @section Find a File |
| 14888 | @cindex Find a File |
| 14889 | |
| 14890 | To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file}) |
| 14891 | command. This command is almost, but not quite right for the lengths |
| 14892 | problem. |
| 14893 | |
| 14894 | @need 1200 |
| 14895 | Let's look at the source for @code{find-file}: |
| 14896 | |
| 14897 | @smallexample |
| 14898 | @group |
| 14899 | (defun find-file (filename) |
| 14900 | "Edit file FILENAME. |
| 14901 | Switch to a buffer visiting file FILENAME, |
| 14902 | creating one if none already exists." |
| 14903 | (interactive "FFind file: ") |
| 14904 | (switch-to-buffer (find-file-noselect filename))) |
| 14905 | @end group |
| 14906 | @end smallexample |
| 14907 | |
| 14908 | @noindent |
| 14909 | (The most recent version of the @code{find-file} function definition |
| 14910 | permits you to specify optional wildcards to visit multiple files; that |
| 14911 | makes the definition more complex and we will not discuss it here, |
| 14912 | since it is not relevant. You can see its source using either |
| 14913 | @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).) |
| 14914 | |
| 14915 | @ignore |
| 14916 | In Emacs 22 |
| 14917 | (defun find-file (filename &optional wildcards) |
| 14918 | "Edit file FILENAME. |
| 14919 | Switch to a buffer visiting file FILENAME, |
| 14920 | creating one if none already exists. |
| 14921 | Interactively, the default if you just type RET is the current directory, |
| 14922 | but the visited file name is available through the minibuffer history: |
| 14923 | type M-n to pull it into the minibuffer. |
| 14924 | |
| 14925 | Interactively, or if WILDCARDS is non-nil in a call from Lisp, |
| 14926 | expand wildcards (if any) and visit multiple files. You can |
| 14927 | suppress wildcard expansion by setting `find-file-wildcards' to nil. |
| 14928 | |
| 14929 | To visit a file without any kind of conversion and without |
| 14930 | automatically choosing a major mode, use \\[find-file-literally]." |
| 14931 | (interactive (find-file-read-args "Find file: " nil)) |
| 14932 | (let ((value (find-file-noselect filename nil nil wildcards))) |
| 14933 | (if (listp value) |
| 14934 | (mapcar 'switch-to-buffer (nreverse value)) |
| 14935 | (switch-to-buffer value)))) |
| 14936 | @end ignore |
| 14937 | |
| 14938 | The definition I am showing possesses short but complete documentation |
| 14939 | and an interactive specification that prompts you for a file name when |
| 14940 | you use the command interactively. The body of the definition |
| 14941 | contains two functions, @code{find-file-noselect} and |
| 14942 | @code{switch-to-buffer}. |
| 14943 | |
| 14944 | According to its documentation as shown by @kbd{C-h f} (the |
| 14945 | @code{describe-function} command), the @code{find-file-noselect} |
| 14946 | function reads the named file into a buffer and returns the buffer. |
| 14947 | (Its most recent version includes an optional wildcards argument, |
| 14948 | too, as well as another to read a file literally and an other you |
| 14949 | suppress warning messages. These optional arguments are irrelevant.) |
| 14950 | |
| 14951 | However, the @code{find-file-noselect} function does not select the |
| 14952 | buffer in which it puts the file. Emacs does not switch its attention |
| 14953 | (or yours if you are using @code{find-file-noselect}) to the selected |
| 14954 | buffer. That is what @code{switch-to-buffer} does: it switches the |
| 14955 | buffer to which Emacs attention is directed; and it switches the |
| 14956 | buffer displayed in the window to the new buffer. We have discussed |
| 14957 | buffer switching elsewhere. (@xref{Switching Buffers}.) |
| 14958 | |
| 14959 | In this histogram project, we do not need to display each file on the |
| 14960 | screen as the program determines the length of each definition within |
| 14961 | it. Instead of employing @code{switch-to-buffer}, we can work with |
| 14962 | @code{set-buffer}, which redirects the attention of the computer |
| 14963 | program to a different buffer but does not redisplay it on the screen. |
| 14964 | So instead of calling on @code{find-file} to do the job, we must write |
| 14965 | our own expression. |
| 14966 | |
| 14967 | The task is easy: use @code{find-file-noselect} and @code{set-buffer}. |
| 14968 | |
| 14969 | @node lengths-list-file |
| 14970 | @section @code{lengths-list-file} in Detail |
| 14971 | |
| 14972 | The core of the @code{lengths-list-file} function is a @code{while} |
| 14973 | loop containing a function to move point forward `defun by defun' and |
| 14974 | a function to count the number of words and symbols in each defun. |
| 14975 | This core must be surrounded by functions that do various other tasks, |
| 14976 | including finding the file, and ensuring that point starts out at the |
| 14977 | beginning of the file. The function definition looks like this: |
| 14978 | @findex lengths-list-file |
| 14979 | |
| 14980 | @smallexample |
| 14981 | @group |
| 14982 | (defun lengths-list-file (filename) |
| 14983 | "Return list of definitions' lengths within FILE. |
| 14984 | The returned list is a list of numbers. |
| 14985 | Each number is the number of words or |
| 14986 | symbols in one function definition." |
| 14987 | @end group |
| 14988 | @group |
| 14989 | (message "Working on `%s' ... " filename) |
| 14990 | (save-excursion |
| 14991 | (let ((buffer (find-file-noselect filename)) |
| 14992 | (lengths-list)) |
| 14993 | (set-buffer buffer) |
| 14994 | (setq buffer-read-only t) |
| 14995 | (widen) |
| 14996 | (goto-char (point-min)) |
| 14997 | (while (re-search-forward "^(defun" nil t) |
| 14998 | (setq lengths-list |
| 14999 | (cons (count-words-in-defun) lengths-list))) |
| 15000 | (kill-buffer buffer) |
| 15001 | lengths-list))) |
| 15002 | @end group |
| 15003 | @end smallexample |
| 15004 | |
| 15005 | @noindent |
| 15006 | The function is passed one argument, the name of the file on which it |
| 15007 | will work. It has four lines of documentation, but no interactive |
| 15008 | specification. Since people worry that a computer is broken if they |
| 15009 | don't see anything going on, the first line of the body is a |
| 15010 | message. |
| 15011 | |
| 15012 | The next line contains a @code{save-excursion} that returns Emacs's |
| 15013 | attention to the current buffer when the function completes. This is |
| 15014 | useful in case you embed this function in another function that |
| 15015 | presumes point is restored to the original buffer. |
| 15016 | |
| 15017 | In the varlist of the @code{let} expression, Emacs finds the file and |
| 15018 | binds the local variable @code{buffer} to the buffer containing the |
| 15019 | file. At the same time, Emacs creates @code{lengths-list} as a local |
| 15020 | variable. |
| 15021 | |
| 15022 | Next, Emacs switches its attention to the buffer. |
| 15023 | |
| 15024 | In the following line, Emacs makes the buffer read-only. Ideally, |
| 15025 | this line is not necessary. None of the functions for counting words |
| 15026 | and symbols in a function definition should change the buffer. |
| 15027 | Besides, the buffer is not going to be saved, even if it were changed. |
| 15028 | This line is entirely the consequence of great, perhaps excessive, |
| 15029 | caution. The reason for the caution is that this function and those |
| 15030 | it calls work on the sources for Emacs and it is inconvenient if they |
| 15031 | are inadvertently modified. It goes without saying that I did not |
| 15032 | realize a need for this line until an experiment went awry and started |
| 15033 | to modify my Emacs source files @dots{} |
| 15034 | |
| 15035 | Next comes a call to widen the buffer if it is narrowed. This |
| 15036 | function is usually not needed---Emacs creates a fresh buffer if none |
| 15037 | already exists; but if a buffer visiting the file already exists Emacs |
| 15038 | returns that one. In this case, the buffer may be narrowed and must |
| 15039 | be widened. If we wanted to be fully `user-friendly', we would |
| 15040 | arrange to save the restriction and the location of point, but we |
| 15041 | won't. |
| 15042 | |
| 15043 | The @code{(goto-char (point-min))} expression moves point to the |
| 15044 | beginning of the buffer. |
| 15045 | |
| 15046 | Then comes a @code{while} loop in which the `work' of the function is |
| 15047 | carried out. In the loop, Emacs determines the length of each |
| 15048 | definition and constructs a lengths' list containing the information. |
| 15049 | |
| 15050 | Emacs kills the buffer after working through it. This is to save |
| 15051 | space inside of Emacs. My version of GNU Emacs 19 contained over 300 |
| 15052 | source files of interest; GNU Emacs 22 contains over a thousand source |
| 15053 | files. Another function will apply @code{lengths-list-file} to each |
| 15054 | of the files. |
| 15055 | |
| 15056 | Finally, the last expression within the @code{let} expression is the |
| 15057 | @code{lengths-list} variable; its value is returned as the value of |
| 15058 | the whole function. |
| 15059 | |
| 15060 | You can try this function by installing it in the usual fashion. Then |
| 15061 | place your cursor after the following expression and type @kbd{C-x |
| 15062 | C-e} (@code{eval-last-sexp}). |
| 15063 | |
| 15064 | @c !!! 22.1.1 lisp sources location here |
| 15065 | @smallexample |
| 15066 | (lengths-list-file |
| 15067 | "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el") |
| 15068 | @end smallexample |
| 15069 | |
| 15070 | @noindent |
| 15071 | (You may need to change the pathname of the file; the one here is for |
| 15072 | GNU Emacs version 22.1.1. To change the expression, copy it to |
| 15073 | the @file{*scratch*} buffer and edit it. |
| 15074 | |
| 15075 | @need 1200 |
| 15076 | @noindent |
| 15077 | (Also, to see the full length of the list, rather than a truncated |
| 15078 | version, you may have to evaluate the following: |
| 15079 | |
| 15080 | @smallexample |
| 15081 | (custom-set-variables '(eval-expression-print-length nil)) |
| 15082 | @end smallexample |
| 15083 | |
| 15084 | @noindent |
| 15085 | (@xref{defcustom, , Specifying Variables using @code{defcustom}}. |
| 15086 | Then evaluate the @code{lengths-list-file} expression.) |
| 15087 | |
| 15088 | @need 1200 |
| 15089 | The lengths' list for @file{debug.el} takes less than a second to |
| 15090 | produce and looks like this in GNU Emacs 22: |
| 15091 | |
| 15092 | @smallexample |
| 15093 | (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587) |
| 15094 | @end smallexample |
| 15095 | |
| 15096 | @need 1500 |
| 15097 | (Using my old machine, the version 19 lengths' list for @file{debug.el} |
| 15098 | took seven seconds to produce and looked like this: |
| 15099 | |
| 15100 | @smallexample |
| 15101 | (75 41 80 62 20 45 44 68 45 12 34 235) |
| 15102 | @end smallexample |
| 15103 | |
| 15104 | (The newer version of @file{debug.el} contains more defuns than the |
| 15105 | earlier one; and my new machine is much faster than the old one.) |
| 15106 | |
| 15107 | Note that the length of the last definition in the file is first in |
| 15108 | the list. |
| 15109 | |
| 15110 | @node Several files |
| 15111 | @section Count Words in @code{defuns} in Different Files |
| 15112 | |
| 15113 | In the previous section, we created a function that returns a list of |
| 15114 | the lengths of each definition in a file. Now, we want to define a |
| 15115 | function to return a master list of the lengths of the definitions in |
| 15116 | a list of files. |
| 15117 | |
| 15118 | Working on each of a list of files is a repetitious act, so we can use |
| 15119 | either a @code{while} loop or recursion. |
| 15120 | |
| 15121 | @menu |
| 15122 | * lengths-list-many-files:: Return a list of the lengths of defuns. |
| 15123 | * append:: Attach one list to another. |
| 15124 | @end menu |
| 15125 | |
| 15126 | @ifnottex |
| 15127 | @node lengths-list-many-files |
| 15128 | @unnumberedsubsec Determine the lengths of @code{defuns} |
| 15129 | @end ifnottex |
| 15130 | |
| 15131 | The design using a @code{while} loop is routine. The argument passed |
| 15132 | the function is a list of files. As we saw earlier (@pxref{Loop |
| 15133 | Example}), you can write a @code{while} loop so that the body of the |
| 15134 | loop is evaluated if such a list contains elements, but to exit the |
| 15135 | loop if the list is empty. For this design to work, the body of the |
| 15136 | loop must contain an expression that shortens the list each time the |
| 15137 | body is evaluated, so that eventually the list is empty. The usual |
| 15138 | technique is to set the value of the list to the value of the @sc{cdr} |
| 15139 | of the list each time the body is evaluated. |
| 15140 | |
| 15141 | @need 800 |
| 15142 | The template looks like this: |
| 15143 | |
| 15144 | @smallexample |
| 15145 | @group |
| 15146 | (while @var{test-whether-list-is-empty} |
| 15147 | @var{body}@dots{} |
| 15148 | @var{set-list-to-cdr-of-list}) |
| 15149 | @end group |
| 15150 | @end smallexample |
| 15151 | |
| 15152 | Also, we remember that a @code{while} loop returns @code{nil} (the |
| 15153 | result of evaluating the true-or-false-test), not the result of any |
| 15154 | evaluation within its body. (The evaluations within the body of the |
| 15155 | loop are done for their side effects.) However, the expression that |
| 15156 | sets the lengths' list is part of the body---and that is the value |
| 15157 | that we want returned by the function as a whole. To do this, we |
| 15158 | enclose the @code{while} loop within a @code{let} expression, and |
| 15159 | arrange that the last element of the @code{let} expression contains |
| 15160 | the value of the lengths' list. (@xref{Incrementing Example, , Loop |
| 15161 | Example with an Incrementing Counter}.) |
| 15162 | |
| 15163 | @findex lengths-list-many-files |
| 15164 | @need 1250 |
| 15165 | These considerations lead us directly to the function itself: |
| 15166 | |
| 15167 | @smallexample |
| 15168 | @group |
| 15169 | ;;; @r{Use @code{while} loop.} |
| 15170 | (defun lengths-list-many-files (list-of-files) |
| 15171 | "Return list of lengths of defuns in LIST-OF-FILES." |
| 15172 | @end group |
| 15173 | @group |
| 15174 | (let (lengths-list) |
| 15175 | |
| 15176 | ;;; @r{true-or-false-test} |
| 15177 | (while list-of-files |
| 15178 | (setq lengths-list |
| 15179 | (append |
| 15180 | lengths-list |
| 15181 | |
| 15182 | ;;; @r{Generate a lengths' list.} |
| 15183 | (lengths-list-file |
| 15184 | (expand-file-name (car list-of-files))))) |
| 15185 | @end group |
| 15186 | |
| 15187 | @group |
| 15188 | ;;; @r{Make files' list shorter.} |
| 15189 | (setq list-of-files (cdr list-of-files))) |
| 15190 | |
| 15191 | ;;; @r{Return final value of lengths' list.} |
| 15192 | lengths-list)) |
| 15193 | @end group |
| 15194 | @end smallexample |
| 15195 | |
| 15196 | @code{expand-file-name} is a built-in function that converts a file |
| 15197 | name to the absolute, long, path name form. The function employs the |
| 15198 | name of the directory in which the function is called. |
| 15199 | |
| 15200 | @c !!! 22.1.1 lisp sources location here |
| 15201 | @need 1500 |
| 15202 | Thus, if @code{expand-file-name} is called on @code{debug.el} when |
| 15203 | Emacs is visiting the |
| 15204 | @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory, |
| 15205 | |
| 15206 | @smallexample |
| 15207 | debug.el |
| 15208 | @end smallexample |
| 15209 | |
| 15210 | @need 800 |
| 15211 | @noindent |
| 15212 | becomes |
| 15213 | |
| 15214 | @c !!! 22.1.1 lisp sources location here |
| 15215 | @smallexample |
| 15216 | /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el |
| 15217 | @end smallexample |
| 15218 | |
| 15219 | The only other new element of this function definition is the as yet |
| 15220 | unstudied function @code{append}, which merits a short section for |
| 15221 | itself. |
| 15222 | |
| 15223 | @node append |
| 15224 | @subsection The @code{append} Function |
| 15225 | |
| 15226 | @need 800 |
| 15227 | The @code{append} function attaches one list to another. Thus, |
| 15228 | |
| 15229 | @smallexample |
| 15230 | (append '(1 2 3 4) '(5 6 7 8)) |
| 15231 | @end smallexample |
| 15232 | |
| 15233 | @need 800 |
| 15234 | @noindent |
| 15235 | produces the list |
| 15236 | |
| 15237 | @smallexample |
| 15238 | (1 2 3 4 5 6 7 8) |
| 15239 | @end smallexample |
| 15240 | |
| 15241 | This is exactly how we want to attach two lengths' lists produced by |
| 15242 | @code{lengths-list-file} to each other. The results contrast with |
| 15243 | @code{cons}, |
| 15244 | |
| 15245 | @smallexample |
| 15246 | (cons '(1 2 3 4) '(5 6 7 8)) |
| 15247 | @end smallexample |
| 15248 | |
| 15249 | @need 1250 |
| 15250 | @noindent |
| 15251 | which constructs a new list in which the first argument to @code{cons} |
| 15252 | becomes the first element of the new list: |
| 15253 | |
| 15254 | @smallexample |
| 15255 | ((1 2 3 4) 5 6 7 8) |
| 15256 | @end smallexample |
| 15257 | |
| 15258 | @node Several files recursively |
| 15259 | @section Recursively Count Words in Different Files |
| 15260 | |
| 15261 | Besides a @code{while} loop, you can work on each of a list of files |
| 15262 | with recursion. A recursive version of @code{lengths-list-many-files} |
| 15263 | is short and simple. |
| 15264 | |
| 15265 | The recursive function has the usual parts: the `do-again-test', the |
| 15266 | `next-step-expression', and the recursive call. The `do-again-test' |
| 15267 | determines whether the function should call itself again, which it |
| 15268 | will do if the @code{list-of-files} contains any remaining elements; |
| 15269 | the `next-step-expression' resets the @code{list-of-files} to the |
| 15270 | @sc{cdr} of itself, so eventually the list will be empty; and the |
| 15271 | recursive call calls itself on the shorter list. The complete |
| 15272 | function is shorter than this description! |
| 15273 | @findex recursive-lengths-list-many-files |
| 15274 | |
| 15275 | @smallexample |
| 15276 | @group |
| 15277 | (defun recursive-lengths-list-many-files (list-of-files) |
| 15278 | "Return list of lengths of each defun in LIST-OF-FILES." |
| 15279 | (if list-of-files ; @r{do-again-test} |
| 15280 | (append |
| 15281 | (lengths-list-file |
| 15282 | (expand-file-name (car list-of-files))) |
| 15283 | (recursive-lengths-list-many-files |
| 15284 | (cdr list-of-files))))) |
| 15285 | @end group |
| 15286 | @end smallexample |
| 15287 | |
| 15288 | @noindent |
| 15289 | In a sentence, the function returns the lengths' list for the first of |
| 15290 | the @code{list-of-files} appended to the result of calling itself on |
| 15291 | the rest of the @code{list-of-files}. |
| 15292 | |
| 15293 | Here is a test of @code{recursive-lengths-list-many-files}, along with |
| 15294 | the results of running @code{lengths-list-file} on each of the files |
| 15295 | individually. |
| 15296 | |
| 15297 | Install @code{recursive-lengths-list-many-files} and |
| 15298 | @code{lengths-list-file}, if necessary, and then evaluate the |
| 15299 | following expressions. You may need to change the files' pathnames; |
| 15300 | those here work when this Info file and the Emacs sources are located |
| 15301 | in their customary places. To change the expressions, copy them to |
| 15302 | the @file{*scratch*} buffer, edit them, and then evaluate them. |
| 15303 | |
| 15304 | The results are shown after the @samp{@result{}}. (These results are |
| 15305 | for files from Emacs version 22.1.1; files from other versions of |
| 15306 | Emacs may produce different results.) |
| 15307 | |
| 15308 | @c !!! 22.1.1 lisp sources location here |
| 15309 | @smallexample |
| 15310 | @group |
| 15311 | (cd "/usr/local/share/emacs/22.1.1/") |
| 15312 | |
| 15313 | (lengths-list-file "./lisp/macros.el") |
| 15314 | @result{} (283 263 480 90) |
| 15315 | @end group |
| 15316 | |
| 15317 | @group |
| 15318 | (lengths-list-file "./lisp/mail/mailalias.el") |
| 15319 | @result{} (38 32 29 95 178 180 321 218 324) |
| 15320 | @end group |
| 15321 | |
| 15322 | @group |
| 15323 | (lengths-list-file "./lisp/makesum.el") |
| 15324 | @result{} (85 181) |
| 15325 | @end group |
| 15326 | |
| 15327 | @group |
| 15328 | (recursive-lengths-list-many-files |
| 15329 | '("./lisp/macros.el" |
| 15330 | "./lisp/mail/mailalias.el" |
| 15331 | "./lisp/makesum.el")) |
| 15332 | @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181) |
| 15333 | @end group |
| 15334 | @end smallexample |
| 15335 | |
| 15336 | The @code{recursive-lengths-list-many-files} function produces the |
| 15337 | output we want. |
| 15338 | |
| 15339 | The next step is to prepare the data in the list for display in a graph. |
| 15340 | |
| 15341 | @node Prepare the data |
| 15342 | @section Prepare the Data for Display in a Graph |
| 15343 | |
| 15344 | The @code{recursive-lengths-list-many-files} function returns a list |
| 15345 | of numbers. Each number records the length of a function definition. |
| 15346 | What we need to do now is transform this data into a list of numbers |
| 15347 | suitable for generating a graph. The new list will tell how many |
| 15348 | functions definitions contain less than 10 words and |
| 15349 | symbols, how many contain between 10 and 19 words and symbols, how |
| 15350 | many contain between 20 and 29 words and symbols, and so on. |
| 15351 | |
| 15352 | In brief, we need to go through the lengths' list produced by the |
| 15353 | @code{recursive-lengths-list-many-files} function and count the number |
| 15354 | of defuns within each range of lengths, and produce a list of those |
| 15355 | numbers. |
| 15356 | |
| 15357 | @menu |
| 15358 | * Data for Display in Detail:: |
| 15359 | * Sorting:: Sorting lists. |
| 15360 | * Files List:: Making a list of files. |
| 15361 | * Counting function definitions:: |
| 15362 | @end menu |
| 15363 | |
| 15364 | @ifnottex |
| 15365 | @node Data for Display in Detail |
| 15366 | @unnumberedsubsec The Data for Display in Detail |
| 15367 | @end ifnottex |
| 15368 | |
| 15369 | Based on what we have done before, we can readily foresee that it |
| 15370 | should not be too hard to write a function that `@sc{cdr}s' down the |
| 15371 | lengths' list, looks at each element, determines which length range it |
| 15372 | is in, and increments a counter for that range. |
| 15373 | |
| 15374 | However, before beginning to write such a function, we should consider |
| 15375 | the advantages of sorting the lengths' list first, so the numbers are |
| 15376 | ordered from smallest to largest. First, sorting will make it easier |
| 15377 | to count the numbers in each range, since two adjacent numbers will |
| 15378 | either be in the same length range or in adjacent ranges. Second, by |
| 15379 | inspecting a sorted list, we can discover the highest and lowest |
| 15380 | number, and thereby determine the largest and smallest length range |
| 15381 | that we will need. |
| 15382 | |
| 15383 | @node Sorting |
| 15384 | @subsection Sorting Lists |
| 15385 | @findex sort |
| 15386 | |
| 15387 | Emacs contains a function to sort lists, called (as you might guess) |
| 15388 | @code{sort}. The @code{sort} function takes two arguments, the list |
| 15389 | to be sorted, and a predicate that determines whether the first of |
| 15390 | two list elements is ``less'' than the second. |
| 15391 | |
| 15392 | As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong |
| 15393 | Type Object as an Argument}), a predicate is a function that |
| 15394 | determines whether some property is true or false. The @code{sort} |
| 15395 | function will reorder a list according to whatever property the |
| 15396 | predicate uses; this means that @code{sort} can be used to sort |
| 15397 | non-numeric lists by non-numeric criteria---it can, for example, |
| 15398 | alphabetize a list. |
| 15399 | |
| 15400 | @need 1250 |
| 15401 | The @code{<} function is used when sorting a numeric list. For example, |
| 15402 | |
| 15403 | @smallexample |
| 15404 | (sort '(4 8 21 17 33 7 21 7) '<) |
| 15405 | @end smallexample |
| 15406 | |
| 15407 | @need 800 |
| 15408 | @noindent |
| 15409 | produces this: |
| 15410 | |
| 15411 | @smallexample |
| 15412 | (4 7 7 8 17 21 21 33) |
| 15413 | @end smallexample |
| 15414 | |
| 15415 | @noindent |
| 15416 | (Note that in this example, both the arguments are quoted so that the |
| 15417 | symbols are not evaluated before being passed to @code{sort} as |
| 15418 | arguments.) |
| 15419 | |
| 15420 | Sorting the list returned by the |
| 15421 | @code{recursive-lengths-list-many-files} function is straightforward; |
| 15422 | it uses the @code{<} function: |
| 15423 | |
| 15424 | @ignore |
| 15425 | 2006 Oct 29 |
| 15426 | In GNU Emacs 22, eval |
| 15427 | (progn |
| 15428 | (cd "/usr/local/share/emacs/22.0.50/") |
| 15429 | (sort |
| 15430 | (recursive-lengths-list-many-files |
| 15431 | '("./lisp/macros.el" |
| 15432 | "./lisp/mail/mailalias.el" |
| 15433 | "./lisp/makesum.el")) |
| 15434 | '<)) |
| 15435 | |
| 15436 | @end ignore |
| 15437 | |
| 15438 | @smallexample |
| 15439 | @group |
| 15440 | (sort |
| 15441 | (recursive-lengths-list-many-files |
| 15442 | '("./lisp/macros.el" |
| 15443 | "./lisp/mailalias.el" |
| 15444 | "./lisp/makesum.el")) |
| 15445 | '<) |
| 15446 | @end group |
| 15447 | @end smallexample |
| 15448 | |
| 15449 | @need 800 |
| 15450 | @noindent |
| 15451 | which produces: |
| 15452 | |
| 15453 | @smallexample |
| 15454 | (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480) |
| 15455 | @end smallexample |
| 15456 | |
| 15457 | @noindent |
| 15458 | (Note that in this example, the first argument to @code{sort} is not |
| 15459 | quoted, since the expression must be evaluated so as to produce the |
| 15460 | list that is passed to @code{sort}.) |
| 15461 | |
| 15462 | @node Files List |
| 15463 | @subsection Making a List of Files |
| 15464 | |
| 15465 | The @code{recursive-lengths-list-many-files} function requires a list |
| 15466 | of files as its argument. For our test examples, we constructed such |
| 15467 | a list by hand; but the Emacs Lisp source directory is too large for |
| 15468 | us to do for that. Instead, we will write a function to do the job |
| 15469 | for us. In this function, we will use both a @code{while} loop and a |
| 15470 | recursive call. |
| 15471 | |
| 15472 | @findex directory-files |
| 15473 | We did not have to write a function like this for older versions of |
| 15474 | GNU Emacs, since they placed all the @samp{.el} files in one |
| 15475 | directory. Instead, we were able to use the @code{directory-files} |
| 15476 | function, which lists the names of files that match a specified |
| 15477 | pattern within a single directory. |
| 15478 | |
| 15479 | However, recent versions of Emacs place Emacs Lisp files in |
| 15480 | sub-directories of the top level @file{lisp} directory. This |
| 15481 | re-arrangement eases navigation. For example, all the mail related |
| 15482 | files are in a @file{lisp} sub-directory called @file{mail}. But at |
| 15483 | the same time, this arrangement forces us to create a file listing |
| 15484 | function that descends into the sub-directories. |
| 15485 | |
| 15486 | @findex files-in-below-directory |
| 15487 | We can create this function, called @code{files-in-below-directory}, |
| 15488 | using familiar functions such as @code{car}, @code{nthcdr}, and |
| 15489 | @code{substring} in conjunction with an existing function called |
| 15490 | @code{directory-files-and-attributes}. This latter function not only |
| 15491 | lists all the filenames in a directory, including the names |
| 15492 | of sub-directories, but also their attributes. |
| 15493 | |
| 15494 | To restate our goal: to create a function that will enable us |
| 15495 | to feed filenames to @code{recursive-lengths-list-many-files} |
| 15496 | as a list that looks like this (but with more elements): |
| 15497 | |
| 15498 | @smallexample |
| 15499 | @group |
| 15500 | ("./lisp/macros.el" |
| 15501 | "./lisp/mail/rmail.el" |
| 15502 | "./lisp/makesum.el") |
| 15503 | @end group |
| 15504 | @end smallexample |
| 15505 | |
| 15506 | The @code{directory-files-and-attributes} function returns a list of |
| 15507 | lists. Each of the lists within the main list consists of 13 |
| 15508 | elements. The first element is a string that contains the name of the |
| 15509 | file---which, in GNU/Linux, may be a `directory file', that is to |
| 15510 | say, a file with the special attributes of a directory. The second |
| 15511 | element of the list is @code{t} for a directory, a string |
| 15512 | for symbolic link (the string is the name linked to), or @code{nil}. |
| 15513 | |
| 15514 | For example, the first @samp{.el} file in the @file{lisp/} directory |
| 15515 | is @file{abbrev.el}. Its name is |
| 15516 | @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a |
| 15517 | directory or a symbolic link. |
| 15518 | |
| 15519 | @need 1000 |
| 15520 | This is how @code{directory-files-and-attributes} lists that file and |
| 15521 | its attributes: |
| 15522 | |
| 15523 | @smallexample |
| 15524 | @group |
| 15525 | ("abbrev.el" |
| 15526 | nil |
| 15527 | 1 |
| 15528 | 1000 |
| 15529 | 100 |
| 15530 | @end group |
| 15531 | @group |
| 15532 | (20615 27034 579989 697000) |
| 15533 | (17905 55681 0 0) |
| 15534 | (20615 26327 734791 805000) |
| 15535 | 13188 |
| 15536 | "-rw-r--r--" |
| 15537 | @end group |
| 15538 | @group |
| 15539 | t |
| 15540 | 2971624 |
| 15541 | 773) |
| 15542 | @end group |
| 15543 | @end smallexample |
| 15544 | |
| 15545 | @need 1200 |
| 15546 | On the other hand, @file{mail/} is a directory within the @file{lisp/} |
| 15547 | directory. The beginning of its listing looks like this: |
| 15548 | |
| 15549 | @smallexample |
| 15550 | @group |
| 15551 | ("mail" |
| 15552 | t |
| 15553 | @dots{} |
| 15554 | ) |
| 15555 | @end group |
| 15556 | @end smallexample |
| 15557 | |
| 15558 | (To learn about the different attributes, look at the documentation of |
| 15559 | @code{file-attributes}. Bear in mind that the @code{file-attributes} |
| 15560 | function does not list the filename, so its first element is |
| 15561 | @code{directory-files-and-attributes}'s second element.) |
| 15562 | |
| 15563 | We will want our new function, @code{files-in-below-directory}, to |
| 15564 | list the @samp{.el} files in the directory it is told to check, and in |
| 15565 | any directories below that directory. |
| 15566 | |
| 15567 | This gives us a hint on how to construct |
| 15568 | @code{files-in-below-directory}: within a directory, the function |
| 15569 | should add @samp{.el} filenames to a list; and if, within a directory, |
| 15570 | the function comes upon a sub-directory, it should go into that |
| 15571 | sub-directory and repeat its actions. |
| 15572 | |
| 15573 | However, we should note that every directory contains a name that |
| 15574 | refers to itself, called @file{.}, (``dot'') and a name that refers to |
| 15575 | its parent directory, called @file{..} (``double dot''). (In |
| 15576 | @file{/}, the root directory, @file{..} refers to itself, since |
| 15577 | @file{/} has no parent.) Clearly, we do not want our |
| 15578 | @code{files-in-below-directory} function to enter those directories, |
| 15579 | since they always lead us, directly or indirectly, to the current |
| 15580 | directory. |
| 15581 | |
| 15582 | Consequently, our @code{files-in-below-directory} function must do |
| 15583 | several tasks: |
| 15584 | |
| 15585 | @itemize @bullet |
| 15586 | @item |
| 15587 | Check to see whether it is looking at a filename that ends in |
| 15588 | @samp{.el}; and if so, add its name to a list. |
| 15589 | |
| 15590 | @item |
| 15591 | Check to see whether it is looking at a filename that is the name of a |
| 15592 | directory; and if so, |
| 15593 | |
| 15594 | @itemize @minus |
| 15595 | @item |
| 15596 | Check to see whether it is looking at @file{.} or @file{..}; and if |
| 15597 | so skip it. |
| 15598 | |
| 15599 | @item |
| 15600 | Or else, go into that directory and repeat the process. |
| 15601 | @end itemize |
| 15602 | @end itemize |
| 15603 | |
| 15604 | Let's write a function definition to do these tasks. We will use a |
| 15605 | @code{while} loop to move from one filename to another within a |
| 15606 | directory, checking what needs to be done; and we will use a recursive |
| 15607 | call to repeat the actions on each sub-directory. The recursive |
| 15608 | pattern is `accumulate' |
| 15609 | (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}), |
| 15610 | using @code{append} as the combiner. |
| 15611 | |
| 15612 | @ignore |
| 15613 | (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$") |
| 15614 | (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'") |
| 15615 | |
| 15616 | (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$") |
| 15617 | (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'") |
| 15618 | @end ignore |
| 15619 | |
| 15620 | @c /usr/local/share/emacs/22.1.1/lisp/ |
| 15621 | |
| 15622 | @need 800 |
| 15623 | Here is the function: |
| 15624 | |
| 15625 | @smallexample |
| 15626 | @group |
| 15627 | (defun files-in-below-directory (directory) |
| 15628 | "List the .el files in DIRECTORY and in its sub-directories." |
| 15629 | ;; Although the function will be used non-interactively, |
| 15630 | ;; it will be easier to test if we make it interactive. |
| 15631 | ;; The directory will have a name such as |
| 15632 | ;; "/usr/local/share/emacs/22.1.1/lisp/" |
| 15633 | (interactive "DDirectory name: ") |
| 15634 | @end group |
| 15635 | @group |
| 15636 | (let (el-files-list |
| 15637 | (current-directory-list |
| 15638 | (directory-files-and-attributes directory t))) |
| 15639 | ;; while we are in the current directory |
| 15640 | (while current-directory-list |
| 15641 | @end group |
| 15642 | @group |
| 15643 | (cond |
| 15644 | ;; check to see whether filename ends in `.el' |
| 15645 | ;; and if so, append its name to a list. |
| 15646 | ((equal ".el" (substring (car (car current-directory-list)) -3)) |
| 15647 | (setq el-files-list |
| 15648 | (cons (car (car current-directory-list)) el-files-list))) |
| 15649 | @end group |
| 15650 | @group |
| 15651 | ;; check whether filename is that of a directory |
| 15652 | ((eq t (car (cdr (car current-directory-list)))) |
| 15653 | ;; decide whether to skip or recurse |
| 15654 | (if |
| 15655 | (equal "." |
| 15656 | (substring (car (car current-directory-list)) -1)) |
| 15657 | ;; then do nothing since filename is that of |
| 15658 | ;; current directory or parent, "." or ".." |
| 15659 | () |
| 15660 | @end group |
| 15661 | @group |
| 15662 | ;; else descend into the directory and repeat the process |
| 15663 | (setq el-files-list |
| 15664 | (append |
| 15665 | (files-in-below-directory |
| 15666 | (car (car current-directory-list))) |
| 15667 | el-files-list))))) |
| 15668 | ;; move to the next filename in the list; this also |
| 15669 | ;; shortens the list so the while loop eventually comes to an end |
| 15670 | (setq current-directory-list (cdr current-directory-list))) |
| 15671 | ;; return the filenames |
| 15672 | el-files-list)) |
| 15673 | @end group |
| 15674 | @end smallexample |
| 15675 | |
| 15676 | @c (files-in-below-directory "/usr/local/src/emacs/lisp/") |
| 15677 | @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/") |
| 15678 | |
| 15679 | The @code{files-in-below-directory} @code{directory-files} function |
| 15680 | takes one argument, the name of a directory. |
| 15681 | |
| 15682 | @need 1250 |
| 15683 | Thus, on my system, |
| 15684 | |
| 15685 | @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/")) |
| 15686 | |
| 15687 | @c !!! 22.1.1 lisp sources location here |
| 15688 | @smallexample |
| 15689 | @group |
| 15690 | (length |
| 15691 | (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")) |
| 15692 | @end group |
| 15693 | @end smallexample |
| 15694 | |
| 15695 | @noindent |
| 15696 | tells me that in and below my Lisp sources directory are 1031 |
| 15697 | @samp{.el} files. |
| 15698 | |
| 15699 | @code{files-in-below-directory} returns a list in reverse alphabetical |
| 15700 | order. An expression to sort the list in alphabetical order looks |
| 15701 | like this: |
| 15702 | |
| 15703 | @smallexample |
| 15704 | @group |
| 15705 | (sort |
| 15706 | (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/") |
| 15707 | 'string-lessp) |
| 15708 | @end group |
| 15709 | @end smallexample |
| 15710 | |
| 15711 | @ignore |
| 15712 | (defun test () |
| 15713 | "Test how long it takes to find lengths of all sorted elisp defuns." |
| 15714 | (insert "\n" (current-time-string) "\n") |
| 15715 | (sit-for 0) |
| 15716 | (sort |
| 15717 | (recursive-lengths-list-many-files |
| 15718 | (files-in-below-directory "/usr/local/src/emacs/lisp/")) |
| 15719 | '<) |
| 15720 | (insert (format "%s" (current-time-string)))) |
| 15721 | @end ignore |
| 15722 | |
| 15723 | @node Counting function definitions |
| 15724 | @subsection Counting function definitions |
| 15725 | |
| 15726 | Our immediate goal is to generate a list that tells us how many |
| 15727 | function definitions contain fewer than 10 words and symbols, how many |
| 15728 | contain between 10 and 19 words and symbols, how many contain between |
| 15729 | 20 and 29 words and symbols, and so on. |
| 15730 | |
| 15731 | With a sorted list of numbers, this is easy: count how many elements |
| 15732 | of the list are smaller than 10, then, after moving past the numbers |
| 15733 | just counted, count how many are smaller than 20, then, after moving |
| 15734 | past the numbers just counted, count how many are smaller than 30, and |
| 15735 | so on. Each of the numbers, 10, 20, 30, 40, and the like, is one |
| 15736 | larger than the top of that range. We can call the list of such |
| 15737 | numbers the @code{top-of-ranges} list. |
| 15738 | |
| 15739 | @need 1200 |
| 15740 | If we wished, we could generate this list automatically, but it is |
| 15741 | simpler to write a list manually. Here it is: |
| 15742 | @vindex top-of-ranges |
| 15743 | |
| 15744 | @smallexample |
| 15745 | @group |
| 15746 | (defvar top-of-ranges |
| 15747 | '(10 20 30 40 50 |
| 15748 | 60 70 80 90 100 |
| 15749 | 110 120 130 140 150 |
| 15750 | 160 170 180 190 200 |
| 15751 | 210 220 230 240 250 |
| 15752 | 260 270 280 290 300) |
| 15753 | "List specifying ranges for `defuns-per-range'.") |
| 15754 | @end group |
| 15755 | @end smallexample |
| 15756 | |
| 15757 | To change the ranges, we edit this list. |
| 15758 | |
| 15759 | Next, we need to write the function that creates the list of the |
| 15760 | number of definitions within each range. Clearly, this function must |
| 15761 | take the @code{sorted-lengths} and the @code{top-of-ranges} lists |
| 15762 | as arguments. |
| 15763 | |
| 15764 | The @code{defuns-per-range} function must do two things again and |
| 15765 | again: it must count the number of definitions within a range |
| 15766 | specified by the current top-of-range value; and it must shift to the |
| 15767 | next higher value in the @code{top-of-ranges} list after counting the |
| 15768 | number of definitions in the current range. Since each of these |
| 15769 | actions is repetitive, we can use @code{while} loops for the job. |
| 15770 | One loop counts the number of definitions in the range defined by the |
| 15771 | current top-of-range value, and the other loop selects each of the |
| 15772 | top-of-range values in turn. |
| 15773 | |
| 15774 | Several entries of the @code{sorted-lengths} list are counted for each |
| 15775 | range; this means that the loop for the @code{sorted-lengths} list |
| 15776 | will be inside the loop for the @code{top-of-ranges} list, like a |
| 15777 | small gear inside a big gear. |
| 15778 | |
| 15779 | The inner loop counts the number of definitions within the range. It |
| 15780 | is a simple counting loop of the type we have seen before. |
| 15781 | (@xref{Incrementing Loop, , A loop with an incrementing counter}.) |
| 15782 | The true-or-false test of the loop tests whether the value from the |
| 15783 | @code{sorted-lengths} list is smaller than the current value of the |
| 15784 | top of the range. If it is, the function increments the counter and |
| 15785 | tests the next value from the @code{sorted-lengths} list. |
| 15786 | |
| 15787 | @need 1250 |
| 15788 | The inner loop looks like this: |
| 15789 | |
| 15790 | @smallexample |
| 15791 | @group |
| 15792 | (while @var{length-element-smaller-than-top-of-range} |
| 15793 | (setq number-within-range (1+ number-within-range)) |
| 15794 | (setq sorted-lengths (cdr sorted-lengths))) |
| 15795 | @end group |
| 15796 | @end smallexample |
| 15797 | |
| 15798 | The outer loop must start with the lowest value of the |
| 15799 | @code{top-of-ranges} list, and then be set to each of the succeeding |
| 15800 | higher values in turn. This can be done with a loop like this: |
| 15801 | |
| 15802 | @smallexample |
| 15803 | @group |
| 15804 | (while top-of-ranges |
| 15805 | @var{body-of-loop}@dots{} |
| 15806 | (setq top-of-ranges (cdr top-of-ranges))) |
| 15807 | @end group |
| 15808 | @end smallexample |
| 15809 | |
| 15810 | @need 1200 |
| 15811 | Put together, the two loops look like this: |
| 15812 | |
| 15813 | @smallexample |
| 15814 | @group |
| 15815 | (while top-of-ranges |
| 15816 | |
| 15817 | ;; @r{Count the number of elements within the current range.} |
| 15818 | (while @var{length-element-smaller-than-top-of-range} |
| 15819 | (setq number-within-range (1+ number-within-range)) |
| 15820 | (setq sorted-lengths (cdr sorted-lengths))) |
| 15821 | |
| 15822 | ;; @r{Move to next range.} |
| 15823 | (setq top-of-ranges (cdr top-of-ranges))) |
| 15824 | @end group |
| 15825 | @end smallexample |
| 15826 | |
| 15827 | In addition, in each circuit of the outer loop, Emacs should record |
| 15828 | the number of definitions within that range (the value of |
| 15829 | @code{number-within-range}) in a list. We can use @code{cons} for |
| 15830 | this purpose. (@xref{cons, , @code{cons}}.) |
| 15831 | |
| 15832 | The @code{cons} function works fine, except that the list it |
| 15833 | constructs will contain the number of definitions for the highest |
| 15834 | range at its beginning and the number of definitions for the lowest |
| 15835 | range at its end. This is because @code{cons} attaches new elements |
| 15836 | of the list to the beginning of the list, and since the two loops are |
| 15837 | working their way through the lengths' list from the lower end first, |
| 15838 | the @code{defuns-per-range-list} will end up largest number first. |
| 15839 | But we will want to print our graph with smallest values first and the |
| 15840 | larger later. The solution is to reverse the order of the |
| 15841 | @code{defuns-per-range-list}. We can do this using the |
| 15842 | @code{nreverse} function, which reverses the order of a list. |
| 15843 | @findex nreverse |
| 15844 | |
| 15845 | @need 800 |
| 15846 | For example, |
| 15847 | |
| 15848 | @smallexample |
| 15849 | (nreverse '(1 2 3 4)) |
| 15850 | @end smallexample |
| 15851 | |
| 15852 | @need 800 |
| 15853 | @noindent |
| 15854 | produces: |
| 15855 | |
| 15856 | @smallexample |
| 15857 | (4 3 2 1) |
| 15858 | @end smallexample |
| 15859 | |
| 15860 | Note that the @code{nreverse} function is ``destructive''---that is, |
| 15861 | it changes the list to which it is applied; this contrasts with the |
| 15862 | @code{car} and @code{cdr} functions, which are non-destructive. In |
| 15863 | this case, we do not want the original @code{defuns-per-range-list}, |
| 15864 | so it does not matter that it is destroyed. (The @code{reverse} |
| 15865 | function provides a reversed copy of a list, leaving the original list |
| 15866 | as is.) |
| 15867 | @findex reverse |
| 15868 | |
| 15869 | @need 1250 |
| 15870 | Put all together, the @code{defuns-per-range} looks like this: |
| 15871 | |
| 15872 | @smallexample |
| 15873 | @group |
| 15874 | (defun defuns-per-range (sorted-lengths top-of-ranges) |
| 15875 | "SORTED-LENGTHS defuns in each TOP-OF-RANGES range." |
| 15876 | (let ((top-of-range (car top-of-ranges)) |
| 15877 | (number-within-range 0) |
| 15878 | defuns-per-range-list) |
| 15879 | @end group |
| 15880 | |
| 15881 | @group |
| 15882 | ;; @r{Outer loop.} |
| 15883 | (while top-of-ranges |
| 15884 | @end group |
| 15885 | |
| 15886 | @group |
| 15887 | ;; @r{Inner loop.} |
| 15888 | (while (and |
| 15889 | ;; @r{Need number for numeric test.} |
| 15890 | (car sorted-lengths) |
| 15891 | (< (car sorted-lengths) top-of-range)) |
| 15892 | @end group |
| 15893 | |
| 15894 | @group |
| 15895 | ;; @r{Count number of definitions within current range.} |
| 15896 | (setq number-within-range (1+ number-within-range)) |
| 15897 | (setq sorted-lengths (cdr sorted-lengths))) |
| 15898 | |
| 15899 | ;; @r{Exit inner loop but remain within outer loop.} |
| 15900 | @end group |
| 15901 | |
| 15902 | @group |
| 15903 | (setq defuns-per-range-list |
| 15904 | (cons number-within-range defuns-per-range-list)) |
| 15905 | (setq number-within-range 0) ; @r{Reset count to zero.} |
| 15906 | @end group |
| 15907 | |
| 15908 | @group |
| 15909 | ;; @r{Move to next range.} |
| 15910 | (setq top-of-ranges (cdr top-of-ranges)) |
| 15911 | ;; @r{Specify next top of range value.} |
| 15912 | (setq top-of-range (car top-of-ranges))) |
| 15913 | @end group |
| 15914 | |
| 15915 | @group |
| 15916 | ;; @r{Exit outer loop and count the number of defuns larger than} |
| 15917 | ;; @r{ the largest top-of-range value.} |
| 15918 | (setq defuns-per-range-list |
| 15919 | (cons |
| 15920 | (length sorted-lengths) |
| 15921 | defuns-per-range-list)) |
| 15922 | @end group |
| 15923 | |
| 15924 | @group |
| 15925 | ;; @r{Return a list of the number of definitions within each range,} |
| 15926 | ;; @r{ smallest to largest.} |
| 15927 | (nreverse defuns-per-range-list))) |
| 15928 | @end group |
| 15929 | @end smallexample |
| 15930 | |
| 15931 | @need 1200 |
| 15932 | @noindent |
| 15933 | The function is straightforward except for one subtle feature. The |
| 15934 | true-or-false test of the inner loop looks like this: |
| 15935 | |
| 15936 | @smallexample |
| 15937 | @group |
| 15938 | (and (car sorted-lengths) |
| 15939 | (< (car sorted-lengths) top-of-range)) |
| 15940 | @end group |
| 15941 | @end smallexample |
| 15942 | |
| 15943 | @need 800 |
| 15944 | @noindent |
| 15945 | instead of like this: |
| 15946 | |
| 15947 | @smallexample |
| 15948 | (< (car sorted-lengths) top-of-range) |
| 15949 | @end smallexample |
| 15950 | |
| 15951 | The purpose of the test is to determine whether the first item in the |
| 15952 | @code{sorted-lengths} list is less than the value of the top of the |
| 15953 | range. |
| 15954 | |
| 15955 | The simple version of the test works fine unless the |
| 15956 | @code{sorted-lengths} list has a @code{nil} value. In that case, the |
| 15957 | @code{(car sorted-lengths)} expression function returns |
| 15958 | @code{nil}. The @code{<} function cannot compare a number to |
| 15959 | @code{nil}, which is an empty list, so Emacs signals an error and |
| 15960 | stops the function from attempting to continue to execute. |
| 15961 | |
| 15962 | The @code{sorted-lengths} list always becomes @code{nil} when the |
| 15963 | counter reaches the end of the list. This means that any attempt to |
| 15964 | use the @code{defuns-per-range} function with the simple version of |
| 15965 | the test will fail. |
| 15966 | |
| 15967 | We solve the problem by using the @code{(car sorted-lengths)} |
| 15968 | expression in conjunction with the @code{and} expression. The |
| 15969 | @code{(car sorted-lengths)} expression returns a non-@code{nil} |
| 15970 | value so long as the list has at least one number within it, but |
| 15971 | returns @code{nil} if the list is empty. The @code{and} expression |
| 15972 | first evaluates the @code{(car sorted-lengths)} expression, and |
| 15973 | if it is @code{nil}, returns false @emph{without} evaluating the |
| 15974 | @code{<} expression. But if the @code{(car sorted-lengths)} |
| 15975 | expression returns a non-@code{nil} value, the @code{and} expression |
| 15976 | evaluates the @code{<} expression, and returns that value as the value |
| 15977 | of the @code{and} expression. |
| 15978 | |
| 15979 | @c colon in printed section title causes problem in Info cross reference |
| 15980 | This way, we avoid an error. |
| 15981 | @iftex |
| 15982 | @noindent |
| 15983 | (For information about @code{and}, see |
| 15984 | @ref{kill-new function, , The @code{kill-new} function}.) |
| 15985 | @end iftex |
| 15986 | @ifinfo |
| 15987 | @noindent |
| 15988 | (@xref{kill-new function, , The @code{kill-new} function}, for |
| 15989 | information about @code{and}.) |
| 15990 | @end ifinfo |
| 15991 | |
| 15992 | Here is a short test of the @code{defuns-per-range} function. First, |
| 15993 | evaluate the expression that binds (a shortened) |
| 15994 | @code{top-of-ranges} list to the list of values, then evaluate the |
| 15995 | expression for binding the @code{sorted-lengths} list, and then |
| 15996 | evaluate the @code{defuns-per-range} function. |
| 15997 | |
| 15998 | @smallexample |
| 15999 | @group |
| 16000 | ;; @r{(Shorter list than we will use later.)} |
| 16001 | (setq top-of-ranges |
| 16002 | '(110 120 130 140 150 |
| 16003 | 160 170 180 190 200)) |
| 16004 | |
| 16005 | (setq sorted-lengths |
| 16006 | '(85 86 110 116 122 129 154 176 179 200 265 300 300)) |
| 16007 | |
| 16008 | (defuns-per-range sorted-lengths top-of-ranges) |
| 16009 | @end group |
| 16010 | @end smallexample |
| 16011 | |
| 16012 | @need 800 |
| 16013 | @noindent |
| 16014 | The list returned looks like this: |
| 16015 | |
| 16016 | @smallexample |
| 16017 | (2 2 2 0 0 1 0 2 0 0 4) |
| 16018 | @end smallexample |
| 16019 | |
| 16020 | @noindent |
| 16021 | Indeed, there are two elements of the @code{sorted-lengths} list |
| 16022 | smaller than 110, two elements between 110 and 119, two elements |
| 16023 | between 120 and 129, and so on. There are four elements with a value |
| 16024 | of 200 or larger. |
| 16025 | |
| 16026 | @c The next step is to turn this numbers' list into a graph. |
| 16027 | @node Readying a Graph |
| 16028 | @chapter Readying a Graph |
| 16029 | @cindex Readying a graph |
| 16030 | @cindex Graph prototype |
| 16031 | @cindex Prototype graph |
| 16032 | @cindex Body of graph |
| 16033 | |
| 16034 | Our goal is to construct a graph showing the numbers of function |
| 16035 | definitions of various lengths in the Emacs lisp sources. |
| 16036 | |
| 16037 | As a practical matter, if you were creating a graph, you would |
| 16038 | probably use a program such as @code{gnuplot} to do the job. |
| 16039 | (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case, |
| 16040 | however, we create one from scratch, and in the process we will |
| 16041 | re-acquaint ourselves with some of what we learned before and learn |
| 16042 | more. |
| 16043 | |
| 16044 | In this chapter, we will first write a simple graph printing function. |
| 16045 | This first definition will be a @dfn{prototype}, a rapidly written |
| 16046 | function that enables us to reconnoiter this unknown graph-making |
| 16047 | territory. We will discover dragons, or find that they are myth. |
| 16048 | After scouting the terrain, we will feel more confident and enhance |
| 16049 | the function to label the axes automatically. |
| 16050 | |
| 16051 | @menu |
| 16052 | * Columns of a graph:: |
| 16053 | * graph-body-print:: How to print the body of a graph. |
| 16054 | * recursive-graph-body-print:: |
| 16055 | * Printed Axes:: |
| 16056 | * Line Graph Exercise:: |
| 16057 | @end menu |
| 16058 | |
| 16059 | @ifnottex |
| 16060 | @node Columns of a graph |
| 16061 | @unnumberedsec Printing the Columns of a Graph |
| 16062 | @end ifnottex |
| 16063 | |
| 16064 | Since Emacs is designed to be flexible and work with all kinds of |
| 16065 | terminals, including character-only terminals, the graph will need to |
| 16066 | be made from one of the `typewriter' symbols. An asterisk will do; as |
| 16067 | we enhance the graph-printing function, we can make the choice of |
| 16068 | symbol a user option. |
| 16069 | |
| 16070 | We can call this function @code{graph-body-print}; it will take a |
| 16071 | @code{numbers-list} as its only argument. At this stage, we will not |
| 16072 | label the graph, but only print its body. |
| 16073 | |
| 16074 | The @code{graph-body-print} function inserts a vertical column of |
| 16075 | asterisks for each element in the @code{numbers-list}. The height of |
| 16076 | each line is determined by the value of that element of the |
| 16077 | @code{numbers-list}. |
| 16078 | |
| 16079 | Inserting columns is a repetitive act; that means that this function can |
| 16080 | be written either with a @code{while} loop or recursively. |
| 16081 | |
| 16082 | Our first challenge is to discover how to print a column of asterisks. |
| 16083 | Usually, in Emacs, we print characters onto a screen horizontally, |
| 16084 | line by line, by typing. We have two routes we can follow: write our |
| 16085 | own column-insertion function or discover whether one exists in Emacs. |
| 16086 | |
| 16087 | To see whether there is one in Emacs, we can use the @kbd{M-x apropos} |
| 16088 | command. This command is like the @kbd{C-h a} (@code{command-apropos}) |
| 16089 | command, except that the latter finds only those functions that are |
| 16090 | commands. The @kbd{M-x apropos} command lists all symbols that match |
| 16091 | a regular expression, including functions that are not interactive. |
| 16092 | @findex apropos |
| 16093 | |
| 16094 | What we want to look for is some command that prints or inserts |
| 16095 | columns. Very likely, the name of the function will contain either |
| 16096 | the word `print' or the word `insert' or the word `column'. |
| 16097 | Therefore, we can simply type @kbd{M-x apropos RET |
| 16098 | print\|insert\|column RET} and look at the result. On my system, this |
| 16099 | command once too takes quite some time, and then produced a list of 79 |
| 16100 | functions and variables. Now it does not take much time at all and |
| 16101 | produces a list of 211 functions and variables. Scanning down the |
| 16102 | list, the only function that looks as if it might do the job is |
| 16103 | @code{insert-rectangle}. |
| 16104 | |
| 16105 | @need 1200 |
| 16106 | Indeed, this is the function we want; its documentation says: |
| 16107 | |
| 16108 | @smallexample |
| 16109 | @group |
| 16110 | insert-rectangle: |
| 16111 | Insert text of RECTANGLE with upper left corner at point. |
| 16112 | RECTANGLE's first line is inserted at point, |
| 16113 | its second line is inserted at a point vertically under point, etc. |
| 16114 | RECTANGLE should be a list of strings. |
| 16115 | After this command, the mark is at the upper left corner |
| 16116 | and point is at the lower right corner. |
| 16117 | @end group |
| 16118 | @end smallexample |
| 16119 | |
| 16120 | We can run a quick test, to make sure it does what we expect of it. |
| 16121 | |
| 16122 | Here is the result of placing the cursor after the |
| 16123 | @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e} |
| 16124 | (@code{eval-last-sexp}). The function inserts the strings |
| 16125 | @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below |
| 16126 | point. Also the function returns @code{nil}. |
| 16127 | |
| 16128 | @smallexample |
| 16129 | @group |
| 16130 | (insert-rectangle '("first" "second" "third"))first |
| 16131 | second |
| 16132 | thirdnil |
| 16133 | @end group |
| 16134 | @end smallexample |
| 16135 | |
| 16136 | @noindent |
| 16137 | Of course, we won't be inserting the text of the |
| 16138 | @code{insert-rectangle} expression itself into the buffer in which we |
| 16139 | are making the graph, but will call the function from our program. We |
| 16140 | shall, however, have to make sure that point is in the buffer at the |
| 16141 | place where the @code{insert-rectangle} function will insert its |
| 16142 | column of strings. |
| 16143 | |
| 16144 | If you are reading this in Info, you can see how this works by |
| 16145 | switching to another buffer, such as the @file{*scratch*} buffer, |
| 16146 | placing point somewhere in the buffer, typing @kbd{M-:}, typing the |
| 16147 | @code{insert-rectangle} expression into the minibuffer at the prompt, |
| 16148 | and then typing @key{RET}. This causes Emacs to evaluate the |
| 16149 | expression in the minibuffer, but to use as the value of point the |
| 16150 | position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the |
| 16151 | keybinding for @code{eval-expression}. Also, @code{nil} does not |
| 16152 | appear in the @file{*scratch*} buffer since the expression is |
| 16153 | evaluated in the minibuffer.) |
| 16154 | |
| 16155 | We find when we do this that point ends up at the end of the last |
| 16156 | inserted line---that is to say, this function moves point as a |
| 16157 | side-effect. If we were to repeat the command, with point at this |
| 16158 | position, the next insertion would be below and to the right of the |
| 16159 | previous insertion. We don't want this! If we are going to make a |
| 16160 | bar graph, the columns need to be beside each other. |
| 16161 | |
| 16162 | So we discover that each cycle of the column-inserting @code{while} |
| 16163 | loop must reposition point to the place we want it, and that place |
| 16164 | will be at the top, not the bottom, of the column. Moreover, we |
| 16165 | remember that when we print a graph, we do not expect all the columns |
| 16166 | to be the same height. This means that the top of each column may be |
| 16167 | at a different height from the previous one. We cannot simply |
| 16168 | reposition point to the same line each time, but moved over to the |
| 16169 | right---or perhaps we can@dots{} |
| 16170 | |
| 16171 | We are planning to make the columns of the bar graph out of asterisks. |
| 16172 | The number of asterisks in the column is the number specified by the |
| 16173 | current element of the @code{numbers-list}. We need to construct a |
| 16174 | list of asterisks of the right length for each call to |
| 16175 | @code{insert-rectangle}. If this list consists solely of the requisite |
| 16176 | number of asterisks, then we will have position point the right number |
| 16177 | of lines above the base for the graph to print correctly. This could |
| 16178 | be difficult. |
| 16179 | |
| 16180 | Alternatively, if we can figure out some way to pass |
| 16181 | @code{insert-rectangle} a list of the same length each time, then we |
| 16182 | can place point on the same line each time, but move it over one |
| 16183 | column to the right for each new column. If we do this, however, some |
| 16184 | of the entries in the list passed to @code{insert-rectangle} must be |
| 16185 | blanks rather than asterisks. For example, if the maximum height of |
| 16186 | the graph is 5, but the height of the column is 3, then |
| 16187 | @code{insert-rectangle} requires an argument that looks like this: |
| 16188 | |
| 16189 | @smallexample |
| 16190 | (" " " " "*" "*" "*") |
| 16191 | @end smallexample |
| 16192 | |
| 16193 | This last proposal is not so difficult, so long as we can determine |
| 16194 | the column height. There are two ways for us to specify the column |
| 16195 | height: we can arbitrarily state what it will be, which would work |
| 16196 | fine for graphs of that height; or we can search through the list of |
| 16197 | numbers and use the maximum height of the list as the maximum height |
| 16198 | of the graph. If the latter operation were difficult, then the former |
| 16199 | procedure would be easiest, but there is a function built into Emacs |
| 16200 | that determines the maximum of its arguments. We can use that |
| 16201 | function. The function is called @code{max} and it returns the |
| 16202 | largest of all its arguments, which must be numbers. Thus, for |
| 16203 | example, |
| 16204 | |
| 16205 | @smallexample |
| 16206 | (max 3 4 6 5 7 3) |
| 16207 | @end smallexample |
| 16208 | |
| 16209 | @noindent |
| 16210 | returns 7. (A corresponding function called @code{min} returns the |
| 16211 | smallest of all its arguments.) |
| 16212 | @findex max |
| 16213 | @findex min |
| 16214 | |
| 16215 | However, we cannot simply call @code{max} on the @code{numbers-list}; |
| 16216 | the @code{max} function expects numbers as its argument, not a list of |
| 16217 | numbers. Thus, the following expression, |
| 16218 | |
| 16219 | @smallexample |
| 16220 | (max '(3 4 6 5 7 3)) |
| 16221 | @end smallexample |
| 16222 | |
| 16223 | @need 800 |
| 16224 | @noindent |
| 16225 | produces the following error message; |
| 16226 | |
| 16227 | @smallexample |
| 16228 | Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3) |
| 16229 | @end smallexample |
| 16230 | |
| 16231 | @findex apply |
| 16232 | We need a function that passes a list of arguments to a function. |
| 16233 | This function is @code{apply}. This function `applies' its first |
| 16234 | argument (a function) to its remaining arguments, the last of which |
| 16235 | may be a list. |
| 16236 | |
| 16237 | @need 1250 |
| 16238 | For example, |
| 16239 | |
| 16240 | @smallexample |
| 16241 | (apply 'max 3 4 7 3 '(4 8 5)) |
| 16242 | @end smallexample |
| 16243 | |
| 16244 | @noindent |
| 16245 | returns 8. |
| 16246 | |
| 16247 | (Incidentally, I don't know how you would learn of this function |
| 16248 | without a book such as this. It is possible to discover other |
| 16249 | functions, like @code{search-forward} or @code{insert-rectangle}, by |
| 16250 | guessing at a part of their names and then using @code{apropos}. Even |
| 16251 | though its base in metaphor is clear---`apply' its first argument to |
| 16252 | the rest---I doubt a novice would come up with that particular word |
| 16253 | when using @code{apropos} or other aid. Of course, I could be wrong; |
| 16254 | after all, the function was first named by someone who had to invent |
| 16255 | it.) |
| 16256 | |
| 16257 | The second and subsequent arguments to @code{apply} are optional, so |
| 16258 | we can use @code{apply} to call a function and pass the elements of a |
| 16259 | list to it, like this, which also returns 8: |
| 16260 | |
| 16261 | @smallexample |
| 16262 | (apply 'max '(4 8 5)) |
| 16263 | @end smallexample |
| 16264 | |
| 16265 | This latter way is how we will use @code{apply}. The |
| 16266 | @code{recursive-lengths-list-many-files} function returns a numbers' |
| 16267 | list to which we can apply @code{max} (we could also apply @code{max} to |
| 16268 | the sorted numbers' list; it does not matter whether the list is |
| 16269 | sorted or not.) |
| 16270 | |
| 16271 | @need 800 |
| 16272 | Hence, the operation for finding the maximum height of the graph is this: |
| 16273 | |
| 16274 | @smallexample |
| 16275 | (setq max-graph-height (apply 'max numbers-list)) |
| 16276 | @end smallexample |
| 16277 | |
| 16278 | Now we can return to the question of how to create a list of strings |
| 16279 | for a column of the graph. Told the maximum height of the graph |
| 16280 | and the number of asterisks that should appear in the column, the |
| 16281 | function should return a list of strings for the |
| 16282 | @code{insert-rectangle} command to insert. |
| 16283 | |
| 16284 | Each column is made up of asterisks or blanks. Since the function is |
| 16285 | passed the value of the height of the column and the number of |
| 16286 | asterisks in the column, the number of blanks can be found by |
| 16287 | subtracting the number of asterisks from the height of the column. |
| 16288 | Given the number of blanks and the number of asterisks, two |
| 16289 | @code{while} loops can be used to construct the list: |
| 16290 | |
| 16291 | @smallexample |
| 16292 | @group |
| 16293 | ;;; @r{First version.} |
| 16294 | (defun column-of-graph (max-graph-height actual-height) |
| 16295 | "Return list of strings that is one column of a graph." |
| 16296 | (let ((insert-list nil) |
| 16297 | (number-of-top-blanks |
| 16298 | (- max-graph-height actual-height))) |
| 16299 | @end group |
| 16300 | |
| 16301 | @group |
| 16302 | ;; @r{Fill in asterisks.} |
| 16303 | (while (> actual-height 0) |
| 16304 | (setq insert-list (cons "*" insert-list)) |
| 16305 | (setq actual-height (1- actual-height))) |
| 16306 | @end group |
| 16307 | |
| 16308 | @group |
| 16309 | ;; @r{Fill in blanks.} |
| 16310 | (while (> number-of-top-blanks 0) |
| 16311 | (setq insert-list (cons " " insert-list)) |
| 16312 | (setq number-of-top-blanks |
| 16313 | (1- number-of-top-blanks))) |
| 16314 | @end group |
| 16315 | |
| 16316 | @group |
| 16317 | ;; @r{Return whole list.} |
| 16318 | insert-list)) |
| 16319 | @end group |
| 16320 | @end smallexample |
| 16321 | |
| 16322 | If you install this function and then evaluate the following |
| 16323 | expression you will see that it returns the list as desired: |
| 16324 | |
| 16325 | @smallexample |
| 16326 | (column-of-graph 5 3) |
| 16327 | @end smallexample |
| 16328 | |
| 16329 | @need 800 |
| 16330 | @noindent |
| 16331 | returns |
| 16332 | |
| 16333 | @smallexample |
| 16334 | (" " " " "*" "*" "*") |
| 16335 | @end smallexample |
| 16336 | |
| 16337 | As written, @code{column-of-graph} contains a major flaw: the symbols |
| 16338 | used for the blank and for the marked entries in the column are |
| 16339 | `hard-coded' as a space and asterisk. This is fine for a prototype, |
| 16340 | but you, or another user, may wish to use other symbols. For example, |
| 16341 | in testing the graph function, you many want to use a period in place |
| 16342 | of the space, to make sure the point is being repositioned properly |
| 16343 | each time the @code{insert-rectangle} function is called; or you might |
| 16344 | want to substitute a @samp{+} sign or other symbol for the asterisk. |
| 16345 | You might even want to make a graph-column that is more than one |
| 16346 | display column wide. The program should be more flexible. The way to |
| 16347 | do that is to replace the blank and the asterisk with two variables |
| 16348 | that we can call @code{graph-blank} and @code{graph-symbol} and define |
| 16349 | those variables separately. |
| 16350 | |
| 16351 | Also, the documentation is not well written. These considerations |
| 16352 | lead us to the second version of the function: |
| 16353 | |
| 16354 | @smallexample |
| 16355 | @group |
| 16356 | (defvar graph-symbol "*" |
| 16357 | "String used as symbol in graph, usually an asterisk.") |
| 16358 | @end group |
| 16359 | |
| 16360 | @group |
| 16361 | (defvar graph-blank " " |
| 16362 | "String used as blank in graph, usually a blank space. |
| 16363 | graph-blank must be the same number of columns wide |
| 16364 | as graph-symbol.") |
| 16365 | @end group |
| 16366 | @end smallexample |
| 16367 | |
| 16368 | @noindent |
| 16369 | (For an explanation of @code{defvar}, see |
| 16370 | @ref{defvar, , Initializing a Variable with @code{defvar}}.) |
| 16371 | |
| 16372 | @smallexample |
| 16373 | @group |
| 16374 | ;;; @r{Second version.} |
| 16375 | (defun column-of-graph (max-graph-height actual-height) |
| 16376 | "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols. |
| 16377 | |
| 16378 | @end group |
| 16379 | @group |
| 16380 | The graph-symbols are contiguous entries at the end |
| 16381 | of the list. |
| 16382 | The list will be inserted as one column of a graph. |
| 16383 | The strings are either graph-blank or graph-symbol." |
| 16384 | @end group |
| 16385 | |
| 16386 | @group |
| 16387 | (let ((insert-list nil) |
| 16388 | (number-of-top-blanks |
| 16389 | (- max-graph-height actual-height))) |
| 16390 | @end group |
| 16391 | |
| 16392 | @group |
| 16393 | ;; @r{Fill in @code{graph-symbols}.} |
| 16394 | (while (> actual-height 0) |
| 16395 | (setq insert-list (cons graph-symbol insert-list)) |
| 16396 | (setq actual-height (1- actual-height))) |
| 16397 | @end group |
| 16398 | |
| 16399 | @group |
| 16400 | ;; @r{Fill in @code{graph-blanks}.} |
| 16401 | (while (> number-of-top-blanks 0) |
| 16402 | (setq insert-list (cons graph-blank insert-list)) |
| 16403 | (setq number-of-top-blanks |
| 16404 | (1- number-of-top-blanks))) |
| 16405 | |
| 16406 | ;; @r{Return whole list.} |
| 16407 | insert-list)) |
| 16408 | @end group |
| 16409 | @end smallexample |
| 16410 | |
| 16411 | If we wished, we could rewrite @code{column-of-graph} a third time to |
| 16412 | provide optionally for a line graph as well as for a bar graph. This |
| 16413 | would not be hard to do. One way to think of a line graph is that it |
| 16414 | is no more than a bar graph in which the part of each bar that is |
| 16415 | below the top is blank. To construct a column for a line graph, the |
| 16416 | function first constructs a list of blanks that is one shorter than |
| 16417 | the value, then it uses @code{cons} to attach a graph symbol to the |
| 16418 | list; then it uses @code{cons} again to attach the `top blanks' to |
| 16419 | the list. |
| 16420 | |
| 16421 | It is easy to see how to write such a function, but since we don't |
| 16422 | need it, we will not do it. But the job could be done, and if it were |
| 16423 | done, it would be done with @code{column-of-graph}. Even more |
| 16424 | important, it is worth noting that few changes would have to be made |
| 16425 | anywhere else. The enhancement, if we ever wish to make it, is |
| 16426 | simple. |
| 16427 | |
| 16428 | Now, finally, we come to our first actual graph printing function. |
| 16429 | This prints the body of a graph, not the labels for the vertical and |
| 16430 | horizontal axes, so we can call this @code{graph-body-print}. |
| 16431 | |
| 16432 | @node graph-body-print |
| 16433 | @section The @code{graph-body-print} Function |
| 16434 | @findex graph-body-print |
| 16435 | |
| 16436 | After our preparation in the preceding section, the |
| 16437 | @code{graph-body-print} function is straightforward. The function |
| 16438 | will print column after column of asterisks and blanks, using the |
| 16439 | elements of a numbers' list to specify the number of asterisks in each |
| 16440 | column. This is a repetitive act, which means we can use a |
| 16441 | decrementing @code{while} loop or recursive function for the job. In |
| 16442 | this section, we will write the definition using a @code{while} loop. |
| 16443 | |
| 16444 | The @code{column-of-graph} function requires the height of the graph |
| 16445 | as an argument, so we should determine and record that as a local variable. |
| 16446 | |
| 16447 | This leads us to the following template for the @code{while} loop |
| 16448 | version of this function: |
| 16449 | |
| 16450 | @smallexample |
| 16451 | @group |
| 16452 | (defun graph-body-print (numbers-list) |
| 16453 | "@var{documentation}@dots{}" |
| 16454 | (let ((height @dots{} |
| 16455 | @dots{})) |
| 16456 | @end group |
| 16457 | |
| 16458 | @group |
| 16459 | (while numbers-list |
| 16460 | @var{insert-columns-and-reposition-point} |
| 16461 | (setq numbers-list (cdr numbers-list))))) |
| 16462 | @end group |
| 16463 | @end smallexample |
| 16464 | |
| 16465 | @noindent |
| 16466 | We need to fill in the slots of the template. |
| 16467 | |
| 16468 | Clearly, we can use the @code{(apply 'max numbers-list)} expression to |
| 16469 | determine the height of the graph. |
| 16470 | |
| 16471 | The @code{while} loop will cycle through the @code{numbers-list} one |
| 16472 | element at a time. As it is shortened by the @code{(setq numbers-list |
| 16473 | (cdr numbers-list))} expression, the @sc{car} of each instance of the |
| 16474 | list is the value of the argument for @code{column-of-graph}. |
| 16475 | |
| 16476 | At each cycle of the @code{while} loop, the @code{insert-rectangle} |
| 16477 | function inserts the list returned by @code{column-of-graph}. Since |
| 16478 | the @code{insert-rectangle} function moves point to the lower right of |
| 16479 | the inserted rectangle, we need to save the location of point at the |
| 16480 | time the rectangle is inserted, move back to that position after the |
| 16481 | rectangle is inserted, and then move horizontally to the next place |
| 16482 | from which @code{insert-rectangle} is called. |
| 16483 | |
| 16484 | If the inserted columns are one character wide, as they will be if |
| 16485 | single blanks and asterisks are used, the repositioning command is |
| 16486 | simply @code{(forward-char 1)}; however, the width of a column may be |
| 16487 | greater than one. This means that the repositioning command should be |
| 16488 | written @code{(forward-char symbol-width)}. The @code{symbol-width} |
| 16489 | itself is the length of a @code{graph-blank} and can be found using |
| 16490 | the expression @code{(length graph-blank)}. The best place to bind |
| 16491 | the @code{symbol-width} variable to the value of the width of graph |
| 16492 | column is in the varlist of the @code{let} expression. |
| 16493 | |
| 16494 | @need 1250 |
| 16495 | These considerations lead to the following function definition: |
| 16496 | |
| 16497 | @smallexample |
| 16498 | @group |
| 16499 | (defun graph-body-print (numbers-list) |
| 16500 | "Print a bar graph of the NUMBERS-LIST. |
| 16501 | The numbers-list consists of the Y-axis values." |
| 16502 | |
| 16503 | (let ((height (apply 'max numbers-list)) |
| 16504 | (symbol-width (length graph-blank)) |
| 16505 | from-position) |
| 16506 | @end group |
| 16507 | |
| 16508 | @group |
| 16509 | (while numbers-list |
| 16510 | (setq from-position (point)) |
| 16511 | (insert-rectangle |
| 16512 | (column-of-graph height (car numbers-list))) |
| 16513 | (goto-char from-position) |
| 16514 | (forward-char symbol-width) |
| 16515 | @end group |
| 16516 | @group |
| 16517 | ;; @r{Draw graph column by column.} |
| 16518 | (sit-for 0) |
| 16519 | (setq numbers-list (cdr numbers-list))) |
| 16520 | @end group |
| 16521 | @group |
| 16522 | ;; @r{Place point for X axis labels.} |
| 16523 | (forward-line height) |
| 16524 | (insert "\n") |
| 16525 | )) |
| 16526 | @end group |
| 16527 | @end smallexample |
| 16528 | |
| 16529 | @noindent |
| 16530 | The one unexpected expression in this function is the |
| 16531 | @w{@code{(sit-for 0)}} expression in the @code{while} loop. This |
| 16532 | expression makes the graph printing operation more interesting to |
| 16533 | watch than it would be otherwise. The expression causes Emacs to |
| 16534 | `sit' or do nothing for a zero length of time and then redraw the |
| 16535 | screen. Placed here, it causes Emacs to redraw the screen column by |
| 16536 | column. Without it, Emacs would not redraw the screen until the |
| 16537 | function exits. |
| 16538 | |
| 16539 | We can test @code{graph-body-print} with a short list of numbers. |
| 16540 | |
| 16541 | @enumerate |
| 16542 | @item |
| 16543 | Install @code{graph-symbol}, @code{graph-blank}, |
| 16544 | @code{column-of-graph}, which are in |
| 16545 | @iftex |
| 16546 | @ref{Readying a Graph, , Readying a Graph}, |
| 16547 | @end iftex |
| 16548 | @ifinfo |
| 16549 | @ref{Columns of a graph}, |
| 16550 | @end ifinfo |
| 16551 | and @code{graph-body-print}. |
| 16552 | |
| 16553 | @need 800 |
| 16554 | @item |
| 16555 | Copy the following expression: |
| 16556 | |
| 16557 | @smallexample |
| 16558 | (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3)) |
| 16559 | @end smallexample |
| 16560 | |
| 16561 | @item |
| 16562 | Switch to the @file{*scratch*} buffer and place the cursor where you |
| 16563 | want the graph to start. |
| 16564 | |
| 16565 | @item |
| 16566 | Type @kbd{M-:} (@code{eval-expression}). |
| 16567 | |
| 16568 | @item |
| 16569 | Yank the @code{graph-body-print} expression into the minibuffer |
| 16570 | with @kbd{C-y} (@code{yank)}. |
| 16571 | |
| 16572 | @item |
| 16573 | Press @key{RET} to evaluate the @code{graph-body-print} expression. |
| 16574 | @end enumerate |
| 16575 | |
| 16576 | @need 800 |
| 16577 | Emacs will print a graph like this: |
| 16578 | |
| 16579 | @smallexample |
| 16580 | @group |
| 16581 | * |
| 16582 | * ** |
| 16583 | * **** |
| 16584 | *** **** |
| 16585 | ********* * |
| 16586 | ************ |
| 16587 | ************* |
| 16588 | @end group |
| 16589 | @end smallexample |
| 16590 | |
| 16591 | @node recursive-graph-body-print |
| 16592 | @section The @code{recursive-graph-body-print} Function |
| 16593 | @findex recursive-graph-body-print |
| 16594 | |
| 16595 | The @code{graph-body-print} function may also be written recursively. |
| 16596 | The recursive solution is divided into two parts: an outside `wrapper' |
| 16597 | that uses a @code{let} expression to determine the values of several |
| 16598 | variables that need only be found once, such as the maximum height of |
| 16599 | the graph, and an inside function that is called recursively to print |
| 16600 | the graph. |
| 16601 | |
| 16602 | @need 1250 |
| 16603 | The `wrapper' is uncomplicated: |
| 16604 | |
| 16605 | @smallexample |
| 16606 | @group |
| 16607 | (defun recursive-graph-body-print (numbers-list) |
| 16608 | "Print a bar graph of the NUMBERS-LIST. |
| 16609 | The numbers-list consists of the Y-axis values." |
| 16610 | (let ((height (apply 'max numbers-list)) |
| 16611 | (symbol-width (length graph-blank)) |
| 16612 | from-position) |
| 16613 | (recursive-graph-body-print-internal |
| 16614 | numbers-list |
| 16615 | height |
| 16616 | symbol-width))) |
| 16617 | @end group |
| 16618 | @end smallexample |
| 16619 | |
| 16620 | The recursive function is a little more difficult. It has four parts: |
| 16621 | the `do-again-test', the printing code, the recursive call, and the |
| 16622 | `next-step-expression'. The `do-again-test' is a @code{when} |
| 16623 | expression that determines whether the @code{numbers-list} contains |
| 16624 | any remaining elements; if it does, the function prints one column of |
| 16625 | the graph using the printing code and calls itself again. The |
| 16626 | function calls itself again according to the value produced by the |
| 16627 | `next-step-expression' which causes the call to act on a shorter |
| 16628 | version of the @code{numbers-list}. |
| 16629 | |
| 16630 | @smallexample |
| 16631 | @group |
| 16632 | (defun recursive-graph-body-print-internal |
| 16633 | (numbers-list height symbol-width) |
| 16634 | "Print a bar graph. |
| 16635 | Used within recursive-graph-body-print function." |
| 16636 | @end group |
| 16637 | |
| 16638 | @group |
| 16639 | (when numbers-list |
| 16640 | (setq from-position (point)) |
| 16641 | (insert-rectangle |
| 16642 | (column-of-graph height (car numbers-list))) |
| 16643 | @end group |
| 16644 | @group |
| 16645 | (goto-char from-position) |
| 16646 | (forward-char symbol-width) |
| 16647 | (sit-for 0) ; @r{Draw graph column by column.} |
| 16648 | (recursive-graph-body-print-internal |
| 16649 | (cdr numbers-list) height symbol-width))) |
| 16650 | @end group |
| 16651 | @end smallexample |
| 16652 | |
| 16653 | @need 1250 |
| 16654 | After installation, this expression can be tested; here is a sample: |
| 16655 | |
| 16656 | @smallexample |
| 16657 | (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1)) |
| 16658 | @end smallexample |
| 16659 | |
| 16660 | @need 800 |
| 16661 | Here is what @code{recursive-graph-body-print} produces: |
| 16662 | |
| 16663 | @smallexample |
| 16664 | @group |
| 16665 | * |
| 16666 | ** * |
| 16667 | **** * |
| 16668 | **** *** |
| 16669 | * ********* |
| 16670 | ************ |
| 16671 | ************* |
| 16672 | @end group |
| 16673 | @end smallexample |
| 16674 | |
| 16675 | Either of these two functions, @code{graph-body-print} or |
| 16676 | @code{recursive-graph-body-print}, create the body of a graph. |
| 16677 | |
| 16678 | @node Printed Axes |
| 16679 | @section Need for Printed Axes |
| 16680 | |
| 16681 | A graph needs printed axes, so you can orient yourself. For a do-once |
| 16682 | project, it may be reasonable to draw the axes by hand using Emacs's |
| 16683 | Picture mode; but a graph drawing function may be used more than once. |
| 16684 | |
| 16685 | For this reason, I have written enhancements to the basic |
| 16686 | @code{print-graph-body} function that automatically print labels for |
| 16687 | the horizontal and vertical axes. Since the label printing functions |
| 16688 | do not contain much new material, I have placed their description in |
| 16689 | an appendix. @xref{Full Graph, , A Graph with Labeled Axes}. |
| 16690 | |
| 16691 | @node Line Graph Exercise |
| 16692 | @section Exercise |
| 16693 | |
| 16694 | Write a line graph version of the graph printing functions. |
| 16695 | |
| 16696 | @node Emacs Initialization |
| 16697 | @chapter Your @file{.emacs} File |
| 16698 | @cindex @file{.emacs} file |
| 16699 | @cindex Customizing your @file{.emacs} file |
| 16700 | @cindex Initialization file |
| 16701 | |
| 16702 | ``You don't have to like Emacs to like it''---this seemingly |
| 16703 | paradoxical statement is the secret of GNU Emacs. The plain, `out of |
| 16704 | the box' Emacs is a generic tool. Most people who use it, customize |
| 16705 | it to suit themselves. |
| 16706 | |
| 16707 | GNU Emacs is mostly written in Emacs Lisp; this means that by writing |
| 16708 | expressions in Emacs Lisp you can change or extend Emacs. |
| 16709 | |
| 16710 | @menu |
| 16711 | * Default Configuration:: |
| 16712 | * Site-wide Init:: You can write site-wide init files. |
| 16713 | * defcustom:: Emacs will write code for you. |
| 16714 | * Beginning init File:: How to write a @file{.emacs} init file. |
| 16715 | * Text and Auto-fill:: Automatically wrap lines. |
| 16716 | * Mail Aliases:: Use abbreviations for email addresses. |
| 16717 | * Indent Tabs Mode:: Don't use tabs with @TeX{} |
| 16718 | * Keybindings:: Create some personal keybindings. |
| 16719 | * Keymaps:: More about key binding. |
| 16720 | * Loading Files:: Load (i.e., evaluate) files automatically. |
| 16721 | * Autoload:: Make functions available. |
| 16722 | * Simple Extension:: Define a function; bind it to a key. |
| 16723 | * X11 Colors:: Colors in X. |
| 16724 | * Miscellaneous:: |
| 16725 | * Mode Line:: How to customize your mode line. |
| 16726 | @end menu |
| 16727 | |
| 16728 | @ifnottex |
| 16729 | @node Default Configuration |
| 16730 | @unnumberedsec Emacs's Default Configuration |
| 16731 | @end ifnottex |
| 16732 | |
| 16733 | There are those who appreciate Emacs's default configuration. After |
| 16734 | all, Emacs starts you in C mode when you edit a C file, starts you in |
| 16735 | Fortran mode when you edit a Fortran file, and starts you in |
| 16736 | Fundamental mode when you edit an unadorned file. This all makes |
| 16737 | sense, if you do not know who is going to use Emacs. Who knows what a |
| 16738 | person hopes to do with an unadorned file? Fundamental mode is the |
| 16739 | right default for such a file, just as C mode is the right default for |
| 16740 | editing C code. (Enough programming languages have syntaxes |
| 16741 | that enable them to share or nearly share features, so C mode is |
| 16742 | now provided by CC mode, the `C Collection'.) |
| 16743 | |
| 16744 | But when you do know who is going to use Emacs---you, |
| 16745 | yourself---then it makes sense to customize Emacs. |
| 16746 | |
| 16747 | For example, I seldom want Fundamental mode when I edit an |
| 16748 | otherwise undistinguished file; I want Text mode. This is why I |
| 16749 | customize Emacs: so it suits me. |
| 16750 | |
| 16751 | You can customize and extend Emacs by writing or adapting a |
| 16752 | @file{~/.emacs} file. This is your personal initialization file; its |
| 16753 | contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You |
| 16754 | may also add @file{.el} to @file{~/.emacs} and call it a |
| 16755 | @file{~/.emacs.el} file. In the past, you were forbidden to type the |
| 16756 | extra keystrokes that the name @file{~/.emacs.el} requires, but now |
| 16757 | you may. The new format is consistent with the Emacs Lisp file |
| 16758 | naming conventions; the old format saves typing.} |
| 16759 | |
| 16760 | A @file{~/.emacs} file contains Emacs Lisp code. You can write this |
| 16761 | code yourself; or you can use Emacs's @code{customize} feature to write |
| 16762 | the code for you. You can combine your own expressions and |
| 16763 | auto-written Customize expressions in your @file{.emacs} file. |
| 16764 | |
| 16765 | (I myself prefer to write my own expressions, except for those, |
| 16766 | particularly fonts, that I find easier to manipulate using the |
| 16767 | @code{customize} command. I combine the two methods.) |
| 16768 | |
| 16769 | Most of this chapter is about writing expressions yourself. It |
| 16770 | describes a simple @file{.emacs} file; for more information, see |
| 16771 | @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and |
| 16772 | @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference |
| 16773 | Manual}. |
| 16774 | |
| 16775 | @node Site-wide Init |
| 16776 | @section Site-wide Initialization Files |
| 16777 | |
| 16778 | @cindex @file{default.el} init file |
| 16779 | @cindex @file{site-init.el} init file |
| 16780 | @cindex @file{site-load.el} init file |
| 16781 | In addition to your personal initialization file, Emacs automatically |
| 16782 | loads various site-wide initialization files, if they exist. These |
| 16783 | have the same form as your @file{.emacs} file, but are loaded by |
| 16784 | everyone. |
| 16785 | |
| 16786 | Two site-wide initialization files, @file{site-load.el} and |
| 16787 | @file{site-init.el}, are loaded into Emacs and then `dumped' if a |
| 16788 | `dumped' version of Emacs is created, as is most common. (Dumped |
| 16789 | copies of Emacs load more quickly. However, once a file is loaded and |
| 16790 | dumped, a change to it does not lead to a change in Emacs unless you |
| 16791 | load it yourself or re-dump Emacs. @xref{Building Emacs, , Building |
| 16792 | Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the |
| 16793 | @file{INSTALL} file.) |
| 16794 | |
| 16795 | Three other site-wide initialization files are loaded automatically |
| 16796 | each time you start Emacs, if they exist. These are |
| 16797 | @file{site-start.el}, which is loaded @emph{before} your @file{.emacs} |
| 16798 | file, and @file{default.el}, and the terminal type file, which are both |
| 16799 | loaded @emph{after} your @file{.emacs} file. |
| 16800 | |
| 16801 | Settings and definitions in your @file{.emacs} file will overwrite |
| 16802 | conflicting settings and definitions in a @file{site-start.el} file, |
| 16803 | if it exists; but the settings and definitions in a @file{default.el} |
| 16804 | or terminal type file will overwrite those in your @file{.emacs} file. |
| 16805 | (You can prevent interference from a terminal type file by setting |
| 16806 | @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A |
| 16807 | Simple Extension}.) |
| 16808 | |
| 16809 | @c Rewritten to avoid overfull hbox. |
| 16810 | The @file{INSTALL} file that comes in the distribution contains |
| 16811 | descriptions of the @file{site-init.el} and @file{site-load.el} files. |
| 16812 | |
| 16813 | The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files |
| 16814 | control loading. These files are in the @file{lisp} directory of the |
| 16815 | Emacs distribution and are worth perusing. |
| 16816 | |
| 16817 | The @file{loaddefs.el} file contains a good many suggestions as to |
| 16818 | what to put into your own @file{.emacs} file, or into a site-wide |
| 16819 | initialization file. |
| 16820 | |
| 16821 | @node defcustom |
| 16822 | @section Specifying Variables using @code{defcustom} |
| 16823 | @findex defcustom |
| 16824 | |
| 16825 | You can specify variables using @code{defcustom} so that you and |
| 16826 | others can then use Emacs's @code{customize} feature to set their |
| 16827 | values. (You cannot use @code{customize} to write function |
| 16828 | definitions; but you can write @code{defuns} in your @file{.emacs} |
| 16829 | file. Indeed, you can write any Lisp expression in your @file{.emacs} |
| 16830 | file.) |
| 16831 | |
| 16832 | The @code{customize} feature depends on the @code{defcustom} macro. |
| 16833 | Although you can use @code{defvar} or @code{setq} for variables that |
| 16834 | users set, the @code{defcustom} macro is designed for the job. |
| 16835 | |
| 16836 | You can use your knowledge of @code{defvar} for writing the |
| 16837 | first three arguments for @code{defcustom}. The first argument to |
| 16838 | @code{defcustom} is the name of the variable. The second argument is |
| 16839 | the variable's initial value, if any; and this value is set only if |
| 16840 | the value has not already been set. The third argument is the |
| 16841 | documentation. |
| 16842 | |
| 16843 | The fourth and subsequent arguments to @code{defcustom} specify types |
| 16844 | and options; these are not featured in @code{defvar}. (These |
| 16845 | arguments are optional.) |
| 16846 | |
| 16847 | Each of these arguments consists of a keyword followed by a value. |
| 16848 | Each keyword starts with the colon character @samp{:}. |
| 16849 | |
| 16850 | @need 1250 |
| 16851 | For example, the customizable user option variable |
| 16852 | @code{text-mode-hook} looks like this: |
| 16853 | |
| 16854 | @smallexample |
| 16855 | @group |
| 16856 | (defcustom text-mode-hook nil |
| 16857 | "Normal hook run when entering Text mode and many related modes." |
| 16858 | :type 'hook |
| 16859 | :options '(turn-on-auto-fill flyspell-mode) |
| 16860 | :group 'wp) |
| 16861 | @end group |
| 16862 | @end smallexample |
| 16863 | |
| 16864 | @noindent |
| 16865 | The name of the variable is @code{text-mode-hook}; it has no default |
| 16866 | value; and its documentation string tells you what it does. |
| 16867 | |
| 16868 | The @code{:type} keyword tells Emacs the kind of data to which |
| 16869 | @code{text-mode-hook} should be set and how to display the value in a |
| 16870 | Customization buffer. |
| 16871 | |
| 16872 | The @code{:options} keyword specifies a suggested list of values for |
| 16873 | the variable. Usually, @code{:options} applies to a hook. |
| 16874 | The list is only a suggestion; it is not exclusive; a person who sets |
| 16875 | the variable may set it to other values; the list shown following the |
| 16876 | @code{:options} keyword is intended to offer convenient choices to a |
| 16877 | user. |
| 16878 | |
| 16879 | Finally, the @code{:group} keyword tells the Emacs Customization |
| 16880 | command in which group the variable is located. This tells where to |
| 16881 | find it. |
| 16882 | |
| 16883 | The @code{defcustom} macro recognizes more than a dozen keywords. |
| 16884 | For more information, see @ref{Customization, , Writing Customization |
| 16885 | Definitions, elisp, The GNU Emacs Lisp Reference Manual}. |
| 16886 | |
| 16887 | Consider @code{text-mode-hook} as an example. |
| 16888 | |
| 16889 | There are two ways to customize this variable. You can use the |
| 16890 | customization command or write the appropriate expressions yourself. |
| 16891 | |
| 16892 | @need 800 |
| 16893 | Using the customization command, you can type: |
| 16894 | |
| 16895 | @smallexample |
| 16896 | M-x customize |
| 16897 | @end smallexample |
| 16898 | |
| 16899 | @noindent |
| 16900 | and find that the group for editing files of data is called `data'. |
| 16901 | Enter that group. Text Mode Hook is the first member. You can click |
| 16902 | on its various options, such as @code{turn-on-auto-fill}, to set the |
| 16903 | values. After you click on the button to |
| 16904 | |
| 16905 | @smallexample |
| 16906 | Save for Future Sessions |
| 16907 | @end smallexample |
| 16908 | |
| 16909 | @noindent |
| 16910 | Emacs will write an expression into your @file{.emacs} file. |
| 16911 | It will look like this: |
| 16912 | |
| 16913 | @smallexample |
| 16914 | @group |
| 16915 | (custom-set-variables |
| 16916 | ;; custom-set-variables was added by Custom. |
| 16917 | ;; If you edit it by hand, you could mess it up, so be careful. |
| 16918 | ;; Your init file should contain only one such instance. |
| 16919 | ;; If there is more than one, they won't work right. |
| 16920 | '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify)))) |
| 16921 | @end group |
| 16922 | @end smallexample |
| 16923 | |
| 16924 | @noindent |
| 16925 | (The @code{text-mode-hook-identify} function tells |
| 16926 | @code{toggle-text-mode-auto-fill} which buffers are in Text mode. |
| 16927 | It comes on automatically.) |
| 16928 | |
| 16929 | The @code{custom-set-variables} function works somewhat differently |
| 16930 | than a @code{setq}. While I have never learned the differences, I |
| 16931 | modify the @code{custom-set-variables} expressions in my @file{.emacs} |
| 16932 | file by hand: I make the changes in what appears to me to be a |
| 16933 | reasonable manner and have not had any problems. Others prefer to use |
| 16934 | the Customization command and let Emacs do the work for them. |
| 16935 | |
| 16936 | Another @code{custom-set-@dots{}} function is @code{custom-set-faces}. |
| 16937 | This function sets the various font faces. Over time, I have set a |
| 16938 | considerable number of faces. Some of the time, I re-set them using |
| 16939 | @code{customize}; other times, I simply edit the |
| 16940 | @code{custom-set-faces} expression in my @file{.emacs} file itself. |
| 16941 | |
| 16942 | The second way to customize your @code{text-mode-hook} is to set it |
| 16943 | yourself in your @file{.emacs} file using code that has nothing to do |
| 16944 | with the @code{custom-set-@dots{}} functions. |
| 16945 | |
| 16946 | @need 800 |
| 16947 | When you do this, and later use @code{customize}, you will see a |
| 16948 | message that says |
| 16949 | |
| 16950 | @smallexample |
| 16951 | CHANGED outside Customize; operating on it here may be unreliable. |
| 16952 | @end smallexample |
| 16953 | |
| 16954 | @need 800 |
| 16955 | This message is only a warning. If you click on the button to |
| 16956 | |
| 16957 | @smallexample |
| 16958 | Save for Future Sessions |
| 16959 | @end smallexample |
| 16960 | |
| 16961 | @noindent |
| 16962 | Emacs will write a @code{custom-set-@dots{}} expression near the end |
| 16963 | of your @file{.emacs} file that will be evaluated after your |
| 16964 | hand-written expression. It will, therefore, overrule your |
| 16965 | hand-written expression. No harm will be done. When you do this, |
| 16966 | however, be careful to remember which expression is active; if you |
| 16967 | forget, you may confuse yourself. |
| 16968 | |
| 16969 | So long as you remember where the values are set, you will have no |
| 16970 | trouble. In any event, the values are always set in your |
| 16971 | initialization file, which is usually called @file{.emacs}. |
| 16972 | |
| 16973 | I myself use @code{customize} for hardly anything. Mostly, I write |
| 16974 | expressions myself. |
| 16975 | |
| 16976 | @findex defsubst |
| 16977 | @findex defconst |
| 16978 | Incidentally, to be more complete concerning defines: @code{defsubst} |
| 16979 | defines an inline function. The syntax is just like that of |
| 16980 | @code{defun}. @code{defconst} defines a symbol as a constant. The |
| 16981 | intent is that neither programs nor users should ever change a value |
| 16982 | set by @code{defconst}. (You can change it; the value set is a |
| 16983 | variable; but please do not.) |
| 16984 | |
| 16985 | @node Beginning init File |
| 16986 | @section Beginning a @file{.emacs} File |
| 16987 | @cindex @file{.emacs} file, beginning of |
| 16988 | |
| 16989 | When you start Emacs, it loads your @file{.emacs} file unless you tell |
| 16990 | it not to by specifying @samp{-q} on the command line. (The |
| 16991 | @code{emacs -q} command gives you a plain, out-of-the-box Emacs.) |
| 16992 | |
| 16993 | A @file{.emacs} file contains Lisp expressions. Often, these are no |
| 16994 | more than expressions to set values; sometimes they are function |
| 16995 | definitions. |
| 16996 | |
| 16997 | @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs |
| 16998 | Manual}, for a short description of initialization files. |
| 16999 | |
| 17000 | This chapter goes over some of the same ground, but is a walk among |
| 17001 | extracts from a complete, long-used @file{.emacs} file---my own. |
| 17002 | |
| 17003 | The first part of the file consists of comments: reminders to myself. |
| 17004 | By now, of course, I remember these things, but when I started, I did |
| 17005 | not. |
| 17006 | |
| 17007 | @need 1200 |
| 17008 | @smallexample |
| 17009 | @group |
| 17010 | ;;;; Bob's .emacs file |
| 17011 | ; Robert J. Chassell |
| 17012 | ; 26 September 1985 |
| 17013 | @end group |
| 17014 | @end smallexample |
| 17015 | |
| 17016 | @noindent |
| 17017 | Look at that date! I started this file a long time ago. I have been |
| 17018 | adding to it ever since. |
| 17019 | |
| 17020 | @smallexample |
| 17021 | @group |
| 17022 | ; Each section in this file is introduced by a |
| 17023 | ; line beginning with four semicolons; and each |
| 17024 | ; entry is introduced by a line beginning with |
| 17025 | ; three semicolons. |
| 17026 | @end group |
| 17027 | @end smallexample |
| 17028 | |
| 17029 | @noindent |
| 17030 | This describes the usual conventions for comments in Emacs Lisp. |
| 17031 | Everything on a line that follows a semicolon is a comment. Two, |
| 17032 | three, and four semicolons are used as subsection and section markers. |
| 17033 | (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for |
| 17034 | more about comments.) |
| 17035 | |
| 17036 | @smallexample |
| 17037 | @group |
| 17038 | ;;;; The Help Key |
| 17039 | ; Control-h is the help key; |
| 17040 | ; after typing control-h, type a letter to |
| 17041 | ; indicate the subject about which you want help. |
| 17042 | ; For an explanation of the help facility, |
| 17043 | ; type control-h two times in a row. |
| 17044 | @end group |
| 17045 | @end smallexample |
| 17046 | |
| 17047 | @noindent |
| 17048 | Just remember: type @kbd{C-h} two times for help. |
| 17049 | |
| 17050 | @smallexample |
| 17051 | @group |
| 17052 | ; To find out about any mode, type control-h m |
| 17053 | ; while in that mode. For example, to find out |
| 17054 | ; about mail mode, enter mail mode and then type |
| 17055 | ; control-h m. |
| 17056 | @end group |
| 17057 | @end smallexample |
| 17058 | |
| 17059 | @noindent |
| 17060 | `Mode help', as I call this, is very helpful. Usually, it tells you |
| 17061 | all you need to know. |
| 17062 | |
| 17063 | Of course, you don't need to include comments like these in your |
| 17064 | @file{.emacs} file. I included them in mine because I kept forgetting |
| 17065 | about Mode help or the conventions for comments---but I was able to |
| 17066 | remember to look here to remind myself. |
| 17067 | |
| 17068 | @node Text and Auto-fill |
| 17069 | @section Text and Auto Fill Mode |
| 17070 | |
| 17071 | Now we come to the part that `turns on' Text mode and |
| 17072 | Auto Fill mode. |
| 17073 | |
| 17074 | @smallexample |
| 17075 | @group |
| 17076 | ;;; Text mode and Auto Fill mode |
| 17077 | ;; The next two lines put Emacs into Text mode |
| 17078 | ;; and Auto Fill mode, and are for writers who |
| 17079 | ;; want to start writing prose rather than code. |
| 17080 | (setq-default major-mode 'text-mode) |
| 17081 | (add-hook 'text-mode-hook 'turn-on-auto-fill) |
| 17082 | @end group |
| 17083 | @end smallexample |
| 17084 | |
| 17085 | Here is the first part of this @file{.emacs} file that does something |
| 17086 | besides remind a forgetful human! |
| 17087 | |
| 17088 | The first of the two lines in parentheses tells Emacs to turn on Text |
| 17089 | mode when you find a file, @emph{unless} that file should go into some |
| 17090 | other mode, such as C mode. |
| 17091 | |
| 17092 | @cindex Per-buffer, local variables list |
| 17093 | @cindex Local variables list, per-buffer, |
| 17094 | @cindex Automatic mode selection |
| 17095 | @cindex Mode selection, automatic |
| 17096 | When Emacs reads a file, it looks at the extension to the file name, |
| 17097 | if any. (The extension is the part that comes after a @samp{.}.) If |
| 17098 | the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns |
| 17099 | on C mode. Also, Emacs looks at first nonblank line of the file; if |
| 17100 | the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs |
| 17101 | possesses a list of extensions and specifications that it uses |
| 17102 | automatically. In addition, Emacs looks near the last page for a |
| 17103 | per-buffer, ``local variables list'', if any. |
| 17104 | |
| 17105 | @ifinfo |
| 17106 | @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU |
| 17107 | Emacs Manual}. |
| 17108 | |
| 17109 | @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs |
| 17110 | Manual}. |
| 17111 | @end ifinfo |
| 17112 | @iftex |
| 17113 | See sections ``How Major Modes are Chosen'' and ``Local Variables in |
| 17114 | Files'' in @cite{The GNU Emacs Manual}. |
| 17115 | @end iftex |
| 17116 | |
| 17117 | Now, back to the @file{.emacs} file. |
| 17118 | |
| 17119 | @need 800 |
| 17120 | Here is the line again; how does it work? |
| 17121 | |
| 17122 | @cindex Text Mode turned on |
| 17123 | @smallexample |
| 17124 | (setq major-mode 'text-mode) |
| 17125 | @end smallexample |
| 17126 | |
| 17127 | @noindent |
| 17128 | This line is a short, but complete Emacs Lisp expression. |
| 17129 | |
| 17130 | We are already familiar with @code{setq}. It sets the following variable, |
| 17131 | @code{major-mode}, to the subsequent value, which is @code{text-mode}. |
| 17132 | The single quote mark before @code{text-mode} tells Emacs to deal directly |
| 17133 | with the @code{text-mode} symbol, not with whatever it might stand for. |
| 17134 | @xref{set & setq, , Setting the Value of a Variable}, |
| 17135 | for a reminder of how @code{setq} works. |
| 17136 | The main point is that there is no difference between the procedure you |
| 17137 | use to set a value in your @file{.emacs} file and the procedure you use |
| 17138 | anywhere else in Emacs. |
| 17139 | |
| 17140 | @need 800 |
| 17141 | Here is the next line: |
| 17142 | |
| 17143 | @cindex Auto Fill mode turned on |
| 17144 | @findex add-hook |
| 17145 | @smallexample |
| 17146 | (add-hook 'text-mode-hook 'turn-on-auto-fill) |
| 17147 | @end smallexample |
| 17148 | |
| 17149 | @noindent |
| 17150 | In this line, the @code{add-hook} command adds |
| 17151 | @code{turn-on-auto-fill} to the variable. |
| 17152 | |
| 17153 | @code{turn-on-auto-fill} is the name of a program, that, you guessed |
| 17154 | it!, turns on Auto Fill mode. |
| 17155 | |
| 17156 | Every time Emacs turns on Text mode, Emacs runs the commands `hooked' |
| 17157 | onto Text mode. So every time Emacs turns on Text mode, Emacs also |
| 17158 | turns on Auto Fill mode. |
| 17159 | |
| 17160 | In brief, the first line causes Emacs to enter Text mode when you edit a |
| 17161 | file, unless the file name extension, a first non-blank line, or local |
| 17162 | variables to tell Emacs otherwise. |
| 17163 | |
| 17164 | Text mode among other actions, sets the syntax table to work |
| 17165 | conveniently for writers. In Text mode, Emacs considers an apostrophe |
| 17166 | as part of a word like a letter; but Emacs does not consider a period |
| 17167 | or a space as part of a word. Thus, @kbd{M-f} moves you over |
| 17168 | @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after |
| 17169 | the @samp{t} of @samp{it's}. |
| 17170 | |
| 17171 | The second line causes Emacs to turn on Auto Fill mode when it turns |
| 17172 | on Text mode. In Auto Fill mode, Emacs automatically breaks a line |
| 17173 | that is too wide and brings the excessively wide part of the line down |
| 17174 | to the next line. Emacs breaks lines between words, not within them. |
| 17175 | |
| 17176 | When Auto Fill mode is turned off, lines continue to the right as you |
| 17177 | type them. Depending on how you set the value of |
| 17178 | @code{truncate-lines}, the words you type either disappear off the |
| 17179 | right side of the screen, or else are shown, in a rather ugly and |
| 17180 | unreadable manner, as a continuation line on the screen. |
| 17181 | |
| 17182 | @need 1250 |
| 17183 | In addition, in this part of my @file{.emacs} file, I tell the Emacs |
| 17184 | fill commands to insert two spaces after a colon: |
| 17185 | |
| 17186 | @smallexample |
| 17187 | (setq colon-double-space t) |
| 17188 | @end smallexample |
| 17189 | |
| 17190 | @node Mail Aliases |
| 17191 | @section Mail Aliases |
| 17192 | |
| 17193 | Here is a @code{setq} that `turns on' mail aliases, along with more |
| 17194 | reminders. |
| 17195 | |
| 17196 | @smallexample |
| 17197 | @group |
| 17198 | ;;; Mail mode |
| 17199 | ; To enter mail mode, type `C-x m' |
| 17200 | ; To enter RMAIL (for reading mail), |
| 17201 | ; type `M-x rmail' |
| 17202 | (setq mail-aliases t) |
| 17203 | @end group |
| 17204 | @end smallexample |
| 17205 | |
| 17206 | @cindex Mail aliases |
| 17207 | @noindent |
| 17208 | This @code{setq} command sets the value of the variable |
| 17209 | @code{mail-aliases} to @code{t}. Since @code{t} means true, the line |
| 17210 | says, in effect, ``Yes, use mail aliases.'' |
| 17211 | |
| 17212 | Mail aliases are convenient short names for long email addresses or |
| 17213 | for lists of email addresses. The file where you keep your `aliases' |
| 17214 | is @file{~/.mailrc}. You write an alias like this: |
| 17215 | |
| 17216 | @smallexample |
| 17217 | alias geo george@@foobar.wiz.edu |
| 17218 | @end smallexample |
| 17219 | |
| 17220 | @noindent |
| 17221 | When you write a message to George, address it to @samp{geo}; the |
| 17222 | mailer will automatically expand @samp{geo} to the full address. |
| 17223 | |
| 17224 | @node Indent Tabs Mode |
| 17225 | @section Indent Tabs Mode |
| 17226 | @cindex Tabs, preventing |
| 17227 | @findex indent-tabs-mode |
| 17228 | |
| 17229 | By default, Emacs inserts tabs in place of multiple spaces when it |
| 17230 | formats a region. (For example, you might indent many lines of text |
| 17231 | all at once with the @code{indent-region} command.) Tabs look fine on |
| 17232 | a terminal or with ordinary printing, but they produce badly indented |
| 17233 | output when you use @TeX{} or Texinfo since @TeX{} ignores tabs. |
| 17234 | |
| 17235 | @need 1250 |
| 17236 | The following turns off Indent Tabs mode: |
| 17237 | |
| 17238 | @smallexample |
| 17239 | @group |
| 17240 | ;;; Prevent Extraneous Tabs |
| 17241 | (setq-default indent-tabs-mode nil) |
| 17242 | @end group |
| 17243 | @end smallexample |
| 17244 | |
| 17245 | Note that this line uses @code{setq-default} rather than the |
| 17246 | @code{setq} command that we have seen before. The @code{setq-default} |
| 17247 | command sets values only in buffers that do not have their own local |
| 17248 | values for the variable. |
| 17249 | |
| 17250 | @ifinfo |
| 17251 | @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}. |
| 17252 | |
| 17253 | @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs |
| 17254 | Manual}. |
| 17255 | @end ifinfo |
| 17256 | @iftex |
| 17257 | See sections ``Tabs vs.@: Spaces'' and ``Local Variables in |
| 17258 | Files'' in @cite{The GNU Emacs Manual}. |
| 17259 | @end iftex |
| 17260 | |
| 17261 | @need 1700 |
| 17262 | @node Keybindings |
| 17263 | @section Some Keybindings |
| 17264 | |
| 17265 | Now for some personal keybindings: |
| 17266 | |
| 17267 | @smallexample |
| 17268 | @group |
| 17269 | ;;; Compare windows |
| 17270 | (global-set-key "\C-cw" 'compare-windows) |
| 17271 | @end group |
| 17272 | @end smallexample |
| 17273 | |
| 17274 | @findex compare-windows |
| 17275 | @code{compare-windows} is a nifty command that compares the text in |
| 17276 | your current window with text in the next window. It makes the |
| 17277 | comparison by starting at point in each window, moving over text in |
| 17278 | each window as far as they match. I use this command all the time. |
| 17279 | |
| 17280 | This also shows how to set a key globally, for all modes. |
| 17281 | |
| 17282 | @cindex Setting a key globally |
| 17283 | @cindex Global set key |
| 17284 | @cindex Key setting globally |
| 17285 | @findex global-set-key |
| 17286 | The command is @code{global-set-key}. It is followed by the |
| 17287 | keybinding. In a @file{.emacs} file, the keybinding is written as |
| 17288 | shown: @code{\C-c} stands for `control-c', which means `press the |
| 17289 | control key and the @key{c} key at the same time'. The @code{w} means |
| 17290 | `press the @key{w} key'. The keybinding is surrounded by double |
| 17291 | quotation marks. In documentation, you would write this as |
| 17292 | @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as |
| 17293 | @kbd{M-c}, rather than a @key{CTRL} key, you would write |
| 17294 | @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, , |
| 17295 | Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for |
| 17296 | details.) |
| 17297 | |
| 17298 | The command invoked by the keys is @code{compare-windows}. Note that |
| 17299 | @code{compare-windows} is preceded by a single quote; otherwise, Emacs |
| 17300 | would first try to evaluate the symbol to determine its value. |
| 17301 | |
| 17302 | These three things, the double quotation marks, the backslash before |
| 17303 | the @samp{C}, and the single quote mark are necessary parts of |
| 17304 | keybinding that I tend to forget. Fortunately, I have come to |
| 17305 | remember that I should look at my existing @file{.emacs} file, and |
| 17306 | adapt what is there. |
| 17307 | |
| 17308 | As for the keybinding itself: @kbd{C-c w}. This combines the prefix |
| 17309 | key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This |
| 17310 | set of keys, @kbd{C-c} followed by a single character, is strictly |
| 17311 | reserved for individuals' own use. (I call these `own' keys, since |
| 17312 | these are for my own use.) You should always be able to create such a |
| 17313 | keybinding for your own use without stomping on someone else's |
| 17314 | keybinding. If you ever write an extension to Emacs, please avoid |
| 17315 | taking any of these keys for public use. Create a key like @kbd{C-c |
| 17316 | C-w} instead. Otherwise, we will run out of `own' keys. |
| 17317 | |
| 17318 | @need 1250 |
| 17319 | Here is another keybinding, with a comment: |
| 17320 | |
| 17321 | @smallexample |
| 17322 | @group |
| 17323 | ;;; Keybinding for `occur' |
| 17324 | ; I use occur a lot, so let's bind it to a key: |
| 17325 | (global-set-key "\C-co" 'occur) |
| 17326 | @end group |
| 17327 | @end smallexample |
| 17328 | |
| 17329 | @findex occur |
| 17330 | The @code{occur} command shows all the lines in the current buffer |
| 17331 | that contain a match for a regular expression. Matching lines are |
| 17332 | shown in a buffer called @file{*Occur*}. That buffer serves as a menu |
| 17333 | to jump to occurrences. |
| 17334 | |
| 17335 | @findex global-unset-key |
| 17336 | @cindex Unbinding key |
| 17337 | @cindex Key unbinding |
| 17338 | @need 1250 |
| 17339 | Here is how to unbind a key, so it does not |
| 17340 | work: |
| 17341 | |
| 17342 | @smallexample |
| 17343 | @group |
| 17344 | ;;; Unbind `C-x f' |
| 17345 | (global-unset-key "\C-xf") |
| 17346 | @end group |
| 17347 | @end smallexample |
| 17348 | |
| 17349 | There is a reason for this unbinding: I found I inadvertently typed |
| 17350 | @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a |
| 17351 | file, as I intended, I accidentally set the width for filled text, |
| 17352 | almost always to a width I did not want. Since I hardly ever reset my |
| 17353 | default width, I simply unbound the key. |
| 17354 | |
| 17355 | @findex list-buffers, @r{rebound} |
| 17356 | @findex buffer-menu, @r{bound to key} |
| 17357 | @need 1250 |
| 17358 | The following rebinds an existing key: |
| 17359 | |
| 17360 | @smallexample |
| 17361 | @group |
| 17362 | ;;; Rebind `C-x C-b' for `buffer-menu' |
| 17363 | (global-set-key "\C-x\C-b" 'buffer-menu) |
| 17364 | @end group |
| 17365 | @end smallexample |
| 17366 | |
| 17367 | By default, @kbd{C-x C-b} runs the |
| 17368 | @code{list-buffers} command. This command lists |
| 17369 | your buffers in @emph{another} window. Since I |
| 17370 | almost always want to do something in that |
| 17371 | window, I prefer the @code{buffer-menu} |
| 17372 | command, which not only lists the buffers, |
| 17373 | but moves point into that window. |
| 17374 | |
| 17375 | @node Keymaps |
| 17376 | @section Keymaps |
| 17377 | @cindex Keymaps |
| 17378 | @cindex Rebinding keys |
| 17379 | |
| 17380 | Emacs uses @dfn{keymaps} to record which keys call which commands. |
| 17381 | When you use @code{global-set-key} to set the keybinding for a single |
| 17382 | command in all parts of Emacs, you are specifying the keybinding in |
| 17383 | @code{current-global-map}. |
| 17384 | |
| 17385 | Specific modes, such as C mode or Text mode, have their own keymaps; |
| 17386 | the mode-specific keymaps override the global map that is shared by |
| 17387 | all buffers. |
| 17388 | |
| 17389 | The @code{global-set-key} function binds, or rebinds, the global |
| 17390 | keymap. For example, the following binds the key @kbd{C-x C-b} to the |
| 17391 | function @code{buffer-menu}: |
| 17392 | |
| 17393 | @smallexample |
| 17394 | (global-set-key "\C-x\C-b" 'buffer-menu) |
| 17395 | @end smallexample |
| 17396 | |
| 17397 | Mode-specific keymaps are bound using the @code{define-key} function, |
| 17398 | which takes a specific keymap as an argument, as well as the key and |
| 17399 | the command. For example, my @file{.emacs} file contains the |
| 17400 | following expression to bind the @code{texinfo-insert-@@group} command |
| 17401 | to @kbd{C-c C-c g}: |
| 17402 | |
| 17403 | @smallexample |
| 17404 | @group |
| 17405 | (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group) |
| 17406 | @end group |
| 17407 | @end smallexample |
| 17408 | |
| 17409 | @noindent |
| 17410 | The @code{texinfo-insert-@@group} function itself is a little extension |
| 17411 | to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I |
| 17412 | use this command all the time and prefer to type the three strokes |
| 17413 | @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}. |
| 17414 | (@samp{@@group} and its matching @samp{@@end group} are commands that |
| 17415 | keep all enclosed text together on one page; many multi-line examples |
| 17416 | in this book are surrounded by @samp{@@group @dots{} @@end group}.) |
| 17417 | |
| 17418 | @need 1250 |
| 17419 | Here is the @code{texinfo-insert-@@group} function definition: |
| 17420 | |
| 17421 | @smallexample |
| 17422 | @group |
| 17423 | (defun texinfo-insert-@@group () |
| 17424 | "Insert the string @@group in a Texinfo buffer." |
| 17425 | (interactive) |
| 17426 | (beginning-of-line) |
| 17427 | (insert "@@group\n")) |
| 17428 | @end group |
| 17429 | @end smallexample |
| 17430 | |
| 17431 | (Of course, I could have used Abbrev mode to save typing, rather than |
| 17432 | write a function to insert a word; but I prefer key strokes consistent |
| 17433 | with other Texinfo mode key bindings.) |
| 17434 | |
| 17435 | You will see numerous @code{define-key} expressions in |
| 17436 | @file{loaddefs.el} as well as in the various mode libraries, such as |
| 17437 | @file{cc-mode.el} and @file{lisp-mode.el}. |
| 17438 | |
| 17439 | @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs |
| 17440 | Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp |
| 17441 | Reference Manual}, for more information about keymaps. |
| 17442 | |
| 17443 | @node Loading Files |
| 17444 | @section Loading Files |
| 17445 | @cindex Loading files |
| 17446 | @c findex load |
| 17447 | |
| 17448 | Many people in the GNU Emacs community have written extensions to |
| 17449 | Emacs. As time goes by, these extensions are often included in new |
| 17450 | releases. For example, the Calendar and Diary packages are now part |
| 17451 | of the standard GNU Emacs, as is Calc. |
| 17452 | |
| 17453 | You can use a @code{load} command to evaluate a complete file and |
| 17454 | thereby install all the functions and variables in the file into Emacs. |
| 17455 | For example: |
| 17456 | |
| 17457 | @c (auto-compression-mode t) |
| 17458 | |
| 17459 | @smallexample |
| 17460 | (load "~/emacs/slowsplit") |
| 17461 | @end smallexample |
| 17462 | |
| 17463 | This evaluates, i.e., loads, the @file{slowsplit.el} file or if it |
| 17464 | exists, the faster, byte compiled @file{slowsplit.elc} file from the |
| 17465 | @file{emacs} sub-directory of your home directory. The file contains |
| 17466 | the function @code{split-window-quietly}, which John Robinson wrote in |
| 17467 | 1989. |
| 17468 | |
| 17469 | The @code{split-window-quietly} function splits a window with the |
| 17470 | minimum of redisplay. I installed it in 1989 because it worked well |
| 17471 | with the slow 1200 baud terminals I was then using. Nowadays, I only |
| 17472 | occasionally come across such a slow connection, but I continue to use |
| 17473 | the function because I like the way it leaves the bottom half of a |
| 17474 | buffer in the lower of the new windows and the top half in the upper |
| 17475 | window. |
| 17476 | |
| 17477 | @need 1250 |
| 17478 | To replace the key binding for the default |
| 17479 | @code{split-window-vertically}, you must also unset that key and bind |
| 17480 | the keys to @code{split-window-quietly}, like this: |
| 17481 | |
| 17482 | @smallexample |
| 17483 | @group |
| 17484 | (global-unset-key "\C-x2") |
| 17485 | (global-set-key "\C-x2" 'split-window-quietly) |
| 17486 | @end group |
| 17487 | @end smallexample |
| 17488 | |
| 17489 | @vindex load-path |
| 17490 | If you load many extensions, as I do, then instead of specifying the |
| 17491 | exact location of the extension file, as shown above, you can specify |
| 17492 | that directory as part of Emacs's @code{load-path}. Then, when Emacs |
| 17493 | loads a file, it will search that directory as well as its default |
| 17494 | list of directories. (The default list is specified in @file{paths.h} |
| 17495 | when Emacs is built.) |
| 17496 | |
| 17497 | @need 1250 |
| 17498 | The following command adds your @file{~/emacs} directory to the |
| 17499 | existing load path: |
| 17500 | |
| 17501 | @smallexample |
| 17502 | @group |
| 17503 | ;;; Emacs Load Path |
| 17504 | (setq load-path (cons "~/emacs" load-path)) |
| 17505 | @end group |
| 17506 | @end smallexample |
| 17507 | |
| 17508 | Incidentally, @code{load-library} is an interactive interface to the |
| 17509 | @code{load} function. The complete function looks like this: |
| 17510 | |
| 17511 | @findex load-library |
| 17512 | @smallexample |
| 17513 | @group |
| 17514 | (defun load-library (library) |
| 17515 | "Load the library named LIBRARY. |
| 17516 | This is an interface to the function `load'." |
| 17517 | (interactive |
| 17518 | (list (completing-read "Load library: " |
| 17519 | (apply-partially 'locate-file-completion-table |
| 17520 | load-path |
| 17521 | (get-load-suffixes))))) |
| 17522 | (load library)) |
| 17523 | @end group |
| 17524 | @end smallexample |
| 17525 | |
| 17526 | The name of the function, @code{load-library}, comes from the use of |
| 17527 | `library' as a conventional synonym for `file'. The source for the |
| 17528 | @code{load-library} command is in the @file{files.el} library. |
| 17529 | |
| 17530 | Another interactive command that does a slightly different job is |
| 17531 | @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for |
| 17532 | Emacs, emacs, The GNU Emacs Manual}, for information on the |
| 17533 | distinction between @code{load-library} and this command. |
| 17534 | |
| 17535 | @node Autoload |
| 17536 | @section Autoloading |
| 17537 | @findex autoload |
| 17538 | |
| 17539 | Instead of installing a function by loading the file that contains it, |
| 17540 | or by evaluating the function definition, you can make the function |
| 17541 | available but not actually install it until it is first called. This |
| 17542 | is called @dfn{autoloading}. |
| 17543 | |
| 17544 | When you execute an autoloaded function, Emacs automatically evaluates |
| 17545 | the file that contains the definition, and then calls the function. |
| 17546 | |
| 17547 | Emacs starts quicker with autoloaded functions, since their libraries |
| 17548 | are not loaded right away; but you need to wait a moment when you |
| 17549 | first use such a function, while its containing file is evaluated. |
| 17550 | |
| 17551 | Rarely used functions are frequently autoloaded. The |
| 17552 | @file{loaddefs.el} library contains hundreds of autoloaded functions, |
| 17553 | from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may |
| 17554 | come to use a `rare' function frequently. When you do, you should |
| 17555 | load that function's file with a @code{load} expression in your |
| 17556 | @file{.emacs} file. |
| 17557 | |
| 17558 | In my @file{.emacs} file, I load 14 libraries that contain functions |
| 17559 | that would otherwise be autoloaded. (Actually, it would have been |
| 17560 | better to include these files in my `dumped' Emacs, but I forgot. |
| 17561 | @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp |
| 17562 | Reference Manual}, and the @file{INSTALL} file for more about |
| 17563 | dumping.) |
| 17564 | |
| 17565 | You may also want to include autoloaded expressions in your @file{.emacs} |
| 17566 | file. @code{autoload} is a built-in function that takes up to five |
| 17567 | arguments, the final three of which are optional. The first argument |
| 17568 | is the name of the function to be autoloaded; the second is the name |
| 17569 | of the file to be loaded. The third argument is documentation for the |
| 17570 | function, and the fourth tells whether the function can be called |
| 17571 | interactively. The fifth argument tells what type of |
| 17572 | object---@code{autoload} can handle a keymap or macro as well as a |
| 17573 | function (the default is a function). |
| 17574 | |
| 17575 | @need 800 |
| 17576 | Here is a typical example: |
| 17577 | |
| 17578 | @smallexample |
| 17579 | @group |
| 17580 | (autoload 'html-helper-mode |
| 17581 | "html-helper-mode" "Edit HTML documents" t) |
| 17582 | @end group |
| 17583 | @end smallexample |
| 17584 | |
| 17585 | @noindent |
| 17586 | (@code{html-helper-mode} is an older alternative to @code{html-mode}, |
| 17587 | which is a standard part of the distribution.) |
| 17588 | |
| 17589 | @noindent |
| 17590 | This expression autoloads the @code{html-helper-mode} function. It |
| 17591 | takes it from the @file{html-helper-mode.el} file (or from the byte |
| 17592 | compiled version @file{html-helper-mode.elc}, if that exists.) The |
| 17593 | file must be located in a directory specified by @code{load-path}. |
| 17594 | The documentation says that this is a mode to help you edit documents |
| 17595 | written in the HyperText Markup Language. You can call this mode |
| 17596 | interactively by typing @kbd{M-x html-helper-mode}. (You need to |
| 17597 | duplicate the function's regular documentation in the autoload |
| 17598 | expression because the regular function is not yet loaded, so its |
| 17599 | documentation is not available.) |
| 17600 | |
| 17601 | @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference |
| 17602 | Manual}, for more information. |
| 17603 | |
| 17604 | @node Simple Extension |
| 17605 | @section A Simple Extension: @code{line-to-top-of-window} |
| 17606 | @findex line-to-top-of-window |
| 17607 | @cindex Simple extension in @file{.emacs} file |
| 17608 | |
| 17609 | Here is a simple extension to Emacs that moves the line point is on to |
| 17610 | the top of the window. I use this all the time, to make text easier |
| 17611 | to read. |
| 17612 | |
| 17613 | You can put the following code into a separate file and then load it |
| 17614 | from your @file{.emacs} file, or you can include it within your |
| 17615 | @file{.emacs} file. |
| 17616 | |
| 17617 | @need 1250 |
| 17618 | Here is the definition: |
| 17619 | |
| 17620 | @smallexample |
| 17621 | @group |
| 17622 | ;;; Line to top of window; |
| 17623 | ;;; replace three keystroke sequence C-u 0 C-l |
| 17624 | (defun line-to-top-of-window () |
| 17625 | "Move the line point is on to top of window." |
| 17626 | (interactive) |
| 17627 | (recenter 0)) |
| 17628 | @end group |
| 17629 | @end smallexample |
| 17630 | |
| 17631 | @need 1250 |
| 17632 | Now for the keybinding. |
| 17633 | |
| 17634 | Nowadays, function keys as well as mouse button events and |
| 17635 | non-@sc{ascii} characters are written within square brackets, without |
| 17636 | quotation marks. (In Emacs version 18 and before, you had to write |
| 17637 | different function key bindings for each different make of terminal.) |
| 17638 | |
| 17639 | I bind @code{line-to-top-of-window} to my @key{F6} function key like |
| 17640 | this: |
| 17641 | |
| 17642 | @smallexample |
| 17643 | (global-set-key [f6] 'line-to-top-of-window) |
| 17644 | @end smallexample |
| 17645 | |
| 17646 | For more information, see @ref{Init Rebinding, , Rebinding Keys in |
| 17647 | Your Init File, emacs, The GNU Emacs Manual}. |
| 17648 | |
| 17649 | @cindex Conditional 'twixt two versions of Emacs |
| 17650 | @cindex Version of Emacs, choosing |
| 17651 | @cindex Emacs version, choosing |
| 17652 | If you run two versions of GNU Emacs, such as versions 22 and 23, and |
| 17653 | use one @file{.emacs} file, you can select which code to evaluate with |
| 17654 | the following conditional: |
| 17655 | |
| 17656 | @smallexample |
| 17657 | @group |
| 17658 | (cond |
| 17659 | ((= 22 emacs-major-version) |
| 17660 | ;; evaluate version 22 code |
| 17661 | ( @dots{} )) |
| 17662 | ((= 23 emacs-major-version) |
| 17663 | ;; evaluate version 23 code |
| 17664 | ( @dots{} ))) |
| 17665 | @end group |
| 17666 | @end smallexample |
| 17667 | |
| 17668 | For example, recent versions blink |
| 17669 | their cursors by default. I hate such blinking, as well as other |
| 17670 | features, so I placed the following in my @file{.emacs} |
| 17671 | file@footnote{When I start instances of Emacs that do not load my |
| 17672 | @file{.emacs} file or any site file, I also turn off blinking: |
| 17673 | |
| 17674 | @smallexample |
| 17675 | emacs -q --no-site-file -eval '(blink-cursor-mode nil)' |
| 17676 | |
| 17677 | @exdent Or nowadays, using an even more sophisticated set of options, |
| 17678 | |
| 17679 | emacs -Q -D |
| 17680 | @end smallexample |
| 17681 | }: |
| 17682 | |
| 17683 | @smallexample |
| 17684 | @group |
| 17685 | (when (>= emacs-major-version 21) |
| 17686 | (blink-cursor-mode 0) |
| 17687 | ;; Insert newline when you press `C-n' (next-line) |
| 17688 | ;; at the end of the buffer |
| 17689 | (setq next-line-add-newlines t) |
| 17690 | @end group |
| 17691 | @group |
| 17692 | ;; Turn on image viewing |
| 17693 | (auto-image-file-mode t) |
| 17694 | @end group |
| 17695 | @group |
| 17696 | ;; Turn on menu bar (this bar has text) |
| 17697 | ;; (Use numeric argument to turn on) |
| 17698 | (menu-bar-mode 1) |
| 17699 | @end group |
| 17700 | @group |
| 17701 | ;; Turn off tool bar (this bar has icons) |
| 17702 | ;; (Use numeric argument to turn on) |
| 17703 | (tool-bar-mode nil) |
| 17704 | @end group |
| 17705 | @group |
| 17706 | ;; Turn off tooltip mode for tool bar |
| 17707 | ;; (This mode causes icon explanations to pop up) |
| 17708 | ;; (Use numeric argument to turn on) |
| 17709 | (tooltip-mode nil) |
| 17710 | ;; If tooltips turned on, make tips appear promptly |
| 17711 | (setq tooltip-delay 0.1) ; default is 0.7 second |
| 17712 | ) |
| 17713 | @end group |
| 17714 | @end smallexample |
| 17715 | |
| 17716 | @node X11 Colors |
| 17717 | @section X11 Colors |
| 17718 | |
| 17719 | You can specify colors when you use Emacs with the MIT X Windowing |
| 17720 | system. |
| 17721 | |
| 17722 | I dislike the default colors and specify my own. |
| 17723 | |
| 17724 | @need 1250 |
| 17725 | Here are the expressions in my @file{.emacs} |
| 17726 | file that set values: |
| 17727 | |
| 17728 | @smallexample |
| 17729 | @group |
| 17730 | ;; Set cursor color |
| 17731 | (set-cursor-color "white") |
| 17732 | |
| 17733 | ;; Set mouse color |
| 17734 | (set-mouse-color "white") |
| 17735 | |
| 17736 | ;; Set foreground and background |
| 17737 | (set-foreground-color "white") |
| 17738 | (set-background-color "darkblue") |
| 17739 | @end group |
| 17740 | |
| 17741 | @group |
| 17742 | ;;; Set highlighting colors for isearch and drag |
| 17743 | (set-face-foreground 'highlight "white") |
| 17744 | (set-face-background 'highlight "blue") |
| 17745 | @end group |
| 17746 | |
| 17747 | @group |
| 17748 | (set-face-foreground 'region "cyan") |
| 17749 | (set-face-background 'region "blue") |
| 17750 | @end group |
| 17751 | |
| 17752 | @group |
| 17753 | (set-face-foreground 'secondary-selection "skyblue") |
| 17754 | (set-face-background 'secondary-selection "darkblue") |
| 17755 | @end group |
| 17756 | |
| 17757 | @group |
| 17758 | ;; Set calendar highlighting colors |
| 17759 | (setq calendar-load-hook |
| 17760 | (lambda () |
| 17761 | (set-face-foreground 'diary-face "skyblue") |
| 17762 | (set-face-background 'holiday-face "slate blue") |
| 17763 | (set-face-foreground 'holiday-face "white"))) |
| 17764 | @end group |
| 17765 | @end smallexample |
| 17766 | |
| 17767 | The various shades of blue soothe my eye and prevent me from seeing |
| 17768 | the screen flicker. |
| 17769 | |
| 17770 | Alternatively, I could have set my specifications in various X |
| 17771 | initialization files. For example, I could set the foreground, |
| 17772 | background, cursor, and pointer (i.e., mouse) colors in my |
| 17773 | @file{~/.Xresources} file like this: |
| 17774 | |
| 17775 | @smallexample |
| 17776 | @group |
| 17777 | Emacs*foreground: white |
| 17778 | Emacs*background: darkblue |
| 17779 | Emacs*cursorColor: white |
| 17780 | Emacs*pointerColor: white |
| 17781 | @end group |
| 17782 | @end smallexample |
| 17783 | |
| 17784 | In any event, since it is not part of Emacs, I set the root color of |
| 17785 | my X window in my @file{~/.xinitrc} file, like this@footnote{I also |
| 17786 | run more modern window managers, such as Enlightenment, Gnome, or KDE; |
| 17787 | in those cases, I often specify an image rather than a plain color.}: |
| 17788 | |
| 17789 | @smallexample |
| 17790 | xsetroot -solid Navy -fg white & |
| 17791 | @end smallexample |
| 17792 | |
| 17793 | @need 1700 |
| 17794 | @node Miscellaneous |
| 17795 | @section Miscellaneous Settings for a @file{.emacs} File |
| 17796 | |
| 17797 | @need 1250 |
| 17798 | Here are a few miscellaneous settings: |
| 17799 | @sp 1 |
| 17800 | |
| 17801 | @itemize @minus |
| 17802 | @item |
| 17803 | Set the shape and color of the mouse cursor: |
| 17804 | |
| 17805 | @smallexample |
| 17806 | @group |
| 17807 | ; Cursor shapes are defined in |
| 17808 | ; `/usr/include/X11/cursorfont.h'; |
| 17809 | ; for example, the `target' cursor is number 128; |
| 17810 | ; the `top_left_arrow' cursor is number 132. |
| 17811 | @end group |
| 17812 | |
| 17813 | @group |
| 17814 | (let ((mpointer (x-get-resource "*mpointer" |
| 17815 | "*emacs*mpointer"))) |
| 17816 | ;; If you have not set your mouse pointer |
| 17817 | ;; then set it, otherwise leave as is: |
| 17818 | (if (eq mpointer nil) |
| 17819 | (setq mpointer "132")) ; top_left_arrow |
| 17820 | @end group |
| 17821 | @group |
| 17822 | (setq x-pointer-shape (string-to-int mpointer)) |
| 17823 | (set-mouse-color "white")) |
| 17824 | @end group |
| 17825 | @end smallexample |
| 17826 | |
| 17827 | @item |
| 17828 | Or you can set the values of a variety of features in an alist, like |
| 17829 | this: |
| 17830 | |
| 17831 | @smallexample |
| 17832 | @group |
| 17833 | (setq-default |
| 17834 | default-frame-alist |
| 17835 | '((cursor-color . "white") |
| 17836 | (mouse-color . "white") |
| 17837 | (foreground-color . "white") |
| 17838 | (background-color . "DodgerBlue4") |
| 17839 | ;; (cursor-type . bar) |
| 17840 | (cursor-type . box) |
| 17841 | @end group |
| 17842 | @group |
| 17843 | (tool-bar-lines . 0) |
| 17844 | (menu-bar-lines . 1) |
| 17845 | (width . 80) |
| 17846 | (height . 58) |
| 17847 | (font . |
| 17848 | "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1") |
| 17849 | )) |
| 17850 | @end group |
| 17851 | @end smallexample |
| 17852 | |
| 17853 | @item |
| 17854 | Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL} |
| 17855 | into @kbd{@key{CTRL}-h}.@* |
| 17856 | (Some older keyboards needed this, although I have not seen the |
| 17857 | problem recently.) |
| 17858 | |
| 17859 | @smallexample |
| 17860 | @group |
| 17861 | ;; Translate `C-h' to <DEL>. |
| 17862 | ; (keyboard-translate ?\C-h ?\C-?) |
| 17863 | |
| 17864 | ;; Translate <DEL> to `C-h'. |
| 17865 | (keyboard-translate ?\C-? ?\C-h) |
| 17866 | @end group |
| 17867 | @end smallexample |
| 17868 | |
| 17869 | @item Turn off a blinking cursor! |
| 17870 | |
| 17871 | @smallexample |
| 17872 | @group |
| 17873 | (if (fboundp 'blink-cursor-mode) |
| 17874 | (blink-cursor-mode -1)) |
| 17875 | @end group |
| 17876 | @end smallexample |
| 17877 | |
| 17878 | @noindent |
| 17879 | or start GNU Emacs with the command @code{emacs -nbc}. |
| 17880 | |
| 17881 | @need 1250 |
| 17882 | @item When using `grep'@* |
| 17883 | @samp{-i}@w{ } Ignore case distinctions@* |
| 17884 | @samp{-n}@w{ } Prefix each line of output with line number@* |
| 17885 | @samp{-H}@w{ } Print the filename for each match.@* |
| 17886 | @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-} |
| 17887 | |
| 17888 | @smallexample |
| 17889 | (setq grep-command "grep -i -nH -e ") |
| 17890 | @end smallexample |
| 17891 | |
| 17892 | @ignore |
| 17893 | @c Evidently, no longer needed in GNU Emacs 22 |
| 17894 | |
| 17895 | item Automatically uncompress compressed files when visiting them |
| 17896 | |
| 17897 | smallexample |
| 17898 | (load "uncompress") |
| 17899 | end smallexample |
| 17900 | |
| 17901 | @end ignore |
| 17902 | |
| 17903 | @item Find an existing buffer, even if it has a different name@* |
| 17904 | This avoids problems with symbolic links. |
| 17905 | |
| 17906 | @smallexample |
| 17907 | (setq find-file-existing-other-name t) |
| 17908 | @end smallexample |
| 17909 | |
| 17910 | @item Set your language environment and default input method |
| 17911 | |
| 17912 | @smallexample |
| 17913 | @group |
| 17914 | (set-language-environment "latin-1") |
| 17915 | ;; Remember you can enable or disable multilingual text input |
| 17916 | ;; with the @code{toggle-input-method'} (@kbd{C-\}) command |
| 17917 | (setq default-input-method "latin-1-prefix") |
| 17918 | @end group |
| 17919 | @end smallexample |
| 17920 | |
| 17921 | If you want to write with Chinese `GB' characters, set this instead: |
| 17922 | |
| 17923 | @smallexample |
| 17924 | @group |
| 17925 | (set-language-environment "Chinese-GB") |
| 17926 | (setq default-input-method "chinese-tonepy") |
| 17927 | @end group |
| 17928 | @end smallexample |
| 17929 | @end itemize |
| 17930 | |
| 17931 | @subsubheading Fixing Unpleasant Key Bindings |
| 17932 | @cindex Key bindings, fixing |
| 17933 | @cindex Bindings, key, fixing unpleasant |
| 17934 | |
| 17935 | Some systems bind keys unpleasantly. Sometimes, for example, the |
| 17936 | @key{CTRL} key appears in an awkward spot rather than at the far left |
| 17937 | of the home row. |
| 17938 | |
| 17939 | Usually, when people fix these sorts of keybindings, they do not |
| 17940 | change their @file{~/.emacs} file. Instead, they bind the proper keys |
| 17941 | on their consoles with the @code{loadkeys} or @code{install-keymap} |
| 17942 | commands in their boot script and then include @code{xmodmap} commands |
| 17943 | in their @file{.xinitrc} or @file{.Xsession} file for X Windows. |
| 17944 | |
| 17945 | @need 1250 |
| 17946 | @noindent |
| 17947 | For a boot script: |
| 17948 | |
| 17949 | @smallexample |
| 17950 | @group |
| 17951 | loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz |
| 17952 | @exdent or |
| 17953 | install-keymap emacs2 |
| 17954 | @end group |
| 17955 | @end smallexample |
| 17956 | |
| 17957 | @need 1250 |
| 17958 | @noindent |
| 17959 | For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps |
| 17960 | Lock} key is at the far left of the home row: |
| 17961 | |
| 17962 | @smallexample |
| 17963 | @group |
| 17964 | # Bind the key labeled `Caps Lock' to `Control' |
| 17965 | # (Such a broken user interface suggests that keyboard manufacturers |
| 17966 | # think that computers are typewriters from 1885.) |
| 17967 | |
| 17968 | xmodmap -e "clear Lock" |
| 17969 | xmodmap -e "add Control = Caps_Lock" |
| 17970 | @end group |
| 17971 | @end smallexample |
| 17972 | |
| 17973 | @need 1250 |
| 17974 | @noindent |
| 17975 | In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT} |
| 17976 | key to a @key{META} key: |
| 17977 | |
| 17978 | @smallexample |
| 17979 | @group |
| 17980 | # Some ill designed keyboards have a key labeled ALT and no Meta |
| 17981 | xmodmap -e "keysym Alt_L = Meta_L Alt_L" |
| 17982 | @end group |
| 17983 | @end smallexample |
| 17984 | |
| 17985 | @need 1700 |
| 17986 | @node Mode Line |
| 17987 | @section A Modified Mode Line |
| 17988 | @vindex mode-line-format |
| 17989 | @cindex Mode line format |
| 17990 | |
| 17991 | Finally, a feature I really like: a modified mode line. |
| 17992 | |
| 17993 | When I work over a network, I forget which machine I am using. Also, |
| 17994 | I tend to I lose track of where I am, and which line point is on. |
| 17995 | |
| 17996 | So I reset my mode line to look like this: |
| 17997 | |
| 17998 | @smallexample |
| 17999 | -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top |
| 18000 | @end smallexample |
| 18001 | |
| 18002 | I am visiting a file called @file{foo.texi}, on my machine |
| 18003 | @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in |
| 18004 | Texinfo mode, and am at the top of the buffer. |
| 18005 | |
| 18006 | @need 1200 |
| 18007 | My @file{.emacs} file has a section that looks like this: |
| 18008 | |
| 18009 | @smallexample |
| 18010 | @group |
| 18011 | ;; Set a Mode Line that tells me which machine, which directory, |
| 18012 | ;; and which line I am on, plus the other customary information. |
| 18013 | (setq-default mode-line-format |
| 18014 | (quote |
| 18015 | (#("-" 0 1 |
| 18016 | (help-echo |
| 18017 | "mouse-1: select window, mouse-2: delete others ...")) |
| 18018 | mode-line-mule-info |
| 18019 | mode-line-modified |
| 18020 | mode-line-frame-identification |
| 18021 | " " |
| 18022 | @end group |
| 18023 | @group |
| 18024 | mode-line-buffer-identification |
| 18025 | " " |
| 18026 | (:eval (substring |
| 18027 | (system-name) 0 (string-match "\\..+" (system-name)))) |
| 18028 | ":" |
| 18029 | default-directory |
| 18030 | #(" " 0 1 |
| 18031 | (help-echo |
| 18032 | "mouse-1: select window, mouse-2: delete others ...")) |
| 18033 | (line-number-mode " Line %l ") |
| 18034 | global-mode-string |
| 18035 | @end group |
| 18036 | @group |
| 18037 | #(" %[(" 0 6 |
| 18038 | (help-echo |
| 18039 | "mouse-1: select window, mouse-2: delete others ...")) |
| 18040 | (:eval (mode-line-mode-name)) |
| 18041 | mode-line-process |
| 18042 | minor-mode-alist |
| 18043 | #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...))) |
| 18044 | ")%] " |
| 18045 | (-3 . "%P") |
| 18046 | ;; "-%-" |
| 18047 | ))) |
| 18048 | @end group |
| 18049 | @end smallexample |
| 18050 | |
| 18051 | @noindent |
| 18052 | Here, I redefine the default mode line. Most of the parts are from |
| 18053 | the original; but I make a few changes. I set the @emph{default} mode |
| 18054 | line format so as to permit various modes, such as Info, to override |
| 18055 | it. |
| 18056 | |
| 18057 | Many elements in the list are self-explanatory: |
| 18058 | @code{mode-line-modified} is a variable that tells whether the buffer |
| 18059 | has been modified, @code{mode-name} tells the name of the mode, and so |
| 18060 | on. However, the format looks complicated because of two features we |
| 18061 | have not discussed. |
| 18062 | |
| 18063 | @cindex Properties, in mode line example |
| 18064 | The first string in the mode line is a dash or hyphen, @samp{-}. In |
| 18065 | the old days, it would have been specified simply as @code{"-"}. But |
| 18066 | nowadays, Emacs can add properties to a string, such as highlighting |
| 18067 | or, as in this case, a help feature. If you place your mouse cursor |
| 18068 | over the hyphen, some help information appears (By default, you must |
| 18069 | wait seven-tenths of a second before the information appears. You can |
| 18070 | change that timing by changing the value of @code{tooltip-delay}.) |
| 18071 | |
| 18072 | @need 1000 |
| 18073 | The new string format has a special syntax: |
| 18074 | |
| 18075 | @smallexample |
| 18076 | #("-" 0 1 (help-echo "mouse-1: select window, ...")) |
| 18077 | @end smallexample |
| 18078 | |
| 18079 | @noindent |
| 18080 | The @code{#(} begins a list. The first element of the list is the |
| 18081 | string itself, just one @samp{-}. The second and third |
| 18082 | elements specify the range over which the fourth element applies. A |
| 18083 | range starts @emph{after} a character, so a zero means the range |
| 18084 | starts just before the first character; a 1 means that the range ends |
| 18085 | just after the first character. The third element is the property for |
| 18086 | the range. It consists of a property list, a |
| 18087 | property name, in this case, @samp{help-echo}, followed by a value, in this |
| 18088 | case, a string. The second, third, and fourth elements of this new |
| 18089 | string format can be repeated. |
| 18090 | |
| 18091 | @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp |
| 18092 | Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format, |
| 18093 | elisp, The GNU Emacs Lisp Reference Manual}, for more information. |
| 18094 | |
| 18095 | @code{mode-line-buffer-identification} |
| 18096 | displays the current buffer name. It is a list |
| 18097 | beginning @code{(#("%12b" 0 4 @dots{}}. |
| 18098 | The @code{#(} begins the list. |
| 18099 | |
| 18100 | The @samp{"%12b"} displays the current buffer name, using the |
| 18101 | @code{buffer-name} function with which we are familiar; the `12' |
| 18102 | specifies the maximum number of characters that will be displayed. |
| 18103 | When a name has fewer characters, whitespace is added to fill out to |
| 18104 | this number. (Buffer names can and often should be longer than 12 |
| 18105 | characters; this length works well in a typical 80 column wide |
| 18106 | window.) |
| 18107 | |
| 18108 | @code{:eval} says to evaluate the following form and use the result as |
| 18109 | a string to display. In this case, the expression displays the first |
| 18110 | component of the full system name. The end of the first component is |
| 18111 | a @samp{.} (`period'), so I use the @code{string-match} function to |
| 18112 | tell me the length of the first component. The substring from the |
| 18113 | zeroth character to that length is the name of the machine. |
| 18114 | |
| 18115 | @need 1250 |
| 18116 | This is the expression: |
| 18117 | |
| 18118 | @smallexample |
| 18119 | @group |
| 18120 | (:eval (substring |
| 18121 | (system-name) 0 (string-match "\\..+" (system-name)))) |
| 18122 | @end group |
| 18123 | @end smallexample |
| 18124 | |
| 18125 | @samp{%[} and @samp{%]} cause a pair of square brackets |
| 18126 | to appear for each recursive editing level. @samp{%n} says `Narrow' |
| 18127 | when narrowing is in effect. @samp{%P} tells you the percentage of |
| 18128 | the buffer that is above the bottom of the window, or `Top', `Bottom', |
| 18129 | or `All'. (A lower case @samp{p} tell you the percentage above the |
| 18130 | @emph{top} of the window.) @samp{%-} inserts enough dashes to fill |
| 18131 | out the line. |
| 18132 | |
| 18133 | Remember, ``You don't have to like Emacs to like it''---your own |
| 18134 | Emacs can have different colors, different commands, and different |
| 18135 | keys than a default Emacs. |
| 18136 | |
| 18137 | On the other hand, if you want to bring up a plain `out of the box' |
| 18138 | Emacs, with no customization, type: |
| 18139 | |
| 18140 | @smallexample |
| 18141 | emacs -q |
| 18142 | @end smallexample |
| 18143 | |
| 18144 | @noindent |
| 18145 | This will start an Emacs that does @emph{not} load your |
| 18146 | @file{~/.emacs} initialization file. A plain, default Emacs. Nothing |
| 18147 | more. |
| 18148 | |
| 18149 | @node Debugging |
| 18150 | @chapter Debugging |
| 18151 | @cindex debugging |
| 18152 | |
| 18153 | GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The |
| 18154 | first is built into the internals of Emacs and is always with you; |
| 18155 | the second requires that you instrument a function before you can use it. |
| 18156 | |
| 18157 | Both debuggers are described extensively in @ref{Debugging, , |
| 18158 | Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}. |
| 18159 | In this chapter, I will walk through a short example of each. |
| 18160 | |
| 18161 | @menu |
| 18162 | * debug:: How to use the built-in debugger. |
| 18163 | * debug-on-entry:: Start debugging when you call a function. |
| 18164 | * debug-on-quit:: Start debugging when you quit with @kbd{C-g}. |
| 18165 | * edebug:: How to use Edebug, a source level debugger. |
| 18166 | * Debugging Exercises:: |
| 18167 | @end menu |
| 18168 | |
| 18169 | @node debug |
| 18170 | @section @code{debug} |
| 18171 | @findex debug |
| 18172 | |
| 18173 | Suppose you have written a function definition that is intended to |
| 18174 | return the sum of the numbers 1 through a given number. (This is the |
| 18175 | @code{triangle} function discussed earlier. @xref{Decrementing |
| 18176 | Example, , Example with Decrementing Counter}, for a discussion.) |
| 18177 | @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.) |
| 18178 | |
| 18179 | However, your function definition has a bug. You have mistyped |
| 18180 | @samp{1=} for @samp{1-}. Here is the broken definition: |
| 18181 | |
| 18182 | @findex triangle-bugged |
| 18183 | @smallexample |
| 18184 | @group |
| 18185 | (defun triangle-bugged (number) |
| 18186 | "Return sum of numbers 1 through NUMBER inclusive." |
| 18187 | (let ((total 0)) |
| 18188 | (while (> number 0) |
| 18189 | (setq total (+ total number)) |
| 18190 | (setq number (1= number))) ; @r{Error here.} |
| 18191 | total)) |
| 18192 | @end group |
| 18193 | @end smallexample |
| 18194 | |
| 18195 | If you are reading this in Info, you can evaluate this definition in |
| 18196 | the normal fashion. You will see @code{triangle-bugged} appear in the |
| 18197 | echo area. |
| 18198 | |
| 18199 | @need 1250 |
| 18200 | Now evaluate the @code{triangle-bugged} function with an |
| 18201 | argument of 4: |
| 18202 | |
| 18203 | @smallexample |
| 18204 | (triangle-bugged 4) |
| 18205 | @end smallexample |
| 18206 | |
| 18207 | @noindent |
| 18208 | In a recent GNU Emacs, you will create and enter a @file{*Backtrace*} |
| 18209 | buffer that says: |
| 18210 | |
| 18211 | @noindent |
| 18212 | @smallexample |
| 18213 | @group |
| 18214 | ---------- Buffer: *Backtrace* ---------- |
| 18215 | Debugger entered--Lisp error: (void-function 1=) |
| 18216 | (1= number) |
| 18217 | (setq number (1= number)) |
| 18218 | (while (> number 0) (setq total (+ total number)) |
| 18219 | (setq number (1= number))) |
| 18220 | (let ((total 0)) (while (> number 0) (setq total ...) |
| 18221 | (setq number ...)) total) |
| 18222 | triangle-bugged(4) |
| 18223 | @end group |
| 18224 | @group |
| 18225 | eval((triangle-bugged 4)) |
| 18226 | eval-last-sexp-1(nil) |
| 18227 | eval-last-sexp(nil) |
| 18228 | call-interactively(eval-last-sexp) |
| 18229 | ---------- Buffer: *Backtrace* ---------- |
| 18230 | @end group |
| 18231 | @end smallexample |
| 18232 | |
| 18233 | @noindent |
| 18234 | (I have reformatted this example slightly; the debugger does not fold |
| 18235 | long lines. As usual, you can quit the debugger by typing @kbd{q} in |
| 18236 | the @file{*Backtrace*} buffer.) |
| 18237 | |
| 18238 | In practice, for a bug as simple as this, the `Lisp error' line will |
| 18239 | tell you what you need to know to correct the definition. The |
| 18240 | function @code{1=} is `void'. |
| 18241 | |
| 18242 | @ignore |
| 18243 | @need 800 |
| 18244 | In GNU Emacs 20 and before, you will see: |
| 18245 | |
| 18246 | @smallexample |
| 18247 | Symbol's function definition is void:@: 1= |
| 18248 | @end smallexample |
| 18249 | |
| 18250 | @noindent |
| 18251 | which has the same meaning as the @file{*Backtrace*} buffer line in |
| 18252 | version 21. |
| 18253 | @end ignore |
| 18254 | |
| 18255 | However, suppose you are not quite certain what is going on? |
| 18256 | You can read the complete backtrace. |
| 18257 | |
| 18258 | In this case, you need to run a recent GNU Emacs, which automatically |
| 18259 | starts the debugger that puts you in the @file{*Backtrace*} buffer; or |
| 18260 | else, you need to start the debugger manually as described below. |
| 18261 | |
| 18262 | Read the @file{*Backtrace*} buffer from the bottom up; it tells you |
| 18263 | what Emacs did that led to the error. Emacs made an interactive call |
| 18264 | to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation |
| 18265 | of the @code{triangle-bugged} expression. Each line above tells you |
| 18266 | what the Lisp interpreter evaluated next. |
| 18267 | |
| 18268 | @need 1250 |
| 18269 | The third line from the top of the buffer is |
| 18270 | |
| 18271 | @smallexample |
| 18272 | (setq number (1= number)) |
| 18273 | @end smallexample |
| 18274 | |
| 18275 | @noindent |
| 18276 | Emacs tried to evaluate this expression; in order to do so, it tried |
| 18277 | to evaluate the inner expression shown on the second line from the |
| 18278 | top: |
| 18279 | |
| 18280 | @smallexample |
| 18281 | (1= number) |
| 18282 | @end smallexample |
| 18283 | |
| 18284 | @need 1250 |
| 18285 | @noindent |
| 18286 | This is where the error occurred; as the top line says: |
| 18287 | |
| 18288 | @smallexample |
| 18289 | Debugger entered--Lisp error: (void-function 1=) |
| 18290 | @end smallexample |
| 18291 | |
| 18292 | @noindent |
| 18293 | You can correct the mistake, re-evaluate the function definition, and |
| 18294 | then run your test again. |
| 18295 | |
| 18296 | @node debug-on-entry |
| 18297 | @section @code{debug-on-entry} |
| 18298 | @findex debug-on-entry |
| 18299 | |
| 18300 | A recent GNU Emacs starts the debugger automatically when your |
| 18301 | function has an error. |
| 18302 | |
| 18303 | @ignore |
| 18304 | GNU Emacs version 20 and before did not; it simply |
| 18305 | presented you with an error message. You had to start the debugger |
| 18306 | manually. |
| 18307 | @end ignore |
| 18308 | |
| 18309 | Incidentally, you can start the debugger manually for all versions of |
| 18310 | Emacs; the advantage is that the debugger runs even if you do not have |
| 18311 | a bug in your code. Sometimes your code will be free of bugs! |
| 18312 | |
| 18313 | You can enter the debugger when you call the function by calling |
| 18314 | @code{debug-on-entry}. |
| 18315 | |
| 18316 | @need 1250 |
| 18317 | @noindent |
| 18318 | Type: |
| 18319 | |
| 18320 | @smallexample |
| 18321 | M-x debug-on-entry RET triangle-bugged RET |
| 18322 | @end smallexample |
| 18323 | |
| 18324 | @need 1250 |
| 18325 | @noindent |
| 18326 | Now, evaluate the following: |
| 18327 | |
| 18328 | @smallexample |
| 18329 | (triangle-bugged 5) |
| 18330 | @end smallexample |
| 18331 | |
| 18332 | @noindent |
| 18333 | All versions of Emacs will create a @file{*Backtrace*} buffer and tell |
| 18334 | you that it is beginning to evaluate the @code{triangle-bugged} |
| 18335 | function: |
| 18336 | |
| 18337 | @smallexample |
| 18338 | @group |
| 18339 | ---------- Buffer: *Backtrace* ---------- |
| 18340 | Debugger entered--entering a function: |
| 18341 | * triangle-bugged(5) |
| 18342 | eval((triangle-bugged 5)) |
| 18343 | @end group |
| 18344 | @group |
| 18345 | eval-last-sexp-1(nil) |
| 18346 | eval-last-sexp(nil) |
| 18347 | call-interactively(eval-last-sexp) |
| 18348 | ---------- Buffer: *Backtrace* ---------- |
| 18349 | @end group |
| 18350 | @end smallexample |
| 18351 | |
| 18352 | In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate |
| 18353 | the first expression in @code{triangle-bugged}; the buffer will look |
| 18354 | like this: |
| 18355 | |
| 18356 | @smallexample |
| 18357 | @group |
| 18358 | ---------- Buffer: *Backtrace* ---------- |
| 18359 | Debugger entered--beginning evaluation of function call form: |
| 18360 | * (let ((total 0)) (while (> number 0) (setq total ...) |
| 18361 | (setq number ...)) total) |
| 18362 | * triangle-bugged(5) |
| 18363 | eval((triangle-bugged 5)) |
| 18364 | @end group |
| 18365 | @group |
| 18366 | eval-last-sexp-1(nil) |
| 18367 | eval-last-sexp(nil) |
| 18368 | call-interactively(eval-last-sexp) |
| 18369 | ---------- Buffer: *Backtrace* ---------- |
| 18370 | @end group |
| 18371 | @end smallexample |
| 18372 | |
| 18373 | @noindent |
| 18374 | Now, type @kbd{d} again, eight times, slowly. Each time you type |
| 18375 | @kbd{d}, Emacs will evaluate another expression in the function |
| 18376 | definition. |
| 18377 | |
| 18378 | @need 1750 |
| 18379 | Eventually, the buffer will look like this: |
| 18380 | |
| 18381 | @smallexample |
| 18382 | @group |
| 18383 | ---------- Buffer: *Backtrace* ---------- |
| 18384 | Debugger entered--beginning evaluation of function call form: |
| 18385 | * (setq number (1= number)) |
| 18386 | * (while (> number 0) (setq total (+ total number)) |
| 18387 | (setq number (1= number))) |
| 18388 | @group |
| 18389 | @end group |
| 18390 | * (let ((total 0)) (while (> number 0) (setq total ...) |
| 18391 | (setq number ...)) total) |
| 18392 | * triangle-bugged(5) |
| 18393 | eval((triangle-bugged 5)) |
| 18394 | @group |
| 18395 | @end group |
| 18396 | eval-last-sexp-1(nil) |
| 18397 | eval-last-sexp(nil) |
| 18398 | call-interactively(eval-last-sexp) |
| 18399 | ---------- Buffer: *Backtrace* ---------- |
| 18400 | @end group |
| 18401 | @end smallexample |
| 18402 | |
| 18403 | @need 1500 |
| 18404 | @noindent |
| 18405 | Finally, after you type @kbd{d} two more times, Emacs will reach the |
| 18406 | error, and the top two lines of the @file{*Backtrace*} buffer will look |
| 18407 | like this: |
| 18408 | |
| 18409 | @smallexample |
| 18410 | @group |
| 18411 | ---------- Buffer: *Backtrace* ---------- |
| 18412 | Debugger entered--Lisp error: (void-function 1=) |
| 18413 | * (1= number) |
| 18414 | @dots{} |
| 18415 | ---------- Buffer: *Backtrace* ---------- |
| 18416 | @end group |
| 18417 | @end smallexample |
| 18418 | |
| 18419 | By typing @kbd{d}, you were able to step through the function. |
| 18420 | |
| 18421 | You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this |
| 18422 | quits the trace, but does not cancel @code{debug-on-entry}. |
| 18423 | |
| 18424 | @findex cancel-debug-on-entry |
| 18425 | To cancel the effect of @code{debug-on-entry}, call |
| 18426 | @code{cancel-debug-on-entry} and the name of the function, like this: |
| 18427 | |
| 18428 | @smallexample |
| 18429 | M-x cancel-debug-on-entry RET triangle-bugged RET |
| 18430 | @end smallexample |
| 18431 | |
| 18432 | @noindent |
| 18433 | (If you are reading this in Info, cancel @code{debug-on-entry} now.) |
| 18434 | |
| 18435 | @node debug-on-quit |
| 18436 | @section @code{debug-on-quit} and @code{(debug)} |
| 18437 | |
| 18438 | In addition to setting @code{debug-on-error} or calling @code{debug-on-entry}, |
| 18439 | there are two other ways to start @code{debug}. |
| 18440 | |
| 18441 | @findex debug-on-quit |
| 18442 | You can start @code{debug} whenever you type @kbd{C-g} |
| 18443 | (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to |
| 18444 | @code{t}. This is useful for debugging infinite loops. |
| 18445 | |
| 18446 | @need 1500 |
| 18447 | @cindex @code{(debug)} in code |
| 18448 | Or, you can insert a line that says @code{(debug)} into your code |
| 18449 | where you want the debugger to start, like this: |
| 18450 | |
| 18451 | @smallexample |
| 18452 | @group |
| 18453 | (defun triangle-bugged (number) |
| 18454 | "Return sum of numbers 1 through NUMBER inclusive." |
| 18455 | (let ((total 0)) |
| 18456 | (while (> number 0) |
| 18457 | (setq total (+ total number)) |
| 18458 | (debug) ; @r{Start debugger.} |
| 18459 | (setq number (1= number))) ; @r{Error here.} |
| 18460 | total)) |
| 18461 | @end group |
| 18462 | @end smallexample |
| 18463 | |
| 18464 | The @code{debug} function is described in detail in @ref{Debugger, , |
| 18465 | The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}. |
| 18466 | |
| 18467 | @node edebug |
| 18468 | @section The @code{edebug} Source Level Debugger |
| 18469 | @cindex Source level debugger |
| 18470 | @findex edebug |
| 18471 | |
| 18472 | Edebug is a source level debugger. Edebug normally displays the |
| 18473 | source of the code you are debugging, with an arrow at the left that |
| 18474 | shows which line you are currently executing. |
| 18475 | |
| 18476 | You can walk through the execution of a function, line by line, or run |
| 18477 | quickly until reaching a @dfn{breakpoint} where execution stops. |
| 18478 | |
| 18479 | Edebug is described in @ref{Edebug, , , elisp, The GNU Emacs |
| 18480 | Lisp Reference Manual}. |
| 18481 | |
| 18482 | @need 1250 |
| 18483 | Here is a bugged function definition for @code{triangle-recursively}. |
| 18484 | @xref{Recursive triangle function, , Recursion in place of a counter}, |
| 18485 | for a review of it. |
| 18486 | |
| 18487 | @smallexample |
| 18488 | @group |
| 18489 | (defun triangle-recursively-bugged (number) |
| 18490 | "Return sum of numbers 1 through NUMBER inclusive. |
| 18491 | Uses recursion." |
| 18492 | (if (= number 1) |
| 18493 | 1 |
| 18494 | (+ number |
| 18495 | (triangle-recursively-bugged |
| 18496 | (1= number))))) ; @r{Error here.} |
| 18497 | @end group |
| 18498 | @end smallexample |
| 18499 | |
| 18500 | @noindent |
| 18501 | Normally, you would install this definition by positioning your cursor |
| 18502 | after the function's closing parenthesis and typing @kbd{C-x C-e} |
| 18503 | (@code{eval-last-sexp}) or else by positioning your cursor within the |
| 18504 | definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default, |
| 18505 | the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp |
| 18506 | Interaction mode.) |
| 18507 | |
| 18508 | @need 1500 |
| 18509 | However, to prepare this function definition for Edebug, you must |
| 18510 | first @dfn{instrument} the code using a different command. You can do |
| 18511 | this by positioning your cursor within or just after the definition |
| 18512 | and typing |
| 18513 | |
| 18514 | @smallexample |
| 18515 | M-x edebug-defun RET |
| 18516 | @end smallexample |
| 18517 | |
| 18518 | @noindent |
| 18519 | This will cause Emacs to load Edebug automatically if it is not |
| 18520 | already loaded, and properly instrument the function. |
| 18521 | |
| 18522 | After instrumenting the function, place your cursor after the |
| 18523 | following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}): |
| 18524 | |
| 18525 | @smallexample |
| 18526 | (triangle-recursively-bugged 3) |
| 18527 | @end smallexample |
| 18528 | |
| 18529 | @noindent |
| 18530 | You will be jumped back to the source for |
| 18531 | @code{triangle-recursively-bugged} and the cursor positioned at the |
| 18532 | beginning of the @code{if} line of the function. Also, you will see |
| 18533 | an arrowhead at the left hand side of that line. The arrowhead marks |
| 18534 | the line where the function is executing. (In the following examples, |
| 18535 | we show the arrowhead with @samp{=>}; in a windowing system, you may |
| 18536 | see the arrowhead as a solid triangle in the window `fringe'.) |
| 18537 | |
| 18538 | @smallexample |
| 18539 | =>@point{}(if (= number 1) |
| 18540 | @end smallexample |
| 18541 | |
| 18542 | @noindent |
| 18543 | @iftex |
| 18544 | In the example, the location of point is displayed with a star, |
| 18545 | @samp{@point{}} (in Info, it is displayed as @samp{-!-}). |
| 18546 | @end iftex |
| 18547 | @ifnottex |
| 18548 | In the example, the location of point is displayed as @samp{@point{}} |
| 18549 | (in a printed book, it is displayed with a five pointed star). |
| 18550 | @end ifnottex |
| 18551 | |
| 18552 | If you now press @key{SPC}, point will move to the next expression to |
| 18553 | be executed; the line will look like this: |
| 18554 | |
| 18555 | @smallexample |
| 18556 | =>(if @point{}(= number 1) |
| 18557 | @end smallexample |
| 18558 | |
| 18559 | @noindent |
| 18560 | As you continue to press @key{SPC}, point will move from expression to |
| 18561 | expression. At the same time, whenever an expression returns a value, |
| 18562 | that value will be displayed in the echo area. For example, after you |
| 18563 | move point past @code{number}, you will see the following: |
| 18564 | |
| 18565 | @smallexample |
| 18566 | Result: 3 (#o3, #x3, ?\C-c) |
| 18567 | @end smallexample |
| 18568 | |
| 18569 | @noindent |
| 18570 | This means the value of @code{number} is 3, which is octal three, |
| 18571 | hexadecimal three, and @sc{ascii} `control-c' (the third letter of the |
| 18572 | alphabet, in case you need to know this information). |
| 18573 | |
| 18574 | You can continue moving through the code until you reach the line with |
| 18575 | the error. Before evaluation, that line looks like this: |
| 18576 | |
| 18577 | @smallexample |
| 18578 | => @point{}(1= number))))) ; @r{Error here.} |
| 18579 | @end smallexample |
| 18580 | |
| 18581 | @need 1250 |
| 18582 | @noindent |
| 18583 | When you press @key{SPC} once again, you will produce an error message |
| 18584 | that says: |
| 18585 | |
| 18586 | @smallexample |
| 18587 | Symbol's function definition is void:@: 1= |
| 18588 | @end smallexample |
| 18589 | |
| 18590 | @noindent |
| 18591 | This is the bug. |
| 18592 | |
| 18593 | Press @kbd{q} to quit Edebug. |
| 18594 | |
| 18595 | To remove instrumentation from a function definition, simply |
| 18596 | re-evaluate it with a command that does not instrument it. |
| 18597 | For example, you could place your cursor after the definition's |
| 18598 | closing parenthesis and type @kbd{C-x C-e}. |
| 18599 | |
| 18600 | Edebug does a great deal more than walk with you through a function. |
| 18601 | You can set it so it races through on its own, stopping only at an |
| 18602 | error or at specified stopping points; you can cause it to display the |
| 18603 | changing values of various expressions; you can find out how many |
| 18604 | times a function is called, and more. |
| 18605 | |
| 18606 | Edebug is described in @ref{Edebug, , , elisp, The GNU Emacs |
| 18607 | Lisp Reference Manual}. |
| 18608 | |
| 18609 | @need 1500 |
| 18610 | @node Debugging Exercises |
| 18611 | @section Debugging Exercises |
| 18612 | |
| 18613 | @itemize @bullet |
| 18614 | @item |
| 18615 | Install the @code{@value{COUNT-WORDS}} function and then cause it to |
| 18616 | enter the built-in debugger when you call it. Run the command on a |
| 18617 | region containing two words. You will need to press @kbd{d} a |
| 18618 | remarkable number of times. On your system, is a `hook' called after |
| 18619 | the command finishes? (For information on hooks, see @ref{Command |
| 18620 | Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference |
| 18621 | Manual}.) |
| 18622 | |
| 18623 | @item |
| 18624 | Copy @code{@value{COUNT-WORDS}} into the @file{*scratch*} buffer, |
| 18625 | instrument the function for Edebug, and walk through its execution. |
| 18626 | The function does not need to have a bug, although you can introduce |
| 18627 | one if you wish. If the function lacks a bug, the walk-through |
| 18628 | completes without problems. |
| 18629 | |
| 18630 | @item |
| 18631 | While running Edebug, type @kbd{?} to see a list of all the Edebug commands. |
| 18632 | (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e., |
| 18633 | @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix |
| 18634 | for commands made outside of the Edebug debugging buffer.) |
| 18635 | |
| 18636 | @item |
| 18637 | In the Edebug debugging buffer, use the @kbd{p} |
| 18638 | (@code{edebug-bounce-point}) command to see where in the region the |
| 18639 | @code{@value{COUNT-WORDS}} is working. |
| 18640 | |
| 18641 | @item |
| 18642 | Move point to some spot further down the function and then type the |
| 18643 | @kbd{h} (@code{edebug-goto-here}) command to jump to that location. |
| 18644 | |
| 18645 | @item |
| 18646 | Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to |
| 18647 | walk through the function on its own; use an upper case @kbd{T} for |
| 18648 | @code{edebug-Trace-fast-mode}. |
| 18649 | |
| 18650 | @item |
| 18651 | Set a breakpoint, then run Edebug in Trace mode until it reaches the |
| 18652 | stopping point. |
| 18653 | @end itemize |
| 18654 | |
| 18655 | @node Conclusion |
| 18656 | @chapter Conclusion |
| 18657 | |
| 18658 | We have now reached the end of this Introduction. You have now |
| 18659 | learned enough about programming in Emacs Lisp to set values, to write |
| 18660 | simple @file{.emacs} files for yourself and your friends, and write |
| 18661 | simple customizations and extensions to Emacs. |
| 18662 | |
| 18663 | This is a place to stop. Or, if you wish, you can now go onward, and |
| 18664 | teach yourself. |
| 18665 | |
| 18666 | You have learned some of the basic nuts and bolts of programming. But |
| 18667 | only some. There are a great many more brackets and hinges that are |
| 18668 | easy to use that we have not touched. |
| 18669 | |
| 18670 | A path you can follow right now lies among the sources to GNU Emacs |
| 18671 | and in |
| 18672 | @ifnotinfo |
| 18673 | @cite{The GNU Emacs Lisp Reference Manual}. |
| 18674 | @end ifnotinfo |
| 18675 | @ifinfo |
| 18676 | @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU |
| 18677 | Emacs Lisp Reference Manual}. |
| 18678 | @end ifinfo |
| 18679 | |
| 18680 | The Emacs Lisp sources are an adventure. When you read the sources and |
| 18681 | come across a function or expression that is unfamiliar, you need to |
| 18682 | figure out or find out what it does. |
| 18683 | |
| 18684 | Go to the Reference Manual. It is a thorough, complete, and fairly |
| 18685 | easy-to-read description of Emacs Lisp. It is written not only for |
| 18686 | experts, but for people who know what you know. (The @cite{Reference |
| 18687 | Manual} comes with the standard GNU Emacs distribution. Like this |
| 18688 | introduction, it comes as a Texinfo source file, so you can read it |
| 18689 | on-line and as a typeset, printed book.) |
| 18690 | |
| 18691 | Go to the other on-line help that is part of GNU Emacs: the on-line |
| 18692 | documentation for all functions and variables, and @code{find-tag}, |
| 18693 | the program that takes you to sources. |
| 18694 | |
| 18695 | Here is an example of how I explore the sources. Because of its name, |
| 18696 | @file{simple.el} is the file I looked at first, a long time ago. As |
| 18697 | it happens some of the functions in @file{simple.el} are complicated, |
| 18698 | or at least look complicated at first sight. The @code{open-line} |
| 18699 | function, for example, looks complicated. |
| 18700 | |
| 18701 | You may want to walk through this function slowly, as we did with the |
| 18702 | @code{forward-sentence} function. (@xref{forward-sentence, The |
| 18703 | @code{forward-sentence} function}.) Or you may want to skip that |
| 18704 | function and look at another, such as @code{split-line}. You don't |
| 18705 | need to read all the functions. According to |
| 18706 | @code{count-words-in-defun}, the @code{split-line} function contains |
| 18707 | 102 words and symbols. |
| 18708 | |
| 18709 | Even though it is short, @code{split-line} contains expressions |
| 18710 | we have not studied: @code{skip-chars-forward}, @code{indent-to}, |
| 18711 | @code{current-column} and @code{insert-and-inherit}. |
| 18712 | |
| 18713 | Consider the @code{skip-chars-forward} function. (It is part of the |
| 18714 | function definition for @code{back-to-indentation}, which is shown in |
| 18715 | @ref{Review, , Review}.) |
| 18716 | |
| 18717 | In GNU Emacs, you can find out more about @code{skip-chars-forward} by |
| 18718 | typing @kbd{C-h f} (@code{describe-function}) and the name of the |
| 18719 | function. This gives you the function documentation. |
| 18720 | |
| 18721 | You may be able to guess what is done by a well named function such as |
| 18722 | @code{indent-to}; or you can look it up, too. Incidentally, the |
| 18723 | @code{describe-function} function itself is in @file{help.el}; it is |
| 18724 | one of those long, but decipherable functions. You can look up |
| 18725 | @code{describe-function} using the @kbd{C-h f} command! |
| 18726 | |
| 18727 | In this instance, since the code is Lisp, the @file{*Help*} buffer |
| 18728 | contains the name of the library containing the function's source. |
| 18729 | You can put point over the name of the library and press the RET key, |
| 18730 | which in this situation is bound to @code{help-follow}, and be taken |
| 18731 | directly to the source, in the same way as @kbd{M-.} |
| 18732 | (@code{find-tag}). |
| 18733 | |
| 18734 | The definition for @code{describe-function} illustrates how to |
| 18735 | customize the @code{interactive} expression without using the standard |
| 18736 | character codes; and it shows how to create a temporary buffer. |
| 18737 | |
| 18738 | (The @code{indent-to} function is written in C rather than Emacs Lisp; |
| 18739 | it is a `built-in' function. @code{help-follow} takes you to its |
| 18740 | source as does @code{find-tag}, when properly set up.) |
| 18741 | |
| 18742 | You can look at a function's source using @code{find-tag}, which is |
| 18743 | bound to @kbd{M-.} Finally, you can find out what the Reference |
| 18744 | Manual has to say by visiting the manual in Info, and typing @kbd{i} |
| 18745 | (@code{Info-index}) and the name of the function, or by looking up the |
| 18746 | function in the index to a printed copy of the manual. |
| 18747 | |
| 18748 | Similarly, you can find out what is meant by |
| 18749 | @code{insert-and-inherit}. |
| 18750 | |
| 18751 | Other interesting source files include @file{paragraphs.el}, |
| 18752 | @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el} |
| 18753 | file includes short, easily understood functions as well as longer |
| 18754 | ones. The @file{loaddefs.el} file contains the many standard |
| 18755 | autoloads and many keymaps. I have never looked at it all; only at |
| 18756 | parts. @file{loadup.el} is the file that loads the standard parts of |
| 18757 | Emacs; it tells you a great deal about how Emacs is built. |
| 18758 | (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp |
| 18759 | Reference Manual}, for more about building.) |
| 18760 | |
| 18761 | As I said, you have learned some nuts and bolts; however, and very |
| 18762 | importantly, we have hardly touched major aspects of programming; I |
| 18763 | have said nothing about how to sort information, except to use the |
| 18764 | predefined @code{sort} function; I have said nothing about how to store |
| 18765 | information, except to use variables and lists; I have said nothing |
| 18766 | about how to write programs that write programs. These are topics for |
| 18767 | another, and different kind of book, a different kind of learning. |
| 18768 | |
| 18769 | What you have done is learn enough for much practical work with GNU |
| 18770 | Emacs. What you have done is get started. This is the end of a |
| 18771 | beginning. |
| 18772 | |
| 18773 | @c ================ Appendix ================ |
| 18774 | |
| 18775 | @node the-the |
| 18776 | @appendix The @code{the-the} Function |
| 18777 | @findex the-the |
| 18778 | @cindex Duplicated words function |
| 18779 | @cindex Words, duplicated |
| 18780 | |
| 18781 | Sometimes when you you write text, you duplicate words---as with ``you |
| 18782 | you'' near the beginning of this sentence. I find that most |
| 18783 | frequently, I duplicate ``the''; hence, I call the function for |
| 18784 | detecting duplicated words, @code{the-the}. |
| 18785 | |
| 18786 | @need 1250 |
| 18787 | As a first step, you could use the following regular expression to |
| 18788 | search for duplicates: |
| 18789 | |
| 18790 | @smallexample |
| 18791 | \\(\\w+[ \t\n]+\\)\\1 |
| 18792 | @end smallexample |
| 18793 | |
| 18794 | @noindent |
| 18795 | This regexp matches one or more word-constituent characters followed |
| 18796 | by one or more spaces, tabs, or newlines. However, it does not detect |
| 18797 | duplicated words on different lines, since the ending of the first |
| 18798 | word, the end of the line, is different from the ending of the second |
| 18799 | word, a space. (For more information about regular expressions, see |
| 18800 | @ref{Regexp Search, , Regular Expression Searches}, as well as |
| 18801 | @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs |
| 18802 | Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp, |
| 18803 | The GNU Emacs Lisp Reference Manual}.) |
| 18804 | |
| 18805 | You might try searching just for duplicated word-constituent |
| 18806 | characters but that does not work since the pattern detects doubles |
| 18807 | such as the two occurrences of `th' in `with the'. |
| 18808 | |
| 18809 | Another possible regexp searches for word-constituent characters |
| 18810 | followed by non-word-constituent characters, reduplicated. Here, |
| 18811 | @w{@samp{\\w+}} matches one or more word-constituent characters and |
| 18812 | @w{@samp{\\W*}} matches zero or more non-word-constituent characters. |
| 18813 | |
| 18814 | @smallexample |
| 18815 | \\(\\(\\w+\\)\\W*\\)\\1 |
| 18816 | @end smallexample |
| 18817 | |
| 18818 | @noindent |
| 18819 | Again, not useful. |
| 18820 | |
| 18821 | Here is the pattern that I use. It is not perfect, but good enough. |
| 18822 | @w{@samp{\\b}} matches the empty string, provided it is at the beginning |
| 18823 | or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of |
| 18824 | any characters that are @emph{not} an @@-sign, space, newline, or tab. |
| 18825 | |
| 18826 | @smallexample |
| 18827 | \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b |
| 18828 | @end smallexample |
| 18829 | |
| 18830 | One can write more complicated expressions, but I found that this |
| 18831 | expression is good enough, so I use it. |
| 18832 | |
| 18833 | Here is the @code{the-the} function, as I include it in my |
| 18834 | @file{.emacs} file, along with a handy global key binding: |
| 18835 | |
| 18836 | @smallexample |
| 18837 | @group |
| 18838 | (defun the-the () |
| 18839 | "Search forward for for a duplicated word." |
| 18840 | (interactive) |
| 18841 | (message "Searching for for duplicated words ...") |
| 18842 | (push-mark) |
| 18843 | @end group |
| 18844 | @group |
| 18845 | ;; This regexp is not perfect |
| 18846 | ;; but is fairly good over all: |
| 18847 | (if (re-search-forward |
| 18848 | "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move) |
| 18849 | (message "Found duplicated word.") |
| 18850 | (message "End of buffer"))) |
| 18851 | @end group |
| 18852 | |
| 18853 | @group |
| 18854 | ;; Bind `the-the' to C-c \ |
| 18855 | (global-set-key "\C-c\\" 'the-the) |
| 18856 | @end group |
| 18857 | @end smallexample |
| 18858 | |
| 18859 | @sp 1 |
| 18860 | Here is test text: |
| 18861 | |
| 18862 | @smallexample |
| 18863 | @group |
| 18864 | one two two three four five |
| 18865 | five six seven |
| 18866 | @end group |
| 18867 | @end smallexample |
| 18868 | |
| 18869 | You can substitute the other regular expressions shown above in the |
| 18870 | function definition and try each of them on this list. |
| 18871 | |
| 18872 | @node Kill Ring |
| 18873 | @appendix Handling the Kill Ring |
| 18874 | @cindex Kill ring handling |
| 18875 | @cindex Handling the kill ring |
| 18876 | @cindex Ring, making a list like a |
| 18877 | |
| 18878 | The kill ring is a list that is transformed into a ring by the |
| 18879 | workings of the @code{current-kill} function. The @code{yank} and |
| 18880 | @code{yank-pop} commands use the @code{current-kill} function. |
| 18881 | |
| 18882 | This appendix describes the @code{current-kill} function as well as |
| 18883 | both the @code{yank} and the @code{yank-pop} commands, but first, |
| 18884 | consider the workings of the kill ring. |
| 18885 | |
| 18886 | @menu |
| 18887 | * What the Kill Ring Does:: |
| 18888 | * current-kill:: |
| 18889 | * yank:: Paste a copy of a clipped element. |
| 18890 | * yank-pop:: Insert element pointed to. |
| 18891 | * ring file:: |
| 18892 | @end menu |
| 18893 | |
| 18894 | @ifnottex |
| 18895 | @node What the Kill Ring Does |
| 18896 | @unnumberedsec What the Kill Ring Does |
| 18897 | @end ifnottex |
| 18898 | |
| 18899 | @need 1250 |
| 18900 | The kill ring has a default maximum length of sixty items; this number |
| 18901 | is too large for an explanation. Instead, set it to four. Please |
| 18902 | evaluate the following: |
| 18903 | |
| 18904 | @smallexample |
| 18905 | @group |
| 18906 | (setq old-kill-ring-max kill-ring-max) |
| 18907 | (setq kill-ring-max 4) |
| 18908 | @end group |
| 18909 | @end smallexample |
| 18910 | |
| 18911 | @noindent |
| 18912 | Then, please copy each line of the following indented example into the |
| 18913 | kill ring. You may kill each line with @kbd{C-k} or mark it and copy |
| 18914 | it with @kbd{M-w}. |
| 18915 | |
| 18916 | @noindent |
| 18917 | (In a read-only buffer, such as the @file{*info*} buffer, the kill |
| 18918 | command, @kbd{C-k} (@code{kill-line}), will not remove the text, |
| 18919 | merely copy it to the kill ring. However, your machine may beep at |
| 18920 | you. Alternatively, for silence, you may copy the region of each line |
| 18921 | with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark |
| 18922 | each line for this command to succeed, but it does not matter at which |
| 18923 | end you put point or mark.) |
| 18924 | |
| 18925 | @need 1250 |
| 18926 | @noindent |
| 18927 | Please invoke the calls in order, so that five elements attempt to |
| 18928 | fill the kill ring: |
| 18929 | |
| 18930 | @smallexample |
| 18931 | @group |
| 18932 | first some text |
| 18933 | second piece of text |
| 18934 | third line |
| 18935 | fourth line of text |
| 18936 | fifth bit of text |
| 18937 | @end group |
| 18938 | @end smallexample |
| 18939 | |
| 18940 | @need 1250 |
| 18941 | @noindent |
| 18942 | Then find the value of @code{kill-ring} by evaluating |
| 18943 | |
| 18944 | @smallexample |
| 18945 | kill-ring |
| 18946 | @end smallexample |
| 18947 | |
| 18948 | @need 800 |
| 18949 | @noindent |
| 18950 | It is: |
| 18951 | |
| 18952 | @smallexample |
| 18953 | @group |
| 18954 | ("fifth bit of text" "fourth line of text" |
| 18955 | "third line" "second piece of text") |
| 18956 | @end group |
| 18957 | @end smallexample |
| 18958 | |
| 18959 | @noindent |
| 18960 | The first element, @samp{first some text}, was dropped. |
| 18961 | |
| 18962 | @need 1250 |
| 18963 | To return to the old value for the length of the kill ring, evaluate: |
| 18964 | |
| 18965 | @smallexample |
| 18966 | (setq kill-ring-max old-kill-ring-max) |
| 18967 | @end smallexample |
| 18968 | |
| 18969 | @node current-kill |
| 18970 | @appendixsec The @code{current-kill} Function |
| 18971 | @findex current-kill |
| 18972 | |
| 18973 | The @code{current-kill} function changes the element in the kill ring |
| 18974 | to which @code{kill-ring-yank-pointer} points. (Also, the |
| 18975 | @code{kill-new} function sets @code{kill-ring-yank-pointer} to point |
| 18976 | to the latest element of the kill ring. The @code{kill-new} |
| 18977 | function is used directly or indirectly by @code{kill-append}, |
| 18978 | @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line}, |
| 18979 | and @code{kill-region}.) |
| 18980 | |
| 18981 | @menu |
| 18982 | * Code for current-kill:: |
| 18983 | * Understanding current-kill:: |
| 18984 | @end menu |
| 18985 | |
| 18986 | @ifnottex |
| 18987 | @node Code for current-kill |
| 18988 | @unnumberedsubsec The code for @code{current-kill} |
| 18989 | @end ifnottex |
| 18990 | |
| 18991 | |
| 18992 | @need 1500 |
| 18993 | The @code{current-kill} function is used by @code{yank} and by |
| 18994 | @code{yank-pop}. Here is the code for @code{current-kill}: |
| 18995 | |
| 18996 | @smallexample |
| 18997 | @group |
| 18998 | (defun current-kill (n &optional do-not-move) |
| 18999 | "Rotate the yanking point by N places, and then return that kill. |
| 19000 | If N is zero, `interprogram-paste-function' is set, and calling it |
| 19001 | returns a string, then that string is added to the front of the |
| 19002 | kill ring and returned as the latest kill. |
| 19003 | @end group |
| 19004 | @group |
| 19005 | If optional arg DO-NOT-MOVE is non-nil, then don't actually move the |
| 19006 | yanking point; just return the Nth kill forward." |
| 19007 | (let ((interprogram-paste (and (= n 0) |
| 19008 | interprogram-paste-function |
| 19009 | (funcall interprogram-paste-function)))) |
| 19010 | @end group |
| 19011 | @group |
| 19012 | (if interprogram-paste |
| 19013 | (progn |
| 19014 | ;; Disable the interprogram cut function when we add the new |
| 19015 | ;; text to the kill ring, so Emacs doesn't try to own the |
| 19016 | ;; selection, with identical text. |
| 19017 | (let ((interprogram-cut-function nil)) |
| 19018 | (kill-new interprogram-paste)) |
| 19019 | interprogram-paste) |
| 19020 | @end group |
| 19021 | @group |
| 19022 | (or kill-ring (error "Kill ring is empty")) |
| 19023 | (let ((ARGth-kill-element |
| 19024 | (nthcdr (mod (- n (length kill-ring-yank-pointer)) |
| 19025 | (length kill-ring)) |
| 19026 | kill-ring))) |
| 19027 | (or do-not-move |
| 19028 | (setq kill-ring-yank-pointer ARGth-kill-element)) |
| 19029 | (car ARGth-kill-element))))) |
| 19030 | @end group |
| 19031 | @end smallexample |
| 19032 | |
| 19033 | Remember also that the @code{kill-new} function sets |
| 19034 | @code{kill-ring-yank-pointer} to the latest element of the kill |
| 19035 | ring, which means that all the functions that call it set the value |
| 19036 | indirectly: @code{kill-append}, @code{copy-region-as-kill}, |
| 19037 | @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}. |
| 19038 | |
| 19039 | @need 1500 |
| 19040 | Here is the line in @code{kill-new}, which is explained in |
| 19041 | @ref{kill-new function, , The @code{kill-new} function}. |
| 19042 | |
| 19043 | @smallexample |
| 19044 | (setq kill-ring-yank-pointer kill-ring) |
| 19045 | @end smallexample |
| 19046 | |
| 19047 | @ifnottex |
| 19048 | @node Understanding current-kill |
| 19049 | @unnumberedsubsec @code{current-kill} in Outline |
| 19050 | @end ifnottex |
| 19051 | |
| 19052 | The @code{current-kill} function looks complex, but as usual, it can |
| 19053 | be understood by taking it apart piece by piece. First look at it in |
| 19054 | skeletal form: |
| 19055 | |
| 19056 | @smallexample |
| 19057 | @group |
| 19058 | (defun current-kill (n &optional do-not-move) |
| 19059 | "Rotate the yanking point by N places, and then return that kill." |
| 19060 | (let @var{varlist} |
| 19061 | @var{body}@dots{}) |
| 19062 | @end group |
| 19063 | @end smallexample |
| 19064 | |
| 19065 | This function takes two arguments, one of which is optional. It has a |
| 19066 | documentation string. It is @emph{not} interactive. |
| 19067 | |
| 19068 | @menu |
| 19069 | * Body of current-kill:: |
| 19070 | * Digression concerning error:: How to mislead humans, but not computers. |
| 19071 | * Determining the Element:: |
| 19072 | @end menu |
| 19073 | |
| 19074 | @ifnottex |
| 19075 | @node Body of current-kill |
| 19076 | @unnumberedsubsubsec The Body of @code{current-kill} |
| 19077 | @end ifnottex |
| 19078 | |
| 19079 | The body of the function definition is a @code{let} expression, which |
| 19080 | itself has a body as well as a @var{varlist}. |
| 19081 | |
| 19082 | The @code{let} expression declares a variable that will be only usable |
| 19083 | within the bounds of this function. This variable is called |
| 19084 | @code{interprogram-paste} and is for copying to another program. It |
| 19085 | is not for copying within this instance of GNU Emacs. Most window |
| 19086 | systems provide a facility for interprogram pasting. Sadly, that |
| 19087 | facility usually provides only for the last element. Most windowing |
| 19088 | systems have not adopted a ring of many possibilities, even though |
| 19089 | Emacs has provided it for decades. |
| 19090 | |
| 19091 | The @code{if} expression has two parts, one if there exists |
| 19092 | @code{interprogram-paste} and one if not. |
| 19093 | |
| 19094 | @need 2000 |
| 19095 | Let us consider the `if not' or else-part of the @code{current-kill} |
| 19096 | function. (The then-part uses the @code{kill-new} function, which |
| 19097 | we have already described. @xref{kill-new function, , The |
| 19098 | @code{kill-new} function}.) |
| 19099 | |
| 19100 | @smallexample |
| 19101 | @group |
| 19102 | (or kill-ring (error "Kill ring is empty")) |
| 19103 | (let ((ARGth-kill-element |
| 19104 | (nthcdr (mod (- n (length kill-ring-yank-pointer)) |
| 19105 | (length kill-ring)) |
| 19106 | kill-ring))) |
| 19107 | (or do-not-move |
| 19108 | (setq kill-ring-yank-pointer ARGth-kill-element)) |
| 19109 | (car ARGth-kill-element)) |
| 19110 | @end group |
| 19111 | @end smallexample |
| 19112 | |
| 19113 | @noindent |
| 19114 | The code first checks whether the kill ring has content; otherwise it |
| 19115 | signals an error. |
| 19116 | |
| 19117 | @need 1000 |
| 19118 | Note that the @code{or} expression is very similar to testing length |
| 19119 | with an @code{if}: |
| 19120 | |
| 19121 | @findex zerop |
| 19122 | @findex error |
| 19123 | @smallexample |
| 19124 | @group |
| 19125 | (if (zerop (length kill-ring)) ; @r{if-part} |
| 19126 | (error "Kill ring is empty")) ; @r{then-part} |
| 19127 | ;; No else-part |
| 19128 | @end group |
| 19129 | @end smallexample |
| 19130 | |
| 19131 | @noindent |
| 19132 | If there is not anything in the kill ring, its length must be zero and |
| 19133 | an error message sent to the user: @samp{Kill ring is empty}. The |
| 19134 | @code{current-kill} function uses an @code{or} expression which is |
| 19135 | simpler. But an @code{if} expression reminds us what goes on. |
| 19136 | |
| 19137 | This @code{if} expression uses the function @code{zerop} which returns |
| 19138 | true if the value it is testing is zero. When @code{zerop} tests |
| 19139 | true, the then-part of the @code{if} is evaluated. The then-part is a |
| 19140 | list starting with the function @code{error}, which is a function that |
| 19141 | is similar to the @code{message} function |
| 19142 | (@pxref{message, , The @code{message} Function}) in that |
| 19143 | it prints a one-line message in the echo area. However, in addition |
| 19144 | to printing a message, @code{error} also stops evaluation of the |
| 19145 | function within which it is embedded. This means that the rest of the |
| 19146 | function will not be evaluated if the length of the kill ring is zero. |
| 19147 | |
| 19148 | Then the @code{current-kill} function selects the element to return. |
| 19149 | The selection depends on the number of places that @code{current-kill} |
| 19150 | rotates and on where @code{kill-ring-yank-pointer} points. |
| 19151 | |
| 19152 | Next, either the optional @code{do-not-move} argument is true or the |
| 19153 | current value of @code{kill-ring-yank-pointer} is set to point to the |
| 19154 | list. Finally, another expression returns the first element of the |
| 19155 | list even if the @code{do-not-move} argument is true. |
| 19156 | |
| 19157 | @ifnottex |
| 19158 | @node Digression concerning error |
| 19159 | @unnumberedsubsubsec Digression about the word `error' |
| 19160 | @end ifnottex |
| 19161 | |
| 19162 | In my opinion, it is slightly misleading, at least to humans, to use |
| 19163 | the term `error' as the name of the @code{error} function. A better |
| 19164 | term would be `cancel'. Strictly speaking, of course, you cannot |
| 19165 | point to, much less rotate a pointer to a list that has no length, so |
| 19166 | from the point of view of the computer, the word `error' is correct. |
| 19167 | But a human expects to attempt this sort of thing, if only to find out |
| 19168 | whether the kill ring is full or empty. This is an act of |
| 19169 | exploration. |
| 19170 | |
| 19171 | From the human point of view, the act of exploration and discovery is |
| 19172 | not necessarily an error, and therefore should not be labeled as one, |
| 19173 | even in the bowels of a computer. As it is, the code in Emacs implies |
| 19174 | that a human who is acting virtuously, by exploring his or her |
| 19175 | environment, is making an error. This is bad. Even though the computer |
| 19176 | takes the same steps as it does when there is an `error', a term such as |
| 19177 | `cancel' would have a clearer connotation. |
| 19178 | |
| 19179 | @ifnottex |
| 19180 | @node Determining the Element |
| 19181 | @unnumberedsubsubsec Determining the Element |
| 19182 | @end ifnottex |
| 19183 | |
| 19184 | Among other actions, the else-part of the @code{if} expression sets |
| 19185 | the value of @code{kill-ring-yank-pointer} to |
| 19186 | @code{ARGth-kill-element} when the kill ring has something in it and |
| 19187 | the value of @code{do-not-move} is @code{nil}. |
| 19188 | |
| 19189 | @need 800 |
| 19190 | The code looks like this: |
| 19191 | |
| 19192 | @smallexample |
| 19193 | @group |
| 19194 | (nthcdr (mod (- n (length kill-ring-yank-pointer)) |
| 19195 | (length kill-ring)) |
| 19196 | kill-ring))) |
| 19197 | @end group |
| 19198 | @end smallexample |
| 19199 | |
| 19200 | This needs some examination. Unless it is not supposed to move the |
| 19201 | pointer, the @code{current-kill} function changes where |
| 19202 | @code{kill-ring-yank-pointer} points. |
| 19203 | That is what the |
| 19204 | @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}} |
| 19205 | expression does. Also, clearly, @code{ARGth-kill-element} is being |
| 19206 | set to be equal to some @sc{cdr} of the kill ring, using the |
| 19207 | @code{nthcdr} function that is described in an earlier section. |
| 19208 | (@xref{copy-region-as-kill}.) How does it do this? |
| 19209 | |
| 19210 | As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function |
| 19211 | works by repeatedly taking the @sc{cdr} of a list---it takes the |
| 19212 | @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{} |
| 19213 | |
| 19214 | @need 800 |
| 19215 | The two following expressions produce the same result: |
| 19216 | |
| 19217 | @smallexample |
| 19218 | @group |
| 19219 | (setq kill-ring-yank-pointer (cdr kill-ring)) |
| 19220 | |
| 19221 | (setq kill-ring-yank-pointer (nthcdr 1 kill-ring)) |
| 19222 | @end group |
| 19223 | @end smallexample |
| 19224 | |
| 19225 | However, the @code{nthcdr} expression is more complicated. It uses |
| 19226 | the @code{mod} function to determine which @sc{cdr} to select. |
| 19227 | |
| 19228 | (You will remember to look at inner functions first; indeed, we will |
| 19229 | have to go inside the @code{mod}.) |
| 19230 | |
| 19231 | The @code{mod} function returns the value of its first argument modulo |
| 19232 | the second; that is to say, it returns the remainder after dividing |
| 19233 | the first argument by the second. The value returned has the same |
| 19234 | sign as the second argument. |
| 19235 | |
| 19236 | @need 800 |
| 19237 | Thus, |
| 19238 | |
| 19239 | @smallexample |
| 19240 | @group |
| 19241 | (mod 12 4) |
| 19242 | @result{} 0 ;; @r{because there is no remainder} |
| 19243 | (mod 13 4) |
| 19244 | @result{} 1 |
| 19245 | @end group |
| 19246 | @end smallexample |
| 19247 | |
| 19248 | @need 1250 |
| 19249 | In this case, the first argument is often smaller than the second. |
| 19250 | That is fine. |
| 19251 | |
| 19252 | @smallexample |
| 19253 | @group |
| 19254 | (mod 0 4) |
| 19255 | @result{} 0 |
| 19256 | (mod 1 4) |
| 19257 | @result{} 1 |
| 19258 | @end group |
| 19259 | @end smallexample |
| 19260 | |
| 19261 | We can guess what the @code{-} function does. It is like @code{+} but |
| 19262 | subtracts instead of adds; the @code{-} function subtracts its second |
| 19263 | argument from its first. Also, we already know what the @code{length} |
| 19264 | function does (@pxref{length}). It returns the length of a list. |
| 19265 | |
| 19266 | And @code{n} is the name of the required argument to the |
| 19267 | @code{current-kill} function. |
| 19268 | |
| 19269 | @need 1250 |
| 19270 | So when the first argument to @code{nthcdr} is zero, the @code{nthcdr} |
| 19271 | expression returns the whole list, as you can see by evaluating the |
| 19272 | following: |
| 19273 | |
| 19274 | @smallexample |
| 19275 | @group |
| 19276 | ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four} |
| 19277 | ;; @r{and} (mod (- 0 4) 4) @result{} 0 |
| 19278 | (nthcdr (mod (- 0 4) 4) |
| 19279 | '("fourth line of text" |
| 19280 | "third line" |
| 19281 | "second piece of text" |
| 19282 | "first some text")) |
| 19283 | @end group |
| 19284 | @end smallexample |
| 19285 | |
| 19286 | @need 1250 |
| 19287 | When the first argument to the @code{current-kill} function is one, |
| 19288 | the @code{nthcdr} expression returns the list without its first |
| 19289 | element. |
| 19290 | |
| 19291 | @smallexample |
| 19292 | @group |
| 19293 | (nthcdr (mod (- 1 4) 4) |
| 19294 | '("fourth line of text" |
| 19295 | "third line" |
| 19296 | "second piece of text" |
| 19297 | "first some text")) |
| 19298 | @end group |
| 19299 | @end smallexample |
| 19300 | |
| 19301 | @cindex @samp{global variable} defined |
| 19302 | @cindex @samp{variable, global}, defined |
| 19303 | Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer} |
| 19304 | are @dfn{global variables}. That means that any expression in Emacs |
| 19305 | Lisp can access them. They are not like the local variables set by |
| 19306 | @code{let} or like the symbols in an argument list. |
| 19307 | Local variables can only be accessed |
| 19308 | within the @code{let} that defines them or the function that specifies |
| 19309 | them in an argument list (and within expressions called by them). |
| 19310 | |
| 19311 | @ignore |
| 19312 | @c texi2dvi fails when the name of the section is within ifnottex ... |
| 19313 | (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and |
| 19314 | @ref{defun, , The @code{defun} Macro}.) |
| 19315 | @end ignore |
| 19316 | |
| 19317 | @node yank |
| 19318 | @appendixsec @code{yank} |
| 19319 | @findex yank |
| 19320 | |
| 19321 | After learning about @code{current-kill}, the code for the |
| 19322 | @code{yank} function is almost easy. |
| 19323 | |
| 19324 | The @code{yank} function does not use the |
| 19325 | @code{kill-ring-yank-pointer} variable directly. It calls |
| 19326 | @code{insert-for-yank} which calls @code{current-kill} which sets the |
| 19327 | @code{kill-ring-yank-pointer} variable. |
| 19328 | |
| 19329 | @need 1250 |
| 19330 | The code looks like this: |
| 19331 | |
| 19332 | @c in GNU Emacs 22 |
| 19333 | @smallexample |
| 19334 | @group |
| 19335 | (defun yank (&optional arg) |
| 19336 | "Reinsert (\"paste\") the last stretch of killed text. |
| 19337 | More precisely, reinsert the stretch of killed text most recently |
| 19338 | killed OR yanked. Put point at end, and set mark at beginning. |
| 19339 | With just \\[universal-argument] as argument, same but put point at |
| 19340 | beginning (and mark at end). With argument N, reinsert the Nth most |
| 19341 | recently killed stretch of killed text. |
| 19342 | |
| 19343 | When this command inserts killed text into the buffer, it honors |
| 19344 | `yank-excluded-properties' and `yank-handler' as described in the |
| 19345 | doc string for `insert-for-yank-1', which see. |
| 19346 | |
| 19347 | See also the command \\[yank-pop]." |
| 19348 | @end group |
| 19349 | @group |
| 19350 | (interactive "*P") |
| 19351 | (setq yank-window-start (window-start)) |
| 19352 | ;; If we don't get all the way thru, make last-command indicate that |
| 19353 | ;; for the following command. |
| 19354 | (setq this-command t) |
| 19355 | (push-mark (point)) |
| 19356 | @end group |
| 19357 | @group |
| 19358 | (insert-for-yank (current-kill (cond |
| 19359 | ((listp arg) 0) |
| 19360 | ((eq arg '-) -2) |
| 19361 | (t (1- arg))))) |
| 19362 | (if (consp arg) |
| 19363 | ;; This is like exchange-point-and-mark, |
| 19364 | ;; but doesn't activate the mark. |
| 19365 | ;; It is cleaner to avoid activation, even though the command |
| 19366 | ;; loop would deactivate the mark because we inserted text. |
| 19367 | (goto-char (prog1 (mark t) |
| 19368 | (set-marker (mark-marker) (point) (current-buffer))))) |
| 19369 | @end group |
| 19370 | @group |
| 19371 | ;; If we do get all the way thru, make this-command indicate that. |
| 19372 | (if (eq this-command t) |
| 19373 | (setq this-command 'yank)) |
| 19374 | nil) |
| 19375 | @end group |
| 19376 | @end smallexample |
| 19377 | |
| 19378 | The key expression is @code{insert-for-yank}, which inserts the string |
| 19379 | returned by @code{current-kill}, but removes some text properties from |
| 19380 | it. |
| 19381 | |
| 19382 | However, before getting to that expression, the function sets the value |
| 19383 | of @code{yank-window-start} to the position returned by the |
| 19384 | @code{(window-start)} expression, the position at which the display |
| 19385 | currently starts. The @code{yank} function also sets |
| 19386 | @code{this-command} and pushes the mark. |
| 19387 | |
| 19388 | After it yanks the appropriate element, if the optional argument is a |
| 19389 | @sc{cons} rather than a number or nothing, it puts point at beginning |
| 19390 | of the yanked text and mark at its end. |
| 19391 | |
| 19392 | (The @code{prog1} function is like @code{progn} but returns the value |
| 19393 | of its first argument rather than the value of its last argument. Its |
| 19394 | first argument is forced to return the buffer's mark as an integer. |
| 19395 | You can see the documentation for these functions by placing point |
| 19396 | over them in this buffer and then typing @kbd{C-h f} |
| 19397 | (@code{describe-function}) followed by a @kbd{RET}; the default is the |
| 19398 | function.) |
| 19399 | |
| 19400 | The last part of the function tells what to do when it succeeds. |
| 19401 | |
| 19402 | @node yank-pop |
| 19403 | @appendixsec @code{yank-pop} |
| 19404 | @findex yank-pop |
| 19405 | |
| 19406 | After understanding @code{yank} and @code{current-kill}, you know how |
| 19407 | to approach the @code{yank-pop} function. Leaving out the |
| 19408 | documentation to save space, it looks like this: |
| 19409 | |
| 19410 | @c GNU Emacs 22 |
| 19411 | @smallexample |
| 19412 | @group |
| 19413 | (defun yank-pop (&optional arg) |
| 19414 | "@dots{}" |
| 19415 | (interactive "*p") |
| 19416 | (if (not (eq last-command 'yank)) |
| 19417 | (error "Previous command was not a yank")) |
| 19418 | @end group |
| 19419 | @group |
| 19420 | (setq this-command 'yank) |
| 19421 | (unless arg (setq arg 1)) |
| 19422 | (let ((inhibit-read-only t) |
| 19423 | (before (< (point) (mark t)))) |
| 19424 | @end group |
| 19425 | @group |
| 19426 | (if before |
| 19427 | (funcall (or yank-undo-function 'delete-region) (point) (mark t)) |
| 19428 | (funcall (or yank-undo-function 'delete-region) (mark t) (point))) |
| 19429 | (setq yank-undo-function nil) |
| 19430 | @end group |
| 19431 | @group |
| 19432 | (set-marker (mark-marker) (point) (current-buffer)) |
| 19433 | (insert-for-yank (current-kill arg)) |
| 19434 | ;; Set the window start back where it was in the yank command, |
| 19435 | ;; if possible. |
| 19436 | (set-window-start (selected-window) yank-window-start t) |
| 19437 | @end group |
| 19438 | @group |
| 19439 | (if before |
| 19440 | ;; This is like exchange-point-and-mark, |
| 19441 | ;; but doesn't activate the mark. |
| 19442 | ;; It is cleaner to avoid activation, even though the command |
| 19443 | ;; loop would deactivate the mark because we inserted text. |
| 19444 | (goto-char (prog1 (mark t) |
| 19445 | (set-marker (mark-marker) |
| 19446 | (point) |
| 19447 | (current-buffer)))))) |
| 19448 | nil) |
| 19449 | @end group |
| 19450 | @end smallexample |
| 19451 | |
| 19452 | The function is interactive with a small @samp{p} so the prefix |
| 19453 | argument is processed and passed to the function. The command can |
| 19454 | only be used after a previous yank; otherwise an error message is |
| 19455 | sent. This check uses the variable @code{last-command} which is set |
| 19456 | by @code{yank} and is discussed elsewhere. |
| 19457 | (@xref{copy-region-as-kill}.) |
| 19458 | |
| 19459 | The @code{let} clause sets the variable @code{before} to true or false |
| 19460 | depending whether point is before or after mark and then the region |
| 19461 | between point and mark is deleted. This is the region that was just |
| 19462 | inserted by the previous yank and it is this text that will be |
| 19463 | replaced. |
| 19464 | |
| 19465 | @code{funcall} calls its first argument as a function, passing |
| 19466 | remaining arguments to it. The first argument is whatever the |
| 19467 | @code{or} expression returns. The two remaining arguments are the |
| 19468 | positions of point and mark set by the preceding @code{yank} command. |
| 19469 | |
| 19470 | There is more, but that is the hardest part. |
| 19471 | |
| 19472 | @node ring file |
| 19473 | @appendixsec The @file{ring.el} File |
| 19474 | @cindex @file{ring.el} file |
| 19475 | |
| 19476 | Interestingly, GNU Emacs posses a file called @file{ring.el} that |
| 19477 | provides many of the features we just discussed. But functions such |
| 19478 | as @code{kill-ring-yank-pointer} do not use this library, possibly |
| 19479 | because they were written earlier. |
| 19480 | |
| 19481 | @node Full Graph |
| 19482 | @appendix A Graph with Labeled Axes |
| 19483 | |
| 19484 | Printed axes help you understand a graph. They convey scale. In an |
| 19485 | earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we |
| 19486 | wrote the code to print the body of a graph. Here we write the code |
| 19487 | for printing and labeling vertical and horizontal axes, along with the |
| 19488 | body itself. |
| 19489 | |
| 19490 | @menu |
| 19491 | * Labeled Example:: |
| 19492 | * print-graph Varlist:: @code{let} expression in @code{print-graph}. |
| 19493 | * print-Y-axis:: Print a label for the vertical axis. |
| 19494 | * print-X-axis:: Print a horizontal label. |
| 19495 | * Print Whole Graph:: The function to print a complete graph. |
| 19496 | @end menu |
| 19497 | |
| 19498 | @ifnottex |
| 19499 | @node Labeled Example |
| 19500 | @unnumberedsec Labeled Example Graph |
| 19501 | @end ifnottex |
| 19502 | |
| 19503 | Since insertions fill a buffer to the right and below point, the new |
| 19504 | graph printing function should first print the Y or vertical axis, |
| 19505 | then the body of the graph, and finally the X or horizontal axis. |
| 19506 | This sequence lays out for us the contents of the function: |
| 19507 | |
| 19508 | @enumerate |
| 19509 | @item |
| 19510 | Set up code. |
| 19511 | |
| 19512 | @item |
| 19513 | Print Y axis. |
| 19514 | |
| 19515 | @item |
| 19516 | Print body of graph. |
| 19517 | |
| 19518 | @item |
| 19519 | Print X axis. |
| 19520 | @end enumerate |
| 19521 | |
| 19522 | @need 800 |
| 19523 | Here is an example of how a finished graph should look: |
| 19524 | |
| 19525 | @smallexample |
| 19526 | @group |
| 19527 | 10 - |
| 19528 | * |
| 19529 | * * |
| 19530 | * ** |
| 19531 | * *** |
| 19532 | 5 - * ******* |
| 19533 | * *** ******* |
| 19534 | ************* |
| 19535 | *************** |
| 19536 | 1 - **************** |
| 19537 | | | | | |
| 19538 | 1 5 10 15 |
| 19539 | @end group |
| 19540 | @end smallexample |
| 19541 | |
| 19542 | @noindent |
| 19543 | In this graph, both the vertical and the horizontal axes are labeled |
| 19544 | with numbers. However, in some graphs, the horizontal axis is time |
| 19545 | and would be better labeled with months, like this: |
| 19546 | |
| 19547 | @smallexample |
| 19548 | @group |
| 19549 | 5 - * |
| 19550 | * ** * |
| 19551 | ******* |
| 19552 | ********** ** |
| 19553 | 1 - ************** |
| 19554 | | ^ | |
| 19555 | Jan June Jan |
| 19556 | @end group |
| 19557 | @end smallexample |
| 19558 | |
| 19559 | Indeed, with a little thought, we can easily come up with a variety of |
| 19560 | vertical and horizontal labeling schemes. Our task could become |
| 19561 | complicated. But complications breed confusion. Rather than permit |
| 19562 | this, it is better choose a simple labeling scheme for our first |
| 19563 | effort, and to modify or replace it later. |
| 19564 | |
| 19565 | @need 1200 |
| 19566 | These considerations suggest the following outline for the |
| 19567 | @code{print-graph} function: |
| 19568 | |
| 19569 | @smallexample |
| 19570 | @group |
| 19571 | (defun print-graph (numbers-list) |
| 19572 | "@var{documentation}@dots{}" |
| 19573 | (let ((height @dots{} |
| 19574 | @dots{})) |
| 19575 | @end group |
| 19576 | @group |
| 19577 | (print-Y-axis height @dots{} ) |
| 19578 | (graph-body-print numbers-list) |
| 19579 | (print-X-axis @dots{} ))) |
| 19580 | @end group |
| 19581 | @end smallexample |
| 19582 | |
| 19583 | We can work on each part of the @code{print-graph} function definition |
| 19584 | in turn. |
| 19585 | |
| 19586 | @node print-graph Varlist |
| 19587 | @appendixsec The @code{print-graph} Varlist |
| 19588 | @cindex @code{print-graph} varlist |
| 19589 | |
| 19590 | In writing the @code{print-graph} function, the first task is to write |
| 19591 | the varlist in the @code{let} expression. (We will leave aside for the |
| 19592 | moment any thoughts about making the function interactive or about the |
| 19593 | contents of its documentation string.) |
| 19594 | |
| 19595 | The varlist should set several values. Clearly, the top of the label |
| 19596 | for the vertical axis must be at least the height of the graph, which |
| 19597 | means that we must obtain this information here. Note that the |
| 19598 | @code{print-graph-body} function also requires this information. There |
| 19599 | is no reason to calculate the height of the graph in two different |
| 19600 | places, so we should change @code{print-graph-body} from the way we |
| 19601 | defined it earlier to take advantage of the calculation. |
| 19602 | |
| 19603 | Similarly, both the function for printing the X axis labels and the |
| 19604 | @code{print-graph-body} function need to learn the value of the width of |
| 19605 | each symbol. We can perform the calculation here and change the |
| 19606 | definition for @code{print-graph-body} from the way we defined it in the |
| 19607 | previous chapter. |
| 19608 | |
| 19609 | The length of the label for the horizontal axis must be at least as long |
| 19610 | as the graph. However, this information is used only in the function |
| 19611 | that prints the horizontal axis, so it does not need to be calculated here. |
| 19612 | |
| 19613 | These thoughts lead us directly to the following form for the varlist |
| 19614 | in the @code{let} for @code{print-graph}: |
| 19615 | |
| 19616 | @smallexample |
| 19617 | @group |
| 19618 | (let ((height (apply 'max numbers-list)) ; @r{First version.} |
| 19619 | (symbol-width (length graph-blank))) |
| 19620 | @end group |
| 19621 | @end smallexample |
| 19622 | |
| 19623 | @noindent |
| 19624 | As we shall see, this expression is not quite right. |
| 19625 | |
| 19626 | @need 2000 |
| 19627 | @node print-Y-axis |
| 19628 | @appendixsec The @code{print-Y-axis} Function |
| 19629 | @cindex Axis, print vertical |
| 19630 | @cindex Y axis printing |
| 19631 | @cindex Vertical axis printing |
| 19632 | @cindex Print vertical axis |
| 19633 | |
| 19634 | The job of the @code{print-Y-axis} function is to print a label for |
| 19635 | the vertical axis that looks like this: |
| 19636 | |
| 19637 | @smallexample |
| 19638 | @group |
| 19639 | 10 - |
| 19640 | |
| 19641 | |
| 19642 | |
| 19643 | |
| 19644 | 5 - |
| 19645 | |
| 19646 | |
| 19647 | |
| 19648 | 1 - |
| 19649 | @end group |
| 19650 | @end smallexample |
| 19651 | |
| 19652 | @noindent |
| 19653 | The function should be passed the height of the graph, and then should |
| 19654 | construct and insert the appropriate numbers and marks. |
| 19655 | |
| 19656 | @menu |
| 19657 | * print-Y-axis in Detail:: |
| 19658 | * Height of label:: What height for the Y axis? |
| 19659 | * Compute a Remainder:: How to compute the remainder of a division. |
| 19660 | * Y Axis Element:: Construct a line for the Y axis. |
| 19661 | * Y-axis-column:: Generate a list of Y axis labels. |
| 19662 | * print-Y-axis Penultimate:: A not quite final version. |
| 19663 | @end menu |
| 19664 | |
| 19665 | @ifnottex |
| 19666 | @node print-Y-axis in Detail |
| 19667 | @unnumberedsubsec The @code{print-Y-axis} Function in Detail |
| 19668 | @end ifnottex |
| 19669 | |
| 19670 | It is easy enough to see in the figure what the Y axis label should |
| 19671 | look like; but to say in words, and then to write a function |
| 19672 | definition to do the job is another matter. It is not quite true to |
| 19673 | say that we want a number and a tic every five lines: there are only |
| 19674 | three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4), |
| 19675 | but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8, |
| 19676 | and 9). It is better to say that we want a number and a tic mark on |
| 19677 | the base line (number 1) and then that we want a number and a tic on |
| 19678 | the fifth line from the bottom and on every line that is a multiple of |
| 19679 | five. |
| 19680 | |
| 19681 | @ifnottex |
| 19682 | @node Height of label |
| 19683 | @unnumberedsubsec What height should the label be? |
| 19684 | @end ifnottex |
| 19685 | |
| 19686 | The next issue is what height the label should be? Suppose the maximum |
| 19687 | height of tallest column of the graph is seven. Should the highest |
| 19688 | label on the Y axis be @samp{5 -}, and should the graph stick up above |
| 19689 | the label? Or should the highest label be @samp{7 -}, and mark the peak |
| 19690 | of the graph? Or should the highest label be @code{10 -}, which is a |
| 19691 | multiple of five, and be higher than the topmost value of the graph? |
| 19692 | |
| 19693 | The latter form is preferred. Most graphs are drawn within rectangles |
| 19694 | whose sides are an integral number of steps long---5, 10, 15, and so |
| 19695 | on for a step distance of five. But as soon as we decide to use a |
| 19696 | step height for the vertical axis, we discover that the simple |
| 19697 | expression in the varlist for computing the height is wrong. The |
| 19698 | expression is @code{(apply 'max numbers-list)}. This returns the |
| 19699 | precise height, not the maximum height plus whatever is necessary to |
| 19700 | round up to the nearest multiple of five. A more complex expression |
| 19701 | is required. |
| 19702 | |
| 19703 | As usual in cases like this, a complex problem becomes simpler if it is |
| 19704 | divided into several smaller problems. |
| 19705 | |
| 19706 | First, consider the case when the highest value of the graph is an |
| 19707 | integral multiple of five---when it is 5, 10, 15, or some higher |
| 19708 | multiple of five. We can use this value as the Y axis height. |
| 19709 | |
| 19710 | A fairly simply way to determine whether a number is a multiple of |
| 19711 | five is to divide it by five and see if the division results in a |
| 19712 | remainder. If there is no remainder, the number is a multiple of |
| 19713 | five. Thus, seven divided by five has a remainder of two, and seven |
| 19714 | is not an integral multiple of five. Put in slightly different |
| 19715 | language, more reminiscent of the classroom, five goes into seven |
| 19716 | once, with a remainder of two. However, five goes into ten twice, |
| 19717 | with no remainder: ten is an integral multiple of five. |
| 19718 | |
| 19719 | @node Compute a Remainder |
| 19720 | @appendixsubsec Side Trip: Compute a Remainder |
| 19721 | |
| 19722 | @findex % @r{(remainder function)} |
| 19723 | @cindex Remainder function, @code{%} |
| 19724 | In Lisp, the function for computing a remainder is @code{%}. The |
| 19725 | function returns the remainder of its first argument divided by its |
| 19726 | second argument. As it happens, @code{%} is a function in Emacs Lisp |
| 19727 | that you cannot discover using @code{apropos}: you find nothing if you |
| 19728 | type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to |
| 19729 | learn of the existence of @code{%} is to read about it in a book such |
| 19730 | as this or in the Emacs Lisp sources. |
| 19731 | |
| 19732 | You can try the @code{%} function by evaluating the following two |
| 19733 | expressions: |
| 19734 | |
| 19735 | @smallexample |
| 19736 | @group |
| 19737 | (% 7 5) |
| 19738 | |
| 19739 | (% 10 5) |
| 19740 | @end group |
| 19741 | @end smallexample |
| 19742 | |
| 19743 | @noindent |
| 19744 | The first expression returns 2 and the second expression returns 0. |
| 19745 | |
| 19746 | To test whether the returned value is zero or some other number, we |
| 19747 | can use the @code{zerop} function. This function returns @code{t} if |
| 19748 | its argument, which must be a number, is zero. |
| 19749 | |
| 19750 | @smallexample |
| 19751 | @group |
| 19752 | (zerop (% 7 5)) |
| 19753 | @result{} nil |
| 19754 | |
| 19755 | (zerop (% 10 5)) |
| 19756 | @result{} t |
| 19757 | @end group |
| 19758 | @end smallexample |
| 19759 | |
| 19760 | Thus, the following expression will return @code{t} if the height |
| 19761 | of the graph is evenly divisible by five: |
| 19762 | |
| 19763 | @smallexample |
| 19764 | (zerop (% height 5)) |
| 19765 | @end smallexample |
| 19766 | |
| 19767 | @noindent |
| 19768 | (The value of @code{height}, of course, can be found from @code{(apply |
| 19769 | 'max numbers-list)}.) |
| 19770 | |
| 19771 | On the other hand, if the value of @code{height} is not a multiple of |
| 19772 | five, we want to reset the value to the next higher multiple of five. |
| 19773 | This is straightforward arithmetic using functions with which we are |
| 19774 | already familiar. First, we divide the value of @code{height} by five |
| 19775 | to determine how many times five goes into the number. Thus, five |
| 19776 | goes into twelve twice. If we add one to this quotient and multiply by |
| 19777 | five, we will obtain the value of the next multiple of five that is |
| 19778 | larger than the height. Five goes into twelve twice. Add one to two, |
| 19779 | and multiply by five; the result is fifteen, which is the next multiple |
| 19780 | of five that is higher than twelve. The Lisp expression for this is: |
| 19781 | |
| 19782 | @smallexample |
| 19783 | (* (1+ (/ height 5)) 5) |
| 19784 | @end smallexample |
| 19785 | |
| 19786 | @noindent |
| 19787 | For example, if you evaluate the following, the result is 15: |
| 19788 | |
| 19789 | @smallexample |
| 19790 | (* (1+ (/ 12 5)) 5) |
| 19791 | @end smallexample |
| 19792 | |
| 19793 | All through this discussion, we have been using `five' as the value |
| 19794 | for spacing labels on the Y axis; but we may want to use some other |
| 19795 | value. For generality, we should replace `five' with a variable to |
| 19796 | which we can assign a value. The best name I can think of for this |
| 19797 | variable is @code{Y-axis-label-spacing}. |
| 19798 | |
| 19799 | @need 1250 |
| 19800 | Using this term, and an @code{if} expression, we produce the |
| 19801 | following: |
| 19802 | |
| 19803 | @smallexample |
| 19804 | @group |
| 19805 | (if (zerop (% height Y-axis-label-spacing)) |
| 19806 | height |
| 19807 | ;; @r{else} |
| 19808 | (* (1+ (/ height Y-axis-label-spacing)) |
| 19809 | Y-axis-label-spacing)) |
| 19810 | @end group |
| 19811 | @end smallexample |
| 19812 | |
| 19813 | @noindent |
| 19814 | This expression returns the value of @code{height} itself if the height |
| 19815 | is an even multiple of the value of the @code{Y-axis-label-spacing} or |
| 19816 | else it computes and returns a value of @code{height} that is equal to |
| 19817 | the next higher multiple of the value of the @code{Y-axis-label-spacing}. |
| 19818 | |
| 19819 | We can now include this expression in the @code{let} expression of the |
| 19820 | @code{print-graph} function (after first setting the value of |
| 19821 | @code{Y-axis-label-spacing}): |
| 19822 | @vindex Y-axis-label-spacing |
| 19823 | |
| 19824 | @smallexample |
| 19825 | @group |
| 19826 | (defvar Y-axis-label-spacing 5 |
| 19827 | "Number of lines from one Y axis label to next.") |
| 19828 | @end group |
| 19829 | |
| 19830 | @group |
| 19831 | @dots{} |
| 19832 | (let* ((height (apply 'max numbers-list)) |
| 19833 | (height-of-top-line |
| 19834 | (if (zerop (% height Y-axis-label-spacing)) |
| 19835 | height |
| 19836 | @end group |
| 19837 | @group |
| 19838 | ;; @r{else} |
| 19839 | (* (1+ (/ height Y-axis-label-spacing)) |
| 19840 | Y-axis-label-spacing))) |
| 19841 | (symbol-width (length graph-blank)))) |
| 19842 | @dots{} |
| 19843 | @end group |
| 19844 | @end smallexample |
| 19845 | |
| 19846 | @noindent |
| 19847 | (Note use of the @code{let*} function: the initial value of height is |
| 19848 | computed once by the @code{(apply 'max numbers-list)} expression and |
| 19849 | then the resulting value of @code{height} is used to compute its |
| 19850 | final value. @xref{fwd-para let, , The @code{let*} expression}, for |
| 19851 | more about @code{let*}.) |
| 19852 | |
| 19853 | @node Y Axis Element |
| 19854 | @appendixsubsec Construct a Y Axis Element |
| 19855 | |
| 19856 | When we print the vertical axis, we want to insert strings such as |
| 19857 | @w{@samp{5 -}} and @w{@samp{10 - }} every five lines. |
| 19858 | Moreover, we want the numbers and dashes to line up, so shorter |
| 19859 | numbers must be padded with leading spaces. If some of the strings |
| 19860 | use two digit numbers, the strings with single digit numbers must |
| 19861 | include a leading blank space before the number. |
| 19862 | |
| 19863 | @findex number-to-string |
| 19864 | To figure out the length of the number, the @code{length} function is |
| 19865 | used. But the @code{length} function works only with a string, not with |
| 19866 | a number. So the number has to be converted from being a number to |
| 19867 | being a string. This is done with the @code{number-to-string} function. |
| 19868 | For example, |
| 19869 | |
| 19870 | @smallexample |
| 19871 | @group |
| 19872 | (length (number-to-string 35)) |
| 19873 | @result{} 2 |
| 19874 | |
| 19875 | (length (number-to-string 100)) |
| 19876 | @result{} 3 |
| 19877 | @end group |
| 19878 | @end smallexample |
| 19879 | |
| 19880 | @noindent |
| 19881 | (@code{number-to-string} is also called @code{int-to-string}; you will |
| 19882 | see this alternative name in various sources.) |
| 19883 | |
| 19884 | In addition, in each label, each number is followed by a string such |
| 19885 | as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker. |
| 19886 | This variable is defined with @code{defvar}: |
| 19887 | |
| 19888 | @vindex Y-axis-tic |
| 19889 | @smallexample |
| 19890 | @group |
| 19891 | (defvar Y-axis-tic " - " |
| 19892 | "String that follows number in a Y axis label.") |
| 19893 | @end group |
| 19894 | @end smallexample |
| 19895 | |
| 19896 | The length of the Y label is the sum of the length of the Y axis tic |
| 19897 | mark and the length of the number of the top of the graph. |
| 19898 | |
| 19899 | @smallexample |
| 19900 | (length (concat (number-to-string height) Y-axis-tic))) |
| 19901 | @end smallexample |
| 19902 | |
| 19903 | This value will be calculated by the @code{print-graph} function in |
| 19904 | its varlist as @code{full-Y-label-width} and passed on. (Note that we |
| 19905 | did not think to include this in the varlist when we first proposed it.) |
| 19906 | |
| 19907 | To make a complete vertical axis label, a tic mark is concatenated |
| 19908 | with a number; and the two together may be preceded by one or more |
| 19909 | spaces depending on how long the number is. The label consists of |
| 19910 | three parts: the (optional) leading spaces, the number, and the tic |
| 19911 | mark. The function is passed the value of the number for the specific |
| 19912 | row, and the value of the width of the top line, which is calculated |
| 19913 | (just once) by @code{print-graph}. |
| 19914 | |
| 19915 | @smallexample |
| 19916 | @group |
| 19917 | (defun Y-axis-element (number full-Y-label-width) |
| 19918 | "Construct a NUMBERed label element. |
| 19919 | A numbered element looks like this ` 5 - ', |
| 19920 | and is padded as needed so all line up with |
| 19921 | the element for the largest number." |
| 19922 | @end group |
| 19923 | @group |
| 19924 | (let* ((leading-spaces |
| 19925 | (- full-Y-label-width |
| 19926 | (length |
| 19927 | (concat (number-to-string number) |
| 19928 | Y-axis-tic))))) |
| 19929 | @end group |
| 19930 | @group |
| 19931 | (concat |
| 19932 | (make-string leading-spaces ? ) |
| 19933 | (number-to-string number) |
| 19934 | Y-axis-tic))) |
| 19935 | @end group |
| 19936 | @end smallexample |
| 19937 | |
| 19938 | The @code{Y-axis-element} function concatenates together the leading |
| 19939 | spaces, if any; the number, as a string; and the tic mark. |
| 19940 | |
| 19941 | To figure out how many leading spaces the label will need, the |
| 19942 | function subtracts the actual length of the label---the length of the |
| 19943 | number plus the length of the tic mark---from the desired label width. |
| 19944 | |
| 19945 | @findex make-string |
| 19946 | Blank spaces are inserted using the @code{make-string} function. This |
| 19947 | function takes two arguments: the first tells it how long the string |
| 19948 | will be and the second is a symbol for the character to insert, in a |
| 19949 | special format. The format is a question mark followed by a blank |
| 19950 | space, like this, @samp{? }. @xref{Character Type, , Character Type, |
| 19951 | elisp, The GNU Emacs Lisp Reference Manual}, for a description of the |
| 19952 | syntax for characters. (Of course, you might want to replace the |
| 19953 | blank space by some other character @dots{} You know what to do.) |
| 19954 | |
| 19955 | The @code{number-to-string} function is used in the concatenation |
| 19956 | expression, to convert the number to a string that is concatenated |
| 19957 | with the leading spaces and the tic mark. |
| 19958 | |
| 19959 | @node Y-axis-column |
| 19960 | @appendixsubsec Create a Y Axis Column |
| 19961 | |
| 19962 | The preceding functions provide all the tools needed to construct a |
| 19963 | function that generates a list of numbered and blank strings to insert |
| 19964 | as the label for the vertical axis: |
| 19965 | |
| 19966 | @findex Y-axis-column |
| 19967 | @smallexample |
| 19968 | @group |
| 19969 | (defun Y-axis-column (height width-of-label) |
| 19970 | "Construct list of Y axis labels and blank strings. |
| 19971 | For HEIGHT of line above base and WIDTH-OF-LABEL." |
| 19972 | (let (Y-axis) |
| 19973 | @group |
| 19974 | @end group |
| 19975 | (while (> height 1) |
| 19976 | (if (zerop (% height Y-axis-label-spacing)) |
| 19977 | ;; @r{Insert label.} |
| 19978 | (setq Y-axis |
| 19979 | (cons |
| 19980 | (Y-axis-element height width-of-label) |
| 19981 | Y-axis)) |
| 19982 | @group |
| 19983 | @end group |
| 19984 | ;; @r{Else, insert blanks.} |
| 19985 | (setq Y-axis |
| 19986 | (cons |
| 19987 | (make-string width-of-label ? ) |
| 19988 | Y-axis))) |
| 19989 | (setq height (1- height))) |
| 19990 | ;; @r{Insert base line.} |
| 19991 | (setq Y-axis |
| 19992 | (cons (Y-axis-element 1 width-of-label) Y-axis)) |
| 19993 | (nreverse Y-axis))) |
| 19994 | @end group |
| 19995 | @end smallexample |
| 19996 | |
| 19997 | In this function, we start with the value of @code{height} and |
| 19998 | repetitively subtract one from its value. After each subtraction, we |
| 19999 | test to see whether the value is an integral multiple of the |
| 20000 | @code{Y-axis-label-spacing}. If it is, we construct a numbered label |
| 20001 | using the @code{Y-axis-element} function; if not, we construct a |
| 20002 | blank label using the @code{make-string} function. The base line |
| 20003 | consists of the number one followed by a tic mark. |
| 20004 | |
| 20005 | @need 2000 |
| 20006 | @node print-Y-axis Penultimate |
| 20007 | @appendixsubsec The Not Quite Final Version of @code{print-Y-axis} |
| 20008 | |
| 20009 | The list constructed by the @code{Y-axis-column} function is passed to |
| 20010 | the @code{print-Y-axis} function, which inserts the list as a column. |
| 20011 | |
| 20012 | @findex print-Y-axis |
| 20013 | @smallexample |
| 20014 | @group |
| 20015 | (defun print-Y-axis (height full-Y-label-width) |
| 20016 | "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH. |
| 20017 | Height must be the maximum height of the graph. |
| 20018 | Full width is the width of the highest label element." |
| 20019 | ;; Value of height and full-Y-label-width |
| 20020 | ;; are passed by `print-graph'. |
| 20021 | @end group |
| 20022 | @group |
| 20023 | (let ((start (point))) |
| 20024 | (insert-rectangle |
| 20025 | (Y-axis-column height full-Y-label-width)) |
| 20026 | ;; @r{Place point ready for inserting graph.} |
| 20027 | (goto-char start) |
| 20028 | ;; @r{Move point forward by value of} full-Y-label-width |
| 20029 | (forward-char full-Y-label-width))) |
| 20030 | @end group |
| 20031 | @end smallexample |
| 20032 | |
| 20033 | The @code{print-Y-axis} uses the @code{insert-rectangle} function to |
| 20034 | insert the Y axis labels created by the @code{Y-axis-column} function. |
| 20035 | In addition, it places point at the correct position for printing the body of |
| 20036 | the graph. |
| 20037 | |
| 20038 | You can test @code{print-Y-axis}: |
| 20039 | |
| 20040 | @enumerate |
| 20041 | @item |
| 20042 | Install |
| 20043 | |
| 20044 | @smallexample |
| 20045 | @group |
| 20046 | Y-axis-label-spacing |
| 20047 | Y-axis-tic |
| 20048 | Y-axis-element |
| 20049 | Y-axis-column |
| 20050 | print-Y-axis |
| 20051 | @end group |
| 20052 | @end smallexample |
| 20053 | |
| 20054 | @item |
| 20055 | Copy the following expression: |
| 20056 | |
| 20057 | @smallexample |
| 20058 | (print-Y-axis 12 5) |
| 20059 | @end smallexample |
| 20060 | |
| 20061 | @item |
| 20062 | Switch to the @file{*scratch*} buffer and place the cursor where you |
| 20063 | want the axis labels to start. |
| 20064 | |
| 20065 | @item |
| 20066 | Type @kbd{M-:} (@code{eval-expression}). |
| 20067 | |
| 20068 | @item |
| 20069 | Yank the @code{graph-body-print} expression into the minibuffer |
| 20070 | with @kbd{C-y} (@code{yank)}. |
| 20071 | |
| 20072 | @item |
| 20073 | Press @key{RET} to evaluate the expression. |
| 20074 | @end enumerate |
| 20075 | |
| 20076 | Emacs will print labels vertically, the top one being @w{@samp{10 -@w{ |
| 20077 | }}}. (The @code{print-graph} function will pass the value of |
| 20078 | @code{height-of-top-line}, which in this case will end up as 15, |
| 20079 | thereby getting rid of what might appear as a bug.) |
| 20080 | |
| 20081 | @need 2000 |
| 20082 | @node print-X-axis |
| 20083 | @appendixsec The @code{print-X-axis} Function |
| 20084 | @cindex Axis, print horizontal |
| 20085 | @cindex X axis printing |
| 20086 | @cindex Print horizontal axis |
| 20087 | @cindex Horizontal axis printing |
| 20088 | |
| 20089 | X axis labels are much like Y axis labels, except that the ticks are on a |
| 20090 | line above the numbers. Labels should look like this: |
| 20091 | |
| 20092 | @smallexample |
| 20093 | @group |
| 20094 | | | | | |
| 20095 | 1 5 10 15 |
| 20096 | @end group |
| 20097 | @end smallexample |
| 20098 | |
| 20099 | The first tic is under the first column of the graph and is preceded by |
| 20100 | several blank spaces. These spaces provide room in rows above for the Y |
| 20101 | axis labels. The second, third, fourth, and subsequent ticks are all |
| 20102 | spaced equally, according to the value of @code{X-axis-label-spacing}. |
| 20103 | |
| 20104 | The second row of the X axis consists of numbers, preceded by several |
| 20105 | blank spaces and also separated according to the value of the variable |
| 20106 | @code{X-axis-label-spacing}. |
| 20107 | |
| 20108 | The value of the variable @code{X-axis-label-spacing} should itself be |
| 20109 | measured in units of @code{symbol-width}, since you may want to change |
| 20110 | the width of the symbols that you are using to print the body of the |
| 20111 | graph without changing the ways the graph is labeled. |
| 20112 | |
| 20113 | @menu |
| 20114 | * Similarities differences:: Much like @code{print-Y-axis}, but not exactly. |
| 20115 | * X Axis Tic Marks:: Create tic marks for the horizontal axis. |
| 20116 | @end menu |
| 20117 | |
| 20118 | @ifnottex |
| 20119 | @node Similarities differences |
| 20120 | @unnumberedsubsec Similarities and differences |
| 20121 | @end ifnottex |
| 20122 | |
| 20123 | The @code{print-X-axis} function is constructed in more or less the |
| 20124 | same fashion as the @code{print-Y-axis} function except that it has |
| 20125 | two lines: the line of tic marks and the numbers. We will write a |
| 20126 | separate function to print each line and then combine them within the |
| 20127 | @code{print-X-axis} function. |
| 20128 | |
| 20129 | This is a three step process: |
| 20130 | |
| 20131 | @enumerate |
| 20132 | @item |
| 20133 | Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}. |
| 20134 | |
| 20135 | @item |
| 20136 | Write a function to print the X numbers, @code{print-X-axis-numbered-line}. |
| 20137 | |
| 20138 | @item |
| 20139 | Write a function to print both lines, the @code{print-X-axis} function, |
| 20140 | using @code{print-X-axis-tic-line} and |
| 20141 | @code{print-X-axis-numbered-line}. |
| 20142 | @end enumerate |
| 20143 | |
| 20144 | @node X Axis Tic Marks |
| 20145 | @appendixsubsec X Axis Tic Marks |
| 20146 | |
| 20147 | The first function should print the X axis tic marks. We must specify |
| 20148 | the tic marks themselves and their spacing: |
| 20149 | |
| 20150 | @smallexample |
| 20151 | @group |
| 20152 | (defvar X-axis-label-spacing |
| 20153 | (if (boundp 'graph-blank) |
| 20154 | (* 5 (length graph-blank)) 5) |
| 20155 | "Number of units from one X axis label to next.") |
| 20156 | @end group |
| 20157 | @end smallexample |
| 20158 | |
| 20159 | @noindent |
| 20160 | (Note that the value of @code{graph-blank} is set by another |
| 20161 | @code{defvar}. The @code{boundp} predicate checks whether it has |
| 20162 | already been set; @code{boundp} returns @code{nil} if it has not. If |
| 20163 | @code{graph-blank} were unbound and we did not use this conditional |
| 20164 | construction, in a recent GNU Emacs, we would enter the debugger and |
| 20165 | see an error message saying @samp{@w{Debugger entered--Lisp error:} |
| 20166 | @w{(void-variable graph-blank)}}.) |
| 20167 | |
| 20168 | @need 1200 |
| 20169 | Here is the @code{defvar} for @code{X-axis-tic-symbol}: |
| 20170 | |
| 20171 | @smallexample |
| 20172 | @group |
| 20173 | (defvar X-axis-tic-symbol "|" |
| 20174 | "String to insert to point to a column in X axis.") |
| 20175 | @end group |
| 20176 | @end smallexample |
| 20177 | |
| 20178 | @need 1250 |
| 20179 | The goal is to make a line that looks like this: |
| 20180 | |
| 20181 | @smallexample |
| 20182 | | | | | |
| 20183 | @end smallexample |
| 20184 | |
| 20185 | The first tic is indented so that it is under the first column, which is |
| 20186 | indented to provide space for the Y axis labels. |
| 20187 | |
| 20188 | A tic element consists of the blank spaces that stretch from one tic to |
| 20189 | the next plus a tic symbol. The number of blanks is determined by the |
| 20190 | width of the tic symbol and the @code{X-axis-label-spacing}. |
| 20191 | |
| 20192 | @need 1250 |
| 20193 | The code looks like this: |
| 20194 | |
| 20195 | @smallexample |
| 20196 | @group |
| 20197 | ;;; X-axis-tic-element |
| 20198 | @dots{} |
| 20199 | (concat |
| 20200 | (make-string |
| 20201 | ;; @r{Make a string of blanks.} |
| 20202 | (- (* symbol-width X-axis-label-spacing) |
| 20203 | (length X-axis-tic-symbol)) |
| 20204 | ? ) |
| 20205 | ;; @r{Concatenate blanks with tic symbol.} |
| 20206 | X-axis-tic-symbol) |
| 20207 | @dots{} |
| 20208 | @end group |
| 20209 | @end smallexample |
| 20210 | |
| 20211 | Next, we determine how many blanks are needed to indent the first tic |
| 20212 | mark to the first column of the graph. This uses the value of |
| 20213 | @code{full-Y-label-width} passed it by the @code{print-graph} function. |
| 20214 | |
| 20215 | @need 1250 |
| 20216 | The code to make @code{X-axis-leading-spaces} |
| 20217 | looks like this: |
| 20218 | |
| 20219 | @smallexample |
| 20220 | @group |
| 20221 | ;; X-axis-leading-spaces |
| 20222 | @dots{} |
| 20223 | (make-string full-Y-label-width ? ) |
| 20224 | @dots{} |
| 20225 | @end group |
| 20226 | @end smallexample |
| 20227 | |
| 20228 | We also need to determine the length of the horizontal axis, which is |
| 20229 | the length of the numbers list, and the number of ticks in the horizontal |
| 20230 | axis: |
| 20231 | |
| 20232 | @smallexample |
| 20233 | @group |
| 20234 | ;; X-length |
| 20235 | @dots{} |
| 20236 | (length numbers-list) |
| 20237 | @end group |
| 20238 | |
| 20239 | @group |
| 20240 | ;; tic-width |
| 20241 | @dots{} |
| 20242 | (* symbol-width X-axis-label-spacing) |
| 20243 | @end group |
| 20244 | |
| 20245 | @group |
| 20246 | ;; number-of-X-ticks |
| 20247 | (if (zerop (% (X-length tic-width))) |
| 20248 | (/ (X-length tic-width)) |
| 20249 | (1+ (/ (X-length tic-width)))) |
| 20250 | @end group |
| 20251 | @end smallexample |
| 20252 | |
| 20253 | @need 1250 |
| 20254 | All this leads us directly to the function for printing the X axis tic line: |
| 20255 | |
| 20256 | @findex print-X-axis-tic-line |
| 20257 | @smallexample |
| 20258 | @group |
| 20259 | (defun print-X-axis-tic-line |
| 20260 | (number-of-X-tics X-axis-leading-spaces X-axis-tic-element) |
| 20261 | "Print ticks for X axis." |
| 20262 | (insert X-axis-leading-spaces) |
| 20263 | (insert X-axis-tic-symbol) ; @r{Under first column.} |
| 20264 | @end group |
| 20265 | @group |
| 20266 | ;; @r{Insert second tic in the right spot.} |
| 20267 | (insert (concat |
| 20268 | (make-string |
| 20269 | (- (* symbol-width X-axis-label-spacing) |
| 20270 | ;; @r{Insert white space up to second tic symbol.} |
| 20271 | (* 2 (length X-axis-tic-symbol))) |
| 20272 | ? ) |
| 20273 | X-axis-tic-symbol)) |
| 20274 | @end group |
| 20275 | @group |
| 20276 | ;; @r{Insert remaining ticks.} |
| 20277 | (while (> number-of-X-tics 1) |
| 20278 | (insert X-axis-tic-element) |
| 20279 | (setq number-of-X-tics (1- number-of-X-tics)))) |
| 20280 | @end group |
| 20281 | @end smallexample |
| 20282 | |
| 20283 | The line of numbers is equally straightforward: |
| 20284 | |
| 20285 | @need 1250 |
| 20286 | First, we create a numbered element with blank spaces before each number: |
| 20287 | |
| 20288 | @findex X-axis-element |
| 20289 | @smallexample |
| 20290 | @group |
| 20291 | (defun X-axis-element (number) |
| 20292 | "Construct a numbered X axis element." |
| 20293 | (let ((leading-spaces |
| 20294 | (- (* symbol-width X-axis-label-spacing) |
| 20295 | (length (number-to-string number))))) |
| 20296 | (concat (make-string leading-spaces ? ) |
| 20297 | (number-to-string number)))) |
| 20298 | @end group |
| 20299 | @end smallexample |
| 20300 | |
| 20301 | Next, we create the function to print the numbered line, starting with |
| 20302 | the number ``1'' under the first column: |
| 20303 | |
| 20304 | @findex print-X-axis-numbered-line |
| 20305 | @smallexample |
| 20306 | @group |
| 20307 | (defun print-X-axis-numbered-line |
| 20308 | (number-of-X-tics X-axis-leading-spaces) |
| 20309 | "Print line of X-axis numbers" |
| 20310 | (let ((number X-axis-label-spacing)) |
| 20311 | (insert X-axis-leading-spaces) |
| 20312 | (insert "1") |
| 20313 | @end group |
| 20314 | @group |
| 20315 | (insert (concat |
| 20316 | (make-string |
| 20317 | ;; @r{Insert white space up to next number.} |
| 20318 | (- (* symbol-width X-axis-label-spacing) 2) |
| 20319 | ? ) |
| 20320 | (number-to-string number))) |
| 20321 | @end group |
| 20322 | @group |
| 20323 | ;; @r{Insert remaining numbers.} |
| 20324 | (setq number (+ number X-axis-label-spacing)) |
| 20325 | (while (> number-of-X-tics 1) |
| 20326 | (insert (X-axis-element number)) |
| 20327 | (setq number (+ number X-axis-label-spacing)) |
| 20328 | (setq number-of-X-tics (1- number-of-X-tics))))) |
| 20329 | @end group |
| 20330 | @end smallexample |
| 20331 | |
| 20332 | Finally, we need to write the @code{print-X-axis} that uses |
| 20333 | @code{print-X-axis-tic-line} and |
| 20334 | @code{print-X-axis-numbered-line}. |
| 20335 | |
| 20336 | The function must determine the local values of the variables used by both |
| 20337 | @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and |
| 20338 | then it must call them. Also, it must print the carriage return that |
| 20339 | separates the two lines. |
| 20340 | |
| 20341 | The function consists of a varlist that specifies five local variables, |
| 20342 | and calls to each of the two line printing functions: |
| 20343 | |
| 20344 | @findex print-X-axis |
| 20345 | @smallexample |
| 20346 | @group |
| 20347 | (defun print-X-axis (numbers-list) |
| 20348 | "Print X axis labels to length of NUMBERS-LIST." |
| 20349 | (let* ((leading-spaces |
| 20350 | (make-string full-Y-label-width ? )) |
| 20351 | @end group |
| 20352 | @group |
| 20353 | ;; symbol-width @r{is provided by} graph-body-print |
| 20354 | (tic-width (* symbol-width X-axis-label-spacing)) |
| 20355 | (X-length (length numbers-list)) |
| 20356 | @end group |
| 20357 | @group |
| 20358 | (X-tic |
| 20359 | (concat |
| 20360 | (make-string |
| 20361 | @end group |
| 20362 | @group |
| 20363 | ;; @r{Make a string of blanks.} |
| 20364 | (- (* symbol-width X-axis-label-spacing) |
| 20365 | (length X-axis-tic-symbol)) |
| 20366 | ? ) |
| 20367 | @end group |
| 20368 | @group |
| 20369 | ;; @r{Concatenate blanks with tic symbol.} |
| 20370 | X-axis-tic-symbol)) |
| 20371 | @end group |
| 20372 | @group |
| 20373 | (tic-number |
| 20374 | (if (zerop (% X-length tic-width)) |
| 20375 | (/ X-length tic-width) |
| 20376 | (1+ (/ X-length tic-width))))) |
| 20377 | @end group |
| 20378 | @group |
| 20379 | (print-X-axis-tic-line tic-number leading-spaces X-tic) |
| 20380 | (insert "\n") |
| 20381 | (print-X-axis-numbered-line tic-number leading-spaces))) |
| 20382 | @end group |
| 20383 | @end smallexample |
| 20384 | |
| 20385 | @need 1250 |
| 20386 | You can test @code{print-X-axis}: |
| 20387 | |
| 20388 | @enumerate |
| 20389 | @item |
| 20390 | Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing}, |
| 20391 | @code{print-X-axis-tic-line}, as well as @code{X-axis-element}, |
| 20392 | @code{print-X-axis-numbered-line}, and @code{print-X-axis}. |
| 20393 | |
| 20394 | @item |
| 20395 | Copy the following expression: |
| 20396 | |
| 20397 | @smallexample |
| 20398 | @group |
| 20399 | (progn |
| 20400 | (let ((full-Y-label-width 5) |
| 20401 | (symbol-width 1)) |
| 20402 | (print-X-axis |
| 20403 | '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16)))) |
| 20404 | @end group |
| 20405 | @end smallexample |
| 20406 | |
| 20407 | @item |
| 20408 | Switch to the @file{*scratch*} buffer and place the cursor where you |
| 20409 | want the axis labels to start. |
| 20410 | |
| 20411 | @item |
| 20412 | Type @kbd{M-:} (@code{eval-expression}). |
| 20413 | |
| 20414 | @item |
| 20415 | Yank the test expression into the minibuffer |
| 20416 | with @kbd{C-y} (@code{yank)}. |
| 20417 | |
| 20418 | @item |
| 20419 | Press @key{RET} to evaluate the expression. |
| 20420 | @end enumerate |
| 20421 | |
| 20422 | @need 1250 |
| 20423 | Emacs will print the horizontal axis like this: |
| 20424 | @sp 1 |
| 20425 | |
| 20426 | @smallexample |
| 20427 | @group |
| 20428 | | | | | | |
| 20429 | 1 5 10 15 20 |
| 20430 | @end group |
| 20431 | @end smallexample |
| 20432 | |
| 20433 | @node Print Whole Graph |
| 20434 | @appendixsec Printing the Whole Graph |
| 20435 | @cindex Printing the whole graph |
| 20436 | @cindex Whole graph printing |
| 20437 | @cindex Graph, printing all |
| 20438 | |
| 20439 | Now we are nearly ready to print the whole graph. |
| 20440 | |
| 20441 | The function to print the graph with the proper labels follows the |
| 20442 | outline we created earlier (@pxref{Full Graph, , A Graph with Labeled |
| 20443 | Axes}), but with additions. |
| 20444 | |
| 20445 | @need 1250 |
| 20446 | Here is the outline: |
| 20447 | |
| 20448 | @smallexample |
| 20449 | @group |
| 20450 | (defun print-graph (numbers-list) |
| 20451 | "@var{documentation}@dots{}" |
| 20452 | (let ((height @dots{} |
| 20453 | @dots{})) |
| 20454 | @end group |
| 20455 | @group |
| 20456 | (print-Y-axis height @dots{} ) |
| 20457 | (graph-body-print numbers-list) |
| 20458 | (print-X-axis @dots{} ))) |
| 20459 | @end group |
| 20460 | @end smallexample |
| 20461 | |
| 20462 | @menu |
| 20463 | * The final version:: A few changes. |
| 20464 | * Test print-graph:: Run a short test. |
| 20465 | * Graphing words in defuns:: Executing the final code. |
| 20466 | * lambda:: How to write an anonymous function. |
| 20467 | * mapcar:: Apply a function to elements of a list. |
| 20468 | * Another Bug:: Yet another bug @dots{} most insidious. |
| 20469 | * Final printed graph:: The graph itself! |
| 20470 | @end menu |
| 20471 | |
| 20472 | @ifnottex |
| 20473 | @node The final version |
| 20474 | @unnumberedsubsec Changes for the Final Version |
| 20475 | @end ifnottex |
| 20476 | |
| 20477 | The final version is different from what we planned in two ways: |
| 20478 | first, it contains additional values calculated once in the varlist; |
| 20479 | second, it carries an option to specify the labels' increment per row. |
| 20480 | This latter feature turns out to be essential; otherwise, a graph may |
| 20481 | have more rows than fit on a display or on a sheet of paper. |
| 20482 | |
| 20483 | @need 1500 |
| 20484 | This new feature requires a change to the @code{Y-axis-column} |
| 20485 | function, to add @code{vertical-step} to it. The function looks like |
| 20486 | this: |
| 20487 | |
| 20488 | @findex Y-axis-column @r{Final version.} |
| 20489 | @smallexample |
| 20490 | @group |
| 20491 | ;;; @r{Final version.} |
| 20492 | (defun Y-axis-column |
| 20493 | (height width-of-label &optional vertical-step) |
| 20494 | "Construct list of labels for Y axis. |
| 20495 | HEIGHT is maximum height of graph. |
| 20496 | WIDTH-OF-LABEL is maximum width of label. |
| 20497 | VERTICAL-STEP, an option, is a positive integer |
| 20498 | that specifies how much a Y axis label increments |
| 20499 | for each line. For example, a step of 5 means |
| 20500 | that each line is five units of the graph." |
| 20501 | @end group |
| 20502 | @group |
| 20503 | (let (Y-axis |
| 20504 | (number-per-line (or vertical-step 1))) |
| 20505 | (while (> height 1) |
| 20506 | (if (zerop (% height Y-axis-label-spacing)) |
| 20507 | @end group |
| 20508 | @group |
| 20509 | ;; @r{Insert label.} |
| 20510 | (setq Y-axis |
| 20511 | (cons |
| 20512 | (Y-axis-element |
| 20513 | (* height number-per-line) |
| 20514 | width-of-label) |
| 20515 | Y-axis)) |
| 20516 | @end group |
| 20517 | @group |
| 20518 | ;; @r{Else, insert blanks.} |
| 20519 | (setq Y-axis |
| 20520 | (cons |
| 20521 | (make-string width-of-label ? ) |
| 20522 | Y-axis))) |
| 20523 | (setq height (1- height))) |
| 20524 | @end group |
| 20525 | @group |
| 20526 | ;; @r{Insert base line.} |
| 20527 | (setq Y-axis (cons (Y-axis-element |
| 20528 | (or vertical-step 1) |
| 20529 | width-of-label) |
| 20530 | Y-axis)) |
| 20531 | (nreverse Y-axis))) |
| 20532 | @end group |
| 20533 | @end smallexample |
| 20534 | |
| 20535 | The values for the maximum height of graph and the width of a symbol |
| 20536 | are computed by @code{print-graph} in its @code{let} expression; so |
| 20537 | @code{graph-body-print} must be changed to accept them. |
| 20538 | |
| 20539 | @findex graph-body-print @r{Final version.} |
| 20540 | @smallexample |
| 20541 | @group |
| 20542 | ;;; @r{Final version.} |
| 20543 | (defun graph-body-print (numbers-list height symbol-width) |
| 20544 | "Print a bar graph of the NUMBERS-LIST. |
| 20545 | The numbers-list consists of the Y-axis values. |
| 20546 | HEIGHT is maximum height of graph. |
| 20547 | SYMBOL-WIDTH is number of each column." |
| 20548 | @end group |
| 20549 | @group |
| 20550 | (let (from-position) |
| 20551 | (while numbers-list |
| 20552 | (setq from-position (point)) |
| 20553 | (insert-rectangle |
| 20554 | (column-of-graph height (car numbers-list))) |
| 20555 | (goto-char from-position) |
| 20556 | (forward-char symbol-width) |
| 20557 | @end group |
| 20558 | @group |
| 20559 | ;; @r{Draw graph column by column.} |
| 20560 | (sit-for 0) |
| 20561 | (setq numbers-list (cdr numbers-list))) |
| 20562 | ;; @r{Place point for X axis labels.} |
| 20563 | (forward-line height) |
| 20564 | (insert "\n"))) |
| 20565 | @end group |
| 20566 | @end smallexample |
| 20567 | |
| 20568 | @need 1250 |
| 20569 | Finally, the code for the @code{print-graph} function: |
| 20570 | |
| 20571 | @findex print-graph @r{Final version.} |
| 20572 | @smallexample |
| 20573 | @group |
| 20574 | ;;; @r{Final version.} |
| 20575 | (defun print-graph |
| 20576 | (numbers-list &optional vertical-step) |
| 20577 | "Print labeled bar graph of the NUMBERS-LIST. |
| 20578 | The numbers-list consists of the Y-axis values. |
| 20579 | @end group |
| 20580 | |
| 20581 | @group |
| 20582 | Optionally, VERTICAL-STEP, a positive integer, |
| 20583 | specifies how much a Y axis label increments for |
| 20584 | each line. For example, a step of 5 means that |
| 20585 | each row is five units." |
| 20586 | @end group |
| 20587 | @group |
| 20588 | (let* ((symbol-width (length graph-blank)) |
| 20589 | ;; @code{height} @r{is both the largest number} |
| 20590 | ;; @r{and the number with the most digits.} |
| 20591 | (height (apply 'max numbers-list)) |
| 20592 | @end group |
| 20593 | @group |
| 20594 | (height-of-top-line |
| 20595 | (if (zerop (% height Y-axis-label-spacing)) |
| 20596 | height |
| 20597 | ;; @r{else} |
| 20598 | (* (1+ (/ height Y-axis-label-spacing)) |
| 20599 | Y-axis-label-spacing))) |
| 20600 | @end group |
| 20601 | @group |
| 20602 | (vertical-step (or vertical-step 1)) |
| 20603 | (full-Y-label-width |
| 20604 | (length |
| 20605 | @end group |
| 20606 | @group |
| 20607 | (concat |
| 20608 | (number-to-string |
| 20609 | (* height-of-top-line vertical-step)) |
| 20610 | Y-axis-tic)))) |
| 20611 | @end group |
| 20612 | |
| 20613 | @group |
| 20614 | (print-Y-axis |
| 20615 | height-of-top-line full-Y-label-width vertical-step) |
| 20616 | @end group |
| 20617 | @group |
| 20618 | (graph-body-print |
| 20619 | numbers-list height-of-top-line symbol-width) |
| 20620 | (print-X-axis numbers-list))) |
| 20621 | @end group |
| 20622 | @end smallexample |
| 20623 | |
| 20624 | @node Test print-graph |
| 20625 | @appendixsubsec Testing @code{print-graph} |
| 20626 | |
| 20627 | @need 1250 |
| 20628 | We can test the @code{print-graph} function with a short list of numbers: |
| 20629 | |
| 20630 | @enumerate |
| 20631 | @item |
| 20632 | Install the final versions of @code{Y-axis-column}, |
| 20633 | @code{graph-body-print}, and @code{print-graph} (in addition to the |
| 20634 | rest of the code.) |
| 20635 | |
| 20636 | @item |
| 20637 | Copy the following expression: |
| 20638 | |
| 20639 | @smallexample |
| 20640 | (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1)) |
| 20641 | @end smallexample |
| 20642 | |
| 20643 | @item |
| 20644 | Switch to the @file{*scratch*} buffer and place the cursor where you |
| 20645 | want the axis labels to start. |
| 20646 | |
| 20647 | @item |
| 20648 | Type @kbd{M-:} (@code{eval-expression}). |
| 20649 | |
| 20650 | @item |
| 20651 | Yank the test expression into the minibuffer |
| 20652 | with @kbd{C-y} (@code{yank)}. |
| 20653 | |
| 20654 | @item |
| 20655 | Press @key{RET} to evaluate the expression. |
| 20656 | @end enumerate |
| 20657 | |
| 20658 | @need 1250 |
| 20659 | Emacs will print a graph that looks like this: |
| 20660 | |
| 20661 | @smallexample |
| 20662 | @group |
| 20663 | 10 - |
| 20664 | |
| 20665 | |
| 20666 | * |
| 20667 | ** * |
| 20668 | 5 - **** * |
| 20669 | **** *** |
| 20670 | * ********* |
| 20671 | ************ |
| 20672 | 1 - ************* |
| 20673 | |
| 20674 | | | | | |
| 20675 | 1 5 10 15 |
| 20676 | @end group |
| 20677 | @end smallexample |
| 20678 | |
| 20679 | @need 1200 |
| 20680 | On the other hand, if you pass @code{print-graph} a |
| 20681 | @code{vertical-step} value of 2, by evaluating this expression: |
| 20682 | |
| 20683 | @smallexample |
| 20684 | (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2) |
| 20685 | @end smallexample |
| 20686 | |
| 20687 | @need 1250 |
| 20688 | @noindent |
| 20689 | The graph looks like this: |
| 20690 | |
| 20691 | @smallexample |
| 20692 | @group |
| 20693 | 20 - |
| 20694 | |
| 20695 | |
| 20696 | * |
| 20697 | ** * |
| 20698 | 10 - **** * |
| 20699 | **** *** |
| 20700 | * ********* |
| 20701 | ************ |
| 20702 | 2 - ************* |
| 20703 | |
| 20704 | | | | | |
| 20705 | 1 5 10 15 |
| 20706 | @end group |
| 20707 | @end smallexample |
| 20708 | |
| 20709 | @noindent |
| 20710 | (A question: is the `2' on the bottom of the vertical axis a bug or a |
| 20711 | feature? If you think it is a bug, and should be a `1' instead, (or |
| 20712 | even a `0'), you can modify the sources.) |
| 20713 | |
| 20714 | @node Graphing words in defuns |
| 20715 | @appendixsubsec Graphing Numbers of Words and Symbols |
| 20716 | |
| 20717 | Now for the graph for which all this code was written: a graph that |
| 20718 | shows how many function definitions contain fewer than 10 words and |
| 20719 | symbols, how many contain between 10 and 19 words and symbols, how |
| 20720 | many contain between 20 and 29 words and symbols, and so on. |
| 20721 | |
| 20722 | This is a multi-step process. First make sure you have loaded all the |
| 20723 | requisite code. |
| 20724 | |
| 20725 | @need 1500 |
| 20726 | It is a good idea to reset the value of @code{top-of-ranges} in case |
| 20727 | you have set it to some different value. You can evaluate the |
| 20728 | following: |
| 20729 | |
| 20730 | @smallexample |
| 20731 | @group |
| 20732 | (setq top-of-ranges |
| 20733 | '(10 20 30 40 50 |
| 20734 | 60 70 80 90 100 |
| 20735 | 110 120 130 140 150 |
| 20736 | 160 170 180 190 200 |
| 20737 | 210 220 230 240 250 |
| 20738 | 260 270 280 290 300) |
| 20739 | @end group |
| 20740 | @end smallexample |
| 20741 | |
| 20742 | @noindent |
| 20743 | Next create a list of the number of words and symbols in each range. |
| 20744 | |
| 20745 | @need 1500 |
| 20746 | @noindent |
| 20747 | Evaluate the following: |
| 20748 | |
| 20749 | @smallexample |
| 20750 | @group |
| 20751 | (setq list-for-graph |
| 20752 | (defuns-per-range |
| 20753 | (sort |
| 20754 | (recursive-lengths-list-many-files |
| 20755 | (directory-files "/usr/local/emacs/lisp" |
| 20756 | t ".+el$")) |
| 20757 | '<) |
| 20758 | top-of-ranges)) |
| 20759 | @end group |
| 20760 | @end smallexample |
| 20761 | |
| 20762 | @noindent |
| 20763 | On my old machine, this took about an hour. It looked though 303 Lisp |
| 20764 | files in my copy of Emacs version 19.23. After all that computing, |
| 20765 | the @code{list-for-graph} had this value: |
| 20766 | |
| 20767 | @smallexample |
| 20768 | @group |
| 20769 | (537 1027 955 785 594 483 349 292 224 199 166 120 116 99 |
| 20770 | 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220) |
| 20771 | @end group |
| 20772 | @end smallexample |
| 20773 | |
| 20774 | @noindent |
| 20775 | This means that my copy of Emacs had 537 function definitions with |
| 20776 | fewer than 10 words or symbols in them, 1,027 function definitions |
| 20777 | with 10 to 19 words or symbols in them, 955 function definitions with |
| 20778 | 20 to 29 words or symbols in them, and so on. |
| 20779 | |
| 20780 | Clearly, just by looking at this list we can see that most function |
| 20781 | definitions contain ten to thirty words and symbols. |
| 20782 | |
| 20783 | Now for printing. We do @emph{not} want to print a graph that is |
| 20784 | 1,030 lines high @dots{} Instead, we should print a graph that is |
| 20785 | fewer than twenty-five lines high. A graph that height can be |
| 20786 | displayed on almost any monitor, and easily printed on a sheet of paper. |
| 20787 | |
| 20788 | This means that each value in @code{list-for-graph} must be reduced to |
| 20789 | one-fiftieth its present value. |
| 20790 | |
| 20791 | Here is a short function to do just that, using two functions we have |
| 20792 | not yet seen, @code{mapcar} and @code{lambda}. |
| 20793 | |
| 20794 | @smallexample |
| 20795 | @group |
| 20796 | (defun one-fiftieth (full-range) |
| 20797 | "Return list, each number one-fiftieth of previous." |
| 20798 | (mapcar (lambda (arg) (/ arg 50)) full-range)) |
| 20799 | @end group |
| 20800 | @end smallexample |
| 20801 | |
| 20802 | @node lambda |
| 20803 | @appendixsubsec A @code{lambda} Expression: Useful Anonymity |
| 20804 | @cindex Anonymous function |
| 20805 | @findex lambda |
| 20806 | |
| 20807 | @code{lambda} is the symbol for an anonymous function, a function |
| 20808 | without a name. Every time you use an anonymous function, you need to |
| 20809 | include its whole body. |
| 20810 | |
| 20811 | @need 1250 |
| 20812 | @noindent |
| 20813 | Thus, |
| 20814 | |
| 20815 | @smallexample |
| 20816 | (lambda (arg) (/ arg 50)) |
| 20817 | @end smallexample |
| 20818 | |
| 20819 | @noindent |
| 20820 | is a function definition that says `return the value resulting from |
| 20821 | dividing whatever is passed to me as @code{arg} by 50'. |
| 20822 | |
| 20823 | @need 1200 |
| 20824 | Earlier, for example, we had a function @code{multiply-by-seven}; it |
| 20825 | multiplied its argument by 7. This function is similar, except it |
| 20826 | divides its argument by 50; and, it has no name. The anonymous |
| 20827 | equivalent of @code{multiply-by-seven} is: |
| 20828 | |
| 20829 | @smallexample |
| 20830 | (lambda (number) (* 7 number)) |
| 20831 | @end smallexample |
| 20832 | |
| 20833 | @noindent |
| 20834 | (@xref{defun, , The @code{defun} Macro}.) |
| 20835 | |
| 20836 | @need 1250 |
| 20837 | @noindent |
| 20838 | If we want to multiply 3 by 7, we can write: |
| 20839 | |
| 20840 | @c clear print-postscript-figures |
| 20841 | @c lambda example diagram #1 |
| 20842 | @ifnottex |
| 20843 | @smallexample |
| 20844 | @group |
| 20845 | (multiply-by-seven 3) |
| 20846 | \_______________/ ^ |
| 20847 | | | |
| 20848 | function argument |
| 20849 | @end group |
| 20850 | @end smallexample |
| 20851 | @end ifnottex |
| 20852 | @ifset print-postscript-figures |
| 20853 | @sp 1 |
| 20854 | @tex |
| 20855 | @center @image{lambda-1} |
| 20856 | @end tex |
| 20857 | @sp 1 |
| 20858 | @end ifset |
| 20859 | @ifclear print-postscript-figures |
| 20860 | @iftex |
| 20861 | @smallexample |
| 20862 | @group |
| 20863 | (multiply-by-seven 3) |
| 20864 | \_______________/ ^ |
| 20865 | | | |
| 20866 | function argument |
| 20867 | @end group |
| 20868 | @end smallexample |
| 20869 | @end iftex |
| 20870 | @end ifclear |
| 20871 | |
| 20872 | @noindent |
| 20873 | This expression returns 21. |
| 20874 | |
| 20875 | @need 1250 |
| 20876 | @noindent |
| 20877 | Similarly, we can write: |
| 20878 | |
| 20879 | @c lambda example diagram #2 |
| 20880 | @ifnottex |
| 20881 | @smallexample |
| 20882 | @group |
| 20883 | ((lambda (number) (* 7 number)) 3) |
| 20884 | \____________________________/ ^ |
| 20885 | | | |
| 20886 | anonymous function argument |
| 20887 | @end group |
| 20888 | @end smallexample |
| 20889 | @end ifnottex |
| 20890 | @ifset print-postscript-figures |
| 20891 | @sp 1 |
| 20892 | @tex |
| 20893 | @center @image{lambda-2} |
| 20894 | @end tex |
| 20895 | @sp 1 |
| 20896 | @end ifset |
| 20897 | @ifclear print-postscript-figures |
| 20898 | @iftex |
| 20899 | @smallexample |
| 20900 | @group |
| 20901 | ((lambda (number) (* 7 number)) 3) |
| 20902 | \____________________________/ ^ |
| 20903 | | | |
| 20904 | anonymous function argument |
| 20905 | @end group |
| 20906 | @end smallexample |
| 20907 | @end iftex |
| 20908 | @end ifclear |
| 20909 | |
| 20910 | @need 1250 |
| 20911 | @noindent |
| 20912 | If we want to divide 100 by 50, we can write: |
| 20913 | |
| 20914 | @c lambda example diagram #3 |
| 20915 | @ifnottex |
| 20916 | @smallexample |
| 20917 | @group |
| 20918 | ((lambda (arg) (/ arg 50)) 100) |
| 20919 | \______________________/ \_/ |
| 20920 | | | |
| 20921 | anonymous function argument |
| 20922 | @end group |
| 20923 | @end smallexample |
| 20924 | @end ifnottex |
| 20925 | @ifset print-postscript-figures |
| 20926 | @sp 1 |
| 20927 | @tex |
| 20928 | @center @image{lambda-3} |
| 20929 | @end tex |
| 20930 | @sp 1 |
| 20931 | @end ifset |
| 20932 | @ifclear print-postscript-figures |
| 20933 | @iftex |
| 20934 | @smallexample |
| 20935 | @group |
| 20936 | ((lambda (arg) (/ arg 50)) 100) |
| 20937 | \______________________/ \_/ |
| 20938 | | | |
| 20939 | anonymous function argument |
| 20940 | @end group |
| 20941 | @end smallexample |
| 20942 | @end iftex |
| 20943 | @end ifclear |
| 20944 | |
| 20945 | @noindent |
| 20946 | This expression returns 2. The 100 is passed to the function, which |
| 20947 | divides that number by 50. |
| 20948 | |
| 20949 | @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs |
| 20950 | Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda |
| 20951 | expressions derive from the Lambda Calculus. |
| 20952 | |
| 20953 | @node mapcar |
| 20954 | @appendixsubsec The @code{mapcar} Function |
| 20955 | @findex mapcar |
| 20956 | |
| 20957 | @code{mapcar} is a function that calls its first argument with each |
| 20958 | element of its second argument, in turn. The second argument must be |
| 20959 | a sequence. |
| 20960 | |
| 20961 | The @samp{map} part of the name comes from the mathematical phrase, |
| 20962 | `mapping over a domain', meaning to apply a function to each of the |
| 20963 | elements in a domain. The mathematical phrase is based on the |
| 20964 | metaphor of a surveyor walking, one step at a time, over an area he is |
| 20965 | mapping. And @samp{car}, of course, comes from the Lisp notion of the |
| 20966 | first of a list. |
| 20967 | |
| 20968 | @need 1250 |
| 20969 | @noindent |
| 20970 | For example, |
| 20971 | |
| 20972 | @smallexample |
| 20973 | @group |
| 20974 | (mapcar '1+ '(2 4 6)) |
| 20975 | @result{} (3 5 7) |
| 20976 | @end group |
| 20977 | @end smallexample |
| 20978 | |
| 20979 | @noindent |
| 20980 | The function @code{1+} which adds one to its argument, is executed on |
| 20981 | @emph{each} element of the list, and a new list is returned. |
| 20982 | |
| 20983 | Contrast this with @code{apply}, which applies its first argument to |
| 20984 | all the remaining. |
| 20985 | (@xref{Readying a Graph, , Readying a Graph}, for a explanation of |
| 20986 | @code{apply}.) |
| 20987 | |
| 20988 | @need 1250 |
| 20989 | In the definition of @code{one-fiftieth}, the first argument is the |
| 20990 | anonymous function: |
| 20991 | |
| 20992 | @smallexample |
| 20993 | (lambda (arg) (/ arg 50)) |
| 20994 | @end smallexample |
| 20995 | |
| 20996 | @noindent |
| 20997 | and the second argument is @code{full-range}, which will be bound to |
| 20998 | @code{list-for-graph}. |
| 20999 | |
| 21000 | @need 1250 |
| 21001 | The whole expression looks like this: |
| 21002 | |
| 21003 | @smallexample |
| 21004 | (mapcar (lambda (arg) (/ arg 50)) full-range)) |
| 21005 | @end smallexample |
| 21006 | |
| 21007 | @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs |
| 21008 | Lisp Reference Manual}, for more about @code{mapcar}. |
| 21009 | |
| 21010 | Using the @code{one-fiftieth} function, we can generate a list in |
| 21011 | which each element is one-fiftieth the size of the corresponding |
| 21012 | element in @code{list-for-graph}. |
| 21013 | |
| 21014 | @smallexample |
| 21015 | @group |
| 21016 | (setq fiftieth-list-for-graph |
| 21017 | (one-fiftieth list-for-graph)) |
| 21018 | @end group |
| 21019 | @end smallexample |
| 21020 | |
| 21021 | @need 1250 |
| 21022 | The resulting list looks like this: |
| 21023 | |
| 21024 | @smallexample |
| 21025 | @group |
| 21026 | (10 20 19 15 11 9 6 5 4 3 3 2 2 |
| 21027 | 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4) |
| 21028 | @end group |
| 21029 | @end smallexample |
| 21030 | |
| 21031 | @noindent |
| 21032 | This, we are almost ready to print! (We also notice the loss of |
| 21033 | information: many of the higher ranges are 0, meaning that fewer than |
| 21034 | 50 defuns had that many words or symbols---but not necessarily meaning |
| 21035 | that none had that many words or symbols.) |
| 21036 | |
| 21037 | @node Another Bug |
| 21038 | @appendixsubsec Another Bug @dots{} Most Insidious |
| 21039 | @cindex Bug, most insidious type |
| 21040 | @cindex Insidious type of bug |
| 21041 | |
| 21042 | I said `almost ready to print'! Of course, there is a bug in the |
| 21043 | @code{print-graph} function @dots{} It has a @code{vertical-step} |
| 21044 | option, but not a @code{horizontal-step} option. The |
| 21045 | @code{top-of-range} scale goes from 10 to 300 by tens. But the |
| 21046 | @code{print-graph} function will print only by ones. |
| 21047 | |
| 21048 | This is a classic example of what some consider the most insidious |
| 21049 | type of bug, the bug of omission. This is not the kind of bug you can |
| 21050 | find by studying the code, for it is not in the code; it is an omitted |
| 21051 | feature. Your best actions are to try your program early and often; |
| 21052 | and try to arrange, as much as you can, to write code that is easy to |
| 21053 | understand and easy to change. Try to be aware, whenever you can, |
| 21054 | that whatever you have written, @emph{will} be rewritten, if not soon, |
| 21055 | eventually. A hard maxim to follow. |
| 21056 | |
| 21057 | It is the @code{print-X-axis-numbered-line} function that needs the |
| 21058 | work; and then the @code{print-X-axis} and the @code{print-graph} |
| 21059 | functions need to be adapted. Not much needs to be done; there is one |
| 21060 | nicety: the numbers ought to line up under the tic marks. This takes |
| 21061 | a little thought. |
| 21062 | |
| 21063 | @need 1250 |
| 21064 | Here is the corrected @code{print-X-axis-numbered-line}: |
| 21065 | |
| 21066 | @smallexample |
| 21067 | @group |
| 21068 | (defun print-X-axis-numbered-line |
| 21069 | (number-of-X-tics X-axis-leading-spaces |
| 21070 | &optional horizontal-step) |
| 21071 | "Print line of X-axis numbers" |
| 21072 | (let ((number X-axis-label-spacing) |
| 21073 | (horizontal-step (or horizontal-step 1))) |
| 21074 | @end group |
| 21075 | @group |
| 21076 | (insert X-axis-leading-spaces) |
| 21077 | ;; @r{Delete extra leading spaces.} |
| 21078 | (delete-char |
| 21079 | (- (1- |
| 21080 | (length (number-to-string horizontal-step))))) |
| 21081 | (insert (concat |
| 21082 | (make-string |
| 21083 | @end group |
| 21084 | @group |
| 21085 | ;; @r{Insert white space.} |
| 21086 | (- (* symbol-width |
| 21087 | X-axis-label-spacing) |
| 21088 | (1- |
| 21089 | (length |
| 21090 | (number-to-string horizontal-step))) |
| 21091 | 2) |
| 21092 | ? ) |
| 21093 | (number-to-string |
| 21094 | (* number horizontal-step)))) |
| 21095 | @end group |
| 21096 | @group |
| 21097 | ;; @r{Insert remaining numbers.} |
| 21098 | (setq number (+ number X-axis-label-spacing)) |
| 21099 | (while (> number-of-X-tics 1) |
| 21100 | (insert (X-axis-element |
| 21101 | (* number horizontal-step))) |
| 21102 | (setq number (+ number X-axis-label-spacing)) |
| 21103 | (setq number-of-X-tics (1- number-of-X-tics))))) |
| 21104 | @end group |
| 21105 | @end smallexample |
| 21106 | |
| 21107 | @need 1500 |
| 21108 | If you are reading this in Info, you can see the new versions of |
| 21109 | @code{print-X-axis} @code{print-graph} and evaluate them. If you are |
| 21110 | reading this in a printed book, you can see the changed lines here |
| 21111 | (the full text is too much to print). |
| 21112 | |
| 21113 | @iftex |
| 21114 | @smallexample |
| 21115 | @group |
| 21116 | (defun print-X-axis (numbers-list horizontal-step) |
| 21117 | @dots{} |
| 21118 | (print-X-axis-numbered-line |
| 21119 | tic-number leading-spaces horizontal-step)) |
| 21120 | @end group |
| 21121 | @end smallexample |
| 21122 | |
| 21123 | @smallexample |
| 21124 | @group |
| 21125 | (defun print-graph |
| 21126 | (numbers-list |
| 21127 | &optional vertical-step horizontal-step) |
| 21128 | @dots{} |
| 21129 | (print-X-axis numbers-list horizontal-step)) |
| 21130 | @end group |
| 21131 | @end smallexample |
| 21132 | @end iftex |
| 21133 | |
| 21134 | @ifnottex |
| 21135 | @smallexample |
| 21136 | @group |
| 21137 | (defun print-X-axis (numbers-list horizontal-step) |
| 21138 | "Print X axis labels to length of NUMBERS-LIST. |
| 21139 | Optionally, HORIZONTAL-STEP, a positive integer, |
| 21140 | specifies how much an X axis label increments for |
| 21141 | each column." |
| 21142 | @end group |
| 21143 | @group |
| 21144 | ;; Value of symbol-width and full-Y-label-width |
| 21145 | ;; are passed by `print-graph'. |
| 21146 | (let* ((leading-spaces |
| 21147 | (make-string full-Y-label-width ? )) |
| 21148 | ;; symbol-width @r{is provided by} graph-body-print |
| 21149 | (tic-width (* symbol-width X-axis-label-spacing)) |
| 21150 | (X-length (length numbers-list)) |
| 21151 | @end group |
| 21152 | @group |
| 21153 | (X-tic |
| 21154 | (concat |
| 21155 | (make-string |
| 21156 | ;; @r{Make a string of blanks.} |
| 21157 | (- (* symbol-width X-axis-label-spacing) |
| 21158 | (length X-axis-tic-symbol)) |
| 21159 | ? ) |
| 21160 | @end group |
| 21161 | @group |
| 21162 | ;; @r{Concatenate blanks with tic symbol.} |
| 21163 | X-axis-tic-symbol)) |
| 21164 | (tic-number |
| 21165 | (if (zerop (% X-length tic-width)) |
| 21166 | (/ X-length tic-width) |
| 21167 | (1+ (/ X-length tic-width))))) |
| 21168 | @end group |
| 21169 | |
| 21170 | @group |
| 21171 | (print-X-axis-tic-line |
| 21172 | tic-number leading-spaces X-tic) |
| 21173 | (insert "\n") |
| 21174 | (print-X-axis-numbered-line |
| 21175 | tic-number leading-spaces horizontal-step))) |
| 21176 | @end group |
| 21177 | @end smallexample |
| 21178 | |
| 21179 | @smallexample |
| 21180 | @group |
| 21181 | (defun print-graph |
| 21182 | (numbers-list &optional vertical-step horizontal-step) |
| 21183 | "Print labeled bar graph of the NUMBERS-LIST. |
| 21184 | The numbers-list consists of the Y-axis values. |
| 21185 | @end group |
| 21186 | |
| 21187 | @group |
| 21188 | Optionally, VERTICAL-STEP, a positive integer, |
| 21189 | specifies how much a Y axis label increments for |
| 21190 | each line. For example, a step of 5 means that |
| 21191 | each row is five units. |
| 21192 | @end group |
| 21193 | |
| 21194 | @group |
| 21195 | Optionally, HORIZONTAL-STEP, a positive integer, |
| 21196 | specifies how much an X axis label increments for |
| 21197 | each column." |
| 21198 | (let* ((symbol-width (length graph-blank)) |
| 21199 | ;; @code{height} @r{is both the largest number} |
| 21200 | ;; @r{and the number with the most digits.} |
| 21201 | (height (apply 'max numbers-list)) |
| 21202 | @end group |
| 21203 | @group |
| 21204 | (height-of-top-line |
| 21205 | (if (zerop (% height Y-axis-label-spacing)) |
| 21206 | height |
| 21207 | ;; @r{else} |
| 21208 | (* (1+ (/ height Y-axis-label-spacing)) |
| 21209 | Y-axis-label-spacing))) |
| 21210 | @end group |
| 21211 | @group |
| 21212 | (vertical-step (or vertical-step 1)) |
| 21213 | (full-Y-label-width |
| 21214 | (length |
| 21215 | (concat |
| 21216 | (number-to-string |
| 21217 | (* height-of-top-line vertical-step)) |
| 21218 | Y-axis-tic)))) |
| 21219 | @end group |
| 21220 | @group |
| 21221 | (print-Y-axis |
| 21222 | height-of-top-line full-Y-label-width vertical-step) |
| 21223 | (graph-body-print |
| 21224 | numbers-list height-of-top-line symbol-width) |
| 21225 | (print-X-axis numbers-list horizontal-step))) |
| 21226 | @end group |
| 21227 | @end smallexample |
| 21228 | @end ifnottex |
| 21229 | |
| 21230 | @c qqq |
| 21231 | @ignore |
| 21232 | Graphing Definitions Re-listed |
| 21233 | |
| 21234 | @need 1250 |
| 21235 | Here are all the graphing definitions in their final form: |
| 21236 | |
| 21237 | @smallexample |
| 21238 | @group |
| 21239 | (defvar top-of-ranges |
| 21240 | '(10 20 30 40 50 |
| 21241 | 60 70 80 90 100 |
| 21242 | 110 120 130 140 150 |
| 21243 | 160 170 180 190 200 |
| 21244 | 210 220 230 240 250) |
| 21245 | "List specifying ranges for `defuns-per-range'.") |
| 21246 | @end group |
| 21247 | |
| 21248 | @group |
| 21249 | (defvar graph-symbol "*" |
| 21250 | "String used as symbol in graph, usually an asterisk.") |
| 21251 | @end group |
| 21252 | |
| 21253 | @group |
| 21254 | (defvar graph-blank " " |
| 21255 | "String used as blank in graph, usually a blank space. |
| 21256 | graph-blank must be the same number of columns wide |
| 21257 | as graph-symbol.") |
| 21258 | @end group |
| 21259 | |
| 21260 | @group |
| 21261 | (defvar Y-axis-tic " - " |
| 21262 | "String that follows number in a Y axis label.") |
| 21263 | @end group |
| 21264 | |
| 21265 | @group |
| 21266 | (defvar Y-axis-label-spacing 5 |
| 21267 | "Number of lines from one Y axis label to next.") |
| 21268 | @end group |
| 21269 | |
| 21270 | @group |
| 21271 | (defvar X-axis-tic-symbol "|" |
| 21272 | "String to insert to point to a column in X axis.") |
| 21273 | @end group |
| 21274 | |
| 21275 | @group |
| 21276 | (defvar X-axis-label-spacing |
| 21277 | (if (boundp 'graph-blank) |
| 21278 | (* 5 (length graph-blank)) 5) |
| 21279 | "Number of units from one X axis label to next.") |
| 21280 | @end group |
| 21281 | @end smallexample |
| 21282 | |
| 21283 | @smallexample |
| 21284 | @group |
| 21285 | (defun count-words-in-defun () |
| 21286 | "Return the number of words and symbols in a defun." |
| 21287 | (beginning-of-defun) |
| 21288 | (let ((count 0) |
| 21289 | (end (save-excursion (end-of-defun) (point)))) |
| 21290 | @end group |
| 21291 | |
| 21292 | @group |
| 21293 | (while |
| 21294 | (and (< (point) end) |
| 21295 | (re-search-forward |
| 21296 | "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" |
| 21297 | end t)) |
| 21298 | (setq count (1+ count))) |
| 21299 | count)) |
| 21300 | @end group |
| 21301 | @end smallexample |
| 21302 | |
| 21303 | @smallexample |
| 21304 | @group |
| 21305 | (defun lengths-list-file (filename) |
| 21306 | "Return list of definitions' lengths within FILE. |
| 21307 | The returned list is a list of numbers. |
| 21308 | Each number is the number of words or |
| 21309 | symbols in one function definition." |
| 21310 | @end group |
| 21311 | |
| 21312 | @group |
| 21313 | (message "Working on `%s' ... " filename) |
| 21314 | (save-excursion |
| 21315 | (let ((buffer (find-file-noselect filename)) |
| 21316 | (lengths-list)) |
| 21317 | (set-buffer buffer) |
| 21318 | (setq buffer-read-only t) |
| 21319 | (widen) |
| 21320 | (goto-char (point-min)) |
| 21321 | @end group |
| 21322 | |
| 21323 | @group |
| 21324 | (while (re-search-forward "^(defun" nil t) |
| 21325 | (setq lengths-list |
| 21326 | (cons (count-words-in-defun) lengths-list))) |
| 21327 | (kill-buffer buffer) |
| 21328 | lengths-list))) |
| 21329 | @end group |
| 21330 | @end smallexample |
| 21331 | |
| 21332 | @smallexample |
| 21333 | @group |
| 21334 | (defun lengths-list-many-files (list-of-files) |
| 21335 | "Return list of lengths of defuns in LIST-OF-FILES." |
| 21336 | (let (lengths-list) |
| 21337 | ;;; @r{true-or-false-test} |
| 21338 | (while list-of-files |
| 21339 | (setq lengths-list |
| 21340 | (append |
| 21341 | lengths-list |
| 21342 | @end group |
| 21343 | @group |
| 21344 | ;;; @r{Generate a lengths' list.} |
| 21345 | (lengths-list-file |
| 21346 | (expand-file-name (car list-of-files))))) |
| 21347 | ;;; @r{Make files' list shorter.} |
| 21348 | (setq list-of-files (cdr list-of-files))) |
| 21349 | ;;; @r{Return final value of lengths' list.} |
| 21350 | lengths-list)) |
| 21351 | @end group |
| 21352 | @end smallexample |
| 21353 | |
| 21354 | @smallexample |
| 21355 | @group |
| 21356 | (defun defuns-per-range (sorted-lengths top-of-ranges) |
| 21357 | "SORTED-LENGTHS defuns in each TOP-OF-RANGES range." |
| 21358 | (let ((top-of-range (car top-of-ranges)) |
| 21359 | (number-within-range 0) |
| 21360 | defuns-per-range-list) |
| 21361 | @end group |
| 21362 | |
| 21363 | @group |
| 21364 | ;; @r{Outer loop.} |
| 21365 | (while top-of-ranges |
| 21366 | |
| 21367 | ;; @r{Inner loop.} |
| 21368 | (while (and |
| 21369 | ;; @r{Need number for numeric test.} |
| 21370 | (car sorted-lengths) |
| 21371 | (< (car sorted-lengths) top-of-range)) |
| 21372 | |
| 21373 | ;; @r{Count number of definitions within current range.} |
| 21374 | (setq number-within-range (1+ number-within-range)) |
| 21375 | (setq sorted-lengths (cdr sorted-lengths))) |
| 21376 | @end group |
| 21377 | |
| 21378 | @group |
| 21379 | ;; @r{Exit inner loop but remain within outer loop.} |
| 21380 | |
| 21381 | (setq defuns-per-range-list |
| 21382 | (cons number-within-range defuns-per-range-list)) |
| 21383 | (setq number-within-range 0) ; @r{Reset count to zero.} |
| 21384 | |
| 21385 | ;; @r{Move to next range.} |
| 21386 | (setq top-of-ranges (cdr top-of-ranges)) |
| 21387 | ;; @r{Specify next top of range value.} |
| 21388 | (setq top-of-range (car top-of-ranges))) |
| 21389 | @end group |
| 21390 | |
| 21391 | @group |
| 21392 | ;; @r{Exit outer loop and count the number of defuns larger than} |
| 21393 | ;; @r{ the largest top-of-range value.} |
| 21394 | (setq defuns-per-range-list |
| 21395 | (cons |
| 21396 | (length sorted-lengths) |
| 21397 | defuns-per-range-list)) |
| 21398 | |
| 21399 | ;; @r{Return a list of the number of definitions within each range,} |
| 21400 | ;; @r{ smallest to largest.} |
| 21401 | (nreverse defuns-per-range-list))) |
| 21402 | @end group |
| 21403 | @end smallexample |
| 21404 | |
| 21405 | @smallexample |
| 21406 | @group |
| 21407 | (defun column-of-graph (max-graph-height actual-height) |
| 21408 | "Return list of MAX-GRAPH-HEIGHT strings; |
| 21409 | ACTUAL-HEIGHT are graph-symbols. |
| 21410 | The graph-symbols are contiguous entries at the end |
| 21411 | of the list. |
| 21412 | The list will be inserted as one column of a graph. |
| 21413 | The strings are either graph-blank or graph-symbol." |
| 21414 | @end group |
| 21415 | |
| 21416 | @group |
| 21417 | (let ((insert-list nil) |
| 21418 | (number-of-top-blanks |
| 21419 | (- max-graph-height actual-height))) |
| 21420 | |
| 21421 | ;; @r{Fill in @code{graph-symbols}.} |
| 21422 | (while (> actual-height 0) |
| 21423 | (setq insert-list (cons graph-symbol insert-list)) |
| 21424 | (setq actual-height (1- actual-height))) |
| 21425 | @end group |
| 21426 | |
| 21427 | @group |
| 21428 | ;; @r{Fill in @code{graph-blanks}.} |
| 21429 | (while (> number-of-top-blanks 0) |
| 21430 | (setq insert-list (cons graph-blank insert-list)) |
| 21431 | (setq number-of-top-blanks |
| 21432 | (1- number-of-top-blanks))) |
| 21433 | |
| 21434 | ;; @r{Return whole list.} |
| 21435 | insert-list)) |
| 21436 | @end group |
| 21437 | @end smallexample |
| 21438 | |
| 21439 | @smallexample |
| 21440 | @group |
| 21441 | (defun Y-axis-element (number full-Y-label-width) |
| 21442 | "Construct a NUMBERed label element. |
| 21443 | A numbered element looks like this ` 5 - ', |
| 21444 | and is padded as needed so all line up with |
| 21445 | the element for the largest number." |
| 21446 | @end group |
| 21447 | @group |
| 21448 | (let* ((leading-spaces |
| 21449 | (- full-Y-label-width |
| 21450 | (length |
| 21451 | (concat (number-to-string number) |
| 21452 | Y-axis-tic))))) |
| 21453 | @end group |
| 21454 | @group |
| 21455 | (concat |
| 21456 | (make-string leading-spaces ? ) |
| 21457 | (number-to-string number) |
| 21458 | Y-axis-tic))) |
| 21459 | @end group |
| 21460 | @end smallexample |
| 21461 | |
| 21462 | @smallexample |
| 21463 | @group |
| 21464 | (defun print-Y-axis |
| 21465 | (height full-Y-label-width &optional vertical-step) |
| 21466 | "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH. |
| 21467 | Height must be the maximum height of the graph. |
| 21468 | Full width is the width of the highest label element. |
| 21469 | Optionally, print according to VERTICAL-STEP." |
| 21470 | @end group |
| 21471 | @group |
| 21472 | ;; Value of height and full-Y-label-width |
| 21473 | ;; are passed by `print-graph'. |
| 21474 | (let ((start (point))) |
| 21475 | (insert-rectangle |
| 21476 | (Y-axis-column height full-Y-label-width vertical-step)) |
| 21477 | @end group |
| 21478 | @group |
| 21479 | ;; @r{Place point ready for inserting graph.} |
| 21480 | (goto-char start) |
| 21481 | ;; @r{Move point forward by value of} full-Y-label-width |
| 21482 | (forward-char full-Y-label-width))) |
| 21483 | @end group |
| 21484 | @end smallexample |
| 21485 | |
| 21486 | @smallexample |
| 21487 | @group |
| 21488 | (defun print-X-axis-tic-line |
| 21489 | (number-of-X-tics X-axis-leading-spaces X-axis-tic-element) |
| 21490 | "Print ticks for X axis." |
| 21491 | (insert X-axis-leading-spaces) |
| 21492 | (insert X-axis-tic-symbol) ; @r{Under first column.} |
| 21493 | @end group |
| 21494 | @group |
| 21495 | ;; @r{Insert second tic in the right spot.} |
| 21496 | (insert (concat |
| 21497 | (make-string |
| 21498 | (- (* symbol-width X-axis-label-spacing) |
| 21499 | ;; @r{Insert white space up to second tic symbol.} |
| 21500 | (* 2 (length X-axis-tic-symbol))) |
| 21501 | ? ) |
| 21502 | X-axis-tic-symbol)) |
| 21503 | @end group |
| 21504 | @group |
| 21505 | ;; @r{Insert remaining ticks.} |
| 21506 | (while (> number-of-X-tics 1) |
| 21507 | (insert X-axis-tic-element) |
| 21508 | (setq number-of-X-tics (1- number-of-X-tics)))) |
| 21509 | @end group |
| 21510 | @end smallexample |
| 21511 | |
| 21512 | @smallexample |
| 21513 | @group |
| 21514 | (defun X-axis-element (number) |
| 21515 | "Construct a numbered X axis element." |
| 21516 | (let ((leading-spaces |
| 21517 | (- (* symbol-width X-axis-label-spacing) |
| 21518 | (length (number-to-string number))))) |
| 21519 | (concat (make-string leading-spaces ? ) |
| 21520 | (number-to-string number)))) |
| 21521 | @end group |
| 21522 | @end smallexample |
| 21523 | |
| 21524 | @smallexample |
| 21525 | @group |
| 21526 | (defun graph-body-print (numbers-list height symbol-width) |
| 21527 | "Print a bar graph of the NUMBERS-LIST. |
| 21528 | The numbers-list consists of the Y-axis values. |
| 21529 | HEIGHT is maximum height of graph. |
| 21530 | SYMBOL-WIDTH is number of each column." |
| 21531 | @end group |
| 21532 | @group |
| 21533 | (let (from-position) |
| 21534 | (while numbers-list |
| 21535 | (setq from-position (point)) |
| 21536 | (insert-rectangle |
| 21537 | (column-of-graph height (car numbers-list))) |
| 21538 | (goto-char from-position) |
| 21539 | (forward-char symbol-width) |
| 21540 | @end group |
| 21541 | @group |
| 21542 | ;; @r{Draw graph column by column.} |
| 21543 | (sit-for 0) |
| 21544 | (setq numbers-list (cdr numbers-list))) |
| 21545 | ;; @r{Place point for X axis labels.} |
| 21546 | (forward-line height) |
| 21547 | (insert "\n"))) |
| 21548 | @end group |
| 21549 | @end smallexample |
| 21550 | |
| 21551 | @smallexample |
| 21552 | @group |
| 21553 | (defun Y-axis-column |
| 21554 | (height width-of-label &optional vertical-step) |
| 21555 | "Construct list of labels for Y axis. |
| 21556 | HEIGHT is maximum height of graph. |
| 21557 | WIDTH-OF-LABEL is maximum width of label. |
| 21558 | @end group |
| 21559 | @group |
| 21560 | VERTICAL-STEP, an option, is a positive integer |
| 21561 | that specifies how much a Y axis label increments |
| 21562 | for each line. For example, a step of 5 means |
| 21563 | that each line is five units of the graph." |
| 21564 | (let (Y-axis |
| 21565 | (number-per-line (or vertical-step 1))) |
| 21566 | @end group |
| 21567 | @group |
| 21568 | (while (> height 1) |
| 21569 | (if (zerop (% height Y-axis-label-spacing)) |
| 21570 | ;; @r{Insert label.} |
| 21571 | (setq Y-axis |
| 21572 | (cons |
| 21573 | (Y-axis-element |
| 21574 | (* height number-per-line) |
| 21575 | width-of-label) |
| 21576 | Y-axis)) |
| 21577 | @end group |
| 21578 | @group |
| 21579 | ;; @r{Else, insert blanks.} |
| 21580 | (setq Y-axis |
| 21581 | (cons |
| 21582 | (make-string width-of-label ? ) |
| 21583 | Y-axis))) |
| 21584 | (setq height (1- height))) |
| 21585 | @end group |
| 21586 | @group |
| 21587 | ;; @r{Insert base line.} |
| 21588 | (setq Y-axis (cons (Y-axis-element |
| 21589 | (or vertical-step 1) |
| 21590 | width-of-label) |
| 21591 | Y-axis)) |
| 21592 | (nreverse Y-axis))) |
| 21593 | @end group |
| 21594 | @end smallexample |
| 21595 | |
| 21596 | @smallexample |
| 21597 | @group |
| 21598 | (defun print-X-axis-numbered-line |
| 21599 | (number-of-X-tics X-axis-leading-spaces |
| 21600 | &optional horizontal-step) |
| 21601 | "Print line of X-axis numbers" |
| 21602 | (let ((number X-axis-label-spacing) |
| 21603 | (horizontal-step (or horizontal-step 1))) |
| 21604 | @end group |
| 21605 | @group |
| 21606 | (insert X-axis-leading-spaces) |
| 21607 | ;; line up number |
| 21608 | (delete-char (- (1- (length (number-to-string horizontal-step))))) |
| 21609 | (insert (concat |
| 21610 | (make-string |
| 21611 | ;; @r{Insert white space up to next number.} |
| 21612 | (- (* symbol-width X-axis-label-spacing) |
| 21613 | (1- (length (number-to-string horizontal-step))) |
| 21614 | 2) |
| 21615 | ? ) |
| 21616 | (number-to-string (* number horizontal-step)))) |
| 21617 | @end group |
| 21618 | @group |
| 21619 | ;; @r{Insert remaining numbers.} |
| 21620 | (setq number (+ number X-axis-label-spacing)) |
| 21621 | (while (> number-of-X-tics 1) |
| 21622 | (insert (X-axis-element (* number horizontal-step))) |
| 21623 | (setq number (+ number X-axis-label-spacing)) |
| 21624 | (setq number-of-X-tics (1- number-of-X-tics))))) |
| 21625 | @end group |
| 21626 | @end smallexample |
| 21627 | |
| 21628 | @smallexample |
| 21629 | @group |
| 21630 | (defun print-X-axis (numbers-list horizontal-step) |
| 21631 | "Print X axis labels to length of NUMBERS-LIST. |
| 21632 | Optionally, HORIZONTAL-STEP, a positive integer, |
| 21633 | specifies how much an X axis label increments for |
| 21634 | each column." |
| 21635 | @end group |
| 21636 | @group |
| 21637 | ;; Value of symbol-width and full-Y-label-width |
| 21638 | ;; are passed by `print-graph'. |
| 21639 | (let* ((leading-spaces |
| 21640 | (make-string full-Y-label-width ? )) |
| 21641 | ;; symbol-width @r{is provided by} graph-body-print |
| 21642 | (tic-width (* symbol-width X-axis-label-spacing)) |
| 21643 | (X-length (length numbers-list)) |
| 21644 | @end group |
| 21645 | @group |
| 21646 | (X-tic |
| 21647 | (concat |
| 21648 | (make-string |
| 21649 | ;; @r{Make a string of blanks.} |
| 21650 | (- (* symbol-width X-axis-label-spacing) |
| 21651 | (length X-axis-tic-symbol)) |
| 21652 | ? ) |
| 21653 | @end group |
| 21654 | @group |
| 21655 | ;; @r{Concatenate blanks with tic symbol.} |
| 21656 | X-axis-tic-symbol)) |
| 21657 | (tic-number |
| 21658 | (if (zerop (% X-length tic-width)) |
| 21659 | (/ X-length tic-width) |
| 21660 | (1+ (/ X-length tic-width))))) |
| 21661 | @end group |
| 21662 | |
| 21663 | @group |
| 21664 | (print-X-axis-tic-line |
| 21665 | tic-number leading-spaces X-tic) |
| 21666 | (insert "\n") |
| 21667 | (print-X-axis-numbered-line |
| 21668 | tic-number leading-spaces horizontal-step))) |
| 21669 | @end group |
| 21670 | @end smallexample |
| 21671 | |
| 21672 | @smallexample |
| 21673 | @group |
| 21674 | (defun one-fiftieth (full-range) |
| 21675 | "Return list, each number of which is 1/50th previous." |
| 21676 | (mapcar (lambda (arg) (/ arg 50)) full-range)) |
| 21677 | @end group |
| 21678 | @end smallexample |
| 21679 | |
| 21680 | @smallexample |
| 21681 | @group |
| 21682 | (defun print-graph |
| 21683 | (numbers-list &optional vertical-step horizontal-step) |
| 21684 | "Print labeled bar graph of the NUMBERS-LIST. |
| 21685 | The numbers-list consists of the Y-axis values. |
| 21686 | @end group |
| 21687 | |
| 21688 | @group |
| 21689 | Optionally, VERTICAL-STEP, a positive integer, |
| 21690 | specifies how much a Y axis label increments for |
| 21691 | each line. For example, a step of 5 means that |
| 21692 | each row is five units. |
| 21693 | @end group |
| 21694 | |
| 21695 | @group |
| 21696 | Optionally, HORIZONTAL-STEP, a positive integer, |
| 21697 | specifies how much an X axis label increments for |
| 21698 | each column." |
| 21699 | (let* ((symbol-width (length graph-blank)) |
| 21700 | ;; @code{height} @r{is both the largest number} |
| 21701 | ;; @r{and the number with the most digits.} |
| 21702 | (height (apply 'max numbers-list)) |
| 21703 | @end group |
| 21704 | @group |
| 21705 | (height-of-top-line |
| 21706 | (if (zerop (% height Y-axis-label-spacing)) |
| 21707 | height |
| 21708 | ;; @r{else} |
| 21709 | (* (1+ (/ height Y-axis-label-spacing)) |
| 21710 | Y-axis-label-spacing))) |
| 21711 | @end group |
| 21712 | @group |
| 21713 | (vertical-step (or vertical-step 1)) |
| 21714 | (full-Y-label-width |
| 21715 | (length |
| 21716 | (concat |
| 21717 | (number-to-string |
| 21718 | (* height-of-top-line vertical-step)) |
| 21719 | Y-axis-tic)))) |
| 21720 | @end group |
| 21721 | @group |
| 21722 | |
| 21723 | (print-Y-axis |
| 21724 | height-of-top-line full-Y-label-width vertical-step) |
| 21725 | (graph-body-print |
| 21726 | numbers-list height-of-top-line symbol-width) |
| 21727 | (print-X-axis numbers-list horizontal-step))) |
| 21728 | @end group |
| 21729 | @end smallexample |
| 21730 | @c qqq |
| 21731 | @end ignore |
| 21732 | |
| 21733 | @page |
| 21734 | @node Final printed graph |
| 21735 | @appendixsubsec The Printed Graph |
| 21736 | |
| 21737 | When made and installed, you can call the @code{print-graph} command |
| 21738 | like this: |
| 21739 | @sp 1 |
| 21740 | |
| 21741 | @smallexample |
| 21742 | @group |
| 21743 | (print-graph fiftieth-list-for-graph 50 10) |
| 21744 | @end group |
| 21745 | @end smallexample |
| 21746 | @sp 1 |
| 21747 | |
| 21748 | @noindent |
| 21749 | Here is the graph: |
| 21750 | @sp 2 |
| 21751 | |
| 21752 | @smallexample |
| 21753 | @group |
| 21754 | 1000 - * |
| 21755 | ** |
| 21756 | ** |
| 21757 | ** |
| 21758 | ** |
| 21759 | 750 - *** |
| 21760 | *** |
| 21761 | *** |
| 21762 | *** |
| 21763 | **** |
| 21764 | 500 - ***** |
| 21765 | ****** |
| 21766 | ****** |
| 21767 | ****** |
| 21768 | ******* |
| 21769 | 250 - ******** |
| 21770 | ********* * |
| 21771 | *********** * |
| 21772 | ************* * |
| 21773 | 50 - ***************** * * |
| 21774 | | | | | | | | | |
| 21775 | 10 50 100 150 200 250 300 350 |
| 21776 | @end group |
| 21777 | @end smallexample |
| 21778 | |
| 21779 | @sp 2 |
| 21780 | |
| 21781 | @noindent |
| 21782 | The largest group of functions contain 10--19 words and symbols each. |
| 21783 | |
| 21784 | @node Free Software and Free Manuals |
| 21785 | @appendix Free Software and Free Manuals |
| 21786 | |
| 21787 | @strong{by Richard M. Stallman} |
| 21788 | @sp 1 |
| 21789 | |
| 21790 | The biggest deficiency in free operating systems is not in the |
| 21791 | software---it is the lack of good free manuals that we can include in |
| 21792 | these systems. Many of our most important programs do not come with |
| 21793 | full manuals. Documentation is an essential part of any software |
| 21794 | package; when an important free software package does not come with a |
| 21795 | free manual, that is a major gap. We have many such gaps today. |
| 21796 | |
| 21797 | Once upon a time, many years ago, I thought I would learn Perl. I got |
| 21798 | a copy of a free manual, but I found it hard to read. When I asked |
| 21799 | Perl users about alternatives, they told me that there were better |
| 21800 | introductory manuals---but those were not free. |
| 21801 | |
| 21802 | Why was this? The authors of the good manuals had written them for |
| 21803 | O'Reilly Associates, which published them with restrictive terms---no |
| 21804 | copying, no modification, source files not available---which exclude |
| 21805 | them from the free software community. |
| 21806 | |
| 21807 | That wasn't the first time this sort of thing has happened, and (to |
| 21808 | our community's great loss) it was far from the last. Proprietary |
| 21809 | manual publishers have enticed a great many authors to restrict their |
| 21810 | manuals since then. Many times I have heard a GNU user eagerly tell me |
| 21811 | about a manual that he is writing, with which he expects to help the |
| 21812 | GNU project---and then had my hopes dashed, as he proceeded to explain |
| 21813 | that he had signed a contract with a publisher that would restrict it |
| 21814 | so that we cannot use it. |
| 21815 | |
| 21816 | Given that writing good English is a rare skill among programmers, we |
| 21817 | can ill afford to lose manuals this way. |
| 21818 | |
| 21819 | Free documentation, like free software, is a matter of freedom, not |
| 21820 | price. The problem with these manuals was not that O'Reilly Associates |
| 21821 | charged a price for printed copies---that in itself is fine. The Free |
| 21822 | Software Foundation @uref{http://shop.fsf.org, sells printed copies} of |
| 21823 | free @uref{http://www.gnu.org/doc/doc.html, GNU manuals}, too. |
| 21824 | But GNU manuals are available in source code form, while these manuals |
| 21825 | are available only on paper. GNU manuals come with permission to copy |
| 21826 | and modify; the Perl manuals do not. These restrictions are the |
| 21827 | problems. |
| 21828 | |
| 21829 | The criterion for a free manual is pretty much the same as for free |
| 21830 | software: it is a matter of giving all users certain |
| 21831 | freedoms. Redistribution (including commercial redistribution) must be |
| 21832 | permitted, so that the manual can accompany every copy of the program, |
| 21833 | on-line or on paper. Permission for modification is crucial too. |
| 21834 | |
| 21835 | As a general rule, I don't believe that it is essential for people to |
| 21836 | have permission to modify all sorts of articles and books. The issues |
| 21837 | for writings are not necessarily the same as those for software. For |
| 21838 | example, I don't think you or I are obliged to give permission to |
| 21839 | modify articles like this one, which describe our actions and our |
| 21840 | views. |
| 21841 | |
| 21842 | But there is a particular reason why the freedom to modify is crucial |
| 21843 | for documentation for free software. When people exercise their right |
| 21844 | to modify the software, and add or change its features, if they are |
| 21845 | conscientious they will change the manual too---so they can provide |
| 21846 | accurate and usable documentation with the modified program. A manual |
| 21847 | which forbids programmers to be conscientious and finish the job, or |
| 21848 | more precisely requires them to write a new manual from scratch if |
| 21849 | they change the program, does not fill our community's needs. |
| 21850 | |
| 21851 | While a blanket prohibition on modification is unacceptable, some |
| 21852 | kinds of limits on the method of modification pose no problem. For |
| 21853 | example, requirements to preserve the original author's copyright |
| 21854 | notice, the distribution terms, or the list of authors, are ok. It is |
| 21855 | also no problem to require modified versions to include notice that |
| 21856 | they were modified, even to have entire sections that may not be |
| 21857 | deleted or changed, as long as these sections deal with nontechnical |
| 21858 | topics. (Some GNU manuals have them.) |
| 21859 | |
| 21860 | These kinds of restrictions are not a problem because, as a practical |
| 21861 | matter, they don't stop the conscientious programmer from adapting the |
| 21862 | manual to fit the modified program. In other words, they don't block |
| 21863 | the free software community from making full use of the manual. |
| 21864 | |
| 21865 | However, it must be possible to modify all the technical content of |
| 21866 | the manual, and then distribute the result in all the usual media, |
| 21867 | through all the usual channels; otherwise, the restrictions do block |
| 21868 | the community, the manual is not free, and so we need another manual. |
| 21869 | |
| 21870 | Unfortunately, it is often hard to find someone to write another |
| 21871 | manual when a proprietary manual exists. The obstacle is that many |
| 21872 | users think that a proprietary manual is good enough---so they don't |
| 21873 | see the need to write a free manual. They do not see that the free |
| 21874 | operating system has a gap that needs filling. |
| 21875 | |
| 21876 | Why do users think that proprietary manuals are good enough? Some have |
| 21877 | not considered the issue. I hope this article will do something to |
| 21878 | change that. |
| 21879 | |
| 21880 | Other users consider proprietary manuals acceptable for the same |
| 21881 | reason so many people consider proprietary software acceptable: they |
| 21882 | judge in purely practical terms, not using freedom as a |
| 21883 | criterion. These people are entitled to their opinions, but since |
| 21884 | those opinions spring from values which do not include freedom, they |
| 21885 | are no guide for those of us who do value freedom. |
| 21886 | |
| 21887 | Please spread the word about this issue. We continue to lose manuals |
| 21888 | to proprietary publishing. If we spread the word that proprietary |
| 21889 | manuals are not sufficient, perhaps the next person who wants to help |
| 21890 | GNU by writing documentation will realize, before it is too late, that |
| 21891 | he must above all make it free. |
| 21892 | |
| 21893 | We can also encourage commercial publishers to sell free, copylefted |
| 21894 | manuals instead of proprietary ones. One way you can help this is to |
| 21895 | check the distribution terms of a manual before you buy it, and prefer |
| 21896 | copylefted manuals to non-copylefted ones. |
| 21897 | |
| 21898 | @sp 2 |
| 21899 | @noindent |
| 21900 | Note: The Free Software Foundation maintains a page on its Web site |
| 21901 | that lists free books available from other publishers:@* |
| 21902 | @uref{http://www.gnu.org/doc/other-free-books.html} |
| 21903 | |
| 21904 | @node GNU Free Documentation License |
| 21905 | @appendix GNU Free Documentation License |
| 21906 | |
| 21907 | @cindex FDL, GNU Free Documentation License |
| 21908 | @include doclicense.texi |
| 21909 | |
| 21910 | @node Index |
| 21911 | @unnumbered Index |
| 21912 | |
| 21913 | @ignore |
| 21914 | MENU ENTRY: NODE NAME. |
| 21915 | @end ignore |
| 21916 | |
| 21917 | @printindex cp |
| 21918 | |
| 21919 | @iftex |
| 21920 | @c Place biographical information on right-hand (verso) page |
| 21921 | |
| 21922 | @tex |
| 21923 | \par\vfill\supereject |
| 21924 | \ifodd\pageno |
| 21925 | \global\evenheadline={\hfil} \global\evenfootline={\hfil} |
| 21926 | \global\oddheadline={\hfil} \global\oddfootline={\hfil} |
| 21927 | %\page\hbox{}\page |
| 21928 | \else |
| 21929 | % \par\vfill\supereject |
| 21930 | \global\evenheadline={\hfil} \global\evenfootline={\hfil} |
| 21931 | \global\oddheadline={\hfil} \global\oddfootline={\hfil} |
| 21932 | %\page\hbox{}%\page |
| 21933 | %\page\hbox{}%\page |
| 21934 | \fi |
| 21935 | @end tex |
| 21936 | |
| 21937 | @c page |
| 21938 | @w{ } |
| 21939 | |
| 21940 | @c ================ Biographical information ================ |
| 21941 | |
| 21942 | @w{ } |
| 21943 | @sp 8 |
| 21944 | @center About the Author |
| 21945 | @sp 1 |
| 21946 | @end iftex |
| 21947 | |
| 21948 | @ifnottex |
| 21949 | @node About the Author |
| 21950 | @unnumbered About the Author |
| 21951 | @end ifnottex |
| 21952 | |
| 21953 | @quotation |
| 21954 | Robert J. Chassell has worked with GNU Emacs since 1985. He writes |
| 21955 | and edits, teaches Emacs and Emacs Lisp, and speaks throughout the |
| 21956 | world on software freedom. Chassell was a founding Director and |
| 21957 | Treasurer of the Free Software Foundation, Inc. He is co-author of |
| 21958 | the @cite{Texinfo} manual, and has edited more than a dozen other |
| 21959 | books. He graduated from Cambridge University, in England. He has an |
| 21960 | abiding interest in social and economic history and flies his own |
| 21961 | airplane. |
| 21962 | @end quotation |
| 21963 | |
| 21964 | @c @page |
| 21965 | @c @w{ } |
| 21966 | @c |
| 21967 | @c @c Prevent page number on blank verso, so eject it first. |
| 21968 | @c @tex |
| 21969 | @c \par\vfill\supereject |
| 21970 | @c @end tex |
| 21971 | |
| 21972 | @c @iftex |
| 21973 | @c @headings off |
| 21974 | @c @evenheading @thispage @| @| @thistitle |
| 21975 | @c @oddheading @| @| @thispage |
| 21976 | @c @end iftex |
| 21977 | |
| 21978 | @bye |