Commit | Line | Data |
---|---|---|
8b096dce EZ |
1 | \input texinfo @c -*-texinfo-*- |
2 | @comment %**start of header | |
3 | @setfilename ../info/eintr | |
4 | @c sethtmlfilename emacs-lisp-intro.html | |
5 | @settitle Programming in Emacs Lisp | |
6 | @syncodeindex vr cp | |
7 | @syncodeindex fn cp | |
8 | @setchapternewpage odd | |
9 | @finalout | |
10 | ||
11 | @c --------- | |
475dc40a EZ |
12 | @c <<<< For hard copy printing, this file is now |
13 | @c set for smallbook, which works for all sizes | |
14 | @c of paper, and with Postscript figures >>>> | |
15 | @smallbook | |
16 | @clear largebook | |
8b096dce | 17 | @set print-postscript-figures |
475dc40a EZ |
18 | @c set largebook |
19 | @c clear print-postscript-figures | |
8b096dce EZ |
20 | @c --------- |
21 | ||
22 | @comment %**end of header | |
23 | ||
475dc40a EZ |
24 | @set edition-number 2.02 |
25 | @set update-date 2001 Nov 25 | |
8b096dce EZ |
26 | |
27 | @ignore | |
28 | ## Summary of shell commands to create various output formats: | |
29 | ||
30 | ## Info output | |
31 | makeinfo --no-split --paragraph-indent=0 --verbose emacs-lisp-intro.texi | |
32 | ||
33 | ## DVI output | |
34 | texi2dvi emacs-lisp-intro.texi | |
35 | ||
36 | ## HTML output | |
8b096dce EZ |
37 | makeinfo --html --no-split --verbose emacs-lisp-intro.texi |
38 | ||
39 | ## Plain text output | |
40 | makeinfo --fill-column=70 --no-split --paragraph-indent=0 \ | |
41 | --verbose --no-headers --output=emacs-lisp-intro.txt emacs-lisp-intro.texi | |
42 | ||
43 | @end ignore | |
44 | ||
45 | @c ================ Included Figures ================ | |
46 | ||
47 | @c Set print-postscript-figures if you print PostScript figures. | |
48 | @c If you clear this, the ten figures will be printed as ASCII diagrams. | |
49 | @c (This is not relevant to Info, since Info only handles ASCII.) | |
50 | @c Your site may require editing changes to print PostScript; in this | |
51 | @c case, search for `print-postscript-figures' and make appropriate changes. | |
52 | ||
53 | ||
54 | @c ================ How to Create an Info file ================ | |
55 | ||
56 | @c If you have `makeinfo' installed, run the following command | |
57 | ||
58 | @c makeinfo emacs-lisp-intro.texi | |
59 | ||
60 | @c or, if you want a single, large Info file, and no paragraph indents: | |
61 | @c makeinfo --no-split --paragraph-indent=0 --verbose emacs-lisp-intro.texi | |
62 | ||
63 | @c After creating the Info file, edit your Info `dir' file, if the | |
0860ed42 | 64 | @c `dircategory' section below does not enable your system to |
8b096dce EZ |
65 | @c install the manual automatically. |
66 | @c (The `dir' file is often in the `/usr/local/info/' directory.) | |
67 | ||
68 | @c ================ How to Create an HTML file ================ | |
69 | ||
70 | @c To convert to HTML format | |
71 | @c makeinfo --html --no-split --verbose emacs-lisp-intro.texi | |
72 | ||
73 | @c ================ How to Print a Book in Various Sizes ================ | |
74 | ||
75 | @c This book can be printed in any of three different sizes. | |
76 | @c In the above header, set @-commands appropriately. | |
77 | ||
78 | @c 7 by 9.25 inches: | |
79 | @c @smallbook | |
80 | @c @clear largebook | |
81 | ||
82 | @c 8.5 by 11 inches: | |
83 | @c @c smallbook | |
84 | @c @set largebook | |
85 | ||
86 | @c European A4 size paper: | |
87 | @c @c smallbook | |
88 | @c @afourpaper | |
89 | @c @set largebook | |
90 | ||
91 | @c ================ How to Typeset and Print ================ | |
92 | ||
93 | @c If you do not include PostScript figures, run either of the | |
94 | @c following command sequences, or similar commands suited to your | |
95 | @c system: | |
96 | ||
97 | @c texi2dvi emacs-lisp-intro.texi | |
98 | @c lpr -d emacs-lisp-intro.dvi | |
99 | ||
100 | @c or else: | |
101 | ||
102 | @c tex emacs-lisp-intro.texi | |
103 | @c texindex emacs-lisp-intro.?? | |
104 | @c tex emacs-lisp-intro.texi | |
105 | @c lpr -d emacs-lisp-intro.dvi | |
106 | ||
475dc40a EZ |
107 | @c If you include the PostScript figures, and you have old software, |
108 | @c you may need to convert the .dvi file to a .ps file before | |
109 | @c printing. Run either of the following command sequences, or one | |
110 | @c similar: | |
8b096dce EZ |
111 | @c |
112 | @c dvips -f < emacs-lisp-intro.dvi > emacs-lisp-intro.ps | |
113 | @c | |
114 | @c or else: | |
115 | @c | |
116 | @c postscript -p < emacs-lisp-intro.dvi > emacs-lisp-intro.ps | |
117 | @c | |
118 | ||
119 | @c (Note: if you edit the book so as to change the length of the | |
120 | @c table of contents, you may have to change the value of `pageno' below.) | |
121 | ||
8b096dce EZ |
122 | @c ================ End of Formatting Sections ================ |
123 | ||
124 | @c For next or subsequent edition: | |
125 | @c create function using with-output-to-temp-buffer | |
126 | @c create a major mode, with keymaps | |
127 | @c run an asynchronous process, like grep or diff | |
128 | ||
129 | @c For smallbook format, use smaller than normal amounts of | |
130 | @c whitespace between chapters, sections, and paragraphs. | |
131 | @tex | |
132 | \global\chapheadingskip = 15pt plus 4pt minus 2pt | |
133 | \global\secheadingskip = 12pt plus 3pt minus 2pt | |
134 | \global\subsecheadingskip = 9pt plus 2pt minus 2pt \global\parskip 2pt | |
135 | plus 1pt | |
136 | @end tex | |
137 | ||
138 | @c For 8.5 by 11 inch format: do not use such a small amount of | |
139 | @c whitespace between paragraphs as above: | |
140 | @ifset largebook | |
141 | @tex | |
142 | \global\parskip 6pt plus 1pt | |
143 | @end tex | |
144 | @end ifset | |
145 | ||
146 | @c For all sized formats: print within-book cross | |
147 | @c reference with ``...'' rather than [...] | |
148 | @tex | |
149 | % Need following so comma appears after section numbers. | |
150 | \global\def\Ysectionnumberandtype{% | |
151 | \ifnum\secno=0 \putwordChapter\xreftie\the\chapno, \space % | |
152 | \else \ifnum \subsecno=0 \putwordSection\xreftie\the\chapno.\the\secno, \space % | |
153 | \else \ifnum \subsubsecno=0 % | |
154 | \putwordSection\xreftie\the\chapno.\the\secno.\the\subsecno, \space % | |
155 | \else % | |
156 | \putwordSection\xreftie\the\chapno.\the\secno.\the\subsecno.\the\subsubsecno, \space% | |
157 | \fi \fi \fi } | |
158 | ||
159 | \global\def\Yappendixletterandtype{% | |
160 | \ifnum\secno=0 \putwordAppendix\xreftie'char\the\appendixno{}, \space% | |
161 | \else \ifnum \subsecno=0 \putwordSection\xreftie'char\the\appendixno.\the\secno, \space % | |
162 | \else \ifnum \subsubsecno=0 % | |
163 | \putwordSection\xreftie'char\the\appendixno.\the\secno.\the\subsecno, \space % | |
164 | \else % | |
165 | \putwordSection\xreftie'char\the\appendixno.\the\secno.\the\subsecno.\the\subsubsecno, \space % | |
166 | \fi \fi \fi } | |
167 | ||
168 | \global\def\xrefX[#1,#2,#3,#4,#5,#6]{\begingroup | |
169 | \def\printedmanual{\ignorespaces #5}% | |
170 | \def\printednodename{\ignorespaces #3}% | |
171 | \setbox1=\hbox{\printedmanual}% | |
172 | \setbox0=\hbox{\printednodename}% | |
173 | \ifdim \wd0 = 0pt | |
174 | % No printed node name was explicitly given. | |
175 | \ifx\SETxref-automatic-section-title\relax % | |
176 | % Use the actual chapter/section title appear inside | |
177 | % the square brackets. Use the real section title if we have it. | |
178 | \ifdim \wd1>0pt% | |
179 | % It is in another manual, so we don't have it. | |
180 | \def\printednodename{\ignorespaces #1}% | |
181 | \else | |
182 | \ifhavexrefs | |
183 | % We know the real title if we have the xref values. | |
184 | \def\printednodename{\refx{#1-title}}% | |
185 | \else | |
186 | % Otherwise just copy the Info node name. | |
187 | \def\printednodename{\ignorespaces #1}% | |
188 | \fi% | |
189 | \fi | |
190 | \def\printednodename{#1-title}% | |
191 | \else | |
192 | % Use the node name inside the square brackets. | |
193 | \def\printednodename{\ignorespaces #1}% | |
194 | \fi | |
195 | \fi | |
196 | % | |
197 | % If we use \unhbox0 and \unhbox1 to print the node names, TeX does not | |
198 | % insert empty discretionaries after hyphens, which means that it will | |
199 | % not find a line break at a hyphen in a node names. Since some manuals | |
200 | % are best written with fairly long node names, containing hyphens, this | |
201 | % is a loss. Therefore, we give the text of the node name again, so it | |
202 | % is as if TeX is seeing it for the first time. | |
203 | \ifdim \wd1 > 0pt | |
204 | \putwordsection{} ``\printednodename'' in \cite{\printedmanual}% | |
205 | \else | |
206 | % _ (for example) has to be the character _ for the purposes of the | |
207 | % control sequence corresponding to the node, but it has to expand | |
208 | % into the usual \leavevmode...\vrule stuff for purposes of | |
209 | % printing. So we \turnoffactive for the \refx-snt, back on for the | |
210 | % printing, back off for the \refx-pg. | |
211 | {\turnoffactive \refx{#1-snt}{}}% | |
212 | % \space [\printednodename],\space % <= original | |
213 | % \putwordsection{} ``\printednodename'',\space | |
214 | ``\printednodename'',\space | |
215 | \turnoffactive \putwordpage\tie\refx{#1-pg}{}% | |
216 | \fi | |
217 | \endgroup} | |
218 | @end tex | |
219 | ||
220 | @c ---------------------------------------------------- | |
221 | ||
d586ab6c EZ |
222 | @dircategory Emacs |
223 | @direntry | |
224 | * Emacs Lisp Intro: (eintr). | |
8b096dce | 225 | A simple introduction to Emacs Lisp programming. |
d586ab6c | 226 | @end direntry |
8b096dce EZ |
227 | |
228 | @ifinfo | |
229 | This is an introduction to @cite{Programming in Emacs Lisp}, for | |
230 | people who are not programmers. | |
231 | ||
232 | Edition @value{edition-number}, @value{update-date} | |
233 | ||
234 | Copyright (C) 1990, '91, '92, '93, '94, '95, '97, 2001 Free Software Foundation, Inc. | |
235 | ||
236 | Permission is granted to copy, distribute and/or modify this document | |
237 | under the terms of the GNU Free Documentation License, Version 1.1 or | |
238 | any later version published by the Free Software Foundation; with the | |
239 | Invariant Section being the Preface, with the Front-Cover Texts being | |
240 | no Front-Cover Texts, and with the Back-Cover Texts being no | |
241 | Back-Cover Texts. A copy of the license is included in the section | |
242 | entitled ``GNU Free Documentation License''. | |
243 | @end ifinfo | |
244 | ||
245 | @c half title; two lines here, so do not use `shorttitlepage' | |
246 | @tex | |
247 | {\begingroup% | |
248 | \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}% | |
249 | \endgroup}% | |
250 | {\begingroup\hbox{}\vskip 0.25in \chaprm% | |
251 | \centerline{Programming in Emacs Lisp}% | |
252 | \endgroup\page\hbox{}\page} | |
253 | @end tex | |
254 | ||
255 | @titlepage | |
256 | @sp 6 | |
257 | @center @titlefont{An Introduction to} | |
258 | @sp 2 | |
259 | @center @titlefont{Programming in Emacs Lisp} | |
260 | @sp 2 | |
261 | @center Second Edition | |
262 | @sp 4 | |
263 | @center by Robert J. Chassell | |
264 | ||
265 | @page | |
266 | @vskip 0pt plus 1filll | |
267 | Copyright @copyright{} 1990, '91, '92, '93, '94, '95, '97, 2001 Free Software Foundation, Inc. | |
268 | @sp 2 | |
269 | ||
270 | Published by the Free Software Foundation, Inc.@* | |
271 | 59 Temple Place, Suite 330@* | |
272 | Boston, MA 02111-1307 USA@* | |
273 | ||
274 | Edition @value{edition-number}, @value{update-date} | |
275 | ||
276 | @c Printed copies are available for $20 each.@* | |
277 | ISBN-1882114-41-8 | |
278 | ||
279 | Permission is granted to copy, distribute and/or modify this document | |
280 | under the terms of the GNU Free Documentation License, Version 1.1 or | |
281 | any later version published by the Free Software Foundation; with the | |
282 | Invariant Section being the Preface, with the Front-Cover Texts being | |
283 | no Front-Cover Texts, and with the Back-Cover Texts being no | |
284 | Back-Cover Texts. A copy of the license is included in the section | |
285 | entitled ``GNU Free Documentation License''. | |
286 | @end titlepage | |
287 | ||
288 | @iftex | |
289 | @headings off | |
290 | @evenheading @thispage @| @| @thischapter | |
291 | @oddheading @thissection @| @| @thispage | |
292 | @end iftex | |
293 | ||
475dc40a | 294 | @ifnothtml |
8b096dce | 295 | |
475dc40a EZ |
296 | @c Keep T.O.C. short by tightening up. |
297 | @ifset largebook | |
298 | @tex | |
299 | \global\parskip 2pt plus 1pt | |
300 | \global\advance\baselineskip by -1pt | |
301 | @end tex | |
302 | @end ifset | |
8b096dce | 303 | |
475dc40a EZ |
304 | @shortcontents |
305 | @contents | |
306 | ||
307 | @ifset largebook | |
308 | @tex | |
309 | \global\parskip 6pt plus 1pt | |
310 | \global\advance\baselineskip by 1pt | |
311 | @end tex | |
312 | @end ifset | |
313 | ||
314 | @end ifnothtml | |
8b096dce EZ |
315 | |
316 | @c >>>> Set pageno appropriately <<<< | |
317 | ||
318 | @c The first page of the Preface is a roman numeral; it is the first | |
319 | @c right handed page after the Table of Contents; hence the following | |
320 | @c setting must be for an odd negative number. | |
321 | ||
322 | @c if largebook, there are 8 pages in Table of Contents | |
323 | @ifset largebook | |
324 | @iftex | |
325 | @pageno = -9 | |
326 | @end iftex | |
327 | @end ifset | |
328 | ||
329 | @c if smallbook, there are 10 pages in Table of Contents | |
330 | @ifclear largebook | |
331 | @iftex | |
332 | @pageno = -11 | |
333 | @end iftex | |
334 | @end ifclear | |
335 | ||
475dc40a EZ |
336 | @ifnottex |
337 | @node Top, Preface, (dir), (dir) | |
338 | @top An Introduction to Programming in Emacs Lisp | |
339 | ||
340 | This is an introduction to @cite{Programming in Emacs Lisp}, for | |
341 | people who are not programmers. | |
342 | ||
343 | This master menu first lists each chapter and index; then it lists | |
344 | every node in every chapter. | |
345 | @end ifnottex | |
346 | ||
8b096dce EZ |
347 | @menu |
348 | * Preface:: What to look for. | |
349 | * List Processing:: What is Lisp? | |
350 | * Practicing Evaluation:: Running several programs. | |
351 | * Writing Defuns:: How to write function definitions. | |
352 | * Buffer Walk Through:: Exploring a few buffer-related functions. | |
353 | * More Complex:: A few, even more complex functions. | |
354 | * Narrowing & Widening:: Restricting your and Emacs attention to | |
355 | a region. | |
356 | * car cdr & cons:: Fundamental functions in Lisp. | |
357 | * Cutting & Storing Text:: Removing text and saving it. | |
358 | * List Implementation:: How lists are implemented in the computer. | |
359 | * Yanking:: Pasting stored text. | |
360 | * Loops & Recursion:: How to repeat a process. | |
361 | * Regexp Search:: Regular expression searches. | |
362 | * Counting Words:: A review of repetition and regexps. | |
363 | * Words in a defun:: Counting words in a @code{defun}. | |
364 | * Readying a Graph:: A prototype graph printing function. | |
365 | * Emacs Initialization:: How to write a @file{.emacs} file. | |
366 | * Debugging:: How to run the Emacs Lisp debuggers. | |
367 | * Conclusion:: Now you have the basics. | |
368 | * the-the:: An appendix: how to find reduplicated words. | |
369 | * Kill Ring:: An appendix: how the kill ring works. | |
370 | * Full Graph:: How to create a graph with labelled axes. | |
371 | * GNU Free Documentation License:: | |
372 | * Index:: | |
373 | * About the Author:: | |
374 | ||
375 | @detailmenu | |
376 | --- The Detailed Node Listing --- | |
377 | ||
378 | Preface | |
379 | ||
380 | * Why:: Why learn Emacs Lisp? | |
381 | * On Reading this Text:: Read, gain familiarity, pick up habits.... | |
382 | * Who You Are:: For whom this is written. | |
383 | * Lisp History:: | |
384 | * Note for Novices:: You can read this as a novice. | |
385 | * Thank You:: | |
386 | ||
387 | List Processing | |
388 | ||
389 | * Lisp Lists:: What are lists? | |
390 | * Run a Program:: Any list in Lisp is a program ready to run. | |
391 | * Making Errors:: Generating an error message. | |
392 | * Names & Definitions:: Names of symbols and function definitions. | |
393 | * Lisp Interpreter:: What the Lisp interpreter does. | |
394 | * Evaluation:: Running a program. | |
395 | * Variables:: Returning a value from a variable. | |
396 | * Arguments:: Passing information to a function. | |
397 | * set & setq:: Setting the value of a variable. | |
398 | * Summary:: The major points. | |
399 | * Error Message Exercises:: | |
400 | ||
401 | Lisp Lists | |
402 | ||
403 | * Numbers Lists:: List have numbers, other lists, in them. | |
404 | * Lisp Atoms:: Elemental entities. | |
405 | * Whitespace in Lists:: Formating lists to be readable. | |
406 | * Typing Lists:: How GNU Emacs helps you type lists. | |
407 | ||
408 | The Lisp Interpreter | |
409 | ||
410 | * Complications:: Variables, Special forms, Lists within. | |
411 | * Byte Compiling:: Specially processing code for speed. | |
412 | ||
413 | Evaluation | |
414 | ||
415 | * Evaluating Inner Lists:: Lists within lists... | |
416 | ||
417 | Variables | |
418 | ||
419 | * fill-column Example:: | |
420 | * Void Function:: The error message for a symbol | |
421 | without a function. | |
422 | * Void Variable:: The error message for a symbol without a value. | |
423 | ||
424 | Arguments | |
425 | ||
426 | * Data types:: Types of data passed to a function. | |
427 | * Args as Variable or List:: An argument can be the value | |
428 | of a variable or list. | |
429 | * Variable Number of Arguments:: Some functions may take a | |
430 | variable number of arguments. | |
431 | * Wrong Type of Argument:: Passing an argument of the wrong type | |
432 | to a function. | |
433 | * message:: A useful function for sending messages. | |
434 | ||
435 | Setting the Value of a Variable | |
436 | ||
437 | * Using set:: Setting values. | |
438 | * Using setq:: Setting a quoted value. | |
439 | * Counting:: Using @code{setq} to count. | |
440 | ||
441 | Practicing Evaluation | |
442 | ||
443 | * How to Evaluate:: Typing editing commands or @kbd{C-x C-e} | |
444 | causes evaluation. | |
445 | * Buffer Names:: Buffers and files are different. | |
446 | * Getting Buffers:: Getting a buffer itself, not merely its name. | |
447 | * Switching Buffers:: How to change to another buffer. | |
448 | * Buffer Size & Locations:: Where point is located and the size of | |
449 | the buffer. | |
450 | * Evaluation Exercise:: | |
451 | ||
452 | How To Write Function Definitions | |
453 | ||
454 | * Primitive Functions:: | |
455 | * defun:: The @code{defun} special form. | |
456 | * Install:: Install a function definition. | |
457 | * Interactive:: Making a function interactive. | |
458 | * Interactive Options:: Different options for @code{interactive}. | |
459 | * Permanent Installation:: Installing code permanently. | |
460 | * let:: Creating and initializing local variables. | |
461 | * if:: What if? | |
462 | * else:: If--then--else expressions. | |
463 | * Truth & Falsehood:: What Lisp considers false and true. | |
464 | * save-excursion:: Keeping track of point, mark, and buffer. | |
465 | * Review:: | |
466 | * defun Exercises:: | |
467 | ||
468 | Install a Function Definition | |
469 | ||
470 | * Effect of installation:: | |
471 | * Change a defun:: How to change a function definition. | |
472 | ||
473 | Make a Function Interactive | |
474 | ||
475 | * Interactive multiply-by-seven:: An overview. | |
476 | * multiply-by-seven in detail:: The interactive version. | |
477 | ||
478 | @code{let} | |
479 | ||
480 | * Prevent confusion:: | |
481 | * Parts of let Expression:: | |
482 | * Sample let Expression:: | |
483 | * Uninitialized let Variables:: | |
484 | ||
485 | The @code{if} Special Form | |
486 | ||
487 | * if in more detail:: | |
488 | * type-of-animal in detail:: An example of an @code{if} expression. | |
489 | ||
490 | Truth and Falsehood in Emacs Lisp | |
491 | ||
492 | * nil explained:: @code{nil} has two meanings. | |
493 | ||
494 | @code{save-excursion} | |
495 | ||
496 | * Point and mark:: A review of various locations. | |
497 | * Template for save-excursion:: | |
498 | ||
499 | A Few Buffer--Related Functions | |
500 | ||
501 | * Finding More:: How to find more information. | |
502 | * simplified-beginning-of-buffer:: Shows @code{goto-char}, | |
503 | @code{point-min}, and @code{push-mark}. | |
504 | * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}. | |
505 | * append-to-buffer:: Uses @code{save-excursion} and | |
506 | @code{insert-buffer-substring}. | |
507 | * Buffer Related Review:: Review. | |
508 | * Buffer Exercises:: | |
509 | ||
510 | The Definition of @code{mark-whole-buffer} | |
511 | ||
512 | * mark-whole-buffer overview:: | |
513 | * Body of mark-whole-buffer:: Only three lines of code. | |
514 | ||
515 | The Definition of @code{append-to-buffer} | |
516 | ||
517 | * append-to-buffer overview:: | |
518 | * append interactive:: A two part interactive expression. | |
519 | * append-to-buffer body:: Incorporates a @code{let} expression. | |
520 | * append save-excursion:: How the @code{save-excursion} works. | |
521 | ||
522 | A Few More Complex Functions | |
523 | ||
524 | * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}. | |
525 | * insert-buffer:: Read-only, and with @code{or}. | |
526 | * beginning-of-buffer:: Shows @code{goto-char}, | |
527 | @code{point-min}, and @code{push-mark}. | |
528 | * Second Buffer Related Review:: | |
529 | * optional Exercise:: | |
530 | ||
531 | The Definition of @code{insert-buffer} | |
532 | ||
533 | * insert-buffer code:: | |
534 | * insert-buffer interactive:: When you can read, but not write. | |
535 | * insert-buffer body:: The body has an @code{or} and a @code{let}. | |
536 | * if & or:: Using an @code{if} instead of an @code{or}. | |
537 | * Insert or:: How the @code{or} expression works. | |
538 | * Insert let:: Two @code{save-excursion} expressions. | |
539 | ||
540 | The Interactive Expression in @code{insert-buffer} | |
541 | ||
542 | * Read-only buffer:: When a buffer cannot be modified. | |
543 | * b for interactive:: An existing buffer or else its name. | |
544 | ||
545 | Complete Definition of @code{beginning-of-buffer} | |
546 | ||
547 | * Optional Arguments:: | |
548 | * beginning-of-buffer opt arg:: Example with optional argument. | |
549 | * beginning-of-buffer complete:: | |
550 | ||
551 | @code{beginning-of-buffer} with an Argument | |
552 | ||
553 | * Disentangle beginning-of-buffer:: | |
554 | * Large buffer case:: | |
555 | * Small buffer case:: | |
556 | ||
557 | Narrowing and Widening | |
558 | ||
559 | * Narrowing advantages:: The advantages of narrowing | |
560 | * save-restriction:: The @code{save-restriction} special form. | |
561 | * what-line:: The number of the line that point is on. | |
562 | * narrow Exercise:: | |
563 | ||
564 | @code{car}, @code{cdr}, @code{cons}: Fundamental Functions | |
565 | ||
566 | * Strange Names:: An historical aside: why the strange names? | |
567 | * car & cdr:: Functions for extracting part of a list. | |
568 | * cons:: Constructing a list. | |
569 | * nthcdr:: Calling @code{cdr} repeatedly. | |
570 | * nth:: | |
571 | * setcar:: Changing the first element of a list. | |
572 | * setcdr:: Changing the rest of a list. | |
573 | * cons Exercise:: | |
574 | ||
575 | @code{cons} | |
576 | ||
577 | * Build a list:: | |
578 | * length:: How to find the length of a list. | |
579 | ||
580 | Cutting and Storing Text | |
581 | ||
582 | * Storing Text:: Text is stored in a list. | |
583 | * zap-to-char:: Cutting out text up to a character. | |
584 | * kill-region:: Cutting text out of a region. | |
585 | * Digression into C:: Minor note on C programming language macros. | |
586 | * defvar:: How to give a variable an initial value. | |
587 | * copy-region-as-kill:: A definition for copying text. | |
588 | * cons & search-fwd Review:: | |
589 | * search Exercises:: | |
590 | ||
591 | @code{zap-to-char} | |
592 | ||
593 | * Complete zap-to-char:: The complete implementation. | |
594 | * zap-to-char interactive:: A three part interactive expression. | |
595 | * zap-to-char body:: A short overview. | |
596 | * search-forward:: How to search for a string. | |
597 | * progn:: The @code{progn} special form. | |
598 | * Summing up zap-to-char:: Using @code{point} and @code{search-forward}. | |
599 | ||
600 | @code{kill-region} | |
601 | ||
602 | * Complete kill-region:: The function definition. | |
603 | * condition-case:: Dealing with a problem. | |
604 | * delete-and-extract-region:: Doing the work. | |
605 | ||
606 | Initializing a Variable with @code{defvar} | |
607 | ||
608 | * See variable current value:: | |
609 | * defvar and asterisk:: An old-time convention. | |
610 | ||
611 | @code{copy-region-as-kill} | |
612 | ||
613 | * Complete copy-region-as-kill:: The complete function definition. | |
614 | * copy-region-as-kill body:: The body of @code{copy-region-as-kill}. | |
615 | ||
616 | The Body of @code{copy-region-as-kill} | |
617 | ||
618 | * last-command & this-command:: | |
619 | * kill-append function:: | |
620 | * kill-new function:: | |
621 | ||
622 | How Lists are Implemented | |
623 | ||
624 | * Lists diagrammed:: | |
625 | * Symbols as Chest:: Exploring a powerful metaphor. | |
626 | * List Exercise:: | |
627 | ||
628 | Yanking Text Back | |
629 | ||
630 | * Kill Ring Overview:: The kill ring is a list. | |
631 | * kill-ring-yank-pointer:: The @code{kill-ring-yank-pointer} variable. | |
632 | * yank nthcdr Exercises:: | |
633 | ||
634 | Loops and Recursion | |
635 | ||
636 | * while:: Causing a stretch of code to repeat. | |
637 | * dolist dotimes:: | |
638 | * Recursion:: Causing a function to call itself. | |
639 | * Looping exercise:: | |
640 | ||
641 | @code{while} | |
642 | ||
643 | * Looping with while:: Repeat so long as test returns true. | |
644 | * Loop Example:: A @code{while} loop that uses a list. | |
645 | * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}. | |
646 | * Incrementing Loop:: A loop with an incrementing counter. | |
647 | * Decrementing Loop:: A loop with a decrementing counter. | |
648 | ||
649 | A Loop with an Incrementing Counter | |
650 | ||
651 | * Incrementing Example:: Counting pebbles in a triangle. | |
652 | * Inc Example parts:: The parts of the function definition. | |
653 | * Inc Example altogether:: Putting the function definition together. | |
654 | ||
655 | Loop with a Decrementing Counter | |
656 | ||
657 | * Decrementing Example:: More pebbles on the beach. | |
658 | * Dec Example parts:: The parts of the function definition. | |
659 | * Dec Example altogether:: Putting the function definition together. | |
660 | ||
661 | Save your time: @code{dolist} and @code{dotimes} | |
662 | ||
663 | * dolist:: | |
664 | * dotimes:: | |
665 | ||
666 | Recursion | |
667 | ||
668 | * Building Robots:: Same model, different serial number ... | |
669 | * Recursive Definition Parts:: Walk until you stop ... | |
670 | * Recursion with list:: Using a list as the test whether to recurse. | |
671 | * Recursive triangle function:: | |
672 | * Recursion with cond:: | |
673 | * Recursive Patterns:: Often used templates. | |
674 | * No Deferment:: Don't store up work ... | |
675 | * No deferment solution:: | |
676 | ||
677 | Recursion in Place of a Counter | |
678 | ||
679 | * Recursive Example arg of 1 or 2:: | |
680 | * Recursive Example arg of 3 or 4:: | |
681 | ||
682 | Recursive Patterns | |
683 | ||
684 | * Every:: | |
685 | * Accumulate:: | |
686 | * Keep:: | |
687 | ||
688 | Regular Expression Searches | |
689 | ||
690 | * sentence-end:: The regular expression for @code{sentence-end}. | |
691 | * re-search-forward:: Very similar to @code{search-forward}. | |
692 | * forward-sentence:: A straightforward example of regexp search. | |
693 | * forward-paragraph:: A somewhat complex example. | |
694 | * etags:: How to create your own @file{TAGS} table. | |
695 | * Regexp Review:: | |
696 | * re-search Exercises:: | |
697 | ||
698 | @code{forward-sentence} | |
699 | ||
700 | * Complete forward-sentence:: | |
701 | * fwd-sentence while loops:: Two @code{while} loops. | |
702 | * fwd-sentence re-search:: A regular expression search. | |
703 | ||
704 | @code{forward-paragraph}: a Goldmine of Functions | |
705 | ||
706 | * forward-paragraph in brief:: Key parts of the function definition. | |
707 | * fwd-para let:: The @code{let*} expression. | |
708 | * fwd-para while:: The forward motion @code{while} loop. | |
709 | * fwd-para between paragraphs:: Movement between paragraphs. | |
710 | * fwd-para within paragraph:: Movement within paragraphs. | |
711 | * fwd-para no fill prefix:: When there is no fill prefix. | |
712 | * fwd-para with fill prefix:: When there is a fill prefix. | |
713 | * fwd-para summary:: Summary of @code{forward-paragraph} code. | |
714 | ||
715 | Counting: Repetition and Regexps | |
716 | ||
717 | * Why Count Words:: | |
718 | * count-words-region:: Use a regexp, but find a problem. | |
719 | * recursive-count-words:: Start with case of no words in region. | |
720 | * Counting Exercise:: | |
721 | ||
722 | The @code{count-words-region} Function | |
723 | ||
724 | * Design count-words-region:: The definition using a @code{while} loop. | |
725 | * Whitespace Bug:: The Whitespace Bug in @code{count-words-region}. | |
726 | ||
727 | Counting Words in a @code{defun} | |
728 | ||
729 | * Divide and Conquer:: | |
730 | * Words and Symbols:: What to count? | |
731 | * Syntax:: What constitutes a word or symbol? | |
732 | * count-words-in-defun:: Very like @code{count-words}. | |
733 | * Several defuns:: Counting several defuns in a file. | |
734 | * Find a File:: Do you want to look at a file? | |
735 | * lengths-list-file:: A list of the lengths of many definitions. | |
736 | * Several files:: Counting in definitions in different files. | |
737 | * Several files recursively:: Recursively counting in different files. | |
738 | * Prepare the data:: Prepare the data for display in a graph. | |
739 | ||
740 | Count Words in @code{defuns} in Different Files | |
741 | ||
742 | * lengths-list-many-files:: Return a list of the lengths of defuns. | |
743 | * append:: Attach one list to another. | |
744 | ||
745 | Prepare the Data for Display in a Graph | |
746 | ||
747 | * Sorting:: Sorting lists. | |
748 | * Files List:: Making a list of files. | |
749 | * Counting function definitions:: | |
750 | ||
751 | Readying a Graph | |
752 | ||
753 | * Columns of a graph:: | |
754 | * graph-body-print:: How to print the body of a graph. | |
755 | * recursive-graph-body-print:: | |
756 | * Printed Axes:: | |
757 | * Line Graph Exercise:: | |
758 | ||
759 | Your @file{.emacs} File | |
760 | ||
761 | * Default Configuration:: | |
762 | * Site-wide Init:: You can write site-wide init files. | |
763 | * defcustom:: Emacs will write code for you. | |
764 | * Beginning a .emacs File:: How to write a @code{.emacs file}. | |
765 | * Text and Auto-fill:: Automatically wrap lines. | |
766 | * Mail Aliases:: Use abbreviations for email addresses. | |
767 | * Indent Tabs Mode:: Don't use tabs with @TeX{} | |
768 | * Keybindings:: Create some personal keybindings. | |
769 | * Keymaps:: More about key binding. | |
770 | * Loading Files:: Load (i.e., evaluate) files automatically. | |
771 | * Autoload:: Make functions available. | |
772 | * Simple Extension:: Define a function; bind it to a key. | |
773 | * X11 Colors:: Colors in version 19 in X. | |
774 | * Miscellaneous:: | |
775 | * Mode Line:: How to customize your mode line. | |
776 | ||
777 | Debugging | |
778 | ||
779 | * debug:: How to use the built-in debugger. | |
780 | * debug-on-entry:: Start debugging when you call a function. | |
781 | * debug-on-quit:: Start debugging when you quit with @kbd{C-g}. | |
782 | * edebug:: How to use Edebug, a source level debugger. | |
783 | * Debugging Exercises:: | |
784 | ||
785 | Handling the Kill Ring | |
786 | ||
787 | * rotate-yank-pointer:: Move a pointer along a list and around. | |
788 | * yank:: Paste a copy of a clipped element. | |
789 | * yank-pop:: Insert first element pointed to. | |
790 | ||
791 | The @code{rotate-yank-pointer} Function | |
792 | ||
793 | * Understanding rotate-yk-ptr:: | |
794 | * rotate-yk-ptr body:: The body of @code{rotate-yank-pointer}. | |
795 | ||
796 | The Body of @code{rotate-yank-pointer} | |
797 | ||
798 | * Digression concerning error:: How to mislead humans, but not computers. | |
799 | * rotate-yk-ptr else-part:: The else-part of the @code{if} expression. | |
800 | * Remainder Function:: The remainder, @code{%}, function. | |
801 | * rotate-yk-ptr remainder:: Using @code{%} in @code{rotate-yank-pointer}. | |
802 | * kill-rng-yk-ptr last elt:: Pointing to the last element. | |
803 | ||
804 | @code{yank} | |
805 | ||
806 | * rotate-yk-ptr arg:: Pass the argument to @code{rotate-yank-pointer}. | |
807 | * rotate-yk-ptr negative arg:: Pass a negative argument. | |
808 | ||
809 | A Graph with Labelled Axes | |
810 | ||
811 | * Labelled Example:: | |
812 | * print-graph Varlist:: @code{let} expression in @code{print-graph}. | |
813 | * print-Y-axis:: Print a label for the vertical axis. | |
814 | * print-X-axis:: Print a horizontal label. | |
815 | * Print Whole Graph:: The function to print a complete graph. | |
816 | ||
817 | The @code{print-Y-axis} Function | |
818 | ||
819 | * Height of label:: What height for the Y axis? | |
820 | * Compute a Remainder:: How to compute the remainder of a division. | |
821 | * Y Axis Element:: Construct a line for the Y axis. | |
822 | * Y-axis-column:: Generate a list of Y axis labels. | |
823 | * print-Y-axis Penultimate:: A not quite final version. | |
824 | ||
825 | The @code{print-X-axis} Function | |
826 | ||
827 | * Similarities differences:: Much like @code{print-Y-axis}, but not exactly. | |
828 | * X Axis Tic Marks:: Create tic marks for the horizontal axis. | |
829 | ||
830 | Printing the Whole Graph | |
831 | ||
832 | * The final version:: A few changes. | |
833 | * Test print-graph:: Run a short test. | |
834 | * Graphing words in defuns:: Executing the final code. | |
835 | * lambda:: How to write an anonymous function. | |
836 | * mapcar:: Apply a function to elements of a list. | |
837 | * Another Bug:: Yet another bug @dots{} most insidious. | |
838 | * Final printed graph:: The graph itself! | |
839 | ||
840 | @end detailmenu | |
841 | @end menu | |
842 | ||
843 | @node Preface, List Processing, Top, Top | |
844 | @comment node-name, next, previous, up | |
845 | @unnumbered Preface | |
846 | ||
847 | Most of the GNU Emacs integrated environment is written in the programming | |
848 | language called Emacs Lisp. The code written in this programming | |
849 | language is the software---the sets of instructions---that tell the | |
850 | computer what to do when you give it commands. Emacs is designed so | |
851 | that you can write new code in Emacs Lisp and easily install it as an | |
852 | extension to the editor. | |
853 | ||
854 | (GNU Emacs is sometimes called an ``extensible editor'', but it does | |
855 | much more than provide editing capabilities. It is better to refer to | |
856 | Emacs as an ``extensible computing environment''. However, that | |
857 | phrase is quite a mouthful. It is easier to refer to Emacs simply as | |
858 | an editor. Moreover, everything you do in Emacs---find the Mayan date | |
859 | and phases of the moon, simplify polynomials, debug code, manage | |
860 | files, read letters, write books---all these activities are kinds of | |
861 | editing in the most general sense of the word.) | |
862 | ||
863 | @menu | |
864 | * Why:: Why learn Emacs Lisp? | |
865 | * On Reading this Text:: Read, gain familiarity, pick up habits.... | |
866 | * Who You Are:: For whom this is written. | |
867 | * Lisp History:: | |
868 | * Note for Novices:: You can read this as a novice. | |
869 | * Thank You:: | |
870 | @end menu | |
871 | ||
872 | @node Why, On Reading this Text, Preface, Preface | |
873 | @ifnottex | |
874 | @unnumberedsec Why Study Emacs Lisp? | |
875 | @end ifnottex | |
876 | ||
877 | Although Emacs Lisp is usually thought of in association only with Emacs, | |
878 | it is a full computer programming language. You can use Emacs Lisp as | |
879 | you would any other programming language. | |
880 | ||
881 | Perhaps you want to understand programming; perhaps you want to extend | |
882 | Emacs; or perhaps you want to become a programmer. This introduction to | |
883 | Emacs Lisp is designed to get you started: to guide you in learning the | |
884 | fundamentals of programming, and more importantly, to show you how you | |
885 | can teach yourself to go further. | |
886 | ||
887 | @node On Reading this Text, Who You Are, Why, Preface | |
888 | @comment node-name, next, previous, up | |
889 | @unnumberedsec On Reading this Text | |
890 | ||
891 | All through this document, you will see little sample programs you can | |
892 | run inside of Emacs. If you read this document in Info inside of GNU | |
893 | Emacs, you can run the programs as they appear. (This is easy to do and | |
894 | is explained when the examples are presented.) Alternatively, you can | |
895 | read this introduction as a printed book while sitting beside a computer | |
896 | running Emacs. (This is what I like to do; I like printed books.) If | |
897 | you don't have a running Emacs beside you, you can still read this book, | |
898 | but in this case, it is best to treat it as a novel or as a travel guide | |
899 | to a country not yet visited: interesting, but not the same as being | |
900 | there. | |
901 | ||
902 | Much of this introduction is dedicated to walk-throughs or guided tours | |
903 | of code used in GNU Emacs. These tours are designed for two purposes: | |
904 | first, to give you familiarity with real, working code (code you use | |
905 | every day); and, second, to give you familiarity with the way Emacs | |
906 | works. It is interesting to see how a working environment is | |
907 | implemented. | |
908 | Also, I | |
909 | hope that you will pick up the habit of browsing through source code. | |
910 | You can learn from it and mine it for ideas. Having GNU Emacs is like | |
911 | having a dragon's cave of treasures. | |
912 | ||
913 | In addition to learning about Emacs as an editor and Emacs Lisp as a | |
914 | programming language, the examples and guided tours will give you an | |
915 | opportunity to get acquainted with Emacs as a Lisp programming | |
916 | environment. GNU Emacs supports programming and provides tools that | |
917 | you will want to become comfortable using, such as @kbd{M-.} (the key | |
918 | which invokes the @code{find-tag} command). You will also learn about | |
919 | buffers and other objects that are part of the environment. | |
920 | Learning about these features of Emacs is like learning new routes | |
921 | around your home town. | |
922 | ||
923 | @ignore | |
924 | In addition, I have written several programs as extended examples. | |
925 | Although these are examples, the programs are real. I use them. | |
926 | Other people use them. You may use them. Beyond the fragments of | |
927 | programs used for illustrations, there is very little in here that is | |
928 | `just for teaching purposes'; what you see is used. This is a great | |
929 | advantage of Emacs Lisp: it is easy to learn to use it for work. | |
930 | @end ignore | |
931 | ||
932 | Finally, I hope to convey some of the skills for using Emacs to | |
933 | learn aspects of programming that you don't know. You can often use | |
934 | Emacs to help you understand what puzzles you or to find out how to do | |
935 | something new. This self-reliance is not only a pleasure, but an | |
936 | advantage. | |
937 | ||
938 | @node Who You Are, Lisp History, On Reading this Text, Preface | |
939 | @comment node-name, next, previous, up | |
940 | @unnumberedsec For Whom This is Written | |
941 | ||
942 | This text is written as an elementary introduction for people who are | |
943 | not programmers. If you are a programmer, you may not be satisfied with | |
944 | this primer. The reason is that you may have become expert at reading | |
945 | reference manuals and be put off by the way this text is organized. | |
946 | ||
947 | An expert programmer who reviewed this text said to me: | |
948 | ||
949 | @quotation | |
950 | @i{I prefer to learn from reference manuals. I ``dive into'' each | |
951 | paragraph, and ``come up for air'' between paragraphs.} | |
952 | ||
953 | @i{When I get to the end of a paragraph, I assume that that subject is | |
954 | done, finished, that I know everything I need (with the | |
955 | possible exception of the case when the next paragraph starts talking | |
956 | about it in more detail). I expect that a well written reference manual | |
957 | will not have a lot of redundancy, and that it will have excellent | |
958 | pointers to the (one) place where the information I want is.} | |
959 | @end quotation | |
960 | ||
961 | This introduction is not written for this person! | |
962 | ||
963 | Firstly, I try to say everything at least three times: first, to | |
964 | introduce it; second, to show it in context; and third, to show it in a | |
965 | different context, or to review it. | |
966 | ||
967 | Secondly, I hardly ever put all the information about a subject in one | |
968 | place, much less in one paragraph. To my way of thinking, that imposes | |
969 | too heavy a burden on the reader. Instead I try to explain only what | |
970 | you need to know at the time. (Sometimes I include a little extra | |
971 | information so you won't be surprised later when the additional | |
972 | information is formally introduced.) | |
973 | ||
974 | When you read this text, you are not expected to learn everything the | |
975 | first time. Frequently, you need only make, as it were, a `nodding | |
976 | acquaintance' with some of the items mentioned. My hope is that I have | |
977 | structured the text and given you enough hints that you will be alert to | |
978 | what is important, and concentrate on it. | |
979 | ||
980 | You will need to ``dive into'' some paragraphs; there is no other way | |
981 | to read them. But I have tried to keep down the number of such | |
982 | paragraphs. This book is intended as an approachable hill, rather than | |
983 | as a daunting mountain. | |
984 | ||
985 | This introduction to @cite{Programming in Emacs Lisp} has a companion | |
986 | document, | |
987 | @iftex | |
988 | @cite{The GNU Emacs Lisp Reference Manual}. | |
989 | @end iftex | |
990 | @ifnottex | |
991 | @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU | |
992 | Emacs Lisp Reference Manual}. | |
993 | @end ifnottex | |
994 | The reference manual has more detail than this introduction. In the | |
995 | reference manual, all the information about one topic is concentrated | |
996 | in one place. You should turn to it if you are like the programmer | |
997 | quoted above. And, of course, after you have read this | |
998 | @cite{Introduction}, you will find the @cite{Reference Manual} useful | |
999 | when you are writing your own programs. | |
1000 | ||
1001 | @node Lisp History, Note for Novices, Who You Are, Preface | |
1002 | @unnumberedsec Lisp History | |
1003 | @cindex Lisp history | |
1004 | ||
1005 | Lisp was first developed in the late 1950s at the Massachusetts | |
1006 | Institute of Technology for research in artificial intelligence. The | |
1007 | great power of the Lisp language makes it superior for other purposes as | |
1008 | well, such as writing editor commands and integrated environments. | |
1009 | ||
1010 | @cindex Maclisp | |
1011 | @cindex Common Lisp | |
1012 | GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT | |
1013 | in the 1960s. It is somewhat inspired by Common Lisp, which became a | |
1014 | standard in the 1980s. However, Emacs Lisp is much simpler than Common | |
1015 | Lisp. (The standard Emacs distribution contains an optional extensions | |
1016 | file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.) | |
1017 | ||
1018 | @node Note for Novices, Thank You, Lisp History, Preface | |
1019 | @comment node-name, next, previous, up | |
1020 | @unnumberedsec A Note for Novices | |
1021 | ||
1022 | If you don't know GNU Emacs, you can still read this document | |
1023 | profitably. However, I recommend you learn Emacs, if only to learn to | |
1024 | move around your computer screen. You can teach yourself how to use | |
1025 | Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This | |
1026 | means you press and release the @key{CTRL} key and the @kbd{h} at the | |
1027 | same time, and then press and release @kbd{t}.) | |
1028 | ||
1029 | Also, I often refer to one of Emacs' standard commands by listing the | |
1030 | keys which you press to invoke the command and then giving the name of | |
1031 | the command in parentheses, like this: @kbd{M-C-\} | |
1032 | (@code{indent-region}). What this means is that the | |
1033 | @code{indent-region} command is customarily invoked by typing | |
1034 | @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to | |
1035 | invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, , | |
1036 | Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your | |
1037 | @key{META} key, @key{CTRL} key and @key{\} key all at the same time. | |
1038 | (On many modern keyboards the @key{META} key is labelled | |
1039 | @key{ALT}.) | |
1040 | Sometimes a combination like this is called a keychord, since it is | |
1041 | similar to the way you play a chord on a piano. If your keyboard does | |
1042 | not have a @key{META} key, the @key{ESC} key prefix is used in place | |
1043 | of it. In this case, @kbd{M-C-\} means that you press and release your | |
1044 | @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at | |
1045 | the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key | |
1046 | along with the key that is labelled @key{ALT} and, at the same time, | |
1047 | press the @key{\} key. | |
1048 | ||
1049 | In addition to typing a lone keychord, you can prefix what you type | |
1050 | with @kbd{C-u}, which is called the `universal argument'. The | |
1051 | @kbd{C-u} keychord passes an argument to the subsequent command. | |
1052 | Thus, to indent a region of plain text by 6 spaces, mark the region, | |
1053 | and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number, | |
1054 | Emacs either passes the number 4 to the command or otherwise runs the | |
1055 | command differently than it would otherwise.) @xref{Arguments, , | |
1056 | Numeric Arguments, emacs, The GNU Emacs Manual}. | |
1057 | ||
1058 | If you are reading this in Info using GNU Emacs, you can read through | |
1059 | this whole document just by pressing the space bar, @key{SPC}. | |
1060 | (To learn about Info, type @kbd{C-h i} and then select Info.) | |
1061 | ||
1062 | A note on terminology: when I use the word Lisp alone, I often am | |
1063 | referring to the various dialects of Lisp in general, but when I speak | |
1064 | of Emacs Lisp, I am referring to GNU Emacs Lisp in particular. | |
1065 | ||
1066 | @node Thank You, , Note for Novices, Preface | |
1067 | @comment node-name, next, previous, up | |
1068 | @unnumberedsec Thank You | |
1069 | ||
1070 | My thanks to all who helped me with this book. My especial thanks to | |
1071 | @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland | |
1072 | McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.@: | |
1073 | Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both | |
1074 | @w{Philip Johnson} and @w{David Stampe} for their patient | |
1075 | encouragement. My mistakes are my own. | |
1076 | ||
1077 | @flushright | |
1078 | Robert J. Chassell | |
1079 | @end flushright | |
1080 | ||
1081 | @c ================ Beginning of main text ================ | |
1082 | ||
1083 | @c Start main text on right-hand (verso) page | |
1084 | ||
1085 | @tex | |
1086 | \par\vfill\supereject | |
1087 | \headings off | |
1088 | \ifodd\pageno | |
1089 | \par\vfill\supereject | |
1090 | \else | |
1091 | \par\vfill\supereject | |
1092 | \page\hbox{}\page | |
1093 | \par\vfill\supereject | |
1094 | \fi | |
1095 | @end tex | |
1096 | ||
1097 | @iftex | |
1098 | @headings off | |
1099 | @evenheading @thispage @| @| @thischapter | |
1100 | @oddheading @thissection @| @| @thispage | |
1101 | @pageno = 1 | |
1102 | @end iftex | |
1103 | ||
1104 | @node List Processing, Practicing Evaluation, Preface, Top | |
1105 | @comment node-name, next, previous, up | |
1106 | @chapter List Processing | |
1107 | ||
1108 | To the untutored eye, Lisp is a strange programming language. In Lisp | |
1109 | code there are parentheses everywhere. Some people even claim that the | |
1110 | name stands for `Lots of Isolated Silly Parentheses'. But the claim is | |
1111 | unwarranted. Lisp stands for LISt Processing, and the programming | |
1112 | language handles @emph{lists} (and lists of lists) by putting them | |
1113 | between parentheses. The parentheses mark the boundaries of the list. | |
1114 | Sometimes a list is preceded by a single apostrophe or quotation mark, | |
1115 | @samp{'}. Lists are the basis of Lisp. | |
1116 | ||
1117 | @menu | |
1118 | * Lisp Lists:: What are lists? | |
1119 | * Run a Program:: Any list in Lisp is a program ready to run. | |
1120 | * Making Errors:: Generating an error message. | |
1121 | * Names & Definitions:: Names of symbols and function definitions. | |
1122 | * Lisp Interpreter:: What the Lisp interpreter does. | |
1123 | * Evaluation:: Running a program. | |
1124 | * Variables:: Returning a value from a variable. | |
1125 | * Arguments:: Passing information to a function. | |
1126 | * set & setq:: Setting the value of a variable. | |
1127 | * Summary:: The major points. | |
1128 | * Error Message Exercises:: | |
1129 | @end menu | |
1130 | ||
1131 | @node Lisp Lists, Run a Program, List Processing, List Processing | |
1132 | @comment node-name, next, previous, up | |
1133 | @section Lisp Lists | |
1134 | @cindex Lisp Lists | |
1135 | ||
1136 | In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}. | |
1137 | This list is preceded by a single apostrophe. It could just as well be | |
1138 | written as follows, which looks more like the kind of list you are likely | |
1139 | to be familiar with: | |
1140 | ||
1141 | @smallexample | |
1142 | @group | |
1143 | '(rose | |
1144 | violet | |
1145 | daisy | |
1146 | buttercup) | |
1147 | @end group | |
1148 | @end smallexample | |
1149 | ||
1150 | @noindent | |
1151 | The elements of this list are the names of the four different flowers, | |
1152 | separated from each other by whitespace and surrounded by parentheses, | |
1153 | like flowers in a field with a stone wall around them. | |
1154 | @cindex Flowers in a field | |
1155 | ||
1156 | @menu | |
1157 | * Numbers Lists:: List have numbers, other lists, in them. | |
1158 | * Lisp Atoms:: Elemental entities. | |
1159 | * Whitespace in Lists:: Formating lists to be readable. | |
1160 | * Typing Lists:: How GNU Emacs helps you type lists. | |
1161 | @end menu | |
1162 | ||
1163 | @node Numbers Lists, Lisp Atoms, Lisp Lists, Lisp Lists | |
1164 | @ifnottex | |
1165 | @unnumberedsubsec Numbers, Lists inside of Lists | |
1166 | @end ifnottex | |
1167 | ||
1168 | Lists can also have numbers in them, as in this list: @code{(+ 2 2)}. | |
1169 | This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each | |
1170 | separated by whitespace. | |
1171 | ||
1172 | In Lisp, both data and programs are represented the same way; that is, | |
1173 | they are both lists of words, numbers, or other lists, separated by | |
1174 | whitespace and surrounded by parentheses. (Since a program looks like | |
1175 | data, one program may easily serve as data for another; this is a very | |
1176 | powerful feature of Lisp.) (Incidentally, these two parenthetical | |
1177 | remarks are @emph{not} Lisp lists, because they contain @samp{;} and | |
1178 | @samp{.} as punctuation marks.) | |
1179 | ||
1180 | @need 1200 | |
1181 | Here is another list, this time with a list inside of it: | |
1182 | ||
1183 | @smallexample | |
1184 | '(this list has (a list inside of it)) | |
1185 | @end smallexample | |
1186 | ||
1187 | The components of this list are the words @samp{this}, @samp{list}, | |
1188 | @samp{has}, and the list @samp{(a list inside of it)}. The interior | |
1189 | list is made up of the words @samp{a}, @samp{list}, @samp{inside}, | |
1190 | @samp{of}, @samp{it}. | |
1191 | ||
1192 | @node Lisp Atoms, Whitespace in Lists, Numbers Lists, Lisp Lists | |
1193 | @comment node-name, next, previous, up | |
1194 | @subsection Lisp Atoms | |
1195 | @cindex Lisp Atoms | |
1196 | ||
1197 | In Lisp, what we have been calling words are called @dfn{atoms}. This | |
1198 | term comes from the historical meaning of the word atom, which means | |
1199 | `indivisible'. As far as Lisp is concerned, the words we have been | |
1200 | using in the lists cannot be divided into any smaller parts and still | |
1201 | mean the same thing as part of a program; likewise with numbers and | |
1202 | single character symbols like @samp{+}. On the other hand, unlike an | |
1203 | atom, a list can be split into parts. (@xref{car cdr & cons, , | |
1204 | @code{car} @code{cdr} & @code{cons} Fundamental Functions}.) | |
1205 | ||
1206 | In a list, atoms are separated from each other by whitespace. They can be | |
1207 | right next to a parenthesis. | |
1208 | ||
1209 | @cindex @samp{empty list} defined | |
1210 | Technically speaking, a list in Lisp consists of parentheses surrounding | |
1211 | atoms separated by whitespace or surrounding other lists or surrounding | |
1212 | both atoms and other lists. A list can have just one atom in it or | |
1213 | have nothing in it at all. A list with nothing in it looks like this: | |
1214 | @code{()}, and is called the @dfn{empty list}. Unlike anything else, an | |
1215 | empty list is considered both an atom and a list at the same time. | |
1216 | ||
1217 | @cindex Symbolic expressions, introduced | |
1218 | @cindex @samp{expression} defined | |
1219 | @cindex @samp{form} defined | |
1220 | The printed representation of both atoms and lists are called | |
1221 | @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}. | |
1222 | The word @dfn{expression} by itself can refer to either the printed | |
1223 | representation, or to the atom or list as it is held internally in the | |
1224 | computer. Often, people use the term @dfn{expression} | |
1225 | indiscriminately. (Also, in many texts, the word @dfn{form} is used | |
1226 | as a synonym for expression.) | |
1227 | ||
1228 | Incidentally, the atoms that make up our universe were named such when | |
1229 | they were thought to be indivisible; but it has been found that physical | |
1230 | atoms are not indivisible. Parts can split off an atom or it can | |
1231 | fission into two parts of roughly equal size. Physical atoms were named | |
1232 | prematurely, before their truer nature was found. In Lisp, certain | |
1233 | kinds of atom, such as an array, can be separated into parts; but the | |
1234 | mechanism for doing this is different from the mechanism for splitting a | |
1235 | list. As far as list operations are concerned, the atoms of a list are | |
1236 | unsplittable. | |
1237 | ||
1238 | As in English, the meanings of the component letters of a Lisp atom | |
1239 | are different from the meaning the letters make as a word. For | |
1240 | example, the word for the South American sloth, the @samp{ai}, is | |
1241 | completely different from the two words, @samp{a}, and @samp{i}. | |
1242 | ||
1243 | There are many kinds of atom in nature but only a few in Lisp: for | |
1244 | example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such | |
1245 | as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have | |
1246 | listed in the examples above are all symbols. In everyday Lisp | |
1247 | conversation, the word ``atom'' is not often used, because programmers | |
1248 | usually try to be more specific about what kind of atom they are dealing | |
1249 | with. Lisp programming is mostly about symbols (and sometimes numbers) | |
1250 | within lists. (Incidentally, the preceding three word parenthetical | |
1251 | remark is a proper list in Lisp, since it consists of atoms, which in | |
1252 | this case are symbols, separated by whitespace and enclosed by | |
1253 | parentheses, without any non-Lisp punctuation.) | |
1254 | ||
1255 | @need 1250 | |
1256 | In addition, text between double quotation marks---even sentences or | |
1257 | paragraphs---is an atom. Here is an example: | |
1258 | @cindex Text between double quotation marks | |
1259 | ||
1260 | @smallexample | |
1261 | '(this list includes "text between quotation marks.") | |
1262 | @end smallexample | |
1263 | ||
1264 | @cindex @samp{string} defined | |
1265 | @noindent | |
1266 | In Lisp, all of the quoted text including the punctuation mark and the | |
1267 | blank spaces is a single atom. This kind of atom is called a | |
1268 | @dfn{string} (for `string of characters') and is the sort of thing that | |
1269 | is used for messages that a computer can print for a human to read. | |
1270 | Strings are a different kind of atom than numbers or symbols and are | |
1271 | used differently. | |
1272 | ||
1273 | @node Whitespace in Lists, Typing Lists, Lisp Atoms, Lisp Lists | |
1274 | @comment node-name, next, previous, up | |
1275 | @subsection Whitespace in Lists | |
1276 | @cindex Whitespace in lists | |
1277 | ||
1278 | @need 1200 | |
1279 | The amount of whitespace in a list does not matter. From the point of view | |
1280 | of the Lisp language, | |
1281 | ||
1282 | @smallexample | |
1283 | @group | |
1284 | '(this list | |
1285 | looks like this) | |
1286 | @end group | |
1287 | @end smallexample | |
1288 | ||
1289 | @need 800 | |
1290 | @noindent | |
1291 | is exactly the same as this: | |
1292 | ||
1293 | @smallexample | |
1294 | '(this list looks like this) | |
1295 | @end smallexample | |
1296 | ||
1297 | Both examples show what to Lisp is the same list, the list made up of | |
1298 | the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and | |
1299 | @samp{this} in that order. | |
1300 | ||
1301 | Extra whitespace and newlines are designed to make a list more readable | |
1302 | by humans. When Lisp reads the expression, it gets rid of all the extra | |
1303 | whitespace (but it needs to have at least one space between atoms in | |
1304 | order to tell them apart.) | |
1305 | ||
1306 | Odd as it seems, the examples we have seen cover almost all of what Lisp | |
1307 | lists look like! Every other list in Lisp looks more or less like one | |
1308 | of these examples, except that the list may be longer and more complex. | |
1309 | In brief, a list is between parentheses, a string is between quotation | |
1310 | marks, a symbol looks like a word, and a number looks like a number. | |
1311 | (For certain situations, square brackets, dots and a few other special | |
1312 | characters may be used; however, we will go quite far without them.) | |
1313 | ||
1314 | @node Typing Lists, , Whitespace in Lists, Lisp Lists | |
1315 | @comment node-name, next, previous, up | |
1316 | @subsection GNU Emacs Helps You Type Lists | |
1317 | @cindex Help typing lists | |
1318 | @cindex Formatting help | |
1319 | ||
1320 | When you type a Lisp expression in GNU Emacs using either Lisp | |
1321 | Interaction mode or Emacs Lisp mode, you have available to you several | |
1322 | commands to format the Lisp expression so it is easy to read. For | |
1323 | example, pressing the @key{TAB} key automatically indents the line the | |
1324 | cursor is on by the right amount. A command to properly indent the | |
1325 | code in a region is customarily bound to @kbd{M-C-\}. Indentation is | |
1326 | designed so that you can see which elements of a list belongs to which | |
1327 | list---elements of a sub-list are indented more than the elements of | |
1328 | the enclosing list. | |
1329 | ||
1330 | In addition, when you type a closing parenthesis, Emacs momentarily | |
1331 | jumps the cursor back to the matching opening parenthesis, so you can | |
1332 | see which one it is. This is very useful, since every list you type | |
1333 | in Lisp must have its closing parenthesis match its opening | |
1334 | parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs | |
1335 | Manual}, for more information about Emacs' modes.) | |
1336 | ||
1337 | @node Run a Program, Making Errors, Lisp Lists, List Processing | |
1338 | @comment node-name, next, previous, up | |
1339 | @section Run a Program | |
1340 | @cindex Run a program | |
1341 | @cindex Program, running one | |
1342 | ||
1343 | @cindex @samp{evaluate} defined | |
1344 | A list in Lisp---any list---is a program ready to run. If you run it | |
1345 | (for which the Lisp jargon is @dfn{evaluate}), the computer will do one | |
1346 | of three things: do nothing except return to you the list itself; send | |
1347 | you an error message; or, treat the first symbol in the list as a | |
1348 | command to do something. (Usually, of course, it is the last of these | |
1349 | three things that you really want!) | |
1350 | ||
1351 | @c use code for the single apostrophe, not samp. | |
1352 | The single apostrophe, @code{'}, that I put in front of some of the | |
1353 | example lists in preceding sections is called a @dfn{quote}; when it | |
1354 | precedes a list, it tells Lisp to do nothing with the list, other than | |
1355 | take it as it is written. But if there is no quote preceding a list, | |
1356 | the first item of the list is special: it is a command for the computer | |
1357 | to obey. (In Lisp, these commands are called @emph{functions}.) The list | |
1358 | @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp | |
1359 | understands that the @code{+} is an instruction to do something with the | |
1360 | rest of the list: add the numbers that follow. | |
1361 | ||
1362 | @need 1250 | |
1363 | If you are reading this inside of GNU Emacs in Info, here is how you can | |
1364 | evaluate such a list: place your cursor immediately after the right | |
1365 | hand parenthesis of the following list and then type @kbd{C-x C-e}: | |
1366 | ||
1367 | @smallexample | |
1368 | (+ 2 2) | |
1369 | @end smallexample | |
1370 | ||
1371 | @c use code for the number four, not samp. | |
1372 | @noindent | |
1373 | You will see the number @code{4} appear in the echo area. (In the | |
1374 | jargon, what you have just done is ``evaluate the list.'' The echo area | |
1375 | is the line at the bottom of the screen that displays or ``echoes'' | |
1376 | text.) Now try the same thing with a quoted list: place the cursor | |
1377 | right after the following list and type @kbd{C-x C-e}: | |
1378 | ||
1379 | @smallexample | |
1380 | '(this is a quoted list) | |
1381 | @end smallexample | |
1382 | ||
1383 | @noindent | |
1384 | You will see @code{(this is a quoted list)} appear in the echo area. | |
1385 | ||
1386 | @cindex Lisp interpreter, explained | |
1387 | @cindex Interpreter, Lisp, explained | |
1388 | In both cases, what you are doing is giving a command to the program | |
1389 | inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the | |
1390 | interpreter a command to evaluate the expression. The name of the Lisp | |
1391 | interpreter comes from the word for the task done by a human who comes | |
1392 | up with the meaning of an expression---who ``interprets'' it. | |
1393 | ||
1394 | You can also evaluate an atom that is not part of a list---one that is | |
1395 | not surrounded by parentheses; again, the Lisp interpreter translates | |
1396 | from the humanly readable expression to the language of the computer. | |
1397 | But before discussing this (@pxref{Variables}), we will discuss what the | |
1398 | Lisp interpreter does when you make an error. | |
1399 | ||
1400 | @node Making Errors, Names & Definitions, Run a Program, List Processing | |
1401 | @comment node-name, next, previous, up | |
1402 | @section Generate an Error Message | |
1403 | @cindex Generate an error message | |
1404 | @cindex Error message generation | |
1405 | ||
1406 | Partly so you won't worry if you do it accidentally, we will now give | |
1407 | a command to the Lisp interpreter that generates an error message. | |
1408 | This is a harmless activity; and indeed, we will often try to generate | |
1409 | error messages intentionally. Once you understand the jargon, error | |
1410 | messages can be informative. Instead of being called ``error'' | |
1411 | messages, they should be called ``help'' messages. They are like | |
1412 | signposts to a traveller in a strange country; deciphering them can be | |
1413 | hard, but once understood, they can point the way. | |
1414 | ||
1415 | The error message is generated by a built-in GNU Emacs debugger. We | |
1416 | will `enter the debugger'. You get out of the debugger by typing @code{q}. | |
1417 | ||
1418 | What we will do is evaluate a list that is not quoted and does not | |
1419 | have a meaningful command as its first element. Here is a list almost | |
1420 | exactly the same as the one we just used, but without the single-quote | |
1421 | in front of it. Position the cursor right after it and type @kbd{C-x | |
1422 | C-e}: | |
1423 | ||
1424 | @smallexample | |
1425 | (this is an unquoted list) | |
1426 | @end smallexample | |
1427 | ||
1428 | @noindent | |
1429 | What you see depends on which version of Emacs you are running. GNU | |
1430 | Emacs version 21 provides more information than version 20 and before. | |
1431 | First, the more recent result of generating an error; then the | |
1432 | earlier, version 20 result. | |
1433 | ||
1434 | @need 1250 | |
1435 | @noindent | |
1436 | In GNU Emacs version 21, a @file{*Backtrace*} window will open up and | |
1437 | you will see the following in it: | |
1438 | ||
1439 | @smallexample | |
1440 | @group | |
1441 | ---------- Buffer: *Backtrace* ---------- | |
1442 | Debugger entered--Lisp error: (void-function this) | |
1443 | (this is an unquoted list) | |
1444 | eval((this is an unquoted list)) | |
1445 | eval-last-sexp-1(nil) | |
1446 | eval-last-sexp(nil) | |
1447 | call-interactively(eval-last-sexp) | |
1448 | ---------- Buffer: *Backtrace* ---------- | |
1449 | @end group | |
1450 | @end smallexample | |
1451 | ||
1452 | @need 1200 | |
1453 | @noindent | |
1454 | Your cursor will be in this window (you may have to wait a few seconds | |
1455 | before it becomes visible). To quit the debugger and make the | |
1456 | debugger window go away, type: | |
1457 | ||
1458 | @smallexample | |
1459 | q | |
1460 | @end smallexample | |
1461 | ||
1462 | @noindent | |
1463 | Please type @kbd{q} right now, so you become confident that you can | |
1464 | get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter | |
1465 | it. | |
1466 | ||
1467 | @cindex @samp{function} defined | |
1468 | Based on what we already know, we can almost read this error message. | |
1469 | ||
1470 | You read the @file{*Backtrace*} buffer from the bottom up; it tells | |
1471 | you what Emacs did. When you typed @kbd{C-x C-e}, you made an | |
1472 | interactive call to the command @code{eval-last-sexp}. @code{eval} is | |
1473 | an abbreviation for `evaluate' and @code{sexp} is an abbreviation for | |
1474 | `symbolic expression'. The command means `evaluate last symbolic | |
1475 | expression', which is the expression just before your cursor. | |
1476 | ||
1477 | Each line above tells you what the Lisp interpreter evaluated next. | |
1478 | The most recent action is at the top. The buffer is called the | |
1479 | @file{*Backtrace*} buffer because it enables you to track Emacs | |
1480 | backwards. | |
1481 | ||
1482 | @need 800 | |
1483 | At the top of the @file{*Backtrace*} buffer, you see the line: | |
1484 | ||
1485 | @smallexample | |
1486 | Debugger entered--Lisp error: (void-function this) | |
1487 | @end smallexample | |
1488 | ||
1489 | @noindent | |
1490 | The Lisp interpreter tried to evaluate the first atom of the list, the | |
1491 | word @samp{this}. It is this action that generated the error message | |
1492 | @samp{void-function this}. | |
1493 | ||
1494 | The message contains the words @samp{void-function} and @samp{this}. | |
1495 | ||
1496 | @cindex @samp{function} defined | |
1497 | The word @samp{function} was mentioned once before. It is a very | |
1498 | important word. For our purposes, we can define it by saying that a | |
1499 | @dfn{function} is a set of instructions to the computer that tell the | |
1500 | computer to do something. | |
1501 | ||
1502 | Now we can begin to understand the error message: @samp{void-function | |
1503 | this}. The function (that is, the word @samp{this}) does not have a | |
1504 | definition of any set of instructions for the computer to carry out. | |
1505 | ||
1506 | The slightly odd word, @samp{void-function}, is designed to cover the | |
1507 | way Emacs Lisp is implemented, which is that when a symbol does not | |
1508 | have a function definition attached to it, the place that should | |
1509 | contain the instructions is `void'. | |
1510 | ||
1511 | On the other hand, since we were able to add 2 plus 2 successfully, by | |
1512 | evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must | |
1513 | have a set of instructions for the computer to obey and those | |
1514 | instructions must be to add the numbers that follow the @code{+}. | |
1515 | ||
1516 | @need 1250 | |
1517 | In GNU Emacs version 20, and in earlier versions, you will see only | |
1518 | one line of error message; it will appear in the echo area and look | |
1519 | like this: | |
1520 | ||
1521 | @smallexample | |
1522 | Symbol's function definition is void:@: this | |
1523 | @end smallexample | |
1524 | ||
1525 | @noindent | |
1526 | (Also, your terminal may beep at you---some do, some don't; and others | |
1527 | blink. This is just a device to get your attention.) The message goes | |
1528 | away as soon as you type another key, even just to move the cursor. | |
1529 | ||
1530 | We know the meaning of the word @samp{Symbol}. It refers to the first | |
1531 | atom of the list, the word @samp{this}. The word @samp{function} | |
1532 | refers to the instructions that tell the computer what to do. | |
1533 | (Technically, the symbol tells the computer where to find the | |
1534 | instructions, but this is a complication we can ignore for the | |
1535 | moment.) | |
1536 | ||
1537 | The error message can be understood: @samp{Symbol's function | |
1538 | definition is void:@: this}. The symbol (that is, the word | |
1539 | @samp{this}) lacks instructions for the computer to carry out. | |
1540 | ||
1541 | @node Names & Definitions, Lisp Interpreter, Making Errors, List Processing | |
1542 | @comment node-name, next, previous, up | |
1543 | @section Symbol Names and Function Definitions | |
1544 | @cindex Symbol names | |
1545 | ||
1546 | We can articulate another characteristic of Lisp based on what we have | |
1547 | discussed so far---an important characteristic: a symbol, like | |
1548 | @code{+}, is not itself the set of instructions for the computer to | |
1549 | carry out. Instead, the symbol is used, perhaps temporarily, as a way | |
1550 | of locating the definition or set of instructions. What we see is the | |
1551 | name through which the instructions can be found. Names of people | |
1552 | work the same way. I can be referred to as @samp{Bob}; however, I am | |
1553 | not the letters @samp{B}, @samp{o}, @samp{b} but am the consciousness | |
1554 | consistently associated with a particular life-form. The name is not | |
1555 | me, but it can be used to refer to me. | |
1556 | ||
1557 | In Lisp, one set of instructions can be attached to several names. | |
1558 | For example, the computer instructions for adding numbers can be | |
1559 | linked to the symbol @code{plus} as well as to the symbol @code{+} | |
1560 | (and are in some dialects of Lisp). Among humans, I can be referred | |
1561 | to as @samp{Robert} as well as @samp{Bob} and by other words as well. | |
1562 | ||
1563 | On the other hand, a symbol can have only one function definition | |
1564 | attached to it at a time. Otherwise, the computer would be confused as | |
1565 | to which definition to use. If this were the case among people, only | |
1566 | one person in the world could be named @samp{Bob}. However, the function | |
1567 | definition to which the name refers can be changed readily. | |
1568 | (@xref{Install, , Install a Function Definition}.) | |
1569 | ||
1570 | Since Emacs Lisp is large, it is customary to name symbols in a way | |
1571 | that identifies the part of Emacs to which the function belongs. | |
1572 | Thus, all the names for functions that deal with Texinfo start with | |
1573 | @samp{texinfo-} and those for functions that deal with reading mail | |
1574 | start with @samp{rmail-}. | |
1575 | ||
1576 | @node Lisp Interpreter, Evaluation, Names & Definitions, List Processing | |
1577 | @comment node-name, next, previous, up | |
1578 | @section The Lisp Interpreter | |
1579 | @cindex Lisp interpreter, what it does | |
1580 | @cindex Interpreter, what it does | |
1581 | ||
1582 | Based on what we have seen, we can now start to figure out what the | |
1583 | Lisp interpreter does when we command it to evaluate a list. | |
1584 | First, it looks to see whether there is a quote before the list; if | |
1585 | there is, the interpreter just gives us the list. On the other | |
1586 | hand, if there is no quote, the interpreter looks at the first element | |
1587 | in the list and sees whether it has a function definition. If it does, | |
1588 | the interpreter carries out the instructions in the function definition. | |
1589 | Otherwise, the interpreter prints an error message. | |
1590 | ||
1591 | This is how Lisp works. Simple. There are added complications which we | |
1592 | will get to in a minute, but these are the fundamentals. Of course, to | |
1593 | write Lisp programs, you need to know how to write function definitions | |
1594 | and attach them to names, and how to do this without confusing either | |
1595 | yourself or the computer. | |
1596 | ||
1597 | @menu | |
1598 | * Complications:: Variables, Special forms, Lists within. | |
1599 | * Byte Compiling:: Specially processing code for speed. | |
1600 | @end menu | |
1601 | ||
1602 | @node Complications, Byte Compiling, Lisp Interpreter, Lisp Interpreter | |
1603 | @ifnottex | |
1604 | @unnumberedsubsec Complications | |
1605 | @end ifnottex | |
1606 | ||
1607 | Now, for the first complication. In addition to lists, the Lisp | |
1608 | interpreter can evaluate a symbol that is not quoted and does not have | |
1609 | parentheses around it. The Lisp interpreter will attempt to determine | |
1610 | the symbol's value as a @dfn{variable}. This situation is described | |
1611 | in the section on variables. (@xref{Variables}.) | |
1612 | ||
1613 | @cindex Special form | |
1614 | The second complication occurs because some functions are unusual and do | |
1615 | not work in the usual manner. Those that don't are called @dfn{special | |
1616 | forms}. They are used for special jobs, like defining a function, and | |
1617 | there are not many of them. In the next few chapters, you will be | |
1618 | introduced to several of the more important special forms. | |
1619 | ||
1620 | The third and final complication is this: if the function that the | |
1621 | Lisp interpreter is looking at is not a special form, and if it is part | |
1622 | of a list, the Lisp interpreter looks to see whether the list has a list | |
1623 | inside of it. If there is an inner list, the Lisp interpreter first | |
1624 | figures out what it should do with the inside list, and then it works on | |
1625 | the outside list. If there is yet another list embedded inside the | |
1626 | inner list, it works on that one first, and so on. It always works on | |
1627 | the innermost list first. The interpreter works on the innermost list | |
1628 | first, to evaluate the result of that list. The result may be | |
1629 | used by the enclosing expression. | |
1630 | ||
1631 | Otherwise, the interpreter works left to right, from one expression to | |
1632 | the next. | |
1633 | ||
1634 | @node Byte Compiling, , Complications, Lisp Interpreter | |
1635 | @subsection Byte Compiling | |
1636 | @cindex Byte compiling | |
1637 | ||
1638 | One other aspect of interpreting: the Lisp interpreter is able to | |
1639 | interpret two kinds of entity: humanly readable code, on which we will | |
1640 | focus exclusively, and specially processed code, called @dfn{byte | |
1641 | compiled} code, which is not humanly readable. Byte compiled code | |
1642 | runs faster than humanly readable code. | |
1643 | ||
1644 | You can transform humanly readable code into byte compiled code by | |
1645 | running one of the compile commands such as @code{byte-compile-file}. | |
1646 | Byte compiled code is usually stored in a file that ends with a | |
1647 | @file{.elc} extension rather than a @file{.el} extension. You will | |
1648 | see both kinds of file in the @file{emacs/lisp} directory; the files | |
1649 | to read are those with @file{.el} extensions. | |
1650 | ||
1651 | As a practical matter, for most things you might do to customize or | |
1652 | extend Emacs, you do not need to byte compile; and I will not discuss | |
1653 | the topic here. @xref{Byte Compilation, , Byte Compilation, elisp, | |
1654 | The GNU Emacs Lisp Reference Manual}, for a full description of byte | |
1655 | compilation. | |
1656 | ||
1657 | @node Evaluation, Variables, Lisp Interpreter, List Processing | |
1658 | @comment node-name, next, previous, up | |
1659 | @section Evaluation | |
1660 | @cindex Evaluation | |
1661 | ||
1662 | When the Lisp interpreter works on an expression, the term for the | |
1663 | activity is called @dfn{evaluation}. We say that the interpreter | |
1664 | `evaluates the expression'. I've used this term several times before. | |
1665 | The word comes from its use in everyday language, `to ascertain the | |
1666 | value or amount of; to appraise', according to @cite{Webster's New | |
1667 | Collegiate Dictionary}. | |
1668 | ||
1669 | After evaluating an expression, the Lisp interpreter will most likely | |
1670 | @dfn{return} the value that the computer produces by carrying out the | |
1671 | instructions it found in the function definition, or perhaps it will | |
1672 | give up on that function and produce an error message. (The interpreter | |
1673 | may also find itself tossed, so to speak, to a different function or it | |
1674 | may attempt to repeat continually what it is doing for ever and ever in | |
1675 | what is called an `infinite loop'. These actions are less common; and | |
1676 | we can ignore them.) Most frequently, the interpreter returns a value. | |
1677 | ||
1678 | @cindex @samp{side effect} defined | |
1679 | At the same time the interpreter returns a value, it may do something | |
1680 | else as well, such as move a cursor or copy a file; this other kind of | |
1681 | action is called a @dfn{side effect}. Actions that we humans think are | |
1682 | important, such as printing results, are often ``side effects'' to the | |
1683 | Lisp interpreter. The jargon can sound peculiar, but it turns out that | |
1684 | it is fairly easy to learn to use side effects. | |
1685 | ||
1686 | In summary, evaluating a symbolic expression most commonly causes the | |
1687 | Lisp interpreter to return a value and perhaps carry out a side effect; | |
1688 | or else produce an error. | |
1689 | ||
1690 | @menu | |
1691 | * Evaluating Inner Lists:: Lists within lists... | |
1692 | @end menu | |
1693 | ||
1694 | @node Evaluating Inner Lists, , Evaluation, Evaluation | |
1695 | @comment node-name, next, previous, up | |
1696 | @subsection Evaluating Inner Lists | |
1697 | @cindex Inner list evaluation | |
1698 | @cindex Evaluating inner lists | |
1699 | ||
1700 | If evaluation applies to a list that is inside another list, the outer | |
1701 | list may use the value returned by the first evaluation as information | |
1702 | when the outer list is evaluated. This explains why inner expressions | |
1703 | are evaluated first: the values they return are used by the outer | |
1704 | expressions. | |
1705 | ||
1706 | @need 1250 | |
1707 | We can investigate this process by evaluating another addition example. | |
1708 | Place your cursor after the following expression and type @kbd{C-x C-e}: | |
1709 | ||
1710 | @smallexample | |
1711 | (+ 2 (+ 3 3)) | |
1712 | @end smallexample | |
1713 | ||
1714 | @noindent | |
1715 | The number 8 will appear in the echo area. | |
1716 | ||
1717 | What happens is that the Lisp interpreter first evaluates the inner | |
1718 | expression, @code{(+ 3 3)}, for which the value 6 is returned; then it | |
1719 | evaluates the outer expression as if it were written @code{(+ 2 6)}, which | |
1720 | returns the value 8. Since there are no more enclosing expressions to | |
1721 | evaluate, the interpreter prints that value in the echo area. | |
1722 | ||
1723 | Now it is easy to understand the name of the command invoked by the | |
1724 | keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The | |
1725 | letters @code{sexp} are an abbreviation for `symbolic expression', and | |
1726 | @code{eval} is an abbreviation for `evaluate'. The command means | |
1727 | `evaluate last symbolic expression'. | |
1728 | ||
1729 | As an experiment, you can try evaluating the expression by putting the | |
1730 | cursor at the beginning of the next line immediately following the | |
1731 | expression, or inside the expression. | |
1732 | ||
1733 | @need 800 | |
1734 | Here is another copy of the expression: | |
1735 | ||
1736 | @smallexample | |
1737 | (+ 2 (+ 3 3)) | |
1738 | @end smallexample | |
1739 | ||
1740 | @noindent | |
1741 | If you place the cursor at the beginning of the blank line that | |
1742 | immediately follows the expression and type @kbd{C-x C-e}, you will | |
1743 | still get the value 8 printed in the echo area. Now try putting the | |
1744 | cursor inside the expression. If you put it right after the next to | |
1745 | last parenthesis (so it appears to sit on top of the last parenthesis), | |
1746 | you will get a 6 printed in the echo area! This is because the command | |
1747 | evaluates the expression @code{(+ 3 3)}. | |
1748 | ||
1749 | Now put the cursor immediately after a number. Type @kbd{C-x C-e} and | |
1750 | you will get the number itself. In Lisp, if you evaluate a number, you | |
1751 | get the number itself---this is how numbers differ from symbols. If you | |
1752 | evaluate a list starting with a symbol like @code{+}, you will get a | |
1753 | value returned that is the result of the computer carrying out the | |
1754 | instructions in the function definition attached to that name. If a | |
1755 | symbol by itself is evaluated, something different happens, as we will | |
1756 | see in the next section. | |
1757 | ||
1758 | @node Variables, Arguments, Evaluation, List Processing | |
1759 | @comment node-name, next, previous, up | |
1760 | @section Variables | |
1761 | @cindex Variables | |
1762 | ||
1763 | In Emacs Lisp, a symbol can have a value attached to it just as it can | |
1764 | have a function definition attached to it. The two are different. | |
1765 | The function definition is a set of instructions that a computer will | |
1766 | obey. A value, on the other hand, is something, such as number or a | |
1767 | name, that can vary (which is why such a symbol is called a variable). | |
1768 | The value of a symbol can be any expression in Lisp, such as a symbol, | |
1769 | number, list, or string. A symbol that has a value is often called a | |
1770 | @dfn{variable}. | |
1771 | ||
1772 | A symbol can have both a function definition and a value attached to | |
1773 | it at the same time. Or it can have just one or the other. | |
1774 | The two are separate. This is somewhat similar | |
1775 | to the way the name Cambridge can refer to the city in Massachusetts | |
1776 | and have some information attached to the name as well, such as | |
1777 | ``great programming center''. | |
1778 | ||
1779 | @ignore | |
1780 | (Incidentally, in Emacs Lisp, a symbol can have two | |
1781 | other things attached to it, too: a property list and a documentation | |
1782 | string; these are discussed later.) | |
1783 | @end ignore | |
1784 | ||
1785 | Another way to think about this is to imagine a symbol as being a chest | |
1786 | of drawers. The function definition is put in one drawer, the value in | |
1787 | another, and so on. What is put in the drawer holding the value can be | |
1788 | changed without affecting the contents of the drawer holding the | |
1789 | function definition, and vice-versa. | |
1790 | ||
1791 | @menu | |
1792 | * fill-column Example:: | |
1793 | * Void Function:: The error message for a symbol | |
1794 | without a function. | |
1795 | * Void Variable:: The error message for a symbol without a value. | |
1796 | @end menu | |
1797 | ||
1798 | @node fill-column Example, Void Function, Variables, Variables | |
1799 | @ifnottex | |
1800 | @unnumberedsubsec @code{fill-column}, an Example Variable | |
1801 | @end ifnottex | |
1802 | ||
1803 | @findex fill-column, @r{an example variable} | |
1804 | @cindex Example variable, @code{fill-column} | |
1805 | @cindex Variable, example of, @code{fill-column} | |
1806 | The variable @code{fill-column} illustrates a symbol with a value | |
1807 | attached to it: in every GNU Emacs buffer, this symbol is set to some | |
1808 | value, usually 72 or 70, but sometimes to some other value. To find the | |
1809 | value of this symbol, evaluate it by itself. If you are reading this in | |
1810 | Info inside of GNU Emacs, you can do this by putting the cursor after | |
1811 | the symbol and typing @kbd{C-x C-e}: | |
1812 | ||
1813 | @smallexample | |
1814 | fill-column | |
1815 | @end smallexample | |
1816 | ||
1817 | @noindent | |
1818 | After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo | |
1819 | area. This is the value for which @code{fill-column} is set for me as I | |
1820 | write this. It may be different for you in your Info buffer. Notice | |
1821 | that the value returned as a variable is printed in exactly the same way | |
1822 | as the value returned by a function carrying out its instructions. From | |
1823 | the point of view of the Lisp interpreter, a value returned is a value | |
1824 | returned. What kind of expression it came from ceases to matter once | |
1825 | the value is known. | |
1826 | ||
1827 | A symbol can have any value attached to it or, to use the jargon, we can | |
1828 | @dfn{bind} the variable to a value: to a number, such as 72; to a | |
1829 | string, @code{"such as this"}; to a list, such as @code{(spruce pine | |
1830 | oak)}; we can even bind a variable to a function definition. | |
1831 | ||
1832 | A symbol can be bound to a value in several ways. @xref{set & setq, , | |
1833 | Setting the Value of a Variable}, for information about one way to do | |
1834 | this. | |
1835 | ||
1836 | @node Void Function, Void Variable, fill-column Example, Variables | |
1837 | @comment node-name, next, previous, up | |
1838 | @subsection Error Message for a Symbol Without a Function | |
1839 | @cindex Symbol without function error | |
1840 | @cindex Error for symbol without function | |
1841 | ||
1842 | When we evaluated @code{fill-column} to find its value as a variable, | |
1843 | we did not place parentheses around the word. This is because we did | |
1844 | not intend to use it as a function name. | |
1845 | ||
1846 | If @code{fill-column} were the first or only element of a list, the | |
1847 | Lisp interpreter would attempt to find the function definition | |
1848 | attached to it. But @code{fill-column} has no function definition. | |
1849 | Try evaluating this: | |
1850 | ||
1851 | @smallexample | |
1852 | (fill-column) | |
1853 | @end smallexample | |
1854 | ||
1855 | @need 1250 | |
1856 | @noindent | |
1857 | In GNU Emacs version 21, you will create a @file{*Backtrace*} buffer | |
1858 | that says: | |
1859 | ||
1860 | @smallexample | |
1861 | @group | |
1862 | ---------- Buffer: *Backtrace* ---------- | |
1863 | Debugger entered--Lisp error: (void-function fill-column) | |
1864 | (fill-column) | |
1865 | eval((fill-column)) | |
1866 | eval-last-sexp-1(nil) | |
1867 | eval-last-sexp(nil) | |
1868 | call-interactively(eval-last-sexp) | |
1869 | ---------- Buffer: *Backtrace* ---------- | |
1870 | @end group | |
1871 | @end smallexample | |
1872 | ||
1873 | @noindent | |
1874 | (Remember, to quit the debugger and make the debugger window go away, | |
1875 | type @kbd{q} in the @file{*Backtrace*} buffer.) | |
1876 | ||
1877 | @need 800 | |
1878 | In GNU Emacs 20 and before, you will produce an error message that says: | |
1879 | ||
1880 | @smallexample | |
1881 | Symbol's function definition is void:@: fill-column | |
1882 | @end smallexample | |
1883 | ||
1884 | @noindent | |
1885 | (The message will go away away as soon as you move the cursor or type | |
1886 | another key.) | |
1887 | ||
1888 | @node Void Variable, , Void Function, Variables | |
1889 | @comment node-name, next, previous, up | |
1890 | @subsection Error Message for a Symbol Without a Value | |
1891 | @cindex Symbol without value error | |
1892 | @cindex Error for symbol without value | |
1893 | ||
1894 | If you attempt to evaluate a symbol that does not have a value bound to | |
1895 | it, you will receive an error message. You can see this by | |
1896 | experimenting with our 2 plus 2 addition. In the following expression, | |
1897 | put your cursor right after the @code{+}, before the first number 2, | |
1898 | type @kbd{C-x C-e}: | |
1899 | ||
1900 | @smallexample | |
1901 | (+ 2 2) | |
1902 | @end smallexample | |
1903 | ||
1904 | @need 1500 | |
1905 | @noindent | |
1906 | In GNU Emacs 21, you will create a @file{*Backtrace*} buffer that | |
1907 | says: | |
1908 | ||
1909 | @smallexample | |
1910 | @group | |
1911 | ---------- Buffer: *Backtrace* ---------- | |
1912 | Debugger entered--Lisp error: (void-variable +) | |
1913 | eval(+) | |
1914 | eval-last-sexp-1(nil) | |
1915 | eval-last-sexp(nil) | |
1916 | call-interactively(eval-last-sexp) | |
1917 | ---------- Buffer: *Backtrace* ---------- | |
1918 | @end group | |
1919 | @end smallexample | |
1920 | ||
1921 | @noindent | |
1922 | (As with the other times we entered the debugger, you can quit by | |
1923 | typing @kbd{q} in the @file{*Backtrace*} buffer.) | |
1924 | ||
1925 | This backtrace is different from the very first error message we saw, | |
1926 | which said, @samp{Debugger entered--Lisp error: (void-function this)}. | |
1927 | In this case, the function does not have a value as a variable; while | |
1928 | in the other error message, the function (the word `this') did not | |
1929 | have a definition. | |
1930 | ||
1931 | In this experiment with the @code{+}, what we did was cause the Lisp | |
1932 | interpreter to evaluate the @code{+} and look for the value of the | |
1933 | variable instead of the function definition. We did this by placing the | |
1934 | cursor right after the symbol rather than after the parenthesis of the | |
1935 | enclosing list as we did before. As a consequence, the Lisp interpreter | |
1936 | evaluated the preceding s-expression, which in this case was the | |
1937 | @code{+} by itself. | |
1938 | ||
1939 | Since @code{+} does not have a value bound to it, just the function | |
1940 | definition, the error message reported that the symbol's value as a | |
1941 | variable was void. | |
1942 | ||
1943 | @need 800 | |
1944 | In GNU Emacs version 20 and before, your error message will say: | |
1945 | ||
1946 | @example | |
1947 | Symbol's value as variable is void:@: + | |
1948 | @end example | |
1949 | ||
1950 | @noindent | |
1951 | The meaning is the same as in GNU Emacs 21. | |
1952 | ||
1953 | @node Arguments, set & setq, Variables, List Processing | |
1954 | @comment node-name, next, previous, up | |
1955 | @section Arguments | |
1956 | @cindex Arguments | |
1957 | @cindex Passing information to functions | |
1958 | ||
1959 | To see how information is passed to functions, let's look again at | |
1960 | our old standby, the addition of two plus two. In Lisp, this is written | |
1961 | as follows: | |
1962 | ||
1963 | @smallexample | |
1964 | (+ 2 2) | |
1965 | @end smallexample | |
1966 | ||
1967 | If you evaluate this expression, the number 4 will appear in your echo | |
1968 | area. What the Lisp interpreter does is add the numbers that follow | |
1969 | the @code{+}. | |
1970 | ||
1971 | @cindex @samp{argument} defined | |
1972 | The numbers added by @code{+} are called the @dfn{arguments} of the | |
1973 | function @code{+}. These numbers are the information that is given to | |
1974 | or @dfn{passed} to the function. | |
1975 | ||
1976 | The word `argument' comes from the way it is used in mathematics and | |
1977 | does not refer to a disputation between two people; instead it refers to | |
1978 | the information presented to the function, in this case, to the | |
1979 | @code{+}. In Lisp, the arguments to a function are the atoms or lists | |
1980 | that follow the function. The values returned by the evaluation of | |
1981 | these atoms or lists are passed to the function. Different functions | |
1982 | require different numbers of arguments; some functions require none at | |
1983 | all.@footnote{It is curious to track the path by which the word `argument' | |
1984 | came to have two different meanings, one in mathematics and the other in | |
1985 | everyday English. According to the @cite{Oxford English Dictionary}, | |
1986 | the word derives from the Latin for @samp{to make clear, prove}; thus it | |
1987 | came to mean, by one thread of derivation, `the evidence offered as | |
1988 | proof', which is to say, `the information offered', which led to its | |
1989 | meaning in Lisp. But in the other thread of derivation, it came to mean | |
1990 | `to assert in a manner against which others may make counter | |
1991 | assertions', which led to the meaning of the word as a disputation. | |
1992 | (Note here that the English word has two different definitions attached | |
1993 | to it at the same time. By contrast, in Emacs Lisp, a symbol cannot | |
1994 | have two different function definitions at the same time.)} | |
1995 | ||
1996 | @menu | |
1997 | * Data types:: Types of data passed to a function. | |
1998 | * Args as Variable or List:: An argument can be the value | |
1999 | of a variable or list. | |
2000 | * Variable Number of Arguments:: Some functions may take a | |
2001 | variable number of arguments. | |
2002 | * Wrong Type of Argument:: Passing an argument of the wrong type | |
2003 | to a function. | |
2004 | * message:: A useful function for sending messages. | |
2005 | @end menu | |
2006 | ||
2007 | @node Data types, Args as Variable or List, Arguments, Arguments | |
2008 | @comment node-name, next, previous, up | |
2009 | @subsection Arguments' Data Types | |
2010 | @cindex Data types | |
2011 | @cindex Types of data | |
2012 | @cindex Arguments' data types | |
2013 | ||
2014 | The type of data that should be passed to a function depends on what | |
2015 | kind of information it uses. The arguments to a function such as | |
2016 | @code{+} must have values that are numbers, since @code{+} adds numbers. | |
2017 | Other functions use different kinds of data for their arguments. | |
2018 | ||
2019 | @findex concat | |
2020 | For example, the @code{concat} function links together or unites two or | |
2021 | more strings of text to produce a string. The arguments are strings. | |
2022 | Concatenating the two character strings @code{abc}, @code{def} produces | |
2023 | the single string @code{abcdef}. This can be seen by evaluating the | |
2024 | following: | |
2025 | ||
2026 | @smallexample | |
2027 | (concat "abc" "def") | |
2028 | @end smallexample | |
2029 | ||
2030 | @noindent | |
2031 | The value produced by evaluating this expression is @code{"abcdef"}. | |
2032 | ||
2033 | A function such as @code{substring} uses both a string and numbers as | |
2034 | arguments. The function returns a part of the string, a substring of | |
2035 | the first argument. This function takes three arguments. Its first | |
2036 | argument is the string of characters, the second and third arguments are | |
2037 | numbers that indicate the beginning and end of the substring. The | |
2038 | numbers are a count of the number of characters (including spaces and | |
2039 | punctuations) from the beginning of the string. | |
2040 | ||
2041 | @need 800 | |
2042 | For example, if you evaluate the following: | |
2043 | ||
2044 | @smallexample | |
2045 | (substring "The quick brown fox jumped." 16 19) | |
2046 | @end smallexample | |
2047 | ||
2048 | @noindent | |
2049 | you will see @code{"fox"} appear in the echo area. The arguments are the | |
2050 | string and the two numbers. | |
2051 | ||
2052 | Note that the string passed to @code{substring} is a single atom even | |
2053 | though it is made up of several words separated by spaces. Lisp counts | |
2054 | everything between the two quotation marks as part of the string, | |
2055 | including the spaces. You can think of the @code{substring} function as | |
2056 | a kind of `atom smasher' since it takes an otherwise indivisible atom | |
2057 | and extracts a part. However, @code{substring} is only able to extract | |
2058 | a substring from an argument that is a string, not from another type of | |
2059 | atom such as a number or symbol. | |
2060 | ||
2061 | @node Args as Variable or List, Variable Number of Arguments, Data types, Arguments | |
2062 | @comment node-name, next, previous, up | |
2063 | @subsection An Argument as the Value of a Variable or List | |
2064 | ||
2065 | An argument can be a symbol that returns a value when it is evaluated. | |
2066 | For example, when the symbol @code{fill-column} by itself is evaluated, | |
2067 | it returns a number. This number can be used in an addition. | |
2068 | ||
2069 | @need 1250 | |
2070 | Position the cursor after the following expression and type @kbd{C-x | |
2071 | C-e}: | |
2072 | ||
2073 | @smallexample | |
2074 | (+ 2 fill-column) | |
2075 | @end smallexample | |
2076 | ||
2077 | @noindent | |
2078 | The value will be a number two more than what you get by evaluating | |
2079 | @code{fill-column} alone. For me, this is 74, because the value of | |
2080 | @code{fill-column} is 72. | |
2081 | ||
2082 | As we have just seen, an argument can be a symbol that returns a value | |
2083 | when evaluated. In addition, an argument can be a list that returns a | |
2084 | value when it is evaluated. For example, in the following expression, | |
2085 | the arguments to the function @code{concat} are the strings | |
2086 | @w{@code{"The "}} and @w{@code{" red foxes."}} and the list | |
2087 | @code{(number-to-string (+ 2 fill-column))}. | |
2088 | ||
2089 | @c For Emacs 21, need number-to-string | |
2090 | @smallexample | |
2091 | (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.") | |
2092 | @end smallexample | |
2093 | ||
2094 | @noindent | |
2095 | If you evaluate this expression---and if, as with my Emacs, | |
2096 | @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will | |
2097 | appear in the echo area. (Note that you must put spaces after the | |
2098 | word @samp{The} and before the word @samp{red} so they will appear in | |
2099 | the final string. The function @code{number-to-string} converts the | |
2100 | integer that the addition function returns to a string. | |
2101 | @code{number-to-string} is also known as @code{int-to-string}.) | |
2102 | ||
2103 | @node Variable Number of Arguments, Wrong Type of Argument, Args as Variable or List, Arguments | |
2104 | @comment node-name, next, previous, up | |
2105 | @subsection Variable Number of Arguments | |
2106 | @cindex Variable number of arguments | |
2107 | @cindex Arguments, variable number of | |
2108 | ||
2109 | Some functions, such as @code{concat}, @code{+} or @code{*}, take any | |
2110 | number of arguments. (The @code{*} is the symbol for multiplication.) | |
2111 | This can be seen by evaluating each of the following expressions in | |
2112 | the usual way. What you will see in the echo area is printed in this | |
2113 | text after @samp{@result{}}, which you may read as `evaluates to'. | |
2114 | ||
2115 | @need 1250 | |
2116 | In the first set, the functions have no arguments: | |
2117 | ||
2118 | @smallexample | |
2119 | @group | |
2120 | (+) @result{} 0 | |
2121 | ||
2122 | (*) @result{} 1 | |
2123 | @end group | |
2124 | @end smallexample | |
2125 | ||
2126 | @need 1250 | |
2127 | In this set, the functions have one argument each: | |
2128 | ||
2129 | @smallexample | |
2130 | @group | |
2131 | (+ 3) @result{} 3 | |
2132 | ||
2133 | (* 3) @result{} 3 | |
2134 | @end group | |
2135 | @end smallexample | |
2136 | ||
2137 | @need 1250 | |
2138 | In this set, the functions have three arguments each: | |
2139 | ||
2140 | @smallexample | |
2141 | @group | |
2142 | (+ 3 4 5) @result{} 12 | |
2143 | ||
2144 | (* 3 4 5) @result{} 60 | |
2145 | @end group | |
2146 | @end smallexample | |
2147 | ||
2148 | @node Wrong Type of Argument, message, Variable Number of Arguments, Arguments | |
2149 | @comment node-name, next, previous, up | |
2150 | @subsection Using the Wrong Type Object as an Argument | |
2151 | @cindex Wrong type of argument | |
2152 | @cindex Argument, wrong type of | |
2153 | ||
2154 | When a function is passed an argument of the wrong type, the Lisp | |
2155 | interpreter produces an error message. For example, the @code{+} | |
2156 | function expects the values of its arguments to be numbers. As an | |
2157 | experiment we can pass it the quoted symbol @code{hello} instead of a | |
2158 | number. Position the cursor after the following expression and type | |
2159 | @kbd{C-x C-e}: | |
2160 | ||
2161 | @smallexample | |
2162 | (+ 2 'hello) | |
2163 | @end smallexample | |
2164 | ||
2165 | @noindent | |
2166 | When you do this you will generate an error message. What has happened | |
2167 | is that @code{+} has tried to add the 2 to the value returned by | |
2168 | @code{'hello}, but the value returned by @code{'hello} is the symbol | |
2169 | @code{hello}, not a number. Only numbers can be added. So @code{+} | |
2170 | could not carry out its addition. | |
2171 | ||
2172 | @need 1250 | |
2173 | In GNU Emacs version 21, you will create and enter a | |
2174 | @file{*Backtrace*} buffer that says: | |
2175 | ||
2176 | @noindent | |
2177 | @smallexample | |
2178 | @group | |
2179 | ---------- Buffer: *Backtrace* ---------- | |
2180 | Debugger entered--Lisp error: | |
2181 | (wrong-type-argument number-or-marker-p hello) | |
2182 | +(2 hello) | |
2183 | eval((+ 2 (quote hello))) | |
2184 | eval-last-sexp-1(nil) | |
2185 | eval-last-sexp(nil) | |
2186 | call-interactively(eval-last-sexp) | |
2187 | ---------- Buffer: *Backtrace* ---------- | |
2188 | @end group | |
2189 | @end smallexample | |
2190 | ||
2191 | @need 1250 | |
2192 | As usual, the error message tries to be helpful and makes sense after you | |
2193 | learn how to read it. | |
2194 | ||
2195 | The first part of the error message is straightforward; it says | |
2196 | @samp{wrong type argument}. Next comes the mysterious jargon word | |
2197 | @w{@samp{number-or-marker-p}}. This word is trying to tell you what | |
2198 | kind of argument the @code{+} expected. | |
2199 | ||
2200 | The symbol @code{number-or-marker-p} says that the Lisp interpreter is | |
2201 | trying to determine whether the information presented it (the value of | |
2202 | the argument) is a number or a marker (a special object representing a | |
2203 | buffer position). What it does is test to see whether the @code{+} is | |
2204 | being given numbers to add. It also tests to see whether the | |
2205 | argument is something called a marker, which is a specific feature of | |
2206 | Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers. | |
2207 | When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command, | |
2208 | its position is kept as a marker. The mark can be considered a | |
2209 | number---the number of characters the location is from the beginning | |
2210 | of the buffer.) In Emacs Lisp, @code{+} can be used to add the | |
2211 | numeric value of marker positions as numbers. | |
2212 | ||
2213 | The @samp{p} of @code{number-or-marker-p} is the embodiment of a | |
2214 | practice started in the early days of Lisp programming. The @samp{p} | |
2215 | stands for `predicate'. In the jargon used by the early Lisp | |
2216 | researchers, a predicate refers to a function to determine whether some | |
2217 | property is true or false. So the @samp{p} tells us that | |
2218 | @code{number-or-marker-p} is the name of a function that determines | |
2219 | whether it is true or false that the argument supplied is a number or | |
2220 | a marker. Other Lisp symbols that end in @samp{p} include @code{zerop}, | |
2221 | a function that tests whether its argument has the value of zero, and | |
2222 | @code{listp}, a function that tests whether its argument is a list. | |
2223 | ||
2224 | Finally, the last part of the error message is the symbol @code{hello}. | |
2225 | This is the value of the argument that was passed to @code{+}. If the | |
2226 | addition had been passed the correct type of object, the value passed | |
2227 | would have been a number, such as 37, rather than a symbol like | |
2228 | @code{hello}. But then you would not have got the error message. | |
2229 | ||
2230 | @need 1250 | |
2231 | In GNU Emacs version 20 and before, the echo area displays an error | |
2232 | message that says: | |
2233 | ||
2234 | @smallexample | |
2235 | Wrong type argument:@: number-or-marker-p, hello | |
2236 | @end smallexample | |
2237 | ||
2238 | This says, in different words, the same as the top line of the | |
2239 | @file{*Backtrace*} buffer. | |
2240 | ||
2241 | @node message, , Wrong Type of Argument, Arguments | |
2242 | @comment node-name, next, previous, up | |
2243 | @subsection The @code{message} Function | |
2244 | @findex message | |
2245 | ||
2246 | Like @code{+}, the @code{message} function takes a variable number of | |
2247 | arguments. It is used to send messages to the user and is so useful | |
2248 | that we will describe it here. | |
2249 | ||
2250 | @need 1250 | |
2251 | A message is printed in the echo area. For example, you can print a | |
2252 | message in your echo area by evaluating the following list: | |
2253 | ||
2254 | @smallexample | |
2255 | (message "This message appears in the echo area!") | |
2256 | @end smallexample | |
2257 | ||
2258 | The whole string between double quotation marks is a single argument | |
2259 | and is printed @i{in toto}. (Note that in this example, the message | |
2260 | itself will appear in the echo area within double quotes; that is | |
2261 | because you see the value returned by the @code{message} function. In | |
2262 | most uses of @code{message} in programs that you write, the text will | |
2263 | be printed in the echo area as a side-effect, without the quotes. | |
2264 | @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in | |
2265 | detail}, for an example of this.) | |
2266 | ||
2267 | However, if there is a @samp{%s} in the quoted string of characters, the | |
2268 | @code{message} function does not print the @samp{%s} as such, but looks | |
2269 | to the argument that follows the string. It evaluates the second | |
2270 | argument and prints the value at the location in the string where the | |
2271 | @samp{%s} is. | |
2272 | ||
2273 | @need 1250 | |
2274 | You can see this by positioning the cursor after the following | |
2275 | expression and typing @kbd{C-x C-e}: | |
2276 | ||
2277 | @smallexample | |
2278 | (message "The name of this buffer is: %s." (buffer-name)) | |
2279 | @end smallexample | |
2280 | ||
2281 | @noindent | |
2282 | In Info, @code{"The name of this buffer is: *info*."} will appear in the | |
2283 | echo area. The function @code{buffer-name} returns the name of the | |
2284 | buffer as a string, which the @code{message} function inserts in place | |
2285 | of @code{%s}. | |
2286 | ||
2287 | To print a value as an integer, use @samp{%d} in the same way as | |
2288 | @samp{%s}. For example, to print a message in the echo area that | |
2289 | states the value of the @code{fill-column}, evaluate the following: | |
2290 | ||
2291 | @smallexample | |
2292 | (message "The value of fill-column is %d." fill-column) | |
2293 | @end smallexample | |
2294 | ||
2295 | @noindent | |
2296 | On my system, when I evaluate this list, @code{"The value of | |
2297 | fill-column is 72."} appears in my echo area@footnote{Actually, you | |
2298 | can use @code{%s} to print a number. It is non-specific. @code{%d} | |
2299 | prints only the part of a number left of a decimal point, and not | |
2300 | anything that is not a number.}. | |
2301 | ||
2302 | If there is more than one @samp{%s} in the quoted string, the value of | |
2303 | the first argument following the quoted string is printed at the | |
2304 | location of the first @samp{%s} and the value of the second argument is | |
2305 | printed at the location of the second @samp{%s}, and so on. | |
2306 | ||
2307 | @need 1250 | |
2308 | For example, if you evaluate the following, | |
2309 | ||
2310 | @smallexample | |
2311 | @group | |
2312 | (message "There are %d %s in the office!" | |
2313 | (- fill-column 14) "pink elephants") | |
2314 | @end group | |
2315 | @end smallexample | |
2316 | ||
2317 | @noindent | |
2318 | a rather whimsical message will appear in your echo area. On my system | |
2319 | it says, @code{"There are 58 pink elephants in the office!"}. | |
2320 | ||
2321 | The expression @code{(- fill-column 14)} is evaluated and the resulting | |
2322 | number is inserted in place of the @samp{%d}; and the string in double | |
2323 | quotes, @code{"pink elephants"}, is treated as a single argument and | |
2324 | inserted in place of the @samp{%s}. (That is to say, a string between | |
2325 | double quotes evaluates to itself, like a number.) | |
2326 | ||
2327 | Finally, here is a somewhat complex example that not only illustrates | |
2328 | the computation of a number, but also shows how you can use an | |
2329 | expression within an expression to generate the text that is substituted | |
2330 | for @samp{%s}: | |
2331 | ||
2332 | @smallexample | |
2333 | @group | |
2334 | (message "He saw %d %s" | |
2335 | (- fill-column 34) | |
2336 | (concat "red " | |
2337 | (substring | |
2338 | "The quick brown foxes jumped." 16 21) | |
2339 | " leaping.")) | |
2340 | @end group | |
2341 | @end smallexample | |
2342 | ||
2343 | In this example, @code{message} has three arguments: the string, | |
2344 | @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and | |
2345 | the expression beginning with the function @code{concat}. The value | |
2346 | resulting from the evaluation of @code{(- fill-column 32)} is inserted | |
2347 | in place of the @samp{%d}; and the value returned by the expression | |
2348 | beginning with @code{concat} is inserted in place of the @samp{%s}. | |
2349 | ||
2350 | When I evaluate the expression, the message @code{"He saw 38 red | |
2351 | foxes leaping."} appears in my echo area. | |
2352 | ||
2353 | @node set & setq, Summary, Arguments, List Processing | |
2354 | @comment node-name, next, previous, up | |
2355 | @section Setting the Value of a Variable | |
2356 | @cindex Variable, setting value | |
2357 | @cindex Setting value of variable | |
2358 | ||
2359 | @cindex @samp{bind} defined | |
2360 | There are several ways by which a variable can be given a value. One of | |
2361 | the ways is to use either the function @code{set} or the function | |
2362 | @code{setq}. Another way is to use @code{let} (@pxref{let}). (The | |
2363 | jargon for this process is to @dfn{bind} a variable to a value.) | |
2364 | ||
2365 | The following sections not only describe how @code{set} and @code{setq} | |
2366 | work but also illustrate how arguments are passed. | |
2367 | ||
2368 | @menu | |
2369 | * Using set:: Setting values. | |
2370 | * Using setq:: Setting a quoted value. | |
2371 | * Counting:: Using @code{setq} to count. | |
2372 | @end menu | |
2373 | ||
2374 | @node Using set, Using setq, set & setq, set & setq | |
2375 | @comment node-name, next, previous, up | |
2376 | @subsection Using @code{set} | |
2377 | @findex set | |
2378 | ||
2379 | To set the value of the symbol @code{flowers} to the list @code{'(rose | |
2380 | violet daisy buttercup)}, evaluate the following expression by | |
2381 | positioning the cursor after the expression and typing @kbd{C-x C-e}. | |
2382 | ||
2383 | @smallexample | |
2384 | (set 'flowers '(rose violet daisy buttercup)) | |
2385 | @end smallexample | |
2386 | ||
2387 | @noindent | |
2388 | The list @code{(rose violet daisy buttercup)} will appear in the echo | |
2389 | area. This is what is @emph{returned} by the @code{set} function. As a | |
2390 | side effect, the symbol @code{flowers} is bound to the list ; that is, | |
2391 | the symbol @code{flowers}, which can be viewed as a variable, is given | |
2392 | the list as its value. (This process, by the way, illustrates how a | |
2393 | side effect to the Lisp interpreter, setting the value, can be the | |
2394 | primary effect that we humans are interested in. This is because every | |
2395 | Lisp function must return a value if it does not get an error, but it | |
2396 | will only have a side effect if it is designed to have one.) | |
2397 | ||
2398 | After evaluating the @code{set} expression, you can evaluate the symbol | |
2399 | @code{flowers} and it will return the value you just set. Here is the | |
2400 | symbol. Place your cursor after it and type @kbd{C-x C-e}. | |
2401 | ||
2402 | @smallexample | |
2403 | flowers | |
2404 | @end smallexample | |
2405 | ||
2406 | @noindent | |
2407 | When you evaluate @code{flowers}, the list | |
2408 | @code{(rose violet daisy buttercup)} appears in the echo area. | |
2409 | ||
2410 | Incidentally, if you evaluate @code{'flowers}, the variable with a quote | |
2411 | in front of it, what you will see in the echo area is the symbol itself, | |
2412 | @code{flowers}. Here is the quoted symbol, so you can try this: | |
2413 | ||
2414 | @smallexample | |
2415 | 'flowers | |
2416 | @end smallexample | |
2417 | ||
2418 | Note also, that when you use @code{set}, you need to quote both | |
2419 | arguments to @code{set}, unless you want them evaluated. Since we do | |
2420 | not want either argument evaluated, neither the variable | |
2421 | @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both | |
2422 | are quoted. (When you use @code{set} without quoting its first | |
2423 | argument, the first argument is evaluated before anything else is | |
2424 | done. If you did this and @code{flowers} did not have a value | |
2425 | already, you would get an error message that the @samp{Symbol's value | |
2426 | as variable is void}; on the other hand, if @code{flowers} did return | |
2427 | a value after it was evaluated, the @code{set} would attempt to set | |
2428 | the value that was returned. There are situations where this is the | |
2429 | right thing for the function to do; but such situations are rare.) | |
2430 | ||
2431 | @node Using setq, Counting, Using set, set & setq | |
2432 | @comment node-name, next, previous, up | |
2433 | @subsection Using @code{setq} | |
2434 | @findex setq | |
2435 | ||
2436 | As a practical matter, you almost always quote the first argument to | |
2437 | @code{set}. The combination of @code{set} and a quoted first argument | |
2438 | is so common that it has its own name: the special form @code{setq}. | |
2439 | This special form is just like @code{set} except that the first argument | |
2440 | is quoted automatically, so you don't need to type the quote mark | |
2441 | yourself. Also, as an added convenience, @code{setq} permits you to set | |
2442 | several different variables to different values, all in one expression. | |
2443 | ||
2444 | To set the value of the variable @code{carnivores} to the list | |
2445 | @code{'(lion tiger leopard)} using @code{setq}, the following expression | |
2446 | is used: | |
2447 | ||
2448 | @smallexample | |
2449 | (setq carnivores '(lion tiger leopard)) | |
2450 | @end smallexample | |
2451 | ||
2452 | @noindent | |
2453 | This is exactly the same as using @code{set} except the first argument | |
2454 | is automatically quoted by @code{setq}. (The @samp{q} in @code{setq} | |
2455 | means @code{quote}.) | |
2456 | ||
2457 | @need 1250 | |
2458 | With @code{set}, the expression would look like this: | |
2459 | ||
2460 | @smallexample | |
2461 | (set 'carnivores '(lion tiger leopard)) | |
2462 | @end smallexample | |
2463 | ||
2464 | Also, @code{setq} can be used to assign different values to | |
2465 | different variables. The first argument is bound to the value | |
2466 | of the second argument, the third argument is bound to the value of the | |
2467 | fourth argument, and so on. For example, you could use the following to | |
2468 | assign a list of trees to the symbol @code{trees} and a list of herbivores | |
2469 | to the symbol @code{herbivores}: | |
2470 | ||
2471 | @smallexample | |
2472 | @group | |
2473 | (setq trees '(pine fir oak maple) | |
2474 | herbivores '(gazelle antelope zebra)) | |
2475 | @end group | |
2476 | @end smallexample | |
2477 | ||
2478 | @noindent | |
2479 | (The expression could just as well have been on one line, but it might | |
2480 | not have fit on a page; and humans find it easier to read nicely | |
2481 | formatted lists.) | |
2482 | ||
2483 | Although I have been using the term `assign', there is another way of | |
2484 | thinking about the workings of @code{set} and @code{setq}; and that is to | |
2485 | say that @code{set} and @code{setq} make the symbol @emph{point} to the | |
2486 | list. This latter way of thinking is very common and in forthcoming | |
2487 | chapters we shall come upon at least one symbol that has `pointer' as | |
2488 | part of its name. The name is chosen because the symbol has a value, | |
2489 | specifically a list, attached to it; or, expressed another way, | |
2490 | the symbol is set to ``point'' to the list. | |
2491 | ||
2492 | @node Counting, , Using setq, set & setq | |
2493 | @comment node-name, next, previous, up | |
2494 | @subsection Counting | |
2495 | @cindex Counting | |
2496 | ||
2497 | Here is an example that shows how to use @code{setq} in a counter. You | |
2498 | might use this to count how many times a part of your program repeats | |
2499 | itself. First set a variable to zero; then add one to the number each | |
2500 | time the program repeats itself. To do this, you need a variable that | |
2501 | serves as a counter, and two expressions: an initial @code{setq} | |
2502 | expression that sets the counter variable to zero; and a second | |
2503 | @code{setq} expression that increments the counter each time it is | |
2504 | evaluated. | |
2505 | ||
2506 | @smallexample | |
2507 | @group | |
2508 | (setq counter 0) ; @r{Let's call this the initializer.} | |
2509 | ||
2510 | (setq counter (+ counter 1)) ; @r{This is the incrementer.} | |
2511 | ||
2512 | counter ; @r{This is the counter.} | |
2513 | @end group | |
2514 | @end smallexample | |
2515 | ||
2516 | @noindent | |
2517 | (The text following the @samp{;} are comments. @xref{Change a | |
2518 | defun, , Change a Function Definition}.) | |
2519 | ||
2520 | If you evaluate the first of these expressions, the initializer, | |
2521 | @code{(setq counter 0)}, and then evaluate the third expression, | |
2522 | @code{counter}, the number @code{0} will appear in the echo area. If | |
2523 | you then evaluate the second expression, the incrementer, @code{(setq | |
2524 | counter (+ counter 1))}, the counter will get the value 1. So if you | |
2525 | again evaluate @code{counter}, the number @code{1} will appear in the | |
2526 | echo area. Each time you evaluate the second expression, the value of | |
2527 | the counter will be incremented. | |
2528 | ||
2529 | When you evaluate the incrementer, @code{(setq counter (+ counter 1))}, | |
2530 | the Lisp interpreter first evaluates the innermost list; this is the | |
2531 | addition. In order to evaluate this list, it must evaluate the variable | |
2532 | @code{counter} and the number @code{1}. When it evaluates the variable | |
2533 | @code{counter}, it receives its current value. It passes this value and | |
2534 | the number @code{1} to the @code{+} which adds them together. The sum | |
2535 | is then returned as the value of the inner list and passed to the | |
2536 | @code{setq} which sets the variable @code{counter} to this new value. | |
2537 | Thus, the value of the variable, @code{counter}, is changed. | |
2538 | ||
2539 | @node Summary, Error Message Exercises, set & setq, List Processing | |
2540 | @comment node-name, next, previous, up | |
2541 | @section Summary | |
2542 | ||
2543 | Learning Lisp is like climbing a hill in which the first part is the | |
2544 | steepest. You have now climbed the most difficult part; what remains | |
2545 | becomes easier as you progress onwards. | |
2546 | ||
2547 | In summary, | |
2548 | ||
2549 | @itemize @bullet | |
2550 | ||
2551 | @item | |
2552 | Lisp programs are made up of expressions, which are lists or single atoms. | |
2553 | ||
2554 | @item | |
2555 | Lists are made up of zero or more atoms or inner lists, separated by whitespace and | |
2556 | surrounded by parentheses. A list can be empty. | |
2557 | ||
2558 | @item | |
2559 | Atoms are multi-character symbols, like @code{forward-paragraph}, single | |
2560 | character symbols like @code{+}, strings of characters between double | |
2561 | quotation marks, or numbers. | |
2562 | ||
2563 | @item | |
2564 | A number evaluates to itself. | |
2565 | ||
2566 | @item | |
2567 | A string between double quotes also evaluates to itself. | |
2568 | ||
2569 | @item | |
2570 | When you evaluate a symbol by itself, its value is returned. | |
2571 | ||
2572 | @item | |
2573 | When you evaluate a list, the Lisp interpreter looks at the first symbol | |
2574 | in the list and then at the function definition bound to that symbol. | |
2575 | Then the instructions in the function definition are carried out. | |
2576 | ||
2577 | @item | |
2578 | A single-quote, @code{'}, tells the Lisp interpreter that it should | |
2579 | return the following expression as written, and not evaluate it as it | |
2580 | would if the quote were not there. | |
2581 | ||
2582 | @item | |
2583 | Arguments are the information passed to a function. The arguments to a | |
2584 | function are computed by evaluating the rest of the elements of the list | |
2585 | of which the function is the first element. | |
2586 | ||
2587 | @item | |
2588 | A function always returns a value when it is evaluated (unless it gets | |
2589 | an error); in addition, it may also carry out some action called a | |
2590 | ``side effect''. In many cases, a function's primary purpose is to | |
2591 | create a side effect. | |
2592 | @end itemize | |
2593 | ||
2594 | @node Error Message Exercises, , Summary, List Processing | |
2595 | @comment node-name, next, previous, up | |
2596 | @section Exercises | |
2597 | ||
2598 | A few simple exercises: | |
2599 | ||
2600 | @itemize @bullet | |
2601 | @item | |
2602 | Generate an error message by evaluating an appropriate symbol that is | |
2603 | not within parentheses. | |
2604 | ||
2605 | @item | |
2606 | Generate an error message by evaluating an appropriate symbol that is | |
2607 | between parentheses. | |
2608 | ||
2609 | @item | |
2610 | Create a counter that increments by two rather than one. | |
2611 | ||
2612 | @item | |
2613 | Write an expression that prints a message in the echo area when | |
2614 | evaluated. | |
2615 | @end itemize | |
2616 | ||
2617 | @node Practicing Evaluation, Writing Defuns, List Processing, Top | |
2618 | @comment node-name, next, previous, up | |
2619 | @chapter Practicing Evaluation | |
2620 | @cindex Practicing evaluation | |
2621 | @cindex Evaluation practice | |
2622 | ||
2623 | Before learning how to write a function definition in Emacs Lisp, it is | |
2624 | useful to spend a little time evaluating various expressions that have | |
2625 | already been written. These expressions will be lists with the | |
2626 | functions as their first (and often only) element. Since some of the | |
2627 | functions associated with buffers are both simple and interesting, we | |
2628 | will start with those. In this section, we will evaluate a few of | |
2629 | these. In another section, we will study the code of several other | |
2630 | buffer-related functions, to see how they were written. | |
2631 | ||
2632 | @menu | |
2633 | * How to Evaluate:: Typing editing commands or @kbd{C-x C-e} | |
2634 | causes evaluation. | |
2635 | * Buffer Names:: Buffers and files are different. | |
2636 | * Getting Buffers:: Getting a buffer itself, not merely its name. | |
2637 | * Switching Buffers:: How to change to another buffer. | |
2638 | * Buffer Size & Locations:: Where point is located and the size of | |
2639 | the buffer. | |
2640 | * Evaluation Exercise:: | |
2641 | @end menu | |
2642 | ||
2643 | @node How to Evaluate, Buffer Names, Practicing Evaluation, Practicing Evaluation | |
2644 | @ifnottex | |
2645 | @unnumberedsec How to Evaluate | |
2646 | @end ifnottex | |
2647 | ||
2648 | @i{Whenever you give an editing command} to Emacs Lisp, such as the | |
2649 | command to move the cursor or to scroll the screen, @i{you are evaluating | |
2650 | an expression,} the first element of which is a function. @i{This is | |
2651 | how Emacs works.} | |
2652 | ||
2653 | @cindex @samp{interactive function} defined | |
2654 | @cindex @samp{command} defined | |
2655 | When you type keys, you cause the Lisp interpreter to evaluate an | |
2656 | expression and that is how you get your results. Even typing plain text | |
2657 | involves evaluating an Emacs Lisp function, in this case, one that uses | |
2658 | @code{self-insert-command}, which simply inserts the character you | |
2659 | typed. The functions you evaluate by typing keystrokes are called | |
2660 | @dfn{interactive} functions, or @dfn{commands}; how you make a function | |
2661 | interactive will be illustrated in the chapter on how to write function | |
2662 | definitions. @xref{Interactive, , Making a Function Interactive}. | |
2663 | ||
2664 | In addition to typing keyboard commands, we have seen a second way to | |
2665 | evaluate an expression: by positioning the cursor after a list and | |
2666 | typing @kbd{C-x C-e}. This is what we will do in the rest of this | |
2667 | section. There are other ways to evaluate an expression as well; these | |
2668 | will be described as we come to them. | |
2669 | ||
2670 | Besides being used for practicing evaluation, the functions shown in the | |
2671 | next few sections are important in their own right. A study of these | |
2672 | functions makes clear the distinction between buffers and files, how to | |
2673 | switch to a buffer, and how to determine a location within it. | |
2674 | ||
2675 | @node Buffer Names, Getting Buffers, How to Evaluate, Practicing Evaluation | |
2676 | @comment node-name, next, previous, up | |
2677 | @section Buffer Names | |
2678 | @findex buffer-name | |
2679 | @findex buffer-file-name | |
2680 | ||
2681 | The two functions, @code{buffer-name} and @code{buffer-file-name}, show | |
2682 | the difference between a file and a buffer. When you evaluate the | |
2683 | following expression, @code{(buffer-name)}, the name of the buffer | |
2684 | appears in the echo area. When you evaluate @code{(buffer-file-name)}, | |
2685 | the name of the file to which the buffer refers appears in the echo | |
2686 | area. Usually, the name returned by @code{(buffer-name)} is the same as | |
2687 | the name of the file to which it refers, and the name returned by | |
2688 | @code{(buffer-file-name)} is the full path-name of the file. | |
2689 | ||
2690 | A file and a buffer are two different entities. A file is information | |
2691 | recorded permanently in the computer (unless you delete it). A buffer, | |
2692 | on the other hand, is information inside of Emacs that will vanish at | |
2693 | the end of the editing session (or when you kill the buffer). Usually, | |
2694 | a buffer contains information that you have copied from a file; we say | |
2695 | the buffer is @dfn{visiting} that file. This copy is what you work on | |
2696 | and modify. Changes to the buffer do not change the file, until you | |
2697 | save the buffer. When you save the buffer, the buffer is copied to the file | |
2698 | and is thus saved permanently. | |
2699 | ||
2700 | @need 1250 | |
2701 | If you are reading this in Info inside of GNU Emacs, you can evaluate | |
2702 | each of the following expressions by positioning the cursor after it and | |
2703 | typing @kbd{C-x C-e}. | |
2704 | ||
2705 | @smallexample | |
2706 | @group | |
2707 | (buffer-name) | |
2708 | ||
2709 | (buffer-file-name) | |
2710 | @end group | |
2711 | @end smallexample | |
2712 | ||
2713 | @noindent | |
2714 | When I do this, @file{"introduction.texinfo"} is the value returned by | |
2715 | evaluating @code{(buffer-name)}, and | |
2716 | @file{"/gnu/work/intro/introduction.texinfo"} is the value returned by | |
2717 | evaluating @code{(buffer-file-name)}. The former is the name of the | |
2718 | buffer and the latter is the name of the file. (In the expressions, the | |
2719 | parentheses tell the Lisp interpreter to treat @code{buffer-name} and | |
2720 | @code{buffer-file-name} as functions; without the parentheses, the | |
2721 | interpreter would attempt to evaluate the symbols as variables. | |
2722 | @xref{Variables}.) | |
2723 | ||
2724 | In spite of the distinction between files and buffers, you will often | |
2725 | find that people refer to a file when they mean a buffer and vice-versa. | |
2726 | Indeed, most people say, ``I am editing a file,'' rather than saying, | |
2727 | ``I am editing a buffer which I will soon save to a file.'' It is | |
2728 | almost always clear from context what people mean. When dealing with | |
2729 | computer programs, however, it is important to keep the distinction in mind, | |
2730 | since the computer is not as smart as a person. | |
2731 | ||
2732 | @cindex Buffer, history of word | |
2733 | The word `buffer', by the way, comes from the meaning of the word as a | |
2734 | cushion that deadens the force of a collision. In early computers, a | |
2735 | buffer cushioned the interaction between files and the computer's | |
2736 | central processing unit. The drums or tapes that held a file and the | |
2737 | central processing unit were pieces of equipment that were very | |
2738 | different from each other, working at their own speeds, in spurts. The | |
2739 | buffer made it possible for them to work together effectively. | |
2740 | Eventually, the buffer grew from being an intermediary, a temporary | |
2741 | holding place, to being the place where work is done. This | |
2742 | transformation is rather like that of a small seaport that grew into a | |
2743 | great city: once it was merely the place where cargo was warehoused | |
2744 | temporarily before being loaded onto ships; then it became a business | |
2745 | and cultural center in its own right. | |
2746 | ||
2747 | Not all buffers are associated with files. For example, when you start | |
2748 | an Emacs session by typing the command @code{emacs} alone, without | |
2749 | naming any files, Emacs will start with the @file{*scratch*} buffer on | |
2750 | the screen. This buffer is not visiting any file. Similarly, a | |
2751 | @file{*Help*} buffer is not associated with any file. | |
2752 | ||
2753 | @cindex @code{nil}, history of word | |
2754 | If you switch to the @file{*scratch*} buffer, type @code{(buffer-name)}, | |
2755 | position the cursor after it, and type @kbd{C-x C-e} to evaluate the | |
2756 | expression, the name @code{"*scratch*"} is returned and will appear in | |
2757 | the echo area. @code{"*scratch*"} is the name of the buffer. However, | |
2758 | if you type @code{(buffer-file-name)} in the @file{*scratch*} buffer and | |
2759 | evaluate that, @code{nil} will appear in the echo area. @code{nil} is | |
2760 | from the Latin word for `nothing'; in this case, it means that the | |
2761 | @file{*scratch*} buffer is not associated with any file. (In Lisp, | |
2762 | @code{nil} is also used to mean `false' and is a synonym for the empty | |
2763 | list, @code{()}.) | |
2764 | ||
2765 | Incidentally, if you are in the @file{*scratch*} buffer and want the | |
2766 | value returned by an expression to appear in the @file{*scratch*} | |
2767 | buffer itself rather than in the echo area, type @kbd{C-u C-x C-e} | |
2768 | instead of @kbd{C-x C-e}. This causes the value returned to appear | |
2769 | after the expression. The buffer will look like this: | |
2770 | ||
2771 | @smallexample | |
2772 | (buffer-name)"*scratch*" | |
2773 | @end smallexample | |
2774 | ||
2775 | @noindent | |
2776 | You cannot do this in Info since Info is read-only and it will not allow | |
2777 | you to change the contents of the buffer. But you can do this in any | |
2778 | buffer you can edit; and when you write code or documentation (such as | |
2779 | this book), this feature is very useful. | |
2780 | ||
2781 | @node Getting Buffers, Switching Buffers, Buffer Names, Practicing Evaluation | |
2782 | @comment node-name, next, previous, up | |
2783 | @section Getting Buffers | |
2784 | @findex current-buffer | |
2785 | @findex other-buffer | |
2786 | @cindex Getting a buffer | |
2787 | ||
2788 | The @code{buffer-name} function returns the @emph{name} of the buffer; | |
2789 | to get the buffer @emph{itself}, a different function is needed: the | |
2790 | @code{current-buffer} function. If you use this function in code, what | |
2791 | you get is the buffer itself. | |
2792 | ||
2793 | A name and the object or entity to which the name refers are different | |
2794 | from each other. You are not your name. You are a person to whom | |
2795 | others refer by name. If you ask to speak to George and someone hands you | |
2796 | a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r}, | |
2797 | @samp{g}, and @samp{e} written on it, you might be amused, but you would | |
2798 | not be satisfied. You do not want to speak to the name, but to the | |
2799 | person to whom the name refers. A buffer is similar: the name of the | |
2800 | scratch buffer is @file{*scratch*}, but the name is not the buffer. To | |
2801 | get a buffer itself, you need to use a function such as | |
2802 | @code{current-buffer}. | |
2803 | ||
2804 | However, there is a slight complication: if you evaluate | |
2805 | @code{current-buffer} in an expression on its own, as we will do here, | |
2806 | what you see is a printed representation of the name of the buffer | |
2807 | without the contents of the buffer. Emacs works this way for two | |
2808 | reasons: the buffer may be thousands of lines long---too long to be | |
2809 | conveniently displayed; and, another buffer may have the same contents | |
2810 | but a different name, and it is important to distinguish between them. | |
2811 | ||
2812 | @need 800 | |
2813 | Here is an expression containing the function: | |
2814 | ||
2815 | @smallexample | |
2816 | (current-buffer) | |
2817 | @end smallexample | |
2818 | ||
2819 | @noindent | |
2820 | If you evaluate the expression in the usual way, @file{#<buffer *info*>} | |
2821 | appears in the echo area. The special format indicates that the | |
2822 | buffer itself is being returned, rather than just its name. | |
2823 | ||
2824 | Incidentally, while you can type a number or symbol into a program, you | |
2825 | cannot do that with the printed representation of a buffer: the only way | |
2826 | to get a buffer itself is with a function such as @code{current-buffer}. | |
2827 | ||
2828 | A related function is @code{other-buffer}. This returns the most | |
2829 | recently selected buffer other than the one you are in currently. If | |
2830 | you have recently switched back and forth from the @file{*scratch*} | |
2831 | buffer, @code{other-buffer} will return that buffer. | |
2832 | ||
2833 | @need 800 | |
2834 | You can see this by evaluating the expression: | |
2835 | ||
2836 | @smallexample | |
2837 | (other-buffer) | |
2838 | @end smallexample | |
2839 | ||
2840 | @noindent | |
2841 | You should see @file{#<buffer *scratch*>} appear in the echo area, or | |
2842 | the name of whatever other buffer you switched back from most | |
2843 | recently@footnote{Actually, by default, if the buffer from which you | |
2844 | just switched is visible to you in another window, @code{other-buffer} | |
2845 | will choose the most recent buffer that you cannot see; this is a | |
2846 | subtlety that I often forget.}. | |
2847 | ||
2848 | @node Switching Buffers, Buffer Size & Locations, Getting Buffers, Practicing Evaluation | |
2849 | @comment node-name, next, previous, up | |
2850 | @section Switching Buffers | |
2851 | @findex switch-to-buffer | |
2852 | @findex set-buffer | |
2853 | @cindex Switching to a buffer | |
2854 | ||
2855 | The @code{other-buffer} function actually provides a buffer when it is | |
2856 | used as an argument to a function that requires one. We can see this | |
2857 | by using @code{other-buffer} and @code{switch-to-buffer} to switch to a | |
2858 | different buffer. | |
2859 | ||
2860 | But first, a brief introduction to the @code{switch-to-buffer} | |
2861 | function. When you switched back and forth from Info to the | |
2862 | @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most | |
2863 | likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or | |
2864 | rather, to save typing, you probably typed just part of the name, such | |
2865 | as @code{*sc}, and then pressed your @kbd{TAB} key to cause it to | |
2866 | expand to the full name; and then typed your @kbd{RET} key.} when | |
2867 | prompted in the minibuffer for the name of the buffer to which you | |
2868 | wanted to switch. The keystrokes, @kbd{C-x b}, cause the Lisp | |
2869 | interpreter to evaluate the interactive function | |
2870 | @code{switch-to-buffer}. As we said before, this is how Emacs works: | |
2871 | different keystrokes call or run different functions. For example, | |
2872 | @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls | |
2873 | @code{forward-sentence}, and so on. | |
2874 | ||
2875 | By writing @code{switch-to-buffer} in an expression, and giving it a | |
2876 | buffer to switch to, we can switch buffers just the way @kbd{C-x b} | |
2877 | does. | |
2878 | ||
2879 | @need 1000 | |
2880 | Here is the Lisp expression: | |
2881 | ||
2882 | @smallexample | |
2883 | (switch-to-buffer (other-buffer)) | |
2884 | @end smallexample | |
2885 | ||
2886 | @noindent | |
2887 | The symbol @code{switch-to-buffer} is the first element of the list, | |
2888 | so the Lisp interpreter will treat it as a function and carry out the | |
2889 | instructions that are attached to it. But before doing that, the | |
2890 | interpreter will note that @code{other-buffer} is inside parentheses | |
2891 | and work on that symbol first. @code{other-buffer} is the first (and | |
2892 | in this case, the only) element of this list, so the Lisp interpreter | |
2893 | calls or runs the function. It returns another buffer. Next, the | |
2894 | interpreter runs @code{switch-to-buffer}, passing to it, as an | |
2895 | argument, the other buffer, which is what Emacs will switch to. If | |
2896 | you are reading this in Info, try this now. Evaluate the expression. | |
2897 | (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this | |
2898 | expression will move you to your most recent other buffer that you | |
2899 | cannot see. If you really want to go to your most recently selected | |
2900 | buffer, even if you can still see it, you need to evaluate the | |
2901 | following more complex expression: | |
2902 | ||
2903 | @smallexample | |
2904 | (switch-to-buffer (other-buffer (current-buffer) t)) | |
2905 | @end smallexample | |
2906 | ||
2907 | @noindent | |
2908 | In this case, the first argument to @code{other-buffer} tells it which | |
2909 | buffer to skip---the current one---and the second argument tells | |
2910 | @code{other-buffer} it is OK to switch to a visible buffer. | |
2911 | In regular use, @code{switch-to-buffer} takes you to an invisible | |
2912 | window since you would most likely use @kbd{C-x o} (@code{other-window}) | |
2913 | to go to another visible buffer.} | |
2914 | ||
2915 | In the programming examples in later sections of this document, you will | |
2916 | see the function @code{set-buffer} more often than | |
2917 | @code{switch-to-buffer}. This is because of a difference between | |
2918 | computer programs and humans: humans have eyes and expect to see the | |
2919 | buffer on which they are working on their computer terminals. This is | |
2920 | so obvious, it almost goes without saying. However, programs do not | |
2921 | have eyes. When a computer program works on a buffer, that buffer does | |
2922 | not need to be visible on the screen. | |
2923 | ||
2924 | @code{switch-to-buffer} is designed for humans and does two different | |
2925 | things: it switches the buffer to which Emacs' attention is directed; and | |
2926 | it switches the buffer displayed in the window to the new buffer. | |
2927 | @code{set-buffer}, on the other hand, does only one thing: it switches | |
2928 | the attention of the computer program to a different buffer. The buffer | |
2929 | on the screen remains unchanged (of course, normally nothing happens | |
2930 | there until the command finishes running). | |
2931 | ||
2932 | @cindex @samp{call} defined | |
2933 | Also, we have just introduced another jargon term, the word @dfn{call}. | |
2934 | When you evaluate a list in which the first symbol is a function, you | |
2935 | are calling that function. The use of the term comes from the notion of | |
2936 | the function as an entity that can do something for you if you `call' | |
2937 | it---just as a plumber is an entity who can fix a leak if you call him | |
2938 | or her. | |
2939 | ||
2940 | @node Buffer Size & Locations, Evaluation Exercise, Switching Buffers, Practicing Evaluation | |
2941 | @comment node-name, next, previous, up | |
2942 | @section Buffer Size and the Location of Point | |
2943 | @cindex Size of buffer | |
2944 | @cindex Buffer size | |
2945 | @cindex Point location | |
2946 | @cindex Location of point | |
2947 | ||
2948 | Finally, let's look at several rather simple functions, | |
2949 | @code{buffer-size}, @code{point}, @code{point-min}, and | |
2950 | @code{point-max}. These give information about the size of a buffer and | |
2951 | the location of point within it. | |
2952 | ||
2953 | The function @code{buffer-size} tells you the size of the current | |
2954 | buffer; that is, the function returns a count of the number of | |
2955 | characters in the buffer. | |
2956 | ||
2957 | @smallexample | |
2958 | (buffer-size) | |
2959 | @end smallexample | |
2960 | ||
2961 | @noindent | |
2962 | You can evaluate this in the usual way, by positioning the | |
2963 | cursor after the expression and typing @kbd{C-x C-e}. | |
2964 | ||
2965 | @cindex @samp{point} defined | |
2966 | In Emacs, the current position of the cursor is called @dfn{point}. | |
2967 | The expression @code{(point)} returns a number that tells you where the | |
2968 | cursor is located as a count of the number of characters from the | |
2969 | beginning of the buffer up to point. | |
2970 | ||
2971 | @need 1250 | |
2972 | You can see the character count for point in this buffer by evaluating | |
2973 | the following expression in the usual way: | |
2974 | ||
2975 | @smallexample | |
2976 | (point) | |
2977 | @end smallexample | |
2978 | ||
2979 | @noindent | |
2980 | As I write this, the value of @code{point} is 65724. The @code{point} | |
2981 | function is frequently used in some of the examples later in this | |
2982 | book. | |
2983 | ||
2984 | @need 1250 | |
2985 | The value of point depends, of course, on its location within the | |
2986 | buffer. If you evaluate point in this spot, the number will be larger: | |
2987 | ||
2988 | @smallexample | |
2989 | (point) | |
2990 | @end smallexample | |
2991 | ||
2992 | @noindent | |
2993 | For me, the value of point in this location is 66043, which means that | |
2994 | there are 319 characters (including spaces) between the two expressions. | |
2995 | ||
2996 | @cindex @samp{narrowing} defined | |
2997 | The function @code{point-min} is somewhat similar to @code{point}, but | |
2998 | it returns the value of the minimum permissible value of point in the | |
2999 | current buffer. This is the number 1 unless @dfn{narrowing} is in | |
3000 | effect. (Narrowing is a mechanism whereby you can restrict yourself, | |
3001 | or a program, to operations on just a part of a buffer. | |
3002 | @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the | |
3003 | function @code{point-max} returns the value of the maximum permissible | |
3004 | value of point in the current buffer. | |
3005 | ||
3006 | @node Evaluation Exercise, , Buffer Size & Locations, Practicing Evaluation | |
3007 | @section Exercise | |
3008 | ||
3009 | Find a file with which you are working and move towards its middle. | |
3010 | Find its buffer name, file name, length, and your position in the file. | |
3011 | ||
3012 | @node Writing Defuns, Buffer Walk Through, Practicing Evaluation, Top | |
3013 | @comment node-name, next, previous, up | |
3014 | @chapter How To Write Function Definitions | |
3015 | @cindex Definition writing | |
3016 | @cindex Function definition writing | |
3017 | @cindex Writing a function definition | |
3018 | ||
3019 | When the Lisp interpreter evaluates a list, it looks to see whether the | |
3020 | first symbol on the list has a function definition attached to it; or, | |
3021 | put another way, whether the symbol points to a function definition. If | |
3022 | it does, the computer carries out the instructions in the definition. A | |
3023 | symbol that has a function definition is called, simply, a function | |
3024 | (although, properly speaking, the definition is the function and the | |
3025 | symbol refers to it.) | |
3026 | ||
3027 | @menu | |
3028 | * Primitive Functions:: | |
3029 | * defun:: The @code{defun} special form. | |
3030 | * Install:: Install a function definition. | |
3031 | * Interactive:: Making a function interactive. | |
3032 | * Interactive Options:: Different options for @code{interactive}. | |
3033 | * Permanent Installation:: Installing code permanently. | |
3034 | * let:: Creating and initializing local variables. | |
3035 | * if:: What if? | |
3036 | * else:: If--then--else expressions. | |
3037 | * Truth & Falsehood:: What Lisp considers false and true. | |
3038 | * save-excursion:: Keeping track of point, mark, and buffer. | |
3039 | * Review:: | |
3040 | * defun Exercises:: | |
3041 | @end menu | |
3042 | ||
3043 | @node Primitive Functions, defun, Writing Defuns, Writing Defuns | |
3044 | @ifnottex | |
3045 | @unnumberedsec An Aside about Primitive Functions | |
3046 | @end ifnottex | |
3047 | @cindex Primitive functions | |
3048 | @cindex Functions, primitive | |
3049 | ||
3050 | @cindex C language primitives | |
3051 | @cindex Primitives written in C | |
3052 | All functions are defined in terms of other functions, except for a few | |
3053 | @dfn{primitive} functions that are written in the C programming | |
3054 | language. When you write functions' definitions, you will write them in | |
3055 | Emacs Lisp and use other functions as your building blocks. Some of the | |
3056 | functions you will use will themselves be written in Emacs Lisp (perhaps | |
3057 | by you) and some will be primitives written in C. The primitive | |
3058 | functions are used exactly like those written in Emacs Lisp and behave | |
3059 | like them. They are written in C so we can easily run GNU Emacs on any | |
3060 | computer that has sufficient power and can run C. | |
3061 | ||
3062 | Let me re-emphasize this: when you write code in Emacs Lisp, you do not | |
3063 | distinguish between the use of functions written in C and the use of | |
3064 | functions written in Emacs Lisp. The difference is irrelevant. I | |
3065 | mention the distinction only because it is interesting to know. Indeed, | |
3066 | unless you investigate, you won't know whether an already-written | |
3067 | function is written in Emacs Lisp or C. | |
3068 | ||
3069 | @node defun, Install, Primitive Functions, Writing Defuns | |
3070 | @comment node-name, next, previous, up | |
3071 | @section The @code{defun} Special Form | |
3072 | @findex defun | |
3073 | @cindex Special form of @code{defun} | |
3074 | ||
3075 | @cindex @samp{function definition} defined | |
3076 | In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to | |
3077 | it that tells the computer what to do when the function is called. | |
3078 | This code is called the @dfn{function definition} and is created by | |
3079 | evaluating a Lisp expression that starts with the symbol @code{defun} | |
3080 | (which is an abbreviation for @emph{define function}). Because | |
3081 | @code{defun} does not evaluate its arguments in the usual way, it is | |
3082 | called a @dfn{special form}. | |
3083 | ||
3084 | In subsequent sections, we will look at function definitions from the | |
3085 | Emacs source code, such as @code{mark-whole-buffer}. In this section, | |
3086 | we will describe a simple function definition so you can see how it | |
3087 | looks. This function definition uses arithmetic because it makes for a | |
3088 | simple example. Some people dislike examples using arithmetic; however, | |
3089 | if you are such a person, do not despair. Hardly any of the code we | |
3090 | will study in the remainder of this introduction involves arithmetic or | |
3091 | mathematics. The examples mostly involve text in one way or another. | |
3092 | ||
3093 | A function definition has up to five parts following the word | |
3094 | @code{defun}: | |
3095 | ||
3096 | @enumerate | |
3097 | @item | |
3098 | The name of the symbol to which the function definition should be | |
3099 | attached. | |
3100 | ||
3101 | @item | |
3102 | A list of the arguments that will be passed to the function. If no | |
3103 | arguments will be passed to the function, this is an empty list, | |
3104 | @code{()}. | |
3105 | ||
3106 | @item | |
3107 | Documentation describing the function. (Technically optional, but | |
3108 | strongly recommended.) | |
3109 | ||
3110 | @item | |
3111 | Optionally, an expression to make the function interactive so you can | |
3112 | use it by typing @kbd{M-x} and then the name of the function; or by | |
3113 | typing an appropriate key or keychord. | |
3114 | ||
3115 | @cindex @samp{body} defined | |
3116 | @item | |
3117 | The code that instructs the computer what to do: the @dfn{body} of the | |
3118 | function definition. | |
3119 | @end enumerate | |
3120 | ||
3121 | It is helpful to think of the five parts of a function definition as | |
3122 | being organized in a template, with slots for each part: | |
3123 | ||
3124 | @smallexample | |
3125 | @group | |
3126 | (defun @var{function-name} (@var{arguments}@dots{}) | |
3127 | "@var{optional-documentation}@dots{}" | |
3128 | (interactive @var{argument-passing-info}) ; @r{optional} | |
3129 | @var{body}@dots{}) | |
3130 | @end group | |
3131 | @end smallexample | |
3132 | ||
3133 | As an example, here is the code for a function that multiplies its | |
3134 | argument by 7. (This example is not interactive. @xref{Interactive, | |
3135 | , Making a Function Interactive}, for that information.) | |
3136 | ||
3137 | @smallexample | |
3138 | @group | |
3139 | (defun multiply-by-seven (number) | |
3140 | "Multiply NUMBER by seven." | |
3141 | (* 7 number)) | |
3142 | @end group | |
3143 | @end smallexample | |
3144 | ||
3145 | This definition begins with a parenthesis and the symbol @code{defun}, | |
3146 | followed by the name of the function. | |
3147 | ||
3148 | @cindex @samp{argument list} defined | |
3149 | The name of the function is followed by a list that contains the | |
3150 | arguments that will be passed to the function. This list is called | |
3151 | the @dfn{argument list}. In this example, the list has only one | |
3152 | element, the symbol, @code{number}. When the function is used, the | |
3153 | symbol will be bound to the value that is used as the argument to the | |
3154 | function. | |
3155 | ||
3156 | Instead of choosing the word @code{number} for the name of the argument, | |
3157 | I could have picked any other name. For example, I could have chosen | |
3158 | the word @code{multiplicand}. I picked the word `number' because it | |
3159 | tells what kind of value is intended for this slot; but I could just as | |
3160 | well have chosen the word `multiplicand' to indicate the role that the | |
3161 | value placed in this slot will play in the workings of the function. I | |
3162 | could have called it @code{foogle}, but that would have been a bad | |
3163 | choice because it would not tell humans what it means. The choice of | |
3164 | name is up to the programmer and should be chosen to make the meaning of | |
3165 | the function clear. | |
3166 | ||
3167 | Indeed, you can choose any name you wish for a symbol in an argument | |
3168 | list, even the name of a symbol used in some other function: the name | |
3169 | you use in an argument list is private to that particular definition. | |
3170 | In that definition, the name refers to a different entity than any use | |
3171 | of the same name outside the function definition. Suppose you have a | |
3172 | nick-name `Shorty' in your family; when your family members refer to | |
3173 | `Shorty', they mean you. But outside your family, in a movie, for | |
3174 | example, the name `Shorty' refers to someone else. Because a name in an | |
3175 | argument list is private to the function definition, you can change the | |
3176 | value of such a symbol inside the body of a function without changing | |
3177 | its value outside the function. The effect is similar to that produced | |
3178 | by a @code{let} expression. (@xref{let, , @code{let}}.) | |
3179 | ||
3180 | @ignore | |
3181 | Note also that we discuss the word `number' in two different ways: as a | |
3182 | symbol that appears in the code, and as the name of something that will | |
3183 | be replaced by a something else during the evaluation of the function. | |
3184 | In the first case, @code{number} is a symbol, not a number; it happens | |
3185 | that within the function, it is a variable who value is the number in | |
3186 | question, but our primary interest in it is as a symbol. On the other | |
3187 | hand, when we are talking about the function, our interest is that we | |
3188 | will substitute a number for the word @var{number}. To keep this | |
3189 | distinction clear, we use different typography for the two | |
3190 | circumstances. When we talk about this function, or about how it works, | |
3191 | we refer to this number by writing @var{number}. In the function | |
3192 | itself, we refer to it by writing @code{number}. | |
3193 | @end ignore | |
3194 | ||
3195 | The argument list is followed by the documentation string that | |
3196 | describes the function. This is what you see when you type | |
3197 | @w{@kbd{C-h f}} and the name of a function. Incidentally, when you | |
3198 | write a documentation string like this, you should make the first line | |
3199 | a complete sentence since some commands, such as @code{apropos}, print | |
3200 | only the first line of a multi-line documentation string. Also, you | |
3201 | should not indent the second line of a documentation string, if you | |
3202 | have one, because that looks odd when you use @kbd{C-h f} | |
3203 | (@code{describe-function}). The documentation string is optional, but | |
3204 | it is so useful, it should be included in almost every function you | |
3205 | write. | |
3206 | ||
3207 | @findex * @r{(multiplication)} | |
3208 | The third line of the example consists of the body of the function | |
3209 | definition. (Most functions' definitions, of course, are longer than | |
3210 | this.) In this function, the body is the list, @code{(* 7 number)}, which | |
3211 | says to multiply the value of @var{number} by 7. (In Emacs Lisp, | |
3212 | @code{*} is the function for multiplication, just as @code{+} is the | |
3213 | function for addition.) | |
3214 | ||
3215 | When you use the @code{multiply-by-seven} function, the argument | |
3216 | @code{number} evaluates to the actual number you want used. Here is an | |
3217 | example that shows how @code{multiply-by-seven} is used; but don't try | |
3218 | to evaluate this yet! | |
3219 | ||
3220 | @smallexample | |
3221 | (multiply-by-seven 3) | |
3222 | @end smallexample | |
3223 | ||
3224 | @noindent | |
3225 | The symbol @code{number}, specified in the function definition in the | |
3226 | next section, is given or ``bound to'' the value 3 in the actual use of | |
3227 | the function. Note that although @code{number} was inside parentheses | |
3228 | in the function definition, the argument passed to the | |
3229 | @code{multiply-by-seven} function is not in parentheses. The | |
3230 | parentheses are written in the function definition so the computer can | |
3231 | figure out where the argument list ends and the rest of the function | |
3232 | definition begins. | |
3233 | ||
3234 | If you evaluate this example, you are likely to get an error message. | |
3235 | (Go ahead, try it!) This is because we have written the function | |
3236 | definition, but not yet told the computer about the definition---we have | |
3237 | not yet installed (or `loaded') the function definition in Emacs. | |
3238 | Installing a function is the process that tells the Lisp interpreter the | |
3239 | definition of the function. Installation is described in the next | |
3240 | section. | |
3241 | ||
3242 | @node Install, Interactive, defun, Writing Defuns | |
3243 | @comment node-name, next, previous, up | |
3244 | @section Install a Function Definition | |
3245 | @cindex Install a Function Definition | |
3246 | @cindex Definition installation | |
3247 | @cindex Function definition installation | |
3248 | ||
3249 | If you are reading this inside of Info in Emacs, you can try out the | |
3250 | @code{multiply-by-seven} function by first evaluating the function | |
3251 | definition and then evaluating @code{(multiply-by-seven 3)}. A copy of | |
3252 | the function definition follows. Place the cursor after the last | |
3253 | parenthesis of the function definition and type @kbd{C-x C-e}. When you | |
3254 | do this, @code{multiply-by-seven} will appear in the echo area. (What | |
3255 | this means is that when a function definition is evaluated, the value it | |
3256 | returns is the name of the defined function.) At the same time, this | |
3257 | action installs the function definition. | |
3258 | ||
3259 | @smallexample | |
3260 | @group | |
3261 | (defun multiply-by-seven (number) | |
3262 | "Multiply NUMBER by seven." | |
3263 | (* 7 number)) | |
3264 | @end group | |
3265 | @end smallexample | |
3266 | ||
3267 | @noindent | |
3268 | By evaluating this @code{defun}, you have just installed | |
3269 | @code{multiply-by-seven} in Emacs. The function is now just as much a | |
3270 | part of Emacs as @code{forward-word} or any other editing function you | |
3271 | use. (@code{multiply-by-seven} will stay installed until you quit | |
3272 | Emacs. To reload code automatically whenever you start Emacs, see | |
3273 | @ref{Permanent Installation, , Installing Code Permanently}.) | |
3274 | ||
3275 | ||
3276 | @menu | |
3277 | * Effect of installation:: | |
3278 | * Change a defun:: How to change a function definition. | |
3279 | @end menu | |
3280 | ||
3281 | @node Effect of installation, Change a defun, Install, Install | |
3282 | @ifnottex | |
3283 | @unnumberedsubsec The effect of installation | |
3284 | @end ifnottex | |
3285 | ||
3286 | ||
3287 | You can see the effect of installing @code{multiply-by-seven} by | |
3288 | evaluating the following sample. Place the cursor after the following | |
3289 | expression and type @kbd{C-x C-e}. The number 21 will appear in the | |
3290 | echo area. | |
3291 | ||
3292 | @smallexample | |
3293 | (multiply-by-seven 3) | |
3294 | @end smallexample | |
3295 | ||
3296 | If you wish, you can read the documentation for the function by typing | |
3297 | @kbd{C-h f} (@code{describe-function}) and then the name of the | |
3298 | function, @code{multiply-by-seven}. When you do this, a | |
3299 | @file{*Help*} window will appear on your screen that says: | |
3300 | ||
3301 | @smallexample | |
3302 | @group | |
3303 | multiply-by-seven: | |
3304 | Multiply NUMBER by seven. | |
3305 | @end group | |
3306 | @end smallexample | |
3307 | ||
3308 | @noindent | |
3309 | (To return to a single window on your screen, type @kbd{C-x 1}.) | |
3310 | ||
3311 | @node Change a defun, , Effect of installation, Install | |
3312 | @comment node-name, next, previous, up | |
3313 | @subsection Change a Function Definition | |
3314 | @cindex Changing a function definition | |
3315 | @cindex Function definition, how to change | |
3316 | @cindex Definition, how to change | |
3317 | ||
3318 | If you want to change the code in @code{multiply-by-seven}, just rewrite | |
3319 | it. To install the new version in place of the old one, evaluate the | |
3320 | function definition again. This is how you modify code in Emacs. It is | |
3321 | very simple. | |
3322 | ||
3323 | As an example, you can change the @code{multiply-by-seven} function to | |
3324 | add the number to itself seven times instead of multiplying the number | |
3325 | by seven. It produces the same answer, but by a different path. At | |
3326 | the same time, we will add a comment to the code; a comment is text | |
3327 | that the Lisp interpreter ignores, but that a human reader may find | |
3328 | useful or enlightening. The comment is that this is the ``second | |
3329 | version''. | |
3330 | ||
3331 | @smallexample | |
3332 | @group | |
3333 | (defun multiply-by-seven (number) ; @r{Second version.} | |
3334 | "Multiply NUMBER by seven." | |
3335 | (+ number number number number number number number)) | |
3336 | @end group | |
3337 | @end smallexample | |
3338 | ||
3339 | @cindex Comments in Lisp code | |
3340 | The comment follows a semicolon, @samp{;}. In Lisp, everything on a | |
3341 | line that follows a semicolon is a comment. The end of the line is the | |
3342 | end of the comment. To stretch a comment over two or more lines, begin | |
3343 | each line with a semicolon. | |
3344 | ||
3345 | @xref{Beginning a .emacs File, , Beginning a @file{.emacs} | |
3346 | File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp | |
3347 | Reference Manual}, for more about comments. | |
3348 | ||
3349 | You can install this version of the @code{multiply-by-seven} function by | |
3350 | evaluating it in the same way you evaluated the first function: place | |
3351 | the cursor after the last parenthesis and type @kbd{C-x C-e}. | |
3352 | ||
3353 | In summary, this is how you write code in Emacs Lisp: you write a | |
3354 | function; install it; test it; and then make fixes or enhancements and | |
3355 | install it again. | |
3356 | ||
3357 | @node Interactive, Interactive Options, Install, Writing Defuns | |
3358 | @comment node-name, next, previous, up | |
3359 | @section Make a Function Interactive | |
3360 | @cindex Interactive functions | |
3361 | @findex interactive | |
3362 | ||
3363 | You make a function interactive by placing a list that begins with | |
3364 | the special form @code{interactive} immediately after the | |
3365 | documentation. A user can invoke an interactive function by typing | |
3366 | @kbd{M-x} and then the name of the function; or by typing the keys to | |
3367 | which it is bound, for example, by typing @kbd{C-n} for | |
3368 | @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}. | |
3369 | ||
3370 | Interestingly, when you call an interactive function interactively, | |
3371 | the value returned is not automatically displayed in the echo area. | |
3372 | This is because you often call an interactive function for its side | |
3373 | effects, such as moving forward by a word or line, and not for the | |
3374 | value returned. If the returned value were displayed in the echo area | |
3375 | each time you typed a key, it would be very distracting. | |
3376 | ||
3377 | @menu | |
3378 | * Interactive multiply-by-seven:: An overview. | |
3379 | * multiply-by-seven in detail:: The interactive version. | |
3380 | @end menu | |
3381 | ||
3382 | @node Interactive multiply-by-seven, multiply-by-seven in detail, Interactive, Interactive | |
3383 | @ifnottex | |
3384 | @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview | |
3385 | @end ifnottex | |
3386 | ||
3387 | Both the use of the special form @code{interactive} and one way to | |
3388 | display a value in the echo area can be illustrated by creating an | |
3389 | interactive version of @code{multiply-by-seven}. | |
3390 | ||
3391 | @need 1250 | |
3392 | Here is the code: | |
3393 | ||
3394 | @smallexample | |
3395 | @group | |
3396 | (defun multiply-by-seven (number) ; @r{Interactive version.} | |
3397 | "Multiply NUMBER by seven." | |
3398 | (interactive "p") | |
3399 | (message "The result is %d" (* 7 number))) | |
3400 | @end group | |
3401 | @end smallexample | |
3402 | ||
3403 | @noindent | |
3404 | You can install this code by placing your cursor after it and typing | |
3405 | @kbd{C-x C-e}. The name of the function will appear in your echo area. | |
3406 | Then, you can use this code by typing @kbd{C-u} and a number and then | |
3407 | typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase | |
3408 | @samp{The result is @dots{}} followed by the product will appear in the | |
3409 | echo area. | |
3410 | ||
3411 | Speaking more generally, you invoke a function like this in either of two | |
3412 | ways: | |
3413 | ||
3414 | @enumerate | |
3415 | @item | |
3416 | By typing a prefix argument that contains the number to be passed, and | |
3417 | then typing @kbd{M-x} and the name of the function, as with | |
3418 | @kbd{C-u 3 M-x forward-sentence}; or, | |
3419 | ||
3420 | @item | |
3421 | By typing whatever key or keychord the function is bound to, as with | |
3422 | @kbd{C-u 3 M-e}. | |
3423 | @end enumerate | |
3424 | ||
3425 | @noindent | |
3426 | Both the examples just mentioned work identically to move point forward | |
3427 | three sentences. (Since @code{multiply-by-seven} is not bound to a key, | |
3428 | it could not be used as an example of key binding.) | |
3429 | ||
3430 | (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command | |
3431 | to a key.) | |
3432 | ||
3433 | A prefix argument is passed to an interactive function by typing the | |
3434 | @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by | |
3435 | typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you | |
3436 | type @kbd{C-u} without a number, it defaults to 4). | |
3437 | ||
3438 | @node multiply-by-seven in detail, , Interactive multiply-by-seven, Interactive | |
3439 | @comment node-name, next, previous, up | |
3440 | @subsection An Interactive @code{multiply-by-seven} | |
3441 | ||
3442 | Let's look at the use of the special form @code{interactive} and then at | |
3443 | the function @code{message} in the interactive version of | |
3444 | @code{multiply-by-seven}. You will recall that the function definition | |
3445 | looks like this: | |
3446 | ||
3447 | @smallexample | |
3448 | @group | |
3449 | (defun multiply-by-seven (number) ; @r{Interactive version.} | |
3450 | "Multiply NUMBER by seven." | |
3451 | (interactive "p") | |
3452 | (message "The result is %d" (* 7 number))) | |
3453 | @end group | |
3454 | @end smallexample | |
3455 | ||
3456 | In this function, the expression, @code{(interactive "p")}, is a list of | |
3457 | two elements. The @code{"p"} tells Emacs to pass the prefix argument to | |
3458 | the function and use its value for the argument of the function. | |
3459 | ||
3460 | @need 1000 | |
3461 | The argument will be a number. This means that the symbol | |
3462 | @code{number} will be bound to a number in the line: | |
3463 | ||
3464 | @smallexample | |
3465 | (message "The result is %d" (* 7 number)) | |
3466 | @end smallexample | |
3467 | ||
3468 | @need 1250 | |
3469 | @noindent | |
3470 | For example, if your prefix argument is 5, the Lisp interpreter will | |
3471 | evaluate the line as if it were: | |
3472 | ||
3473 | @smallexample | |
3474 | (message "The result is %d" (* 7 5)) | |
3475 | @end smallexample | |
3476 | ||
3477 | @noindent | |
3478 | (If you are reading this in GNU Emacs, you can evaluate this expression | |
3479 | yourself.) First, the interpreter will evaluate the inner list, which | |
3480 | is @code{(* 7 5)}. This returns a value of 35. Next, it | |
3481 | will evaluate the outer list, passing the values of the second and | |
3482 | subsequent elements of the list to the function @code{message}. | |
3483 | ||
3484 | As we have seen, @code{message} is an Emacs Lisp function especially | |
3485 | designed for sending a one line message to a user. (@xref{message, , The | |
3486 | @code{message} function}.) | |
3487 | In summary, the @code{message} function prints its first argument in the | |
3488 | echo area as is, except for occurrences of @samp{%d}, @samp{%s}, or | |
3489 | @samp{%c}. When it sees one of these control sequences, the function | |
3490 | looks to the second and subsequent arguments and prints the value of the | |
3491 | argument in the location in the string where the control sequence is | |
3492 | located. | |
3493 | ||
3494 | In the interactive @code{multiply-by-seven} function, the control string | |
3495 | is @samp{%d}, which requires a number, and the value returned by | |
3496 | evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35 | |
3497 | is printed in place of the @samp{%d} and the message is @samp{The result | |
3498 | is 35}. | |
3499 | ||
3500 | (Note that when you call the function @code{multiply-by-seven}, the | |
3501 | message is printed without quotes, but when you call @code{message}, the | |
3502 | text is printed in double quotes. This is because the value returned by | |
3503 | @code{message} is what appears in the echo area when you evaluate an | |
3504 | expression whose first element is @code{message}; but when embedded in a | |
3505 | function, @code{message} prints the text as a side effect without | |
3506 | quotes.) | |
3507 | ||
3508 | @node Interactive Options, Permanent Installation, Interactive, Writing Defuns | |
3509 | @comment node-name, next, previous, up | |
3510 | @section Different Options for @code{interactive} | |
3511 | @cindex Options for @code{interactive} | |
3512 | @cindex Interactive options | |
3513 | ||
3514 | In the example, @code{multiply-by-seven} used @code{"p"} as the | |
3515 | argument to @code{interactive}. This argument told Emacs to interpret | |
3516 | your typing either @kbd{C-u} followed by a number or @key{META} | |
3517 | followed by a number as a command to pass that number to the function | |
3518 | as its argument. Emacs has more than twenty characters predefined for | |
3519 | use with @code{interactive}. In almost every case, one of these | |
3520 | options will enable you to pass the right information interactively to | |
3521 | a function. (@xref{Interactive Codes, , Code Characters for | |
3522 | @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.) | |
3523 | ||
3524 | @need 1250 | |
3525 | For example, the character @samp{r} causes Emacs to pass the beginning | |
3526 | and end of the region (the current values of point and mark) to the | |
3527 | function as two separate arguments. It is used as follows: | |
3528 | ||
3529 | @smallexample | |
3530 | (interactive "r") | |
3531 | @end smallexample | |
3532 | ||
3533 | On the other hand, a @samp{B} tells Emacs to ask for the name of a | |
3534 | buffer that will be passed to the function. When it sees a @samp{B}, | |
3535 | Emacs will ask for the name by prompting the user in the minibuffer, | |
3536 | using a string that follows the @samp{B}, as in @code{"BAppend to | |
3537 | buffer:@: "}. Not only will Emacs prompt for the name, but Emacs will | |
3538 | complete the name if you type enough of it and press @key{TAB}. | |
3539 | ||
3540 | A function with two or more arguments can have information passed to | |
3541 | each argument by adding parts to the string that follows | |
3542 | @code{interactive}. When you do this, the information is passed to | |
3543 | each argument in the same order it is specified in the | |
3544 | @code{interactive} list. In the string, each part is separated from | |
3545 | the next part by a @samp{\n}, which is a newline. For example, you | |
3546 | could follow @code{"BAppend to buffer:@: "} with a @samp{\n}) and an | |
3547 | @samp{r}. This would cause Emacs to pass the values of point and mark | |
3548 | to the function as well as prompt you for the buffer---three arguments | |
3549 | in all. | |
3550 | ||
3551 | In this case, the function definition would look like the following, | |
3552 | where @code{buffer}, @code{start}, and @code{end} are the symbols to | |
3553 | which @code{interactive} binds the buffer and the current values of the | |
3554 | beginning and ending of the region: | |
3555 | ||
3556 | @smallexample | |
3557 | @group | |
3558 | (defun @var{name-of-function} (buffer start end) | |
3559 | "@var{documentation}@dots{}" | |
3560 | (interactive "BAppend to buffer:@: \nr") | |
3561 | @var{body-of-function}@dots{}) | |
3562 | @end group | |
3563 | @end smallexample | |
3564 | ||
3565 | @noindent | |
3566 | (The space after the colon in the prompt makes it look better when you | |
3567 | are prompted. The @code{append-to-buffer} function looks exactly like | |
3568 | this. @xref{append-to-buffer, , The Definition of | |
3569 | @code{append-to-buffer}}.) | |
3570 | ||
3571 | If a function does not have arguments, then @code{interactive} does not | |
3572 | require any. Such a function contains the simple expression | |
3573 | @code{(interactive)}. The @code{mark-whole-buffer} function is like | |
3574 | this. | |
3575 | ||
3576 | Alternatively, if the special letter-codes are not right for your | |
3577 | application, you can pass your own arguments to @code{interactive} as | |
3578 | a list. @xref{interactive, , Using @code{Interactive}, elisp, The | |
3579 | GNU Emacs Lisp Reference Manual}, for more information about this advanced | |
3580 | technique. | |
3581 | ||
3582 | @node Permanent Installation, let, Interactive Options, Writing Defuns | |
3583 | @comment node-name, next, previous, up | |
3584 | @section Install Code Permanently | |
3585 | @cindex Install code permanently | |
3586 | @cindex Permanent code installation | |
3587 | @cindex Code installation | |
3588 | ||
3589 | When you install a function definition by evaluating it, it will stay | |
3590 | installed until you quit Emacs. The next time you start a new session | |
3591 | of Emacs, the function will not be installed unless you evaluate the | |
3592 | function definition again. | |
3593 | ||
3594 | At some point, you may want to have code installed automatically | |
3595 | whenever you start a new session of Emacs. There are several ways of | |
3596 | doing this: | |
3597 | ||
3598 | @itemize @bullet | |
3599 | @item | |
3600 | If you have code that is just for yourself, you can put the code for the | |
3601 | function definition in your @file{.emacs} initialization file. When you | |
3602 | start Emacs, your @file{.emacs} file is automatically evaluated and all | |
3603 | the function definitions within it are installed. | |
3604 | @xref{Emacs Initialization, , Your @file{.emacs} File}. | |
3605 | ||
3606 | @item | |
3607 | Alternatively, you can put the function definitions that you want | |
3608 | installed in one or more files of their own and use the @code{load} | |
3609 | function to cause Emacs to evaluate and thereby install each of the | |
3610 | functions in the files. | |
3611 | @xref{Loading Files, , Loading Files}. | |
3612 | ||
3613 | @item | |
3614 | On the other hand, if you have code that your whole site will use, it | |
3615 | is usual to put it in a file called @file{site-init.el} that is loaded | |
3616 | when Emacs is built. This makes the code available to everyone who | |
3617 | uses your machine. (See the @file{INSTALL} file that is part of the | |
3618 | Emacs distribution.) | |
3619 | @end itemize | |
3620 | ||
3621 | Finally, if you have code that everyone who uses Emacs may want, you | |
3622 | can post it on a computer network or send a copy to the Free Software | |
3623 | Foundation. (When you do this, please license the code and its | |
3624 | documentation under a license that permits other people to run, copy, | |
3625 | study, modify, and redistribute the code and which protects you from | |
3626 | having your work taken from you.) If you send a copy of your code to | |
3627 | the Free Software Foundation, and properly protect yourself and | |
3628 | others, it may be included in the next release of Emacs. In large | |
3629 | part, this is how Emacs has grown over the past years, by donations. | |
3630 | ||
3631 | @node let, if, Permanent Installation, Writing Defuns | |
3632 | @comment node-name, next, previous, up | |
3633 | @section @code{let} | |
3634 | @findex let | |
3635 | ||
3636 | The @code{let} expression is a special form in Lisp that you will need | |
3637 | to use in most function definitions. | |
3638 | ||
3639 | @code{let} is used to attach or bind a symbol to a value in such a way | |
3640 | that the Lisp interpreter will not confuse the variable with a | |
3641 | variable of the same name that is not part of the function. | |
3642 | ||
3643 | To understand why the @code{let} special form is necessary, consider | |
3644 | the situation in which you own a home that you generally refer to as | |
3645 | `the house', as in the sentence, ``The house needs painting.'' If you | |
3646 | are visiting a friend and your host refers to `the house', he is | |
3647 | likely to be referring to @emph{his} house, not yours, that is, to a | |
3648 | different house. | |
3649 | ||
3650 | If your friend is referring to his house and you think he is referring | |
3651 | to your house, you may be in for some confusion. The same thing could | |
3652 | happen in Lisp if a variable that is used inside of one function has | |
3653 | the same name as a variable that is used inside of another function, | |
3654 | and the two are not intended to refer to the same value. The | |
3655 | @code{let} special form prevents this kind of confusion. | |
3656 | ||
3657 | @menu | |
3658 | * Prevent confusion:: | |
3659 | * Parts of let Expression:: | |
3660 | * Sample let Expression:: | |
3661 | * Uninitialized let Variables:: | |
3662 | @end menu | |
3663 | ||
3664 | @node Prevent confusion, Parts of let Expression, let, let | |
3665 | @ifnottex | |
3666 | @unnumberedsubsec @code{let} Prevents Confusion | |
3667 | @end ifnottex | |
3668 | ||
3669 | @cindex @samp{local variable} defined | |
3670 | The @code{let} special form prevents confusion. @code{let} creates a | |
3671 | name for a @dfn{local variable} that overshadows any use of the same | |
3672 | name outside the @code{let} expression. This is like understanding | |
3673 | that whenever your host refers to `the house', he means his house, not | |
3674 | yours. (Symbols used in argument lists work the same way. | |
3675 | @xref{defun, , The @code{defun} Special Form}.) | |
3676 | ||
3677 | Local variables created by a @code{let} expression retain their value | |
3678 | @emph{only} within the @code{let} expression itself (and within | |
3679 | expressions called within the @code{let} expression); the local | |
3680 | variables have no effect outside the @code{let} expression. | |
3681 | ||
3682 | Another way to think about @code{let} is that it is like a @code{setq} | |
3683 | that is temporary and local. The values set by @code{let} are | |
3684 | automatically undone when the @code{let} is finished. The setting | |
3685 | only effects expressions that are inside the bounds of the @code{let} | |
3686 | expression. In computer science jargon, we would say ``the binding of | |
3687 | a symbol is visible only in functions called in the @code{let} form; | |
3688 | in Emacs Lisp, scoping is dynamic, not lexical.'' | |
3689 | ||
3690 | @code{let} can create more than one variable at once. Also, | |
3691 | @code{let} gives each variable it creates an initial value, either a | |
3692 | value specified by you, or @code{nil}. (In the jargon, this is called | |
3693 | `binding the variable to the value'.) After @code{let} has created | |
3694 | and bound the variables, it executes the code in the body of the | |
3695 | @code{let}, and returns the value of the last expression in the body, | |
3696 | as the value of the whole @code{let} expression. (`Execute' is a jargon | |
3697 | term that means to evaluate a list; it comes from the use of the word | |
3698 | meaning `to give practical effect to' (@cite{Oxford English | |
3699 | Dictionary}). Since you evaluate an expression to perform an action, | |
3700 | `execute' has evolved as a synonym to `evaluate'.) | |
3701 | ||
3702 | @node Parts of let Expression, Sample let Expression, Prevent confusion, let | |
3703 | @comment node-name, next, previous, up | |
3704 | @subsection The Parts of a @code{let} Expression | |
3705 | @cindex @code{let} expression, parts of | |
3706 | @cindex Parts of @code{let} expression | |
3707 | ||
3708 | @cindex @samp{varlist} defined | |
3709 | A @code{let} expression is a list of three parts. The first part is | |
3710 | the symbol @code{let}. The second part is a list, called a | |
3711 | @dfn{varlist}, each element of which is either a symbol by itself or a | |
3712 | two-element list, the first element of which is a symbol. The third | |
3713 | part of the @code{let} expression is the body of the @code{let}. The | |
3714 | body usually consists of one or more lists. | |
3715 | ||
3716 | @need 800 | |
3717 | A template for a @code{let} expression looks like this: | |
3718 | ||
3719 | @smallexample | |
3720 | (let @var{varlist} @var{body}@dots{}) | |
3721 | @end smallexample | |
3722 | ||
3723 | @noindent | |
3724 | The symbols in the varlist are the variables that are given initial | |
3725 | values by the @code{let} special form. Symbols by themselves are given | |
3726 | the initial value of @code{nil}; and each symbol that is the first | |
3727 | element of a two-element list is bound to the value that is returned | |
3728 | when the Lisp interpreter evaluates the second element. | |
3729 | ||
3730 | Thus, a varlist might look like this: @code{(thread (needles 3))}. In | |
3731 | this case, in a @code{let} expression, Emacs binds the symbol | |
3732 | @code{thread} to an initial value of @code{nil}, and binds the symbol | |
3733 | @code{needles} to an initial value of 3. | |
3734 | ||
3735 | When you write a @code{let} expression, what you do is put the | |
3736 | appropriate expressions in the slots of the @code{let} expression | |
3737 | template. | |
3738 | ||
3739 | If the varlist is composed of two-element lists, as is often the case, | |
3740 | the template for the @code{let} expression looks like this: | |
3741 | ||
3742 | @smallexample | |
3743 | @group | |
3744 | (let ((@var{variable} @var{value}) | |
3745 | (@var{variable} @var{value}) | |
3746 | @dots{}) | |
3747 | @var{body}@dots{}) | |
3748 | @end group | |
3749 | @end smallexample | |
3750 | ||
3751 | @node Sample let Expression, Uninitialized let Variables, Parts of let Expression, let | |
3752 | @comment node-name, next, previous, up | |
3753 | @subsection Sample @code{let} Expression | |
3754 | @cindex Sample @code{let} expression | |
3755 | @cindex @code{let} expression sample | |
3756 | ||
3757 | The following expression creates and gives initial values | |
3758 | to the two variables @code{zebra} and @code{tiger}. The body of the | |
3759 | @code{let} expression is a list which calls the @code{message} function. | |
3760 | ||
3761 | @smallexample | |
3762 | @group | |
3763 | (let ((zebra 'stripes) | |
3764 | (tiger 'fierce)) | |
3765 | (message "One kind of animal has %s and another is %s." | |
3766 | zebra tiger)) | |
3767 | @end group | |
3768 | @end smallexample | |
3769 | ||
3770 | Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}. | |
3771 | ||
3772 | The two variables are @code{zebra} and @code{tiger}. Each variable is | |
3773 | the first element of a two-element list and each value is the second | |
3774 | element of its two-element list. In the varlist, Emacs binds the | |
3775 | variable @code{zebra} to the value @code{stripes}, and binds the | |
3776 | variable @code{tiger} to the value @code{fierce}. In this example, | |
3777 | both values are symbols preceded by a quote. The values could just as | |
3778 | well have been another list or a string. The body of the @code{let} | |
3779 | follows after the list holding the variables. In this example, the body | |
3780 | is a list that uses the @code{message} function to print a string in | |
3781 | the echo area. | |
3782 | ||
3783 | @need 1500 | |
3784 | You may evaluate the example in the usual fashion, by placing the | |
3785 | cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do | |
3786 | this, the following will appear in the echo area: | |
3787 | ||
3788 | @smallexample | |
3789 | "One kind of animal has stripes and another is fierce." | |
3790 | @end smallexample | |
3791 | ||
3792 | As we have seen before, the @code{message} function prints its first | |
3793 | argument, except for @samp{%s}. In this example, the value of the variable | |
3794 | @code{zebra} is printed at the location of the first @samp{%s} and the | |
3795 | value of the variable @code{tiger} is printed at the location of the | |
3796 | second @samp{%s}. | |
3797 | ||
3798 | @node Uninitialized let Variables, , Sample let Expression, let | |
3799 | @comment node-name, next, previous, up | |
3800 | @subsection Uninitialized Variables in a @code{let} Statement | |
3801 | @cindex Uninitialized @code{let} variables | |
3802 | @cindex @code{let} variables uninitialized | |
3803 | ||
3804 | If you do not bind the variables in a @code{let} statement to specific | |
3805 | initial values, they will automatically be bound to an initial value of | |
3806 | @code{nil}, as in the following expression: | |
3807 | ||
3808 | @smallexample | |
3809 | @group | |
3810 | (let ((birch 3) | |
3811 | pine | |
3812 | fir | |
3813 | (oak 'some)) | |
3814 | (message | |
3815 | "Here are %d variables with %s, %s, and %s value." | |
3816 | birch pine fir oak)) | |
3817 | @end group | |
3818 | @end smallexample | |
3819 | ||
3820 | @noindent | |
3821 | Here, the varlist is @code{((birch 3) pine fir (oak 'some))}. | |
3822 | ||
3823 | @need 1250 | |
3824 | If you evaluate this expression in the usual way, the following will | |
3825 | appear in your echo area: | |
3826 | ||
3827 | @smallexample | |
3828 | "Here are 3 variables with nil, nil, and some value." | |
3829 | @end smallexample | |
3830 | ||
3831 | @noindent | |
3832 | In this example, Emacs binds the symbol @code{birch} to the number 3, | |
3833 | binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds | |
3834 | the symbol @code{oak} to the value @code{some}. | |
3835 | ||
3836 | Note that in the first part of the @code{let}, the variables @code{pine} | |
3837 | and @code{fir} stand alone as atoms that are not surrounded by | |
3838 | parentheses; this is because they are being bound to @code{nil}, the | |
3839 | empty list. But @code{oak} is bound to @code{some} and so is a part of | |
3840 | the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the | |
3841 | number 3 and so is in a list with that number. (Since a number | |
3842 | evaluates to itself, the number does not need to be quoted. Also, the | |
3843 | number is printed in the message using a @samp{%d} rather than a | |
3844 | @samp{%s}.) The four variables as a group are put into a list to | |
3845 | delimit them from the body of the @code{let}. | |
3846 | ||
3847 | @node if, else, let, Writing Defuns | |
3848 | @comment node-name, next, previous, up | |
3849 | @section The @code{if} Special Form | |
3850 | @findex if | |
3851 | @cindex Conditional with @code{if} | |
3852 | ||
3853 | A third special form, in addition to @code{defun} and @code{let}, is the | |
3854 | conditional @code{if}. This form is used to instruct the computer to | |
3855 | make decisions. You can write function definitions without using | |
3856 | @code{if}, but it is used often enough, and is important enough, to be | |
3857 | included here. It is used, for example, in the code for the | |
3858 | function @code{beginning-of-buffer}. | |
3859 | ||
3860 | The basic idea behind an @code{if}, is that ``@emph{if} a test is true, | |
3861 | @emph{then} an expression is evaluated.'' If the test is not true, the | |
3862 | expression is not evaluated. For example, you might make a decision | |
3863 | such as, ``if it is warm and sunny, then go to the beach!'' | |
3864 | ||
3865 | @menu | |
3866 | * if in more detail:: | |
3867 | * type-of-animal in detail:: An example of an @code{if} expression. | |
3868 | @end menu | |
3869 | ||
3870 | @node if in more detail, type-of-animal in detail, if, if | |
3871 | @ifnottex | |
3872 | @unnumberedsubsec @code{if} in more detail | |
3873 | @end ifnottex | |
3874 | ||
3875 | @cindex @samp{if-part} defined | |
3876 | @cindex @samp{then-part} defined | |
3877 | An @code{if} expression written in Lisp does not use the word `then'; | |
3878 | the test and the action are the second and third elements of the list | |
3879 | whose first element is @code{if}. Nonetheless, the test part of an | |
3880 | @code{if} expression is often called the @dfn{if-part} and the second | |
3881 | argument is often called the @dfn{then-part}. | |
3882 | ||
3883 | Also, when an @code{if} expression is written, the true-or-false-test | |
3884 | is usually written on the same line as the symbol @code{if}, but the | |
3885 | action to carry out if the test is true, the ``then-part'', is written | |
3886 | on the second and subsequent lines. This makes the @code{if} | |
3887 | expression easier to read. | |
3888 | ||
3889 | @smallexample | |
3890 | @group | |
3891 | (if @var{true-or-false-test} | |
3892 | @var{action-to-carry-out-if-test-is-true}) | |
3893 | @end group | |
3894 | @end smallexample | |
3895 | ||
3896 | @noindent | |
3897 | The true-or-false-test will be an expression that | |
3898 | is evaluated by the Lisp interpreter. | |
3899 | ||
3900 | Here is an example that you can evaluate in the usual manner. The test | |
3901 | is whether the number 5 is greater than the number 4. Since it is, the | |
3902 | message @samp{5 is greater than 4!} will be printed. | |
3903 | ||
3904 | @smallexample | |
3905 | @group | |
3906 | (if (> 5 4) ; @r{if-part} | |
3907 | (message "5 is greater than 4!")) ; @r{then-part} | |
3908 | @end group | |
3909 | @end smallexample | |
3910 | ||
3911 | @noindent | |
3912 | (The function @code{>} tests whether its first argument is greater than | |
3913 | its second argument and returns true if it is.) | |
3914 | @findex > (greater than) | |
3915 | ||
3916 | Of course, in actual use, the test in an @code{if} expression will not | |
3917 | be fixed for all time as it is by the expression @code{(> 5 4)}. | |
3918 | Instead, at least one of the variables used in the test will be bound to | |
3919 | a value that is not known ahead of time. (If the value were known ahead | |
3920 | of time, we would not need to run the test!) | |
3921 | ||
3922 | For example, the value may be bound to an argument of a function | |
3923 | definition. In the following function definition, the character of the | |
3924 | animal is a value that is passed to the function. If the value bound to | |
3925 | @code{characteristic} is @code{fierce}, then the message, @samp{It's a | |
3926 | tiger!} will be printed; otherwise, @code{nil} will be returned. | |
3927 | ||
3928 | @smallexample | |
3929 | @group | |
3930 | (defun type-of-animal (characteristic) | |
3931 | "Print message in echo area depending on CHARACTERISTIC. | |
3932 | If the CHARACTERISTIC is the symbol `fierce', | |
3933 | then warn of a tiger." | |
3934 | (if (equal characteristic 'fierce) | |
3935 | (message "It's a tiger!"))) | |
3936 | @end group | |
3937 | @end smallexample | |
3938 | ||
3939 | @need 1500 | |
3940 | @noindent | |
3941 | If you are reading this inside of GNU Emacs, you can evaluate the | |
3942 | function definition in the usual way to install it in Emacs, and then you | |
3943 | can evaluate the following two expressions to see the results: | |
3944 | ||
3945 | @smallexample | |
3946 | @group | |
3947 | (type-of-animal 'fierce) | |
3948 | ||
3949 | (type-of-animal 'zebra) | |
3950 | ||
3951 | @end group | |
3952 | @end smallexample | |
3953 | ||
3954 | @c Following sentences rewritten to prevent overfull hbox. | |
3955 | @noindent | |
3956 | When you evaluate @code{(type-of-animal 'fierce)}, you will see the | |
3957 | following message printed in the echo area: @code{"It's a tiger!"}; and | |
3958 | when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil} | |
3959 | printed in the echo area. | |
3960 | ||
3961 | @node type-of-animal in detail, , if in more detail, if | |
3962 | @comment node-name, next, previous, up | |
3963 | @subsection The @code{type-of-animal} Function in Detail | |
3964 | ||
3965 | Let's look at the @code{type-of-animal} function in detail. | |
3966 | ||
3967 | The function definition for @code{type-of-animal} was written by filling | |
3968 | the slots of two templates, one for a function definition as a whole, and | |
3969 | a second for an @code{if} expression. | |
3970 | ||
3971 | @need 1250 | |
3972 | The template for every function that is not interactive is: | |
3973 | ||
3974 | @smallexample | |
3975 | @group | |
3976 | (defun @var{name-of-function} (@var{argument-list}) | |
3977 | "@var{documentation}@dots{}" | |
3978 | @var{body}@dots{}) | |
3979 | @end group | |
3980 | @end smallexample | |
3981 | ||
3982 | @need 800 | |
3983 | The parts of the function that match this template look like this: | |
3984 | ||
3985 | @smallexample | |
3986 | @group | |
3987 | (defun type-of-animal (characteristic) | |
3988 | "Print message in echo area depending on CHARACTERISTIC. | |
3989 | If the CHARACTERISTIC is the symbol `fierce', | |
3990 | then warn of a tiger." | |
3991 | @var{body: the} @code{if} @var{expression}) | |
3992 | @end group | |
3993 | @end smallexample | |
3994 | ||
3995 | The name of function is @code{type-of-animal}; it is passed the value | |
3996 | of one argument. The argument list is followed by a multi-line | |
3997 | documentation string. The documentation string is included in the | |
3998 | example because it is a good habit to write documentation string for | |
3999 | every function definition. The body of the function definition | |
4000 | consists of the @code{if} expression. | |
4001 | ||
4002 | @need 800 | |
4003 | The template for an @code{if} expression looks like this: | |
4004 | ||
4005 | @smallexample | |
4006 | @group | |
4007 | (if @var{true-or-false-test} | |
4008 | @var{action-to-carry-out-if-the-test-returns-true}) | |
4009 | @end group | |
4010 | @end smallexample | |
4011 | ||
4012 | @need 1250 | |
4013 | In the @code{type-of-animal} function, the code for the @code{if} | |
4014 | looks like this: | |
4015 | ||
4016 | @smallexample | |
4017 | @group | |
4018 | (if (equal characteristic 'fierce) | |
4019 | (message "It's a tiger!"))) | |
4020 | @end group | |
4021 | @end smallexample | |
4022 | ||
4023 | @need 800 | |
4024 | Here, the true-or-false-test is the expression: | |
4025 | ||
4026 | @smallexample | |
4027 | (equal characteristic 'fierce) | |
4028 | @end smallexample | |
4029 | ||
4030 | @noindent | |
4031 | In Lisp, @code{equal} is a function that determines whether its first | |
4032 | argument is equal to its second argument. The second argument is the | |
4033 | quoted symbol @code{'fierce} and the first argument is the value of the | |
4034 | symbol @code{characteristic}---in other words, the argument passed to | |
4035 | this function. | |
4036 | ||
4037 | In the first exercise of @code{type-of-animal}, the argument | |
4038 | @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce} | |
4039 | is equal to @code{fierce}, the expression, @code{(equal characteristic | |
4040 | 'fierce)}, returns a value of true. When this happens, the @code{if} | |
4041 | evaluates the second argument or then-part of the @code{if}: | |
4042 | @code{(message "It's tiger!")}. | |
4043 | ||
4044 | On the other hand, in the second exercise of @code{type-of-animal}, the | |
4045 | argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra} | |
4046 | is not equal to @code{fierce}, so the then-part is not evaluated and | |
4047 | @code{nil} is returned by the @code{if} expression. | |
4048 | ||
4049 | @node else, Truth & Falsehood, if, Writing Defuns | |
4050 | @comment node-name, next, previous, up | |
4051 | @section If--then--else Expressions | |
4052 | @cindex Else | |
4053 | ||
4054 | An @code{if} expression may have an optional third argument, called | |
4055 | the @dfn{else-part}, for the case when the true-or-false-test returns | |
4056 | false. When this happens, the second argument or then-part of the | |
4057 | overall @code{if} expression is @emph{not} evaluated, but the third or | |
4058 | else-part @emph{is} evaluated. You might think of this as the cloudy | |
4059 | day alternative for the decision `if it is warm and sunny, then go to | |
4060 | the beach, else read a book!''. | |
4061 | ||
4062 | The word ``else'' is not written in the Lisp code; the else-part of an | |
4063 | @code{if} expression comes after the then-part. In the written Lisp, the | |
4064 | else-part is usually written to start on a line of its own and is | |
4065 | indented less than the then-part: | |
4066 | ||
4067 | @smallexample | |
4068 | @group | |
4069 | (if @var{true-or-false-test} | |
4070 | @var{action-to-carry-out-if-the-test-returns-true} | |
4071 | @var{action-to-carry-out-if-the-test-returns-false}) | |
4072 | @end group | |
4073 | @end smallexample | |
4074 | ||
4075 | For example, the following @code{if} expression prints the message @samp{4 | |
4076 | is not greater than 5!} when you evaluate it in the usual way: | |
4077 | ||
4078 | @smallexample | |
4079 | @group | |
4080 | (if (> 4 5) ; @r{if-part} | |
4081 | (message "5 is greater than 4!") ; @r{then-part} | |
4082 | (message "4 is not greater than 5!")) ; @r{else-part} | |
4083 | @end group | |
4084 | @end smallexample | |
4085 | ||
4086 | @noindent | |
4087 | Note that the different levels of indentation make it easy to | |
4088 | distinguish the then-part from the else-part. (GNU Emacs has several | |
4089 | commands that automatically indent @code{if} expressions correctly. | |
4090 | @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.) | |
4091 | ||
4092 | We can extend the @code{type-of-animal} function to include an | |
4093 | else-part by simply incorporating an additional part to the @code{if} | |
4094 | expression. | |
4095 | ||
4096 | @need 1500 | |
4097 | You can see the consequences of doing this if you evaluate the following | |
4098 | version of the @code{type-of-animal} function definition to install it | |
4099 | and then evaluate the two subsequent expressions to pass different | |
4100 | arguments to the function. | |
4101 | ||
4102 | @smallexample | |
4103 | @group | |
4104 | (defun type-of-animal (characteristic) ; @r{Second version.} | |
4105 | "Print message in echo area depending on CHARACTERISTIC. | |
4106 | If the CHARACTERISTIC is the symbol `fierce', | |
4107 | then warn of a tiger; | |
4108 | else say it's not fierce." | |
4109 | (if (equal characteristic 'fierce) | |
4110 | (message "It's a tiger!") | |
4111 | (message "It's not fierce!"))) | |
4112 | @end group | |
4113 | @end smallexample | |
4114 | @sp 1 | |
4115 | ||
4116 | @smallexample | |
4117 | @group | |
4118 | (type-of-animal 'fierce) | |
4119 | ||
4120 | (type-of-animal 'zebra) | |
4121 | ||
4122 | @end group | |
4123 | @end smallexample | |
4124 | ||
4125 | @c Following sentence rewritten to prevent overfull hbox. | |
4126 | @noindent | |
4127 | When you evaluate @code{(type-of-animal 'fierce)}, you will see the | |
4128 | following message printed in the echo area: @code{"It's a tiger!"}; but | |
4129 | when you evaluate @code{(type-of-animal 'zebra)}, you will see | |
4130 | @code{"It's not fierce!"}. | |
4131 | ||
4132 | (Of course, if the @var{characteristic} were @code{ferocious}, the | |
4133 | message @code{"It's not fierce!"} would be printed; and it would be | |
4134 | misleading! When you write code, you need to take into account the | |
4135 | possibility that some such argument will be tested by the @code{if} and | |
4136 | write your program accordingly.) | |
4137 | ||
4138 | @node Truth & Falsehood, save-excursion, else, Writing Defuns | |
4139 | @comment node-name, next, previous, up | |
4140 | @section Truth and Falsehood in Emacs Lisp | |
4141 | @cindex Truth and falsehood in Emacs Lisp | |
4142 | @cindex Falsehood and truth in Emacs Lisp | |
4143 | @findex nil | |
4144 | ||
4145 | There is an important aspect to the truth test in an @code{if} | |
4146 | expression. So far, we have spoken of `true' and `false' as values of | |
4147 | predicates as if they were new kinds of Emacs Lisp objects. In fact, | |
4148 | `false' is just our old friend @code{nil}. Anything else---anything | |
4149 | at all---is `true'. | |
4150 | ||
4151 | The expression that tests for truth is interpreted as @dfn{true} | |
4152 | if the result of evaluating it is a value that is not @code{nil}. In | |
4153 | other words, the result of the test is considered true if the value | |
4154 | returned is a number such as 47, a string such as @code{"hello"}, or a | |
4155 | symbol (other than @code{nil}) such as @code{flowers}, or a list, or | |
4156 | even a buffer! | |
4157 | ||
4158 | @menu | |
4159 | * nil explained:: @code{nil} has two meanings. | |
4160 | @end menu | |
4161 | ||
4162 | @node nil explained, , Truth & Falsehood, Truth & Falsehood | |
4163 | @ifnottex | |
4164 | @unnumberedsubsec An explanation of @code{nil} | |
4165 | @end ifnottex | |
4166 | ||
4167 | Before illustrating a test for truth, we need an explanation of @code{nil}. | |
4168 | ||
4169 | In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the | |
4170 | empty list. Second, it means false and is the value returned when a | |
4171 | true-or-false-test tests false. @code{nil} can be written as an empty | |
4172 | list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is | |
4173 | concerned, @code{()} and @code{nil} are the same. Humans, however, tend | |
4174 | to use @code{nil} for false and @code{()} for the empty list. | |
4175 | ||
4176 | In Emacs Lisp, any value that is not @code{nil}---is not the empty | |
4177 | list---is considered true. This means that if an evaluation returns | |
4178 | something that is not an empty list, an @code{if} expression will test | |
4179 | true. For example, if a number is put in the slot for the test, it | |
4180 | will be evaluated and will return itself, since that is what numbers | |
4181 | do when evaluated. In this conditional, the @code{if} expression will | |
4182 | test true. The expression tests false only when @code{nil}, an empty | |
4183 | list, is returned by evaluating the expression. | |
4184 | ||
4185 | You can see this by evaluating the two expressions in the following examples. | |
4186 | ||
4187 | In the first example, the number 4 is evaluated as the test in the | |
4188 | @code{if} expression and returns itself; consequently, the then-part | |
4189 | of the expression is evaluated and returned: @samp{true} appears in | |
4190 | the echo area. In the second example, the @code{nil} indicates false; | |
4191 | consequently, the else-part of the expression is evaluated and | |
4192 | returned: @samp{false} appears in the echo area. | |
4193 | ||
4194 | @smallexample | |
4195 | @group | |
4196 | (if 4 | |
4197 | 'true | |
4198 | 'false) | |
4199 | @end group | |
4200 | ||
4201 | @group | |
4202 | (if nil | |
4203 | 'true | |
4204 | 'false) | |
4205 | @end group | |
4206 | @end smallexample | |
4207 | ||
4208 | @need 1250 | |
4209 | Incidentally, if some other useful value is not available for a test that | |
4210 | returns true, then the Lisp interpreter will return the symbol @code{t} | |
4211 | for true. For example, the expression @code{(> 5 4)} returns @code{t} | |
4212 | when evaluated, as you can see by evaluating it in the usual way: | |
4213 | ||
4214 | @smallexample | |
4215 | (> 5 4) | |
4216 | @end smallexample | |
4217 | ||
4218 | @need 1250 | |
4219 | @noindent | |
4220 | On the other hand, this function returns @code{nil} if the test is false. | |
4221 | ||
4222 | @smallexample | |
4223 | (> 4 5) | |
4224 | @end smallexample | |
4225 | ||
4226 | @node save-excursion, Review, Truth & Falsehood, Writing Defuns | |
4227 | @comment node-name, next, previous, up | |
4228 | @section @code{save-excursion} | |
4229 | @findex save-excursion | |
4230 | @cindex Region, what it is | |
4231 | @cindex Preserving point, mark, and buffer | |
4232 | @cindex Point, mark, buffer preservation | |
4233 | @findex point | |
4234 | @findex mark | |
4235 | ||
4236 | The @code{save-excursion} function is the fourth and final special form | |
4237 | that we will discuss in this chapter. | |
4238 | ||
4239 | In Emacs Lisp programs used for editing, the @code{save-excursion} | |
4240 | function is very common. It saves the location of point and mark, | |
4241 | executes the body of the function, and then restores point and mark to | |
4242 | their previous positions if their locations were changed. Its primary | |
4243 | purpose is to keep the user from being surprised and disturbed by | |
4244 | unexpected movement of point or mark. | |
4245 | ||
4246 | @menu | |
4247 | * Point and mark:: A review of various locations. | |
4248 | * Template for save-excursion:: | |
4249 | @end menu | |
4250 | ||
4251 | @node Point and mark, Template for save-excursion, save-excursion, save-excursion | |
4252 | @ifnottex | |
4253 | @unnumberedsubsec Point and Mark | |
4254 | @end ifnottex | |
4255 | ||
4256 | Before discussing @code{save-excursion}, however, it may be useful | |
4257 | first to review what point and mark are in GNU Emacs. @dfn{Point} is | |
4258 | the current location of the cursor. Wherever the cursor | |
4259 | is, that is point. More precisely, on terminals where the cursor | |
4260 | appears to be on top of a character, point is immediately before the | |
4261 | character. In Emacs Lisp, point is an integer. The first character in | |
4262 | a buffer is number one, the second is number two, and so on. The | |
4263 | function @code{point} returns the current position of the cursor as a | |
4264 | number. Each buffer has its own value for point. | |
4265 | ||
4266 | The @dfn{mark} is another position in the buffer; its value can be set | |
4267 | with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If | |
4268 | a mark has been set, you can use the command @kbd{C-x C-x} | |
4269 | (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark | |
4270 | and set the mark to be the previous position of point. In addition, if | |
4271 | you set another mark, the position of the previous mark is saved in the | |
4272 | mark ring. Many mark positions can be saved this way. You can jump the | |
4273 | cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more | |
4274 | times. | |
4275 | ||
4276 | The part of the buffer between point and mark is called @dfn{the | |
4277 | region}. Numerous commands work on the region, including | |
4278 | @code{center-region}, @code{count-lines-region}, @code{kill-region}, and | |
4279 | @code{print-region}. | |
4280 | ||
4281 | The @code{save-excursion} special form saves the locations of point and | |
4282 | mark and restores those positions after the code within the body of the | |
4283 | special form is evaluated by the Lisp interpreter. Thus, if point were | |
4284 | in the beginning of a piece of text and some code moved point to the end | |
4285 | of the buffer, the @code{save-excursion} would put point back to where | |
4286 | it was before, after the expressions in the body of the function were | |
4287 | evaluated. | |
4288 | ||
4289 | In Emacs, a function frequently moves point as part of its internal | |
4290 | workings even though a user would not expect this. For example, | |
4291 | @code{count-lines-region} moves point. To prevent the user from being | |
4292 | bothered by jumps that are both unexpected and (from the user's point of | |
4293 | view) unnecessary, @code{save-excursion} is often used to keep point and | |
4294 | mark in the location expected by the user. The use of | |
4295 | @code{save-excursion} is good housekeeping. | |
4296 | ||
4297 | To make sure the house stays clean, @code{save-excursion} restores the | |
4298 | values of point and mark even if something goes wrong in the code inside | |
4299 | of it (or, to be more precise and to use the proper jargon, ``in case of | |
4300 | abnormal exit''). This feature is very helpful. | |
4301 | ||
4302 | In addition to recording the values of point and mark, | |
4303 | @code{save-excursion} keeps track of the current buffer, and restores | |
4304 | it, too. This means you can write code that will change the buffer and | |
4305 | have @code{save-excursion} switch you back to the original buffer. This | |
4306 | is how @code{save-excursion} is used in @code{append-to-buffer}. | |
4307 | (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.) | |
4308 | ||
4309 | @node Template for save-excursion, , Point and mark, save-excursion | |
4310 | @comment node-name, next, previous, up | |
4311 | @subsection Template for a @code{save-excursion} Expression | |
4312 | ||
4313 | @need 800 | |
4314 | The template for code using @code{save-excursion} is simple: | |
4315 | ||
4316 | @smallexample | |
4317 | @group | |
4318 | (save-excursion | |
4319 | @var{body}@dots{}) | |
4320 | @end group | |
4321 | @end smallexample | |
4322 | ||
4323 | @noindent | |
4324 | The body of the function is one or more expressions that will be | |
4325 | evaluated in sequence by the Lisp interpreter. If there is more than | |
4326 | one expression in the body, the value of the last one will be returned | |
4327 | as the value of the @code{save-excursion} function. The other | |
4328 | expressions in the body are evaluated only for their side effects; and | |
4329 | @code{save-excursion} itself is used only for its side effect (which | |
4330 | is restoring the positions of point and mark). | |
4331 | ||
4332 | @need 1250 | |
4333 | In more detail, the template for a @code{save-excursion} expression | |
4334 | looks like this: | |
4335 | ||
4336 | @smallexample | |
4337 | @group | |
4338 | (save-excursion | |
4339 | @var{first-expression-in-body} | |
4340 | @var{second-expression-in-body} | |
4341 | @var{third-expression-in-body} | |
4342 | @dots{} | |
4343 | @var{last-expression-in-body}) | |
4344 | @end group | |
4345 | @end smallexample | |
4346 | ||
4347 | @noindent | |
4348 | An expression, of course, may be a symbol on its own or a list. | |
4349 | ||
4350 | In Emacs Lisp code, a @code{save-excursion} expression often occurs | |
4351 | within the body of a @code{let} expression. It looks like this: | |
4352 | ||
4353 | @smallexample | |
4354 | @group | |
4355 | (let @var{varlist} | |
4356 | (save-excursion | |
4357 | @var{body}@dots{})) | |
4358 | @end group | |
4359 | @end smallexample | |
4360 | ||
4361 | @node Review, defun Exercises, save-excursion, Writing Defuns | |
4362 | @comment node-name, next, previous, up | |
4363 | @section Review | |
4364 | ||
4365 | In the last few chapters we have introduced a fair number of functions | |
4366 | and special forms. Here they are described in brief, along with a few | |
4367 | similar functions that have not been mentioned yet. | |
4368 | ||
4369 | @table @code | |
4370 | @item eval-last-sexp | |
4371 | Evaluate the last symbolic expression before the current location of | |
4372 | point. The value is printed in the echo area unless the function is | |
4373 | invoked with an argument; in that case, the output is printed in the | |
4374 | current buffer. This command is normally bound to @kbd{C-x C-e}. | |
4375 | ||
4376 | @item defun | |
4377 | Define function. This special form has up to five parts: the name, | |
4378 | a template for the arguments that will be passed to the function, | |
4379 | documentation, an optional interactive declaration, and the body of the | |
4380 | definition. | |
4381 | ||
4382 | @need 1250 | |
4383 | For example: | |
4384 | ||
4385 | @smallexample | |
4386 | @group | |
4387 | (defun back-to-indentation () | |
4388 | "Move point to first visible character on line." | |
4389 | (interactive) | |
4390 | (beginning-of-line 1) | |
4391 | (skip-chars-forward " \t")) | |
4392 | @end group | |
4393 | @end smallexample | |
4394 | ||
4395 | @item interactive | |
4396 | Declare to the interpreter that the function can be used | |
4397 | interactively. This special form may be followed by a string with one | |
4398 | or more parts that pass the information to the arguments of the | |
4399 | function, in sequence. These parts may also tell the interpreter to | |
4400 | prompt for information. Parts of the string are separated by | |
4401 | newlines, @samp{\n}. | |
4402 | ||
4403 | Common code characters are: | |
4404 | ||
4405 | @table @code | |
4406 | @item b | |
4407 | The name of an existing buffer. | |
4408 | ||
4409 | @item f | |
4410 | The name of an existing file. | |
4411 | ||
4412 | @item p | |
4413 | The numeric prefix argument. (Note that this `p' is lower case.) | |
4414 | ||
4415 | @item r | |
4416 | Point and the mark, as two numeric arguments, smallest first. This | |
4417 | is the only code letter that specifies two successive arguments | |
4418 | rather than one. | |
4419 | @end table | |
4420 | ||
4421 | @xref{Interactive Codes, , Code Characters for @samp{interactive}, | |
4422 | elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of | |
4423 | code characters. | |
4424 | ||
4425 | @item let | |
4426 | Declare that a list of variables is for use within the body of the | |
4427 | @code{let} and give them an initial value, either @code{nil} or a | |
4428 | specified value; then evaluate the rest of the expressions in the body | |
4429 | of the @code{let} and return the value of the last one. Inside the | |
4430 | body of the @code{let}, the Lisp interpreter does not see the values of | |
4431 | the variables of the same names that are bound outside of the | |
4432 | @code{let}. | |
4433 | ||
4434 | @need 1250 | |
4435 | For example, | |
4436 | ||
4437 | @smallexample | |
4438 | @group | |
4439 | (let ((foo (buffer-name)) | |
4440 | (bar (buffer-size))) | |
4441 | (message | |
4442 | "This buffer is %s and has %d characters." | |
4443 | foo bar)) | |
4444 | @end group | |
4445 | @end smallexample | |
4446 | ||
4447 | @item save-excursion | |
4448 | Record the values of point and mark and the current buffer before | |
4449 | evaluating the body of this special form. Restore the values of point | |
4450 | and mark and buffer afterward. | |
4451 | ||
4452 | @need 1250 | |
4453 | For example, | |
4454 | ||
4455 | @smallexample | |
4456 | @group | |
4457 | (message "We are %d characters into this buffer." | |
4458 | (- (point) | |
4459 | (save-excursion | |
4460 | (goto-char (point-min)) (point)))) | |
4461 | @end group | |
4462 | @end smallexample | |
4463 | ||
4464 | @item if | |
4465 | Evaluate the first argument to the function; if it is true, evaluate | |
4466 | the second argument; else evaluate the third argument, if there is one. | |
4467 | ||
4468 | The @code{if} special form is called a @dfn{conditional}. There are | |
4469 | other conditionals in Emacs Lisp, but @code{if} is perhaps the most | |
4470 | commonly used. | |
4471 | ||
4472 | @need 1250 | |
4473 | For example, | |
4474 | ||
4475 | @smallexample | |
4476 | @group | |
4477 | (if (string-equal | |
4478 | (number-to-string 21) | |
4479 | (substring (emacs-version) 10 12)) | |
4480 | (message "This is version 21 Emacs") | |
4481 | (message "This is not version 21 Emacs")) | |
4482 | @end group | |
4483 | @end smallexample | |
4484 | ||
4485 | @item equal | |
4486 | @itemx eq | |
4487 | Test whether two objects are the same. @code{equal} uses one meaning | |
4488 | of the word `same' and @code{eq} uses another: @code{equal} returns | |
4489 | true if the two objects have a similar structure and contents, such as | |
4490 | two copies of the same book. On the other hand, @code{eq}, returns | |
4491 | true if both arguments are actually the same object. | |
4492 | @findex equal | |
4493 | @findex eq | |
4494 | ||
4495 | @need 1250 | |
4496 | @item < | |
4497 | @itemx > | |
4498 | @itemx <= | |
4499 | @itemx >= | |
4500 | The @code{<} function tests whether its first argument is smaller than | |
4501 | its second argument. A corresponding function, @code{>}, tests whether | |
4502 | the first argument is greater than the second. Likewise, @code{<=} | |
4503 | tests whether the first argument is less than or equal to the second and | |
4504 | @code{>=} tests whether the first argument is greater than or equal to | |
4505 | the second. In all cases, both arguments must be numbers or markers | |
4506 | (markers indicate positions in buffers). | |
4507 | ||
4508 | @item string< | |
4509 | @itemx string-lessp | |
4510 | @itemx string= | |
4511 | @itemx string-equal | |
4512 | The @code{string-lessp} function tests whether its first argument is | |
4513 | smaller than the second argument. A shorter, alternative name for the | |
4514 | same function (a @code{defalias}) is @code{string<}. | |
4515 | ||
4516 | The arguments to @code{string-lessp} must be strings or symbols; the | |
4517 | ordering is lexicographic, so case is significant. The print names of | |
4518 | symbols are used instead of the symbols themselves. | |
4519 | ||
4520 | @code{string-equal} provides the corresponding test for equality. Its | |
4521 | shorter, alternative name is @code{string=}. There are no string test | |
4522 | functions that correspond to @var{>}, @code{>=}, or @code{<=}. | |
4523 | ||
4524 | @item message | |
4525 | Print a message in the echo area. The first argument is a string that | |
4526 | can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of | |
4527 | arguments that follow the string. The argument used by @samp{%s} must | |
4528 | be a string or a symbol; the argument used by @samp{%d} must be a | |
4529 | number. The argument used by @samp{%c} must be an ascii code number; | |
4530 | it will be printed as the character with that @sc{ascii} code. | |
4531 | ||
4532 | @item setq | |
4533 | @itemx set | |
4534 | The @code{setq} function sets the value of its first argument to the | |
4535 | value of the second argument. The first argument is automatically | |
4536 | quoted by @code{setq}. It does the same for succeeding pairs of | |
4537 | arguments. Another function, @code{set}, takes only two arguments and | |
4538 | evaluates both of them before setting the value returned by its first | |
4539 | argument to the value returned by its second argument. | |
4540 | ||
4541 | @item buffer-name | |
4542 | Without an argument, return the name of the buffer, as a string. | |
4543 | ||
4544 | @itemx buffer-file-name | |
4545 | Without an argument, return the name of the file the buffer is | |
4546 | visiting. | |
4547 | ||
4548 | @item current-buffer | |
4549 | Return the buffer in which Emacs is active; it may not be | |
4550 | the buffer that is visible on the screen. | |
4551 | ||
4552 | @item other-buffer | |
4553 | Return the most recently selected buffer (other than the buffer passed | |
4554 | to @code{other-buffer} as an argument and other than the current | |
4555 | buffer). | |
4556 | ||
4557 | @item switch-to-buffer | |
4558 | Select a buffer for Emacs to be active in and display it in the current | |
4559 | window so users can look at it. Usually bound to @kbd{C-x b}. | |
4560 | ||
4561 | @item set-buffer | |
4562 | Switch Emacs' attention to a buffer on which programs will run. Don't | |
4563 | alter what the window is showing. | |
4564 | ||
4565 | @item buffer-size | |
4566 | Return the number of characters in the current buffer. | |
4567 | ||
4568 | @item point | |
4569 | Return the value of the current position of the cursor, as an | |
4570 | integer counting the number of characters from the beginning of the | |
4571 | buffer. | |
4572 | ||
4573 | @item point-min | |
4574 | Return the minimum permissible value of point in | |
4575 | the current buffer. This is 1, unless narrowing is in effect. | |
4576 | ||
4577 | @item point-max | |
4578 | Return the value of the maximum permissible value of point in the | |
4579 | current buffer. This is the end of the buffer, unless narrowing is in | |
4580 | effect. | |
4581 | @end table | |
4582 | ||
4583 | @need 1500 | |
4584 | @node defun Exercises, , Review, Writing Defuns | |
4585 | @section Exercises | |
4586 | ||
4587 | @itemize @bullet | |
4588 | @item | |
4589 | Write a non-interactive function that doubles the value of its | |
4590 | argument, a number. Make that function interactive. | |
4591 | ||
4592 | @item | |
4593 | Write a function that tests whether the current value of | |
4594 | @code{fill-column} is greater than the argument passed to the function, | |
4595 | and if so, prints an appropriate message. | |
4596 | @end itemize | |
4597 | ||
4598 | @node Buffer Walk Through, More Complex, Writing Defuns, Top | |
4599 | @comment node-name, next, previous, up | |
4600 | @chapter A Few Buffer--Related Functions | |
4601 | ||
4602 | In this chapter we study in detail several of the functions used in GNU | |
4603 | Emacs. This is called a ``walk-through''. These functions are used as | |
4604 | examples of Lisp code, but are not imaginary examples; with the | |
4605 | exception of the first, simplified function definition, these functions | |
4606 | show the actual code used in GNU Emacs. You can learn a great deal from | |
4607 | these definitions. The functions described here are all related to | |
4608 | buffers. Later, we will study other functions. | |
4609 | ||
4610 | @menu | |
4611 | * Finding More:: How to find more information. | |
4612 | * simplified-beginning-of-buffer:: Shows @code{goto-char}, | |
4613 | @code{point-min}, and @code{push-mark}. | |
4614 | * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}. | |
4615 | * append-to-buffer:: Uses @code{save-excursion} and | |
4616 | @code{insert-buffer-substring}. | |
4617 | * Buffer Related Review:: Review. | |
4618 | * Buffer Exercises:: | |
4619 | @end menu | |
4620 | ||
4621 | @node Finding More, simplified-beginning-of-buffer, Buffer Walk Through, Buffer Walk Through | |
4622 | @section Finding More Information | |
4623 | ||
4624 | @findex describe-function, @r{introduced} | |
4625 | @cindex Find function documentation | |
4626 | In this walk-through, I will describe each new function as we come to | |
4627 | it, sometimes in detail and sometimes briefly. If you are interested, | |
4628 | you can get the full documentation of any Emacs Lisp function at any | |
4629 | time by typing @kbd{C-h f} and then the name of the function (and then | |
4630 | @key{RET}). Similarly, you can get the full documentation for a | |
4631 | variable by typing @kbd{C-h v} and then the name of the variable (and | |
4632 | then @key{RET}). | |
4633 | ||
4634 | @cindex Find source of function | |
4635 | In versions 20 and higher, when a function is written in Emacs Lisp, | |
4636 | @code{describe-function} will also tell you the location of the | |
4637 | function definition. If you move point over the file name and press | |
4638 | the @key{RET} key, which is this case means @code{help-follow} rather | |
4639 | than `return' or `enter', Emacs will take you directly to the function | |
4640 | definition. | |
4641 | ||
4642 | More generally, if you want to see a function in its original source | |
4643 | file, you can use the @code{find-tags} function to jump to it. | |
4644 | @code{find-tags} works with a wide variety of languages, not just | |
4645 | Lisp, and C, and it works with non-programming text as well. For | |
4646 | example, @code{find-tags} will jump to the various nodes in the | |
4647 | Texinfo source file of this document. | |
4648 | ||
4649 | The @code{find-tags} function depends on `tags tables' that record | |
4650 | the locations of the functions, variables, and other items to which | |
4651 | @code{find-tags} jumps. | |
4652 | ||
4653 | To use the @code{find-tags} command, type @kbd{M-.} (i.e., type the | |
4654 | @key{META} key and the period key at the same time, or else type the | |
4655 | @key{ESC} key and then type the period key), and then, at the prompt, | |
4656 | type in the name of the function whose source code you want to see, | |
4657 | such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will | |
4658 | switch buffers and display the source code for the function on your | |
4659 | screen. To switch back to your current buffer, type @kbd{C-x b | |
4660 | @key{RET}}. (On some keyboards, the @key{META} key is labelled | |
4661 | @key{ALT}.) | |
4662 | ||
4663 | @c !!! 21.0.100 tags table location in this paragraph | |
4664 | @cindex TAGS table, specifying | |
4665 | @findex find-tags | |
4666 | Depending on how the initial default values of your copy of Emacs are | |
4667 | set, you may also need to specify the location of your `tags table', | |
4668 | which is a file called @file{TAGS}. For example, if you are | |
4669 | interested in Emacs sources, the tags table you will most likely want, | |
4670 | if it has already been created for you, will be in a subdirectory of | |
4671 | the @file{/usr/local/share/emacs/} directory; thus you would use the | |
4672 | @code{M-x visit-tags-table} command and specify a pathname such as | |
4673 | @file{/usr/local/share/emacs/21.0.100/lisp/TAGS} or | |
4674 | @file{/usr/local/src/emacs/lisp/TAGS}. If the tags table has | |
4675 | not already been created, you will have to create it yourself. | |
4676 | ||
4677 | @need 1250 | |
4678 | To create a @file{TAGS} file in a specific directory, switch to that | |
4679 | directory in Emacs using @kbd{M-x cd} command, or list the directory | |
4680 | with @kbd{C-x d} (@code{dired}). Then run the compile command, with | |
4681 | @w{@code{etags *.el}} as the command to execute | |
4682 | ||
4683 | @smallexample | |
4684 | M-x compile RET etags *.el RET | |
4685 | @end smallexample | |
4686 | ||
4687 | For more information, see @ref{etags, , Create Your Own @file{TAGS} File}. | |
4688 | ||
4689 | After you become more familiar with Emacs Lisp, you will find that you will | |
4690 | frequently use @code{find-tags} to navigate your way around source code; | |
4691 | and you will create your own @file{TAGS} tables. | |
4692 | ||
4693 | @cindex Library, as term for `file' | |
4694 | Incidentally, the files that contain Lisp code are conventionally | |
4695 | called @dfn{libraries}. The metaphor is derived from that of a | |
4696 | specialized library, such as a law library or an engineering library, | |
4697 | rather than a general library. Each library, or file, contains | |
4698 | functions that relate to a particular topic or activity, such as | |
4699 | @file{abbrev.el} for handling abbreviations and other typing | |
4700 | shortcuts, and @file{help.el} for on-line help. (Sometimes several | |
4701 | libraries provide code for a single activity, as the various | |
4702 | @file{rmail@dots{}} files provide code for reading electronic mail.) | |
4703 | In @cite{The GNU Emacs Manual}, you will see sentences such as ``The | |
4704 | @kbd{C-h p} command lets you search the standard Emacs Lisp libraries | |
4705 | by topic keywords.'' | |
4706 | ||
4707 | @node simplified-beginning-of-buffer, mark-whole-buffer, Finding More, Buffer Walk Through | |
4708 | @comment node-name, next, previous, up | |
4709 | @section A Simplified @code{beginning-of-buffer} Definition | |
4710 | @findex simplified-beginning-of-buffer | |
4711 | ||
4712 | The @code{beginning-of-buffer} command is a good function to start with | |
4713 | since you are likely to be familiar with it and it is easy to | |
4714 | understand. Used as an interactive command, @code{beginning-of-buffer} | |
4715 | moves the cursor to the beginning of the buffer, leaving the mark at the | |
4716 | previous position. It is generally bound to @kbd{M-<}. | |
4717 | ||
4718 | In this section, we will discuss a shortened version of the function | |
4719 | that shows how it is most frequently used. This shortened function | |
4720 | works as written, but it does not contain the code for a complex option. | |
4721 | In another section, we will describe the entire function. | |
4722 | (@xref{beginning-of-buffer, , Complete Definition of | |
4723 | @code{beginning-of-buffer}}.) | |
4724 | ||
4725 | Before looking at the code, let's consider what the function | |
4726 | definition has to contain: it must include an expression that makes | |
4727 | the function interactive so it can be called by typing @kbd{M-x | |
4728 | beginning-of-buffer} or by typing a keychord such as @kbd{C-<}; it | |
4729 | must include code to leave a mark at the original position in the | |
4730 | buffer; and it must include code to move the cursor to the beginning | |
4731 | of the buffer. | |
4732 | ||
4733 | @need 1250 | |
4734 | Here is the complete text of the shortened version of the function: | |
4735 | ||
4736 | @smallexample | |
4737 | @group | |
4738 | (defun simplified-beginning-of-buffer () | |
4739 | "Move point to the beginning of the buffer; | |
4740 | leave mark at previous position." | |
4741 | (interactive) | |
4742 | (push-mark) | |
4743 | (goto-char (point-min))) | |
4744 | @end group | |
4745 | @end smallexample | |
4746 | ||
4747 | Like all function definitions, this definition has five parts following | |
4748 | the special form @code{defun}: | |
4749 | ||
4750 | @enumerate | |
4751 | @item | |
4752 | The name: in this example, @code{simplified-beginning-of-buffer}. | |
4753 | ||
4754 | @item | |
4755 | A list of the arguments: in this example, an empty list, @code{()}, | |
4756 | ||
4757 | @item | |
4758 | The documentation string. | |
4759 | ||
4760 | @item | |
4761 | The interactive expression. | |
4762 | ||
4763 | @item | |
4764 | The body. | |
4765 | @end enumerate | |
4766 | ||
4767 | @noindent | |
4768 | In this function definition, the argument list is empty; this means that | |
4769 | this function does not require any arguments. (When we look at the | |
4770 | definition for the complete function, we will see that it may be passed | |
4771 | an optional argument.) | |
4772 | ||
4773 | The interactive expression tells Emacs that the function is intended to | |
4774 | be used interactively. In this example, @code{interactive} does not have | |
4775 | an argument because @code{simplified-beginning-of-buffer} does not | |
4776 | require one. | |
4777 | ||
4778 | @need 800 | |
4779 | The body of the function consists of the two lines: | |
4780 | ||
4781 | @smallexample | |
4782 | @group | |
4783 | (push-mark) | |
4784 | (goto-char (point-min)) | |
4785 | @end group | |
4786 | @end smallexample | |
4787 | ||
4788 | The first of these lines is the expression, @code{(push-mark)}. When | |
4789 | this expression is evaluated by the Lisp interpreter, it sets a mark at | |
4790 | the current position of the cursor, wherever that may be. The position | |
4791 | of this mark is saved in the mark ring. | |
4792 | ||
4793 | The next line is @code{(goto-char (point-min))}. This expression | |
4794 | jumps the cursor to the minimum point in the buffer, that is, to the | |
4795 | beginning of the buffer (or to the beginning of the accessible portion | |
4796 | of the buffer if it is narrowed. @xref{Narrowing & Widening, , | |
4797 | Narrowing and Widening}.) | |
4798 | ||
4799 | The @code{push-mark} command sets a mark at the place where the cursor | |
4800 | was located before it was moved to the beginning of the buffer by the | |
4801 | @code{(goto-char (point-min))} expression. Consequently, you can, if | |
4802 | you wish, go back to where you were originally by typing @kbd{C-x C-x}. | |
4803 | ||
4804 | That is all there is to the function definition! | |
4805 | ||
4806 | @findex describe-function | |
4807 | When you are reading code such as this and come upon an unfamiliar | |
4808 | function, such as @code{goto-char}, you can find out what it does by | |
4809 | using the @code{describe-function} command. To use this command, type | |
4810 | @kbd{C-h f} and then type in the name of the function and press | |
4811 | @key{RET}. The @code{describe-function} command will print the | |
4812 | function's documentation string in a @file{*Help*} window. For | |
4813 | example, the documentation for @code{goto-char} is: | |
4814 | ||
4815 | @smallexample | |
4816 | @group | |
4817 | One arg, a number. Set point to that number. | |
4818 | Beginning of buffer is position (point-min), | |
4819 | end is (point-max). | |
4820 | @end group | |
4821 | @end smallexample | |
4822 | ||
4823 | @noindent | |
4824 | (The prompt for @code{describe-function} will offer you the symbol | |
4825 | under or preceding the cursor, so you can save typing by positioning | |
4826 | the cursor right over or after the function and then typing @kbd{C-h f | |
4827 | @key{RET}}.) | |
4828 | ||
4829 | The @code{end-of-buffer} function definition is written in the same way as | |
4830 | the @code{beginning-of-buffer} definition except that the body of the | |
4831 | function contains the expression @code{(goto-char (point-max))} in place | |
4832 | of @code{(goto-char (point-min))}. | |
4833 | ||
4834 | @node mark-whole-buffer, append-to-buffer, simplified-beginning-of-buffer, Buffer Walk Through | |
4835 | @comment node-name, next, previous, up | |
4836 | @section The Definition of @code{mark-whole-buffer} | |
4837 | @findex mark-whole-buffer | |
4838 | ||
4839 | The @code{mark-whole-buffer} function is no harder to understand than the | |
4840 | @code{simplified-beginning-of-buffer} function. In this case, however, | |
4841 | we will look at the complete function, not a shortened version. | |
4842 | ||
4843 | The @code{mark-whole-buffer} function is not as commonly used as the | |
4844 | @code{beginning-of-buffer} function, but is useful nonetheless: it | |
4845 | marks a whole buffer as a region by putting point at the beginning and | |
4846 | a mark at the end of the buffer. It is generally bound to @kbd{C-x | |
4847 | h}. | |
4848 | ||
4849 | ||
4850 | @menu | |
4851 | * mark-whole-buffer overview:: | |
4852 | * Body of mark-whole-buffer:: Only three lines of code. | |
4853 | @end menu | |
4854 | ||
4855 | ||
4856 | @node mark-whole-buffer overview, Body of mark-whole-buffer, mark-whole-buffer, mark-whole-buffer | |
4857 | @ifnottex | |
4858 | @unnumberedsubsec An overview of @code{mark-whole-buffer} | |
4859 | @end ifnottex | |
4860 | ||
4861 | @need 1250 | |
4862 | In GNU Emacs 20, the code for the complete function looks like this: | |
4863 | ||
4864 | @smallexample | |
4865 | @group | |
4866 | (defun mark-whole-buffer () | |
4867 | "Put point at beginning and mark at end of buffer." | |
4868 | (interactive) | |
4869 | (push-mark (point)) | |
4870 | (push-mark (point-max)) | |
4871 | (goto-char (point-min))) | |
4872 | @end group | |
4873 | @end smallexample | |
4874 | ||
4875 | @need 1250 | |
4876 | Like all other functions, the @code{mark-whole-buffer} function fits | |
4877 | into the template for a function definition. The template looks like | |
4878 | this: | |
4879 | ||
4880 | @smallexample | |
4881 | @group | |
4882 | (defun @var{name-of-function} (@var{argument-list}) | |
4883 | "@var{documentation}@dots{}" | |
4884 | (@var{interactive-expression}@dots{}) | |
4885 | @var{body}@dots{}) | |
4886 | @end group | |
4887 | @end smallexample | |
4888 | ||
4889 | Here is how the function works: the name of the function is | |
4890 | @code{mark-whole-buffer}; it is followed by an empty argument list, | |
4891 | @samp{()}, which means that the function does not require arguments. | |
4892 | The documentation comes next. | |
4893 | ||
4894 | The next line is an @code{(interactive)} expression that tells Emacs | |
4895 | that the function will be used interactively. These details are similar | |
4896 | to the @code{simplified-beginning-of-buffer} function described in the | |
4897 | previous section. | |
4898 | ||
4899 | @node Body of mark-whole-buffer, , mark-whole-buffer overview, mark-whole-buffer | |
4900 | @comment node-name, next, previous, up | |
4901 | @subsection Body of @code{mark-whole-buffer} | |
4902 | ||
4903 | The body of the @code{mark-whole-buffer} function consists of three | |
4904 | lines of code: | |
4905 | ||
4906 | @smallexample | |
4907 | @group | |
4908 | (push-mark (point)) | |
4909 | (push-mark (point-max)) | |
4910 | (goto-char (point-min)) | |
4911 | @end group | |
4912 | @end smallexample | |
4913 | ||
4914 | The first of these lines is the expression, @code{(push-mark (point))}. | |
4915 | ||
4916 | This line does exactly the same job as the first line of the body of | |
4917 | the @code{simplified-beginning-of-buffer} function, which is written | |
4918 | @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark | |
4919 | at the current position of the cursor. | |
4920 | ||
4921 | I don't know why the expression in @code{mark-whole-buffer} is written | |
4922 | @code{(push-mark (point))} and the expression in | |
4923 | @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps | |
4924 | whoever wrote the code did not know that the arguments for | |
4925 | @code{push-mark} are optional and that if @code{push-mark} is not | |
4926 | passed an argument, the function automatically sets mark at the | |
4927 | location of point by default. Or perhaps the expression was written | |
4928 | so as to parallel the structure of the next line. In any case, the | |
4929 | line causes Emacs to determine the position of point and set a mark | |
4930 | there. | |
4931 | ||
4932 | The next line of @code{mark-whole-buffer} is @code{(push-mark (point-max)}. | |
4933 | This expression sets a mark at the point in the buffer | |
4934 | that has the highest number. This will be the end of the buffer (or, | |
4935 | if the buffer is narrowed, the end of the accessible portion of the | |
4936 | buffer. @xref{Narrowing & Widening, , Narrowing and Widening}, for | |
4937 | more about narrowing.) After this mark has been set, the previous | |
4938 | mark, the one set at point, is no longer set, but Emacs remembers its | |
4939 | position, just as all other recent marks are always remembered. This | |
4940 | means that you can, if you wish, go back to that position by typing | |
4941 | @kbd{C-u C-@key{SPC}} twice. | |
4942 | ||
4943 | (In GNU Emacs 21, the @code{(push-mark (point-max)} is slightly more | |
4944 | complicated than shown here. The line reads | |
4945 | ||
4946 | @smallexample | |
4947 | (push-mark (point-max) nil t) | |
4948 | @end smallexample | |
4949 | ||
4950 | @noindent | |
4951 | (The expression works nearly the same as before. It sets a mark at | |
4952 | the highest numbered place in the buffer that it can. However, in | |
4953 | this version, @code{push-mark} has two additional arguments The second | |
4954 | argument to @code{push-mark} is @code{nil}. This tells the function | |
4955 | it should @emph{not} display a message that says `Mark set' when it | |
4956 | pushes the mark. The third argument is @code{t}. This tells | |
4957 | @code{push-mark} to activate the mark when Transient Mark mode is | |
4958 | turned on. Transient Mark mode highlights the currently active | |
4959 | region. It is usually turned off.) | |
4960 | ||
4961 | Finally, the last line of the function is @code{(goto-char | |
4962 | (point-min)))}. This is written exactly the same way as it is written | |
4963 | in @code{beginning-of-buffer}. The expression moves the cursor to | |
4964 | the minimum point in the buffer, that is, to the beginning of the buffer | |
4965 | (or to the beginning of the accessible portion of the buffer). As a | |
4966 | result of this, point is placed at the beginning of the buffer and mark | |
4967 | is set at the end of the buffer. The whole buffer is, therefore, the | |
4968 | region. | |
4969 | ||
4970 | @node append-to-buffer, Buffer Related Review, mark-whole-buffer, Buffer Walk Through | |
4971 | @comment node-name, next, previous, up | |
4972 | @section The Definition of @code{append-to-buffer} | |
4973 | @findex append-to-buffer | |
4974 | ||
4975 | The @code{append-to-buffer} command is very nearly as simple as the | |
4976 | @code{mark-whole-buffer} command. What it does is copy the region (that | |
4977 | is, the part of the buffer between point and mark) from the current | |
4978 | buffer to a specified buffer. | |
4979 | ||
4980 | @menu | |
4981 | * append-to-buffer overview:: | |
4982 | * append interactive:: A two part interactive expression. | |
4983 | * append-to-buffer body:: Incorporates a @code{let} expression. | |
4984 | * append save-excursion:: How the @code{save-excursion} works. | |
4985 | @end menu | |
4986 | ||
4987 | @node append-to-buffer overview, append interactive, append-to-buffer, append-to-buffer | |
4988 | @ifnottex | |
4989 | @unnumberedsubsec An Overview of @code{append-to-buffer} | |
4990 | @end ifnottex | |
4991 | ||
4992 | @findex insert-buffer-substring | |
4993 | The @code{append-to-buffer} command uses the | |
4994 | @code{insert-buffer-substring} function to copy the region. | |
4995 | @code{insert-buffer-substring} is described by its name: it takes a | |
4996 | string of characters from part of a buffer, a ``substring'', and | |
4997 | inserts them into another buffer. Most of @code{append-to-buffer} is | |
4998 | concerned with setting up the conditions for | |
4999 | @code{insert-buffer-substring} to work: the code must specify both the | |
5000 | buffer to which the text will go and the region that will be copied. | |
5001 | Here is the complete text of the function: | |
5002 | ||
5003 | @smallexample | |
5004 | @group | |
5005 | (defun append-to-buffer (buffer start end) | |
5006 | "Append to specified buffer the text of the region. | |
5007 | It is inserted into that buffer before its point. | |
5008 | @end group | |
5009 | ||
5010 | @group | |
5011 | When calling from a program, give three arguments: | |
5012 | a buffer or the name of one, and two character numbers | |
5013 | specifying the portion of the current buffer to be copied." | |
5014 | (interactive "BAppend to buffer:@: \nr") | |
5015 | (let ((oldbuf (current-buffer))) | |
5016 | (save-excursion | |
5017 | (set-buffer (get-buffer-create buffer)) | |
5018 | (insert-buffer-substring oldbuf start end)))) | |
5019 | @end group | |
5020 | @end smallexample | |
5021 | ||
5022 | The function can be understood by looking at it as a series of | |
5023 | filled-in templates. | |
5024 | ||
5025 | The outermost template is for the function definition. In this | |
5026 | function, it looks like this (with several slots filled in): | |
5027 | ||
5028 | @smallexample | |
5029 | @group | |
5030 | (defun append-to-buffer (buffer start end) | |
5031 | "@var{documentation}@dots{}" | |
5032 | (interactive "BAppend to buffer:@: \nr") | |
5033 | @var{body}@dots{}) | |
5034 | @end group | |
5035 | @end smallexample | |
5036 | ||
5037 | The first line of the function includes its name and three arguments. | |
5038 | The arguments are the @code{buffer} to which the text will be copied, and | |
5039 | the @code{start} and @code{end} of the region in the current buffer that | |
5040 | will be copied. | |
5041 | ||
5042 | The next part of the function is the documentation, which is clear and | |
5043 | complete. | |
5044 | ||
5045 | @node append interactive, append-to-buffer body, append-to-buffer overview, append-to-buffer | |
5046 | @comment node-name, next, previous, up | |
5047 | @subsection The @code{append-to-buffer} Interactive Expression | |
5048 | ||
5049 | Since the @code{append-to-buffer} function will be used interactively, | |
5050 | the function must have an @code{interactive} expression. (For a | |
5051 | review of @code{interactive}, see @ref{Interactive, , Making a | |
5052 | Function Interactive}.) The expression reads as follows: | |
5053 | ||
5054 | @smallexample | |
5055 | (interactive "BAppend to buffer:@: \nr") | |
5056 | @end smallexample | |
5057 | ||
5058 | @noindent | |
5059 | This expression has an argument inside of quotation marks and that | |
5060 | argument has two parts, separated by @samp{\n}. | |
5061 | ||
5062 | The first part is @samp{BAppend to buffer:@: }. Here, the @samp{B} | |
5063 | tells Emacs to ask for the name of the buffer that will be passed to the | |
5064 | function. Emacs will ask for the name by prompting the user in the | |
5065 | minibuffer, using the string following the @samp{B}, which is the string | |
5066 | @samp{Append to buffer:@: }. Emacs then binds the variable @code{buffer} | |
5067 | in the function's argument list to the specified buffer. | |
5068 | ||
5069 | The newline, @samp{\n}, separates the first part of the argument from | |
5070 | the second part. It is followed by an @samp{r} that tells Emacs to bind | |
5071 | the two arguments that follow the symbol @code{buffer} in the function's | |
5072 | argument list (that is, @code{start} and @code{end}) to the values of | |
5073 | point and mark. | |
5074 | ||
5075 | @node append-to-buffer body, append save-excursion, append interactive, append-to-buffer | |
5076 | @comment node-name, next, previous, up | |
5077 | @subsection The Body of @code{append-to-buffer} | |
5078 | ||
5079 | The body of the @code{append-to-buffer} function begins with @code{let}. | |
5080 | ||
5081 | As we have seen before (@pxref{let, , @code{let}}), the purpose of a | |
5082 | @code{let} expression is to create and give initial values to one or | |
5083 | more variables that will only be used within the body of the | |
5084 | @code{let}. This means that such a variable will not be confused with | |
5085 | any variable of the same name outside the @code{let} expression. | |
5086 | ||
5087 | We can see how the @code{let} expression fits into the function as a | |
5088 | whole by showing a template for @code{append-to-buffer} with the | |
5089 | @code{let} expression in outline: | |
5090 | ||
5091 | @smallexample | |
5092 | @group | |
5093 | (defun append-to-buffer (buffer start end) | |
5094 | "@var{documentation}@dots{}" | |
5095 | (interactive "BAppend to buffer:@: \nr") | |
5096 | (let ((@var{variable} @var{value})) | |
5097 | @var{body}@dots{}) | |
5098 | @end group | |
5099 | @end smallexample | |
5100 | ||
5101 | The @code{let} expression has three elements: | |
5102 | ||
5103 | @enumerate | |
5104 | @item | |
5105 | The symbol @code{let}; | |
5106 | ||
5107 | @item | |
5108 | A varlist containing, in this case, a single two-element list, | |
5109 | @code{(@var{variable} @var{value})}; | |
5110 | ||
5111 | @item | |
5112 | The body of the @code{let} expression. | |
5113 | @end enumerate | |
5114 | ||
5115 | @need 800 | |
5116 | In the @code{append-to-buffer} function, the varlist looks like this: | |
5117 | ||
5118 | @smallexample | |
5119 | (oldbuf (current-buffer)) | |
5120 | @end smallexample | |
5121 | ||
5122 | @noindent | |
5123 | In this part of the @code{let} expression, the one variable, | |
5124 | @code{oldbuf}, is bound to the value returned by the | |
5125 | @code{(current-buffer)} expression. The variable, @code{oldbuf}, is | |
5126 | used to keep track of the buffer in which you are working and from | |
5127 | which you will copy. | |
5128 | ||
5129 | The element or elements of a varlist are surrounded by a set of | |
5130 | parentheses so the Lisp interpreter can distinguish the varlist from | |
5131 | the body of the @code{let}. As a consequence, the two-element list | |
5132 | within the varlist is surrounded by a circumscribing set of parentheses. | |
5133 | The line looks like this: | |
5134 | ||
5135 | @smallexample | |
5136 | @group | |
5137 | (let ((oldbuf (current-buffer))) | |
5138 | @dots{} ) | |
5139 | @end group | |
5140 | @end smallexample | |
5141 | ||
5142 | @noindent | |
5143 | The two parentheses before @code{oldbuf} might surprise you if you did | |
5144 | not realize that the first parenthesis before @code{oldbuf} marks the | |
5145 | boundary of the varlist and the second parenthesis marks the beginning | |
5146 | of the two-element list, @code{(oldbuf (current-buffer))}. | |
5147 | ||
5148 | @node append save-excursion, , append-to-buffer body, append-to-buffer | |
5149 | @comment node-name, next, previous, up | |
5150 | @subsection @code{save-excursion} in @code{append-to-buffer} | |
5151 | ||
5152 | The body of the @code{let} expression in @code{append-to-buffer} | |
5153 | consists of a @code{save-excursion} expression. | |
5154 | ||
5155 | The @code{save-excursion} function saves the locations of point and | |
5156 | mark, and restores them to those positions after the expressions in the | |
5157 | body of the @code{save-excursion} complete execution. In addition, | |
5158 | @code{save-excursion} keeps track of the original buffer, and | |
5159 | restores it. This is how @code{save-excursion} is used in | |
5160 | @code{append-to-buffer}. | |
5161 | ||
5162 | @need 1500 | |
5163 | @cindex Indentation for formatting | |
5164 | @cindex Formatting convention | |
5165 | Incidentally, it is worth noting here that a Lisp function is normally | |
5166 | formatted so that everything that is enclosed in a multi-line spread is | |
5167 | indented more to the right than the first symbol. In this function | |
5168 | definition, the @code{let} is indented more than the @code{defun}, and | |
5169 | the @code{save-excursion} is indented more than the @code{let}, like | |
5170 | this: | |
5171 | ||
5172 | @smallexample | |
5173 | @group | |
5174 | (defun @dots{} | |
5175 | @dots{} | |
5176 | @dots{} | |
5177 | (let@dots{} | |
5178 | (save-excursion | |
5179 | @dots{} | |
5180 | @end group | |
5181 | @end smallexample | |
5182 | ||
5183 | @need 1500 | |
5184 | @noindent | |
5185 | This formatting convention makes it easy to see that the two lines in | |
5186 | the body of the @code{save-excursion} are enclosed by the parentheses | |
5187 | associated with @code{save-excursion}, just as the | |
5188 | @code{save-excursion} itself is enclosed by the parentheses associated | |
5189 | with the @code{let}: | |
5190 | ||
5191 | @smallexample | |
5192 | @group | |
5193 | (let ((oldbuf (current-buffer))) | |
5194 | (save-excursion | |
5195 | (set-buffer (get-buffer-create buffer)) | |
5196 | (insert-buffer-substring oldbuf start end)))) | |
5197 | @end group | |
5198 | @end smallexample | |
5199 | ||
5200 | @need 1200 | |
5201 | The use of the @code{save-excursion} function can be viewed as a process | |
5202 | of filling in the slots of a template: | |
5203 | ||
5204 | @smallexample | |
5205 | @group | |
5206 | (save-excursion | |
5207 | @var{first-expression-in-body} | |
5208 | @var{second-expression-in-body} | |
5209 | @dots{} | |
5210 | @var{last-expression-in-body}) | |
5211 | @end group | |
5212 | @end smallexample | |
5213 | ||
5214 | @need 1200 | |
5215 | @noindent | |
5216 | In this function, the body of the @code{save-excursion} contains only | |
5217 | two expressions. The body looks like this: | |
5218 | ||
5219 | @smallexample | |
5220 | @group | |
5221 | (set-buffer (get-buffer-create buffer)) | |
5222 | (insert-buffer-substring oldbuf start end) | |
5223 | @end group | |
5224 | @end smallexample | |
5225 | ||
5226 | When the @code{append-to-buffer} function is evaluated, the two | |
5227 | expressions in the body of the @code{save-excursion} are evaluated in | |
5228 | sequence. The value of the last expression is returned as the value of | |
5229 | the @code{save-excursion} function; the other expression is evaluated | |
5230 | only for its side effects. | |
5231 | ||
5232 | The first line in the body of the @code{save-excursion} uses the | |
5233 | @code{set-buffer} function to change the current buffer to the one | |
5234 | specified in the first argument to @code{append-to-buffer}. (Changing | |
5235 | the buffer is the side effect; as we have said before, in Lisp, a side | |
5236 | effect is often the primary thing we want.) The second line does the | |
5237 | primary work of the function. | |
5238 | ||
5239 | The @code{set-buffer} function changes Emacs' attention to the buffer to | |
5240 | which the text will be copied and from which @code{save-excursion} will | |
5241 | return. | |
5242 | ||
5243 | @need 800 | |
5244 | The line looks like this: | |
5245 | ||
5246 | @smallexample | |
5247 | (set-buffer (get-buffer-create buffer)) | |
5248 | @end smallexample | |
5249 | ||
5250 | The innermost expression of this list is @code{(get-buffer-create | |
5251 | buffer)}. This expression uses the @code{get-buffer-create} function, | |
5252 | which either gets the named buffer, or if it does not exist, creates one | |
5253 | with the given name. This means you can use @code{append-to-buffer} to | |
5254 | put text into a buffer that did not previously exist. | |
5255 | ||
5256 | @code{get-buffer-create} also keeps @code{set-buffer} from getting an | |
5257 | unnecessary error: @code{set-buffer} needs a buffer to go to; if you | |
5258 | were to specify a buffer that does not exist, Emacs would baulk. | |
5259 | Since @code{get-buffer-create} will create a buffer if none exists, | |
5260 | @code{set-buffer} is always provided with a buffer. | |
5261 | ||
5262 | @need 1250 | |
5263 | The last line of @code{append-to-buffer} does the work of appending | |
5264 | the text: | |
5265 | ||
5266 | @smallexample | |
5267 | (insert-buffer-substring oldbuf start end) | |
5268 | @end smallexample | |
5269 | ||
5270 | @noindent | |
5271 | The @code{insert-buffer-substring} function copies a string @emph{from} | |
5272 | the buffer specified as its first argument and inserts the string into | |
5273 | the present buffer. In this case, the argument to | |
5274 | @code{insert-buffer-substring} is the value of the variable created and | |
5275 | bound by the @code{let}, namely the value of @code{oldbuf}, which was | |
5276 | the current buffer when you gave the @code{append-to-buffer} command. | |
5277 | ||
5278 | After @code{insert-buffer-substring} has done its work, | |
5279 | @code{save-excursion} will restore the action to the original buffer and | |
5280 | @code{append-to-buffer} will have done its job. | |
5281 | ||
5282 | @need 800 | |
5283 | Written in skeletal form, the workings of the body look like this: | |
5284 | ||
5285 | @smallexample | |
5286 | @group | |
5287 | (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer}) | |
5288 | (save-excursion ; @r{Keep track of buffer.} | |
5289 | @var{change-buffer} | |
5290 | @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer}) | |
5291 | ||
5292 | @var{change-back-to-original-buffer-when-finished} | |
5293 | @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished} | |
5294 | ||
5295 | @end group | |
5296 | @end smallexample | |
5297 | ||
5298 | In summary, @code{append-to-buffer} works as follows: it saves the value | |
5299 | of the current buffer in the variable called @code{oldbuf}. It gets the | |
5300 | new buffer, creating one if need be, and switches Emacs to it. Using | |
5301 | the value of @code{oldbuf}, it inserts the region of text from the old | |
5302 | buffer into the new buffer; and then using @code{save-excursion}, it | |
5303 | brings you back to your original buffer. | |
5304 | ||
5305 | In looking at @code{append-to-buffer}, you have explored a fairly | |
5306 | complex function. It shows how to use @code{let} and | |
5307 | @code{save-excursion}, and how to change to and come back from another | |
5308 | buffer. Many function definitions use @code{let}, | |
5309 | @code{save-excursion}, and @code{set-buffer} this way. | |
5310 | ||
5311 | @node Buffer Related Review, Buffer Exercises, append-to-buffer, Buffer Walk Through | |
5312 | @comment node-name, next, previous, up | |
5313 | @section Review | |
5314 | ||
5315 | Here is a brief summary of the various functions discussed in this chapter. | |
5316 | ||
5317 | @table @code | |
5318 | @item describe-function | |
5319 | @itemx describe-variable | |
5320 | Print the documentation for a function or variable. | |
5321 | Conventionally bound to @kbd{C-h f} and @kbd{C-h v}. | |
5322 | ||
5323 | @item find-tag | |
5324 | Find the file containing the source for a function or variable and | |
5325 | switch buffers to it, positioning point at the beginning of the item. | |
5326 | Conventionally bound to @kbd{M-.} (that's a period following the | |
5327 | @key{META} key). | |
5328 | ||
5329 | @item save-excursion | |
5330 | Save the location of point and mark and restore their values after the | |
5331 | arguments to @code{save-excursion} have been evaluated. Also, remember | |
5332 | the current buffer and return to it. | |
5333 | ||
5334 | @item push-mark | |
5335 | Set mark at a location and record the value of the previous mark on the | |
5336 | mark ring. The mark is a location in the buffer that will keep its | |
5337 | relative position even if text is added to or removed from the buffer. | |
5338 | ||
5339 | @item goto-char | |
5340 | Set point to the location specified by the value of the argument, which | |
5341 | can be a number, a marker, or an expression that returns the number of | |
5342 | a position, such as @code{(point-min)}. | |
5343 | ||
5344 | @item insert-buffer-substring | |
5345 | Copy a region of text from a buffer that is passed to the function as | |
5346 | an argument and insert the region into the current buffer. | |
5347 | ||
5348 | @item mark-whole-buffer | |
5349 | Mark the whole buffer as a region. Normally bound to @kbd{C-x h}. | |
5350 | ||
5351 | @item set-buffer | |
5352 | Switch the attention of Emacs to another buffer, but do not change the | |
5353 | window being displayed. Used when the program rather than a human is | |
5354 | to work on a different buffer. | |
5355 | ||
5356 | @item get-buffer-create | |
5357 | @itemx get-buffer | |
5358 | Find a named buffer or create one if a buffer of that name does not | |
5359 | exist. The @code{get-buffer} function returns @code{nil} if the named | |
5360 | buffer does not exist. | |
5361 | @end table | |
5362 | ||
5363 | @need 1500 | |
5364 | @node Buffer Exercises, , Buffer Related Review, Buffer Walk Through | |
5365 | @section Exercises | |
5366 | ||
5367 | @itemize @bullet | |
5368 | @item | |
5369 | Write your own @code{simplified-end-of-buffer} function definition; | |
5370 | then test it to see whether it works. | |
5371 | ||
5372 | @item | |
5373 | Use @code{if} and @code{get-buffer} to write a function that prints a | |
5374 | message telling you whether a buffer exists. | |
5375 | ||
5376 | @item | |
5377 | Using @code{find-tag}, find the source for the @code{copy-to-buffer} | |
5378 | function. | |
5379 | @end itemize | |
5380 | ||
5381 | @node More Complex, Narrowing & Widening, Buffer Walk Through, Top | |
5382 | @comment node-name, next, previous, up | |
5383 | @chapter A Few More Complex Functions | |
5384 | ||
5385 | In this chapter, we build on what we have learned in previous chapters | |
5386 | by looking at more complex functions. The @code{copy-to-buffer} | |
5387 | function illustrates use of two @code{save-excursion} expressions in | |
5388 | one definition, while the @code{insert-buffer} function illustrates | |
5389 | use of an asterisk in an @code{interactive} expression, use of | |
5390 | @code{or}, and the important distinction between a name and the object | |
5391 | to which the name refers. | |
5392 | ||
5393 | @menu | |
5394 | * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}. | |
5395 | * insert-buffer:: Read-only, and with @code{or}. | |
5396 | * beginning-of-buffer:: Shows @code{goto-char}, | |
5397 | @code{point-min}, and @code{push-mark}. | |
5398 | * Second Buffer Related Review:: | |
5399 | * optional Exercise:: | |
5400 | @end menu | |
5401 | ||
5402 | @node copy-to-buffer, insert-buffer, More Complex, More Complex | |
5403 | @comment node-name, next, previous, up | |
5404 | @section The Definition of @code{copy-to-buffer} | |
5405 | @findex copy-to-buffer | |
5406 | ||
5407 | After understanding how @code{append-to-buffer} works, it is easy to | |
5408 | understand @code{copy-to-buffer}. This function copies text into a | |
5409 | buffer, but instead of adding to the second buffer, it replaces the | |
5410 | previous text in the second buffer. The code for the | |
5411 | @code{copy-to-buffer} function is almost the same as the code for | |
5412 | @code{append-to-buffer}, except that @code{erase-buffer} and a second | |
5413 | @code{save-excursion} are used. (@xref{append-to-buffer, , The | |
5414 | Definition of @code{append-to-buffer}}, for the description of | |
5415 | @code{append-to-buffer}.) | |
5416 | ||
5417 | @need 800 | |
5418 | The body of @code{copy-to-buffer} looks like this | |
5419 | ||
5420 | @smallexample | |
5421 | @group | |
5422 | @dots{} | |
5423 | (interactive "BCopy to buffer:@: \nr") | |
5424 | (let ((oldbuf (current-buffer))) | |
5425 | (save-excursion | |
5426 | (set-buffer (get-buffer-create buffer)) | |
5427 | (erase-buffer) | |
5428 | (save-excursion | |
5429 | (insert-buffer-substring oldbuf start end))))) | |
5430 | @end group | |
5431 | @end smallexample | |
5432 | ||
5433 | This code is similar to the code in @code{append-to-buffer}: it is | |
5434 | only after changing to the buffer to which the text will be copied | |
5435 | that the definition for this function diverges from the definition for | |
5436 | @code{append-to-buffer}: the @code{copy-to-buffer} function erases the | |
5437 | buffer's former contents. (This is what is meant by `replacement'; to | |
5438 | replace text, Emacs erases the previous text and then inserts new | |
5439 | text.) After erasing the previous contents of the buffer, | |
5440 | @code{save-excursion} is used for a second time and the new text is | |
5441 | inserted. | |
5442 | ||
5443 | Why is @code{save-excursion} used twice? Consider again what the | |
5444 | function does. | |
5445 | ||
5446 | @need 1250 | |
5447 | In outline, the body of @code{copy-to-buffer} looks like this: | |
5448 | ||
5449 | @smallexample | |
5450 | @group | |
5451 | (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer}) | |
5452 | (save-excursion ; @r{First use of @code{save-excursion}.} | |
5453 | @var{change-buffer} | |
5454 | (erase-buffer) | |
5455 | (save-excursion ; @r{Second use of @code{save-excursion}.} | |
5456 | @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer}))) | |
5457 | @end group | |
5458 | @end smallexample | |
5459 | ||
5460 | The first use of @code{save-excursion} returns Emacs to the buffer from | |
5461 | which the text is being copied. That is clear, and is just like its use | |
5462 | in @code{append-to-buffer}. Why the second use? The reason is that | |
5463 | @code{insert-buffer-substring} always leaves point at the @emph{end} of | |
5464 | the region being inserted. The second @code{save-excursion} causes | |
5465 | Emacs to leave point at the beginning of the text being inserted. In | |
5466 | most circumstances, users prefer to find point at the beginning of | |
5467 | inserted text. (Of course, the @code{copy-to-buffer} function returns | |
5468 | the user to the original buffer when done---but if the user @emph{then} | |
5469 | switches to the copied-to buffer, point will go to the beginning of the | |
5470 | text. Thus, this use of a second @code{save-excursion} is a little | |
5471 | nicety.) | |
5472 | ||
5473 | @node insert-buffer, beginning-of-buffer, copy-to-buffer, More Complex | |
5474 | @comment node-name, next, previous, up | |
5475 | @section The Definition of @code{insert-buffer} | |
5476 | @findex insert-buffer | |
5477 | ||
5478 | @code{insert-buffer} is yet another buffer-related function. This | |
5479 | command copies another buffer @emph{into} the current buffer. It is the | |
5480 | reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they | |
5481 | copy a region of text @emph{from} the current buffer to another buffer. | |
5482 | ||
5483 | In addition, this code illustrates the use of @code{interactive} with a | |
5484 | buffer that might be @dfn{read-only} and the important distinction | |
5485 | between the name of an object and the object actually referred to. | |
5486 | ||
5487 | @menu | |
5488 | * insert-buffer code:: | |
5489 | * insert-buffer interactive:: When you can read, but not write. | |
5490 | * insert-buffer body:: The body has an @code{or} and a @code{let}. | |
5491 | * if & or:: Using an @code{if} instead of an @code{or}. | |
5492 | * Insert or:: How the @code{or} expression works. | |
5493 | * Insert let:: Two @code{save-excursion} expressions. | |
5494 | @end menu | |
5495 | ||
5496 | @node insert-buffer code, insert-buffer interactive, insert-buffer, insert-buffer | |
5497 | @ifnottex | |
5498 | @unnumberedsubsec The Code for @code{insert-buffer} | |
5499 | @end ifnottex | |
5500 | ||
5501 | @need 800 | |
5502 | Here is the code: | |
5503 | ||
5504 | @smallexample | |
5505 | @group | |
5506 | (defun insert-buffer (buffer) | |
5507 | "Insert after point the contents of BUFFER. | |
5508 | Puts mark after the inserted text. | |
5509 | BUFFER may be a buffer or a buffer name." | |
5510 | (interactive "*bInsert buffer:@: ") | |
5511 | @end group | |
5512 | @group | |
5513 | (or (bufferp buffer) | |
5514 | (setq buffer (get-buffer buffer))) | |
5515 | (let (start end newmark) | |
5516 | (save-excursion | |
5517 | (save-excursion | |
5518 | (set-buffer buffer) | |
5519 | (setq start (point-min) end (point-max))) | |
5520 | @end group | |
5521 | @group | |
5522 | (insert-buffer-substring buffer start end) | |
5523 | (setq newmark (point))) | |
5524 | (push-mark newmark))) | |
5525 | @end group | |
5526 | @end smallexample | |
5527 | ||
5528 | @need 1200 | |
5529 | As with other function definitions, you can use a template to see an | |
5530 | outline of the function: | |
5531 | ||
5532 | @smallexample | |
5533 | @group | |
5534 | (defun insert-buffer (buffer) | |
5535 | "@var{documentation}@dots{}" | |
5536 | (interactive "*bInsert buffer:@: ") | |
5537 | @var{body}@dots{}) | |
5538 | @end group | |
5539 | @end smallexample | |
5540 | ||
5541 | @node insert-buffer interactive, insert-buffer body, insert-buffer code, insert-buffer | |
5542 | @comment node-name, next, previous, up | |
5543 | @subsection The Interactive Expression in @code{insert-buffer} | |
5544 | @findex interactive, @r{example use of} | |
5545 | ||
5546 | In @code{insert-buffer}, the argument to the @code{interactive} | |
5547 | declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert | |
5548 | buffer:@: }. | |
5549 | ||
5550 | @menu | |
5551 | * Read-only buffer:: When a buffer cannot be modified. | |
5552 | * b for interactive:: An existing buffer or else its name. | |
5553 | @end menu | |
5554 | ||
5555 | @node Read-only buffer, b for interactive, insert-buffer interactive, insert-buffer interactive | |
5556 | @comment node-name, next, previous, up | |
5557 | @unnumberedsubsubsec A Read-only Buffer | |
5558 | @cindex Read-only buffer | |
5559 | @cindex Asterisk for read-only buffer | |
5560 | @findex * @r{for read-only buffer} | |
5561 | ||
5562 | The asterisk is for the situation when the buffer is a read-only | |
5563 | buffer---a buffer that cannot be modified. If @code{insert-buffer} is | |
5564 | called on a buffer that is read-only, a message to this effect is | |
5565 | printed in the echo area and the terminal may beep or blink at you; | |
5566 | you will not be permitted to insert anything into current buffer. The | |
5567 | asterisk does not need to be followed by a newline to separate it from | |
5568 | the next argument. | |
5569 | ||
5570 | @node b for interactive, , Read-only buffer, insert-buffer interactive | |
5571 | @comment node-name, next, previous, up | |
5572 | @unnumberedsubsubsec @samp{b} in an Interactive Expression | |
5573 | ||
5574 | The next argument in the interactive expression starts with a lower | |
5575 | case @samp{b}. (This is different from the code for | |
5576 | @code{append-to-buffer}, which uses an upper-case @samp{B}. | |
5577 | @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.) | |
5578 | The lower-case @samp{b} tells the Lisp interpreter that the argument | |
5579 | for @code{insert-buffer} should be an existing buffer or else its | |
5580 | name. (The upper-case @samp{B} option provides for the possibility | |
5581 | that the buffer does not exist.) Emacs will prompt you for the name | |
5582 | of the buffer, offering you a default buffer, with name completion | |
5583 | enabled. If the buffer does not exist, you receive a message that | |
5584 | says ``No match''; your terminal may beep at you as well. | |
5585 | ||
5586 | @node insert-buffer body, if & or, insert-buffer interactive, insert-buffer | |
5587 | @comment node-name, next, previous, up | |
5588 | @subsection The Body of the @code{insert-buffer} Function | |
5589 | ||
5590 | The body of the @code{insert-buffer} function has two major parts: an | |
5591 | @code{or} expression and a @code{let} expression. The purpose of the | |
5592 | @code{or} expression is to ensure that the argument @code{buffer} is | |
5593 | bound to a buffer and not just the name of a buffer. The body of the | |
5594 | @code{let} expression contains the code which copies the other buffer | |
5595 | into the current buffer. | |
5596 | ||
5597 | @need 1250 | |
5598 | In outline, the two expressions fit into the @code{insert-buffer} | |
5599 | function like this: | |
5600 | ||
5601 | @smallexample | |
5602 | @group | |
5603 | (defun insert-buffer (buffer) | |
5604 | "@var{documentation}@dots{}" | |
5605 | (interactive "*bInsert buffer:@: ") | |
5606 | (or @dots{} | |
5607 | @dots{} | |
5608 | @end group | |
5609 | @group | |
5610 | (let (@var{varlist}) | |
5611 | @var{body-of-}@code{let}@dots{} ) | |
5612 | @end group | |
5613 | @end smallexample | |
5614 | ||
5615 | To understand how the @code{or} expression ensures that the argument | |
5616 | @code{buffer} is bound to a buffer and not to the name of a buffer, it | |
5617 | is first necessary to understand the @code{or} function. | |
5618 | ||
5619 | Before doing this, let me rewrite this part of the function using | |
5620 | @code{if} so that you can see what is done in a manner that will be familiar. | |
5621 | ||
5622 | @node if & or, Insert or, insert-buffer body, insert-buffer | |
5623 | @comment node-name, next, previous, up | |
5624 | @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or} | |
5625 | ||
5626 | The job to be done is to make sure the value of @code{buffer} is a | |
5627 | buffer itself and not the name of a buffer. If the value is the name, | |
5628 | then the buffer itself must be got. | |
5629 | ||
5630 | You can imagine yourself at a conference where an usher is wandering | |
5631 | around holding a list with your name on it and looking for you: the | |
5632 | usher is ``bound'' to your name, not to you; but when the usher finds | |
5633 | you and takes your arm, the usher becomes ``bound'' to you. | |
5634 | ||
5635 | @need 800 | |
5636 | In Lisp, you might describe this situation like this: | |
5637 | ||
5638 | @smallexample | |
5639 | @group | |
5640 | (if (not (holding-on-to-guest)) | |
5641 | (find-and-take-arm-of-guest)) | |
5642 | @end group | |
5643 | @end smallexample | |
5644 | ||
5645 | We want to do the same thing with a buffer---if we do not have the | |
5646 | buffer itself, we want to get it. | |
5647 | ||
5648 | @need 1200 | |
5649 | Using a predicate called @code{bufferp} that tells us whether we have a | |
5650 | buffer (rather than its name), we can write the code like this: | |
5651 | ||
5652 | @smallexample | |
5653 | @group | |
5654 | (if (not (bufferp buffer)) ; @r{if-part} | |
5655 | (setq buffer (get-buffer buffer))) ; @r{then-part} | |
5656 | @end group | |
5657 | @end smallexample | |
5658 | ||
5659 | @noindent | |
5660 | Here, the true-or-false-test of the @code{if} expression is | |
5661 | @w{@code{(not (bufferp buffer))}}; and the then-part is the expression | |
5662 | @w{@code{(setq buffer (get-buffer buffer))}}. | |
5663 | ||
5664 | In the test, the function @code{bufferp} returns true if its argument is | |
5665 | a buffer---but false if its argument is the name of the buffer. (The | |
5666 | last character of the function name @code{bufferp} is the character | |
5667 | @samp{p}; as we saw earlier, such use of @samp{p} is a convention that | |
5668 | indicates that the function is a predicate, which is a term that means | |
5669 | that the function will determine whether some property is true or false. | |
5670 | @xref{Wrong Type of Argument, , Using the Wrong Type Object as an | |
5671 | Argument}.) | |
5672 | ||
5673 | @need 1200 | |
5674 | The function @code{not} precedes the expression @code{(bufferp buffer)}, | |
5675 | so the true-or-false-test looks like this: | |
5676 | ||
5677 | @smallexample | |
5678 | (not (bufferp buffer)) | |
5679 | @end smallexample | |
5680 | ||
5681 | @noindent | |
5682 | @code{not} is a function that returns true if its argument is false | |
5683 | and false if its argument is true. So if @code{(bufferp buffer)} | |
5684 | returns true, the @code{not} expression returns false and vice-versa: | |
5685 | what is ``not true'' is false and what is ``not false'' is true. | |
5686 | ||
5687 | Using this test, the @code{if} expression works as follows: when the | |
5688 | value of the variable @code{buffer} is actually a buffer rather then | |
5689 | its name, the true-or-false-test returns false and the @code{if} | |
5690 | expression does not evaluate the then-part. This is fine, since we do | |
5691 | not need to do anything to the variable @code{buffer} if it really is | |
5692 | a buffer. | |
5693 | ||
5694 | On the other hand, when the value of @code{buffer} is not a buffer | |
5695 | itself, but the name of a buffer, the true-or-false-test returns true | |
5696 | and the then-part of the expression is evaluated. In this case, the | |
5697 | then-part is @code{(setq buffer (get-buffer buffer))}. This | |
5698 | expression uses the @code{get-buffer} function to return an actual | |
5699 | buffer itself, given its name. The @code{setq} then sets the variable | |
5700 | @code{buffer} to the value of the buffer itself, replacing its previous | |
5701 | value (which was the name of the buffer). | |
5702 | ||
5703 | @node Insert or, Insert let, if & or, insert-buffer | |
5704 | @comment node-name, next, previous, up | |
5705 | @subsection The @code{or} in the Body | |
5706 | ||
5707 | The purpose of the @code{or} expression in the @code{insert-buffer} | |
5708 | function is to ensure that the argument @code{buffer} is bound to a | |
5709 | buffer and not just to the name of a buffer. The previous section shows | |
5710 | how the job could have been done using an @code{if} expression. | |
5711 | However, the @code{insert-buffer} function actually uses @code{or}. | |
5712 | To understand this, it is necessary to understand how @code{or} works. | |
5713 | ||
5714 | @findex or | |
5715 | An @code{or} function can have any number of arguments. It evaluates | |
5716 | each argument in turn and returns the value of the first of its | |
5717 | arguments that is not @code{nil}. Also, and this is a crucial feature | |
5718 | of @code{or}, it does not evaluate any subsequent arguments after | |
5719 | returning the first non-@code{nil} value. | |
5720 | ||
5721 | @need 800 | |
5722 | The @code{or} expression looks like this: | |
5723 | ||
5724 | @smallexample | |
5725 | @group | |
5726 | (or (bufferp buffer) | |
5727 | (setq buffer (get-buffer buffer))) | |
5728 | @end group | |
5729 | @end smallexample | |
5730 | ||
5731 | @noindent | |
5732 | The first argument to @code{or} is the expression @code{(bufferp buffer)}. | |
5733 | This expression returns true (a non-@code{nil} value) if the buffer is | |
5734 | actually a buffer, and not just the name of a buffer. In the @code{or} | |
5735 | expression, if this is the case, the @code{or} expression returns this | |
5736 | true value and does not evaluate the next expression---and this is fine | |
5737 | with us, since we do not want to do anything to the value of | |
5738 | @code{buffer} if it really is a buffer. | |
5739 | ||
5740 | On the other hand, if the value of @code{(bufferp buffer)} is @code{nil}, | |
5741 | which it will be if the value of @code{buffer} is the name of a buffer, | |
5742 | the Lisp interpreter evaluates the next element of the @code{or} | |
5743 | expression. This is the expression @code{(setq buffer (get-buffer | |
5744 | buffer))}. This expression returns a non-@code{nil} value, which | |
5745 | is the value to which it sets the variable @code{buffer}---and this | |
5746 | value is a buffer itself, not the name of a buffer. | |
5747 | ||
5748 | The result of all this is that the symbol @code{buffer} is always | |
5749 | bound to a buffer itself rather than to the name of a buffer. All | |
5750 | this is necessary because the @code{set-buffer} function in a | |
5751 | following line only works with a buffer itself, not with the name to a | |
5752 | buffer. | |
5753 | ||
5754 | @need 1250 | |
5755 | Incidentally, using @code{or}, the situation with the usher would be | |
5756 | written like this: | |
5757 | ||
5758 | @smallexample | |
5759 | (or (holding-on-to-guest) (find-and-take-arm-of-guest)) | |
5760 | @end smallexample | |
5761 | ||
5762 | @node Insert let, , Insert or, insert-buffer | |
5763 | @comment node-name, next, previous, up | |
5764 | @subsection The @code{let} Expression in @code{insert-buffer} | |
5765 | ||
5766 | After ensuring that the variable @code{buffer} refers to a buffer itself | |
5767 | and not just to the name of a buffer, the @code{insert-buffer function} | |
5768 | continues with a @code{let} expression. This specifies three local | |
5769 | variables, @code{start}, @code{end}, and @code{newmark} and binds them | |
5770 | to the initial value @code{nil}. These variables are used inside the | |
5771 | remainder of the @code{let} and temporarily hide any other occurrence of | |
5772 | variables of the same name in Emacs until the end of the @code{let}. | |
5773 | ||
5774 | @need 1200 | |
5775 | The body of the @code{let} contains two @code{save-excursion} | |
5776 | expressions. First, we will look at the inner @code{save-excursion} | |
5777 | expression in detail. The expression looks like this: | |
5778 | ||
5779 | @smallexample | |
5780 | @group | |
5781 | (save-excursion | |
5782 | (set-buffer buffer) | |
5783 | (setq start (point-min) end (point-max))) | |
5784 | @end group | |
5785 | @end smallexample | |
5786 | ||
5787 | @noindent | |
5788 | The expression @code{(set-buffer buffer)} changes Emacs' attention | |
5789 | from the current buffer to the one from which the text will copied. | |
5790 | In that buffer, the variables @code{start} and @code{end} are set to | |
5791 | the beginning and end of the buffer, using the commands | |
5792 | @code{point-min} and @code{point-max}. Note that we have here an | |
5793 | illustration of how @code{setq} is able to set two variables in the | |
5794 | same expression. The first argument of @code{setq} is set to the | |
5795 | value of its second, and its third argument is set to the value of its | |
5796 | fourth. | |
5797 | ||
5798 | After the body of the inner @code{save-excursion} is evaluated, the | |
5799 | @code{save-excursion} restores the original buffer, but @code{start} and | |
5800 | @code{end} remain set to the values of the beginning and end of the | |
5801 | buffer from which the text will be copied. | |
5802 | ||
5803 | @need 1250 | |
5804 | The outer @code{save-excursion} expression looks like this: | |
5805 | ||
5806 | @smallexample | |
5807 | @group | |
5808 | (save-excursion | |
5809 | (@var{inner-}@code{save-excursion}@var{-expression} | |
5810 | (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end}) | |
5811 | (insert-buffer-substring buffer start end) | |
5812 | (setq newmark (point))) | |
5813 | @end group | |
5814 | @end smallexample | |
5815 | ||
5816 | @noindent | |
5817 | The @code{insert-buffer-substring} function copies the text | |
5818 | @emph{into} the current buffer @emph{from} the region indicated by | |
5819 | @code{start} and @code{end} in @code{buffer}. Since the whole of the | |
5820 | second buffer lies between @code{start} and @code{end}, the whole of | |
5821 | the second buffer is copied into the buffer you are editing. Next, | |
5822 | the value of point, which will be at the end of the inserted text, is | |
5823 | recorded in the variable @code{newmark}. | |
5824 | ||
5825 | After the body of the outer @code{save-excursion} is evaluated, point | |
5826 | and mark are relocated to their original places. | |
5827 | ||
5828 | However, it is convenient to locate a mark at the end of the newly | |
5829 | inserted text and locate point at its beginning. The @code{newmark} | |
5830 | variable records the end of the inserted text. In the last line of | |
5831 | the @code{let} expression, the @code{(push-mark newmark)} expression | |
5832 | function sets a mark to this location. (The previous location of the | |
5833 | mark is still accessible; it is recorded on the mark ring and you can | |
5834 | go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is | |
5835 | located at the beginning of the inserted text, which is where it was | |
5836 | before you called the insert function. | |
5837 | ||
5838 | @need 1250 | |
5839 | The whole @code{let} expression looks like this: | |
5840 | ||
5841 | @smallexample | |
5842 | @group | |
5843 | (let (start end newmark) | |
5844 | (save-excursion | |
5845 | (save-excursion | |
5846 | (set-buffer buffer) | |
5847 | (setq start (point-min) end (point-max))) | |
5848 | (insert-buffer-substring buffer start end) | |
5849 | (setq newmark (point))) | |
5850 | (push-mark newmark)) | |
5851 | @end group | |
5852 | @end smallexample | |
5853 | ||
5854 | Like the @code{append-to-buffer} function, the @code{insert-buffer} | |
5855 | function uses @code{let}, @code{save-excursion}, and | |
5856 | @code{set-buffer}. In addition, the function illustrates one way to | |
5857 | use @code{or}. All these functions are building blocks that we will | |
5858 | find and use again and again. | |
5859 | ||
5860 | @node beginning-of-buffer, Second Buffer Related Review, insert-buffer, More Complex | |
5861 | @comment node-name, next, previous, up | |
5862 | @section Complete Definition of @code{beginning-of-buffer} | |
5863 | @findex beginning-of-buffer | |
5864 | ||
5865 | The basic structure of the @code{beginning-of-buffer} function has | |
5866 | already been discussed. (@xref{simplified-beginning-of-buffer, , A | |
5867 | Simplified @code{beginning-of-buffer} Definition}.) | |
5868 | This section describes the complex part of the definition. | |
5869 | ||
5870 | As previously described, when invoked without an argument, | |
5871 | @code{beginning-of-buffer} moves the cursor to the beginning of the | |
5872 | buffer, leaving the mark at the previous position. However, when the | |
5873 | command is invoked with a number between one and ten, the function | |
5874 | considers that number to be a fraction of the length of the buffer, | |
5875 | measured in tenths, and Emacs moves the cursor that fraction of the way | |
5876 | from the beginning of the buffer. Thus, you can either call this | |
5877 | function with the key command @kbd{M-<}, which will move the cursor to | |
5878 | the beginning of the buffer, or with a key command such as @kbd{C-u 7 | |
5879 | M-<} which will move the cursor to a point 70% of the way through the | |
5880 | buffer. If a number bigger than ten is used for the argument, it moves | |
5881 | to the end of the buffer. | |
5882 | ||
5883 | The @code{beginning-of-buffer} function can be called with or without an | |
5884 | argument. The use of the argument is optional. | |
5885 | ||
5886 | @menu | |
5887 | * Optional Arguments:: | |
5888 | * beginning-of-buffer opt arg:: Example with optional argument. | |
5889 | * beginning-of-buffer complete:: | |
5890 | @end menu | |
5891 | ||
5892 | @node Optional Arguments, beginning-of-buffer opt arg, beginning-of-buffer, beginning-of-buffer | |
5893 | @subsection Optional Arguments | |
5894 | ||
5895 | Unless told otherwise, Lisp expects that a function with an argument in | |
5896 | its function definition will be called with a value for that argument. | |
5897 | If that does not happen, you get an error and a message that says | |
5898 | @samp{Wrong number of arguments}. | |
5899 | ||
5900 | @cindex Optional arguments | |
5901 | @cindex Keyword | |
5902 | @findex optional | |
5903 | However, optional arguments are a feature of Lisp: a @dfn{keyword} may | |
5904 | be used to tell the Lisp interpreter that an argument is optional. | |
5905 | The keyword is @code{&optional}. (The @samp{&} in front of | |
5906 | @samp{optional} is part of the keyword.) In a function definition, if | |
5907 | an argument follows the keyword @code{&optional}, a value does not | |
5908 | need to be passed to that argument when the function is called. | |
5909 | ||
5910 | @need 1200 | |
5911 | The first line of the function definition of @code{beginning-of-buffer} | |
5912 | therefore looks like this: | |
5913 | ||
5914 | @smallexample | |
5915 | (defun beginning-of-buffer (&optional arg) | |
5916 | @end smallexample | |
5917 | ||
5918 | @need 1250 | |
5919 | In outline, the whole function looks like this: | |
5920 | ||
5921 | @smallexample | |
5922 | @group | |
5923 | (defun beginning-of-buffer (&optional arg) | |
5924 | "@var{documentation}@dots{}" | |
5925 | (interactive "P") | |
5926 | (push-mark) | |
5927 | (goto-char | |
5928 | (@var{if-there-is-an-argument} | |
5929 | @var{figure-out-where-to-go} | |
5930 | @var{else-go-to} | |
5931 | (point-min)))) | |
5932 | @end group | |
5933 | @end smallexample | |
5934 | ||
5935 | The function is similar to the @code{simplified-beginning-of-buffer} | |
5936 | function except that the @code{interactive} expression has @code{"P"} | |
5937 | as an argument and the @code{goto-char} function is followed by an | |
5938 | if-then-else expression that figures out where to put the cursor if | |
5939 | there is an argument. | |
5940 | ||
5941 | The @code{"P"} in the @code{interactive} expression tells Emacs to pass | |
5942 | a prefix argument, if there is one, to the function. A prefix argument | |
5943 | is made by typing the @key{META} key followed by a number, or by typing | |
5944 | @kbd{C-u} and then a number (if you don't type a number, @kbd{C-u} | |
5945 | defaults to 4). | |
5946 | ||
5947 | The true-or-false-test of the @code{if} expression is simple: it is | |
5948 | simply the argument @code{arg}. If @code{arg} has a value that is not | |
5949 | @code{nil}, which will be the case if @code{beginning-of-buffer} is | |
5950 | called with an argument, then this true-or-false-test will return true | |
5951 | and the then-part of the @code{if} expression will be evaluated. On the | |
5952 | other hand, if @code{beginning-of-buffer} is not called with an | |
5953 | argument, the value of @code{arg} will be @code{nil} and the else-part | |
5954 | of the @code{if} expression will be evaluated. The else-part is simply | |
5955 | @code{point-min}, and when this is the outcome, the whole | |
5956 | @code{goto-char} expression is @code{(goto-char (point-min))}, which is | |
5957 | how we saw the @code{beginning-of-buffer} function in its simplified | |
5958 | form. | |
5959 | ||
5960 | @node beginning-of-buffer opt arg, beginning-of-buffer complete, Optional Arguments, beginning-of-buffer | |
5961 | @subsection @code{beginning-of-buffer} with an Argument | |
5962 | ||
5963 | When @code{beginning-of-buffer} is called with an argument, an | |
5964 | expression is evaluated which calculates what value to pass to | |
5965 | @code{goto-char}. This expression is rather complicated at first sight. | |
5966 | It includes an inner @code{if} expression and much arithmetic. It looks | |
5967 | like this: | |
5968 | ||
5969 | @smallexample | |
5970 | @group | |
5971 | (if (> (buffer-size) 10000) | |
5972 | ;; @r{Avoid overflow for large buffer sizes!} | |
5973 | (* (prefix-numeric-value arg) (/ (buffer-size) 10)) | |
5974 | (/ | |
5975 | (+ 10 | |
5976 | (* | |
5977 | (buffer-size) (prefix-numeric-value arg))) 10)) | |
5978 | @end group | |
5979 | @end smallexample | |
5980 | ||
5981 | @menu | |
5982 | * Disentangle beginning-of-buffer:: | |
5983 | * Large buffer case:: | |
5984 | * Small buffer case:: | |
5985 | @end menu | |
5986 | ||
5987 | @node Disentangle beginning-of-buffer, Large buffer case, beginning-of-buffer opt arg, beginning-of-buffer opt arg | |
5988 | @ifnottex | |
5989 | @unnumberedsubsubsec Disentangle @code{beginning-of-buffer} | |
5990 | @end ifnottex | |
5991 | ||
5992 | Like other complex-looking expressions, the conditional expression | |
5993 | within @code{beginning-of-buffer} can be disentangled by looking at it | |
5994 | as parts of a template, in this case, the template for an if-then-else | |
5995 | expression. In skeletal form, the expression looks like this: | |
5996 | ||
5997 | @smallexample | |
5998 | @group | |
5999 | (if (@var{buffer-is-large} | |
6000 | @var{divide-buffer-size-by-10-and-multiply-by-arg} | |
6001 | @var{else-use-alternate-calculation} | |
6002 | @end group | |
6003 | @end smallexample | |
6004 | ||
6005 | The true-or-false-test of this inner @code{if} expression checks the | |
6006 | size of the buffer. The reason for this is that the old Version 18 | |
6007 | Emacs used numbers that are no bigger than eight million or so | |
6008 | and in the computation that followed, the programmer feared that Emacs | |
6009 | might try to use over-large numbers if the buffer were large. The | |
6010 | term `overflow', mentioned in the comment, means numbers that are over | |
6011 | large. Version 21 Emacs uses larger numbers, but this code has not | |
6012 | been touched, if only because people now look at buffers that are far, | |
6013 | far larger than ever before. | |
6014 | ||
6015 | There are two cases: if the buffer is large and if it is not. | |
6016 | ||
6017 | @node Large buffer case, Small buffer case, Disentangle beginning-of-buffer, beginning-of-buffer opt arg | |
6018 | @comment node-name, next, previous, up | |
6019 | @unnumberedsubsubsec What happens in a large buffer | |
6020 | ||
6021 | In @code{beginning-of-buffer}, the inner @code{if} expression tests | |
6022 | whether the size of the buffer is greater than 10,000 characters. To do | |
6023 | this, it uses the @code{>} function and the @code{buffer-size} function. | |
6024 | ||
6025 | @need 800 | |
6026 | The line looks like this: | |
6027 | ||
6028 | @smallexample | |
6029 | (if (> (buffer-size) 10000) | |
6030 | @end smallexample | |
6031 | ||
6032 | @need 1200 | |
6033 | @noindent | |
6034 | When the buffer is large, the then-part of the @code{if} expression is | |
6035 | evaluated. It reads like this (after formatting for easy reading): | |
6036 | ||
6037 | @smallexample | |
6038 | @group | |
6039 | (* | |
6040 | (prefix-numeric-value arg) | |
6041 | (/ (buffer-size) 10)) | |
6042 | @end group | |
6043 | @end smallexample | |
6044 | ||
6045 | @noindent | |
6046 | This expression is a multiplication, with two arguments to the function | |
6047 | @code{*}. | |
6048 | ||
6049 | The first argument is @code{(prefix-numeric-value arg)}. When | |
6050 | @code{"P"} is used as the argument for @code{interactive}, the value | |
6051 | passed to the function as its argument is passed a ``raw prefix | |
6052 | argument'', and not a number. (It is a number in a list.) To perform | |
6053 | the arithmetic, a conversion is necessary, and | |
6054 | @code{prefix-numeric-value} does the job. | |
6055 | ||
6056 | @findex / @r{(division)} | |
6057 | @cindex Division | |
6058 | The second argument is @code{(/ (buffer-size) 10)}. This expression | |
6059 | divides the numeric value of the buffer by ten. This produces a number | |
6060 | that tells how many characters make up one tenth of the buffer size. | |
6061 | (In Lisp, @code{/} is used for division, just as @code{*} is | |
6062 | used for multiplication.) | |
6063 | ||
6064 | @need 1200 | |
6065 | In the multiplication expression as a whole, this amount is multiplied | |
6066 | by the value of the prefix argument---the multiplication looks like this: | |
6067 | ||
6068 | @smallexample | |
6069 | @group | |
6070 | (* @var{numeric-value-of-prefix-arg} | |
6071 | @var{number-of-characters-in-one-tenth-of-the-buffer}) | |
6072 | @end group | |
6073 | @end smallexample | |
6074 | ||
6075 | @noindent | |
6076 | If, for example, the prefix argument is @samp{7}, the one-tenth value | |
6077 | will be multiplied by 7 to give a position 70% of the way through the | |
6078 | buffer. | |
6079 | ||
6080 | @need 1200 | |
6081 | The result of all this is that if the buffer is large, the | |
6082 | @code{goto-char} expression reads like this: | |
6083 | ||
6084 | @smallexample | |
6085 | @group | |
6086 | (goto-char (* (prefix-numeric-value arg) | |
6087 | (/ (buffer-size) 10))) | |
6088 | @end group | |
6089 | @end smallexample | |
6090 | ||
6091 | This puts the cursor where we want it. | |
6092 | ||
6093 | @node Small buffer case, , Large buffer case, beginning-of-buffer opt arg | |
6094 | @comment node-name, next, previous, up | |
6095 | @unnumberedsubsubsec What happens in a small buffer | |
6096 | ||
6097 | If the buffer contains fewer than 10,000 characters, a slightly | |
6098 | different computation is performed. You might think this is not | |
6099 | necessary, since the first computation could do the job. However, in | |
6100 | a small buffer, the first method may not put the cursor on exactly the | |
6101 | desired line; the second method does a better job. | |
6102 | ||
6103 | @need 800 | |
6104 | The code looks like this: | |
6105 | ||
6106 | @c Keep this on one line. | |
6107 | @smallexample | |
6108 | (/ (+ 10 (* (buffer-size) (prefix-numeric-value arg))) 10)) | |
6109 | @end smallexample | |
6110 | ||
6111 | @need 1200 | |
6112 | @noindent | |
6113 | This is code in which you figure out what happens by discovering how the | |
6114 | functions are embedded in parentheses. It is easier to read if you | |
6115 | reformat it with each expression indented more deeply than its | |
6116 | enclosing expression: | |
6117 | ||
6118 | @smallexample | |
6119 | @group | |
6120 | (/ | |
6121 | (+ 10 | |
6122 | (* | |
6123 | (buffer-size) | |
6124 | (prefix-numeric-value arg))) | |
6125 | 10)) | |
6126 | @end group | |
6127 | @end smallexample | |
6128 | ||
6129 | @need 1200 | |
6130 | @noindent | |
6131 | Looking at parentheses, we see that the innermost operation is | |
6132 | @code{(prefix-numeric-value arg)}, which converts the raw argument to a | |
6133 | number. This number is multiplied by the buffer size in the following | |
6134 | expression: | |
6135 | ||
6136 | @smallexample | |
6137 | (* (buffer-size) (prefix-numeric-value arg) | |
6138 | @end smallexample | |
6139 | ||
6140 | @noindent | |
6141 | This multiplication creates a number that may be larger than the size of | |
6142 | the buffer---seven times larger if the argument is 7, for example. Ten | |
6143 | is then added to this number and finally the large number is divided by | |
6144 | ten to provide a value that is one character larger than the percentage | |
6145 | position in the buffer. | |
6146 | ||
6147 | The number that results from all this is passed to @code{goto-char} and | |
6148 | the cursor is moved to that point. | |
6149 | ||
6150 | @node beginning-of-buffer complete, , beginning-of-buffer opt arg, beginning-of-buffer | |
6151 | @comment node-name, next, previous, up | |
6152 | @subsection The Complete @code{beginning-of-buffer} | |
6153 | ||
6154 | @need 800 | |
6155 | Here is the complete text of the @code{beginning-of-buffer} function: | |
6156 | ||
6157 | @smallexample | |
6158 | @group | |
6159 | (defun beginning-of-buffer (&optional arg) | |
6160 | "Move point to the beginning of the buffer; | |
6161 | leave mark at previous position. | |
6162 | With arg N, put point N/10 of the way | |
6163 | from the true beginning. | |
6164 | Don't use this in Lisp programs! | |
6165 | \(goto-char (point-min)) is faster | |
6166 | and does not set the mark." | |
6167 | (interactive "P") | |
6168 | (push-mark) | |
6169 | @end group | |
6170 | @group | |
6171 | (goto-char | |
6172 | (if arg | |
6173 | (if (> (buffer-size) 10000) | |
6174 | ;; @r{Avoid overflow for large buffer sizes!} | |
6175 | (* (prefix-numeric-value arg) | |
6176 | (/ (buffer-size) 10)) | |
6177 | @end group | |
6178 | @group | |
6179 | (/ (+ 10 (* (buffer-size) | |
6180 | (prefix-numeric-value arg))) | |
6181 | 10)) | |
6182 | (point-min))) | |
6183 | (if arg (forward-line 1))) | |
6184 | @end group | |
6185 | @end smallexample | |
6186 | ||
6187 | @noindent | |
6188 | Except for two small points, the previous discussion shows how this | |
6189 | function works. The first point deals with a detail in the | |
6190 | documentation string, and the second point concerns the last line of | |
6191 | the function. | |
6192 | ||
6193 | @need 800 | |
6194 | In the documentation string, there is reference to an expression: | |
6195 | ||
6196 | @smallexample | |
6197 | \(goto-char (point-min)) | |
6198 | @end smallexample | |
6199 | ||
6200 | @noindent | |
6201 | A @samp{\} is used before the first parenthesis of this expression. | |
6202 | This @samp{\} tells the Lisp interpreter that the expression should be | |
6203 | printed as shown in the documentation rather than evaluated as a | |
6204 | symbolic expression, which is what it looks like. | |
6205 | ||
6206 | @need 1200 | |
6207 | Finally, the last line of the @code{beginning-of-buffer} command says to | |
6208 | move point to the beginning of the next line if the command is | |
6209 | invoked with an argument: | |
6210 | ||
6211 | @smallexample | |
6212 | (if arg (forward-line 1))) | |
6213 | @end smallexample | |
6214 | ||
6215 | @noindent | |
6216 | This puts the cursor at the beginning of the first line after the | |
6217 | appropriate tenths position in the buffer. This is a flourish that | |
6218 | means that the cursor is always located @emph{at least} the requested | |
6219 | tenths of the way through the buffer, which is a nicety that is, | |
6220 | perhaps, not necessary, but which, if it did not occur, would be sure to | |
6221 | draw complaints. | |
6222 | ||
6223 | @node Second Buffer Related Review, optional Exercise, beginning-of-buffer, More Complex | |
6224 | @comment node-name, next, previous, up | |
6225 | @section Review | |
6226 | ||
6227 | Here is a brief summary of some of the topics covered in this chapter. | |
6228 | ||
6229 | @table @code | |
6230 | @item or | |
6231 | Evaluate each argument in sequence, and return the value of the first | |
6232 | argument that is not @code{nil}; if none return a value that is not | |
6233 | @code{nil}, return @code{nil}. In brief, return the first true value | |
6234 | of the arguments; return a true value if one @emph{or} any of the | |
6235 | other are true. | |
6236 | ||
6237 | @item and | |
6238 | Evaluate each argument in sequence, and if any are @code{nil}, return | |
6239 | @code{nil}; if none are @code{nil}, return the value of the last | |
6240 | argument. In brief, return a true value only if all the arguments are | |
6241 | true; return a true value if one @emph{and} each of the others is | |
6242 | true. | |
6243 | ||
6244 | @item &optional | |
6245 | A keyword used to indicate that an argument to a function definition | |
6246 | is optional; this means that the function can be evaluated without the | |
6247 | argument, if desired. | |
6248 | ||
6249 | @item prefix-numeric-value | |
6250 | Convert the `raw prefix argument' produced by @code{(interactive | |
6251 | "P")} to a numeric value. | |
6252 | ||
6253 | @item forward-line | |
6254 | Move point forward to the beginning of the next line, or if the argument | |
6255 | is greater than one, forward that many lines. If it can't move as far | |
6256 | forward as it is supposed to, @code{forward-line} goes forward as far as | |
6257 | it can and then returns a count of the number of additional lines it was | |
6258 | supposed to move but couldn't. | |
6259 | ||
6260 | @item erase-buffer | |
6261 | Delete the entire contents of the current buffer. | |
6262 | ||
6263 | @item bufferp | |
6264 | Return @code{t} if its argument is a buffer; otherwise return @code{nil}. | |
6265 | @end table | |
6266 | ||
6267 | @node optional Exercise, , Second Buffer Related Review, More Complex | |
6268 | @section @code{optional} Argument Exercise | |
6269 | ||
6270 | Write an interactive function with an optional argument that tests | |
6271 | whether its argument, a number, is greater or less than the value of | |
6272 | @code{fill-column}, and tells you which, in a message. However, if you | |
6273 | do not pass an argument to the function, use 56 as a default value. | |
6274 | ||
6275 | @node Narrowing & Widening, car cdr & cons, More Complex, Top | |
6276 | @comment node-name, next, previous, up | |
6277 | @chapter Narrowing and Widening | |
6278 | @cindex Focusing attention (narrowing) | |
6279 | @cindex Narrowing | |
6280 | @cindex Widening | |
6281 | ||
6282 | Narrowing is a feature of Emacs that makes it possible for you to focus | |
6283 | on a specific part of a buffer, and work without accidentally changing | |
6284 | other parts. Narrowing is normally disabled since it can confuse | |
6285 | novices. | |
6286 | ||
6287 | @menu | |
6288 | * Narrowing advantages:: The advantages of narrowing | |
6289 | * save-restriction:: The @code{save-restriction} special form. | |
6290 | * what-line:: The number of the line that point is on. | |
6291 | * narrow Exercise:: | |
6292 | @end menu | |
6293 | ||
6294 | @node Narrowing advantages, save-restriction, Narrowing & Widening, Narrowing & Widening | |
6295 | @ifnottex | |
6296 | @unnumberedsec The Advantages of Narrowing | |
6297 | @end ifnottex | |
6298 | ||
6299 | With narrowing, the rest of a buffer is made invisible, as if it weren't | |
6300 | there. This is an advantage if, for example, you want to replace a word | |
6301 | in one part of a buffer but not in another: you narrow to the part you want | |
6302 | and the replacement is carried out only in that section, not in the rest | |
6303 | of the buffer. Searches will only work within a narrowed region, not | |
6304 | outside of one, so if you are fixing a part of a document, you can keep | |
6305 | yourself from accidentally finding parts you do not need to fix by | |
6306 | narrowing just to the region you want. | |
6307 | (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.) | |
6308 | ||
6309 | However, narrowing does make the rest of the buffer invisible, which | |
6310 | can scare people who inadvertently invoke narrowing and think they | |
6311 | have deleted a part of their file. Moreover, the @code{undo} command | |
6312 | (which is usually bound to @kbd{C-x u}) does not turn off narrowing | |
6313 | (nor should it), so people can become quite desperate if they do not | |
6314 | know that they can return the rest of a buffer to visibility with the | |
6315 | @code{widen} command. | |
6316 | (The key binding for @code{widen} is @kbd{C-x n w}.) | |
6317 | ||
6318 | Narrowing is just as useful to the Lisp interpreter as to a human. | |
6319 | Often, an Emacs Lisp function is designed to work on just part of a | |
6320 | buffer; or conversely, an Emacs Lisp function needs to work on all of a | |
6321 | buffer that has been narrowed. The @code{what-line} function, for | |
6322 | example, removes the narrowing from a buffer, if it has any narrowing | |
6323 | and when it has finished its job, restores the narrowing to what it was. | |
6324 | On the other hand, the @code{count-lines} function, which is called by | |
6325 | @code{what-line}, uses narrowing to restrict itself to just that portion | |
6326 | of the buffer in which it is interested and then restores the previous | |
6327 | situation. | |
6328 | ||
6329 | @node save-restriction, what-line, Narrowing advantages, Narrowing & Widening | |
6330 | @comment node-name, next, previous, up | |
6331 | @section The @code{save-restriction} Special Form | |
6332 | @findex save-restriction | |
6333 | ||
6334 | In Emacs Lisp, you can use the @code{save-restriction} special form to | |
6335 | keep track of whatever narrowing is in effect, if any. When the Lisp | |
6336 | interpreter meets with @code{save-restriction}, it executes the code | |
6337 | in the body of the @code{save-restriction} expression, and then undoes | |
6338 | any changes to narrowing that the code caused. If, for example, the | |
6339 | buffer is narrowed and the code that follows @code{save-restriction} | |
6340 | gets rid of the narrowing, @code{save-restriction} returns the buffer | |
6341 | to its narrowed region afterwards. In the @code{what-line} command, | |
6342 | any narrowing the buffer may have is undone by the @code{widen} | |
6343 | command that immediately follows the @code{save-restriction} command. | |
6344 | Any original narrowing is restored just before the completion of the | |
6345 | function. | |
6346 | ||
6347 | @need 1250 | |
6348 | The template for a @code{save-restriction} expression is simple: | |
6349 | ||
6350 | @smallexample | |
6351 | @group | |
6352 | (save-restriction | |
6353 | @var{body}@dots{} ) | |
6354 | @end group | |
6355 | @end smallexample | |
6356 | ||
6357 | @noindent | |
6358 | The body of the @code{save-restriction} is one or more expressions that | |
6359 | will be evaluated in sequence by the Lisp interpreter. | |
6360 | ||
6361 | Finally, a point to note: when you use both @code{save-excursion} and | |
6362 | @code{save-restriction}, one right after the other, you should use | |
6363 | @code{save-excursion} outermost. If you write them in reverse order, | |
6364 | you may fail to record narrowing in the buffer to which Emacs switches | |
6365 | after calling @code{save-excursion}. Thus, when written together, | |
6366 | @code{save-excursion} and @code{save-restriction} should be written | |
6367 | like this: | |
6368 | ||
6369 | @smallexample | |
6370 | @group | |
6371 | (save-excursion | |
6372 | (save-restriction | |
6373 | @var{body}@dots{})) | |
6374 | @end group | |
6375 | @end smallexample | |
6376 | ||
6377 | In other circumstances, when not written together, the | |
6378 | @code{save-excursion} and @code{save-restriction} special forms must | |
6379 | be written in the order appropriate to the function. | |
6380 | ||
6381 | @need 1250 | |
6382 | For example, | |
6383 | ||
6384 | @smallexample | |
6385 | @group | |
6386 | (save-restriction | |
6387 | (widen) | |
6388 | (save-excursion | |
6389 | @var{body}@dots{})) | |
6390 | @end group | |
6391 | @end smallexample | |
6392 | ||
6393 | @node what-line, narrow Exercise, save-restriction, Narrowing & Widening | |
6394 | @comment node-name, next, previous, up | |
6395 | @section @code{what-line} | |
6396 | @findex what-line | |
6397 | @cindex Widening, example of | |
6398 | ||
6399 | The @code{what-line} command tells you the number of the line in which | |
6400 | the cursor is located. The function illustrates the use of the | |
6401 | @code{save-restriction} and @code{save-excursion} commands. Here is the | |
6402 | text of the function in full: | |
6403 | ||
6404 | @smallexample | |
6405 | @group | |
6406 | (defun what-line () | |
6407 | "Print the current line number (in the buffer) of point." | |
6408 | (interactive) | |
6409 | (save-restriction | |
6410 | (widen) | |
6411 | (save-excursion | |
6412 | (beginning-of-line) | |
6413 | (message "Line %d" | |
6414 | (1+ (count-lines 1 (point))))))) | |
6415 | @end group | |
6416 | @end smallexample | |
6417 | ||
6418 | The function has a documentation line and is interactive, as you would | |
6419 | expect. The next two lines use the functions @code{save-restriction} and | |
6420 | @code{widen}. | |
6421 | ||
6422 | The @code{save-restriction} special form notes whatever narrowing is in | |
6423 | effect, if any, in the current buffer and restores that narrowing after | |
6424 | the code in the body of the @code{save-restriction} has been evaluated. | |
6425 | ||
6426 | The @code{save-restriction} special form is followed by @code{widen}. | |
6427 | This function undoes any narrowing the current buffer may have had | |
6428 | when @code{what-line} was called. (The narrowing that was there is | |
6429 | the narrowing that @code{save-restriction} remembers.) This widening | |
6430 | makes it possible for the line counting commands to count from the | |
6431 | beginning of the buffer. Otherwise, they would have been limited to | |
6432 | counting within the accessible region. Any original narrowing is | |
6433 | restored just before the completion of the function by the | |
6434 | @code{save-restriction} special form. | |
6435 | ||
6436 | The call to @code{widen} is followed by @code{save-excursion}, which | |
6437 | saves the location of the cursor (i.e., of point) and of the mark, and | |
6438 | restores them after the code in the body of the @code{save-excursion} | |
6439 | uses the @code{beginning-of-line} function to move point. | |
6440 | ||
6441 | (Note that the @code{(widen)} expression comes between the | |
6442 | @code{save-restriction} and @code{save-excursion} special forms. When | |
6443 | you write the two @code{save- @dots{}} expressions in sequence, write | |
6444 | @code{save-excursion} outermost.) | |
6445 | ||
6446 | @need 1200 | |
6447 | The last two lines of the @code{what-line} function are functions to | |
6448 | count the number of lines in the buffer and then print the number in the | |
6449 | echo area. | |
6450 | ||
6451 | @smallexample | |
6452 | @group | |
6453 | (message "Line %d" | |
6454 | (1+ (count-lines 1 (point))))))) | |
6455 | @end group | |
6456 | @end smallexample | |
6457 | ||
6458 | The @code{message} function prints a one-line message at the bottom of the | |
6459 | Emacs screen. The first argument is inside of quotation marks and is | |
6460 | printed as a string of characters. However, it may contain @samp{%d}, | |
6461 | @samp{%s}, or @samp{%c} to print arguments that follow the string. | |
6462 | @samp{%d} prints the argument as a decimal, so the message will say | |
6463 | something such as @samp{Line 243}. | |
6464 | ||
6465 | @need 1200 | |
6466 | The number that is printed in place of the @samp{%d} is computed by the | |
6467 | last line of the function: | |
6468 | ||
6469 | @smallexample | |
6470 | (1+ (count-lines 1 (point))) | |
6471 | @end smallexample | |
6472 | ||
6473 | @noindent | |
6474 | What this does is count the lines from the first position of the | |
6475 | buffer, indicated by the @code{1}, up to @code{(point)}, and then add | |
6476 | one to that number. (The @code{1+} function adds one to its | |
6477 | argument.) We add one to it because line 2 has only one line before | |
6478 | it, and @code{count-lines} counts only the lines @emph{before} the | |
6479 | current line. | |
6480 | ||
6481 | After @code{count-lines} has done its job, and the message has been | |
6482 | printed in the echo area, the @code{save-excursion} restores point and | |
6483 | mark to their original positions; and @code{save-restriction} restores | |
6484 | the original narrowing, if any. | |
6485 | ||
6486 | @node narrow Exercise, , what-line, Narrowing & Widening | |
6487 | @section Exercise with Narrowing | |
6488 | ||
6489 | Write a function that will display the first 60 characters of the | |
6490 | current buffer, even if you have narrowed the buffer to its latter | |
6491 | half so that the first line is inaccessible. Restore point, mark, | |
6492 | and narrowing. For this exercise, you need to use | |
6493 | @code{save-restriction}, @code{widen}, @code{goto-char}, | |
6494 | @code{point-min}, @code{buffer-substring}, @code{message}, and other | |
6495 | functions, a whole potpourri. | |
6496 | ||
6497 | @node car cdr & cons, Cutting & Storing Text, Narrowing & Widening, Top | |
6498 | @comment node-name, next, previous, up | |
6499 | @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions | |
6500 | @findex car, @r{introduced} | |
6501 | @findex cdr, @r{introduced} | |
6502 | ||
6503 | In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental | |
6504 | functions. The @code{cons} function is used to construct lists, and | |
6505 | the @code{car} and @code{cdr} functions are used to take them apart. | |
6506 | ||
6507 | In the walk through of the @code{copy-region-as-kill} function, we | |
6508 | will see @code{cons} as well as two variants on @code{cdr}, | |
6509 | namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.) | |
6510 | ||
6511 | @menu | |
6512 | * Strange Names:: An historical aside: why the strange names? | |
6513 | * car & cdr:: Functions for extracting part of a list. | |
6514 | * cons:: Constructing a list. | |
6515 | * nthcdr:: Calling @code{cdr} repeatedly. | |
6516 | * nth:: | |
6517 | * setcar:: Changing the first element of a list. | |
6518 | * setcdr:: Changing the rest of a list. | |
6519 | * cons Exercise:: | |
6520 | @end menu | |
6521 | ||
6522 | @node Strange Names, car & cdr, car cdr & cons, car cdr & cons | |
6523 | @ifnottex | |
6524 | @unnumberedsec Strange Names | |
6525 | @end ifnottex | |
6526 | ||
6527 | The name of the @code{cons} function is not unreasonable: it is an | |
6528 | abbreviation of the word `construct'. The origins of the names for | |
6529 | @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car} | |
6530 | is an acronym from the phrase `Contents of the Address part of the | |
6531 | Register'; and @code{cdr} (pronounced `could-er') is an acronym from | |
6532 | the phrase `Contents of the Decrement part of the Register'. These | |
6533 | phrases refer to specific pieces of hardware on the very early | |
6534 | computer on which the original Lisp was developed. Besides being | |
6535 | obsolete, the phrases have been completely irrelevant for more than 25 | |
6536 | years to anyone thinking about Lisp. Nonetheless, although a few | |
6537 | brave scholars have begun to use more reasonable names for these | |
6538 | functions, the old terms are still in use. In particular, since the | |
6539 | terms are used in the Emacs Lisp source code, we will use them in this | |
6540 | introduction. | |
6541 | ||
6542 | @node car & cdr, cons, Strange Names, car cdr & cons | |
6543 | @comment node-name, next, previous, up | |
6544 | @section @code{car} and @code{cdr} | |
6545 | ||
6546 | The @sc{car} of a list is, quite simply, the first item in the list. | |
6547 | Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is | |
6548 | @code{rose}. | |
6549 | ||
6550 | @need 1200 | |
6551 | If you are reading this in Info in GNU Emacs, you can see this by | |
6552 | evaluating the following: | |
6553 | ||
6554 | @smallexample | |
6555 | (car '(rose violet daisy buttercup)) | |
6556 | @end smallexample | |
6557 | ||
6558 | @noindent | |
6559 | After evaluating the expression, @code{rose} will appear in the echo | |
6560 | area. | |
6561 | ||
6562 | Clearly, a more reasonable name for the @code{car} function would be | |
6563 | @code{first} and this is often suggested. | |
6564 | ||
6565 | @code{car} does not remove the first item from the list; it only reports | |
6566 | what it is. After @code{car} has been applied to a list, the list is | |
6567 | still the same as it was. In the jargon, @code{car} is | |
6568 | `non-destructive'. This feature turns out to be important. | |
6569 | ||
6570 | The @sc{cdr} of a list is the rest of the list, that is, the | |
6571 | @code{cdr} function returns the part of the list that follows the | |
6572 | first item. Thus, while the @sc{car} of the list @code{'(rose violet | |
6573 | daisy buttercup)} is @code{rose}, the rest of the list, the value | |
6574 | returned by the @code{cdr} function, is @code{(violet daisy | |
6575 | buttercup)}. | |
6576 | ||
6577 | @need 1250 | |
6578 | You can see this by evaluating the following in the usual way: | |
6579 | ||
6580 | @smallexample | |
6581 | (cdr '(rose violet daisy buttercup)) | |
6582 | @end smallexample | |
6583 | ||
6584 | @noindent | |
6585 | When you evaluate this, @code{(violet daisy buttercup)} will appear in | |
6586 | the echo area. | |
6587 | ||
6588 | Like @code{car}, @code{cdr} does not remove any elements from the | |
6589 | list---it just returns a report of what the second and subsequent | |
6590 | elements are. | |
6591 | ||
6592 | Incidentally, in the example, the list of flowers is quoted. If it were | |
6593 | not, the Lisp interpreter would try to evaluate the list by calling | |
6594 | @code{rose} as a function. In this example, we do not want to do that. | |
6595 | ||
6596 | Clearly, a more reasonable name for @code{cdr} would be @code{rest}. | |
6597 | ||
6598 | (There is a lesson here: when you name new functions, consider very | |
6599 | carefully what you are doing, since you may be stuck with the names | |
6600 | for far longer than you expect. The reason this document perpetuates | |
6601 | these names is that the Emacs Lisp source code uses them, and if I did | |
6602 | not use them, you would have a hard time reading the code; but do, | |
6603 | please, try to avoid using these terms yourself. The people who come | |
6604 | after you will be grateful to you.) | |
6605 | ||
6606 | When @code{car} and @code{cdr} are applied to a list made up of symbols, | |
6607 | such as the list @code{(pine fir oak maple)}, the element of the list | |
6608 | returned by the function @code{car} is the symbol @code{pine} without | |
6609 | any parentheses around it. @code{pine} is the first element in the | |
6610 | list. However, the @sc{cdr} of the list is a list itself, @code{(fir | |
6611 | oak maple)}, as you can see by evaluating the following expressions in | |
6612 | the usual way: | |
6613 | ||
6614 | @smallexample | |
6615 | @group | |
6616 | (car '(pine fir oak maple)) | |
6617 | ||
6618 | (cdr '(pine fir oak maple)) | |
6619 | @end group | |
6620 | @end smallexample | |
6621 | ||
6622 | On the other hand, in a list of lists, the first element is itself a | |
6623 | list. @code{car} returns this first element as a list. For example, | |
6624 | the following list contains three sub-lists, a list of carnivores, a | |
6625 | list of herbivores and a list of sea mammals: | |
6626 | ||
6627 | @smallexample | |
6628 | @group | |
6629 | (car '((lion tiger cheetah) | |
6630 | (gazelle antelope zebra) | |
6631 | (whale dolphin seal))) | |
6632 | @end group | |
6633 | @end smallexample | |
6634 | ||
6635 | @noindent | |
6636 | In this example, the first element or @sc{car} of the list is the list of | |
6637 | carnivores, @code{(lion tiger cheetah)}, and the rest of the list is | |
6638 | @code{((gazelle antelope zebra) (whale dolphin seal))}. | |
6639 | ||
6640 | @smallexample | |
6641 | @group | |
6642 | (cdr '((lion tiger cheetah) | |
6643 | (gazelle antelope zebra) | |
6644 | (whale dolphin seal))) | |
6645 | @end group | |
6646 | @end smallexample | |
6647 | ||
6648 | It is worth saying again that @code{car} and @code{cdr} are | |
6649 | non-destructive---that is, they do not modify or change lists to which | |
6650 | they are applied. This is very important for how they are used. | |
6651 | ||
6652 | Also, in the first chapter, in the discussion about atoms, I said that | |
6653 | in Lisp, ``certain kinds of atom, such as an array, can be separated | |
6654 | into parts; but the mechanism for doing this is different from the | |
6655 | mechanism for splitting a list. As far as Lisp is concerned, the | |
6656 | atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The | |
6657 | @code{car} and @code{cdr} functions are used for splitting lists and | |
6658 | are considered fundamental to Lisp. Since they cannot split or gain | |
6659 | access to the parts of an array, an array is considered an atom. | |
6660 | Conversely, the other fundamental function, @code{cons}, can put | |
6661 | together or construct a list, but not an array. (Arrays are handled | |
6662 | by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU | |
6663 | Emacs Lisp Reference Manual}.) | |
6664 | ||
6665 | @node cons, nthcdr, car & cdr, car cdr & cons | |
6666 | @comment node-name, next, previous, up | |
6667 | @section @code{cons} | |
6668 | @findex cons, @r{introduced} | |
6669 | ||
6670 | The @code{cons} function constructs lists; it is the inverse of | |
6671 | @code{car} and @code{cdr}. For example, @code{cons} can be used to make | |
6672 | a four element list from the three element list, @code{(fir oak maple)}: | |
6673 | ||
6674 | @smallexample | |
6675 | (cons 'pine '(fir oak maple)) | |
6676 | @end smallexample | |
6677 | ||
6678 | @need 800 | |
6679 | @noindent | |
6680 | After evaluating this list, you will see | |
6681 | ||
6682 | @smallexample | |
6683 | (pine fir oak maple) | |
6684 | @end smallexample | |
6685 | ||
6686 | @noindent | |
6687 | appear in the echo area. @code{cons} puts a new element at the | |
6688 | beginning of a list; it attaches or pushes elements onto the list. | |
6689 | ||
6690 | @menu | |
6691 | * Build a list:: | |
6692 | * length:: How to find the length of a list. | |
6693 | @end menu | |
6694 | ||
6695 | @node Build a list, length, cons, cons | |
6696 | @ifnottex | |
6697 | @unnumberedsubsec Build a list | |
6698 | @end ifnottex | |
6699 | ||
6700 | @code{cons} must have a list to attach to.@footnote{Actually, you can | |
6701 | @code{cons} an element to an atom to produce a dotted pair. Dotted | |
6702 | pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted | |
6703 | Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You | |
6704 | cannot start from absolutely nothing. If you are building a list, you | |
6705 | need to provide at least an empty list at the beginning. Here is a | |
6706 | series of @code{cons} expressions that build up a list of flowers. If | |
6707 | you are reading this in Info in GNU Emacs, you can evaluate each of | |
6708 | the expressions in the usual way; the value is printed in this text | |
6709 | after @samp{@result{}}, which you may read as `evaluates to'. | |
6710 | ||
6711 | @smallexample | |
6712 | @group | |
6713 | (cons 'buttercup ()) | |
6714 | @result{} (buttercup) | |
6715 | @end group | |
6716 | ||
6717 | @group | |
6718 | (cons 'daisy '(buttercup)) | |
6719 | @result{} (daisy buttercup) | |
6720 | @end group | |
6721 | ||
6722 | @group | |
6723 | (cons 'violet '(daisy buttercup)) | |
6724 | @result{} (violet daisy buttercup) | |
6725 | @end group | |
6726 | ||
6727 | @group | |
6728 | (cons 'rose '(violet daisy buttercup)) | |
6729 | @result{} (rose violet daisy buttercup) | |
6730 | @end group | |
6731 | @end smallexample | |
6732 | ||
6733 | @noindent | |
6734 | In the first example, the empty list is shown as @code{()} and a list | |
6735 | made up of @code{buttercup} followed by the empty list is constructed. | |
6736 | As you can see, the empty list is not shown in the list that was | |
6737 | constructed. All that you see is @code{(buttercup)}. The empty list is | |
6738 | not counted as an element of a list because there is nothing in an empty | |
6739 | list. Generally speaking, an empty list is invisible. | |
6740 | ||
6741 | The second example, @code{(cons 'daisy '(buttercup))} constructs a new, | |
6742 | two element list by putting @code{daisy} in front of @code{buttercup}; | |
6743 | and the third example constructs a three element list by putting | |
6744 | @code{violet} in front of @code{daisy} and @code{buttercup}. | |
6745 | ||
6746 | @node length, , Build a list, cons | |
6747 | @comment node-name, next, previous, up | |
6748 | @subsection Find the Length of a List: @code{length} | |
6749 | @findex length | |
6750 | ||
6751 | You can find out how many elements there are in a list by using the Lisp | |
6752 | function @code{length}, as in the following examples: | |
6753 | ||
6754 | @smallexample | |
6755 | @group | |
6756 | (length '(buttercup)) | |
6757 | @result{} 1 | |
6758 | @end group | |
6759 | ||
6760 | @group | |
6761 | (length '(daisy buttercup)) | |
6762 | @result{} 2 | |
6763 | @end group | |
6764 | ||
6765 | @group | |
6766 | (length (cons 'violet '(daisy buttercup))) | |
6767 | @result{} 3 | |
6768 | @end group | |
6769 | @end smallexample | |
6770 | ||
6771 | @noindent | |
6772 | In the third example, the @code{cons} function is used to construct a | |
6773 | three element list which is then passed to the @code{length} function as | |
6774 | its argument. | |
6775 | ||
6776 | @need 1200 | |
6777 | We can also use @code{length} to count the number of elements in an | |
6778 | empty list: | |
6779 | ||
6780 | @smallexample | |
6781 | @group | |
6782 | (length ()) | |
6783 | @result{} 0 | |
6784 | @end group | |
6785 | @end smallexample | |
6786 | ||
6787 | @noindent | |
6788 | As you would expect, the number of elements in an empty list is zero. | |
6789 | ||
6790 | An interesting experiment is to find out what happens if you try to find | |
6791 | the length of no list at all; that is, if you try to call @code{length} | |
6792 | without giving it an argument, not even an empty list: | |
6793 | ||
6794 | @smallexample | |
6795 | (length ) | |
6796 | @end smallexample | |
6797 | ||
6798 | @need 800 | |
6799 | @noindent | |
6800 | What you see, if you evaluate this, is the error message | |
6801 | ||
6802 | @smallexample | |
6803 | Wrong number of arguments: #<subr length>, 0 | |
6804 | @end smallexample | |
6805 | ||
6806 | @noindent | |
6807 | This means that the function receives the wrong number of | |
6808 | arguments, zero, when it expects some other number of arguments. In | |
6809 | this case, one argument is expected, the argument being a list whose | |
6810 | length the function is measuring. (Note that @emph{one} list is | |
6811 | @emph{one} argument, even if the list has many elements inside it.) | |
6812 | ||
6813 | The part of the error message that says @samp{#<subr length>} is the | |
6814 | name of the function. This is written with a special notation, | |
6815 | @samp{#<subr}, that indicates that the function @code{length} is one | |
6816 | of the primitive functions written in C rather than in Emacs Lisp. | |
6817 | (@samp{subr} is an abbreviation for `subroutine'.) @xref{What Is a | |
6818 | Function, , What Is a Function?, elisp , The GNU Emacs Lisp Reference | |
6819 | Manual}, for more about subroutines. | |
6820 | ||
6821 | @node nthcdr, nth, cons, car cdr & cons | |
6822 | @comment node-name, next, previous, up | |
6823 | @section @code{nthcdr} | |
6824 | @findex nthcdr | |
6825 | ||
6826 | The @code{nthcdr} function is associated with the @code{cdr} function. | |
6827 | What it does is take the @sc{cdr} of a list repeatedly. | |
6828 | ||
6829 | If you take the @sc{cdr} of the list @code{(pine fir | |
6830 | oak maple)}, you will be returned the list @code{(fir oak maple)}. If you | |
6831 | repeat this on what was returned, you will be returned the list | |
6832 | @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original | |
6833 | list will just give you the original @sc{cdr} since the function does | |
6834 | not change the list. You need to evaluate the @sc{cdr} of the | |
6835 | @sc{cdr} and so on.) If you continue this, eventually you will be | |
6836 | returned an empty list, which in this case, instead of being shown as | |
6837 | @code{()} is shown as @code{nil}. | |
6838 | ||
6839 | @need 1200 | |
6840 | For review, here is a series of repeated @sc{cdr}s, the text following | |
6841 | the @samp{@result{}} shows what is returned. | |
6842 | ||
6843 | @smallexample | |
6844 | @group | |
6845 | (cdr '(pine fir oak maple)) | |
6846 | @result{}(fir oak maple) | |
6847 | @end group | |
6848 | ||
6849 | @group | |
6850 | (cdr '(fir oak maple)) | |
6851 | @result{} (oak maple) | |
6852 | @end group | |
6853 | ||
6854 | @group | |
6855 | (cdr '(oak maple)) | |
6856 | @result{}(maple) | |
6857 | @end group | |
6858 | ||
6859 | @group | |
6860 | (cdr '(maple)) | |
6861 | @result{} nil | |
6862 | @end group | |
6863 | ||
6864 | @group | |
6865 | (cdr 'nil) | |
6866 | @result{} nil | |
6867 | @end group | |
6868 | ||
6869 | @group | |
6870 | (cdr ()) | |
6871 | @result{} nil | |
6872 | @end group | |
6873 | @end smallexample | |
6874 | ||
6875 | @need 1200 | |
6876 | You can also do several @sc{cdr}s without printing the values in | |
6877 | between, like this: | |
6878 | ||
6879 | @smallexample | |
6880 | @group | |
6881 | (cdr (cdr '(pine fir oak maple))) | |
6882 | @result{} (oak maple) | |
6883 | @end group | |
6884 | @end smallexample | |
6885 | ||
6886 | @noindent | |
6887 | In this example, the Lisp interpreter evaluates the innermost list first. | |
6888 | The innermost list is quoted, so it just passes the list as it is to the | |
6889 | innermost @code{cdr}. This @code{cdr} passes a list made up of the | |
6890 | second and subsequent elements of the list to the outermost @code{cdr}, | |
6891 | which produces a list composed of the third and subsequent elements of | |
6892 | the original list. In this example, the @code{cdr} function is repeated | |
6893 | and returns a list that consists of the original list without its | |
6894 | first two elements. | |
6895 | ||
6896 | The @code{nthcdr} function does the same as repeating the call to | |
6897 | @code{cdr}. In the following example, the argument 2 is passed to the | |
6898 | function @code{nthcdr}, along with the list, and the value returned is | |
6899 | the list without its first two items, which is exactly the same | |
6900 | as repeating @code{cdr} twice on the list: | |
6901 | ||
6902 | @smallexample | |
6903 | @group | |
6904 | (nthcdr 2 '(pine fir oak maple)) | |
6905 | @result{} (oak maple) | |
6906 | @end group | |
6907 | @end smallexample | |
6908 | ||
6909 | @need 1200 | |
6910 | Using the original four element list, we can see what happens when | |
6911 | various numeric arguments are passed to @code{nthcdr}, including 0, 1, | |
6912 | and 5: | |
6913 | ||
6914 | @smallexample | |
6915 | @group | |
6916 | ;; @r{Leave the list as it was.} | |
6917 | (nthcdr 0 '(pine fir oak maple)) | |
6918 | @result{} (pine fir oak maple) | |
6919 | @end group | |
6920 | ||
6921 | @group | |
6922 | ;; @r{Return a copy without the first element.} | |
6923 | (nthcdr 1 '(pine fir oak maple)) | |
6924 | @result{} (fir oak maple) | |
6925 | @end group | |
6926 | ||
6927 | @group | |
6928 | ;; @r{Return a copy of the list without three elements.} | |
6929 | (nthcdr 3 '(pine fir oak maple)) | |
6930 | @result{} (maple) | |
6931 | @end group | |
6932 | ||
6933 | @group | |
6934 | ;; @r{Return a copy lacking all four elements.} | |
6935 | (nthcdr 4 '(pine fir oak maple)) | |
6936 | @result{} nil | |
6937 | @end group | |
6938 | ||
6939 | @group | |
6940 | ;; @r{Return a copy lacking all elements.} | |
6941 | (nthcdr 5 '(pine fir oak maple)) | |
6942 | @result{} nil | |
6943 | @end group | |
6944 | @end smallexample | |
6945 | ||
6946 | @node nth, setcar, nthcdr, car cdr & cons | |
6947 | @comment node-name, next, previous, up | |
6948 | @section @code{nth} | |
6949 | @findex nth | |
6950 | ||
6951 | The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly. | |
6952 | The @code{nth} function takes the @sc{car} of the result returned by | |
6953 | @code{nthcdr}. It returns the Nth element of the list. | |
6954 | ||
6955 | @need 1500 | |
6956 | Thus, if it were not defined in C for speed, the definition of | |
6957 | @code{nth} would be: | |
6958 | ||
6959 | @smallexample | |
6960 | @group | |
6961 | (defun nth (n list) | |
6962 | "Returns the Nth element of LIST. | |
6963 | N counts from zero. If LIST is not that long, nil is returned." | |
6964 | (car (nthcdr n list))) | |
6965 | @end group | |
6966 | @end smallexample | |
6967 | ||
6968 | @noindent | |
6969 | (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el}, | |
6970 | but its definition was redone in C in the 1980s.) | |
6971 | ||
6972 | The @code{nth} function returns a single element of a list. | |
6973 | This can be very convenient. | |
6974 | ||
6975 | Note that the elements are numbered from zero, not one. That is to | |
6976 | say, the first element of a list, its @sc{car} is the zeroth element. | |
6977 | This is called `zero-based' counting and often bothers people who | |
6978 | are accustomed to the first element in a list being number one, which | |
6979 | is `one-based'. | |
6980 | ||
6981 | @need 1250 | |
6982 | For example: | |
6983 | ||
6984 | @smallexample | |
6985 | @group | |
6986 | (nth 0 '("one" "two" "three")) | |
6987 | @result{} "one" | |
6988 | ||
6989 | (nth 1 '("one" "two" "three")) | |
6990 | @result{} "two" | |
6991 | @end group | |
6992 | @end smallexample | |
6993 | ||
6994 | It is worth mentioning that @code{nth}, like @code{nthcdr} and | |
6995 | @code{cdr}, does not change the original list---the function is | |
6996 | non-destructive. This is in sharp contrast to the @code{setcar} and | |
6997 | @code{setcdr} functions. | |
6998 | ||
6999 | @node setcar, setcdr, nth, car cdr & cons | |
7000 | @comment node-name, next, previous, up | |
7001 | @section @code{setcar} | |
7002 | @findex setcar | |
7003 | ||
7004 | As you might guess from their names, the @code{setcar} and @code{setcdr} | |
7005 | functions set the @sc{car} or the @sc{cdr} of a list to a new value. | |
7006 | They actually change the original list, unlike @code{car} and @code{cdr} | |
7007 | which leave the original list as it was. One way to find out how this | |
7008 | works is to experiment. We will start with the @code{setcar} function. | |
7009 | ||
7010 | @need 1200 | |
7011 | First, we can make a list and then set the value of a variable to the | |
7012 | list, using the @code{setq} function. Here is a list of animals: | |
7013 | ||
7014 | @smallexample | |
7015 | (setq animals '(antelope giraffe lion tiger)) | |
7016 | @end smallexample | |
7017 | ||
7018 | @noindent | |
7019 | If you are reading this in Info inside of GNU Emacs, you can evaluate | |
7020 | this expression in the usual fashion, by positioning the cursor after | |
7021 | the expression and typing @kbd{C-x C-e}. (I'm doing this right here as | |
7022 | I write this. This is one of the advantages of having the interpreter | |
7023 | built into the computing environment.) | |
7024 | ||
7025 | @need 1200 | |
7026 | When we evaluate the variable @code{animals}, we see that it is bound to | |
7027 | the list @code{(antelope giraffe lion tiger)}: | |
7028 | ||
7029 | @smallexample | |
7030 | @group | |
7031 | animals | |
7032 | @result{} (antelope giraffe lion tiger) | |
7033 | @end group | |
7034 | @end smallexample | |
7035 | ||
7036 | @noindent | |
7037 | Put another way, the variable @code{animals} points to the list | |
7038 | @code{(antelope giraffe lion tiger)}. | |
7039 | ||
7040 | Next, evaluate the function @code{setcar} while passing it two | |
7041 | arguments, the variable @code{animals} and the quoted symbol | |
7042 | @code{hippopotamus}; this is done by writing the three element list | |
7043 | @code{(setcar animals 'hippopotamus)} and then evaluating it in the | |
7044 | usual fashion: | |
7045 | ||
7046 | @smallexample | |
7047 | (setcar animals 'hippopotamus) | |
7048 | @end smallexample | |
7049 | ||
7050 | @need 1200 | |
7051 | @noindent | |
7052 | After evaluating this expression, evaluate the variable @code{animals} | |
7053 | again. You will see that the list of animals has changed: | |
7054 | ||
7055 | @smallexample | |
7056 | @group | |
7057 | animals | |
7058 | @result{} (hippopotamus giraffe lion tiger) | |
7059 | @end group | |
7060 | @end smallexample | |
7061 | ||
7062 | @noindent | |
7063 | The first element on the list, @code{antelope} is replaced by | |
7064 | @code{hippopotamus}. | |
7065 | ||
7066 | So we can see that @code{setcar} did not add a new element to the list | |
7067 | as @code{cons} would have; it replaced @code{giraffe} with | |
7068 | @code{hippopotamus}; it @emph{changed} the list. | |
7069 | ||
7070 | @node setcdr, cons Exercise, setcar, car cdr & cons | |
7071 | @comment node-name, next, previous, up | |
7072 | @section @code{setcdr} | |
7073 | @findex setcdr | |
7074 | ||
7075 | The @code{setcdr} function is similar to the @code{setcar} function, | |
7076 | except that the function replaces the second and subsequent elements of | |
7077 | a list rather than the first element. | |
7078 | ||
7079 | @need 1200 | |
7080 | To see how this works, set the value of the variable to a list of | |
7081 | domesticated animals by evaluating the following expression: | |
7082 | ||
7083 | @smallexample | |
7084 | (setq domesticated-animals '(horse cow sheep goat)) | |
7085 | @end smallexample | |
7086 | ||
7087 | @need 1200 | |
7088 | @noindent | |
7089 | If you now evaluate the list, you will be returned the list | |
7090 | @code{(horse cow sheep goat)}: | |
7091 | ||
7092 | @smallexample | |
7093 | @group | |
7094 | domesticated-animals | |
7095 | @result{} (horse cow sheep goat) | |
7096 | @end group | |
7097 | @end smallexample | |
7098 | ||
7099 | @need 1200 | |
7100 | Next, evaluate @code{setcdr} with two arguments, the name of the | |
7101 | variable which has a list as its value, and the list to which the | |
7102 | @sc{cdr} of the first list will be set; | |
7103 | ||
7104 | @smallexample | |
7105 | (setcdr domesticated-animals '(cat dog)) | |
7106 | @end smallexample | |
7107 | ||
7108 | @noindent | |
7109 | If you evaluate this expression, the list @code{(cat dog)} will appear | |
7110 | in the echo area. This is the value returned by the function. The | |
7111 | result we are interested in is the ``side effect'', which we can see by | |
7112 | evaluating the variable @code{domesticated-animals}: | |
7113 | ||
7114 | @smallexample | |
7115 | @group | |
7116 | domesticated-animals | |
7117 | @result{} (horse cat dog) | |
7118 | @end group | |
7119 | @end smallexample | |
7120 | ||
7121 | @noindent | |
7122 | Indeed, the list is changed from @code{(horse cow sheep goat)} to | |
7123 | @code{(horse cat dog)}. The @sc{cdr} of the list is changed from | |
7124 | @code{(cow sheep goat)} to @code{(cat dog)}. | |
7125 | ||
7126 | @node cons Exercise, , setcdr, car cdr & cons | |
7127 | @section Exercise | |
7128 | ||
7129 | Construct a list of four birds by evaluating several expressions with | |
7130 | @code{cons}. Find out what happens when you @code{cons} a list onto | |
7131 | itself. Replace the first element of the list of four birds with a | |
7132 | fish. Replace the rest of that list with a list of other fish. | |
7133 | @node Cutting & Storing Text, List Implementation, car cdr & cons, Top | |
7134 | @comment node-name, next, previous, up | |
7135 | @chapter Cutting and Storing Text | |
7136 | @cindex Cutting and storing text | |
7137 | @cindex Storing and cutting text | |
7138 | @cindex Killing text | |
7139 | @cindex Clipping text | |
7140 | @cindex Erasing text | |
7141 | @cindex Deleting text | |
7142 | ||
7143 | Whenever you cut or clip text out of a buffer with a `kill' command in | |
7144 | GNU Emacs, it is stored in a list and you can bring it back with a | |
7145 | `yank' command. | |
7146 | ||
7147 | (The use of the word `kill' in Emacs for processes which specifically | |
7148 | @emph{do not} destroy the values of the entities is an unfortunate | |
7149 | historical accident. A much more appropriate word would be `clip' since | |
7150 | that is what the kill commands do; they clip text out of a buffer and | |
7151 | put it into storage from which it can be brought back. I have often | |
7152 | been tempted to replace globally all occurrences of `kill' in the Emacs | |
7153 | sources with `clip' and all occurrences of `killed' with `clipped'.) | |
7154 | ||
7155 | @menu | |
7156 | * Storing Text:: Text is stored in a list. | |
7157 | * zap-to-char:: Cutting out text up to a character. | |
7158 | * kill-region:: Cutting text out of a region. | |
7159 | * Digression into C:: Minor note on C programming language macros. | |
7160 | * defvar:: How to give a variable an initial value. | |
7161 | * copy-region-as-kill:: A definition for copying text. | |
7162 | * cons & search-fwd Review:: | |
7163 | * search Exercises:: | |
7164 | @end menu | |
7165 | ||
7166 | @node Storing Text, zap-to-char, Cutting & Storing Text, Cutting & Storing Text | |
7167 | @ifnottex | |
7168 | @unnumberedsec Storing Text in a List | |
7169 | @end ifnottex | |
7170 | ||
7171 | When text is cut out of a buffer, it is stored on a list. Successive | |
7172 | pieces of text are stored on the list successively, so the list might | |
7173 | look like this: | |
7174 | ||
7175 | @smallexample | |
7176 | ("a piece of text" "previous piece") | |
7177 | @end smallexample | |
7178 | ||
7179 | @need 1200 | |
7180 | @noindent | |
7181 | The function @code{cons} can be used to add a piece of text to the list, | |
7182 | like this: | |
7183 | ||
7184 | @smallexample | |
7185 | @group | |
7186 | (cons "another piece" | |
7187 | '("a piece of text" "previous piece")) | |
7188 | @end group | |
7189 | @end smallexample | |
7190 | ||
7191 | @need 1200 | |
7192 | @noindent | |
7193 | If you evaluate this expression, a list of three elements will appear in | |
7194 | the echo area: | |
7195 | ||
7196 | @smallexample | |
7197 | ("another piece" "a piece of text" "previous piece") | |
7198 | @end smallexample | |
7199 | ||
7200 | With the @code{car} and @code{nthcdr} functions, you can retrieve | |
7201 | whichever piece of text you want. For example, in the following code, | |
7202 | @code{nthcdr 1 @dots{}} returns the list with the first item removed; | |
7203 | and the @code{car} returns the first element of that remainder---the | |
7204 | second element of the original list: | |
7205 | ||
7206 | @smallexample | |
7207 | @group | |
7208 | (car (nthcdr 1 '("another piece" | |
7209 | "a piece of text" | |
7210 | "previous piece"))) | |
7211 | @result{} "a piece of text" | |
7212 | @end group | |
7213 | @end smallexample | |
7214 | ||
7215 | The actual functions in Emacs are more complex than this, of course. | |
7216 | The code for cutting and retrieving text has to be written so that | |
7217 | Emacs can figure out which element in the list you want---the first, | |
7218 | second, third, or whatever. In addition, when you get to the end of | |
7219 | the list, Emacs should give you the first element of the list, rather | |
7220 | than nothing at all. | |
7221 | ||
7222 | The list that holds the pieces of text is called the @dfn{kill ring}. | |
7223 | This chapter leads up to a description of the kill ring and how it is | |
7224 | used by first tracing how the @code{zap-to-char} function works. This | |
7225 | function uses (or `calls') a function that invokes a function that | |
7226 | manipulates the kill ring. Thus, before reaching the mountains, we | |
7227 | climb the foothills. | |
7228 | ||
7229 | A subsequent chapter describes how text that is cut from the buffer is | |
7230 | retrieved. @xref{Yanking, , Yanking Text Back}. | |
7231 | ||
7232 | @node zap-to-char, kill-region, Storing Text, Cutting & Storing Text | |
7233 | @comment node-name, next, previous, up | |
7234 | @section @code{zap-to-char} | |
7235 | @findex zap-to-char | |
7236 | ||
7237 | The @code{zap-to-char} function barely changed between GNU Emacs | |
7238 | version 19 and GNU Emacs version 21. However, @code{zap-to-char} | |
7239 | calls another function, @code{kill-region}, which enjoyed a major rewrite | |
7240 | on the way to version 21. | |
7241 | ||
7242 | The @code{kill-region} function in Emacs 19 is complex, but does not | |
7243 | use code that is important at this time. We will skip it. | |
7244 | ||
7245 | The @code{kill-region} function in Emacs 21 is easier to read than the | |
7246 | same function in Emacs 19 and introduces a very important concept, | |
7247 | that of error handling. We will walk through the function. | |
7248 | ||
7249 | But first, let us look at the interactive @code{zap-to-char} function. | |
7250 | ||
7251 | @menu | |
7252 | * Complete zap-to-char:: The complete implementation. | |
7253 | * zap-to-char interactive:: A three part interactive expression. | |
7254 | * zap-to-char body:: A short overview. | |
7255 | * search-forward:: How to search for a string. | |
7256 | * progn:: The @code{progn} special form. | |
7257 | * Summing up zap-to-char:: Using @code{point} and @code{search-forward}. | |
7258 | @end menu | |
7259 | ||
7260 | @node Complete zap-to-char, zap-to-char interactive, zap-to-char, zap-to-char | |
7261 | @ifnottex | |
7262 | @unnumberedsubsec The Complete @code{zap-to-char} Implementation | |
7263 | @end ifnottex | |
7264 | ||
7265 | The GNU Emacs version 19 and version 21 implementations of the | |
7266 | @code{zap-to-char} function are nearly identical in form, and they | |
7267 | work alike. The function removes the text in the region between the | |
7268 | location of the cursor (i.e., of point) up to and including the next | |
7269 | occurrence of a specified character. The text that @code{zap-to-char} | |
7270 | removes is put in the kill ring; and it can be retrieved from the kill | |
7271 | ring by typing @kbd{C-y} (@code{yank}). If the command is given an | |
7272 | argument, it removes text through that number of occurrences. Thus, | |
7273 | if the cursor were at the beginning of this sentence and the character | |
7274 | were @samp{s}, @samp{Thus} would be removed. If the argument were | |
7275 | two, @samp{Thus, if the curs} would be removed, up to and including | |
7276 | the @samp{s} in @samp{cursor}. | |
7277 | ||
7278 | If the specified character is not found, @code{zap-to-char} will say | |
7279 | ``Search failed'', tell you the character you typed, and not remove | |
7280 | any text. | |
7281 | ||
7282 | In order to determine how much text to remove, @code{zap-to-char} uses | |
7283 | a search function. Searches are used extensively in code that | |
7284 | manipulates text, and we will focus attention on them as well as on the | |
7285 | deletion command. | |
7286 | ||
7287 | @need 800 | |
7288 | Here is the complete text of the version 19 implementation of the function: | |
7289 | ||
7290 | @c v 19 | |
7291 | @smallexample | |
7292 | @group | |
7293 | (defun zap-to-char (arg char) ; version 19 implementation | |
7294 | "Kill up to and including ARG'th occurrence of CHAR. | |
7295 | Goes backward if ARG is negative; error if CHAR not found." | |
7296 | (interactive "*p\ncZap to char: ") | |
7297 | (kill-region (point) | |
7298 | (progn | |
7299 | (search-forward | |
7300 | (char-to-string char) nil nil arg) | |
7301 | (point)))) | |
7302 | @end group | |
7303 | @end smallexample | |
7304 | ||
7305 | @node zap-to-char interactive, zap-to-char body, Complete zap-to-char, zap-to-char | |
7306 | @comment node-name, next, previous, up | |
7307 | @subsection The @code{interactive} Expression | |
7308 | ||
7309 | @need 800 | |
7310 | The interactive expression in the @code{zap-to-char} command looks like | |
7311 | this: | |
7312 | ||
7313 | @smallexample | |
7314 | (interactive "*p\ncZap to char: ") | |
7315 | @end smallexample | |
7316 | ||
7317 | The part within quotation marks, @code{"*p\ncZap to char:@: "}, specifies | |
7318 | three different things. First, and most simply, the asterisk, @samp{*}, | |
7319 | causes an error to be signalled if the buffer is read-only. This means that | |
7320 | if you try @code{zap-to-char} in a read-only buffer you will not be able to | |
7321 | remove text, and you will receive a message that says ``Buffer is | |
7322 | read-only''; your terminal may beep at you as well. | |
7323 | ||
7324 | The version 21 implementation does not have the asterisk, @samp{*}. The | |
7325 | function works the same as in version 19: in both cases, it cannot | |
7326 | remove text from a read-only buffer but the function does copy the | |
7327 | text that would have been removed to the kill ring. Also, in both | |
7328 | cases, you see an error message. | |
7329 | ||
7330 | However, the version 19 implementation copies text from a read-only | |
7331 | buffer only because of a mistake in the implementation of | |
7332 | @code{interactive}. According to the documentation for | |
7333 | @code{interactive}, the asterisk, @samp{*}, should prevent the | |
7334 | @code{zap-to-char} function from doing anything at all when the buffer | |
7335 | is read only. The function should not copy the text to the kill ring. | |
7336 | It is a bug that it does. | |
7337 | ||
7338 | In version 21, @code{interactive} is implemented correctly. So the | |
7339 | asterisk, @samp{*}, had to be removed from the interactive | |
7340 | specification. If you insert an @samp{*} and evaluate the function | |
7341 | definition, then the next time you run the @code{zap-to-char} function | |
7342 | on a read-only buffer, you will not copy any text. | |
7343 | ||
7344 | That change aside, and a change to the documentation, the two versions | |
7345 | of the @code{zap-to-char} function are identical. | |
7346 | ||
7347 | Let us continue with the interactive specification. | |
7348 | ||
7349 | The second part of @code{"*p\ncZap to char:@: "} is the @samp{p}. | |
7350 | This part is separated from the next part by a newline, @samp{\n}. | |
7351 | The @samp{p} means that the first argument to the function will be | |
7352 | passed the value of a `processed prefix'. The prefix argument is | |
7353 | passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If | |
7354 | the function is called interactively without a prefix, 1 is passed to | |
7355 | this argument. | |
7356 | ||
7357 | The third part of @code{"*p\ncZap to char:@: "} is @samp{cZap to char:@: | |
7358 | }. In this part, the lower case @samp{c} indicates that | |
7359 | @code{interactive} expects a prompt and that the argument will be a | |
7360 | character. The prompt follows the @samp{c} and is the string @samp{Zap | |
7361 | to char:@: } (with a space after the colon to make it look good). | |
7362 | ||
7363 | What all this does is prepare the arguments to @code{zap-to-char} so they | |
7364 | are of the right type, and give the user a prompt. | |
7365 | ||
7366 | @node zap-to-char body, search-forward, zap-to-char interactive, zap-to-char | |
7367 | @comment node-name, next, previous, up | |
7368 | @subsection The Body of @code{zap-to-char} | |
7369 | ||
7370 | The body of the @code{zap-to-char} function contains the code that | |
7371 | kills (that is, removes) the text in the region from the current | |
7372 | position of the cursor up to and including the specified character. | |
7373 | The first part of the code looks like this: | |
7374 | ||
7375 | @smallexample | |
7376 | (kill-region (point) @dots{} | |
7377 | @end smallexample | |
7378 | ||
7379 | @noindent | |
7380 | @code{(point)} is the current position of the cursor. | |
7381 | ||
7382 | The next part of the code is an expression using @code{progn}. The body | |
7383 | of the @code{progn} consists of calls to @code{search-forward} and | |
7384 | @code{point}. | |
7385 | ||
7386 | It is easier to understand how @code{progn} works after learning about | |
7387 | @code{search-forward}, so we will look at @code{search-forward} and | |
7388 | then at @code{progn}. | |
7389 | ||
7390 | @node search-forward, progn, zap-to-char body, zap-to-char | |
7391 | @comment node-name, next, previous, up | |
7392 | @subsection The @code{search-forward} Function | |
7393 | @findex search-forward | |
7394 | ||
7395 | The @code{search-forward} function is used to locate the | |
7396 | zapped-for-character in @code{zap-to-char}. If the search is | |
7397 | successful, @code{search-forward} leaves point immediately after the | |
7398 | last character in the target string. (In @code{zap-to-char}, the | |
7399 | target string is just one character long.) If the search is | |
7400 | backwards, @code{search-forward} leaves point just before the first | |
7401 | character in the target. Also, @code{search-forward} returns @code{t} | |
7402 | for true. (Moving point is therefore a `side effect'.) | |
7403 | ||
7404 | @need 1250 | |
7405 | In @code{zap-to-char}, the @code{search-forward} function looks like this: | |
7406 | ||
7407 | @smallexample | |
7408 | (search-forward (char-to-string char) nil nil arg) | |
7409 | @end smallexample | |
7410 | ||
7411 | The @code{search-forward} function takes four arguments: | |
7412 | ||
7413 | @enumerate | |
7414 | @item | |
7415 | The first argument is the target, what is searched for. This must be a | |
7416 | string, such as @samp{"z"}. | |
7417 | ||
7418 | As it happens, the argument passed to @code{zap-to-char} is a single | |
7419 | character. Because of the way computers are built, the Lisp | |
7420 | interpreter may treat a single character as being different from a | |
7421 | string of characters. Inside the computer, a single character has a | |
7422 | different electronic format than a string of one character. (A single | |
7423 | character can often be recorded in the computer using exactly one | |
7424 | byte; but a string may be longer, and the computer needs to be ready | |
7425 | for this.) Since the @code{search-forward} function searches for a | |
7426 | string, the character that the @code{zap-to-char} function receives as | |
7427 | its argument must be converted inside the computer from one format to | |
7428 | the other; otherwise the @code{search-forward} function will fail. | |
7429 | The @code{char-to-string} function is used to make this conversion. | |
7430 | ||
7431 | @item | |
7432 | The second argument bounds the search; it is specified as a position in | |
7433 | the buffer. In this case, the search can go to the end of the buffer, | |
7434 | so no bound is set and the second argument is @code{nil}. | |
7435 | ||
7436 | @item | |
7437 | The third argument tells the function what it should do if the search | |
7438 | fails---it can signal an error (and print a message) or it can return | |
7439 | @code{nil}. A @code{nil} as the third argument causes the function to | |
7440 | signal an error when the search fails. | |
7441 | ||
7442 | @item | |
7443 | The fourth argument to @code{search-forward} is the repeat count---how | |
7444 | many occurrences of the string to look for. This argument is optional | |
7445 | and if the function is called without a repeat count, this argument is | |
7446 | passed the value 1. If this argument is negative, the search goes | |
7447 | backwards. | |
7448 | @end enumerate | |
7449 | ||
7450 | @need 800 | |
7451 | In template form, a @code{search-forward} expression looks like this: | |
7452 | ||
7453 | @smallexample | |
7454 | @group | |
7455 | (search-forward "@var{target-string}" | |
7456 | @var{limit-of-search} | |
7457 | @var{what-to-do-if-search-fails} | |
7458 | @var{repeat-count}) | |
7459 | @end group | |
7460 | @end smallexample | |
7461 | ||
7462 | We will look at @code{progn} next. | |
7463 | ||
7464 | @node progn, Summing up zap-to-char, search-forward, zap-to-char | |
7465 | @comment node-name, next, previous, up | |
7466 | @subsection The @code{progn} Special Form | |
7467 | @findex progn | |
7468 | ||
7469 | @code{progn} is a special form that causes each of its arguments to be | |
7470 | evaluated in sequence and then returns the value of the last one. The | |
7471 | preceding expressions are evaluated only for the side effects they | |
7472 | perform. The values produced by them are discarded. | |
7473 | ||
7474 | @need 800 | |
7475 | The template for a @code{progn} expression is very simple: | |
7476 | ||
7477 | @smallexample | |
7478 | @group | |
7479 | (progn | |
7480 | @var{body}@dots{}) | |
7481 | @end group | |
7482 | @end smallexample | |
7483 | ||
7484 | In @code{zap-to-char}, the @code{progn} expression has to do two things: | |
7485 | put point in exactly the right position; and return the location of | |
7486 | point so that @code{kill-region} will know how far to kill to. | |
7487 | ||
7488 | The first argument to the @code{progn} is @code{search-forward}. When | |
7489 | @code{search-forward} finds the string, the function leaves point | |
7490 | immediately after the last character in the target string. (In this | |
7491 | case the target string is just one character long.) If the search is | |
7492 | backwards, @code{search-forward} leaves point just before the first | |
7493 | character in the target. The movement of point is a side effect. | |
7494 | ||
7495 | The second and last argument to @code{progn} is the expression | |
7496 | @code{(point)}. This expression returns the value of point, which in | |
7497 | this case will be the location to which it has been moved by | |
7498 | @code{search-forward}. This value is returned by the @code{progn} | |
7499 | expression and is passed to @code{kill-region} as @code{kill-region}'s | |
7500 | second argument. | |
7501 | ||
7502 | @node Summing up zap-to-char, , progn, zap-to-char | |
7503 | @comment node-name, next, previous, up | |
7504 | @subsection Summing up @code{zap-to-char} | |
7505 | ||
7506 | Now that we have seen how @code{search-forward} and @code{progn} work, | |
7507 | we can see how the @code{zap-to-char} function works as a whole. | |
7508 | ||
7509 | The first argument to @code{kill-region} is the position of the cursor | |
7510 | when the @code{zap-to-char} command is given---the value of point at | |
7511 | that time. Within the @code{progn}, the search function then moves | |
7512 | point to just after the zapped-to-character and @code{point} returns the | |
7513 | value of this location. The @code{kill-region} function puts together | |
7514 | these two values of point, the first one as the beginning of the region | |
7515 | and the second one as the end of the region, and removes the region. | |
7516 | ||
7517 | The @code{progn} special form is necessary because the @code{kill-region} | |
7518 | command takes two arguments; and it would fail if @code{search-forward} | |
7519 | and @code{point} expressions were written in sequence as two | |
7520 | additional arguments. The @code{progn} expression is a single argument | |
7521 | to @code{kill-region} and returns the one value that @code{kill-region} | |
7522 | needs for its second argument. | |
7523 | ||
7524 | @node kill-region, Digression into C, zap-to-char, Cutting & Storing Text | |
7525 | @comment node-name, next, previous, up | |
7526 | @section @code{kill-region} | |
7527 | @findex kill-region | |
7528 | ||
7529 | The @code{zap-to-char} function uses the @code{kill-region} function. | |
7530 | This function clips text from a region and copies that text to | |
7531 | the kill ring, from which it may be retrieved. | |
7532 | ||
7533 | The Emacs 21 version of that function uses @code{condition-case} and | |
7534 | @code{copy-region-as-kill}, both of which we will explain. | |
7535 | @code{condition-case} is an important special form. | |
7536 | ||
7537 | In essence, the @code{kill-region} function calls | |
7538 | @code{condition-case}, which takes three arguments. In this function, | |
7539 | the first argument does nothing. The second argument contains the | |
7540 | code that does the work when all goes well. The third argument | |
7541 | contains the code that is called in the event of an error. | |
7542 | ||
7543 | @menu | |
7544 | * Complete kill-region:: The function definition. | |
7545 | * condition-case:: Dealing with a problem. | |
7546 | * delete-and-extract-region:: Doing the work. | |
7547 | @end menu | |
7548 | ||
7549 | @node Complete kill-region, condition-case, kill-region, kill-region | |
7550 | @ifnottex | |
7551 | @unnumberedsubsec The Complete @code{kill-region} Definition | |
7552 | @end ifnottex | |
7553 | ||
7554 | @need 1200 | |
7555 | We will go through the @code{condition-case} code in a moment. First, | |
7556 | let us look at the complete definition of @code{kill-region}, with | |
7557 | comments added: | |
7558 | ||
7559 | @c v 21 | |
7560 | @smallexample | |
7561 | @group | |
7562 | (defun kill-region (beg end) | |
7563 | "Kill between point and mark. | |
7564 | The text is deleted but saved in the kill ring." | |
7565 | (interactive "r") | |
7566 | @end group | |
7567 | ||
7568 | @group | |
7569 | ;; 1. `condition-case' takes three arguments. | |
7570 | ;; If the first argument is nil, as it is here, | |
7571 | ;; information about the error signal is not | |
7572 | ;; stored for use by another function. | |
7573 | (condition-case nil | |
7574 | @end group | |
7575 | ||
7576 | @group | |
7577 | ;; 2. The second argument to `condition-case' | |
7578 | ;; tells the Lisp interpreter what to do when all goes well. | |
7579 | @end group | |
7580 | ||
7581 | @group | |
7582 | ;; The `delete-and-extract-region' function usually does the | |
7583 | ;; work. If the beginning and ending of the region are both | |
7584 | ;; the same, then the variable `string' will be empty, or nil | |
7585 | (let ((string (delete-and-extract-region beg end))) | |
7586 | @end group | |
7587 | ||
7588 | @group | |
7589 | ;; `when' is an `if' clause that cannot take an `else-part'. | |
7590 | ;; Emacs normally sets the value of `last-command' to the | |
7591 | ;; previous command. | |
7592 | @end group | |
7593 | @group | |
7594 | ;; `kill-append' concatenates the new string and the old. | |
7595 | ;; `kill-new' inserts text into a new item in the kill ring. | |
7596 | (when string | |
7597 | (if (eq last-command 'kill-region) | |
7598 | ;; if true, prepend string | |
7599 | (kill-append string (< end beg)) | |
7600 | (kill-new string))) | |
7601 | (setq this-command 'kill-region)) | |
7602 | @end group | |
7603 | ||
7604 | @group | |
7605 | ;; 3. The third argument to `condition-case' tells the interpreter | |
7606 | ;; what to do with an error. | |
7607 | @end group | |
7608 | @group | |
7609 | ;; The third argument has a conditions part and a body part. | |
7610 | ;; If the conditions are met (in this case, | |
7611 | ;; if text or buffer is read-only) | |
7612 | ;; then the body is executed. | |
7613 | @end group | |
7614 | @group | |
7615 | ((buffer-read-only text-read-only) ;; this is the if-part | |
7616 | ;; then... | |
7617 | (copy-region-as-kill beg end) | |
7618 | @end group | |
7619 | @group | |
7620 | (if kill-read-only-ok ;; usually this variable is nil | |
7621 | (message "Read only text copied to kill ring") | |
7622 | ;; or else, signal an error if the buffer is read-only; | |
7623 | (barf-if-buffer-read-only) | |
7624 | ;; and, in any case, signal that the text is read-only. | |
7625 | (signal 'text-read-only (list (current-buffer))))))) | |
7626 | @end group | |
7627 | @end smallexample | |
7628 | ||
7629 | @node condition-case, delete-and-extract-region, Complete kill-region, kill-region | |
7630 | @comment node-name, next, previous, up | |
7631 | @subsection @code{condition-case} | |
7632 | @findex condition-case | |
7633 | ||
7634 | As we have seen earlier (@pxref{Making Errors, , Generate an Error | |
7635 | Message}), when the Emacs Lisp interpreter has trouble evaluating an | |
7636 | expression, it provides you with help; in the jargon, this is called | |
7637 | ``signaling an error''. Usually, the computer stops the program and | |
7638 | shows you a message. | |
7639 | ||
7640 | However, some programs undertake complicated actions. They should not | |
7641 | simply stop on an error. In the @code{kill-region} function, the most | |
7642 | likely error is that you will try to kill text that is read-only and | |
7643 | cannot be removed. So the @code{kill-region} function contains code | |
7644 | to handle this circumstance. This code, which makes up the body of | |
7645 | the @code{kill-region} function, is inside of a @code{condition-case} | |
7646 | special form. | |
7647 | ||
7648 | @need 800 | |
7649 | The template for @code{condition-case} looks like this: | |
7650 | ||
7651 | @smallexample | |
7652 | @group | |
7653 | (condition-case | |
7654 | @var{var} | |
7655 | @var{bodyform} | |
7656 | @var{error-handler}@dots{}) | |
7657 | @end group | |
7658 | @end smallexample | |
7659 | ||
7660 | The second argument, @var{bodyform}, is straightforward. The | |
7661 | @code{condition-case} special form causes the Lisp interpreter to | |
7662 | evaluate the code in @var{bodyform}. If no error occurs, the special | |
7663 | form returns the code's value and produces the side-effects, if any. | |
7664 | ||
7665 | In short, the @var{bodyform} part of a @code{condition-case} | |
7666 | expression determines what should happen when everything works | |
7667 | correctly. | |
7668 | ||
7669 | However, if an error occurs, among its other actions, the function | |
7670 | generating the error signal will define one or more error condition | |
7671 | names. | |
7672 | ||
7673 | An error handler is the third argument to @code{condition case}. | |
7674 | An error handler has two parts, a @var{condition-name} and a | |
7675 | @var{body}. If the @var{condition-name} part of an error handler | |
7676 | matches a condition name generated by an error, then the @var{body} | |
7677 | part of the error handler is run. | |
7678 | ||
7679 | As you will expect, the @var{condition-name} part of an error handler | |
7680 | may be either a single condition name or a list of condition names. | |
7681 | ||
7682 | Also, a complete @code{condition-case} expression may contain more | |
7683 | than one error handler. When an error occurs, the first applicable | |
7684 | handler is run. | |
7685 | ||
7686 | Lastly, the first argument to the @code{condition-case} expression, | |
7687 | the @var{var} argument, is sometimes bound to a variable that | |
7688 | contains information about the error. However, if that argument is | |
7689 | nil, as is the case in @code{kill-region}, that information is | |
7690 | discarded. | |
7691 | ||
7692 | @need 1200 | |
7693 | In brief, in the @code{kill-region} function, the code | |
7694 | @code{condition-case} works like this: | |
7695 | ||
7696 | @smallexample | |
7697 | @group | |
7698 | @var{If no errors}, @var{run only this code} | |
7699 | @var{but}, @var{if errors}, @var{run this other code}. | |
7700 | @end group | |
7701 | @end smallexample | |
7702 | ||
7703 | @node delete-and-extract-region, , condition-case, kill-region | |
7704 | @comment node-name, next, previous, up | |
7705 | @subsection @code{delete-and-extract-region} | |
7706 | @findex delete-and-extract-region | |
7707 | ||
7708 | A @code{condition-case} expression has two parts, a part that is | |
7709 | evaluated in the expectation that all will go well, but which may | |
7710 | generate an error; and a part that is evaluated when there is an | |
7711 | error. | |
7712 | ||
7713 | First, let us look at the code in @code{kill-region} that is run in | |
7714 | the expectation that all goes well. This is the core of the function. | |
7715 | The code looks like this: | |
7716 | ||
7717 | @smallexample | |
7718 | @group | |
7719 | (let ((string (delete-and-extract-region beg end))) | |
7720 | (when string | |
7721 | (if (eq last-command 'kill-region) | |
7722 | (kill-append string (< end beg)) | |
7723 | (kill-new string))) | |
7724 | (setq this-command 'kill-region)) | |
7725 | @end group | |
7726 | @end smallexample | |
7727 | ||
7728 | It looks complicated because we have the new functions | |
7729 | @code{delete-and-extract-region}, @code{kill-append}, and | |
7730 | @code{kill-new} as well as the new variables, | |
7731 | @code{last-command} and @code{this-command}. | |
7732 | ||
7733 | The @code{delete-and-extract-region} function is straightforward. It | |
7734 | is a built-in function that deletes the text in a region (a side | |
7735 | effect) and also returns that text. This is the function that | |
7736 | actually removes the text. (And if it cannot do that, it signals the | |
7737 | error.) | |
7738 | ||
7739 | In this @code{let} expression, the text that | |
7740 | @code{delete-and-extract-region} returns is placed in the local | |
7741 | variable called @samp{string}. This is the text that is removed from | |
7742 | the buffer. (To be more precise, the variable is set to point to the | |
7743 | address of the extracted text; to say it is `placed in' the variable | |
7744 | is simply a shorthand.) | |
7745 | ||
7746 | If the variable @samp{string} does point to text, that text is added | |
7747 | to the kill ring. The variable will have a @code{nil} value if no | |
7748 | text was removed. | |
7749 | ||
7750 | The code uses @code{when} to determine whether the variable | |
7751 | @samp{string} points to text. A @code{when} statement is simply a | |
7752 | programmers' convenience. A @code{when} statement is an @code{if} | |
7753 | statement without the possibility of an else clause. In your mind, you | |
7754 | can replace @code{when} with @code{if} and understand what goes on. | |
7755 | That is what the Lisp interpreter does. | |
7756 | ||
7757 | @cindex Macro, lisp | |
7758 | @cindex Lisp macro | |
7759 | Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro} | |
7760 | enables you to define new control constructs and other language | |
7761 | features. It tells the interpreter how to compute another Lisp | |
7762 | expression which will in turn compute the value. In this case, the | |
7763 | `other expression' is an @code{if} expression. For more about Lisp | |
7764 | macros, see @ref{Macros, , Macros, elisp, The GNU Emacs Lisp Reference | |
7765 | Manual}. The C programming language also provides macros. These are | |
7766 | different, but also useful. We will briefly look at C macros in | |
7767 | @ref{Digression into C, , @code{delete-and-extract-region}: | |
7768 | Digressing into C}. | |
7769 | ||
7770 | @need 1200 | |
7771 | If the string has content, then another conditional expression is | |
7772 | executed. This is an @code{if} with both a then-part and an else-part. | |
7773 | ||
7774 | @smallexample | |
7775 | @group | |
7776 | (if (eq last-command 'kill-region) | |
7777 | (kill-append string (< end beg)) | |
7778 | (kill-new string))) | |
7779 | @end group | |
7780 | @end smallexample | |
7781 | ||
7782 | The then-part is evaluated if the previous command was another call to | |
7783 | @code{kill-region}; if not, the else-part is evaluated. | |
7784 | ||
7785 | @code{last-command} is a variable that comes with Emacs that we have | |
7786 | not seen before. Normally, whenever a function is executed, Emacs | |
7787 | sets the value of @code{last-command} to the previous command. | |
7788 | ||
7789 | @need 1200 | |
7790 | In this segment of the definition, the @code{if} expression checks | |
7791 | whether the previous command was @code{kill-region}. If it was, | |
7792 | ||
7793 | @smallexample | |
7794 | (kill-append string (< end beg)) | |
7795 | @end smallexample | |
7796 | ||
7797 | @noindent | |
7798 | concatenates a copy of the newly clipped text to the just previously | |
7799 | clipped text in the kill ring. (If the @w{@code{(< end beg))}} | |
7800 | expression is true, @code{kill-append} prepends the string to the just | |
7801 | previously clipped text. For a detailed discussion, see | |
7802 | @ref{kill-append function, , The @code{kill-append} function}.) | |
7803 | ||
7804 | If you then yank back the text, i.e., `paste' it, you get both | |
7805 | pieces of text at once. That way, if you delete two words in a row, | |
7806 | and then yank them back, you get both words, in their proper order, | |
7807 | with one yank. (The @w{@code{(< end beg))}} expression makes sure the | |
7808 | order is correct.) | |
7809 | ||
7810 | On the other hand, if the previous command is not @code{kill-region}, | |
7811 | then the @code{kill-new} function is called, which adds the text to | |
7812 | the kill ring as the latest item, and sets the | |
7813 | @code{kill-ring-yank-pointer} variable to point to it. | |
7814 | ||
7815 | @node Digression into C, defvar, kill-region, Cutting & Storing Text | |
7816 | @comment node-name, next, previous, up | |
7817 | @section @code{delete-and-extract-region}: Digressing into C | |
7818 | @findex delete-and-extract-region | |
7819 | @cindex C, a digression into | |
7820 | @cindex Digression into C | |
7821 | ||
7822 | The @code{zap-to-char} command uses the | |
7823 | @code{delete-and-extract-region} function, which in turn uses two | |
7824 | other functions, @code{copy-region-as-kill} and | |
7825 | @code{del_range_1}. The @code{copy-region-as-kill} function will be | |
7826 | described in a following section; it puts a copy of the region in the | |
7827 | kill ring so it can be yanked back. (@xref{copy-region-as-kill, , | |
7828 | @code{copy-region-as-kill}}.) | |
7829 | ||
7830 | The @code{delete-and-extract-region} function removes the contents of | |
7831 | a region and you cannot get them back. | |
7832 | ||
7833 | Unlike the other code discussed here, @code{delete-and-extract-region} | |
7834 | is not written in Emacs Lisp; it is written in C and is one of the | |
7835 | primitives of the GNU Emacs system. Since it is very simple, I will | |
7836 | digress briefly from Lisp and describe it here. | |
7837 | ||
7838 | @need 1500 | |
7839 | Like many of the other Emacs primitives, | |
7840 | @code{delete-and-extract-region} is written as an instance of a C | |
7841 | macro, a macro being a template for code. The complete macro looks | |
7842 | like this: | |
7843 | ||
7844 | @c /usr/local/src/emacs/src/editfns.c | |
7845 | @smallexample | |
7846 | @group | |
7847 | DEFUN ("delete-and-extract-region", Fdelete_and_extract_region, | |
7848 | Sdelete_and_extract_region, 2, 2, 0, | |
7849 | "Delete the text between START and END and return it.") | |
7850 | (start, end) | |
7851 | Lisp_Object start, end; | |
7852 | @{ | |
7853 | validate_region (&start, &end); | |
7854 | return del_range_1 (XINT (start), XINT (end), 1, 1); | |
7855 | @} | |
7856 | @end group | |
7857 | @end smallexample | |
7858 | ||
7859 | Without going into the details of the macro writing process, let me | |
7860 | point out that this macro starts with the word @code{DEFUN}. The word | |
7861 | @code{DEFUN} was chosen since the code serves the same purpose as | |
7862 | @code{defun} does in Lisp. The word @code{DEFUN} is followed by seven | |
7863 | parts inside of parentheses: | |
7864 | ||
7865 | @itemize @bullet | |
7866 | @item | |
7867 | The first part is the name given to the function in Lisp, | |
7868 | @code{delete-and-extract-region}. | |
7869 | ||
7870 | @item | |
7871 | The second part is the name of the function in C, | |
7872 | @code{Fdelete_and_extract_region}. By convention, it starts with | |
7873 | @samp{F}. Since C does not use hyphens in names, underscores are used | |
7874 | instead. | |
7875 | ||
7876 | @item | |
7877 | The third part is the name for the C constant structure that records | |
7878 | information on this function for internal use. It is the name of the | |
7879 | function in C but begins with an @samp{S} instead of an @samp{F}. | |
7880 | ||
7881 | @item | |
7882 | The fourth and fifth parts specify the minimum and maximum number of | |
7883 | arguments the function can have. This function demands exactly 2 | |
7884 | arguments. | |
7885 | ||
7886 | @item | |
7887 | The sixth part is nearly like the argument that follows the | |
7888 | @code{interactive} declaration in a function written in Lisp: a letter | |
7889 | followed, perhaps, by a prompt. The only difference from the Lisp is | |
7890 | when the macro is called with no arguments. Then you write a @code{0} | |
7891 | (which is a `null string'), as in this macro. | |
7892 | ||
7893 | If you were to specify arguments, you would place them between | |
7894 | quotation marks. The C macro for @code{goto-char} includes | |
7895 | @code{"NGoto char: "} in this position to indicate that the function | |
7896 | expects a raw prefix, in this case, a numerical location in a buffer, | |
7897 | and provides a prompt. | |
7898 | ||
7899 | @item | |
7900 | The seventh part is a documentation string, just like the one for a | |
7901 | function written in Emacs Lisp, except that every newline must be | |
7902 | written explicitly as @samp{\n} followed by a backslash and carriage | |
7903 | return. | |
7904 | ||
7905 | @need 1000 | |
7906 | Thus, the first two lines of documentation for @code{goto-char} are | |
7907 | written like this: | |
7908 | ||
7909 | @smallexample | |
7910 | @group | |
7911 | "Set point to POSITION, a number or marker.\n\ | |
7912 | Beginning of buffer is position (point-min), end is (point-max). | |
7913 | @end group | |
7914 | @end smallexample | |
7915 | @end itemize | |
7916 | ||
7917 | @need 1200 | |
7918 | In a C macro, the formal parameters come next, with a statement of | |
7919 | what kind of object they are, followed by what might be called the `body' | |
7920 | of the macro. For @code{delete-and-extract-region} the `body' | |
7921 | consists of the following two lines: | |
7922 | ||
7923 | @smallexample | |
7924 | @group | |
7925 | validate_region (&start, &end); | |
7926 | return del_range_1 (XINT (start), XINT (end), 1, 1); | |
7927 | @end group | |
7928 | @end smallexample | |
7929 | ||
7930 | The first function, @code{validate_region} checks whether the values | |
7931 | passed as the beginning and end of the region are the proper type and | |
7932 | are within range. The second function, @code{del_range_1}, actually | |
7933 | deletes the text. | |
7934 | ||
7935 | @code{del_range_1} is a complex function we will not look into. It | |
7936 | updates the buffer and does other things. | |
7937 | ||
7938 | However, it is worth looking at the two arguments passed to | |
7939 | @code{del_range}. These are @w{@code{XINT (start)}} and @w{@code{XINT | |
7940 | (end)}}. | |
7941 | ||
7942 | As far as the C language is concerned, @code{start} and @code{end} are | |
7943 | two integers that mark the beginning and end of the region to be | |
7944 | deleted@footnote{More precisely, and requiring more expert knowledge | |
7945 | to understand, the two integers are of type `Lisp_Object', which can | |
7946 | also be a C union instead of an integer type.}. | |
7947 | ||
7948 | In early versions of Emacs, these two numbers were thirty-two bits | |
7949 | long, but the code is slowly being generalized to handle other | |
7950 | lengths. Three of the available bits are used to specify the type of | |
7951 | information and a fourth bit is used for handling the computer's | |
7952 | memory; the remaining bits are used as `content'. | |
7953 | ||
7954 | @samp{XINT} is a C macro that extracts the relevant number from the | |
7955 | longer collection of bits; the four other bits are discarded. | |
7956 | ||
7957 | @need 800 | |
7958 | The command in @code{delete-and-extract-region} looks like this: | |
7959 | ||
7960 | @smallexample | |
7961 | del_range_1 (XINT (start), XINT (end), 1, 1); | |
7962 | @end smallexample | |
7963 | ||
7964 | @noindent | |
7965 | It deletes the region between the beginning position, @code{start}, | |
7966 | and the ending position, @code{end}. | |
7967 | ||
7968 | From the point of view of the person writing Lisp, Emacs is all very | |
7969 | simple; but hidden underneath is a great deal of complexity to make it | |
7970 | all work. | |
7971 | ||
7972 | @node defvar, copy-region-as-kill, Digression into C, Cutting & Storing Text | |
7973 | @comment node-name, next, previous, up | |
7974 | @section Initializing a Variable with @code{defvar} | |
7975 | @findex defvar | |
7976 | @cindex Initializing a variable | |
7977 | @cindex Variable initialization | |
7978 | ||
7979 | Unlike the @code{delete-and-extract-region} function, the | |
7980 | @code{copy-region-as-kill} function is written in Emacs Lisp. Two | |
7981 | functions within it, @code{kill-append} and @code{kill-new}, copy a | |
7982 | region in a buffer and save it in a variable called the | |
7983 | @code{kill-ring}. This section describes how the @code{kill-ring} | |
7984 | variable is created and initialized using the @code{defvar} special | |
7985 | form. | |
7986 | ||
7987 | (Again we note that the term @code{kill-ring} is a misnomer. The text | |
7988 | that is clipped out of the buffer can be brought back; it is not a ring | |
7989 | of corpses, but a ring of resurrectable text.) | |
7990 | ||
7991 | In Emacs Lisp, a variable such as the @code{kill-ring} is created and | |
7992 | given an initial value by using the @code{defvar} special form. The | |
7993 | name comes from ``define variable''. | |
7994 | ||
7995 | The @code{defvar} special form is similar to @code{setq} in that it sets | |
7996 | the value of a variable. It is unlike @code{setq} in two ways: first, | |
7997 | it only sets the value of the variable if the variable does not already | |
7998 | have a value. If the variable already has a value, @code{defvar} does | |
7999 | not override the existing value. Second, @code{defvar} has a | |
8000 | documentation string. | |
8001 | ||
8002 | (Another special form, @code{defcustom}, is designed for variables | |
8003 | that people customize. It has more features than @code{defvar}. | |
8004 | (@xref{defcustom, , Setting Variables with @code{defcustom}}.) | |
8005 | ||
8006 | @menu | |
8007 | * See variable current value:: | |
8008 | * defvar and asterisk:: An old-time convention. | |
8009 | @end menu | |
8010 | ||
8011 | @node See variable current value, defvar and asterisk, defvar, defvar | |
8012 | @ifnottex | |
8013 | @unnumberedsubsec Seeing the Current Value of a Variable | |
8014 | @end ifnottex | |
8015 | ||
8016 | You can see the current value of a variable, any variable, by using | |
8017 | the @code{describe-variable} function, which is usually invoked by | |
8018 | typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring} | |
8019 | (followed by @key{RET}) when prompted, you will see what is in your | |
8020 | current kill ring---this may be quite a lot! Conversely, if you have | |
8021 | been doing nothing this Emacs session except read this document, you | |
8022 | may have nothing in it. Also, you will see the documentation for | |
8023 | @code{kill-ring}: | |
8024 | ||
8025 | @smallexample | |
8026 | @group | |
8027 | Documentation: | |
8028 | List of killed text sequences. | |
8029 | Since the kill ring is supposed to interact nicely with cut-and-paste | |
8030 | facilities offered by window systems, use of this variable should | |
8031 | @end group | |
8032 | @group | |
8033 | interact nicely with `interprogram-cut-function' and | |
8034 | `interprogram-paste-function'. The functions `kill-new', | |
8035 | `kill-append', and `current-kill' are supposed to implement this | |
8036 | interaction; you may want to use them instead of manipulating the kill | |
8037 | ring directly. | |
8038 | @end group | |
8039 | @end smallexample | |
8040 | ||
8041 | @need 800 | |
8042 | The kill ring is defined by a @code{defvar} in the following way: | |
8043 | ||
8044 | @smallexample | |
8045 | @group | |
8046 | (defvar kill-ring nil | |
8047 | "List of killed text sequences. | |
8048 | @dots{}") | |
8049 | @end group | |
8050 | @end smallexample | |
8051 | ||
8052 | @noindent | |
8053 | In this variable definition, the variable is given an initial value of | |
8054 | @code{nil}, which makes sense, since if you have saved nothing, you want | |
8055 | nothing back if you give a @code{yank} command. The documentation | |
8056 | string is written just like the documentation string of a @code{defun}. | |
8057 | As with the documentation string of the @code{defun}, the first line of | |
8058 | the documentation should be a complete sentence, since some commands, | |
8059 | like @code{apropos}, print only the first line of documentation. | |
8060 | Succeeding lines should not be indented; otherwise they look odd when | |
8061 | you use @kbd{C-h v} (@code{describe-variable}). | |
8062 | ||
8063 | @node defvar and asterisk, , See variable current value, defvar | |
8064 | @subsection @code{defvar} and an asterisk | |
8065 | @findex defvar @r{for a user customizable variable} | |
8066 | @findex defvar @r{with an asterisk} | |
8067 | ||
8068 | In the past, Emacs used the @code{defvar} special form both for | |
8069 | internal variables that you would not expect a user to change and for | |
8070 | variables that you do expect a user to change. Although you can still | |
8071 | use @code{defvar} for user customizable variables, please use | |
8072 | @code{defcustom} instead, since that special form provides a path into | |
8073 | the Customization commands. (@xref{defcustom, , Setting Variables | |
8074 | with @code{defcustom}}.) | |
8075 | ||
8076 | When you specified a variable using the @code{defvar} special form, | |
8077 | you could distinguish a readily settable variable from others by | |
8078 | typing an asterisk, @samp{*}, in the first column of its documentation | |
8079 | string. For example: | |
8080 | ||
8081 | @smallexample | |
8082 | @group | |
8083 | (defvar shell-command-default-error-buffer nil | |
8084 | "*Buffer name for `shell-command' @dots{} error output. | |
8085 | @dots{} ") | |
8086 | @end group | |
8087 | @end smallexample | |
8088 | ||
8089 | @noindent | |
8090 | This means that you could (and still can) use the @code{edit-options} | |
8091 | command to change the value of | |
8092 | @code{shell-command-default-error-buffer} temporarily. | |
8093 | ||
8094 | @findex edit-options | |
8095 | However, options set using @code{edit-options} are set only for the | |
8096 | duration of your editing session. The new values are not saved | |
8097 | between sessions. Each time Emacs starts, it reads the original | |
8098 | value, unless you change the value within your @file{.emacs} file, | |
8099 | either by setting it manually or by using @code{customize}. | |
8100 | @xref{Emacs Initialization, , Your @file{.emacs} File}. | |
8101 | ||
8102 | For me, the major use of the @code{edit-options} command is to suggest | |
8103 | variables that I might want to set in my @file{.emacs} file. I urge | |
8104 | you to look through the list. (@xref{Edit Options, , Editing Variable | |
8105 | Values, emacs, The GNU Emacs Manual}.) | |
8106 | ||
8107 | @node copy-region-as-kill, cons & search-fwd Review, defvar, Cutting & Storing Text | |
8108 | @comment node-name, next, previous, up | |
8109 | @section @code{copy-region-as-kill} | |
8110 | @findex copy-region-as-kill | |
8111 | @findex nthcdr | |
8112 | ||
8113 | The @code{copy-region-as-kill} function copies a region of text from a | |
8114 | buffer and (via either @code{kill-append} or @code{kill-new}) saves it | |
8115 | in the @code{kill-ring}. | |
8116 | ||
8117 | If you call @code{copy-region-as-kill} immediately after a | |
8118 | @code{kill-region} command, Emacs appends the newly copied text to the | |
8119 | previously copied text. This means that if you yank back the text, you | |
8120 | get it all, from both this and the previous operation. On the other | |
8121 | hand, if some other command precedes the @code{copy-region-as-kill}, | |
8122 | the function copies the text into a separate entry in the kill ring. | |
8123 | ||
8124 | @menu | |
8125 | * Complete copy-region-as-kill:: The complete function definition. | |
8126 | * copy-region-as-kill body:: The body of @code{copy-region-as-kill}. | |
8127 | @end menu | |
8128 | ||
8129 | @node Complete copy-region-as-kill, copy-region-as-kill body, copy-region-as-kill, copy-region-as-kill | |
8130 | @ifnottex | |
8131 | @unnumberedsubsec The complete @code{copy-region-as-kill} function definition | |
8132 | @end ifnottex | |
8133 | ||
8134 | @need 1200 | |
8135 | Here is the complete text of the version 21 @code{copy-region-as-kill} | |
8136 | function: | |
8137 | ||
8138 | @smallexample | |
8139 | @group | |
8140 | (defun copy-region-as-kill (beg end) | |
8141 | "Save the region as if killed, but don't kill it. | |
8142 | In Transient Mark mode, deactivate the mark. | |
8143 | If `interprogram-cut-function' is non-nil, also save | |
8144 | the text for a window system cut and paste." | |
8145 | (interactive "r") | |
8146 | @end group | |
8147 | @group | |
8148 | (if (eq last-command 'kill-region) | |
8149 | (kill-append (buffer-substring beg end) (< end beg)) | |
8150 | (kill-new (buffer-substring beg end))) | |
8151 | @end group | |
8152 | @group | |
8153 | (if transient-mark-mode | |
8154 | (setq deactivate-mark t)) | |
8155 | nil) | |
8156 | @end group | |
8157 | @end smallexample | |
8158 | ||
8159 | @need 800 | |
8160 | As usual, this function can be divided into its component parts: | |
8161 | ||
8162 | @smallexample | |
8163 | @group | |
8164 | (defun copy-region-as-kill (@var{argument-list}) | |
8165 | "@var{documentation}@dots{}" | |
8166 | (interactive "r") | |
8167 | @var{body}@dots{}) | |
8168 | @end group | |
8169 | @end smallexample | |
8170 | ||
8171 | The arguments are @code{beg} and @code{end} and the function is | |
8172 | interactive with @code{"r"}, so the two arguments must refer to the | |
8173 | beginning and end of the region. If you have been reading though this | |
8174 | document from the beginning, understanding these parts of a function is | |
8175 | almost becoming routine. | |
8176 | ||
8177 | The documentation is somewhat confusing unless you remember that the | |
8178 | word `kill' has a meaning different from its usual meaning. The | |
8179 | `Transient Mark' and @code{interprogram-cut-function} comments explain | |
8180 | certain side-effects. | |
8181 | ||
8182 | After you once set a mark, a buffer always contains a region. If you | |
8183 | wish, you can use Transient Mark mode to highlight the region | |
8184 | temporarily. (No one wants to highlight the region all the time, so | |
8185 | Transient Mark mode highlights it only at appropriate times. Many | |
8186 | people turn off Transient Mark mode, so the region is never | |
8187 | highlighted.) | |
8188 | ||
8189 | Also, a windowing system allows you to copy, cut, and paste among | |
8190 | different programs. In the X windowing system, for example, the | |
8191 | @code{interprogram-cut-function} function is @code{x-select-text}, | |
8192 | which works with the windowing system's equivalent of the Emacs kill | |
8193 | ring. | |
8194 | ||
8195 | The body of the @code{copy-region-as-kill} function starts with an | |
8196 | @code{if} clause. What this clause does is distinguish between two | |
8197 | different situations: whether or not this command is executed | |
8198 | immediately after a previous @code{kill-region} command. In the first | |
8199 | case, the new region is appended to the previously copied text. | |
8200 | Otherwise, it is inserted into the beginning of the kill ring as a | |
8201 | separate piece of text from the previous piece. | |
8202 | ||
8203 | The last two lines of the function prevent the region from lighting up | |
8204 | if Transient Mark mode is turned on. | |
8205 | ||
8206 | The body of @code{copy-region-as-kill} merits discussion in detail. | |
8207 | ||
8208 | @node copy-region-as-kill body, , Complete copy-region-as-kill, copy-region-as-kill | |
8209 | @comment node-name, next, previous, up | |
8210 | @subsection The Body of @code{copy-region-as-kill} | |
8211 | ||
8212 | The @code{copy-region-as-kill} function works in much the same way as | |
8213 | the @code{kill-region} function (@pxref{kill-region, | |
8214 | ,@code{kill-region}}). Both are written so that two or more kills in | |
8215 | a row combine their text into a single entry. If you yank back the | |
8216 | text from the kill ring, you get it all in one piece. Moreover, kills | |
8217 | that kill forward from the current position of the cursor are added to | |
8218 | the end of the previously copied text and commands that copy text | |
8219 | backwards add it to the beginning of the previously copied text. This | |
8220 | way, the words in the text stay in the proper order. | |
8221 | ||
8222 | Like @code{kill-region}, the @code{copy-region-as-kill} function makes | |
8223 | use of the @code{last-command} variable that keeps track of the | |
8224 | previous Emacs command. | |
8225 | ||
8226 | @menu | |
8227 | * last-command & this-command:: | |
8228 | * kill-append function:: | |
8229 | * kill-new function:: | |
8230 | @end menu | |
8231 | ||
8232 | @node last-command & this-command, kill-append function, copy-region-as-kill body, copy-region-as-kill body | |
8233 | @ifnottex | |
8234 | @unnumberedsubsubsec @code{last-command} and @code{this-command} | |
8235 | @end ifnottex | |
8236 | ||
8237 | Normally, whenever a function is executed, Emacs sets the value of | |
8238 | @code{this-command} to the function being executed (which in this case | |
8239 | would be @code{copy-region-as-kill}). At the same time, Emacs sets | |
8240 | the value of @code{last-command} to the previous value of | |
8241 | @code{this-command}. | |
8242 | ||
8243 | In the first part of the body of the @code{copy-region-as-kill} | |
8244 | function, an @code{if} expression determines whether the value of | |
8245 | @code{last-command} is @code{kill-region}. If so, the then-part of | |
8246 | the @code{if} expression is evaluated; it uses the @code{kill-append} | |
8247 | function to concatenate the text copied at this call to the function | |
8248 | with the text already in the first element (the @sc{car}) of the kill | |
8249 | ring. On the other hand, if the value of @code{last-command} is not | |
8250 | @code{kill-region}, then the @code{copy-region-as-kill} function | |
8251 | attaches a new element to the kill ring using the @code{kill-new} | |
8252 | function. | |
8253 | ||
8254 | @need 1250 | |
8255 | The @code{if} expression reads as follows; it uses @code{eq}, which is | |
8256 | a function we have not yet seen: | |
8257 | ||
8258 | @smallexample | |
8259 | @group | |
8260 | (if (eq last-command 'kill-region) | |
8261 | ;; @r{then-part} | |
8262 | (kill-append (buffer-substring beg end) (< end beg)) | |
8263 | ;; @r{else-part} | |
8264 | (kill-new (buffer-substring beg end))) | |
8265 | @end group | |
8266 | @end smallexample | |
8267 | ||
8268 | @findex eq @r{(example of use)} | |
8269 | @noindent | |
8270 | The @code{eq} function tests whether its first argument is the same Lisp | |
8271 | object as its second argument. The @code{eq} function is similar to the | |
8272 | @code{equal} function in that it is used to test for equality, but | |
8273 | differs in that it determines whether two representations are actually | |
8274 | the same object inside the computer, but with different names. | |
8275 | @code{equal} determines whether the structure and contents of two | |
8276 | expressions are the same. | |
8277 | ||
8278 | If the previous command was @code{kill-region}, then the Emacs Lisp | |
8279 | interpreter calls the @code{kill-append} function | |
8280 | ||
8281 | @node kill-append function, kill-new function, last-command & this-command, copy-region-as-kill body | |
8282 | @unnumberedsubsubsec The @code{kill-append} function | |
8283 | @findex kill-append | |
8284 | ||
8285 | @need 800 | |
8286 | The @code{kill-append} function looks like this: | |
8287 | ||
8288 | @smallexample | |
8289 | @group | |
8290 | (defun kill-append (string before-p) | |
8291 | "Append STRING to the end of the latest kill in the kill ring. | |
8292 | If BEFORE-P is non-nil, prepend STRING to the kill. | |
8293 | If `interprogram-cut-function' is set, pass the resulting kill to | |
8294 | it." | |
8295 | (kill-new (if before-p | |
8296 | (concat string (car kill-ring)) | |
8297 | (concat (car kill-ring) string)) | |
8298 | t)) | |
8299 | @end group | |
8300 | @end smallexample | |
8301 | ||
8302 | @noindent | |
8303 | The @code{kill-append} function is fairly straightforward. It uses | |
8304 | the @code{kill-new} function, which we will discuss in more detail in | |
8305 | a moment. | |
8306 | ||
8307 | First, let us look at the conditional that is one of the two arguments | |
8308 | to @code{kill-new}. It uses @code{concat} to concatenate the new text | |
8309 | to the @sc{car} of the kill ring. Whether it prepends or appends the | |
8310 | text depends on the results of an @code{if} expression: | |
8311 | ||
8312 | @smallexample | |
8313 | @group | |
8314 | (if before-p ; @r{if-part} | |
8315 | (concat string (car kill-ring)) ; @r{then-part} | |
8316 | (concat (car kill-ring) string)) ; @r{else-part} | |
8317 | @end group | |
8318 | @end smallexample | |
8319 | ||
8320 | @noindent | |
8321 | If the region being killed is before the region that was killed in the | |
8322 | last command, then it should be prepended before the material that was | |
8323 | saved in the previous kill; and conversely, if the killed text follows | |
8324 | what was just killed, it should be appended after the previous text. | |
8325 | The @code{if} expression depends on the predicate @code{before-p} to | |
8326 | decide whether the newly saved text should be put before or after the | |
8327 | previously saved text. | |
8328 | ||
8329 | The symbol @code{before-p} is the name of one of the arguments to | |
8330 | @code{kill-append}. When the @code{kill-append} function is | |
8331 | evaluated, it is bound to the value returned by evaluating the actual | |
8332 | argument. In this case, this is the expression @code{(< end beg)}. | |
8333 | This expression does not directly determine whether the killed text in | |
8334 | this command is located before or after the kill text of the last | |
8335 | command; what is does is determine whether the value of the variable | |
8336 | @code{end} is less than the value of the variable @code{beg}. If it | |
8337 | is, it means that the user is most likely heading towards the | |
8338 | beginning of the buffer. Also, the result of evaluating the predicate | |
8339 | expression, @code{(< end beg)}, will be true and the text will be | |
8340 | prepended before the previous text. On the other hand, if the value of | |
8341 | the variable @code{end} is greater than the value of the variable | |
8342 | @code{beg}, the text will be appended after the previous text. | |
8343 | ||
8344 | @need 800 | |
8345 | When the newly saved text will be prepended, then the string with the new | |
8346 | text will be concatenated before the old text: | |
8347 | ||
8348 | @smallexample | |
8349 | (concat string (car kill-ring)) | |
8350 | @end smallexample | |
8351 | ||
8352 | @need 1200 | |
8353 | @noindent | |
8354 | But if the text will be appended, it will be concatenated | |
8355 | after the old text: | |
8356 | ||
8357 | @smallexample | |
8358 | (concat (car kill-ring) string)) | |
8359 | @end smallexample | |
8360 | ||
8361 | To understand how this works, we first need to review the | |
8362 | @code{concat} function. The @code{concat} function links together or | |
8363 | unites two strings of text. The result is a string. For example: | |
8364 | ||
8365 | @smallexample | |
8366 | @group | |
8367 | (concat "abc" "def") | |
8368 | @result{} "abcdef" | |
8369 | @end group | |
8370 | ||
8371 | @group | |
8372 | (concat "new " | |
8373 | (car '("first element" "second element"))) | |
8374 | @result{} "new first element" | |
8375 | ||
8376 | (concat (car | |
8377 | '("first element" "second element")) " modified") | |
8378 | @result{} "first element modified" | |
8379 | @end group | |
8380 | @end smallexample | |
8381 | ||
8382 | We can now make sense of @code{kill-append}: it modifies the contents | |
8383 | of the kill ring. The kill ring is a list, each element of which is | |
8384 | saved text. The @code{kill-append} function uses the @code{kill-new} | |
8385 | function which in turn uses the @code{setcar} function. | |
8386 | ||
8387 | @node kill-new function, , kill-append function, copy-region-as-kill body | |
8388 | @unnumberedsubsubsec The @code{kill-new} function | |
8389 | @findex kill-new | |
8390 | ||
8391 | @need 1200 | |
8392 | The @code{kill-new} function looks like this: | |
8393 | ||
8394 | @smallexample | |
8395 | @group | |
8396 | (defun kill-new (string &optional replace) | |
8397 | "Make STRING the latest kill in the kill ring. | |
8398 | Set the kill-ring-yank pointer to point to it. | |
8399 | If `interprogram-cut-function' is non-nil, apply it to STRING. | |
8400 | Optional second argument REPLACE non-nil means that STRING will replace | |
8401 | the front of the kill ring, rather than being added to the list." | |
8402 | @end group | |
8403 | @group | |
8404 | (and (fboundp 'menu-bar-update-yank-menu) | |
8405 | (menu-bar-update-yank-menu string (and replace (car kill-ring)))) | |
8406 | @end group | |
8407 | @group | |
8408 | (if (and replace kill-ring) | |
8409 | (setcar kill-ring string) | |
8410 | (setq kill-ring (cons string kill-ring)) | |
8411 | (if (> (length kill-ring) kill-ring-max) | |
8412 | (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))) | |
8413 | @end group | |
8414 | @group | |
8415 | (setq kill-ring-yank-pointer kill-ring) | |
8416 | (if interprogram-cut-function | |
8417 | (funcall interprogram-cut-function string (not replace)))) | |
8418 | @end group | |
8419 | @end smallexample | |
8420 | ||
8421 | As usual, we can look at this function in parts. | |
8422 | ||
8423 | @need 1200 | |
8424 | The first line of the documentation makes sense: | |
8425 | ||
8426 | @smallexample | |
8427 | Make STRING the latest kill in the kill ring. | |
8428 | @end smallexample | |
8429 | ||
8430 | @noindent | |
8431 | Let's skip over the rest of the documentation for the moment. | |
8432 | ||
8433 | Also, let's skip over the first two lines of code, those involving | |
8434 | @code{menu-bar-update-yank-menu}. We will explain them below. | |
8435 | ||
8436 | @need 1200 | |
8437 | The critical lines are these: | |
8438 | ||
8439 | @smallexample | |
8440 | @group | |
8441 | (if (and replace kill-ring) | |
8442 | ;; @r{then} | |
8443 | (setcar kill-ring string) | |
8444 | @end group | |
8445 | @group | |
8446 | ;; @r{else} | |
8447 | (setq kill-ring (cons string kill-ring)) | |
8448 | (if (> (length kill-ring) kill-ring-max) | |
8449 | ;; @r{avoid overly long kill ring} | |
8450 | (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))) | |
8451 | @end group | |
8452 | @group | |
8453 | (setq kill-ring-yank-pointer kill-ring) | |
8454 | (if interprogram-cut-function | |
8455 | (funcall interprogram-cut-function string (not replace)))) | |
8456 | @end group | |
8457 | @end smallexample | |
8458 | ||
8459 | The conditional test is @w{@code{(and replace kill-ring)}}. | |
8460 | This will be true when two conditions are met: the kill ring has | |
8461 | something in it, and the @code{replace} variable is true. | |
8462 | ||
8463 | @need 1250 | |
8464 | The @code{kill-append} function sets @code{replace} to be true; then, | |
8465 | when the kill ring has at least one item in it, the @code{setcar} | |
8466 | expression is executed: | |
8467 | ||
8468 | @smallexample | |
8469 | (setcar kill-ring string) | |
8470 | @end smallexample | |
8471 | ||
8472 | The @code{setcar} function actually changes the first element of the | |
8473 | @code{kill-ring} list to the value of @code{string}. It replaces the | |
8474 | first element. | |
8475 | ||
8476 | On the other hand, if the kill ring is empty, or replace is false, the | |
8477 | else-part of the condition is executed: | |
8478 | ||
8479 | @smallexample | |
8480 | @group | |
8481 | (setq kill-ring (cons string kill-ring)) | |
8482 | (if (> (length kill-ring) kill-ring-max) | |
8483 | (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)) | |
8484 | @end group | |
8485 | @end smallexample | |
8486 | ||
8487 | @noindent | |
8488 | This expression first constructs a new version of the kill ring by | |
8489 | prepending @code{string} to the existing kill ring as a new element. | |
8490 | Then it executes a second @code{if} clause. This second @code{if} | |
8491 | clause keeps the kill ring from growing too long. | |
8492 | ||
8493 | Let's look at these two expressions in order. | |
8494 | ||
8495 | The @code{setq} line of the else-part sets the new value of the kill | |
8496 | ring to what results from adding the string being killed to the old kill | |
8497 | ring. | |
8498 | ||
8499 | @need 800 | |
8500 | We can see how this works with an example: | |
8501 | ||
8502 | @smallexample | |
8503 | (setq example-list '("here is a clause" "another clause")) | |
8504 | @end smallexample | |
8505 | ||
8506 | @need 1200 | |
8507 | @noindent | |
8508 | After evaluating this expression with @kbd{C-x C-e}, you can evaluate | |
8509 | @code{example-list} and see what it returns: | |
8510 | ||
8511 | @smallexample | |
8512 | @group | |
8513 | example-list | |
8514 | @result{} ("here is a clause" "another clause") | |
8515 | @end group | |
8516 | @end smallexample | |
8517 | ||
8518 | @need 1200 | |
8519 | @noindent | |
8520 | Now, we can add a new element on to this list by evaluating the | |
8521 | following expression: | |
8522 | @findex cons, @r{example} | |
8523 | ||
8524 | @smallexample | |
8525 | (setq example-list (cons "a third clause" example-list)) | |
8526 | @end smallexample | |
8527 | ||
8528 | @need 800 | |
8529 | @noindent | |
8530 | When we evaluate @code{example-list}, we find its value is: | |
8531 | ||
8532 | @smallexample | |
8533 | @group | |
8534 | example-list | |
8535 | @result{} ("a third clause" "here is a clause" "another clause") | |
8536 | @end group | |
8537 | @end smallexample | |
8538 | ||
8539 | @noindent | |
8540 | Thus, the third clause was added to the list by @code{cons}. | |
8541 | ||
8542 | @need 1200 | |
8543 | This is exactly similar to what the @code{setq} and @code{cons} do in | |
8544 | the function. Here is the line again: | |
8545 | ||
8546 | @smallexample | |
8547 | (setq kill-ring (cons string kill-ring)) | |
8548 | @end smallexample | |
8549 | ||
8550 | @need 1200 | |
8551 | Now for the second part of the @code{if} clause. This expression | |
8552 | keeps the kill ring from growing too long. It looks like this: | |
8553 | ||
8554 | @smallexample | |
8555 | @group | |
8556 | (if (> (length kill-ring) kill-ring-max) | |
8557 | (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)) | |
8558 | @end group | |
8559 | @end smallexample | |
8560 | ||
8561 | The code checks whether the length of the kill ring is greater than | |
8562 | the maximum permitted length. This is the value of | |
8563 | @code{kill-ring-max} (which is 60, by default). If the length of the | |
8564 | kill ring is too long, then this code sets the last element of the | |
8565 | kill ring to @code{nil}. It does this by using two functions, | |
8566 | @code{nthcdr} and @code{setcdr}. | |
8567 | ||
8568 | We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}). | |
8569 | It sets the @sc{cdr} of a list, just as @code{setcar} sets the | |
8570 | @sc{car} of a list. In this case, however, @code{setcdr} will not be | |
8571 | setting the @sc{cdr} of the whole kill ring; the @code{nthcdr} | |
8572 | function is used to cause it to set the @sc{cdr} of the next to last | |
8573 | element of the kill ring---this means that since the @sc{cdr} of the | |
8574 | next to last element is the last element of the kill ring, it will set | |
8575 | the last element of the kill ring. | |
8576 | ||
8577 | @findex nthcdr, @r{example} | |
8578 | The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a | |
8579 | list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr} | |
8580 | @dots{} It does this @var{N} times and returns the results. | |
8581 | ||
8582 | @findex setcdr, @r{example} | |
8583 | Thus, if we had a four element list that was supposed to be three | |
8584 | elements long, we could set the @sc{cdr} of the next to last element | |
8585 | to @code{nil}, and thereby shorten the list. | |
8586 | ||
8587 | You can see this by evaluating the following three expressions in turn. | |
8588 | First set the value of @code{trees} to @code{(maple oak pine birch)}, | |
8589 | then set the @sc{cdr} of its second @sc{cdr} to @code{nil} and then | |
8590 | find the value of @code{trees}: | |
8591 | ||
8592 | @smallexample | |
8593 | @group | |
8594 | (setq trees '(maple oak pine birch)) | |
8595 | @result{} (maple oak pine birch) | |
8596 | @end group | |
8597 | ||
8598 | @group | |
8599 | (setcdr (nthcdr 2 trees) nil) | |
8600 | @result{} nil | |
8601 | ||
8602 | trees | |
8603 | @result{} (maple oak pine) | |
8604 | @end group | |
8605 | @end smallexample | |
8606 | ||
8607 | @noindent | |
8608 | (The value returned by the @code{setcdr} expression is @code{nil} since | |
8609 | that is what the @sc{cdr} is set to.) | |
8610 | ||
8611 | To repeat, in @code{kill-new}, the @code{nthcdr} function takes the | |
8612 | @sc{cdr} a number of times that is one less than the maximum permitted | |
8613 | size of the kill ring and sets the @sc{cdr} of that element (which | |
8614 | will be the rest of the elements in the kill ring) to @code{nil}. | |
8615 | This prevents the kill ring from growing too long. | |
8616 | ||
8617 | @need 800 | |
8618 | The next to last expression in the @code{kill-new} function is | |
8619 | ||
8620 | @smallexample | |
8621 | (setq kill-ring-yank-pointer kill-ring) | |
8622 | @end smallexample | |
8623 | ||
8624 | The @code{kill-ring-yank-pointer} is a global variable that is set to be | |
8625 | the @code{kill-ring}. | |
8626 | ||
8627 | Even though the @code{kill-ring-yank-pointer} is called a | |
8628 | @samp{pointer}, it is a variable just like the kill ring. However, the | |
8629 | name has been chosen to help humans understand how the variable is used. | |
8630 | The variable is used in functions such as @code{yank} and | |
8631 | @code{yank-pop} (@pxref{Yanking, , Yanking Text Back}). | |
8632 | ||
8633 | @need 1200 | |
8634 | Now, to return to the first two lines in the body of the function: | |
8635 | ||
8636 | @smallexample | |
8637 | @group | |
8638 | (and (fboundp 'menu-bar-update-yank-menu) | |
8639 | (menu-bar-update-yank-menu string (and replace (car kill-ring)))) | |
8640 | @end group | |
8641 | @end smallexample | |
8642 | ||
8643 | @noindent | |
8644 | This is an expression whose first element is the function @code{and}. | |
8645 | ||
8646 | @findex and, @r{introduced} | |
8647 | The @code{and} special form evaluates each of its arguments until one of | |
8648 | the arguments returns a value of @code{nil}, in which case the | |
8649 | @code{and} expression returns @code{nil}; however, if none of the | |
8650 | arguments returns a value of @code{nil}, the value resulting from | |
8651 | evaluating the last argument is returned. (Since such a value is not | |
8652 | @code{nil}, it is considered true in Emacs Lisp.) In other words, an | |
8653 | @code{and} expression returns a true value only if all its arguments | |
8654 | are true. | |
8655 | @findex and | |
8656 | ||
8657 | In this case, the expression tests first to see whether | |
8658 | @code{menu-bar-update-yank-menu} exists as a function, and if so, | |
8659 | calls it. The @code{fboundp} function returns true if the symbol it | |
8660 | is testing has a function definition that `is not void'. If the | |
8661 | symbol's function definition were void, we would receive an error | |
8662 | message, as we did when we created errors intentionally (@pxref{Making | |
8663 | Errors, , Generate an Error Message}). | |
8664 | ||
8665 | @need 1200 | |
8666 | Essentially, the @code{and} is an @code{if} expression that reads like | |
8667 | this: | |
8668 | ||
8669 | @smallexample | |
8670 | @group | |
8671 | if @var{the-menu-bar-function-exists} | |
8672 | then @var{execute-it} | |
8673 | @end group | |
8674 | @end smallexample | |
8675 | ||
8676 | @code{menu-bar-update-yank-menu} is one of the functions that make it | |
8677 | possible to use the `Select and Paste' menu in the Edit item of a menu | |
8678 | bar; using a mouse, you can look at the various pieces of text you | |
8679 | have saved and select one piece to paste. | |
8680 | ||
8681 | Finally, the last expression in the @code{kill-new} function adds the | |
8682 | newly copied string to whatever facility exists for copying and | |
8683 | pasting among different programs running in a windowing system. In | |
8684 | the X Windowing system, for example, the @code{x-select-text} function | |
8685 | takes the string and stores it in memory operated by X. You can paste | |
8686 | the string in another program, such as an Xterm. | |
8687 | ||
8688 | @need 1200 | |
8689 | The expression looks like this: | |
8690 | ||
8691 | @smallexample | |
8692 | @group | |
8693 | (if interprogram-cut-function | |
8694 | (funcall interprogram-cut-function string (not replace)))) | |
8695 | @end group | |
8696 | @end smallexample | |
8697 | ||
8698 | If an @code{interprogram-cut-function} exists, then Emacs executes | |
8699 | @code{funcall}, which in turn calls its first argument as a function | |
8700 | and passes the remaining arguments to it. (Incidentally, as far as I | |
8701 | can see, this @code{if} expression could be replaced by an @code{and} | |
8702 | expression similar to the one in the first part of the function.) | |
8703 | ||
8704 | We are not going to discuss windowing systems and other programs | |
8705 | further, but merely note that this is a mechanism that enables GNU | |
8706 | Emacs to work easily and well with other programs. | |
8707 | ||
8708 | This code for placing text in the kill ring, either concatenated with | |
8709 | an existing element or as a new element, leads us to the code for | |
8710 | bringing back text that has been cut out of the buffer---the yank | |
8711 | commands. However, before discussing the yank commands, it is better | |
8712 | to learn how lists are implemented in a computer. This will make | |
8713 | clear such mysteries as the use of the term `pointer'. | |
8714 | ||
8715 | @node cons & search-fwd Review, search Exercises, copy-region-as-kill, Cutting & Storing Text | |
8716 | @comment node-name, next, previous, up | |
8717 | @section Review | |
8718 | ||
8719 | Here is a brief summary of some recently introduced functions. | |
8720 | ||
8721 | @table @code | |
8722 | @item car | |
8723 | @itemx cdr | |
8724 | @code{car} returns the first element of a list; @code{cdr} returns the | |
8725 | second and subsequent elements of a list. | |
8726 | ||
8727 | @need 1250 | |
8728 | For example: | |
8729 | ||
8730 | @smallexample | |
8731 | @group | |
8732 | (car '(1 2 3 4 5 6 7)) | |
8733 | @result{} 1 | |
8734 | (cdr '(1 2 3 4 5 6 7)) | |
8735 | @result{} (2 3 4 5 6 7) | |
8736 | @end group | |
8737 | @end smallexample | |
8738 | ||
8739 | @item cons | |
8740 | @code{cons} constructs a list by prepending its first argument to its | |
8741 | second argument. | |
8742 | ||
8743 | @need 1250 | |
8744 | For example: | |
8745 | ||
8746 | @smallexample | |
8747 | @group | |
8748 | (cons 1 '(2 3 4)) | |
8749 | @result{} (1 2 3 4) | |
8750 | @end group | |
8751 | @end smallexample | |
8752 | ||
8753 | @item nthcdr | |
8754 | Return the result of taking @sc{cdr} `n' times on a list. | |
8755 | @iftex | |
8756 | The | |
8757 | @tex | |
8758 | $n^{th}$ | |
8759 | @end tex | |
8760 | @code{cdr}. | |
8761 | @end iftex | |
8762 | The `rest of the rest', as it were. | |
8763 | ||
8764 | @need 1250 | |
8765 | For example: | |
8766 | ||
8767 | @smallexample | |
8768 | @group | |
8769 | (nthcdr 3 '(1 2 3 4 5 6 7)) | |
8770 | @result{} (4 5 6 7) | |
8771 | @end group | |
8772 | @end smallexample | |
8773 | ||
8774 | @item setcar | |
8775 | @itemx setcdr | |
8776 | @code{setcar} changes the first element of a list; @code{setcdr} | |
8777 | changes the second and subsequent elements of a list. | |
8778 | ||
8779 | @need 1250 | |
8780 | For example: | |
8781 | ||
8782 | @smallexample | |
8783 | @group | |
8784 | (setq triple '(1 2 3)) | |
8785 | ||
8786 | (setcar triple '37) | |
8787 | ||
8788 | triple | |
8789 | @result{} (37 2 3) | |
8790 | ||
8791 | (setcdr triple '("foo" "bar")) | |
8792 | ||
8793 | triple | |
8794 | @result{} (37 "foo" "bar") | |
8795 | @end group | |
8796 | @end smallexample | |
8797 | ||
8798 | @item progn | |
8799 | Evaluate each argument in sequence and then return the value of the | |
8800 | last. | |
8801 | ||
8802 | @need 1250 | |
8803 | For example: | |
8804 | ||
8805 | @smallexample | |
8806 | @group | |
8807 | (progn 1 2 3 4) | |
8808 | @result{} 4 | |
8809 | @end group | |
8810 | @end smallexample | |
8811 | ||
8812 | @item save-restriction | |
8813 | Record whatever narrowing is in effect in the current buffer, if any, | |
8814 | and restore that narrowing after evaluating the arguments. | |
8815 | ||
8816 | @item search-forward | |
8817 | Search for a string, and if the string is found, move point. | |
8818 | ||
8819 | @need 1250 | |
8820 | @noindent | |
8821 | Takes four arguments: | |
8822 | ||
8823 | @enumerate | |
8824 | @item | |
8825 | The string to search for. | |
8826 | ||
8827 | @item | |
8828 | Optionally, the limit of the search. | |
8829 | ||
8830 | @item | |
8831 | Optionally, what to do if the search fails, return @code{nil} or an | |
8832 | error message. | |
8833 | ||
8834 | @item | |
8835 | Optionally, how many times to repeat the search; if negative, the | |
8836 | search goes backwards. | |
8837 | @end enumerate | |
8838 | ||
8839 | @item kill-region | |
8840 | @itemx delete-region | |
8841 | @itemx copy-region-as-kill | |
8842 | ||
8843 | @code{kill-region} cuts the text between point and mark from the | |
8844 | buffer and stores that text in the kill ring, so you can get it back | |
8845 | by yanking. | |
8846 | ||
8847 | @code{delete-and-extract-region} removes the text between point and | |
8848 | mark from the buffer and throws it away. You cannot get it back. | |
8849 | ||
8850 | @code{copy-region-as-kill} copies the text between point and mark into | |
8851 | the kill ring, from which you can get it by yanking. The function | |
8852 | does not cut or remove the text from the buffer. | |
8853 | @end table | |
8854 | ||
8855 | @need 1500 | |
8856 | @node search Exercises, , cons & search-fwd Review, Cutting & Storing Text | |
8857 | @section Searching Exercises | |
8858 | ||
8859 | @itemize @bullet | |
8860 | @item | |
8861 | Write an interactive function that searches for a string. If the | |
8862 | search finds the string, leave point after it and display a message | |
8863 | that says ``Found!''. (Do not use @code{search-forward} for the name | |
8864 | of this function; if you do, you will overwrite the existing version of | |
8865 | @code{search-forward} that comes with Emacs. Use a name such as | |
8866 | @code{test-search} instead.) | |
8867 | ||
8868 | @item | |
8869 | Write a function that prints the third element of the kill ring in the | |
8870 | echo area, if any; if the kill ring does not contain a third element, | |
8871 | print an appropriate message. | |
8872 | @end itemize | |
8873 | ||
8874 | @node List Implementation, Yanking, Cutting & Storing Text, Top | |
8875 | @comment node-name, next, previous, up | |
8876 | @chapter How Lists are Implemented | |
8877 | @cindex Lists in a computer | |
8878 | ||
8879 | In Lisp, atoms are recorded in a straightforward fashion; if the | |
8880 | implementation is not straightforward in practice, it is, nonetheless, | |
8881 | straightforward in theory. The atom @samp{rose}, for example, is | |
8882 | recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s}, | |
8883 | @samp{e}. A list, on the other hand, is kept differently. The mechanism | |
8884 | is equally simple, but it takes a moment to get used to the idea. A | |
8885 | list is kept using a series of pairs of pointers. In the series, the | |
8886 | first pointer in each pair points to an atom or to another list, and the | |
8887 | second pointer in each pair points to the next pair, or to the symbol | |
8888 | @code{nil}, which marks the end of the list. | |
8889 | ||
8890 | A pointer itself is quite simply the electronic address of what is | |
8891 | pointed to. Hence, a list is kept as a series of electronic addresses. | |
8892 | ||
8893 | @menu | |
8894 | * Lists diagrammed:: | |
8895 | * Symbols as Chest:: Exploring a powerful metaphor. | |
8896 | * List Exercise:: | |
8897 | @end menu | |
8898 | ||
8899 | @node Lists diagrammed, Symbols as Chest, List Implementation, List Implementation | |
8900 | @ifnottex | |
8901 | @unnumberedsec Lists diagrammed | |
8902 | @end ifnottex | |
8903 | ||
8904 | For example, the list @code{(rose violet buttercup)} has three elements, | |
8905 | @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the | |
8906 | electronic address of @samp{rose} is recorded in a segment of computer | |
8907 | memory along with the address that gives the electronic address of where | |
8908 | the atom @samp{violet} is located; and that address (the one that tells | |
8909 | where @samp{violet} is located) is kept along with an address that tells | |
8910 | where the address for the atom @samp{buttercup} is located. | |
8911 | ||
8912 | @need 1200 | |
8913 | This sounds more complicated than it is and is easier seen in a diagram: | |
8914 | ||
8915 | @c clear print-postscript-figures | |
8916 | @c !!! cons-cell-diagram #1 | |
8917 | @ifnottex | |
8918 | @smallexample | |
8919 | @group | |
8920 | ___ ___ ___ ___ ___ ___ | |
8921 | |___|___|--> |___|___|--> |___|___|--> nil | |
8922 | | | | | |
8923 | | | | | |
8924 | --> rose --> violet --> buttercup | |
8925 | @end group | |
8926 | @end smallexample | |
8927 | @end ifnottex | |
8928 | @ifset print-postscript-figures | |
8929 | @sp 1 | |
8930 | @tex | |
8931 | @image{cons-1} | |
8932 | %%%% old method of including an image | |
8933 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
8934 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}} | |
8935 | % \catcode`\@=0 % | |
8936 | @end tex | |
8937 | @sp 1 | |
8938 | @end ifset | |
8939 | @ifclear print-postscript-figures | |
8940 | @iftex | |
8941 | @smallexample | |
8942 | @group | |
8943 | ___ ___ ___ ___ ___ ___ | |
8944 | |___|___|--> |___|___|--> |___|___|--> nil | |
8945 | | | | | |
8946 | | | | | |
8947 | --> rose --> violet --> buttercup | |
8948 | @end group | |
8949 | @end smallexample | |
8950 | @end iftex | |
8951 | @end ifclear | |
8952 | ||
8953 | @noindent | |
8954 | In the diagram, each box represents a word of computer memory that | |
8955 | holds a Lisp object, usually in the form of a memory address. The boxes, | |
8956 | i.e.@: the addresses, are in pairs. Each arrow points to what the address | |
8957 | is the address of, either an atom or another pair of addresses. The | |
8958 | first box is the electronic address of @samp{rose} and the arrow points | |
8959 | to @samp{rose}; the second box is the address of the next pair of boxes, | |
8960 | the first part of which is the address of @samp{violet} and the second | |
8961 | part of which is the address of the next pair. The very last box | |
8962 | points to the symbol @code{nil}, which marks the end of the list. | |
8963 | ||
8964 | @need 1200 | |
8965 | When a variable is set to a list with a function such as @code{setq}, | |
8966 | it stores the address of the first box in the variable. Thus, | |
8967 | evaluation of the expression | |
8968 | ||
8969 | @smallexample | |
8970 | (setq bouquet '(rose violet buttercup)) | |
8971 | @end smallexample | |
8972 | ||
8973 | @need 1250 | |
8974 | @noindent | |
8975 | creates a situation like this: | |
8976 | ||
8977 | @c cons-cell-diagram #2 | |
8978 | @ifnottex | |
8979 | @smallexample | |
8980 | @group | |
8981 | bouquet | |
8982 | | | |
8983 | | ___ ___ ___ ___ ___ ___ | |
8984 | --> |___|___|--> |___|___|--> |___|___|--> nil | |
8985 | | | | | |
8986 | | | | | |
8987 | --> rose --> violet --> buttercup | |
8988 | @end group | |
8989 | @end smallexample | |
8990 | @end ifnottex | |
8991 | @ifset print-postscript-figures | |
8992 | @sp 1 | |
8993 | @tex | |
8994 | @image{cons-2} | |
8995 | %%%% old method of including an image | |
8996 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
8997 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}} | |
8998 | % \catcode`\@=0 % | |
8999 | @end tex | |
9000 | @sp 1 | |
9001 | @end ifset | |
9002 | @ifclear print-postscript-figures | |
9003 | @iftex | |
9004 | @smallexample | |
9005 | @group | |
9006 | bouquet | |
9007 | | | |
9008 | | ___ ___ ___ ___ ___ ___ | |
9009 | --> |___|___|--> |___|___|--> |___|___|--> nil | |
9010 | | | | | |
9011 | | | | | |
9012 | --> rose --> violet --> buttercup | |
9013 | @end group | |
9014 | @end smallexample | |
9015 | @end iftex | |
9016 | @end ifclear | |
9017 | ||
9018 | @noindent | |
9019 | In this example, the symbol @code{bouquet} holds the address of the first | |
9020 | pair of boxes. | |
9021 | ||
9022 | @need 1200 | |
9023 | This same list can be illustrated in a different sort of box notation | |
9024 | like this: | |
9025 | ||
9026 | @c cons-cell-diagram #2a | |
9027 | @ifnottex | |
9028 | @smallexample | |
9029 | @group | |
9030 | bouquet | |
9031 | | | |
9032 | | -------------- --------------- ---------------- | |
9033 | | | car | cdr | | car | cdr | | car | cdr | | |
9034 | -->| rose | o------->| violet | o------->| butter- | nil | | |
9035 | | | | | | | | cup | | | |
9036 | -------------- --------------- ---------------- | |
9037 | @end group | |
9038 | @end smallexample | |
9039 | @end ifnottex | |
9040 | @ifset print-postscript-figures | |
9041 | @sp 1 | |
9042 | @tex | |
9043 | @image{cons-2a} | |
9044 | %%%% old method of including an image | |
9045 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
9046 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}} | |
9047 | % \catcode`\@=0 % | |
9048 | @end tex | |
9049 | @sp 1 | |
9050 | @end ifset | |
9051 | @ifclear print-postscript-figures | |
9052 | @iftex | |
9053 | @smallexample | |
9054 | @group | |
9055 | bouquet | |
9056 | | | |
9057 | | -------------- --------------- ---------------- | |
9058 | | | car | cdr | | car | cdr | | car | cdr | | |
9059 | -->| rose | o------->| violet | o------->| butter- | nil | | |
9060 | | | | | | | | cup | | | |
9061 | -------------- --------------- ---------------- | |
9062 | @end group | |
9063 | @end smallexample | |
9064 | @end iftex | |
9065 | @end ifclear | |
9066 | ||
9067 | (Symbols consist of more than pairs of addresses, but the structure of | |
9068 | a symbol is made up of addresses. Indeed, the symbol @code{bouquet} | |
9069 | consists of a group of address-boxes, one of which is the address of | |
9070 | the printed word @samp{bouquet}, a second of which is the address of a | |
9071 | function definition attached to the symbol, if any, a third of which | |
9072 | is the address of the first pair of address-boxes for the list | |
9073 | @code{(rose violet buttercup)}, and so on. Here we are showing that | |
9074 | the symbol's third address-box points to the first pair of | |
9075 | address-boxes for the list.) | |
9076 | ||
9077 | If a symbol is set to the @sc{cdr} of a list, the list itself is not | |
9078 | changed; the symbol simply has an address further down the list. (In | |
9079 | the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus, | |
9080 | evaluation of the following expression | |
9081 | ||
9082 | @smallexample | |
9083 | (setq flowers (cdr bouquet)) | |
9084 | @end smallexample | |
9085 | ||
9086 | @need 800 | |
9087 | @noindent | |
9088 | produces this: | |
9089 | ||
9090 | @c cons-cell-diagram #3 | |
9091 | @ifnottex | |
9092 | @sp 1 | |
9093 | @smallexample | |
9094 | @group | |
9095 | bouquet flowers | |
9096 | | | | |
9097 | | ___ ___ | ___ ___ ___ ___ | |
9098 | --> | | | --> | | | | | | | |
9099 | |___|___|----> |___|___|--> |___|___|--> nil | |
9100 | | | | | |
9101 | | | | | |
9102 | --> rose --> violet --> buttercup | |
9103 | @end group | |
9104 | @end smallexample | |
9105 | @sp 1 | |
9106 | @end ifnottex | |
9107 | @ifset print-postscript-figures | |
9108 | @sp 1 | |
9109 | @tex | |
9110 | @image{cons-3} | |
9111 | %%%% old method of including an image | |
9112 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
9113 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}} | |
9114 | % \catcode`\@=0 % | |
9115 | @end tex | |
9116 | @sp 1 | |
9117 | @end ifset | |
9118 | @ifclear print-postscript-figures | |
9119 | @iftex | |
9120 | @sp 1 | |
9121 | @smallexample | |
9122 | @group | |
9123 | bouquet flowers | |
9124 | | | | |
9125 | | ___ ___ | ___ ___ ___ ___ | |
9126 | --> | | | --> | | | | | | | |
9127 | |___|___|----> |___|___|--> |___|___|--> nil | |
9128 | | | | | |
9129 | | | | | |
9130 | --> rose --> violet --> buttercup | |
9131 | @end group | |
9132 | @end smallexample | |
9133 | @sp 1 | |
9134 | @end iftex | |
9135 | @end ifclear | |
9136 | ||
9137 | @noindent | |
9138 | The value of @code{flowers} is @code{(violet buttercup)}, which is | |
9139 | to say, the symbol @code{flowers} holds the address of the pair of | |
9140 | address-boxes, the first of which holds the address of @code{violet}, | |
9141 | and the second of which holds the address of @code{buttercup}. | |
9142 | ||
9143 | A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted | |
9144 | pair}. @xref{List Type, , List Type , elisp, The GNU Emacs Lisp | |
9145 | Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair | |
9146 | Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more | |
9147 | information about cons cells and dotted pairs. | |
9148 | ||
9149 | @need 1200 | |
9150 | The function @code{cons} adds a new pair of addresses to the front of | |
9151 | a series of addresses like that shown above. For example, evaluating | |
9152 | the expression | |
9153 | ||
9154 | @smallexample | |
9155 | (setq bouquet (cons 'lily bouquet)) | |
9156 | @end smallexample | |
9157 | ||
9158 | @need 1500 | |
9159 | @noindent | |
9160 | produces: | |
9161 | ||
9162 | @c cons-cell-diagram #4 | |
9163 | @ifnottex | |
9164 | @sp 1 | |
9165 | @smallexample | |
9166 | @group | |
9167 | bouquet flowers | |
9168 | | | | |
9169 | | ___ ___ ___ ___ | ___ ___ ___ ___ | |
9170 | --> | | | | | | --> | | | | | | | |
9171 | |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil | |
9172 | | | | | | |
9173 | | | | | | |
9174 | --> lily --> rose --> violet --> buttercup | |
9175 | @end group | |
9176 | @end smallexample | |
9177 | @sp 1 | |
9178 | @end ifnottex | |
9179 | @ifset print-postscript-figures | |
9180 | @sp 1 | |
9181 | @tex | |
9182 | @image{cons-4} | |
9183 | %%%% old method of including an image | |
9184 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
9185 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}} | |
9186 | % \catcode`\@=0 % | |
9187 | @end tex | |
9188 | @sp 1 | |
9189 | @end ifset | |
9190 | @ifclear print-postscript-figures | |
9191 | @iftex | |
9192 | @sp 1 | |
9193 | @smallexample | |
9194 | @group | |
9195 | bouquet flowers | |
9196 | | | | |
9197 | | ___ ___ ___ ___ | ___ ___ ___ ___ | |
9198 | --> | | | | | | --> | | | | | | | |
9199 | |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil | |
9200 | | | | | | |
9201 | | | | | | |
9202 | --> lily --> rose --> violet --> buttercup | |
9203 | @end group | |
9204 | @end smallexample | |
9205 | @sp 1 | |
9206 | @end iftex | |
9207 | @end ifclear | |
9208 | ||
9209 | @need 1200 | |
9210 | @noindent | |
9211 | However, this does not change the value of the symbol | |
9212 | @code{flowers}, as you can see by evaluating the following, | |
9213 | ||
9214 | @smallexample | |
9215 | (eq (cdr (cdr bouquet)) flowers) | |
9216 | @end smallexample | |
9217 | ||
9218 | @noindent | |
9219 | which returns @code{t} for true. | |
9220 | ||
9221 | Until it is reset, @code{flowers} still has the value | |
9222 | @code{(violet buttercup)}; that is, it has the address of the cons | |
9223 | cell whose first address is of @code{violet}. Also, this does not | |
9224 | alter any of the pre-existing cons cells; they are all still there. | |
9225 | ||
9226 | Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address | |
9227 | of the next cons cell in the series; to get the @sc{car} of a list, | |
9228 | you get the address of the first element of the list; to @code{cons} a | |
9229 | new element on a list, you add a new cons cell to the front of the list. | |
9230 | That is all there is to it! The underlying structure of Lisp is | |
9231 | brilliantly simple! | |
9232 | ||
9233 | And what does the last address in a series of cons cells refer to? It | |
9234 | is the address of the empty list, of @code{nil}. | |
9235 | ||
9236 | In summary, when a Lisp variable is set to a value, it is provided with | |
9237 | the address of the list to which the variable refers. | |
9238 | ||
9239 | @node Symbols as Chest, List Exercise, Lists diagrammed, List Implementation | |
9240 | @section Symbols as a Chest of Drawers | |
9241 | @cindex Symbols as a Chest of Drawers | |
9242 | @cindex Chest of Drawers, metaphor for a symbol | |
9243 | @cindex Drawers, Chest of, metaphor for a symbol | |
9244 | ||
9245 | In an earlier section, I suggested that you might imagine a symbol as | |
9246 | being a chest of drawers. The function definition is put in one | |
9247 | drawer, the value in another, and so on. What is put in the drawer | |
9248 | holding the value can be changed without affecting the contents of the | |
9249 | drawer holding the function definition, and vice-versa. | |
9250 | ||
9251 | Actually, what is put in each drawer is the address of the value or | |
9252 | function definition. It is as if you found an old chest in the attic, | |
9253 | and in one of its drawers you found a map giving you directions to | |
9254 | where the buried treasure lies. | |
9255 | ||
9256 | (In addition to its name, symbol definition, and variable value, a | |
9257 | symbol has a `drawer' for a @dfn{property list} which can be used to | |
9258 | record other information. Property lists are not discussed here; see | |
9259 | @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp | |
9260 | Reference Manual}.) | |
9261 | ||
9262 | @need 1500 | |
9263 | Here is a fanciful representation: | |
9264 | ||
9265 | @c chest-of-drawers diagram | |
9266 | @ifnottex | |
9267 | @sp 1 | |
9268 | @smallexample | |
9269 | @group | |
9270 | Chest of Drawers Contents of Drawers | |
9271 | ||
9272 | __ o0O0o __ | |
9273 | / \ | |
9274 | --------------------- | |
9275 | | directions to | [map to] | |
9276 | | symbol name | bouquet | |
9277 | | | | |
9278 | +---------------------+ | |
9279 | | directions to | | |
9280 | | symbol definition | [none] | |
9281 | | | | |
9282 | +---------------------+ | |
9283 | | directions to | [map to] | |
9284 | | variable value | (rose violet buttercup) | |
9285 | | | | |
9286 | +---------------------+ | |
9287 | | directions to | | |
9288 | | property list | [not described here] | |
9289 | | | | |
9290 | +---------------------+ | |
9291 | |/ \| | |
9292 | @end group | |
9293 | @end smallexample | |
9294 | @sp 1 | |
9295 | @end ifnottex | |
9296 | @ifset print-postscript-figures | |
9297 | @sp 1 | |
9298 | @tex | |
9299 | @image{drawers} | |
9300 | %%%% old method of including an image | |
9301 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
9302 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}} | |
9303 | % \catcode`\@=0 % | |
9304 | @end tex | |
9305 | @sp 1 | |
9306 | @end ifset | |
9307 | @ifclear print-postscript-figures | |
9308 | @iftex | |
9309 | @sp 1 | |
9310 | @smallexample | |
9311 | @group | |
9312 | Chest of Drawers Contents of Drawers | |
9313 | ||
9314 | __ o0O0o __ | |
9315 | / \ | |
9316 | --------------------- | |
9317 | | directions to | [map to] | |
9318 | | symbol name | bouquet | |
9319 | | | | |
9320 | +---------------------+ | |
9321 | | directions to | | |
9322 | | symbol definition | [none] | |
9323 | | | | |
9324 | +---------------------+ | |
9325 | | directions to | [map to] | |
9326 | | variable value | (rose violet buttercup) | |
9327 | | | | |
9328 | +---------------------+ | |
9329 | | directions to | | |
9330 | | property list | [not described here] | |
9331 | | | | |
9332 | +---------------------+ | |
9333 | |/ \| | |
9334 | @end group | |
9335 | @end smallexample | |
9336 | @sp 1 | |
9337 | @end iftex | |
9338 | @end ifclear | |
9339 | ||
9340 | @node List Exercise, , Symbols as Chest, List Implementation | |
9341 | @section Exercise | |
9342 | ||
9343 | Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two | |
9344 | more flowers on to this list and set this new list to | |
9345 | @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish. | |
9346 | What does the @code{more-flowers} list now contain? | |
9347 | ||
9348 | @node Yanking, Loops & Recursion, List Implementation, Top | |
9349 | @comment node-name, next, previous, up | |
9350 | @chapter Yanking Text Back | |
9351 | @findex yank | |
9352 | @findex rotate-yank-pointer | |
9353 | @cindex Text retrieval | |
9354 | @cindex Retrieving text | |
9355 | @cindex Pasting text | |
9356 | ||
9357 | Whenever you cut text out of a buffer with a `kill' command in GNU Emacs, | |
9358 | you can bring it back with a `yank' command. The text that is cut out of | |
9359 | the buffer is put in the kill ring and the yank commands insert the | |
9360 | appropriate contents of the kill ring back into a buffer (not necessarily | |
9361 | the original buffer). | |
9362 | ||
9363 | A simple @kbd{C-y} (@code{yank}) command inserts the first item from | |
9364 | the kill ring into the current buffer. If the @kbd{C-y} command is | |
9365 | followed immediately by @kbd{M-y}, the first element is replaced by | |
9366 | the second element. Successive @kbd{M-y} commands replace the second | |
9367 | element with the third, fourth, or fifth element, and so on. When the | |
9368 | last element in the kill ring is reached, it is replaced by the first | |
9369 | element and the cycle is repeated. (Thus the kill ring is called a | |
9370 | `ring' rather than just a `list'. However, the actual data structure | |
9371 | that holds the text is a list. | |
9372 | @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the | |
9373 | list is handled as a ring.) | |
9374 | ||
9375 | @menu | |
9376 | * Kill Ring Overview:: The kill ring is a list. | |
9377 | * kill-ring-yank-pointer:: The @code{kill-ring-yank-pointer} variable. | |
9378 | * yank nthcdr Exercises:: | |
9379 | @end menu | |
9380 | ||
9381 | @node Kill Ring Overview, kill-ring-yank-pointer, Yanking, Yanking | |
9382 | @comment node-name, next, previous, up | |
9383 | @section Kill Ring Overview | |
9384 | @cindex Kill ring overview | |
9385 | ||
9386 | The kill ring is a list of textual strings. This is what it looks like: | |
9387 | ||
9388 | @smallexample | |
9389 | ("some text" "a different piece of text" "yet more text") | |
9390 | @end smallexample | |
9391 | ||
9392 | If this were the contents of my kill ring and I pressed @kbd{C-y}, the | |
9393 | string of characters saying @samp{some text} would be inserted in this | |
9394 | buffer where my cursor is located. | |
9395 | ||
9396 | The @code{yank} command is also used for duplicating text by copying it. | |
9397 | The copied text is not cut from the buffer, but a copy of it is put on the | |
9398 | kill ring and is inserted by yanking it back. | |
9399 | ||
9400 | Three functions are used for bringing text back from the kill ring: | |
9401 | @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop}, | |
9402 | which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer}, | |
9403 | which is used by the two other functions. | |
9404 | ||
9405 | These functions refer to the kill ring through a variable called the | |
9406 | @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the | |
9407 | @code{yank} and @code{yank-pop} functions is: | |
9408 | ||
9409 | @smallexample | |
9410 | (insert (car kill-ring-yank-pointer)) | |
9411 | @end smallexample | |
9412 | ||
9413 | To begin to understand how @code{yank} and @code{yank-pop} work, it is | |
9414 | first necessary to look at the @code{kill-ring-yank-pointer} variable | |
9415 | and the @code{rotate-yank-pointer} function. | |
9416 | ||
9417 | @node kill-ring-yank-pointer, yank nthcdr Exercises, Kill Ring Overview, Yanking | |
9418 | @comment node-name, next, previous, up | |
9419 | @section The @code{kill-ring-yank-pointer} Variable | |
9420 | ||
9421 | @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is | |
9422 | a variable. It points to something by being bound to the value of what | |
9423 | it points to, like any other Lisp variable. | |
9424 | ||
9425 | @need 1000 | |
9426 | Thus, if the value of the kill ring is: | |
9427 | ||
9428 | @smallexample | |
9429 | ("some text" "a different piece of text" "yet more text") | |
9430 | @end smallexample | |
9431 | ||
9432 | @need 1250 | |
9433 | @noindent | |
9434 | and the @code{kill-ring-yank-pointer} points to the second clause, the | |
9435 | value of @code{kill-ring-yank-pointer} is: | |
9436 | ||
9437 | @smallexample | |
9438 | ("a different piece of text" "yet more text") | |
9439 | @end smallexample | |
9440 | ||
9441 | As explained in the previous chapter (@pxref{List Implementation}), the | |
9442 | computer does not keep two different copies of the text being pointed to | |
9443 | by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The | |
9444 | words ``a different piece of text'' and ``yet more text'' are not | |
9445 | duplicated. Instead, the two Lisp variables point to the same pieces of | |
9446 | text. Here is a diagram: | |
9447 | ||
9448 | @c cons-cell-diagram #5 | |
9449 | @ifnottex | |
9450 | @smallexample | |
9451 | @group | |
9452 | kill-ring kill-ring-yank-pointer | |
9453 | | | | |
9454 | | ___ ___ | ___ ___ ___ ___ | |
9455 | ---> | | | --> | | | | | | | |
9456 | |___|___|----> |___|___|--> |___|___|--> nil | |
9457 | | | | | |
9458 | | | | | |
9459 | | | --> "yet more text" | |
9460 | | | | |
9461 | | --> "a different piece of text | |
9462 | | | |
9463 | --> "some text" | |
9464 | @end group | |
9465 | @end smallexample | |
9466 | @sp 1 | |
9467 | @end ifnottex | |
9468 | @ifset print-postscript-figures | |
9469 | @sp 1 | |
9470 | @tex | |
9471 | @image{cons-5} | |
9472 | %%%% old method of including an image | |
9473 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
9474 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}} | |
9475 | % \catcode`\@=0 % | |
9476 | @end tex | |
9477 | @sp 1 | |
9478 | @end ifset | |
9479 | @ifclear print-postscript-figures | |
9480 | @iftex | |
9481 | @smallexample | |
9482 | @group | |
9483 | kill-ring kill-ring-yank-pointer | |
9484 | | | | |
9485 | | ___ ___ | ___ ___ ___ ___ | |
9486 | ---> | | | --> | | | | | | | |
9487 | |___|___|----> |___|___|--> |___|___|--> nil | |
9488 | | | | | |
9489 | | | | | |
9490 | | | --> "yet more text" | |
9491 | | | | |
9492 | | --> "a different piece of text | |
9493 | | | |
9494 | --> "some text" | |
9495 | @end group | |
9496 | @end smallexample | |
9497 | @sp 1 | |
9498 | @end iftex | |
9499 | @end ifclear | |
9500 | ||
9501 | Both the variable @code{kill-ring} and the variable | |
9502 | @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is | |
9503 | usually described as if it were actually what it is composed of. The | |
9504 | @code{kill-ring} is spoken of as if it were the list rather than that it | |
9505 | points to the list. Conversely, the @code{kill-ring-yank-pointer} is | |
9506 | spoken of as pointing to a list. | |
9507 | ||
9508 | These two ways of talking about the same thing sound confusing at first but | |
9509 | make sense on reflection. The kill ring is generally thought of as the | |
9510 | complete structure of data that holds the information of what has recently | |
9511 | been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer} | |
9512 | on the other hand, serves to indicate---that is, to `point to'---that part | |
9513 | of the kill ring of which the first element (the @sc{car}) will be | |
9514 | inserted. | |
9515 | ||
9516 | The @code{rotate-yank-pointer} function changes the element in the | |
9517 | kill ring to which the @code{kill-ring-yank-pointer} points; when the | |
9518 | pointer is set to point to the next element beyond the end of the kill | |
9519 | ring, it automatically sets it to point to the first element of the | |
9520 | kill ring. This is how the list is transformed into a ring. The | |
9521 | @code{rotate-yank-pointer} function itself is not difficult, but | |
9522 | contains many details. It and the much simpler @code{yank} and | |
9523 | @code{yank-pop} functions are described in an appendix. | |
9524 | @xref{Kill Ring, , Handling the Kill Ring}. | |
9525 | ||
9526 | @need 1500 | |
9527 | @node yank nthcdr Exercises, , kill-ring-yank-pointer, Yanking | |
9528 | @section Exercises with @code{yank} and @code{nthcdr} | |
9529 | ||
9530 | @itemize @bullet | |
9531 | @item | |
9532 | Using @kbd{C-h v} (@code{describe-variable}), look at the value of | |
9533 | your kill ring. Add several items to your kill ring; look at its | |
9534 | value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way | |
9535 | around the kill ring. How many items were in your kill ring? Find | |
9536 | the value of @code{kill-ring-max}. Was your kill ring full, or could | |
9537 | you have kept more blocks of text within it? | |
9538 | ||
9539 | @item | |
9540 | Using @code{nthcdr} and @code{car}, construct a series of expressions | |
9541 | to return the first, second, third, and fourth elements of a list. | |
9542 | @end itemize | |
9543 | ||
9544 | @node Loops & Recursion, Regexp Search, Yanking, Top | |
9545 | @comment node-name, next, previous, up | |
9546 | @chapter Loops and Recursion | |
9547 | @cindex Loops and recursion | |
9548 | @cindex Recursion and loops | |
9549 | @cindex Repetition (loops) | |
9550 | ||
9551 | Emacs Lisp has two primary ways to cause an expression, or a series of | |
9552 | expressions, to be evaluated repeatedly: one uses a @code{while} | |
9553 | loop, and the other uses @dfn{recursion}. | |
9554 | ||
9555 | Repetition can be very valuable. For example, to move forward four | |
9556 | sentences, you need only write a program that will move forward one | |
9557 | sentence and then repeat the process four times. Since a computer does | |
9558 | not get bored or tired, such repetitive action does not have the | |
9559 | deleterious effects that excessive or the wrong kinds of repetition can | |
9560 | have on humans. | |
9561 | ||
9562 | People mostly write Emacs Lisp functions using @code{while} loops and | |
9563 | their kin; but you can use recursion, which provides a very powerful | |
9564 | way to think about and then to solve problems@footnote{You can write | |
9565 | recursive functions to be frugal or wasteful of mental or computer | |
9566 | resources; as it happens, methods that people find easy---that are | |
9567 | frugal of `mental resources'---sometimes use considerable computer | |
9568 | resources. Emacs was designed to run on machines that we now consider | |
9569 | limited and its default settings are conservative. You may want to | |
9570 | increase the values of @code{max-specpdl-size} and | |
9571 | @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to | |
9572 | 15 and 30 times their default value.}. | |
9573 | ||
9574 | @menu | |
9575 | * while:: Causing a stretch of code to repeat. | |
9576 | * dolist dotimes:: | |
9577 | * Recursion:: Causing a function to call itself. | |
9578 | * Looping exercise:: | |
9579 | @end menu | |
9580 | ||
9581 | @node while, dolist dotimes, Loops & Recursion, Loops & Recursion | |
9582 | @comment node-name, next, previous, up | |
9583 | @section @code{while} | |
9584 | @cindex Loops | |
9585 | @findex while | |
9586 | ||
9587 | The @code{while} special form tests whether the value returned by | |
9588 | evaluating its first argument is true or false. This is similar to what | |
9589 | the Lisp interpreter does with an @code{if}; what the interpreter does | |
9590 | next, however, is different. | |
9591 | ||
9592 | In a @code{while} expression, if the value returned by evaluating the | |
9593 | first argument is false, the Lisp interpreter skips the rest of the | |
9594 | expression (the @dfn{body} of the expression) and does not evaluate it. | |
9595 | However, if the value is true, the Lisp interpreter evaluates the body | |
9596 | of the expression and then again tests whether the first argument to | |
9597 | @code{while} is true or false. If the value returned by evaluating the | |
9598 | first argument is again true, the Lisp interpreter again evaluates the | |
9599 | body of the expression. | |
9600 | ||
9601 | @need 1200 | |
9602 | The template for a @code{while} expression looks like this: | |
9603 | ||
9604 | @smallexample | |
9605 | @group | |
9606 | (while @var{true-or-false-test} | |
9607 | @var{body}@dots{}) | |
9608 | @end group | |
9609 | @end smallexample | |
9610 | ||
9611 | @menu | |
9612 | * Looping with while:: Repeat so long as test returns true. | |
9613 | * Loop Example:: A @code{while} loop that uses a list. | |
9614 | * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}. | |
9615 | * Incrementing Loop:: A loop with an incrementing counter. | |
9616 | * Decrementing Loop:: A loop with a decrementing counter. | |
9617 | @end menu | |
9618 | ||
9619 | @node Looping with while, Loop Example, while, while | |
9620 | @ifnottex | |
9621 | @unnumberedsubsec Looping with @code{while} | |
9622 | @end ifnottex | |
9623 | ||
9624 | So long as the true-or-false-test of the @code{while} expression | |
9625 | returns a true value when it is evaluated, the body is repeatedly | |
9626 | evaluated. This process is called a loop since the Lisp interpreter | |
9627 | repeats the same thing again and again, like an airplane doing a loop. | |
9628 | When the result of evaluating the true-or-false-test is false, the | |
9629 | Lisp interpreter does not evaluate the rest of the @code{while} | |
9630 | expression and `exits the loop'. | |
9631 | ||
9632 | Clearly, if the value returned by evaluating the first argument to | |
9633 | @code{while} is always true, the body following will be evaluated | |
9634 | again and again @dots{} and again @dots{} forever. Conversely, if the | |
9635 | value returned is never true, the expressions in the body will never | |
9636 | be evaluated. The craft of writing a @code{while} loop consists of | |
9637 | choosing a mechanism such that the true-or-false-test returns true | |
9638 | just the number of times that you want the subsequent expressions to | |
9639 | be evaluated, and then have the test return false. | |
9640 | ||
9641 | The value returned by evaluating a @code{while} is the value of the | |
9642 | true-or-false-test. An interesting consequence of this is that a | |
9643 | @code{while} loop that evaluates without error will return @code{nil} | |
9644 | or false regardless of whether it has looped 1 or 100 times or none at | |
9645 | all. A @code{while} expression that evaluates successfully never | |
9646 | returns a true value! What this means is that @code{while} is always | |
9647 | evaluated for its side effects, which is to say, the consequences of | |
9648 | evaluating the expressions within the body of the @code{while} loop. | |
9649 | This makes sense. It is not the mere act of looping that is desired, | |
9650 | but the consequences of what happens when the expressions in the loop | |
9651 | are repeatedly evaluated. | |
9652 | ||
9653 | @node Loop Example, print-elements-of-list, Looping with while, while | |
9654 | @comment node-name, next, previous, up | |
9655 | @subsection A @code{while} Loop and a List | |
9656 | ||
9657 | A common way to control a @code{while} loop is to test whether a list | |
9658 | has any elements. If it does, the loop is repeated; but if it does not, | |
9659 | the repetition is ended. Since this is an important technique, we will | |
9660 | create a short example to illustrate it. | |
9661 | ||
9662 | A simple way to test whether a list has elements is to evaluate the | |
9663 | list: if it has no elements, it is an empty list and will return the | |
9664 | empty list, @code{()}, which is a synonym for @code{nil} or false. On | |
9665 | the other hand, a list with elements will return those elements when it | |
9666 | is evaluated. Since Emacs Lisp considers as true any value that is not | |
9667 | @code{nil}, a list that returns elements will test true in a | |
9668 | @code{while} loop. | |
9669 | ||
9670 | @need 1200 | |
9671 | For example, you can set the variable @code{empty-list} to @code{nil} by | |
9672 | evaluating the following @code{setq} expression: | |
9673 | ||
9674 | @smallexample | |
9675 | (setq empty-list ()) | |
9676 | @end smallexample | |
9677 | ||
9678 | @noindent | |
9679 | After evaluating the @code{setq} expression, you can evaluate the | |
9680 | variable @code{empty-list} in the usual way, by placing the cursor after | |
9681 | the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your | |
9682 | echo area: | |
9683 | ||
9684 | @smallexample | |
9685 | empty-list | |
9686 | @end smallexample | |
9687 | ||
9688 | On the other hand, if you set a variable to be a list with elements, the | |
9689 | list will appear when you evaluate the variable, as you can see by | |
9690 | evaluating the following two expressions: | |
9691 | ||
9692 | @smallexample | |
9693 | @group | |
9694 | (setq animals '(gazelle giraffe lion tiger)) | |
9695 | ||
9696 | animals | |
9697 | @end group | |
9698 | @end smallexample | |
9699 | ||
9700 | Thus, to create a @code{while} loop that tests whether there are any | |
9701 | items in the list @code{animals}, the first part of the loop will be | |
9702 | written like this: | |
9703 | ||
9704 | @smallexample | |
9705 | @group | |
9706 | (while animals | |
9707 | @dots{} | |
9708 | @end group | |
9709 | @end smallexample | |
9710 | ||
9711 | @noindent | |
9712 | When the @code{while} tests its first argument, the variable | |
9713 | @code{animals} is evaluated. It returns a list. So long as the list | |
9714 | has elements, the @code{while} considers the results of the test to be | |
9715 | true; but when the list is empty, it considers the results of the test | |
9716 | to be false. | |
9717 | ||
9718 | To prevent the @code{while} loop from running forever, some mechanism | |
9719 | needs to be provided to empty the list eventually. An oft-used | |
9720 | technique is to have one of the subsequent forms in the @code{while} | |
9721 | expression set the value of the list to be the @sc{cdr} of the list. | |
9722 | Each time the @code{cdr} function is evaluated, the list will be made | |
9723 | shorter, until eventually only the empty list will be left. At this | |
9724 | point, the test of the @code{while} loop will return false, and the | |
9725 | arguments to the @code{while} will no longer be evaluated. | |
9726 | ||
9727 | For example, the list of animals bound to the variable @code{animals} | |
9728 | can be set to be the @sc{cdr} of the original list with the | |
9729 | following expression: | |
9730 | ||
9731 | @smallexample | |
9732 | (setq animals (cdr animals)) | |
9733 | @end smallexample | |
9734 | ||
9735 | @noindent | |
9736 | If you have evaluated the previous expressions and then evaluate this | |
9737 | expression, you will see @code{(giraffe lion tiger)} appear in the echo | |
9738 | area. If you evaluate the expression again, @code{(lion tiger)} will | |
9739 | appear in the echo area. If you evaluate it again and yet again, | |
9740 | @code{(tiger)} appears and then the empty list, shown by @code{nil}. | |
9741 | ||
9742 | A template for a @code{while} loop that uses the @code{cdr} function | |
9743 | repeatedly to cause the true-or-false-test eventually to test false | |
9744 | looks like this: | |
9745 | ||
9746 | @smallexample | |
9747 | @group | |
9748 | (while @var{test-whether-list-is-empty} | |
9749 | @var{body}@dots{} | |
9750 | @var{set-list-to-cdr-of-list}) | |
9751 | @end group | |
9752 | @end smallexample | |
9753 | ||
9754 | This test and use of @code{cdr} can be put together in a function that | |
9755 | goes through a list and prints each element of the list on a line of its | |
9756 | own. | |
9757 | ||
9758 | @node print-elements-of-list, Incrementing Loop, Loop Example, while | |
9759 | @subsection An Example: @code{print-elements-of-list} | |
9760 | @findex print-elements-of-list | |
9761 | ||
9762 | The @code{print-elements-of-list} function illustrates a @code{while} | |
9763 | loop with a list. | |
9764 | ||
9765 | @cindex @file{*scratch*} buffer | |
9766 | The function requires several lines for its output. If you are | |
9767 | reading this in Emacs 21 or a later version, you can evaluate the | |
9768 | following expression inside of Info, as usual. | |
9769 | ||
9770 | If you are using an earlier version of Emacs, you need to copy the | |
9771 | necessary expressions to your @file{*scratch*} buffer and evaluate | |
9772 | them there. This is because the echo area had only one line in the | |
9773 | earlier versions. | |
9774 | ||
9775 | You can copy the expressions by marking the beginning of the region | |
9776 | with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to | |
9777 | the end of the region and then copying the region using @kbd{M-w} | |
9778 | (@code{copy-region-as-kill}). In the @file{*scratch*} buffer, you can | |
9779 | yank the expressions back by typing @kbd{C-y} (@code{yank}). | |
9780 | ||
9781 | After you have copied the expressions to the @file{*scratch*} buffer, | |
9782 | evaluate each expression in turn. Be sure to evaluate the last | |
9783 | expression, @code{(print-elements-of-list animals)}, by typing | |
9784 | @kbd{C-u C-x C-e}, that is, by giving an argument to | |
9785 | @code{eval-last-sexp}. This will cause the result of the evaluation | |
9786 | to be printed in the @file{*scratch*} buffer instead of being printed | |
9787 | in the echo area. (Otherwise you will see something like this in your | |
9788 | echo area: @code{^Jgiraffe^J^Jgazelle^J^Jlion^J^Jtiger^Jnil}, in which | |
9789 | each @samp{^J} stands for a `newline'.) | |
9790 | ||
9791 | @need 1500 | |
9792 | If you are using Emacs 21 or later, you can evaluate these expressions | |
9793 | directly in the Info buffer, and the echo area will grow to show the | |
9794 | results. | |
9795 | ||
9796 | @smallexample | |
9797 | @group | |
9798 | (setq animals '(gazelle giraffe lion tiger)) | |
9799 | ||
9800 | (defun print-elements-of-list (list) | |
9801 | "Print each element of LIST on a line of its own." | |
9802 | (while list | |
9803 | (print (car list)) | |
9804 | (setq list (cdr list)))) | |
9805 | ||
9806 | (print-elements-of-list animals) | |
9807 | @end group | |
9808 | @end smallexample | |
9809 | ||
9810 | @need 1200 | |
9811 | @noindent | |
9812 | When you evaluate the three expressions in sequence, you will see | |
9813 | this: | |
9814 | ||
9815 | @smallexample | |
9816 | @group | |
9817 | giraffe | |
9818 | ||
9819 | gazelle | |
9820 | ||
9821 | lion | |
9822 | ||
9823 | tiger | |
9824 | nil | |
9825 | @end group | |
9826 | @end smallexample | |
9827 | ||
9828 | Each element of the list is printed on a line of its own (that is what | |
9829 | the function @code{print} does) and then the value returned by the | |
9830 | function is printed. Since the last expression in the function is the | |
9831 | @code{while} loop, and since @code{while} loops always return | |
9832 | @code{nil}, a @code{nil} is printed after the last element of the list. | |
9833 | ||
9834 | @node Incrementing Loop, Decrementing Loop, print-elements-of-list, while | |
9835 | @comment node-name, next, previous, up | |
9836 | @subsection A Loop with an Incrementing Counter | |
9837 | ||
9838 | A loop is not useful unless it stops when it ought. Besides | |
9839 | controlling a loop with a list, a common way of stopping a loop is to | |
9840 | write the first argument as a test that returns false when the correct | |
9841 | number of repetitions are complete. This means that the loop must | |
9842 | have a counter---an expression that counts how many times the loop | |
9843 | repeats itself. | |
9844 | ||
9845 | The test can be an expression such as @code{(< count desired-number)} | |
9846 | which returns @code{t} for true if the value of @code{count} is less | |
9847 | than the @code{desired-number} of repetitions and @code{nil} for false if | |
9848 | the value of @code{count} is equal to or is greater than the | |
9849 | @code{desired-number}. The expression that increments the count can be | |
9850 | a simple @code{setq} such as @code{(setq count (1+ count))}, where | |
9851 | @code{1+} is a built-in function in Emacs Lisp that adds 1 to its | |
9852 | argument. (The expression @code{(1+ count)} has the same result as | |
9853 | @code{(+ count 1)}, but is easier for a human to read.) | |
9854 | ||
9855 | @need 1250 | |
9856 | The template for a @code{while} loop controlled by an incrementing | |
9857 | counter looks like this: | |
9858 | ||
9859 | @smallexample | |
9860 | @group | |
9861 | @var{set-count-to-initial-value} | |
9862 | (while (< count desired-number) ; @r{true-or-false-test} | |
9863 | @var{body}@dots{} | |
9864 | (setq count (1+ count))) ; @r{incrementer} | |
9865 | @end group | |
9866 | @end smallexample | |
9867 | ||
9868 | @noindent | |
9869 | Note that you need to set the initial value of @code{count}; usually it | |
9870 | is set to 1. | |
9871 | ||
9872 | @menu | |
9873 | * Incrementing Example:: Counting pebbles in a triangle. | |
9874 | * Inc Example parts:: The parts of the function definition. | |
9875 | * Inc Example altogether:: Putting the function definition together. | |
9876 | @end menu | |
9877 | ||
9878 | @node Incrementing Example, Inc Example parts, Incrementing Loop, Incrementing Loop | |
9879 | @unnumberedsubsubsec Example with incrementing counter | |
9880 | ||
9881 | Suppose you are playing on the beach and decide to make a triangle of | |
9882 | pebbles, putting one pebble in the first row, two in the second row, | |
9883 | three in the third row and so on, like this: | |
9884 | ||
9885 | @sp 1 | |
9886 | @c pebble diagram | |
9887 | @ifnottex | |
9888 | @smallexample | |
9889 | @group | |
9890 | * | |
9891 | * * | |
9892 | * * * | |
9893 | * * * * | |
9894 | @end group | |
9895 | @end smallexample | |
9896 | @end ifnottex | |
9897 | @iftex | |
9898 | @smallexample | |
9899 | @group | |
9900 | @bullet{} | |
9901 | @bullet{} @bullet{} | |
9902 | @bullet{} @bullet{} @bullet{} | |
9903 | @bullet{} @bullet{} @bullet{} @bullet{} | |
9904 | @end group | |
9905 | @end smallexample | |
9906 | @end iftex | |
9907 | @sp 1 | |
9908 | ||
9909 | @noindent | |
9910 | (About 2500 years ago, Pythagoras and others developed the beginnings of | |
9911 | number theory by considering questions such as this.) | |
9912 | ||
9913 | Suppose you want to know how many pebbles you will need to make a | |
9914 | triangle with 7 rows? | |
9915 | ||
9916 | Clearly, what you need to do is add up the numbers from 1 to 7. There | |
9917 | are two ways to do this; start with the smallest number, one, and add up | |
9918 | the list in sequence, 1, 2, 3, 4 and so on; or start with the largest | |
9919 | number and add the list going down: 7, 6, 5, 4 and so on. Because both | |
9920 | mechanisms illustrate common ways of writing @code{while} loops, we will | |
9921 | create two examples, one counting up and the other counting down. In | |
9922 | this first example, we will start with 1 and add 2, 3, 4 and so on. | |
9923 | ||
9924 | If you are just adding up a short list of numbers, the easiest way to do | |
9925 | it is to add up all the numbers at once. However, if you do not know | |
9926 | ahead of time how many numbers your list will have, or if you want to be | |
9927 | prepared for a very long list, then you need to design your addition so | |
9928 | that what you do is repeat a simple process many times instead of doing | |
9929 | a more complex process once. | |
9930 | ||
9931 | For example, instead of adding up all the pebbles all at once, what you | |
9932 | can do is add the number of pebbles in the first row, 1, to the number | |
9933 | in the second row, 2, and then add the total of those two rows to the | |
9934 | third row, 3. Then you can add the number in the fourth row, 4, to the | |
9935 | total of the first three rows; and so on. | |
9936 | ||
9937 | The critical characteristic of the process is that each repetitive | |
9938 | action is simple. In this case, at each step we add only two numbers, | |
9939 | the number of pebbles in the row and the total already found. This | |
9940 | process of adding two numbers is repeated again and again until the last | |
9941 | row has been added to the total of all the preceding rows. In a more | |
9942 | complex loop the repetitive action might not be so simple, but it will | |
9943 | be simpler than doing everything all at once. | |
9944 | ||
9945 | @node Inc Example parts, Inc Example altogether, Incrementing Example, Incrementing Loop | |
9946 | @unnumberedsubsubsec The parts of the function definition | |
9947 | ||
9948 | The preceding analysis gives us the bones of our function definition: | |
9949 | first, we will need a variable that we can call @code{total} that will | |
9950 | be the total number of pebbles. This will be the value returned by | |
9951 | the function. | |
9952 | ||
9953 | Second, we know that the function will require an argument: this | |
9954 | argument will be the total number of rows in the triangle. It can be | |
9955 | called @code{number-of-rows}. | |
9956 | ||
9957 | Finally, we need a variable to use as a counter. We could call this | |
9958 | variable @code{counter}, but a better name is @code{row-number}. | |
9959 | That is because what the counter does is count rows, and a program | |
9960 | should be written to be as understandable as possible. | |
9961 | ||
9962 | When the Lisp interpreter first starts evaluating the expressions in the | |
9963 | function, the value of @code{total} should be set to zero, since we have | |
9964 | not added anything to it. Then the function should add the number of | |
9965 | pebbles in the first row to the total, and then add the number of | |
9966 | pebbles in the second to the total, and then add the number of | |
9967 | pebbles in the third row to the total, and so on, until there are no | |
9968 | more rows left to add. | |
9969 | ||
9970 | Both @code{total} and @code{row-number} are used only inside the | |
9971 | function, so they can be declared as local variables with @code{let} | |
9972 | and given initial values. Clearly, the initial value for @code{total} | |
9973 | should be 0. The initial value of @code{row-number} should be 1, | |
9974 | since we start with the first row. This means that the @code{let} | |
9975 | statement will look like this: | |
9976 | ||
9977 | @smallexample | |
9978 | @group | |
9979 | (let ((total 0) | |
9980 | (row-number 1)) | |
9981 | @var{body}@dots{}) | |
9982 | @end group | |
9983 | @end smallexample | |
9984 | ||
9985 | After the internal variables are declared and bound to their initial | |
9986 | values, we can begin the @code{while} loop. The expression that serves | |
9987 | as the test should return a value of @code{t} for true so long as the | |
9988 | @code{row-number} is less than or equal to the @code{number-of-rows}. | |
9989 | (If the expression tests true only so long as the row number is less | |
9990 | than the number of rows in the triangle, the last row will never be | |
9991 | added to the total; hence the row number has to be either less than or | |
9992 | equal to the number of rows.) | |
9993 | ||
9994 | @need 1500 | |
9995 | @findex <= @r{(less than or equal)} | |
9996 | Lisp provides the @code{<=} function that returns true if the value of | |
9997 | its first argument is less than or equal to the value of its second | |
9998 | argument and false otherwise. So the expression that the @code{while} | |
9999 | will evaluate as its test should look like this: | |
10000 | ||
10001 | @smallexample | |
10002 | (<= row-number number-of-rows) | |
10003 | @end smallexample | |
10004 | ||
10005 | The total number of pebbles can be found by repeatedly adding the number | |
10006 | of pebbles in a row to the total already found. Since the number of | |
10007 | pebbles in the row is equal to the row number, the total can be found by | |
10008 | adding the row number to the total. (Clearly, in a more complex | |
10009 | situation, the number of pebbles in the row might be related to the row | |
10010 | number in a more complicated way; if this were the case, the row number | |
10011 | would be replaced by the appropriate expression.) | |
10012 | ||
10013 | @smallexample | |
10014 | (setq total (+ total row-number)) | |
10015 | @end smallexample | |
10016 | ||
10017 | @noindent | |
10018 | What this does is set the new value of @code{total} to be equal to the | |
10019 | sum of adding the number of pebbles in the row to the previous total. | |
10020 | ||
10021 | After setting the value of @code{total}, the conditions need to be | |
10022 | established for the next repetition of the loop, if there is one. This | |
10023 | is done by incrementing the value of the @code{row-number} variable, | |
10024 | which serves as a counter. After the @code{row-number} variable has | |
10025 | been incremented, the true-or-false-test at the beginning of the | |
10026 | @code{while} loop tests whether its value is still less than or equal to | |
10027 | the value of the @code{number-of-rows} and if it is, adds the new value | |
10028 | of the @code{row-number} variable to the @code{total} of the previous | |
10029 | repetition of the loop. | |
10030 | ||
10031 | @need 1200 | |
10032 | The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the | |
10033 | @code{row-number} variable can be incremented with this expression: | |
10034 | ||
10035 | @smallexample | |
10036 | (setq row-number (1+ row-number)) | |
10037 | @end smallexample | |
10038 | ||
10039 | @node Inc Example altogether, , Inc Example parts, Incrementing Loop | |
10040 | @unnumberedsubsubsec Putting the function definition together | |
10041 | ||
10042 | We have created the parts for the function definition; now we need to | |
10043 | put them together. | |
10044 | ||
10045 | @need 800 | |
10046 | First, the contents of the @code{while} expression: | |
10047 | ||
10048 | @smallexample | |
10049 | @group | |
10050 | (while (<= row-number number-of-rows) ; @r{true-or-false-test} | |
10051 | (setq total (+ total row-number)) | |
10052 | (setq row-number (1+ row-number))) ; @r{incrementer} | |
10053 | @end group | |
10054 | @end smallexample | |
10055 | ||
10056 | Along with the @code{let} expression varlist, this very nearly | |
10057 | completes the body of the function definition. However, it requires | |
10058 | one final element, the need for which is somewhat subtle. | |
10059 | ||
10060 | The final touch is to place the variable @code{total} on a line by | |
10061 | itself after the @code{while} expression. Otherwise, the value returned | |
10062 | by the whole function is the value of the last expression that is | |
10063 | evaluated in the body of the @code{let}, and this is the value | |
10064 | returned by the @code{while}, which is always @code{nil}. | |
10065 | ||
10066 | This may not be evident at first sight. It almost looks as if the | |
10067 | incrementing expression is the last expression of the whole function. | |
10068 | But that expression is part of the body of the @code{while}; it is the | |
10069 | last element of the list that starts with the symbol @code{while}. | |
10070 | Moreover, the whole of the @code{while} loop is a list within the body | |
10071 | of the @code{let}. | |
10072 | ||
10073 | @need 1250 | |
10074 | In outline, the function will look like this: | |
10075 | ||
10076 | @smallexample | |
10077 | @group | |
10078 | (defun @var{name-of-function} (@var{argument-list}) | |
10079 | "@var{documentation}@dots{}" | |
10080 | (let (@var{varlist}) | |
10081 | (while (@var{true-or-false-test}) | |
10082 | @var{body-of-while}@dots{} ) | |
10083 | @dots{} ) ; @r{Need final expression here.} | |
10084 | @end group | |
10085 | @end smallexample | |
10086 | ||
10087 | The result of evaluating the @code{let} is what is going to be returned | |
10088 | by the @code{defun} since the @code{let} is not embedded within any | |
10089 | containing list, except for the @code{defun} as a whole. However, if | |
10090 | the @code{while} is the last element of the @code{let} expression, the | |
10091 | function will always return @code{nil}. This is not what we want! | |
10092 | Instead, what we want is the value of the variable @code{total}. This | |
10093 | is returned by simply placing the symbol as the last element of the list | |
10094 | starting with @code{let}. It gets evaluated after the preceding | |
10095 | elements of the list are evaluated, which means it gets evaluated after | |
10096 | it has been assigned the correct value for the total. | |
10097 | ||
10098 | It may be easier to see this by printing the list starting with | |
10099 | @code{let} all on one line. This format makes it evident that the | |
10100 | @var{varlist} and @code{while} expressions are the second and third | |
10101 | elements of the list starting with @code{let}, and the @code{total} is | |
10102 | the last element: | |
10103 | ||
10104 | @smallexample | |
10105 | @group | |
10106 | (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total) | |
10107 | @end group | |
10108 | @end smallexample | |
10109 | ||
10110 | @need 1200 | |
10111 | Putting everything together, the @code{triangle} function definition | |
10112 | looks like this: | |
10113 | ||
10114 | @smallexample | |
10115 | @group | |
10116 | (defun triangle (number-of-rows) ; @r{Version with} | |
10117 | ; @r{ incrementing counter.} | |
10118 | "Add up the number of pebbles in a triangle. | |
10119 | The first row has one pebble, the second row two pebbles, | |
10120 | the third row three pebbles, and so on. | |
10121 | The argument is NUMBER-OF-ROWS." | |
10122 | @end group | |
10123 | @group | |
10124 | (let ((total 0) | |
10125 | (row-number 1)) | |
10126 | (while (<= row-number number-of-rows) | |
10127 | (setq total (+ total row-number)) | |
10128 | (setq row-number (1+ row-number))) | |
10129 | total)) | |
10130 | @end group | |
10131 | @end smallexample | |
10132 | ||
10133 | @need 1200 | |
10134 | After you have installed @code{triangle} by evaluating the function, you | |
10135 | can try it out. Here are two examples: | |
10136 | ||
10137 | @smallexample | |
10138 | @group | |
10139 | (triangle 4) | |
10140 | ||
10141 | (triangle 7) | |
10142 | @end group | |
10143 | @end smallexample | |
10144 | ||
10145 | @noindent | |
10146 | The sum of the first four numbers is 10 and the sum of the first seven | |
10147 | numbers is 28. | |
10148 | ||
10149 | @node Decrementing Loop, , Incrementing Loop, while | |
10150 | @comment node-name, next, previous, up | |
10151 | @subsection Loop with a Decrementing Counter | |
10152 | ||
10153 | Another common way to write a @code{while} loop is to write the test | |
10154 | so that it determines whether a counter is greater than zero. So long | |
10155 | as the counter is greater than zero, the loop is repeated. But when | |
10156 | the counter is equal to or less than zero, the loop is stopped. For | |
10157 | this to work, the counter has to start out greater than zero and then | |
10158 | be made smaller and smaller by a form that is evaluated | |
10159 | repeatedly. | |
10160 | ||
10161 | The test will be an expression such as @code{(> counter 0)} which | |
10162 | returns @code{t} for true if the value of @code{counter} is greater | |
10163 | than zero, and @code{nil} for false if the value of @code{counter} is | |
10164 | equal to or less than zero. The expression that makes the number | |
10165 | smaller and smaller can be a simple @code{setq} such as @code{(setq | |
10166 | counter (1- counter))}, where @code{1-} is a built-in function in | |
10167 | Emacs Lisp that subtracts 1 from its argument. | |
10168 | ||
10169 | @need 1250 | |
10170 | The template for a decrementing @code{while} loop looks like this: | |
10171 | ||
10172 | @smallexample | |
10173 | @group | |
10174 | (while (> counter 0) ; @r{true-or-false-test} | |
10175 | @var{body}@dots{} | |
10176 | (setq counter (1- counter))) ; @r{decrementer} | |
10177 | @end group | |
10178 | @end smallexample | |
10179 | ||
10180 | @menu | |
10181 | * Decrementing Example:: More pebbles on the beach. | |
10182 | * Dec Example parts:: The parts of the function definition. | |
10183 | * Dec Example altogether:: Putting the function definition together. | |
10184 | @end menu | |
10185 | ||
10186 | @node Decrementing Example, Dec Example parts, Decrementing Loop, Decrementing Loop | |
10187 | @unnumberedsubsubsec Example with decrementing counter | |
10188 | ||
10189 | To illustrate a loop with a decrementing counter, we will rewrite the | |
10190 | @code{triangle} function so the counter decreases to zero. | |
10191 | ||
10192 | This is the reverse of the earlier version of the function. In this | |
10193 | case, to find out how many pebbles are needed to make a triangle with | |
10194 | 3 rows, add the number of pebbles in the third row, 3, to the number | |
10195 | in the preceding row, 2, and then add the total of those two rows to | |
10196 | the row that precedes them, which is 1. | |
10197 | ||
10198 | Likewise, to find the number of pebbles in a triangle with 7 rows, add | |
10199 | the number of pebbles in the seventh row, 7, to the number in the | |
10200 | preceding row, which is 6, and then add the total of those two rows to | |
10201 | the row that precedes them, which is 5, and so on. As in the previous | |
10202 | example, each addition only involves adding two numbers, the total of | |
10203 | the rows already added up and the number of pebbles in the row that is | |
10204 | being added to the total. This process of adding two numbers is | |
10205 | repeated again and again until there are no more pebbles to add. | |
10206 | ||
10207 | We know how many pebbles to start with: the number of pebbles in the | |
10208 | last row is equal to the number of rows. If the triangle has seven | |
10209 | rows, the number of pebbles in the last row is 7. Likewise, we know how | |
10210 | many pebbles are in the preceding row: it is one less than the number in | |
10211 | the row. | |
10212 | ||
10213 | @node Dec Example parts, Dec Example altogether, Decrementing Example, Decrementing Loop | |
10214 | @unnumberedsubsubsec The parts of the function definition | |
10215 | ||
10216 | We start with three variables: the total number of rows in the | |
10217 | triangle; the number of pebbles in a row; and the total number of | |
10218 | pebbles, which is what we want to calculate. These variables can be | |
10219 | named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and | |
10220 | @code{total}, respectively. | |
10221 | ||
10222 | Both @code{total} and @code{number-of-pebbles-in-row} are used only | |
10223 | inside the function and are declared with @code{let}. The initial | |
10224 | value of @code{total} should, of course, be zero. However, the | |
10225 | initial value of @code{number-of-pebbles-in-row} should be equal to | |
10226 | the number of rows in the triangle, since the addition will start with | |
10227 | the longest row. | |
10228 | ||
10229 | @need 1250 | |
10230 | This means that the beginning of the @code{let} expression will look | |
10231 | like this: | |
10232 | ||
10233 | @smallexample | |
10234 | @group | |
10235 | (let ((total 0) | |
10236 | (number-of-pebbles-in-row number-of-rows)) | |
10237 | @var{body}@dots{}) | |
10238 | @end group | |
10239 | @end smallexample | |
10240 | ||
10241 | The total number of pebbles can be found by repeatedly adding the number | |
10242 | of pebbles in a row to the total already found, that is, by repeatedly | |
10243 | evaluating the following expression: | |
10244 | ||
10245 | @smallexample | |
10246 | (setq total (+ total number-of-pebbles-in-row)) | |
10247 | @end smallexample | |
10248 | ||
10249 | @noindent | |
10250 | After the @code{number-of-pebbles-in-row} is added to the @code{total}, | |
10251 | the @code{number-of-pebbles-in-row} should be decremented by one, since | |
10252 | the next time the loop repeats, the preceding row will be | |
10253 | added to the total. | |
10254 | ||
10255 | The number of pebbles in a preceding row is one less than the number of | |
10256 | pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be | |
10257 | used to compute the number of pebbles in the preceding row. This can be | |
10258 | done with the following expression: | |
10259 | ||
10260 | @smallexample | |
10261 | @group | |
10262 | (setq number-of-pebbles-in-row | |
10263 | (1- number-of-pebbles-in-row)) | |
10264 | @end group | |
10265 | @end smallexample | |
10266 | ||
10267 | Finally, we know that the @code{while} loop should stop making repeated | |
10268 | additions when there are no pebbles in a row. So the test for | |
10269 | the @code{while} loop is simply: | |
10270 | ||
10271 | @smallexample | |
10272 | (while (> number-of-pebbles-in-row 0) | |
10273 | @end smallexample | |
10274 | ||
10275 | @node Dec Example altogether, , Dec Example parts, Decrementing Loop | |
10276 | @unnumberedsubsubsec Putting the function definition together | |
10277 | ||
10278 | We can put these expressions together to create a function definition | |
10279 | that works. However, on examination, we find that one of the local | |
10280 | variables is unneeded! | |
10281 | ||
10282 | @need 1250 | |
10283 | The function definition looks like this: | |
10284 | ||
10285 | @smallexample | |
10286 | @group | |
10287 | ;;; @r{First subtractive version.} | |
10288 | (defun triangle (number-of-rows) | |
10289 | "Add up the number of pebbles in a triangle." | |
10290 | (let ((total 0) | |
10291 | (number-of-pebbles-in-row number-of-rows)) | |
10292 | (while (> number-of-pebbles-in-row 0) | |
10293 | (setq total (+ total number-of-pebbles-in-row)) | |
10294 | (setq number-of-pebbles-in-row | |
10295 | (1- number-of-pebbles-in-row))) | |
10296 | total)) | |
10297 | @end group | |
10298 | @end smallexample | |
10299 | ||
10300 | As written, this function works. | |
10301 | ||
10302 | However, we do not need @code{number-of-pebbles-in-row}. | |
10303 | ||
10304 | @cindex Argument as local variable | |
10305 | When the @code{triangle} function is evaluated, the symbol | |
10306 | @code{number-of-rows} will be bound to a number, giving it an initial | |
10307 | value. That number can be changed in the body of the function as if | |
10308 | it were a local variable, without any fear that such a change will | |
10309 | effect the value of the variable outside of the function. This is a | |
10310 | very useful characteristic of Lisp; it means that the variable | |
10311 | @code{number-of-rows} can be used anywhere in the function where | |
10312 | @code{number-of-pebbles-in-row} is used. | |
10313 | ||
10314 | @need 800 | |
10315 | Here is a second version of the function written a bit more cleanly: | |
10316 | ||
10317 | @smallexample | |
10318 | @group | |
10319 | (defun triangle (number) ; @r{Second version.} | |
10320 | "Return sum of numbers 1 through NUMBER inclusive." | |
10321 | (let ((total 0)) | |
10322 | (while (> number 0) | |
10323 | (setq total (+ total number)) | |
10324 | (setq number (1- number))) | |
10325 | total)) | |
10326 | @end group | |
10327 | @end smallexample | |
10328 | ||
10329 | In brief, a properly written @code{while} loop will consist of three parts: | |
10330 | ||
10331 | @enumerate | |
10332 | @item | |
10333 | A test that will return false after the loop has repeated itself the | |
10334 | correct number of times. | |
10335 | ||
10336 | @item | |
10337 | An expression the evaluation of which will return the value desired | |
10338 | after being repeatedly evaluated. | |
10339 | ||
10340 | @item | |
10341 | An expression to change the value passed to the true-or-false-test so | |
10342 | that the test returns false after the loop has repeated itself the right | |
10343 | number of times. | |
10344 | @end enumerate | |
10345 | ||
10346 | @node dolist dotimes, Recursion, while, Loops & Recursion | |
10347 | @comment node-name, next, previous, up | |
10348 | @section Save your time: @code{dolist} and @code{dotimes} | |
10349 | ||
10350 | In addition to @code{while}, both @code{dolist} and @code{dotimes} | |
10351 | provide for looping. Sometimes these are quicker to write than the | |
10352 | equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, , | |
10353 | Macros, elisp, The GNU Emacs Lisp Reference Manual}. ) | |
10354 | ||
10355 | @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a | |
10356 | list': @code{dolist} automatically shortens the list each time it | |
10357 | loops---takes the @sc{cdr} of the list---and binds the @sc{car} of | |
10358 | each shorter version of the list to the first of its arguments. | |
10359 | ||
10360 | @code{dotimes} loops a specific number of time: you specify the number. | |
10361 | ||
10362 | @menu | |
10363 | * dolist:: | |
10364 | * dotimes:: | |
10365 | @end menu | |
10366 | ||
10367 | @node dolist, dotimes, dolist dotimes, dolist dotimes | |
10368 | @unnumberedsubsubsec The @code{dolist} Macro | |
10369 | @findex dolist | |
10370 | ||
10371 | Suppose, for example, you want to reverse a list, so that | |
10372 | ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''. | |
10373 | ||
10374 | @need 1250 | |
10375 | In practice, you would use the @code{reverse} function, like this: | |
10376 | ||
10377 | @smallexample | |
10378 | @group | |
10379 | (setq animals '(gazelle giraffe lion tiger)) | |
10380 | ||
10381 | (reverse animals) | |
10382 | @end group | |
10383 | @end smallexample | |
10384 | ||
10385 | @need 800 | |
10386 | @noindent | |
10387 | Here is how you could reverse the list using a @code{while} loop: | |
10388 | ||
10389 | @smallexample | |
10390 | @group | |
10391 | (setq animals '(gazelle giraffe lion tiger)) | |
10392 | ||
10393 | (defun reverse-list-with-while (list) | |
10394 | "Using while, reverse the order of LIST." | |
10395 | (let (value) ; make sure list starts empty | |
10396 | (while list | |
10397 | (setq value (cons (car list) value)) | |
10398 | (setq list (cdr list))) | |
10399 | value)) | |
10400 | ||
10401 | (reverse-list-with-while animals) | |
10402 | @end group | |
10403 | @end smallexample | |
10404 | ||
10405 | @need 800 | |
10406 | @noindent | |
10407 | And here is how you could use the @code{dolist} macro: | |
10408 | ||
10409 | @smallexample | |
10410 | @group | |
10411 | (setq animals '(gazelle giraffe lion tiger)) | |
10412 | ||
10413 | (defun reverse-list-with-dolist (list) | |
10414 | "Using dolist, reverse the order of LIST." | |
10415 | (let (value) ; make sure list starts empty | |
10416 | (dolist (element list value) | |
10417 | (setq value (cons element value))))) | |
10418 | ||
10419 | (reverse-list-with-dolist animals) | |
10420 | @end group | |
10421 | @end smallexample | |
10422 | ||
10423 | @need 1250 | |
10424 | @noindent | |
10425 | In Info, you can place your cursor after the closing parenthesis of | |
10426 | each expression and type @kbd{C-x C-e}; in each case, you should see | |
10427 | ||
10428 | @smallexample | |
10429 | (tiger lion giraffe gazelle) | |
10430 | @end smallexample | |
10431 | ||
10432 | @noindent | |
10433 | in the echo area. | |
10434 | ||
10435 | For this example, the existing @code{reverse} function is obviously best. | |
10436 | The @code{while} loop is just like our first example (@pxref{Loop | |
10437 | Example, , A @code{while} Loop and a List}). The @code{while} first | |
10438 | checks whether the list has elements; if so, it constructs a new list | |
10439 | by adding the first element of the list to the existing list (which in | |
10440 | the first iteration of the loop is @code{nil}). Since the second | |
10441 | element is prepended in front of the first element, and the third | |
10442 | element is prepended in front of the second element, the list is reversed. | |
10443 | ||
10444 | In the expression using a @code{while} loop, | |
10445 | the @w{@code{(setq list (cdr list))}} | |
10446 | expression shortens the list, so the @code{while} loop eventually | |
10447 | stops. In addition, it provides the @code{cons} expression with a new | |
10448 | first element by creating a new and shorter list at each repetition of | |
10449 | the loop. | |
10450 | ||
10451 | The @code{dolist} expression does very much the same as the | |
10452 | @code{while} expression, except that the @code{dolist} macro does some | |
10453 | of the work you have to do when writing a @code{while} expression. | |
10454 | ||
10455 | Like a @code{while} loop, a @code{dolist} loops. What is different is | |
10456 | that it automatically shortens the list each time it loops --- it | |
10457 | `@sc{cdr}s down the list' on its own --- and it automatically binds | |
10458 | the @sc{car} of each shorter version of the list to the first of its | |
10459 | arguments. | |
10460 | ||
10461 | In the example, the @sc{car} of each shorter version of the list is | |
10462 | referred to using the symbol @samp{element}, the list itself is called | |
10463 | @samp{list}, and the value returned is called @samp{value}. The | |
10464 | remainder of the @code{dolist} expression is the body. | |
10465 | ||
10466 | The @code{dolist} expression binds the @sc{car} of each shorter | |
10467 | version of the list to @code{element} and then evaluates the body of | |
10468 | the expression; and repeats the loop. The result is returned in | |
10469 | @code{value}. | |
10470 | ||
10471 | @node dotimes, , dolist, dolist dotimes | |
10472 | @unnumberedsubsubsec The @code{dotimes} Macro | |
10473 | @findex dotimes | |
10474 | ||
10475 | The @code{dotimes} macro is similar to @code{dolist}, except that it | |
10476 | loops a specific number of times. | |
10477 | ||
10478 | The first argument to @code{dotimes} is assigned the numbers 0, 1, 2 | |
10479 | and so forth each time around the loop, and the value of the third | |
10480 | argument is returned. You need to provide the value of the second | |
10481 | argument, which is how many times the macro loops. | |
10482 | ||
10483 | @need 1250 | |
10484 | For example, the following binds the numbers from 0 up to, but not | |
10485 | including, the number 3 to the first argument, @var{number}, and then | |
10486 | constructs a list of the three numbers. (The first number is 0, the | |
10487 | second number is 1, and the third number is 2; this makes a total of | |
10488 | three numbers in all, starting with zero as the first number.) | |
10489 | ||
10490 | @smallexample | |
10491 | @group | |
10492 | (let (value) ; otherwise a value is a void variable | |
10493 | (dotimes (number 3 value) | |
10494 | (setq value (cons number value)))) | |
10495 | ||
10496 | @result{} (2 1 0) | |
10497 | @end group | |
10498 | @end smallexample | |
10499 | ||
10500 | @noindent | |
10501 | @code{dotimes} returns @code{value}, so the way to use | |
10502 | @code{dotimes} is to operate on some expression @var{number} number of | |
10503 | times and then return the result, either as a list or an atom. | |
10504 | ||
10505 | @need 1250 | |
10506 | Here is an example of a @code{defun} that uses @code{dotimes} to add | |
10507 | up the number of pebbles in a triangle. | |
10508 | ||
10509 | @smallexample | |
10510 | @group | |
10511 | (defun triangle-using-dotimes (number-of-rows) | |
10512 | "Using dotimes, add up the number of pebbles in a triangle." | |
10513 | (let ((total 0)) ; otherwise a total is a void variable | |
10514 | (dotimes (number number-of-rows total) | |
10515 | (setq total (+ total (1+ number)))))) | |
10516 | ||
10517 | (triangle-using-dotimes 4) | |
10518 | @end group | |
10519 | @end smallexample | |
10520 | ||
10521 | @node Recursion, Looping exercise, dolist dotimes, Loops & Recursion | |
10522 | @comment node-name, next, previous, up | |
10523 | @section Recursion | |
10524 | @cindex Recursion | |
10525 | ||
10526 | A recursive function contains code that tells the Lisp interpreter to | |
10527 | call a program that runs exactly like itself, but with slightly | |
10528 | different arguments. The code runs exactly the same because it has | |
10529 | the same name. However, even though it has the same name, it is not | |
10530 | the same thread of execution. It is different. In the jargon, it is | |
10531 | a different `instance'. | |
10532 | ||
10533 | Eventually, if the program is written correctly, the `slightly | |
10534 | different arguments' will become sufficiently different from the first | |
10535 | arguments that the final instance will stop. | |
10536 | ||
10537 | @menu | |
10538 | * Building Robots:: Same model, different serial number ... | |
10539 | * Recursive Definition Parts:: Walk until you stop ... | |
10540 | * Recursion with list:: Using a list as the test whether to recurse. | |
10541 | * Recursive triangle function:: | |
10542 | * Recursion with cond:: | |
10543 | * Recursive Patterns:: Often used templates. | |
10544 | * No Deferment:: Don't store up work ... | |
10545 | * No deferment solution:: | |
10546 | @end menu | |
10547 | ||
10548 | @node Building Robots, Recursive Definition Parts, Recursion, Recursion | |
10549 | @comment node-name, next, previous, up | |
10550 | @subsection Building Robots: Extending the Metaphor | |
10551 | @cindex Building robots | |
10552 | @cindex Robots, building | |
10553 | ||
10554 | It is sometimes helpful to think of a running program as a robot that | |
10555 | does a job. In doing its job, a recursive function calls on a second | |
10556 | robot to help it. The second robot is identical to the first in every | |
10557 | way, except that the second robot helps the first and has been | |
10558 | passed different arguments than the first. | |
10559 | ||
10560 | In a recursive function, the second robot may call a third; and the | |
10561 | third may call a fourth, and so on. Each of these is a different | |
10562 | entity; but all are clones. | |
10563 | ||
10564 | Since each robot has slightly different instructions---the arguments | |
10565 | will differ from one robot to the next---the last robot should know | |
10566 | when to stop. | |
10567 | ||
10568 | Let's expand on the metaphor in which a computer program is a robot. | |
10569 | ||
10570 | A function definition provides the blueprints for a robot. When you | |
10571 | install a function definition, that is, when you evaluate a | |
10572 | @code{defun} special form, you install the necessary equipment to | |
10573 | build robots. It is as if you were in a factory, setting up an | |
10574 | assembly line. Robots with the same name are built according to the | |
10575 | same blueprints. So they have, as it were, the same `model number', | |
10576 | but a different `serial number'. | |
10577 | ||
10578 | We often say that a recursive function `calls itself'. What we mean | |
10579 | is that the instructions in a recursive function cause the Lisp | |
10580 | interpreter to run a different function that has the same name and | |
10581 | does the same job as the first, but with different arguments. | |
10582 | ||
10583 | It is important that the arguments differ from one instance to the | |
10584 | next; otherwise, the process will never stop. | |
10585 | ||
10586 | @node Recursive Definition Parts, Recursion with list, Building Robots, Recursion | |
10587 | @comment node-name, next, previous, up | |
10588 | @subsection The Parts of a Recursive Definition | |
10589 | @cindex Parts of a Recursive Definition | |
10590 | @cindex Recursive Definition Parts | |
10591 | ||
10592 | A recursive function typically contains a conditional expression which | |
10593 | has three parts: | |
10594 | ||
10595 | @enumerate | |
10596 | @item | |
10597 | A true-or-false-test that determines whether the function is called | |
10598 | again, here called the @dfn{do-again-test}. | |
10599 | ||
10600 | @item | |
10601 | The name of the function. When this name is called, a new instance of | |
10602 | the function---a new robot, as it were---is created and told what to do. | |
10603 | ||
10604 | @item | |
10605 | An expression that returns a different value each time the function is | |
10606 | called, here called the @dfn{next-step-expression}. Consequently, the | |
10607 | argument (or arguments) passed to the new instance of the function | |
10608 | will be different from that passed to the previous instance. This | |
10609 | causes the conditional expression, the @dfn{do-again-test}, to test | |
10610 | false after the correct number of repetitions. | |
10611 | @end enumerate | |
10612 | ||
10613 | Recursive functions can be much simpler than any other kind of | |
10614 | function. Indeed, when people first start to use them, they often look | |
10615 | so mysteriously simple as to be incomprehensible. Like riding a | |
10616 | bicycle, reading a recursive function definition takes a certain knack | |
10617 | which is hard at first but then seems simple. | |
10618 | ||
10619 | @need 1200 | |
10620 | There are several different common recursive patterns. A very simple | |
10621 | pattern looks like this: | |
10622 | ||
10623 | @smallexample | |
10624 | @group | |
10625 | (defun @var{name-of-recursive-function} (@var{argument-list}) | |
10626 | "@var{documentation}@dots{}" | |
10627 | (if @var{do-again-test} | |
10628 | @var{body}@dots{} | |
10629 | (@var{name-of-recursive-function} | |
10630 | @var{next-step-expression}))) | |
10631 | @end group | |
10632 | @end smallexample | |
10633 | ||
10634 | Each time a recursive function is evaluated, a new instance of it is | |
10635 | created and told what to do. The arguments tell the instance what to do. | |
10636 | ||
10637 | An argument is bound to the value of the next-step-expression. Each | |
10638 | instance runs with a different value of the next-step-expression. | |
10639 | ||
10640 | The value in the next-step-expression is used in the do-again-test. | |
10641 | ||
10642 | The value returned by the next-step-expression is passed to the new | |
10643 | instance of the function, which evaluates it (or some | |
10644 | transmogrification of it) to determine whether to continue or stop. | |
10645 | The next-step-expression is designed so that the do-again-test returns | |
10646 | false when the function should no longer be repeated. | |
10647 | ||
10648 | The do-again-test is sometimes called the @dfn{stop condition}, | |
10649 | since it stops the repetitions when it tests false. | |
10650 | ||
10651 | @node Recursion with list, Recursive triangle function, Recursive Definition Parts, Recursion | |
10652 | @comment node-name, next, previous, up | |
10653 | @subsection Recursion with a List | |
10654 | ||
10655 | The example of a @code{while} loop that printed the elements of a list | |
10656 | of numbers can be written recursively. Here is the code, including | |
10657 | an expression to set the value of the variable @code{animals} to a list. | |
10658 | ||
10659 | If you are using Emacs 20 or before, this example must be copied to | |
10660 | the @file{*scratch*} buffer and each expression must be evaluated | |
10661 | there. Use @kbd{C-u C-x C-e} to evaluate the | |
10662 | @code{(print-elements-recursively animals)} expression so that the | |
10663 | results are printed in the buffer; otherwise the Lisp interpreter will | |
10664 | try to squeeze the results into the one line of the echo area. | |
10665 | ||
10666 | Also, place your cursor immediately after the last closing parenthesis | |
10667 | of the @code{print-elements-recursively} function, before the comment. | |
10668 | Otherwise, the Lisp interpreter will try to evaluate the comment. | |
10669 | ||
10670 | If you are using Emacs 21 or later, you can evaluate this expression | |
10671 | directly in Info. | |
10672 | ||
10673 | @findex print-elements-recursively | |
10674 | @smallexample | |
10675 | @group | |
10676 | (setq animals '(gazelle giraffe lion tiger)) | |
10677 | ||
10678 | (defun print-elements-recursively (list) | |
10679 | "Print each element of LIST on a line of its own. | |
10680 | Uses recursion." | |
10681 | (if list ; @r{do-again-test} | |
10682 | (progn | |
10683 | (print (car list)) ; @r{body} | |
10684 | (print-elements-recursively ; @r{recursive call} | |
10685 | (cdr list))))) ; @r{next-step-expression} | |
10686 | ||
10687 | (print-elements-recursively animals) | |
10688 | @end group | |
10689 | @end smallexample | |
10690 | ||
10691 | The @code{print-elements-recursively} function first tests whether | |
10692 | there is any content in the list; if there is, the function prints the | |
10693 | first element of the list, the @sc{car} of the list. Then the | |
10694 | function `invokes itself', but gives itself as its argument, not the | |
10695 | whole list, but the second and subsequent elements of the list, the | |
10696 | @sc{cdr} of the list. | |
10697 | ||
10698 | Put another way, if the list is not empty, the function invokes | |
10699 | another instance of code that is similar to the initial code, but is a | |
10700 | different thread of execution, with different arguments than the first | |
10701 | instance. | |
10702 | ||
10703 | Put in yet another way, if the list is not empty, the first robot | |
10704 | assemblies a second robot and tells it what to do; the second robot is | |
10705 | a different individual from the first, but is the same model. | |
10706 | ||
10707 | When the second evaluation occurs, the @code{if} expression is | |
10708 | evaluated and if true, prints the first element of the list it | |
10709 | receives as its argument (which is the second element of the original | |
10710 | list). Then the function `calls itself' with the @sc{cdr} of the list | |
10711 | it is invoked with, which (the second time around) is the @sc{cdr} of | |
10712 | the @sc{cdr} of the original list. | |
10713 | ||
10714 | Note that although we say that the function `calls itself', what we | |
10715 | mean is that the Lisp interpreter assembles and instructs a new | |
10716 | instance of the program. The new instance is a clone of the first, | |
10717 | but is a separate individual. | |
10718 | ||
10719 | Each time the function `invokes itself', it invokes itself on a | |
10720 | shorter version of the original list. It creates a new instance that | |
10721 | works on a shorter list. | |
10722 | ||
10723 | Eventually, the function invokes itself on an empty list. It creates | |
10724 | a new instance whose argument is @code{nil}. The conditional expression | |
10725 | tests the value of @code{list}. Since the value of @code{list} is | |
10726 | @code{nil}, the @code{if} expression tests false so the then-part is | |
10727 | not evaluated. The function as a whole then returns @code{nil}. | |
10728 | ||
10729 | @need 1200 | |
10730 | When you evaluate @code{(print-elements-recursively animals)} in the | |
10731 | @file{*scratch*} buffer, you see this result: | |
10732 | ||
10733 | @smallexample | |
10734 | @group | |
10735 | giraffe | |
10736 | ||
10737 | gazelle | |
10738 | ||
10739 | lion | |
10740 | ||
10741 | tiger | |
10742 | nil | |
10743 | @end group | |
10744 | @end smallexample | |
10745 | ||
10746 | @node Recursive triangle function, Recursion with cond, Recursion with list, Recursion | |
10747 | @comment node-name, next, previous, up | |
10748 | @subsection Recursion in Place of a Counter | |
10749 | @findex triangle-recursively | |
10750 | ||
10751 | @need 1200 | |
10752 | The @code{triangle} function described in a previous section can also | |
10753 | be written recursively. It looks like this: | |
10754 | ||
10755 | @smallexample | |
10756 | @group | |
10757 | (defun triangle-recursively (number) | |
10758 | "Return the sum of the numbers 1 through NUMBER inclusive. | |
10759 | Uses recursion." | |
10760 | (if (= number 1) ; @r{do-again-test} | |
10761 | 1 ; @r{then-part} | |
10762 | (+ number ; @r{else-part} | |
10763 | (triangle-recursively ; @r{recursive call} | |
10764 | (1- number))))) ; @r{next-step-expression} | |
10765 | ||
10766 | (triangle-recursively 7) | |
10767 | @end group | |
10768 | @end smallexample | |
10769 | ||
10770 | @noindent | |
10771 | You can install this function by evaluating it and then try it by | |
10772 | evaluating @code{(triangle-recursively 7)}. (Remember to put your | |
10773 | cursor immediately after the last parenthesis of the function | |
10774 | definition, before the comment.) The function evaluates to 28. | |
10775 | ||
10776 | To understand how this function works, let's consider what happens in the | |
10777 | various cases when the function is passed 1, 2, 3, or 4 as the value of | |
10778 | its argument. | |
10779 | ||
10780 | @menu | |
10781 | * Recursive Example arg of 1 or 2:: | |
10782 | * Recursive Example arg of 3 or 4:: | |
10783 | @end menu | |
10784 | ||
10785 | @node Recursive Example arg of 1 or 2, Recursive Example arg of 3 or 4, Recursive triangle function, Recursive triangle function | |
10786 | @ifnottex | |
10787 | @unnumberedsubsubsec An argument of 1 or 2 | |
10788 | @end ifnottex | |
10789 | ||
10790 | First, what happens if the value of the argument is 1? | |
10791 | ||
10792 | The function has an @code{if} expression after the documentation | |
10793 | string. It tests whether the value of @code{number} is equal to 1; if | |
10794 | so, Emacs evaluates the then-part of the @code{if} expression, which | |
10795 | returns the number 1 as the value of the function. (A triangle with | |
10796 | one row has one pebble in it.) | |
10797 | ||
10798 | Suppose, however, that the value of the argument is 2. In this case, | |
10799 | Emacs evaluates the else-part of the @code{if} expression. | |
10800 | ||
10801 | @need 1200 | |
10802 | The else-part consists of an addition, the recursive call to | |
10803 | @code{triangle-recursively} and a decrementing action; and it looks like | |
10804 | this: | |
10805 | ||
10806 | @smallexample | |
10807 | (+ number (triangle-recursively (1- number))) | |
10808 | @end smallexample | |
10809 | ||
10810 | When Emacs evaluates this expression, the innermost expression is | |
10811 | evaluated first; then the other parts in sequence. Here are the steps | |
10812 | in detail: | |
10813 | ||
10814 | @table @i | |
10815 | @item Step 1 @w{ } Evaluate the innermost expression. | |
10816 | ||
10817 | The innermost expression is @code{(1- number)} so Emacs decrements the | |
10818 | value of @code{number} from 2 to 1. | |
10819 | ||
10820 | @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function. | |
10821 | ||
10822 | The Lisp interpreter creates an individual instance of | |
10823 | @code{triangle-recursively}. It does not matter that this function is | |
10824 | contained within itself. Emacs passes the result Step 1 as the | |
10825 | argument used by this instance of the @code{triangle-recursively} | |
10826 | function | |
10827 | ||
10828 | In this case, Emacs evaluates @code{triangle-recursively} with an | |
10829 | argument of 1. This means that this evaluation of | |
10830 | @code{triangle-recursively} returns 1. | |
10831 | ||
10832 | @item Step 3 @w{ } Evaluate the value of @code{number}. | |
10833 | ||
10834 | The variable @code{number} is the second element of the list that | |
10835 | starts with @code{+}; its value is 2. | |
10836 | ||
10837 | @item Step 4 @w{ } Evaluate the @code{+} expression. | |
10838 | ||
10839 | The @code{+} expression receives two arguments, the first | |
10840 | from the evaluation of @code{number} (Step 3) and the second from the | |
10841 | evaluation of @code{triangle-recursively} (Step 2). | |
10842 | ||
10843 | The result of the addition is the sum of 2 plus 1, and the number 3 is | |
10844 | returned, which is correct. A triangle with two rows has three | |
10845 | pebbles in it. | |
10846 | @end table | |
10847 | ||
10848 | @node Recursive Example arg of 3 or 4, , Recursive Example arg of 1 or 2, Recursive triangle function | |
10849 | @unnumberedsubsubsec An argument of 3 or 4 | |
10850 | ||
10851 | Suppose that @code{triangle-recursively} is called with an argument of | |
10852 | 3. | |
10853 | ||
10854 | @table @i | |
10855 | @item Step 1 @w{ } Evaluate the do-again-test. | |
10856 | ||
10857 | The @code{if} expression is evaluated first. This is the do-again | |
10858 | test and returns false, so the else-part of the @code{if} expression | |
10859 | is evaluated. (Note that in this example, the do-again-test causes | |
10860 | the function to call itself when it tests false, not when it tests | |
10861 | true.) | |
10862 | ||
10863 | @item Step 2 @w{ } Evaluate the innermost expression of the else-part. | |
10864 | ||
10865 | The innermost expression of the else-part is evaluated, which decrements | |
10866 | 3 to 2. This is the next-step-expression. | |
10867 | ||
10868 | @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function. | |
10869 | ||
10870 | The number 2 is passed to the @code{triangle-recursively} function. | |
10871 | ||
10872 | We know what happens when Emacs evaluates @code{triangle-recursively} with | |
10873 | an argument of 2. After going through the sequence of actions described | |
10874 | earlier, it returns a value of 3. So that is what will happen here. | |
10875 | ||
10876 | @item Step 4 @w{ } Evaluate the addition. | |
10877 | ||
10878 | 3 will be passed as an argument to the addition and will be added to the | |
10879 | number with which the function was called, which is 3. | |
10880 | @end table | |
10881 | ||
10882 | @noindent | |
10883 | The value returned by the function as a whole will be 6. | |
10884 | ||
10885 | Now that we know what will happen when @code{triangle-recursively} is | |
10886 | called with an argument of 3, it is evident what will happen if it is | |
10887 | called with an argument of 4: | |
10888 | ||
10889 | @quotation | |
10890 | @need 800 | |
10891 | In the recursive call, the evaluation of | |
10892 | ||
10893 | @smallexample | |
10894 | (triangle-recursively (1- 4)) | |
10895 | @end smallexample | |
10896 | ||
10897 | @need 800 | |
10898 | @noindent | |
10899 | will return the value of evaluating | |
10900 | ||
10901 | @smallexample | |
10902 | (triangle-recursively 3) | |
10903 | @end smallexample | |
10904 | ||
10905 | @noindent | |
10906 | which is 6 and this value will be added to 4 by the addition in the | |
10907 | third line. | |
10908 | @end quotation | |
10909 | ||
10910 | @noindent | |
10911 | The value returned by the function as a whole will be 10. | |
10912 | ||
10913 | Each time @code{triangle-recursively} is evaluated, it evaluates a | |
10914 | version of itself---a different instance of itself---with a smaller | |
10915 | argument, until the argument is small enough so that it does not | |
10916 | evaluate itself. | |
10917 | ||
10918 | Note that this particular design for a recursive function | |
10919 | requires that operations be deferred. | |
10920 | ||
10921 | Before @code{(triangle-recursively 7)} can calculate its answer, it | |
10922 | must call @code{(triangle-recursively 6)}; and before | |
10923 | @code{(triangle-recursively 6)} can calculate its answer, it must call | |
10924 | @code{(triangle-recursively 5)}; and so on. That is to say, the | |
10925 | calculation that @code{(triangle-recursively 7)} makes must be | |
10926 | deferred until @code{(triangle-recursively 6)} makes its calculation; | |
10927 | and @code{(triangle-recursively 6)} must defer until | |
10928 | @code{(triangle-recursively 5)} completes; and so on. | |
10929 | ||
10930 | If each of these instances of @code{triangle-recursively} are thought | |
10931 | of as different robots, the first robot must wait for the second to | |
10932 | complete its job, which must wait until the third completes, and so | |
10933 | on. | |
10934 | ||
10935 | There is a way around this kind of waiting, which we will discuss in | |
10936 | @ref{No Deferment, , Recursion without Deferments}. | |
10937 | ||
10938 | @node Recursion with cond, Recursive Patterns, Recursive triangle function, Recursion | |
10939 | @comment node-name, next, previous, up | |
10940 | @subsection Recursion Example Using @code{cond} | |
10941 | @findex cond | |
10942 | ||
10943 | The version of @code{triangle-recursively} described earlier is written | |
10944 | with the @code{if} special form. It can also be written using another | |
10945 | special form called @code{cond}. The name of the special form | |
10946 | @code{cond} is an abbreviation of the word @samp{conditional}. | |
10947 | ||
10948 | Although the @code{cond} special form is not used as often in the | |
10949 | Emacs Lisp sources as @code{if}, it is used often enough to justify | |
10950 | explaining it. | |
10951 | ||
10952 | @need 800 | |
10953 | The template for a @code{cond} expression looks like this: | |
10954 | ||
10955 | @smallexample | |
10956 | @group | |
10957 | (cond | |
10958 | @var{body}@dots{}) | |
10959 | @end group | |
10960 | @end smallexample | |
10961 | ||
10962 | @noindent | |
10963 | where the @var{body} is a series of lists. | |
10964 | ||
10965 | @need 800 | |
10966 | Written out more fully, the template looks like this: | |
10967 | ||
10968 | @smallexample | |
10969 | @group | |
10970 | (cond | |
10971 | (@var{first-true-or-false-test} @var{first-consequent}) | |
10972 | (@var{second-true-or-false-test} @var{second-consequent}) | |
10973 | (@var{third-true-or-false-test} @var{third-consequent}) | |
10974 | @dots{}) | |
10975 | @end group | |
10976 | @end smallexample | |
10977 | ||
10978 | When the Lisp interpreter evaluates the @code{cond} expression, it | |
10979 | evaluates the first element (the @sc{car} or true-or-false-test) of | |
10980 | the first expression in a series of expressions within the body of the | |
10981 | @code{cond}. | |
10982 | ||
10983 | If the true-or-false-test returns @code{nil} the rest of that | |
10984 | expression, the consequent, is skipped and the true-or-false-test of the | |
10985 | next expression is evaluated. When an expression is found whose | |
10986 | true-or-false-test returns a value that is not @code{nil}, the | |
10987 | consequent of that expression is evaluated. The consequent can be one | |
10988 | or more expressions. If the consequent consists of more than one | |
10989 | expression, the expressions are evaluated in sequence and the value of | |
10990 | the last one is returned. If the expression does not have a consequent, | |
10991 | the value of the true-or-false-test is returned. | |
10992 | ||
10993 | If none of the true-or-false-tests test true, the @code{cond} expression | |
10994 | returns @code{nil}. | |
10995 | ||
10996 | @need 1250 | |
10997 | Written using @code{cond}, the @code{triangle} function looks like this: | |
10998 | ||
10999 | @smallexample | |
11000 | @group | |
11001 | (defun triangle-using-cond (number) | |
11002 | (cond ((<= number 0) 0) | |
11003 | ((= number 1) 1) | |
11004 | ((> number 1) | |
11005 | (+ number (triangle-using-cond (1- number)))))) | |
11006 | @end group | |
11007 | @end smallexample | |
11008 | ||
11009 | @noindent | |
11010 | In this example, the @code{cond} returns 0 if the number is less than or | |
11011 | equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+ | |
11012 | number (triangle-using-cond (1- number)))} if the number is greater than | |
11013 | 1. | |
11014 | ||
11015 | @node Recursive Patterns, No Deferment, Recursion with cond, Recursion | |
11016 | @comment node-name, next, previous, up | |
11017 | @subsection Recursive Patterns | |
11018 | @cindex Recursive Patterns | |
11019 | ||
11020 | Here are three common recursive patterns. Each involves a list. | |
11021 | Recursion does not need to involve lists, but Lisp is designed for lists | |
11022 | and this provides a sense of its primal capabilities. | |
11023 | ||
11024 | @menu | |
11025 | * Every:: | |
11026 | * Accumulate:: | |
11027 | * Keep:: | |
11028 | @end menu | |
11029 | ||
11030 | @node Every, Accumulate, Recursive Patterns, Recursive Patterns | |
11031 | @comment node-name, next, previous, up | |
11032 | @unnumberedsubsubsec Recursive Pattern: @emph{every} | |
11033 | @cindex Every, type of recursive pattern | |
11034 | @cindex Recursive pattern: every | |
11035 | ||
11036 | In the @code{every} recursive pattern, an action is performed on every | |
11037 | element of a list. | |
11038 | ||
11039 | @need 1500 | |
11040 | The basic pattern is: | |
11041 | ||
11042 | @itemize @bullet | |
11043 | @item | |
11044 | If a list be empty, return @code{nil}. | |
11045 | @item | |
11046 | Else, act on the beginning of the list (the @sc{car} of the list) | |
11047 | @itemize @minus | |
11048 | @item | |
11049 | through a recursive call by the function on the rest (the | |
11050 | @sc{cdr}) of the list, | |
11051 | @item | |
11052 | and, optionally, combine the acted-on element, using @code{cons}, | |
11053 | with the results of acting on the rest. | |
11054 | @end itemize | |
11055 | @end itemize | |
11056 | ||
11057 | @need 1500 | |
11058 | Here is example: | |
11059 | ||
11060 | @smallexample | |
11061 | @group | |
11062 | (defun square-each (numbers-list) | |
11063 | "Square each of a NUMBERS LIST, recursively." | |
11064 | (if (not numbers-list) ; do-again-test | |
11065 | nil | |
11066 | (cons | |
11067 | (* (car numbers-list) (car numbers-list)) | |
11068 | (square-each (cdr numbers-list))))) ; next-step-expression | |
11069 | @end group | |
11070 | ||
11071 | @group | |
11072 | (square-each '(1 2 3)) | |
11073 | @result{} (1 4 9) | |
11074 | @end group | |
11075 | @end smallexample | |
11076 | ||
11077 | @need 1200 | |
11078 | @noindent | |
11079 | If @code{numbers-list} is empty, do nothing. But if it has content, | |
11080 | construct a list combining the square of the first number in the list | |
11081 | with the result of the recursive call. | |
11082 | ||
11083 | (The example follows the pattern exactly: @code{nil} is returned if | |
11084 | the numbers' list is empty. In practice, you would write the | |
11085 | conditional so it carries out the action when the numbers' list is not | |
11086 | empty.) | |
11087 | ||
11088 | The @code{print-elements-recursively} function (@pxref{Recursion with | |
11089 | list, , Recursion with a List}) is another example of an @code{every} | |
11090 | pattern, except in this case, rather than bring the results together | |
11091 | using @code{cons}, we print each element of output. | |
11092 | ||
11093 | @need 1250 | |
11094 | The @code{print-elements-recursively} function looks like this: | |
11095 | ||
11096 | @smallexample | |
11097 | @group | |
11098 | (setq animals '(gazelle giraffe lion tiger)) | |
11099 | @end group | |
11100 | ||
11101 | @group | |
11102 | (defun print-elements-recursively (list) | |
11103 | "Print each element of LIST on a line of its own. | |
11104 | Uses recursion." | |
11105 | (if list ; @r{do-again-test} | |
11106 | (progn | |
11107 | (print (car list)) ; @r{body} | |
11108 | (print-elements-recursively ; @r{recursive call} | |
11109 | (cdr list))))) ; @r{next-step-expression} | |
11110 | ||
11111 | (print-elements-recursively animals) | |
11112 | @end group | |
11113 | @end smallexample | |
11114 | ||
11115 | @need 1500 | |
11116 | The pattern for @code{print-elements-recursively} is: | |
11117 | ||
11118 | @itemize @bullet | |
11119 | @item | |
11120 | If the list be empty, do nothing. | |
11121 | @item | |
11122 | But if the list has at least one element, | |
11123 | @itemize @minus | |
11124 | @item | |
11125 | act on the beginning of the list (the @sc{car} of the list), | |
11126 | @item | |
11127 | and make a recursive call on the rest (the @sc{cdr}) of the list. | |
11128 | @end itemize | |
11129 | @end itemize | |
11130 | ||
11131 | @node Accumulate, Keep, Every, Recursive Patterns | |
11132 | @comment node-name, next, previous, up | |
11133 | @unnumberedsubsubsec Recursive Pattern: @emph{accumulate} | |
11134 | @cindex Accumulate, type of recursive pattern | |
11135 | @cindex Recursive pattern: accumulate | |
11136 | ||
11137 | Another recursive pattern is called the @code{accumulate} pattern. In | |
11138 | the @code{accumulate} recursive pattern, an action is performed on | |
11139 | every element of a list and the result of that action is accumulated | |
11140 | with the results of performing the action on the other elements. | |
11141 | ||
11142 | This is very like the `every' pattern using @code{cons}, except that | |
11143 | @code{cons} is not used, but some other combiner. | |
11144 | ||
11145 | @need 1500 | |
11146 | The pattern is: | |
11147 | ||
11148 | @itemize @bullet | |
11149 | @item | |
11150 | If a list be empty, return zero or some other constant. | |
11151 | @item | |
11152 | Else, act on the beginning of the list (the @sc{car} of the list), | |
11153 | @itemize @minus | |
11154 | @item | |
11155 | and combine that acted-on element, using @code{+} or | |
11156 | some other combining function, with | |
11157 | @item | |
11158 | a recursive call by the function on the rest (the @sc{cdr}) of the list. | |
11159 | @end itemize | |
11160 | @end itemize | |
11161 | ||
11162 | @need 1500 | |
11163 | Here is an example: | |
11164 | ||
11165 | @smallexample | |
11166 | @group | |
11167 | (defun add-elements (numbers-list) | |
11168 | "Add the elements of NUMBERS-LIST together." | |
11169 | (if (not numbers-list) | |
11170 | 0 | |
11171 | (+ (car numbers-list) (add-elements (cdr numbers-list))))) | |
11172 | @end group | |
11173 | ||
11174 | @group | |
11175 | (add-elements '(1 2 3 4)) | |
11176 | @result{} 10 | |
11177 | @end group | |
11178 | @end smallexample | |
11179 | ||
11180 | @xref{Files List, , Making a List of Files}, for an example of the | |
11181 | accumulate pattern. | |
11182 | ||
11183 | @node Keep, , Accumulate, Recursive Patterns | |
11184 | @comment node-name, next, previous, up | |
11185 | @unnumberedsubsubsec Recursive Pattern: @emph{keep} | |
11186 | @cindex Keep, type of recursive pattern | |
11187 | @cindex Recursive pattern: keep | |
11188 | ||
11189 | A third recursive pattern is called the @code{keep} pattern. | |
11190 | In the @code{keep} recursive pattern, each element of a list is tested; | |
11191 | the element is acted on and the results are kept only if the element | |
11192 | meets a criterion. | |
11193 | ||
11194 | Again, this is very like the `every' pattern, except the element is | |
11195 | skipped unless it meets a criterion. | |
11196 | ||
11197 | @need 1500 | |
11198 | The pattern has three parts: | |
11199 | ||
11200 | @itemize @bullet | |
11201 | @item | |
11202 | If a list be empty, return @code{nil}. | |
11203 | @item | |
11204 | Else, if the beginning of the list (the @sc{car} of the list) passes | |
11205 | a test | |
11206 | @itemize @minus | |
11207 | @item | |
11208 | act on that element and combine it, using @code{cons} with | |
11209 | @item | |
11210 | a recursive call by the function on the rest (the @sc{cdr}) of the list. | |
11211 | @end itemize | |
11212 | @item | |
11213 | Otherwise, if the beginning of the list (the @sc{car} of the list) fails | |
11214 | the test | |
11215 | @itemize @minus | |
11216 | @item | |
11217 | skip on that element, | |
11218 | @item | |
11219 | and, recursively call the function on the rest (the @sc{cdr}) of the list. | |
11220 | @end itemize | |
11221 | @end itemize | |
11222 | ||
11223 | @need 1500 | |
11224 | Here is an example that uses @code{cond}: | |
11225 | ||
11226 | @smallexample | |
11227 | @group | |
11228 | (defun keep-three-letter-words (word-list) | |
11229 | "Keep three letter words in WORD-LIST." | |
11230 | (cond | |
11231 | ;; First do-again-test: stop-condition | |
11232 | ((not word-list) nil) | |
11233 | ||
11234 | ;; Second do-again-test: when to act | |
11235 | ((eq 3 (length (symbol-name (car word-list)))) | |
11236 | ;; combine acted-on element with recursive call on shorter list | |
11237 | (cons (car word-list) (keep-three-letter-words (cdr word-list)))) | |
11238 | ||
11239 | ;; Third do-again-test: when to skip element; | |
11240 | ;; recursively call shorter list with next-step expression | |
11241 | (t (keep-three-letter-words (cdr word-list))))) | |
11242 | @end group | |
11243 | ||
11244 | @group | |
11245 | (keep-three-letter-words '(one two three four five six)) | |
11246 | @result{} (one two six) | |
11247 | @end group | |
11248 | @end smallexample | |
11249 | ||
11250 | It goes without saying that you need not use @code{nil} as the test for | |
11251 | when to stop; and you can, of course, combine these patterns. | |
11252 | ||
11253 | @node No Deferment, No deferment solution, Recursive Patterns, Recursion | |
11254 | @subsection Recursion without Deferments | |
11255 | @cindex Deferment in recursion | |
11256 | @cindex Recursion without Deferments | |
11257 | ||
11258 | Let's consider again what happens with the @code{triangle-recursively} | |
11259 | function. We will find that the intermediate calculations are | |
11260 | deferred until all can be done. | |
11261 | ||
11262 | @need 800 | |
11263 | Here is the function definition: | |
11264 | ||
11265 | @smallexample | |
11266 | @group | |
11267 | (defun triangle-recursively (number) | |
11268 | "Return the sum of the numbers 1 through NUMBER inclusive. | |
11269 | Uses recursion." | |
11270 | (if (= number 1) ; @r{do-again-test} | |
11271 | 1 ; @r{then-part} | |
11272 | (+ number ; @r{else-part} | |
11273 | (triangle-recursively ; @r{recursive call} | |
11274 | (1- number))))) ; @r{next-step-expression} | |
11275 | @end group | |
11276 | @end smallexample | |
11277 | ||
11278 | What happens when we call this function with a argument of 7? | |
11279 | ||
11280 | The first instance of the @code{triangle-recursively} function adds | |
11281 | the number 7 to the value returned by a second instance of | |
11282 | @code{triangle-recursively}, an instance that has been passed an | |
11283 | argument of 6. That is to say, the first calculation is: | |
11284 | ||
11285 | @smallexample | |
11286 | (+ 7 (triangle-recursively 6) | |
11287 | @end smallexample | |
11288 | ||
11289 | @noindent | |
11290 | The first instance of @code{triangle-recursively}---you may want to | |
11291 | think of it as a little robot---cannot complete its job. It must hand | |
11292 | off the calculation for @code{(triangle-recursively 6)} to a second | |
11293 | instance of the program, to a second robot. This second individual is | |
11294 | completely different from the first one; it is, in the jargon, a | |
11295 | `different instantiation'. Or, put another way, it is a different | |
11296 | robot. It is the same model as the first; it calculates triangle | |
11297 | numbers recursively; but it has a different serial number. | |
11298 | ||
11299 | And what does @code{(triangle-recursively 6)} return? It returns the | |
11300 | number 6 added to the value returned by evaluating | |
11301 | @code{triangle-recursively} with an argument of 5. Using the robot | |
11302 | metaphor, it asks yet another robot to help it. | |
11303 | ||
11304 | @need 800 | |
11305 | Now the total is: | |
11306 | ||
11307 | @smallexample | |
11308 | (+ 7 6 (triangle-recursively 5) | |
11309 | @end smallexample | |
11310 | ||
11311 | @need 800 | |
11312 | And what happens next? | |
11313 | ||
11314 | @smallexample | |
11315 | (+ 7 6 5 (triangle-recursively 4) | |
11316 | @end smallexample | |
11317 | ||
11318 | Each time @code{triangle-recursively} is called, except for the last | |
11319 | time, it creates another instance of the program---another robot---and | |
11320 | asks it to make a calculation. | |
11321 | ||
11322 | @need 800 | |
11323 | Eventually, the full addition is set up and performed: | |
11324 | ||
11325 | @smallexample | |
11326 | (+ 7 6 5 4 3 2 1) | |
11327 | @end smallexample | |
11328 | ||
11329 | This design for the function defers the calculation of the first step | |
11330 | until the second can be done, and defers that until the third can be | |
11331 | done, and so on. Each deferment means the computer must remember what | |
11332 | is being waited on. This is not a problem when there are only a few | |
11333 | steps, as in this example. But it can be a problem when there are | |
11334 | more steps. | |
11335 | ||
11336 | @node No deferment solution, , No Deferment, Recursion | |
11337 | @subsection No Deferment Solution | |
11338 | @cindex No deferment solution | |
11339 | @cindex Defermentless solution | |
11340 | @cindex Solution without deferment | |
11341 | ||
11342 | The solution to the problem of deferred operations is to write in a | |
11343 | manner that does not defer operations@footnote{The phrase @dfn{tail | |
11344 | recursive} is used to describe such a process, one that uses | |
11345 | `constant space'.}. This requires | |
11346 | writing to a different pattern, often one that involves writing two | |
11347 | function definitions, an `initialization' function and a `helper' | |
11348 | function. | |
11349 | ||
11350 | The `initialization' function sets up the job; the `helper' function | |
11351 | does the work. | |
11352 | ||
11353 | @need 1200 | |
11354 | Here are the two function definitions for adding up numbers. They are | |
11355 | so simple, I find them hard to understand. | |
11356 | ||
11357 | @smallexample | |
11358 | @group | |
11359 | (defun triangle-initialization (number) | |
11360 | "Return the sum of the numbers 1 through NUMBER inclusive. | |
11361 | This is the `initialization' component of a two function | |
11362 | duo that uses recursion." | |
11363 | (triangle-recursive-helper 0 0 number)) | |
11364 | @end group | |
11365 | @end smallexample | |
11366 | ||
11367 | @smallexample | |
11368 | @group | |
11369 | (defun triangle-recursive-helper (sum counter number) | |
11370 | "Return SUM, using COUNTER, through NUMBER inclusive. | |
11371 | This is the `helper' component of a two function duo | |
11372 | that uses recursion." | |
11373 | (if (> counter number) | |
11374 | sum | |
11375 | (triangle-recursive-helper (+ sum counter) ; @r{sum} | |
11376 | (1+ counter) ; @r{counter} | |
11377 | number))) ; @r{number} | |
11378 | @end group | |
11379 | @end smallexample | |
11380 | ||
11381 | @need 1250 | |
11382 | Install both function definitions by evaluating them, then call | |
11383 | @code{triangle-initialization} with 2 rows: | |
11384 | ||
11385 | @smallexample | |
11386 | @group | |
11387 | (triangle-initialization 2) | |
11388 | @result{} 3 | |
11389 | @end group | |
11390 | @end smallexample | |
11391 | ||
11392 | The `initialization' function calls the first instance of the `helper' | |
11393 | function with three arguments: zero, zero, and a number which is the | |
11394 | number of rows in the triangle. | |
11395 | ||
11396 | The first two arguments passed to the `helper' function are | |
11397 | initialization values. These values are changed when | |
11398 | @code{triangle-recursive-helper} invokes new instances.@footnote{The | |
11399 | jargon is mildly confusing: @code{triangle-recursive-helper} uses a | |
11400 | process that is iterative in a procedure that is recursive. The | |
11401 | process is called iterative because the computer need only record the | |
11402 | three values, @code{sum}, @code{counter}, and @code{number}; the | |
11403 | procedure is recursive because the function `calls itself'. On the | |
11404 | other hand, both the process and the procedure used by | |
11405 | @code{triangle-recursively} are called recursive. The word | |
11406 | `recursive' has different meanings in the two contexts.} | |
11407 | ||
11408 | Let's see what happens when we have a triangle that has one row. (This | |
11409 | triangle will have one pebble in it!) | |
11410 | ||
11411 | @need 1200 | |
11412 | @code{triangle-initialization} will call its helper with | |
11413 | the arguments @w{@code{0 0 1}}. That function will run the conditional | |
11414 | test whether @code{(> counter number)}: | |
11415 | ||
11416 | @smallexample | |
11417 | (> 0 1) | |
11418 | @end smallexample | |
11419 | ||
11420 | @need 1200 | |
11421 | @noindent | |
11422 | and find that the result is false, so it will invoke | |
11423 | the then-part of the @code{if} clause: | |
11424 | ||
11425 | @smallexample | |
11426 | @group | |
11427 | (triangle-recursive-helper | |
11428 | (+ sum counter) ; @r{sum plus counter} @result{} @r{sum} | |
11429 | (1+ counter) ; @r{increment counter} @result{} @r{counter} | |
11430 | number) ; @r{number stays the same} | |
11431 | @end group | |
11432 | @end smallexample | |
11433 | ||
11434 | @need 800 | |
11435 | @noindent | |
11436 | which will first compute: | |
11437 | ||
11438 | @smallexample | |
11439 | @group | |
11440 | (triangle-recursive-helper (+ 0 0) ; @r{sum} | |
11441 | (1+ 0) ; @r{counter} | |
11442 | 1) ; @r{number} | |
11443 | @exdent which is: | |
11444 | ||
11445 | (triangle-recursive-helper 0 1 1) | |
11446 | @end group | |
11447 | @end smallexample | |
11448 | ||
11449 | Again, @code{(> counter number)} will be false, so again, the Lisp | |
11450 | interpreter will evaluate @code{triangle-recursive-helper}, creating a | |
11451 | new instance with new arguments. | |
11452 | ||
11453 | @need 800 | |
11454 | This new instance will be; | |
11455 | ||
11456 | @smallexample | |
11457 | @group | |
11458 | (triangle-recursive-helper | |
11459 | (+ sum counter) ; @r{sum plus counter} @result{} @r{sum} | |
11460 | (1+ counter) ; @r{increment counter} @result{} @r{counter} | |
11461 | number) ; @r{number stays the same} | |
11462 | ||
11463 | @exdent which is: | |
11464 | ||
11465 | (triangle-recursive-helper 1 2 1) | |
11466 | @end group | |
11467 | @end smallexample | |
11468 | ||
11469 | In this case, the @code{(> counter number)} test will be true! So the | |
11470 | instance will return the value of the sum, which will be 1, as | |
11471 | expected. | |
11472 | ||
11473 | Now, let's pass @code{triangle-initialization} an argument | |
11474 | of 2, to find out how many pebbles there are in a triangle with two rows. | |
11475 | ||
11476 | That function calls @code{(triangle-recursive-helper 0 0 2)}. | |
11477 | ||
11478 | @need 800 | |
11479 | In stages, the instances called will be: | |
11480 | ||
11481 | @smallexample | |
11482 | @group | |
11483 | @r{sum counter number} | |
11484 | (triangle-recursive-helper 0 1 2) | |
11485 | ||
11486 | (triangle-recursive-helper 1 2 2) | |
11487 | ||
11488 | (triangle-recursive-helper 3 3 2) | |
11489 | @end group | |
11490 | @end smallexample | |
11491 | ||
11492 | When the last instance is called, the @code{(> counter number)} test | |
11493 | will be true, so the instance will return the value of @code{sum}, | |
11494 | which will be 3. | |
11495 | ||
11496 | This kind of pattern helps when you are writing functions that can use | |
11497 | many resources in a computer. | |
11498 | ||
11499 | @need 1500 | |
11500 | @node Looping exercise, , Recursion, Loops & Recursion | |
11501 | @section Looping Exercise | |
11502 | ||
11503 | @itemize @bullet | |
11504 | @item | |
11505 | Write a function similar to @code{triangle} in which each row has a | |
11506 | value which is the square of the row number. Use a @code{while} loop. | |
11507 | ||
11508 | @item | |
11509 | Write a function similar to @code{triangle} that multiplies instead of | |
11510 | adds the values. | |
11511 | ||
11512 | @item | |
11513 | Rewrite these two functions recursively. Rewrite these functions | |
11514 | using @code{cond}. | |
11515 | ||
11516 | @c comma in printed title causes problem in Info cross reference | |
11517 | @item | |
11518 | Write a function for Texinfo mode that creates an index entry at the | |
11519 | beginning of a paragraph for every @samp{@@dfn} within the paragraph. | |
11520 | (In a Texinfo file, @samp{@@dfn} marks a definition. For more | |
11521 | information, see | |
11522 | @ifinfo | |
11523 | @ref{Indicating, , Indicating Definitions, texinfo}.) | |
11524 | @end ifinfo | |
11525 | @ifhtml | |
11526 | @ref{Indicating, , Indicating, texinfo, Texinfo Manual}.) | |
11527 | @end ifhtml | |
11528 | @iftex | |
11529 | ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU | |
11530 | Documentation Format}.) | |
11531 | @end iftex | |
11532 | @end itemize | |
11533 | ||
11534 | @node Regexp Search, Counting Words, Loops & Recursion, Top | |
11535 | @comment node-name, next, previous, up | |
11536 | @chapter Regular Expression Searches | |
11537 | @cindex Searches, illustrating | |
11538 | @cindex Regular expression searches | |
11539 | @cindex Patterns, searching for | |
11540 | @cindex Motion by sentence and paragraph | |
11541 | @cindex Sentences, movement by | |
11542 | @cindex Paragraphs, movement by | |
11543 | ||
11544 | Regular expression searches are used extensively in GNU Emacs. The | |
11545 | two functions, @code{forward-sentence} and @code{forward-paragraph}, | |
11546 | illustrate these searches well. They use regular expressions to find | |
11547 | where to move point. The phrase `regular expression' is often written | |
11548 | as `regexp'. | |
11549 | ||
11550 | Regular expression searches are described in @ref{Regexp Search, , | |
11551 | Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in | |
11552 | @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference | |
11553 | Manual}. In writing this chapter, I am presuming that you have at | |
11554 | least a mild acquaintance with them. The major point to remember is | |
11555 | that regular expressions permit you to search for patterns as well as | |
11556 | for literal strings of characters. For example, the code in | |
11557 | @code{forward-sentence} searches for the pattern of possible | |
11558 | characters that could mark the end of a sentence, and moves point to | |
11559 | that spot. | |
11560 | ||
11561 | Before looking at the code for the @code{forward-sentence} function, it | |
11562 | is worth considering what the pattern that marks the end of a sentence | |
11563 | must be. The pattern is discussed in the next section; following that | |
11564 | is a description of the regular expression search function, | |
11565 | @code{re-search-forward}. The @code{forward-sentence} function | |
11566 | is described in the section following. Finally, the | |
11567 | @code{forward-paragraph} function is described in the last section of | |
11568 | this chapter. @code{forward-paragraph} is a complex function that | |
11569 | introduces several new features. | |
11570 | ||
11571 | @menu | |
11572 | * sentence-end:: The regular expression for @code{sentence-end}. | |
11573 | * re-search-forward:: Very similar to @code{search-forward}. | |
11574 | * forward-sentence:: A straightforward example of regexp search. | |
11575 | * forward-paragraph:: A somewhat complex example. | |
11576 | * etags:: How to create your own @file{TAGS} table. | |
11577 | * Regexp Review:: | |
11578 | * re-search Exercises:: | |
11579 | @end menu | |
11580 | ||
11581 | @node sentence-end, re-search-forward, Regexp Search, Regexp Search | |
11582 | @comment node-name, next, previous, up | |
11583 | @section The Regular Expression for @code{sentence-end} | |
11584 | @findex sentence-end | |
11585 | ||
11586 | The symbol @code{sentence-end} is bound to the pattern that marks the | |
11587 | end of a sentence. What should this regular expression be? | |
11588 | ||
11589 | Clearly, a sentence may be ended by a period, a question mark, or an | |
11590 | exclamation mark. Indeed, only clauses that end with one of those three | |
11591 | characters should be considered the end of a sentence. This means that | |
11592 | the pattern should include the character set: | |
11593 | ||
11594 | @smallexample | |
11595 | [.?!] | |
11596 | @end smallexample | |
11597 | ||
11598 | However, we do not want @code{forward-sentence} merely to jump to a | |
11599 | period, a question mark, or an exclamation mark, because such a character | |
11600 | might be used in the middle of a sentence. A period, for example, is | |
11601 | used after abbreviations. So other information is needed. | |
11602 | ||
11603 | According to convention, you type two spaces after every sentence, but | |
11604 | only one space after a period, a question mark, or an exclamation mark in | |
11605 | the body of a sentence. So a period, a question mark, or an exclamation | |
11606 | mark followed by two spaces is a good indicator of an end of sentence. | |
11607 | However, in a file, the two spaces may instead be a tab or the end of a | |
11608 | line. This means that the regular expression should include these three | |
11609 | items as alternatives. | |
11610 | ||
11611 | @need 800 | |
11612 | This group of alternatives will look like this: | |
11613 | ||
11614 | @smallexample | |
11615 | @group | |
11616 | \\($\\| \\| \\) | |
11617 | ^ ^^ | |
11618 | TAB SPC | |
11619 | @end group | |
11620 | @end smallexample | |
11621 | ||
11622 | @noindent | |
11623 | Here, @samp{$} indicates the end of the line, and I have pointed out | |
11624 | where the tab and two spaces are inserted in the expression. Both are | |
11625 | inserted by putting the actual characters into the expression. | |
11626 | ||
11627 | Two backslashes, @samp{\\}, are required before the parentheses and | |
11628 | vertical bars: the first backslash quotes the following backslash in | |
11629 | Emacs; and the second indicates that the following character, the | |
11630 | parenthesis or the vertical bar, is special. | |
11631 | ||
11632 | @need 1000 | |
11633 | Also, a sentence may be followed by one or more carriage returns, like | |
11634 | this: | |
11635 | ||
11636 | @smallexample | |
11637 | @group | |
11638 | [ | |
11639 | ]* | |
11640 | @end group | |
11641 | @end smallexample | |
11642 | ||
11643 | @noindent | |
11644 | Like tabs and spaces, a carriage return is inserted into a regular | |
11645 | expression by inserting it literally. The asterisk indicates that the | |
11646 | @key{RET} is repeated zero or more times. | |
11647 | ||
11648 | But a sentence end does not consist only of a period, a question mark or | |
11649 | an exclamation mark followed by appropriate space: a closing quotation | |
11650 | mark or a closing brace of some kind may precede the space. Indeed more | |
11651 | than one such mark or brace may precede the space. These require a | |
11652 | expression that looks like this: | |
11653 | ||
11654 | @smallexample | |
11655 | []\"')@}]* | |
11656 | @end smallexample | |
11657 | ||
11658 | In this expression, the first @samp{]} is the first character in the | |
11659 | expression; the second character is @samp{"}, which is preceded by a | |
11660 | @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last | |
11661 | three characters are @samp{'}, @samp{)}, and @samp{@}}. | |
11662 | ||
11663 | All this suggests what the regular expression pattern for matching the | |
11664 | end of a sentence should be; and, indeed, if we evaluate | |
11665 | @code{sentence-end} we find that it returns the following value: | |
11666 | ||
11667 | @smallexample | |
11668 | @group | |
11669 | sentence-end | |
11670 | @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[ | |
11671 | ]*" | |
11672 | @end group | |
11673 | @end smallexample | |
11674 | ||
11675 | @ignore | |
11676 | ||
11677 | @noindent | |
11678 | (Note that here the @key{TAB}, two spaces, and @key{RET} are shown | |
11679 | literally in the pattern.) | |
11680 | ||
11681 | This regular expression can be decyphered as follows: | |
11682 | ||
11683 | @table @code | |
11684 | @item [.?!] | |
11685 | The first part of the pattern is the three characters, a period, a question | |
11686 | mark and an exclamation mark, within square brackets. The pattern must | |
11687 | begin with one or other of these characters. | |
11688 | ||
11689 | @item []\"')@}]* | |
11690 | The second part of the pattern is the group of closing braces and | |
11691 | quotation marks, which can appear zero or more times. These may follow | |
11692 | the period, question mark or exclamation mark. In a regular expression, | |
11693 | the backslash, @samp{\}, followed by the double quotation mark, | |
11694 | @samp{"}, indicates the class of string-quote characters. Usually, the | |
11695 | double quotation mark is the only character in this class. The | |
11696 | asterisk, @samp{*}, indicates that the items in the previous group (the | |
11697 | group surrounded by square brackets, @samp{[]}) may be repeated zero or | |
11698 | more times. | |
11699 | ||
11700 | @item \\($\\| \\| \\) | |
11701 | The third part of the pattern is one or other of: either the end of a | |
11702 | line, or two blank spaces, or a tab. The double back-slashes are used | |
11703 | to prevent Emacs from reading the parentheses and vertical bars as part | |
11704 | of the search pattern; the parentheses are used to mark the group and | |
11705 | the vertical bars are used to indicated that the patterns to either side | |
11706 | of them are alternatives. The dollar sign is used to indicate the end | |
11707 | of a line and both the two spaces and the tab are each inserted as is to | |
11708 | indicate what they are. | |
11709 | ||
11710 | @item [@key{RET}]* | |
11711 | Finally, the last part of the pattern indicates that the end of the line | |
11712 | or the whitespace following the period, question mark or exclamation | |
11713 | mark may, but need not, be followed by one or more carriage returns. In | |
11714 | the pattern, the carriage return is inserted as an actual carriage | |
11715 | return between square brackets but here it is shown as @key{RET}. | |
11716 | @end table | |
11717 | ||
11718 | @end ignore | |
11719 | ||
11720 | @node re-search-forward, forward-sentence, sentence-end, Regexp Search | |
11721 | @comment node-name, next, previous, up | |
11722 | @section The @code{re-search-forward} Function | |
11723 | @findex re-search-forward | |
11724 | ||
11725 | The @code{re-search-forward} function is very like the | |
11726 | @code{search-forward} function. (@xref{search-forward, , The | |
11727 | @code{search-forward} Function}.) | |
11728 | ||
11729 | @code{re-search-forward} searches for a regular expression. If the | |
11730 | search is successful, it leaves point immediately after the last | |
11731 | character in the target. If the search is backwards, it leaves point | |
11732 | just before the first character in the target. You may tell | |
11733 | @code{re-search-forward} to return @code{t} for true. (Moving point | |
11734 | is therefore a `side effect'.) | |
11735 | ||
11736 | Like @code{search-forward}, the @code{re-search-forward} function takes | |
11737 | four arguments: | |
11738 | ||
11739 | @enumerate | |
11740 | @item | |
11741 | The first argument is the regular expression that the function searches | |
11742 | for. The regular expression will be a string between quotations marks. | |
11743 | ||
11744 | @item | |
11745 | The optional second argument limits how far the function will search; it is a | |
11746 | bound, which is specified as a position in the buffer. | |
11747 | ||
11748 | @item | |
11749 | The optional third argument specifies how the function responds to | |
11750 | failure: @code{nil} as the third argument causes the function to | |
11751 | signal an error (and print a message) when the search fails; any other | |
11752 | value causes it to return @code{nil} if the search fails and @code{t} | |
11753 | if the search succeeds. | |
11754 | ||
11755 | @item | |
11756 | The optional fourth argument is the repeat count. A negative repeat | |
11757 | count causes @code{re-search-forward} to search backwards. | |
11758 | @end enumerate | |
11759 | ||
11760 | @need 800 | |
11761 | The template for @code{re-search-forward} looks like this: | |
11762 | ||
11763 | @smallexample | |
11764 | @group | |
11765 | (re-search-forward "@var{regular-expression}" | |
11766 | @var{limit-of-search} | |
11767 | @var{what-to-do-if-search-fails} | |
11768 | @var{repeat-count}) | |
11769 | @end group | |
11770 | @end smallexample | |
11771 | ||
11772 | The second, third, and fourth arguments are optional. However, if you | |
11773 | want to pass a value to either or both of the last two arguments, you | |
11774 | must also pass a value to all the preceding arguments. Otherwise, the | |
11775 | Lisp interpreter will mistake which argument you are passing the value | |
11776 | to. | |
11777 | ||
11778 | @need 1200 | |
11779 | In the @code{forward-sentence} function, the regular expression will be | |
11780 | the value of the variable @code{sentence-end}, namely: | |
11781 | ||
11782 | @smallexample | |
11783 | @group | |
11784 | "[.?!][]\"')@}]*\\($\\| \\| \\)[ | |
11785 | ]*" | |
11786 | @end group | |
11787 | @end smallexample | |
11788 | ||
11789 | @noindent | |
11790 | The limit of the search will be the end of the paragraph (since a | |
11791 | sentence cannot go beyond a paragraph). If the search fails, the | |
11792 | function will return @code{nil}; and the repeat count will be provided | |
11793 | by the argument to the @code{forward-sentence} function. | |
11794 | ||
11795 | @node forward-sentence, forward-paragraph, re-search-forward, Regexp Search | |
11796 | @comment node-name, next, previous, up | |
11797 | @section @code{forward-sentence} | |
11798 | @findex forward-sentence | |
11799 | ||
11800 | The command to move the cursor forward a sentence is a straightforward | |
11801 | illustration of how to use regular expression searches in Emacs Lisp. | |
11802 | Indeed, the function looks longer and more complicated than it is; this | |
11803 | is because the function is designed to go backwards as well as forwards; | |
11804 | and, optionally, over more than one sentence. The function is usually | |
11805 | bound to the key command @kbd{M-e}. | |
11806 | ||
11807 | @menu | |
11808 | * Complete forward-sentence:: | |
11809 | * fwd-sentence while loops:: Two @code{while} loops. | |
11810 | * fwd-sentence re-search:: A regular expression search. | |
11811 | @end menu | |
11812 | ||
11813 | @node Complete forward-sentence, fwd-sentence while loops, forward-sentence, forward-sentence | |
11814 | @ifnottex | |
11815 | @unnumberedsubsec Complete @code{forward-sentence} function definition | |
11816 | @end ifnottex | |
11817 | ||
11818 | @need 1250 | |
11819 | Here is the code for @code{forward-sentence}: | |
11820 | ||
11821 | @smallexample | |
11822 | @group | |
11823 | (defun forward-sentence (&optional arg) | |
11824 | "Move forward to next sentence-end. With argument, repeat. | |
11825 | With negative argument, move backward repeatedly to sentence-beginning. | |
11826 | Sentence ends are identified by the value of sentence-end | |
11827 | treated as a regular expression. Also, every paragraph boundary | |
11828 | terminates sentences as well." | |
11829 | @end group | |
11830 | @group | |
11831 | (interactive "p") | |
11832 | (or arg (setq arg 1)) | |
11833 | (while (< arg 0) | |
11834 | (let ((par-beg | |
11835 | (save-excursion (start-of-paragraph-text) (point)))) | |
11836 | (if (re-search-backward | |
11837 | (concat sentence-end "[^ \t\n]") par-beg t) | |
11838 | (goto-char (1- (match-end 0))) | |
11839 | (goto-char par-beg))) | |
11840 | (setq arg (1+ arg))) | |
11841 | (while (> arg 0) | |
11842 | (let ((par-end | |
11843 | (save-excursion (end-of-paragraph-text) (point)))) | |
11844 | (if (re-search-forward sentence-end par-end t) | |
11845 | (skip-chars-backward " \t\n") | |
11846 | (goto-char par-end))) | |
11847 | (setq arg (1- arg)))) | |
11848 | @end group | |
11849 | @end smallexample | |
11850 | ||
11851 | The function looks long at first sight and it is best to look at its | |
11852 | skeleton first, and then its muscle. The way to see the skeleton is to | |
11853 | look at the expressions that start in the left-most columns: | |
11854 | ||
11855 | @smallexample | |
11856 | @group | |
11857 | (defun forward-sentence (&optional arg) | |
11858 | "@var{documentation}@dots{}" | |
11859 | (interactive "p") | |
11860 | (or arg (setq arg 1)) | |
11861 | (while (< arg 0) | |
11862 | @var{body-of-while-loop} | |
11863 | (while (> arg 0) | |
11864 | @var{body-of-while-loop} | |
11865 | @end group | |
11866 | @end smallexample | |
11867 | ||
11868 | This looks much simpler! The function definition consists of | |
11869 | documentation, an @code{interactive} expression, an @code{or} | |
11870 | expression, and @code{while} loops. | |
11871 | ||
11872 | Let's look at each of these parts in turn. | |
11873 | ||
11874 | We note that the documentation is thorough and understandable. | |
11875 | ||
11876 | The function has an @code{interactive "p"} declaration. This means | |
11877 | that the processed prefix argument, if any, is passed to the | |
11878 | function as its argument. (This will be a number.) If the function | |
11879 | is not passed an argument (it is optional) then the argument | |
11880 | @code{arg} will be bound to 1. When @code{forward-sentence} is called | |
11881 | non-interactively without an argument, @code{arg} is bound to | |
11882 | @code{nil}. | |
11883 | ||
11884 | The @code{or} expression handles the prefix argument. What it does is | |
11885 | either leave the value of @code{arg} as it is, but only if @code{arg} | |
11886 | is bound to a value; or it sets the value of @code{arg} to 1, in the | |
11887 | case when @code{arg} is bound to @code{nil}. | |
11888 | ||
11889 | @node fwd-sentence while loops, fwd-sentence re-search, Complete forward-sentence, forward-sentence | |
11890 | @unnumberedsubsec The @code{while} loops | |
11891 | ||
11892 | Two @code{while} loops follow the @code{or} expression. The first | |
11893 | @code{while} has a true-or-false-test that tests true if the prefix | |
11894 | argument for @code{forward-sentence} is a negative number. This is for | |
11895 | going backwards. The body of this loop is similar to the body of the | |
11896 | second @code{while} clause, but it is not exactly the same. We will | |
11897 | skip this @code{while} loop and concentrate on the second @code{while} | |
11898 | loop. | |
11899 | ||
11900 | @need 1500 | |
11901 | The second @code{while} loop is for moving point forward. Its skeleton | |
11902 | looks like this: | |
11903 | ||
11904 | @smallexample | |
11905 | @group | |
11906 | (while (> arg 0) ; @r{true-or-false-test} | |
11907 | (let @var{varlist} | |
11908 | (if (@var{true-or-false-test}) | |
11909 | @var{then-part} | |
11910 | @var{else-part} | |
11911 | (setq arg (1- arg)))) ; @code{while} @r{loop decrementer} | |
11912 | @end group | |
11913 | @end smallexample | |
11914 | ||
11915 | The @code{while} loop is of the decrementing kind. | |
11916 | (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It | |
11917 | has a true-or-false-test that tests true so long as the counter (in | |
11918 | this case, the variable @code{arg}) is greater than zero; and it has a | |
11919 | decrementer that subtracts 1 from the value of the counter every time | |
11920 | the loop repeats. | |
11921 | ||
11922 | If no prefix argument is given to @code{forward-sentence}, which is | |
11923 | the most common way the command is used, this @code{while} loop will | |
11924 | run once, since the value of @code{arg} will be 1. | |
11925 | ||
11926 | The body of the @code{while} loop consists of a @code{let} expression, | |
11927 | which creates and binds a local variable, and has, as its body, an | |
11928 | @code{if} expression. | |
11929 | ||
11930 | @need 1250 | |
11931 | The body of the @code{while} loop looks like this: | |
11932 | ||
11933 | @smallexample | |
11934 | @group | |
11935 | (let ((par-end | |
11936 | (save-excursion (end-of-paragraph-text) (point)))) | |
11937 | (if (re-search-forward sentence-end par-end t) | |
11938 | (skip-chars-backward " \t\n") | |
11939 | (goto-char par-end))) | |
11940 | @end group | |
11941 | @end smallexample | |
11942 | ||
11943 | The @code{let} expression creates and binds the local variable | |
11944 | @code{par-end}. As we shall see, this local variable is designed to | |
11945 | provide a bound or limit to the regular expression search. If the | |
11946 | search fails to find a proper sentence ending in the paragraph, it will | |
11947 | stop on reaching the end of the paragraph. | |
11948 | ||
11949 | But first, let us examine how @code{par-end} is bound to the value of | |
11950 | the end of the paragraph. What happens is that the @code{let} sets the | |
11951 | value of @code{par-end} to the value returned when the Lisp interpreter | |
11952 | evaluates the expression | |
11953 | ||
11954 | @smallexample | |
11955 | @group | |
11956 | (save-excursion (end-of-paragraph-text) (point)) | |
11957 | @end group | |
11958 | @end smallexample | |
11959 | ||
11960 | @noindent | |
11961 | In this expression, @code{(end-of-paragraph-text)} moves point to the | |
11962 | end of the paragraph, @code{(point)} returns the value of point, and then | |
11963 | @code{save-excursion} restores point to its original position. Thus, | |
11964 | the @code{let} binds @code{par-end} to the value returned by the | |
11965 | @code{save-excursion} expression, which is the position of the end of | |
11966 | the paragraph. (The @code{(end-of-paragraph-text)} function uses | |
11967 | @code{forward-paragraph}, which we will discuss shortly.) | |
11968 | ||
11969 | @need 1200 | |
11970 | Emacs next evaluates the body of the @code{let}, which is an @code{if} | |
11971 | expression that looks like this: | |
11972 | ||
11973 | @smallexample | |
11974 | @group | |
11975 | (if (re-search-forward sentence-end par-end t) ; @r{if-part} | |
11976 | (skip-chars-backward " \t\n") ; @r{then-part} | |
11977 | (goto-char par-end))) ; @r{else-part} | |
11978 | @end group | |
11979 | @end smallexample | |
11980 | ||
11981 | The @code{if} tests whether its first argument is true and if so, | |
11982 | evaluates its then-part; otherwise, the Emacs Lisp interpreter | |
11983 | evaluates the else-part. The true-or-false-test of the @code{if} | |
11984 | expression is the regular expression search. | |
11985 | ||
11986 | It may seem odd to have what looks like the `real work' of | |
11987 | the @code{forward-sentence} function buried here, but this is a common | |
11988 | way this kind of operation is carried out in Lisp. | |
11989 | ||
11990 | @node fwd-sentence re-search, , fwd-sentence while loops, forward-sentence | |
11991 | @unnumberedsubsec The regular expression search | |
11992 | ||
11993 | The @code{re-search-forward} function searches for the end of the | |
11994 | sentence, that is, for the pattern defined by the @code{sentence-end} | |
11995 | regular expression. If the pattern is found---if the end of the sentence is | |
11996 | found---then the @code{re-search-forward} function does two things: | |
11997 | ||
11998 | @enumerate | |
11999 | @item | |
12000 | The @code{re-search-forward} function carries out a side effect, which | |
12001 | is to move point to the end of the occurrence found. | |
12002 | ||
12003 | @item | |
12004 | The @code{re-search-forward} function returns a value of true. This is | |
12005 | the value received by the @code{if}, and means that the search was | |
12006 | successful. | |
12007 | @end enumerate | |
12008 | ||
12009 | @noindent | |
12010 | The side effect, the movement of point, is completed before the | |
12011 | @code{if} function is handed the value returned by the successful | |
12012 | conclusion of the search. | |
12013 | ||
12014 | When the @code{if} function receives the value of true from a successful | |
12015 | call to @code{re-search-forward}, the @code{if} evaluates the then-part, | |
12016 | which is the expression @code{(skip-chars-backward " \t\n")}. This | |
12017 | expression moves backwards over any blank spaces, tabs or carriage | |
12018 | returns until a printed character is found and then leaves point after | |
12019 | the character. Since point has already been moved to the end of the | |
12020 | pattern that marks the end of the sentence, this action leaves point | |
12021 | right after the closing printed character of the sentence, which is | |
12022 | usually a period. | |
12023 | ||
12024 | On the other hand, if the @code{re-search-forward} function fails to | |
12025 | find a pattern marking the end of the sentence, the function returns | |
12026 | false. The false then causes the @code{if} to evaluate its third | |
12027 | argument, which is @code{(goto-char par-end)}: it moves point to the | |
12028 | end of the paragraph. | |
12029 | ||
12030 | Regular expression searches are exceptionally useful and the pattern | |
12031 | illustrated by @code{re-search-forward}, in which the search is the | |
12032 | test of an @code{if} expression, is handy. You will see or write code | |
12033 | incorporating this pattern often. | |
12034 | ||
12035 | @node forward-paragraph, etags, forward-sentence, Regexp Search | |
12036 | @comment node-name, next, previous, up | |
12037 | @section @code{forward-paragraph}: a Goldmine of Functions | |
12038 | @findex forward-paragraph | |
12039 | ||
12040 | The @code{forward-paragraph} function moves point forward to the end | |
12041 | of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a | |
12042 | number of functions that are important in themselves, including | |
12043 | @code{let*}, @code{match-beginning}, and @code{looking-at}. | |
12044 | ||
12045 | The function definition for @code{forward-paragraph} is considerably | |
12046 | longer than the function definition for @code{forward-sentence} | |
12047 | because it works with a paragraph, each line of which may begin with a | |
12048 | fill prefix. | |
12049 | ||
12050 | A fill prefix consists of a string of characters that are repeated at | |
12051 | the beginning of each line. For example, in Lisp code, it is a | |
12052 | convention to start each line of a paragraph-long comment with | |
12053 | @samp{;;; }. In Text mode, four blank spaces make up another common | |
12054 | fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , , | |
12055 | emacs, The GNU Emacs Manual}, for more information about fill | |
12056 | prefixes.) | |
12057 | ||
12058 | The existence of a fill prefix means that in addition to being able to | |
12059 | find the end of a paragraph whose lines begin on the left-most | |
12060 | column, the @code{forward-paragraph} function must be able to find the | |
12061 | end of a paragraph when all or many of the lines in the buffer begin | |
12062 | with the fill prefix. | |
12063 | ||
12064 | Moreover, it is sometimes practical to ignore a fill prefix that | |
12065 | exists, especially when blank lines separate paragraphs. | |
12066 | This is an added complication. | |
12067 | ||
12068 | @menu | |
12069 | * forward-paragraph in brief:: Key parts of the function definition. | |
12070 | * fwd-para let:: The @code{let*} expression. | |
12071 | * fwd-para while:: The forward motion @code{while} loop. | |
12072 | * fwd-para between paragraphs:: Movement between paragraphs. | |
12073 | * fwd-para within paragraph:: Movement within paragraphs. | |
12074 | * fwd-para no fill prefix:: When there is no fill prefix. | |
12075 | * fwd-para with fill prefix:: When there is a fill prefix. | |
12076 | * fwd-para summary:: Summary of @code{forward-paragraph} code. | |
12077 | @end menu | |
12078 | ||
12079 | @node forward-paragraph in brief, fwd-para let, forward-paragraph, forward-paragraph | |
12080 | @ifnottex | |
12081 | @unnumberedsubsec Shortened @code{forward-paragraph} function definition | |
12082 | @end ifnottex | |
12083 | ||
12084 | Rather than print all of the @code{forward-paragraph} function, we | |
12085 | will only print parts of it. Read without preparation, the function | |
12086 | can be daunting! | |
12087 | ||
12088 | @need 800 | |
12089 | In outline, the function looks like this: | |
12090 | ||
12091 | @smallexample | |
12092 | @group | |
12093 | (defun forward-paragraph (&optional arg) | |
12094 | "@var{documentation}@dots{}" | |
12095 | (interactive "p") | |
12096 | (or arg (setq arg 1)) | |
12097 | (let* | |
12098 | @var{varlist} | |
12099 | (while (< arg 0) ; @r{backward-moving-code} | |
12100 | @dots{} | |
12101 | (setq arg (1+ arg))) | |
12102 | (while (> arg 0) ; @r{forward-moving-code} | |
12103 | @dots{} | |
12104 | (setq arg (1- arg))))) | |
12105 | @end group | |
12106 | @end smallexample | |
12107 | ||
12108 | The first parts of the function are routine: the function's argument | |
12109 | list consists of one optional argument. Documentation follows. | |
12110 | ||
12111 | The lower case @samp{p} in the @code{interactive} declaration means | |
12112 | that the processed prefix argument, if any, is passed to the function. | |
12113 | This will be a number, and is the repeat count of how many paragraphs | |
12114 | point will move. The @code{or} expression in the next line handles | |
12115 | the common case when no argument is passed to the function, which occurs | |
12116 | if the function is called from other code rather than interactively. | |
12117 | This case was described earlier. (@xref{forward-sentence, The | |
12118 | @code{forward-sentence} function}.) Now we reach the end of the | |
12119 | familiar part of this function. | |
12120 | ||
12121 | @node fwd-para let, fwd-para while, forward-paragraph in brief, forward-paragraph | |
12122 | @unnumberedsubsec The @code{let*} expression | |
12123 | ||
12124 | The next line of the @code{forward-paragraph} function begins a | |
12125 | @code{let*} expression. This is a different kind of expression than | |
12126 | we have seen so far. The symbol is @code{let*} not @code{let}. | |
12127 | ||
12128 | The @code{let*} special form is like @code{let} except that Emacs sets | |
12129 | each variable in sequence, one after another, and variables in the | |
12130 | latter part of the varlist can make use of the values to which Emacs | |
12131 | set variables in the earlier part of the varlist. | |
12132 | ||
12133 | In the @code{let*} expression in this function, Emacs binds two | |
12134 | variables: @code{fill-prefix-regexp} and @code{paragraph-separate}. | |
12135 | The value to which @code{paragraph-separate} is bound depends on the | |
12136 | value of @code{fill-prefix-regexp}. | |
12137 | ||
12138 | @need 1200 | |
12139 | Let's look at each in turn. The symbol @code{fill-prefix-regexp} is | |
12140 | set to the value returned by evaluating the following list: | |
12141 | ||
12142 | @smallexample | |
12143 | @group | |
12144 | (and fill-prefix | |
12145 | (not (equal fill-prefix "")) | |
12146 | (not paragraph-ignore-fill-prefix) | |
12147 | (regexp-quote fill-prefix)) | |
12148 | @end group | |
12149 | @end smallexample | |
12150 | ||
12151 | @noindent | |
12152 | This is an expression whose first element is the @code{and} special form. | |
12153 | ||
12154 | As we learned earlier (@pxref{kill-new function, , The @code{kill-new} | |
12155 | function}), the @code{and} special form evaluates each of its | |
12156 | arguments until one of the arguments returns a value of @code{nil}, in | |
12157 | which case the @code{and} expression returns @code{nil}; however, if | |
12158 | none of the arguments returns a value of @code{nil}, the value | |
12159 | resulting from evaluating the last argument is returned. (Since such | |
12160 | a value is not @code{nil}, it is considered true in Lisp.) In other | |
12161 | words, an @code{and} expression returns a true value only if all its | |
12162 | arguments are true. | |
12163 | @findex and | |
12164 | ||
12165 | In this case, the variable @code{fill-prefix-regexp} is bound to a | |
12166 | non-@code{nil} value only if the following four expressions produce a | |
12167 | true (i.e., a non-@code{nil}) value when they are evaluated; otherwise, | |
12168 | @code{fill-prefix-regexp} is bound to @code{nil}. | |
12169 | ||
12170 | @table @code | |
12171 | @item fill-prefix | |
12172 | When this variable is evaluated, the value of the fill prefix, if any, | |
12173 | is returned. If there is no fill prefix, this variable returns | |
12174 | @code{nil}. | |
12175 | ||
12176 | @item (not (equal fill-prefix "") | |
12177 | This expression checks whether an existing fill prefix is an empty | |
12178 | string, that is, a string with no characters in it. An empty string is | |
12179 | not a useful fill prefix. | |
12180 | ||
12181 | @item (not paragraph-ignore-fill-prefix) | |
12182 | This expression returns @code{nil} if the variable | |
12183 | @code{paragraph-ignore-fill-prefix} has been turned on by being set to a | |
12184 | true value such as @code{t}. | |
12185 | ||
12186 | @item (regexp-quote fill-prefix) | |
12187 | This is the last argument to the @code{and} special form. If all the | |
12188 | arguments to the @code{and} are true, the value resulting from | |
12189 | evaluating this expression will be returned by the @code{and} expression | |
12190 | and bound to the variable @code{fill-prefix-regexp}, | |
12191 | @end table | |
12192 | ||
12193 | @findex regexp-quote | |
12194 | @noindent | |
12195 | The result of evaluating this @code{and} expression successfully is that | |
12196 | @code{fill-prefix-regexp} will be bound to the value of | |
12197 | @code{fill-prefix} as modified by the @code{regexp-quote} function. | |
12198 | What @code{regexp-quote} does is read a string and return a regular | |
12199 | expression that will exactly match the string and match nothing else. | |
12200 | This means that @code{fill-prefix-regexp} will be set to a value that | |
12201 | will exactly match the fill prefix if the fill prefix exists. | |
12202 | Otherwise, the variable will be set to @code{nil}. | |
12203 | ||
12204 | The second local variable in the @code{let*} expression is | |
12205 | @code{paragraph-separate}. It is bound to the value returned by | |
12206 | evaluating the expression: | |
12207 | ||
12208 | @smallexample | |
12209 | @group | |
12210 | (if fill-prefix-regexp | |
12211 | (concat paragraph-separate | |
12212 | "\\|^" fill-prefix-regexp "[ \t]*$") | |
12213 | paragraph-separate))) | |
12214 | @end group | |
12215 | @end smallexample | |
12216 | ||
12217 | This expression shows why @code{let*} rather than @code{let} was used. | |
12218 | The true-or-false-test for the @code{if} depends on whether the variable | |
12219 | @code{fill-prefix-regexp} evaluates to @code{nil} or some other value. | |
12220 | ||
12221 | If @code{fill-prefix-regexp} does not have a value, Emacs evaluates | |
12222 | the else-part of the @code{if} expression and binds | |
12223 | @code{paragraph-separate} to its local value. | |
12224 | (@code{paragraph-separate} is a regular expression that matches what | |
12225 | separates paragraphs.) | |
12226 | ||
12227 | But if @code{fill-prefix-regexp} does have a value, Emacs evaluates | |
12228 | the then-part of the @code{if} expression and binds | |
12229 | @code{paragraph-separate} to a regular expression that includes the | |
12230 | @code{fill-prefix-regexp} as part of the pattern. | |
12231 | ||
12232 | Specifically, @code{paragraph-separate} is set to the original value | |
12233 | of the paragraph separate regular expression concatenated with an | |
12234 | alternative expression that consists of the @code{fill-prefix-regexp} | |
12235 | followed by a blank line. The @samp{^} indicates that the | |
12236 | @code{fill-prefix-regexp} must begin a line, and the optional | |
12237 | whitespace to the end of the line is defined by @w{@code{"[ \t]*$"}}.) | |
12238 | The @samp{\\|} defines this portion of the regexp as an alternative to | |
12239 | @code{paragraph-separate}. | |
12240 | ||
12241 | Now we get into the body of the @code{let*}. The first part of the body | |
12242 | of the @code{let*} deals with the case when the function is given a | |
12243 | negative argument and is therefore moving backwards. We will skip this | |
12244 | section. | |
12245 | ||
12246 | @node fwd-para while, fwd-para between paragraphs, fwd-para let, forward-paragraph | |
12247 | @unnumberedsubsec The forward motion @code{while} loop | |
12248 | ||
12249 | The second part of the body of the @code{let*} deals with forward | |
12250 | motion. It is a @code{while} loop that repeats itself so long as the | |
12251 | value of @code{arg} is greater than zero. In the most common use of | |
12252 | the function, the value of the argument is 1, so the body of the | |
12253 | @code{while} loop is evaluated exactly once, and the cursor moves | |
12254 | forward one paragraph. | |
12255 | ||
12256 | This part handles three situations: when point is between paragraphs, | |
12257 | when point is within a paragraph and there is a fill prefix, and | |
12258 | when point is within a paragraph and there is no fill prefix. | |
12259 | ||
12260 | @need 800 | |
12261 | The @code{while} loop looks like this: | |
12262 | ||
12263 | @smallexample | |
12264 | @group | |
12265 | (while (> arg 0) | |
12266 | (beginning-of-line) | |
12267 | ||
12268 | ;; @r{between paragraphs} | |
12269 | (while (prog1 (and (not (eobp)) | |
12270 | (looking-at paragraph-separate)) | |
12271 | (forward-line 1))) | |
12272 | @end group | |
12273 | ||
12274 | @group | |
12275 | ;; @r{within paragraphs, with a fill prefix} | |
12276 | (if fill-prefix-regexp | |
12277 | ;; @r{There is a fill prefix; it overrides paragraph-start.} | |
12278 | (while (and (not (eobp)) | |
12279 | (not (looking-at paragraph-separate)) | |
12280 | (looking-at fill-prefix-regexp)) | |
12281 | (forward-line 1)) | |
12282 | @end group | |
12283 | ||
12284 | @group | |
12285 | ;; @r{within paragraphs, no fill prefix} | |
12286 | (if (re-search-forward paragraph-start nil t) | |
12287 | (goto-char (match-beginning 0)) | |
12288 | (goto-char (point-max)))) | |
12289 | ||
12290 | (setq arg (1- arg))) | |
12291 | @end group | |
12292 | @end smallexample | |
12293 | ||
12294 | We can see immediately that this is a decrementing counter @code{while} | |
12295 | loop, using the expression @code{(setq arg (1- arg))} as the decrementer. | |
12296 | ||
12297 | @need 800 | |
12298 | The body of the loop consists of three expressions: | |
12299 | ||
12300 | @smallexample | |
12301 | @group | |
12302 | ;; @r{between paragraphs} | |
12303 | (beginning-of-line) | |
12304 | (while | |
12305 | @var{body-of-while}) | |
12306 | @end group | |
12307 | ||
12308 | @group | |
12309 | ;; @r{within paragraphs, with fill prefix} | |
12310 | (if @var{true-or-false-test} | |
12311 | @var{then-part} | |
12312 | @end group | |
12313 | ||
12314 | @group | |
12315 | ;; @r{within paragraphs, no fill prefix} | |
12316 | @var{else-part} | |
12317 | @end group | |
12318 | @end smallexample | |
12319 | ||
12320 | @noindent | |
12321 | When the Emacs Lisp interpreter evaluates the body of the | |
12322 | @code{while} loop, the first thing it does is evaluate the | |
12323 | @code{(beginning-of-line)} expression and move point to the beginning | |
12324 | of the line. Then there is an inner @code{while} loop. This | |
12325 | @code{while} loop is designed to move the cursor out of the blank | |
12326 | space between paragraphs, if it should happen to be there. Finally, | |
12327 | there is an @code{if} expression that actually moves point to the end | |
12328 | of the paragraph. | |
12329 | ||
12330 | @node fwd-para between paragraphs, fwd-para within paragraph, fwd-para while, forward-paragraph | |
12331 | @unnumberedsubsec Between paragraphs | |
12332 | ||
12333 | First, let us look at the inner @code{while} loop. This loop handles | |
12334 | the case when point is between paragraphs; it uses three functions | |
12335 | that are new to us: @code{prog1}, @code{eobp} and @code{looking-at}. | |
12336 | @findex prog1 | |
12337 | @findex eobp | |
12338 | @findex looking-at | |
12339 | ||
12340 | @itemize @bullet | |
12341 | @item | |
12342 | @code{prog1} is similar to the @code{progn} special form, | |
12343 | except that @code{prog1} evaluates its arguments in sequence and then | |
12344 | returns the value of its first argument as the value of the whole | |
12345 | expression. (@code{progn} returns the value of its last argument as the | |
12346 | value of the expression.) The second and subsequent arguments to | |
12347 | @code{prog1} are evaluated only for their side effects. | |
12348 | ||
12349 | @item | |
12350 | @code{eobp} is an abbreviation of @samp{End Of Buffer P} and is a | |
12351 | function that returns true if point is at the end of the buffer. | |
12352 | ||
12353 | @item | |
12354 | @code{looking-at} is a function that returns true if the text following | |
12355 | point matches the regular expression passed @code{looking-at} as its | |
12356 | argument. | |
12357 | @end itemize | |
12358 | ||
12359 | @need 800 | |
12360 | The @code{while} loop we are studying looks like this: | |
12361 | ||
12362 | @smallexample | |
12363 | @group | |
12364 | (while (prog1 (and (not (eobp)) | |
12365 | (looking-at paragraph-separate)) | |
12366 | (forward-line 1))) | |
12367 | @end group | |
12368 | @end smallexample | |
12369 | ||
12370 | @need 1200 | |
12371 | @noindent | |
12372 | This is a @code{while} loop with no body! The true-or-false-test of the | |
12373 | loop is the expression: | |
12374 | ||
12375 | @smallexample | |
12376 | @group | |
12377 | (prog1 (and (not (eobp)) | |
12378 | (looking-at paragraph-separate)) | |
12379 | (forward-line 1)) | |
12380 | @end group | |
12381 | @end smallexample | |
12382 | ||
12383 | @noindent | |
12384 | The first argument to the @code{prog1} is the @code{and} expression. It | |
12385 | has within in it a test of whether point is at the end of the buffer and | |
12386 | also a test of whether the pattern following point matches the regular | |
12387 | expression for separating paragraphs. | |
12388 | ||
12389 | If the cursor is not at the end of the buffer and if the characters | |
12390 | following the cursor mark the separation between two paragraphs, then | |
12391 | the @code{and} expression is true. After evaluating the @code{and} | |
12392 | expression, the Lisp interpreter evaluates the second argument to | |
12393 | @code{prog1}, which is @code{forward-line}. This moves point forward | |
12394 | one line. The value returned by the @code{prog1} however, is the | |
12395 | value of its first argument, so the @code{while} loop continues so | |
12396 | long as point is not at the end of the buffer and is between | |
12397 | paragraphs. When, finally, point is moved to a paragraph, the | |
12398 | @code{and} expression tests false. Note however, that the | |
12399 | @code{forward-line} command is carried out anyhow. This means that | |
12400 | when point is moved from between paragraphs to a paragraph, it is left | |
12401 | at the beginning of the second line of the paragraph. | |
12402 | ||
12403 | @node fwd-para within paragraph, fwd-para no fill prefix, fwd-para between paragraphs, forward-paragraph | |
12404 | @unnumberedsubsec Within paragraphs | |
12405 | ||
12406 | The next expression in the outer @code{while} loop is an @code{if} | |
12407 | expression. The Lisp interpreter evaluates the then-part of the | |
12408 | @code{if} when the @code{fill-prefix-regexp} variable has a value other | |
12409 | than @code{nil}, and it evaluates the else-part when the value of | |
12410 | @code{if fill-prefix-regexp} is @code{nil}, that is, when there is no | |
12411 | fill prefix. | |
12412 | ||
12413 | @node fwd-para no fill prefix, fwd-para with fill prefix, fwd-para within paragraph, forward-paragraph | |
12414 | @unnumberedsubsec No fill prefix | |
12415 | ||
12416 | It is simplest to look at the code for the case when there is no fill | |
12417 | prefix first. This code consists of yet another inner @code{if} | |
12418 | expression, and reads as follows: | |
12419 | ||
12420 | @smallexample | |
12421 | @group | |
12422 | (if (re-search-forward paragraph-start nil t) | |
12423 | (goto-char (match-beginning 0)) | |
12424 | (goto-char (point-max))) | |
12425 | @end group | |
12426 | @end smallexample | |
12427 | ||
12428 | @noindent | |
12429 | This expression actually does the work that most people think of as | |
12430 | the primary purpose of the @code{forward-paragraph} command: it causes | |
12431 | a regular expression search to occur that searches forward to the | |
12432 | start of the next paragraph and if it is found, moves point there; but | |
12433 | if the start of another paragraph if not found, it moves point to the | |
12434 | end of the accessible region of the buffer. | |
12435 | ||
12436 | The only unfamiliar part of this is the use of @code{match-beginning}. | |
12437 | This is another function that is new to us. The | |
12438 | @code{match-beginning} function returns a number specifying the | |
12439 | location of the start of the text that was matched by the last regular | |
12440 | expression search. | |
12441 | ||
12442 | The @code{match-beginning} function is used here because of a | |
12443 | characteristic of a forward search: a successful forward search, | |
12444 | regardless of whether it is a plain search or a regular expression | |
12445 | search, will move point to the end of the text that is found. In this | |
12446 | case, a successful search will move point to the end of the pattern for | |
12447 | @code{paragraph-start}, which will be the beginning of the next | |
12448 | paragraph rather than the end of the current one. | |
12449 | ||
12450 | However, we want to put point at the end of the current paragraph, not at | |
12451 | the beginning of the next one. The two positions may be different, | |
12452 | because there may be several blank lines between paragraphs. | |
12453 | ||
12454 | @findex match-beginning | |
12455 | When given an argument of 0, @code{match-beginning} returns the position | |
12456 | that is the start of the text that the most recent regular | |
12457 | expression search matched. In this case, the most recent regular | |
12458 | expression search is the one looking for @code{paragraph-start}, so | |
12459 | @code{match-beginning} returns the beginning position of the pattern, | |
12460 | rather than the end of the pattern. The beginning position is the end | |
12461 | of the paragraph. | |
12462 | ||
12463 | (Incidentally, when passed a positive number as an argument, the | |
12464 | @code{match-beginning} function will place point at that parenthesized | |
12465 | expression in the last regular expression. It is a useful function.) | |
12466 | ||
12467 | @node fwd-para with fill prefix, fwd-para summary, fwd-para no fill prefix, forward-paragraph | |
12468 | @unnumberedsubsec With a fill prefix | |
12469 | ||
12470 | The inner @code{if} expression just discussed is the else-part of an enclosing | |
12471 | @code{if} expression which tests whether there is a fill prefix. If | |
12472 | there is a fill prefix, the then-part of this @code{if} is evaluated. | |
12473 | It looks like this: | |
12474 | ||
12475 | @smallexample | |
12476 | @group | |
12477 | (while (and (not (eobp)) | |
12478 | (not (looking-at paragraph-separate)) | |
12479 | (looking-at fill-prefix-regexp)) | |
12480 | (forward-line 1)) | |
12481 | @end group | |
12482 | @end smallexample | |
12483 | ||
12484 | @noindent | |
12485 | What this expression does is move point forward line by line so long | |
12486 | as three conditions are true: | |
12487 | ||
12488 | @enumerate | |
12489 | @item | |
12490 | Point is not at the end of the buffer. | |
12491 | ||
12492 | @item | |
12493 | The text following point does not separate paragraphs. | |
12494 | ||
12495 | @item | |
12496 | The pattern following point is the fill prefix regular expression. | |
12497 | @end enumerate | |
12498 | ||
12499 | The last condition may be puzzling, until you remember that point was | |
12500 | moved to the beginning of the line early in the @code{forward-paragraph} | |
12501 | function. This means that if the text has a fill prefix, the | |
12502 | @code{looking-at} function will see it. | |
12503 | ||
12504 | @node fwd-para summary, , fwd-para with fill prefix, forward-paragraph | |
12505 | @unnumberedsubsec Summary | |
12506 | ||
12507 | In summary, when moving forward, the @code{forward-paragraph} function | |
12508 | does the following: | |
12509 | ||
12510 | @itemize @bullet | |
12511 | @item | |
12512 | Move point to the beginning of the line. | |
12513 | ||
12514 | @item | |
12515 | Skip over lines between paragraphs. | |
12516 | ||
12517 | @item | |
12518 | Check whether there is a fill prefix, and if there is: | |
12519 | ||
12520 | @itemize --- | |
12521 | ||
12522 | @item | |
12523 | Go forward line by line so long as the line is not a paragraph | |
12524 | separating line. | |
12525 | @end itemize | |
12526 | ||
12527 | @item | |
12528 | But if there is no fill prefix, | |
12529 | ||
12530 | @itemize --- | |
12531 | ||
12532 | @item | |
12533 | Search for the next paragraph start pattern. | |
12534 | ||
12535 | @item | |
12536 | Go to the beginning of the paragraph start pattern, which will be the | |
12537 | end of the previous paragraph. | |
12538 | ||
12539 | @item | |
12540 | Or else go to the end of the accessible portion of the buffer. | |
12541 | @end itemize | |
12542 | @end itemize | |
12543 | ||
12544 | @need 1200 | |
12545 | For review, here is the code we have just been discussing, formatted | |
12546 | for clarity: | |
12547 | ||
12548 | @smallexample | |
12549 | @group | |
12550 | (interactive "p") | |
12551 | (or arg (setq arg 1)) | |
12552 | (let* ( | |
12553 | (fill-prefix-regexp | |
12554 | (and fill-prefix (not (equal fill-prefix "")) | |
12555 | (not paragraph-ignore-fill-prefix) | |
12556 | (regexp-quote fill-prefix))) | |
12557 | @end group | |
12558 | ||
12559 | @group | |
12560 | (paragraph-separate | |
12561 | (if fill-prefix-regexp | |
12562 | (concat paragraph-separate | |
12563 | "\\|^" | |
12564 | fill-prefix-regexp | |
12565 | "[ \t]*$") | |
12566 | paragraph-separate))) | |
12567 | ||
12568 | @var{omitted-backward-moving-code} @dots{} | |
12569 | @end group | |
12570 | ||
12571 | @group | |
12572 | (while (> arg 0) ; @r{forward-moving-code} | |
12573 | (beginning-of-line) | |
12574 | ||
12575 | (while (prog1 (and (not (eobp)) | |
12576 | (looking-at paragraph-separate)) | |
12577 | (forward-line 1))) | |
12578 | @end group | |
12579 | ||
12580 | @group | |
12581 | (if fill-prefix-regexp | |
12582 | (while (and (not (eobp)) ; @r{then-part} | |
12583 | (not (looking-at paragraph-separate)) | |
12584 | (looking-at fill-prefix-regexp)) | |
12585 | (forward-line 1)) | |
12586 | @end group | |
12587 | @group | |
12588 | ; @r{else-part: the inner-if} | |
12589 | (if (re-search-forward paragraph-start nil t) | |
12590 | (goto-char (match-beginning 0)) | |
12591 | (goto-char (point-max)))) | |
12592 | ||
12593 | (setq arg (1- arg))))) ; @r{decrementer} | |
12594 | @end group | |
12595 | @end smallexample | |
12596 | ||
12597 | The full definition for the @code{forward-paragraph} function not only | |
12598 | includes this code for going forwards, but also code for going backwards. | |
12599 | ||
12600 | If you are reading this inside of GNU Emacs and you want to see the | |
12601 | whole function, you can type @kbd{C-h f} (@code{describe-function}) | |
12602 | and the name of the function. This gives you the function | |
12603 | documentation and the name of the library containing the function's | |
12604 | source. Place point over the name of the library and press the RET | |
12605 | key; you will be taken directly to the source. (Be sure to install | |
12606 | your sources! Without them, you are like a person who tries to drive | |
12607 | a car with his eyes shut!) | |
12608 | ||
12609 | @c !!! again, 21.0.100 tags table location in this paragraph | |
12610 | Or -- a good habit to get into -- you can type @kbd{M-.} | |
12611 | (@code{find-tag}) and the name of the function when prompted for it. | |
12612 | This will take you directly to the source. If the @code{find-tag} | |
12613 | function first asks you for the name of a @file{TAGS} table, give it | |
12614 | the name of the @file{TAGS} file such as | |
12615 | @file{/usr/local/share/emacs/21.0.100/lisp/TAGS}. (The exact path to your | |
12616 | @file{TAGS} file depends on how your copy of Emacs was installed.) | |
12617 | ||
12618 | You can also create your own @file{TAGS} file for directories that | |
12619 | lack one. | |
12620 | @ifnottex | |
12621 | @xref{etags, , Create Your Own @file{TAGS} File}. | |
12622 | @end ifnottex | |
12623 | ||
12624 | @node etags, Regexp Review, forward-paragraph, Regexp Search | |
12625 | @section Create Your Own @file{TAGS} File | |
12626 | @findex etags | |
12627 | @cindex @file{TAGS} file, create own | |
12628 | ||
12629 | The @kbd{M-.} (@code{find-tag}) command takes you directly to the | |
12630 | source for a function, variable, node, or other source. The function | |
12631 | depends on tags tables to tell it where to go. | |
12632 | ||
12633 | You often need to build and install tags tables yourself. They are | |
12634 | not built automatically. A tags table is called a @file{TAGS} file; | |
12635 | the name is in upper case letters. | |
12636 | ||
12637 | You can create a @file{TAGS} file by calling the @code{etags} program | |
12638 | that comes as a part of the Emacs distribution. Usually, @code{etags} | |
12639 | is compiled and installed when Emacs is built. (@code{etags} is not | |
12640 | an Emacs Lisp function or a part of Emacs; it is a C program.) | |
12641 | ||
12642 | @need 1250 | |
12643 | To create a @file{TAGS} file, first switch to the directory in which | |
12644 | you want to create the file. In Emacs you can do this with the | |
12645 | @kbd{M-x cd} command, or by visiting a file in the directory, or by | |
12646 | listing the directory with @kbd{C-x d} (@code{dired}). Then run the | |
12647 | compile command, with @w{@code{etags *.el}} as the command to execute | |
12648 | ||
12649 | @smallexample | |
12650 | M-x compile RET etags *.el RET | |
12651 | @end smallexample | |
12652 | ||
12653 | @noindent | |
12654 | to create a @file{TAGS} file. | |
12655 | ||
12656 | For example, if you have a large number of files in your | |
12657 | @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it, | |
12658 | of which I load 12---you can create a @file{TAGS} file for the Emacs | |
12659 | Lisp files in that directory. | |
12660 | ||
12661 | @need 1250 | |
12662 | The @code{etags} program takes all the | |
12663 | usual shell `wildcards'. For example, if you have two directories for | |
12664 | which you want a single @file{TAGS file}, type | |
12665 | @w{@code{etags *.el ../elisp/*.el}}, | |
12666 | where @file{../elisp/} is the second directory: | |
12667 | ||
12668 | @smallexample | |
12669 | M-x compile RET etags *.el ../elisp/*.el RET | |
12670 | @end smallexample | |
12671 | ||
12672 | @need 1250 | |
12673 | Type | |
12674 | ||
12675 | @smallexample | |
12676 | M-x compile RET etags --help RET | |
12677 | @end smallexample | |
12678 | ||
12679 | @noindent | |
12680 | to see a list of the options accepted by @code{etags} as well as a | |
12681 | list of supported languages. | |
12682 | ||
12683 | The @code{etags} program handles more than 20 languages, including | |
12684 | Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, Java, LaTeX, | |
12685 | Pascal, Perl, Python, Texinfo, makefiles, and most assemblers. The | |
12686 | program has no switches for specifying the language; it recognizes the | |
12687 | language in an input file according to its file name and contents. | |
12688 | ||
12689 | @file{etags} is very helpful when you are writing code yourself and | |
12690 | want to refer back to functions you have already written. Just run | |
12691 | @code{etags} again at intervals as you write new functions, so they | |
12692 | become part of the @file{TAGS} file. | |
12693 | ||
12694 | If you think an appropriate @file{TAGS} file already exists for what | |
12695 | you want, but do not know where it is, you can use the @code{locate} | |
12696 | program to attempt to find it. | |
12697 | ||
12698 | Type @w{@kbd{M-x locate RET TAGS RET}} and Emacs will list for you the | |
12699 | full path names of all your @file{TAGS} files. On my system, this | |
12700 | command lists 34 @file{TAGS} files. On the other hand, a `plain | |
12701 | vanilla' system I recently installed did not contain any @file{TAGS} | |
12702 | files. | |
12703 | ||
12704 | If the tags table you want has been created, you can use the @code{M-x | |
12705 | visit-tags-table} command to specify it. Otherwise, you will need to | |
12706 | create the tag table yourself and then use @code{M-x | |
12707 | visit-tags-table}. | |
12708 | ||
12709 | @subsubheading Building Tags in the Emacs sources | |
12710 | @cindex Building Tags in the Emacs sources | |
12711 | @cindex Tags in the Emacs sources | |
12712 | @findex make tags | |
12713 | ||
12714 | The GNU Emacs sources come with a @file{Makefile} that contains a | |
12715 | sophisticated @code{etags} command that creates, collects, and merges | |
12716 | tags tables from all over the Emacs sources and puts the information | |
12717 | into one @file{TAGS} file in the @file{src/} directory below the top | |
12718 | level of your Emacs source directory. | |
12719 | ||
12720 | @need 1250 | |
12721 | To build this @file{TAGS} file, go to the top level of your Emacs | |
12722 | source directory and run the compile command @code{make tags}: | |
12723 | ||
12724 | @smallexample | |
12725 | M-x compile RET make tags RET | |
12726 | @end smallexample | |
12727 | ||
12728 | @noindent | |
12729 | (The @code{make tags} command works well with the GNU Emacs sources, | |
12730 | as well as with some other source packages.) | |
12731 | ||
12732 | For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs | |
12733 | Manual}. | |
12734 | ||
12735 | @node Regexp Review, re-search Exercises, etags, Regexp Search | |
12736 | @comment node-name, next, previous, up | |
12737 | @section Review | |
12738 | ||
12739 | Here is a brief summary of some recently introduced functions. | |
12740 | ||
12741 | @table @code | |
12742 | @item while | |
12743 | Repeatedly evaluate the body of the expression so long as the first | |
12744 | element of the body tests true. Then return @code{nil}. (The | |
12745 | expression is evaluated only for its side effects.) | |
12746 | ||
12747 | @need 1250 | |
12748 | For example: | |
12749 | ||
12750 | @smallexample | |
12751 | @group | |
12752 | (let ((foo 2)) | |
12753 | (while (> foo 0) | |
12754 | (insert (format "foo is %d.\n" foo)) | |
12755 | (setq foo (1- foo)))) | |
12756 | ||
12757 | @result{} foo is 2. | |
12758 | foo is 1. | |
12759 | nil | |
12760 | @end group | |
12761 | @end smallexample | |
12762 | @noindent | |
12763 | (The @code{insert} function inserts its arguments at point; the | |
12764 | @code{format} function returns a string formatted from its arguments | |
12765 | the way @code{message} formats its arguments; @code{\n} produces a new | |
12766 | line.) | |
12767 | ||
12768 | @item re-search-forward | |
12769 | Search for a pattern, and if the pattern is found, move point to rest | |
12770 | just after it. | |
12771 | ||
12772 | @noindent | |
12773 | Takes four arguments, like @code{search-forward}: | |
12774 | ||
12775 | @enumerate | |
12776 | @item | |
12777 | A regular expression that specifies the pattern to search for. | |
12778 | ||
12779 | @item | |
12780 | Optionally, the limit of the search. | |
12781 | ||
12782 | @item | |
12783 | Optionally, what to do if the search fails, return @code{nil} or an | |
12784 | error message. | |
12785 | ||
12786 | @item | |
12787 | Optionally, how many times to repeat the search; if negative, the | |
12788 | search goes backwards. | |
12789 | @end enumerate | |
12790 | ||
12791 | @item let* | |
12792 | Bind some variables locally to particular values, | |
12793 | and then evaluate the remaining arguments, returning the value of the | |
12794 | last one. While binding the local variables, use the local values of | |
12795 | variables bound earlier, if any. | |
12796 | ||
12797 | @need 1250 | |
12798 | For example: | |
12799 | ||
12800 | @smallexample | |
12801 | @group | |
12802 | (let* ((foo 7) | |
12803 | (bar (* 3 foo))) | |
12804 | (message "`bar' is %d." bar)) | |
12805 | @result{} `bar' is 21. | |
12806 | @end group | |
12807 | @end smallexample | |
12808 | ||
12809 | @item match-beginning | |
12810 | Return the position of the start of the text found by the last regular | |
12811 | expression search. | |
12812 | ||
12813 | @item looking-at | |
12814 | Return @code{t} for true if the text after point matches the argument, | |
12815 | which should be a regular expression. | |
12816 | ||
12817 | @item eobp | |
12818 | Return @code{t} for true if point is at the end of the accessible part | |
12819 | of a buffer. The end of the accessible part is the end of the buffer | |
12820 | if the buffer is not narrowed; it is the end of the narrowed part if | |
12821 | the buffer is narrowed. | |
12822 | ||
12823 | @item prog1 | |
12824 | Evaluate each argument in sequence and then return the value of the | |
12825 | @emph{first}. | |
12826 | ||
12827 | @need 1250 | |
12828 | For example: | |
12829 | ||
12830 | @smallexample | |
12831 | @group | |
12832 | (prog1 1 2 3 4) | |
12833 | @result{} 1 | |
12834 | @end group | |
12835 | @end smallexample | |
12836 | @end table | |
12837 | ||
12838 | @need 1500 | |
12839 | @node re-search Exercises, , Regexp Review, Regexp Search | |
12840 | @section Exercises with @code{re-search-forward} | |
12841 | ||
12842 | @itemize @bullet | |
12843 | @item | |
12844 | Write a function to search for a regular expression that matches two | |
12845 | or more blank lines in sequence. | |
12846 | ||
12847 | @item | |
12848 | Write a function to search for duplicated words, such as `the the'. | |
12849 | @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs | |
12850 | Manual}, for information on how to write a regexp (a regular | |
12851 | expression) to match a string that is composed of two identical | |
12852 | halves. You can devise several regexps; some are better than others. | |
12853 | The function I use is described in an appendix, along with several | |
12854 | regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}. | |
12855 | @end itemize | |
12856 | ||
12857 | @node Counting Words, Words in a defun, Regexp Search, Top | |
12858 | @chapter Counting: Repetition and Regexps | |
12859 | @cindex Repetition for word counting | |
12860 | @cindex Regular expressions for word counting | |
12861 | ||
12862 | Repetition and regular expression searches are powerful tools that you | |
12863 | often use when you write code in Emacs Lisp. This chapter illustrates | |
12864 | the use of regular expression searches through the construction of | |
12865 | word count commands using @code{while} loops and recursion. | |
12866 | ||
12867 | @menu | |
12868 | * Why Count Words:: | |
12869 | * count-words-region:: Use a regexp, but find a problem. | |
12870 | * recursive-count-words:: Start with case of no words in region. | |
12871 | * Counting Exercise:: | |
12872 | @end menu | |
12873 | ||
12874 | @node Why Count Words, count-words-region, Counting Words, Counting Words | |
12875 | @ifnottex | |
12876 | @unnumberedsec Counting words | |
12877 | @end ifnottex | |
12878 | ||
12879 | The standard Emacs distribution contains a function for counting the | |
12880 | number of lines within a region. However, there is no corresponding | |
12881 | function for counting words. | |
12882 | ||
12883 | Certain types of writing ask you to count words. Thus, if you write | |
12884 | an essay, you may be limited to 800 words; if you write a novel, you | |
12885 | may discipline yourself to write 1000 words a day. It seems odd to me | |
12886 | that Emacs lacks a word count command. Perhaps people use Emacs | |
12887 | mostly for code or types of documentation that do not require word | |
12888 | counts; or perhaps they restrict themselves to the operating system | |
12889 | word count command, @code{wc}. Alternatively, people may follow | |
12890 | the publishers' convention and compute a word count by dividing the | |
12891 | number of characters in a document by five. In any event, here are | |
12892 | commands to count words. | |
12893 | ||
12894 | @node count-words-region, recursive-count-words, Why Count Words, Counting Words | |
12895 | @comment node-name, next, previous, up | |
12896 | @section The @code{count-words-region} Function | |
12897 | @findex count-words-region | |
12898 | ||
12899 | A word count command could count words in a line, paragraph, region, | |
12900 | or buffer. What should the command cover? You could design the | |
12901 | command to count the number of words in a complete buffer. However, | |
12902 | the Emacs tradition encourages flexibility---you may want to count | |
12903 | words in just a section, rather than all of a buffer. So it makes | |
12904 | more sense to design the command to count the number of words in a | |
12905 | region. Once you have a @code{count-words-region} command, you can, | |
12906 | if you wish, count words in a whole buffer by marking it with @kbd{C-x | |
12907 | h} (@code{mark-whole-buffer}). | |
12908 | ||
12909 | Clearly, counting words is a repetitive act: starting from the | |
12910 | beginning of the region, you count the first word, then the second | |
12911 | word, then the third word, and so on, until you reach the end of the | |
12912 | region. This means that word counting is ideally suited to recursion | |
12913 | or to a @code{while} loop. | |
12914 | ||
12915 | @menu | |
12916 | * Design count-words-region:: The definition using a @code{while} loop. | |
12917 | * Whitespace Bug:: The Whitespace Bug in @code{count-words-region}. | |
12918 | @end menu | |
12919 | ||
12920 | @node Design count-words-region, Whitespace Bug, count-words-region, count-words-region | |
12921 | @ifnottex | |
12922 | @unnumberedsubsec Designing @code{count-words-region} | |
12923 | @end ifnottex | |
12924 | ||
12925 | First, we will implement the word count command with a @code{while} | |
12926 | loop, then with recursion. The command will, of course, be | |
12927 | interactive. | |
12928 | ||
12929 | @need 800 | |
12930 | The template for an interactive function definition is, as always: | |
12931 | ||
12932 | @smallexample | |
12933 | @group | |
12934 | (defun @var{name-of-function} (@var{argument-list}) | |
12935 | "@var{documentation}@dots{}" | |
12936 | (@var{interactive-expression}@dots{}) | |
12937 | @var{body}@dots{}) | |
12938 | @end group | |
12939 | @end smallexample | |
12940 | ||
12941 | What we need to do is fill in the slots. | |
12942 | ||
12943 | The name of the function should be self-explanatory and similar to the | |
12944 | existing @code{count-lines-region} name. This makes the name easier | |
12945 | to remember. @code{count-words-region} is a good choice. | |
12946 | ||
12947 | The function counts words within a region. This means that the | |
12948 | argument list must contain symbols that are bound to the two | |
12949 | positions, the beginning and end of the region. These two positions | |
12950 | can be called @samp{beginning} and @samp{end} respectively. The first | |
12951 | line of the documentation should be a single sentence, since that is | |
12952 | all that is printed as documentation by a command such as | |
12953 | @code{apropos}. The interactive expression will be of the form | |
12954 | @samp{(interactive "r")}, since that will cause Emacs to pass the | |
12955 | beginning and end of the region to the function's argument list. All | |
12956 | this is routine. | |
12957 | ||
12958 | The body of the function needs to be written to do three tasks: | |
12959 | first, to set up conditions under which the @code{while} loop can | |
12960 | count words, second, to run the @code{while} loop, and third, to send | |
12961 | a message to the user. | |
12962 | ||
12963 | When a user calls @code{count-words-region}, point may be at the | |
12964 | beginning or the end of the region. However, the counting process | |
12965 | must start at the beginning of the region. This means we will want | |
12966 | to put point there if it is not already there. Executing | |
12967 | @code{(goto-char beginning)} ensures this. Of course, we will want to | |
12968 | return point to its expected position when the function finishes its | |
12969 | work. For this reason, the body must be enclosed in a | |
12970 | @code{save-excursion} expression. | |
12971 | ||
12972 | The central part of the body of the function consists of a | |
12973 | @code{while} loop in which one expression jumps point forward word by | |
12974 | word, and another expression counts those jumps. The true-or-false-test | |
12975 | of the @code{while} loop should test true so long as point should jump | |
12976 | forward, and false when point is at the end of the region. | |
12977 | ||
12978 | We could use @code{(forward-word 1)} as the expression for moving point | |
12979 | forward word by word, but it is easier to see what Emacs identifies as a | |
12980 | `word' if we use a regular expression search. | |
12981 | ||
12982 | A regular expression search that finds the pattern for which it is | |
12983 | searching leaves point after the last character matched. This means | |
12984 | that a succession of successful word searches will move point forward | |
12985 | word by word. | |
12986 | ||
12987 | As a practical matter, we want the regular expression search to jump | |
12988 | over whitespace and punctuation between words as well as over the | |
12989 | words themselves. A regexp that refuses to jump over interword | |
12990 | whitespace would never jump more than one word! This means that | |
12991 | the regexp should include the whitespace and punctuation that follows | |
12992 | a word, if any, as well as the word itself. (A word may end a buffer | |
12993 | and not have any following whitespace or punctuation, so that part of | |
12994 | the regexp must be optional.) | |
12995 | ||
12996 | Thus, what we want for the regexp is a pattern defining one or more | |
12997 | word constituent characters followed, optionally, by one or more | |
12998 | characters that are not word constituents. The regular expression for | |
12999 | this is: | |
13000 | ||
13001 | @smallexample | |
13002 | \w+\W* | |
13003 | @end smallexample | |
13004 | ||
13005 | @noindent | |
13006 | The buffer's syntax table determines which characters are and are not | |
13007 | word constituents. (@xref{Syntax, , What Constitutes a Word or | |
13008 | Symbol?}, for more about syntax. Also, see @ref{Syntax, Syntax, The | |
13009 | Syntax Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, , | |
13010 | Syntax Tables, elisp, The GNU Emacs Lisp Reference Manual}.) | |
13011 | ||
13012 | @need 800 | |
13013 | The search expression looks like this: | |
13014 | ||
13015 | @smallexample | |
13016 | (re-search-forward "\\w+\\W*") | |
13017 | @end smallexample | |
13018 | ||
13019 | @noindent | |
13020 | (Note that paired backslashes precede the @samp{w} and @samp{W}. A | |
13021 | single backslash has special meaning to the Emacs Lisp interpreter. It | |
13022 | indicates that the following character is interpreted differently than | |
13023 | usual. For example, the two characters, @samp{\n}, stand for | |
13024 | @samp{newline}, rather than for a backslash followed by @samp{n}. Two | |
13025 | backslashes in a row stand for an ordinary, `unspecial' backslash.) | |
13026 | ||
13027 | We need a counter to count how many words there are; this variable | |
13028 | must first be set to 0 and then incremented each time Emacs goes | |
13029 | around the @code{while} loop. The incrementing expression is simply: | |
13030 | ||
13031 | @smallexample | |
13032 | (setq count (1+ count)) | |
13033 | @end smallexample | |
13034 | ||
13035 | Finally, we want to tell the user how many words there are in the | |
13036 | region. The @code{message} function is intended for presenting this | |
13037 | kind of information to the user. The message has to be phrased so | |
13038 | that it reads properly regardless of how many words there are in the | |
13039 | region: we don't want to say that ``there are 1 words in the region''. | |
13040 | The conflict between singular and plural is ungrammatical. We can | |
13041 | solve this problem by using a conditional expression that evaluates | |
13042 | different messages depending on the number of words in the region. | |
13043 | There are three possibilities: no words in the region, one word in the | |
13044 | region, and more than one word. This means that the @code{cond} | |
13045 | special form is appropriate. | |
13046 | ||
13047 | @need 1500 | |
13048 | All this leads to the following function definition: | |
13049 | ||
13050 | @smallexample | |
13051 | @group | |
13052 | ;;; @r{First version; has bugs!} | |
13053 | (defun count-words-region (beginning end) | |
13054 | "Print number of words in the region. | |
13055 | Words are defined as at least one word-constituent | |
13056 | character followed by at least one character that | |
13057 | is not a word-constituent. The buffer's syntax | |
13058 | table determines which characters these are." | |
13059 | (interactive "r") | |
13060 | (message "Counting words in region ... ") | |
13061 | @end group | |
13062 | ||
13063 | @group | |
13064 | ;;; @r{1. Set up appropriate conditions.} | |
13065 | (save-excursion | |
13066 | (goto-char beginning) | |
13067 | (let ((count 0)) | |
13068 | @end group | |
13069 | ||
13070 | @group | |
13071 | ;;; @r{2. Run the} while @r{loop.} | |
13072 | (while (< (point) end) | |
13073 | (re-search-forward "\\w+\\W*") | |
13074 | (setq count (1+ count))) | |
13075 | @end group | |
13076 | ||
13077 | @group | |
13078 | ;;; @r{3. Send a message to the user.} | |
13079 | (cond ((zerop count) | |
13080 | (message | |
13081 | "The region does NOT have any words.")) | |
13082 | ((= 1 count) | |
13083 | (message | |
13084 | "The region has 1 word.")) | |
13085 | (t | |
13086 | (message | |
13087 | "The region has %d words." count)))))) | |
13088 | @end group | |
13089 | @end smallexample | |
13090 | ||
13091 | @noindent | |
13092 | As written, the function works, but not in all circumstances. | |
13093 | ||
13094 | @node Whitespace Bug, , Design count-words-region, count-words-region | |
13095 | @comment node-name, next, previous, up | |
13096 | @subsection The Whitespace Bug in @code{count-words-region} | |
13097 | ||
13098 | The @code{count-words-region} command described in the preceding | |
13099 | section has two bugs, or rather, one bug with two manifestations. | |
13100 | First, if you mark a region containing only whitespace in the middle | |
13101 | of some text, the @code{count-words-region} command tells you that the | |
13102 | region contains one word! Second, if you mark a region containing | |
13103 | only whitespace at the end of the buffer or the accessible portion of | |
13104 | a narrowed buffer, the command displays an error message that looks | |
13105 | like this: | |
13106 | ||
13107 | @smallexample | |
13108 | Search failed: "\\w+\\W*" | |
13109 | @end smallexample | |
13110 | ||
13111 | If you are reading this in Info in GNU Emacs, you can test for these | |
13112 | bugs yourself. | |
13113 | ||
13114 | First, evaluate the function in the usual manner to install it. | |
13115 | @ifinfo | |
13116 | Here is a copy of the definition. Place your cursor after the closing | |
13117 | parenthesis and type @kbd{C-x C-e} to install it. | |
13118 | ||
13119 | @smallexample | |
13120 | @group | |
13121 | ;; @r{First version; has bugs!} | |
13122 | (defun count-words-region (beginning end) | |
13123 | "Print number of words in the region. | |
13124 | Words are defined as at least one word-constituent character followed | |
13125 | by at least one character that is not a word-constituent. The buffer's | |
13126 | syntax table determines which characters these are." | |
13127 | @end group | |
13128 | @group | |
13129 | (interactive "r") | |
13130 | (message "Counting words in region ... ") | |
13131 | @end group | |
13132 | ||
13133 | @group | |
13134 | ;;; @r{1. Set up appropriate conditions.} | |
13135 | (save-excursion | |
13136 | (goto-char beginning) | |
13137 | (let ((count 0)) | |
13138 | @end group | |
13139 | ||
13140 | @group | |
13141 | ;;; @r{2. Run the} while @r{loop.} | |
13142 | (while (< (point) end) | |
13143 | (re-search-forward "\\w+\\W*") | |
13144 | (setq count (1+ count))) | |
13145 | @end group | |
13146 | ||
13147 | @group | |
13148 | ;;; @r{3. Send a message to the user.} | |
13149 | (cond ((zerop count) | |
13150 | (message "The region does NOT have any words.")) | |
13151 | ((= 1 count) (message "The region has 1 word.")) | |
13152 | (t (message "The region has %d words." count)))))) | |
13153 | @end group | |
13154 | @end smallexample | |
13155 | @end ifinfo | |
13156 | ||
13157 | @need 1000 | |
13158 | If you wish, you can also install this keybinding by evaluating it: | |
13159 | ||
13160 | @smallexample | |
13161 | (global-set-key "\C-c=" 'count-words-region) | |
13162 | @end smallexample | |
13163 | ||
13164 | To conduct the first test, set mark and point to the beginning and end | |
13165 | of the following line and then type @kbd{C-c =} (or @kbd{M-x | |
13166 | count-words-region} if you have not bound @kbd{C-c =}): | |
13167 | ||
13168 | @smallexample | |
13169 | one two three | |
13170 | @end smallexample | |
13171 | ||
13172 | @noindent | |
13173 | Emacs will tell you, correctly, that the region has three words. | |
13174 | ||
13175 | Repeat the test, but place mark at the beginning of the line and place | |
13176 | point just @emph{before} the word @samp{one}. Again type the command | |
13177 | @kbd{C-c =} (or @kbd{M-x count-words-region}). Emacs should tell you | |
13178 | that the region has no words, since it is composed only of the | |
13179 | whitespace at the beginning of the line. But instead Emacs tells you | |
13180 | that the region has one word! | |
13181 | ||
13182 | For the third test, copy the sample line to the end of the | |
13183 | @file{*scratch*} buffer and then type several spaces at the end of the | |
13184 | line. Place mark right after the word @samp{three} and point at the | |
13185 | end of line. (The end of the line will be the end of the buffer.) | |
13186 | Type @kbd{C-c =} (or @kbd{M-x count-words-region}) as you did before. | |
13187 | Again, Emacs should tell you that the region has no words, since it is | |
13188 | composed only of the whitespace at the end of the line. Instead, | |
13189 | Emacs displays an error message saying @samp{Search failed}. | |
13190 | ||
13191 | The two bugs stem from the same problem. | |
13192 | ||
13193 | Consider the first manifestation of the bug, in which the command | |
13194 | tells you that the whitespace at the beginning of the line contains | |
13195 | one word. What happens is this: The @code{M-x count-words-region} | |
13196 | command moves point to the beginning of the region. The @code{while} | |
13197 | tests whether the value of point is smaller than the value of | |
13198 | @code{end}, which it is. Consequently, the regular expression search | |
13199 | looks for and finds the first word. It leaves point after the word. | |
13200 | @code{count} is set to one. The @code{while} loop repeats; but this | |
13201 | time the value of point is larger than the value of @code{end}, the | |
13202 | loop is exited; and the function displays a message saying the number | |
13203 | of words in the region is one. In brief, the regular expression | |
13204 | search looks for and finds the word even though it is outside | |
13205 | the marked region. | |
13206 | ||
13207 | In the second manifestation of the bug, the region is whitespace at | |
13208 | the end of the buffer. Emacs says @samp{Search failed}. What happens | |
13209 | is that the true-or-false-test in the @code{while} loop tests true, so | |
13210 | the search expression is executed. But since there are no more words | |
13211 | in the buffer, the search fails. | |
13212 | ||
13213 | In both manifestations of the bug, the search extends or attempts to | |
13214 | extend outside of the region. | |
13215 | ||
13216 | The solution is to limit the search to the region---this is a fairly | |
13217 | simple action, but as you may have come to expect, it is not quite as | |
13218 | simple as you might think. | |
13219 | ||
13220 | As we have seen, the @code{re-search-forward} function takes a search | |
13221 | pattern as its first argument. But in addition to this first, | |
13222 | mandatory argument, it accepts three optional arguments. The optional | |
13223 | second argument bounds the search. The optional third argument, if | |
13224 | @code{t}, causes the function to return @code{nil} rather than signal | |
13225 | an error if the search fails. The optional fourth argument is a | |
13226 | repeat count. (In Emacs, you can see a function's documentation by | |
13227 | typing @kbd{C-h f}, the name of the function, and then @key{RET}.) | |
13228 | ||
13229 | In the @code{count-words-region} definition, the value of the end of | |
13230 | the region is held by the variable @code{end} which is passed as an | |
13231 | argument to the function. Thus, we can add @code{end} as an argument | |
13232 | to the regular expression search expression: | |
13233 | ||
13234 | @smallexample | |
13235 | (re-search-forward "\\w+\\W*" end) | |
13236 | @end smallexample | |
13237 | ||
13238 | However, if you make only this change to the @code{count-words-region} | |
13239 | definition and then test the new version of the definition on a | |
13240 | stretch of whitespace, you will receive an error message saying | |
13241 | @samp{Search failed}. | |
13242 | ||
13243 | What happens is this: the search is limited to the region, and fails | |
13244 | as you expect because there are no word-constituent characters in the | |
13245 | region. Since it fails, we receive an error message. But we do not | |
13246 | want to receive an error message in this case; we want to receive the | |
13247 | message that "The region does NOT have any words." | |
13248 | ||
13249 | The solution to this problem is to provide @code{re-search-forward} | |
13250 | with a third argument of @code{t}, which causes the function to return | |
13251 | @code{nil} rather than signal an error if the search fails. | |
13252 | ||
13253 | However, if you make this change and try it, you will see the message | |
13254 | ``Counting words in region ... '' and @dots{} you will keep on seeing | |
13255 | that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}). | |
13256 | ||
13257 | Here is what happens: the search is limited to the region, as before, | |
13258 | and it fails because there are no word-constituent characters in the | |
13259 | region, as expected. Consequently, the @code{re-search-forward} | |
13260 | expression returns @code{nil}. It does nothing else. In particular, | |
13261 | it does not move point, which it does as a side effect if it finds the | |
13262 | search target. After the @code{re-search-forward} expression returns | |
13263 | @code{nil}, the next expression in the @code{while} loop is evaluated. | |
13264 | This expression increments the count. Then the loop repeats. The | |
13265 | true-or-false-test tests true because the value of point is still less | |
13266 | than the value of end, since the @code{re-search-forward} expression | |
13267 | did not move point. @dots{} and the cycle repeats @dots{} | |
13268 | ||
13269 | The @code{count-words-region} definition requires yet another | |
13270 | modification, to cause the true-or-false-test of the @code{while} loop | |
13271 | to test false if the search fails. Put another way, there are two | |
13272 | conditions that must be satisfied in the true-or-false-test before the | |
13273 | word count variable is incremented: point must still be within the | |
13274 | region and the search expression must have found a word to count. | |
13275 | ||
13276 | Since both the first condition and the second condition must be true | |
13277 | together, the two expressions, the region test and the search | |
13278 | expression, can be joined with an @code{and} special form and embedded in | |
13279 | the @code{while} loop as the true-or-false-test, like this: | |
13280 | ||
13281 | @smallexample | |
13282 | (and (< (point) end) (re-search-forward "\\w+\\W*" end t)) | |
13283 | @end smallexample | |
13284 | ||
13285 | @c colon in printed section title causes problem in Info cross reference | |
13286 | @c also trouble with an overfull hbox | |
13287 | @iftex | |
13288 | @noindent | |
13289 | (For information about @code{and}, see | |
13290 | @ref{forward-paragraph, , @code{forward-paragraph}: a Goldmine of | |
13291 | Functions}.) | |
13292 | @end iftex | |
13293 | @ifinfo | |
13294 | @noindent | |
13295 | (@xref{forward-paragraph}, for information about @code{and}.) | |
13296 | @end ifinfo | |
13297 | ||
13298 | The @code{re-search-forward} expression returns @code{t} if the search | |
13299 | succeeds and as a side effect moves point. Consequently, as words are | |
13300 | found, point is moved through the region. When the search | |
13301 | expression fails to find another word, or when point reaches the end | |
13302 | of the region, the true-or-false-test tests false, the @code{while} | |
13303 | loop exists, and the @code{count-words-region} function displays one | |
13304 | or other of its messages. | |
13305 | ||
13306 | After incorporating these final changes, the @code{count-words-region} | |
13307 | works without bugs (or at least, without bugs that I have found!). | |
13308 | Here is what it looks like: | |
13309 | ||
13310 | @smallexample | |
13311 | @group | |
13312 | ;;; @r{Final version:} @code{while} | |
13313 | (defun count-words-region (beginning end) | |
13314 | "Print number of words in the region." | |
13315 | (interactive "r") | |
13316 | (message "Counting words in region ... ") | |
13317 | @end group | |
13318 | ||
13319 | @group | |
13320 | ;;; @r{1. Set up appropriate conditions.} | |
13321 | (save-excursion | |
13322 | (let ((count 0)) | |
13323 | (goto-char beginning) | |
13324 | @end group | |
13325 | ||
13326 | @group | |
13327 | ;;; @r{2. Run the} while @r{loop.} | |
13328 | (while (and (< (point) end) | |
13329 | (re-search-forward "\\w+\\W*" end t)) | |
13330 | (setq count (1+ count))) | |
13331 | @end group | |
13332 | ||
13333 | @group | |
13334 | ;;; @r{3. Send a message to the user.} | |
13335 | (cond ((zerop count) | |
13336 | (message | |
13337 | "The region does NOT have any words.")) | |
13338 | ((= 1 count) | |
13339 | (message | |
13340 | "The region has 1 word.")) | |
13341 | (t | |
13342 | (message | |
13343 | "The region has %d words." count)))))) | |
13344 | @end group | |
13345 | @end smallexample | |
13346 | ||
13347 | @node recursive-count-words, Counting Exercise, count-words-region, Counting Words | |
13348 | @comment node-name, next, previous, up | |
13349 | @section Count Words Recursively | |
13350 | @cindex Count words recursively | |
13351 | @cindex Recursively counting words | |
13352 | @cindex Words, counted recursively | |
13353 | ||
13354 | You can write the function for counting words recursively as well as | |
13355 | with a @code{while} loop. Let's see how this is done. | |
13356 | ||
13357 | First, we need to recognize that the @code{count-words-region} | |
13358 | function has three jobs: it sets up the appropriate conditions for | |
13359 | counting to occur; it counts the words in the region; and it sends a | |
13360 | message to the user telling how many words there are. | |
13361 | ||
13362 | If we write a single recursive function to do everything, we will | |
13363 | receive a message for every recursive call. If the region contains 13 | |
13364 | words, we will receive thirteen messages, one right after the other. | |
13365 | We don't want this! Instead, we must write two functions to do the | |
13366 | job, one of which (the recursive function) will be used inside of the | |
13367 | other. One function will set up the conditions and display the | |
13368 | message; the other will return the word count. | |
13369 | ||
13370 | Let us start with the function that causes the message to be displayed. | |
13371 | We can continue to call this @code{count-words-region}. | |
13372 | ||
13373 | This is the function that the user will call. It will be interactive. | |
13374 | Indeed, it will be similar to our previous versions of this | |
13375 | function, except that it will call @code{recursive-count-words} to | |
13376 | determine how many words are in the region. | |
13377 | ||
13378 | @need 1250 | |
13379 | We can readily construct a template for this function, based on our | |
13380 | previous versions: | |
13381 | ||
13382 | @smallexample | |
13383 | @group | |
13384 | ;; @r{Recursive version; uses regular expression search} | |
13385 | (defun count-words-region (beginning end) | |
13386 | "@var{documentation}@dots{}" | |
13387 | (@var{interactive-expression}@dots{}) | |
13388 | @end group | |
13389 | @group | |
13390 | ||
13391 | ;;; @r{1. Set up appropriate conditions.} | |
13392 | (@var{explanatory message}) | |
13393 | (@var{set-up functions}@dots{} | |
13394 | @end group | |
13395 | @group | |
13396 | ||
13397 | ;;; @r{2. Count the words.} | |
13398 | @var{recursive call} | |
13399 | @end group | |
13400 | @group | |
13401 | ||
13402 | ;;; @r{3. Send a message to the user.} | |
13403 | @var{message providing word count})) | |
13404 | @end group | |
13405 | @end smallexample | |
13406 | ||
13407 | The definition looks straightforward, except that somehow the count | |
13408 | returned by the recursive call must be passed to the message | |
13409 | displaying the word count. A little thought suggests that this can be | |
13410 | done by making use of a @code{let} expression: we can bind a variable | |
13411 | in the varlist of a @code{let} expression to the number of words in | |
13412 | the region, as returned by the recursive call; and then the | |
13413 | @code{cond} expression, using binding, can display the value to the | |
13414 | user. | |
13415 | ||
13416 | Often, one thinks of the binding within a @code{let} expression as | |
13417 | somehow secondary to the `primary' work of a function. But in this | |
13418 | case, what you might consider the `primary' job of the function, | |
13419 | counting words, is done within the @code{let} expression. | |
13420 | ||
13421 | @need 1250 | |
13422 | Using @code{let}, the function definition looks like this: | |
13423 | ||
13424 | @smallexample | |
13425 | @group | |
13426 | (defun count-words-region (beginning end) | |
13427 | "Print number of words in the region." | |
13428 | (interactive "r") | |
13429 | @end group | |
13430 | ||
13431 | @group | |
13432 | ;;; @r{1. Set up appropriate conditions.} | |
13433 | (message "Counting words in region ... ") | |
13434 | (save-excursion | |
13435 | (goto-char beginning) | |
13436 | @end group | |
13437 | ||
13438 | @group | |
13439 | ;;; @r{2. Count the words.} | |
13440 | (let ((count (recursive-count-words end))) | |
13441 | @end group | |
13442 | ||
13443 | @group | |
13444 | ;;; @r{3. Send a message to the user.} | |
13445 | (cond ((zerop count) | |
13446 | (message | |
13447 | "The region does NOT have any words.")) | |
13448 | ((= 1 count) | |
13449 | (message | |
13450 | "The region has 1 word.")) | |
13451 | (t | |
13452 | (message | |
13453 | "The region has %d words." count)))))) | |
13454 | @end group | |
13455 | @end smallexample | |
13456 | ||
13457 | Next, we need to write the recursive counting function. | |
13458 | ||
13459 | A recursive function has at least three parts: the `do-again-test', the | |
13460 | `next-step-expression', and the recursive call. | |
13461 | ||
13462 | The do-again-test determines whether the function will or will not be | |
13463 | called again. Since we are counting words in a region and can use a | |
13464 | function that moves point forward for every word, the do-again-test | |
13465 | can check whether point is still within the region. The do-again-test | |
13466 | should find the value of point and determine whether point is before, | |
13467 | at, or after the value of the end of the region. We can use the | |
13468 | @code{point} function to locate point. Clearly, we must pass the | |
13469 | value of the end of the region to the recursive counting function as an | |
13470 | argument. | |
13471 | ||
13472 | In addition, the do-again-test should also test whether the search finds a | |
13473 | word. If it does not, the function should not call itself again. | |
13474 | ||
13475 | The next-step-expression changes a value so that when the recursive | |
13476 | function is supposed to stop calling itself, it stops. More | |
13477 | precisely, the next-step-expression changes a value so that at the | |
13478 | right time, the do-again-test stops the recursive function from | |
13479 | calling itself again. In this case, the next-step-expression can be | |
13480 | the expression that moves point forward, word by word. | |
13481 | ||
13482 | The third part of a recursive function is the recursive call. | |
13483 | ||
13484 | Somewhere, also, we also need a part that does the `work' of the | |
13485 | function, a part that does the counting. A vital part! | |
13486 | ||
13487 | @need 1250 | |
13488 | But already, we have an outline of the recursive counting function: | |
13489 | ||
13490 | @smallexample | |
13491 | @group | |
13492 | (defun recursive-count-words (region-end) | |
13493 | "@var{documentation}@dots{}" | |
13494 | @var{do-again-test} | |
13495 | @var{next-step-expression} | |
13496 | @var{recursive call}) | |
13497 | @end group | |
13498 | @end smallexample | |
13499 | ||
13500 | Now we need to fill in the slots. Let's start with the simplest cases | |
13501 | first: if point is at or beyond the end of the region, there cannot | |
13502 | be any words in the region, so the function should return zero. | |
13503 | Likewise, if the search fails, there are no words to count, so the | |
13504 | function should return zero. | |
13505 | ||
13506 | On the other hand, if point is within the region and the search | |
13507 | succeeds, the function should call itself again. | |
13508 | ||
13509 | @need 800 | |
13510 | Thus, the do-again-test should look like this: | |
13511 | ||
13512 | @smallexample | |
13513 | @group | |
13514 | (and (< (point) region-end) | |
13515 | (re-search-forward "\\w+\\W*" region-end t)) | |
13516 | @end group | |
13517 | @end smallexample | |
13518 | ||
13519 | Note that the search expression is part of the do-again-test---the | |
13520 | function returns @code{t} if its search succeeds and @code{nil} if it | |
13521 | fails. (@xref{Whitespace Bug, , The Whitespace Bug in | |
13522 | @code{count-words-region}}, for an explanation of how | |
13523 | @code{re-search-forward} works.) | |
13524 | ||
13525 | The do-again-test is the true-or-false test of an @code{if} clause. | |
13526 | Clearly, if the do-again-test succeeds, the then-part of the @code{if} | |
13527 | clause should call the function again; but if it fails, the else-part | |
13528 | should return zero since either point is outside the region or the | |
13529 | search failed because there were no words to find. | |
13530 | ||
13531 | But before considering the recursive call, we need to consider the | |
13532 | next-step-expression. What is it? Interestingly, it is the search | |
13533 | part of the do-again-test. | |
13534 | ||
13535 | In addition to returning @code{t} or @code{nil} for the | |
13536 | do-again-test, @code{re-search-forward} moves point forward as a side | |
13537 | effect of a successful search. This is the action that changes the | |
13538 | value of point so that the recursive function stops calling itself | |
13539 | when point completes its movement through the region. Consequently, | |
13540 | the @code{re-search-forward} expression is the next-step-expression. | |
13541 | ||
13542 | @need 1200 | |
13543 | In outline, then, the body of the @code{recursive-count-words} | |
13544 | function looks like this: | |
13545 | ||
13546 | @smallexample | |
13547 | @group | |
13548 | (if @var{do-again-test-and-next-step-combined} | |
13549 | ;; @r{then} | |
13550 | @var{recursive-call-returning-count} | |
13551 | ;; @r{else} | |
13552 | @var{return-zero}) | |
13553 | @end group | |
13554 | @end smallexample | |
13555 | ||
13556 | How to incorporate the mechanism that counts? | |
13557 | ||
13558 | If you are not used to writing recursive functions, a question like | |
13559 | this can be troublesome. But it can and should be approached | |
13560 | systematically. | |
13561 | ||
13562 | We know that the counting mechanism should be associated in some way | |
13563 | with the recursive call. Indeed, since the next-step-expression moves | |
13564 | point forward by one word, and since a recursive call is made for | |
13565 | each word, the counting mechanism must be an expression that adds one | |
13566 | to the value returned by a call to @code{recursive-count-words}. | |
13567 | ||
13568 | Consider several cases: | |
13569 | ||
13570 | @itemize @bullet | |
13571 | @item | |
13572 | If there are two words in the region, the function should return | |
13573 | a value resulting from adding one to the value returned when it counts | |
13574 | the first word, plus the number returned when it counts the remaining | |
13575 | words in the region, which in this case is one. | |
13576 | ||
13577 | @item | |
13578 | If there is one word in the region, the function should return | |
13579 | a value resulting from adding one to the value returned when it counts | |
13580 | that word, plus the number returned when it counts the remaining | |
13581 | words in the region, which in this case is zero. | |
13582 | ||
13583 | @item | |
13584 | If there are no words in the region, the function should return zero. | |
13585 | @end itemize | |
13586 | ||
13587 | From the sketch we can see that the else-part of the @code{if} returns | |
13588 | zero for the case of no words. This means that the then-part of the | |
13589 | @code{if} must return a value resulting from adding one to the value | |
13590 | returned from a count of the remaining words. | |
13591 | ||
13592 | @need 1200 | |
13593 | The expression will look like this, where @code{1+} is a function that | |
13594 | adds one to its argument. | |
13595 | ||
13596 | @smallexample | |
13597 | (1+ (recursive-count-words region-end)) | |
13598 | @end smallexample | |
13599 | ||
13600 | @need 1200 | |
13601 | The whole @code{recursive-count-words} function will then look like | |
13602 | this: | |
13603 | ||
13604 | @smallexample | |
13605 | @group | |
13606 | (defun recursive-count-words (region-end) | |
13607 | "@var{documentation}@dots{}" | |
13608 | ||
13609 | ;;; @r{1. do-again-test} | |
13610 | (if (and (< (point) region-end) | |
13611 | (re-search-forward "\\w+\\W*" region-end t)) | |
13612 | @end group | |
13613 | ||
13614 | @group | |
13615 | ;;; @r{2. then-part: the recursive call} | |
13616 | (1+ (recursive-count-words region-end)) | |
13617 | ||
13618 | ;;; @r{3. else-part} | |
13619 | 0)) | |
13620 | @end group | |
13621 | @end smallexample | |
13622 | ||
13623 | @need 1250 | |
13624 | Let's examine how this works: | |
13625 | ||
13626 | If there are no words in the region, the else part of the @code{if} | |
13627 | expression is evaluated and consequently the function returns zero. | |
13628 | ||
13629 | If there is one word in the region, the value of point is less than | |
13630 | the value of @code{region-end} and the search succeeds. In this case, | |
13631 | the true-or-false-test of the @code{if} expression tests true, and the | |
13632 | then-part of the @code{if} expression is evaluated. The counting | |
13633 | expression is evaluated. This expression returns a value (which will | |
13634 | be the value returned by the whole function) that is the sum of one | |
13635 | added to the value returned by a recursive call. | |
13636 | ||
13637 | Meanwhile, the next-step-expression has caused point to jump over the | |
13638 | first (and in this case only) word in the region. This means that | |
13639 | when @code{(recursive-count-words region-end)} is evaluated a second | |
13640 | time, as a result of the recursive call, the value of point will be | |
13641 | equal to or greater than the value of region end. So this time, | |
13642 | @code{recursive-count-words} will return zero. The zero will be added | |
13643 | to one, and the original evaluation of @code{recursive-count-words} | |
13644 | will return one plus zero, which is one, which is the correct amount. | |
13645 | ||
13646 | Clearly, if there are two words in the region, the first call to | |
13647 | @code{recursive-count-words} returns one added to the value returned | |
13648 | by calling @code{recursive-count-words} on a region containing the | |
13649 | remaining word---that is, it adds one to one, producing two, which is | |
13650 | the correct amount. | |
13651 | ||
13652 | Similarly, if there are three words in the region, the first call to | |
13653 | @code{recursive-count-words} returns one added to the value returned | |
13654 | by calling @code{recursive-count-words} on a region containing the | |
13655 | remaining two words---and so on and so on. | |
13656 | ||
13657 | @need 1250 | |
13658 | @noindent | |
13659 | With full documentation the two functions look like this: | |
13660 | ||
13661 | @need 1250 | |
13662 | @noindent | |
13663 | The recursive function: | |
13664 | ||
13665 | @findex recursive-count-words | |
13666 | @smallexample | |
13667 | @group | |
13668 | (defun recursive-count-words (region-end) | |
13669 | "Number of words between point and REGION-END." | |
13670 | @end group | |
13671 | ||
13672 | @group | |
13673 | ;;; @r{1. do-again-test} | |
13674 | (if (and (< (point) region-end) | |
13675 | (re-search-forward "\\w+\\W*" region-end t)) | |
13676 | @end group | |
13677 | ||
13678 | @group | |
13679 | ;;; @r{2. then-part: the recursive call} | |
13680 | (1+ (recursive-count-words region-end)) | |
13681 | ||
13682 | ;;; @r{3. else-part} | |
13683 | 0)) | |
13684 | @end group | |
13685 | @end smallexample | |
13686 | ||
13687 | @need 800 | |
13688 | @noindent | |
13689 | The wrapper: | |
13690 | ||
13691 | @smallexample | |
13692 | @group | |
13693 | ;;; @r{Recursive version} | |
13694 | (defun count-words-region (beginning end) | |
13695 | "Print number of words in the region. | |
13696 | @end group | |
13697 | ||
13698 | @group | |
13699 | Words are defined as at least one word-constituent | |
13700 | character followed by at least one character that is | |
13701 | not a word-constituent. The buffer's syntax table | |
13702 | determines which characters these are." | |
13703 | @end group | |
13704 | @group | |
13705 | (interactive "r") | |
13706 | (message "Counting words in region ... ") | |
13707 | (save-excursion | |
13708 | (goto-char beginning) | |
13709 | (let ((count (recursive-count-words end))) | |
13710 | @end group | |
13711 | @group | |
13712 | (cond ((zerop count) | |
13713 | (message | |
13714 | "The region does NOT have any words.")) | |
13715 | @end group | |
13716 | @group | |
13717 | ((= 1 count) | |
13718 | (message "The region has 1 word.")) | |
13719 | (t | |
13720 | (message | |
13721 | "The region has %d words." count)))))) | |
13722 | @end group | |
13723 | @end smallexample | |
13724 | ||
13725 | @node Counting Exercise, , recursive-count-words, Counting Words | |
13726 | @section Exercise: Counting Punctuation | |
13727 | ||
13728 | Using a @code{while} loop, write a function to count the number of | |
13729 | punctuation marks in a region---period, comma, semicolon, colon, | |
13730 | exclamation mark, and question mark. Do the same using recursion. | |
13731 | ||
13732 | @node Words in a defun, Readying a Graph, Counting Words, Top | |
13733 | @chapter Counting Words in a @code{defun} | |
13734 | @cindex Counting words in a @code{defun} | |
13735 | @cindex Word counting in a @code{defun} | |
13736 | ||
13737 | Our next project is to count the number of words in a function | |
13738 | definition. Clearly, this can be done using some variant of | |
13739 | @code{count-word-region}. @xref{Counting Words, , Counting Words: | |
13740 | Repetition and Regexps}. If we are just going to count the words in | |
13741 | one definition, it is easy enough to mark the definition with the | |
13742 | @kbd{C-M-h} (@code{mark-defun}) command, and then call | |
13743 | @code{count-word-region}. | |
13744 | ||
13745 | However, I am more ambitious: I want to count the words and symbols in | |
13746 | every definition in the Emacs sources and then print a graph that | |
13747 | shows how many functions there are of each length: how many contain 40 | |
13748 | to 49 words or symbols, how many contain 50 to 59 words or symbols, | |
13749 | and so on. I have often been curious how long a typical function is, | |
13750 | and this will tell. | |
13751 | ||
13752 | @menu | |
13753 | * Divide and Conquer:: | |
13754 | * Words and Symbols:: What to count? | |
13755 | * Syntax:: What constitutes a word or symbol? | |
13756 | * count-words-in-defun:: Very like @code{count-words}. | |
13757 | * Several defuns:: Counting several defuns in a file. | |
13758 | * Find a File:: Do you want to look at a file? | |
13759 | * lengths-list-file:: A list of the lengths of many definitions. | |
13760 | * Several files:: Counting in definitions in different files. | |
13761 | * Several files recursively:: Recursively counting in different files. | |
13762 | * Prepare the data:: Prepare the data for display in a graph. | |
13763 | @end menu | |
13764 | ||
13765 | @node Divide and Conquer, Words and Symbols, Words in a defun, Words in a defun | |
13766 | @ifnottex | |
13767 | @unnumberedsec Divide and Conquer | |
13768 | @end ifnottex | |
13769 | ||
13770 | Described in one phrase, the histogram project is daunting; but | |
13771 | divided into numerous small steps, each of which we can take one at a | |
13772 | time, the project becomes less fearsome. Let us consider what the | |
13773 | steps must be: | |
13774 | ||
13775 | @itemize @bullet | |
13776 | @item | |
13777 | First, write a function to count the words in one definition. This | |
13778 | includes the problem of handling symbols as well as words. | |
13779 | ||
13780 | @item | |
13781 | Second, write a function to list the numbers of words in each function | |
13782 | in a file. This function can use the @code{count-words-in-defun} | |
13783 | function. | |
13784 | ||
13785 | @item | |
13786 | Third, write a function to list the numbers of words in each function | |
13787 | in each of several files. This entails automatically finding the | |
13788 | various files, switching to them, and counting the words in the | |
13789 | definitions within them. | |
13790 | ||
13791 | @item | |
13792 | Fourth, write a function to convert the list of numbers that we | |
13793 | created in step three to a form that will be suitable for printing as | |
13794 | a graph. | |
13795 | ||
13796 | @item | |
13797 | Fifth, write a function to print the results as a graph. | |
13798 | @end itemize | |
13799 | ||
13800 | This is quite a project! But if we take each step slowly, it will not | |
13801 | be difficult. | |
13802 | ||
13803 | @node Words and Symbols, Syntax, Divide and Conquer, Words in a defun | |
13804 | @section What to Count? | |
13805 | @cindex Words and symbols in defun | |
13806 | ||
13807 | When we first start thinking about how to count the words in a | |
13808 | function definition, the first question is (or ought to be) what are | |
13809 | we going to count? When we speak of `words' with respect to a Lisp | |
13810 | function definition, we are actually speaking, in large part, of | |
13811 | `symbols'. For example, the following @code{multiply-by-seven} | |
13812 | function contains the five symbols @code{defun}, | |
13813 | @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In | |
13814 | addition, in the documentation string, it contains the four words | |
13815 | @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The | |
13816 | symbol @samp{number} is repeated, so the definition contains a total | |
13817 | of ten words and symbols. | |
13818 | ||
13819 | @smallexample | |
13820 | @group | |
13821 | (defun multiply-by-seven (number) | |
13822 | "Multiply NUMBER by seven." | |
13823 | (* 7 number)) | |
13824 | @end group | |
13825 | @end smallexample | |
13826 | ||
13827 | @noindent | |
13828 | However, if we mark the @code{multiply-by-seven} definition with | |
13829 | @kbd{C-M-h} (@code{mark-defun}), and then call | |
13830 | @code{count-words-region} on it, we will find that | |
13831 | @code{count-words-region} claims the definition has eleven words, not | |
13832 | ten! Something is wrong! | |
13833 | ||
13834 | The problem is twofold: @code{count-words-region} does not count the | |
13835 | @samp{*} as a word, and it counts the single symbol, | |
13836 | @code{multiply-by-seven}, as containing three words. The hyphens are | |
13837 | treated as if they were interword spaces rather than intraword | |
13838 | connectors: @samp{multiply-by-seven} is counted as if it were written | |
13839 | @samp{multiply by seven}. | |
13840 | ||
13841 | The cause of this confusion is the regular expression search within | |
13842 | the @code{count-words-region} definition that moves point forward word | |
13843 | by word. In the canonical version of @code{count-words-region}, the | |
13844 | regexp is: | |
13845 | ||
13846 | @smallexample | |
13847 | "\\w+\\W*" | |
13848 | @end smallexample | |
13849 | ||
13850 | @noindent | |
13851 | This regular expression is a pattern defining one or more word | |
13852 | constituent characters possibly followed by one or more characters | |
13853 | that are not word constituents. What is meant by `word constituent | |
13854 | characters' brings us to the issue of syntax, which is worth a section | |
13855 | of its own. | |
13856 | ||
13857 | @node Syntax, count-words-in-defun, Words and Symbols, Words in a defun | |
13858 | @section What Constitutes a Word or Symbol? | |
13859 | @cindex Syntax categories and tables | |
13860 | ||
13861 | Emacs treats different characters as belonging to different | |
13862 | @dfn{syntax categories}. For example, the regular expression, | |
13863 | @samp{\\w+}, is a pattern specifying one or more @emph{word | |
13864 | constituent} characters. Word constituent characters are members of | |
13865 | one syntax category. Other syntax categories include the class of | |
13866 | punctuation characters, such as the period and the comma, and the | |
13867 | class of whitespace characters, such as the blank space and the tab | |
13868 | character. (For more information, see @ref{Syntax, Syntax, The Syntax | |
13869 | Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, , Syntax | |
13870 | Tables, elisp, The GNU Emacs Lisp Reference Manual}.) | |
13871 | ||
13872 | Syntax tables specify which characters belong to which categories. | |
13873 | Usually, a hyphen is not specified as a `word constituent character'. | |
13874 | Instead, it is specified as being in the `class of characters that are | |
13875 | part of symbol names but not words.' This means that the | |
13876 | @code{count-words-region} function treats it in the same way it treats | |
13877 | an interword white space, which is why @code{count-words-region} | |
13878 | counts @samp{multiply-by-seven} as three words. | |
13879 | ||
13880 | There are two ways to cause Emacs to count @samp{multiply-by-seven} as | |
13881 | one symbol: modify the syntax table or modify the regular expression. | |
13882 | ||
13883 | We could redefine a hyphen as a word constituent character by | |
13884 | modifying the syntax table that Emacs keeps for each mode. This | |
13885 | action would serve our purpose, except that a hyphen is merely the | |
13886 | most common character within symbols that is not typically a word | |
13887 | constituent character; there are others, too. | |
13888 | ||
13889 | Alternatively, we can redefine the regular expression used in the | |
13890 | @code{count-words} definition so as to include symbols. This | |
13891 | procedure has the merit of clarity, but the task is a little tricky. | |
13892 | ||
13893 | @need 1200 | |
13894 | The first part is simple enough: the pattern must match ``at least one | |
13895 | character that is a word or symbol constituent''. Thus: | |
13896 | ||
13897 | @smallexample | |
13898 | "\\(\\w\\|\\s_\\)+" | |
13899 | @end smallexample | |
13900 | ||
13901 | @noindent | |
13902 | The @samp{\\(} is the first part of the grouping construct that | |
13903 | includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated | |
13904 | by the @samp{\\|}. The @samp{\\w} matches any word-constituent | |
13905 | character and the @samp{\\s_} matches any character that is part of a | |
13906 | symbol name but not a word-constituent character. The @samp{+} | |
13907 | following the group indicates that the word or symbol constituent | |
13908 | characters must be matched at least once. | |
13909 | ||
13910 | However, the second part of the regexp is more difficult to design. | |
13911 | What we want is to follow the first part with ``optionally one or more | |
13912 | characters that are not constituents of a word or symbol''. At first, | |
13913 | I thought I could define this with the following: | |
13914 | ||
13915 | @smallexample | |
13916 | "\\(\\W\\|\\S_\\)*" | |
13917 | @end smallexample | |
13918 | ||
13919 | @noindent | |
13920 | The upper case @samp{W} and @samp{S} match characters that are | |
13921 | @emph{not} word or symbol constituents. Unfortunately, this | |
13922 | expression matches any character that is either not a word constituent | |
13923 | or not a symbol constituent. This matches any character! | |
13924 | ||
13925 | I then noticed that every word or symbol in my test region was | |
13926 | followed by white space (blank space, tab, or newline). So I tried | |
13927 | placing a pattern to match one or more blank spaces after the pattern | |
13928 | for one or more word or symbol constituents. This failed, too. Words | |
13929 | and symbols are often separated by whitespace, but in actual code | |
13930 | parentheses may follow symbols and punctuation may follow words. So | |
13931 | finally, I designed a pattern in which the word or symbol constituents | |
13932 | are followed optionally by characters that are not white space and | |
13933 | then followed optionally by white space. | |
13934 | ||
13935 | @need 800 | |
13936 | Here is the full regular expression: | |
13937 | ||
13938 | @smallexample | |
13939 | "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" | |
13940 | @end smallexample | |
13941 | ||
13942 | @node count-words-in-defun, Several defuns, Syntax, Words in a defun | |
13943 | @section The @code{count-words-in-defun} Function | |
13944 | @cindex Counting words in a @code{defun} | |
13945 | ||
13946 | We have seen that there are several ways to write a | |
13947 | @code{count-word-region} function. To write a | |
13948 | @code{count-words-in-defun}, we need merely adapt one of these | |
13949 | versions. | |
13950 | ||
13951 | The version that uses a @code{while} loop is easy to understand, so I | |
13952 | am going to adapt that. Because @code{count-words-in-defun} will be | |
13953 | part of a more complex program, it need not be interactive and it need | |
13954 | not display a message but just return the count. These considerations | |
13955 | simplify the definition a little. | |
13956 | ||
13957 | On the other hand, @code{count-words-in-defun} will be used within a | |
13958 | buffer that contains function definitions. Consequently, it is | |
13959 | reasonable to ask that the function determine whether it is called | |
13960 | when point is within a function definition, and if it is, to return | |
13961 | the count for that definition. This adds complexity to the | |
13962 | definition, but saves us from needing to pass arguments to the | |
13963 | function. | |
13964 | ||
13965 | @need 1250 | |
13966 | These considerations lead us to prepare the following template: | |
13967 | ||
13968 | @smallexample | |
13969 | @group | |
13970 | (defun count-words-in-defun () | |
13971 | "@var{documentation}@dots{}" | |
13972 | (@var{set up}@dots{} | |
13973 | (@var{while loop}@dots{}) | |
13974 | @var{return count}) | |
13975 | @end group | |
13976 | @end smallexample | |
13977 | ||
13978 | @noindent | |
13979 | As usual, our job is to fill in the slots. | |
13980 | ||
13981 | First, the set up. | |
13982 | ||
13983 | We are presuming that this function will be called within a buffer | |
13984 | containing function definitions. Point will either be within a | |
13985 | function definition or not. For @code{count-words-in-defun} to work, | |
13986 | point must move to the beginning of the definition, a counter must | |
13987 | start at zero, and the counting loop must stop when point reaches the | |
13988 | end of the definition. | |
13989 | ||
13990 | The @code{beginning-of-defun} function searches backwards for an | |
13991 | opening delimiter such as a @samp{(} at the beginning of a line, and | |
13992 | moves point to that position, or else to the limit of the search. In | |
13993 | practice, this means that @code{beginning-of-defun} moves point to the | |
13994 | beginning of an enclosing or preceding function definition, or else to | |
13995 | the beginning of the buffer. We can use @code{beginning-of-defun} to | |
13996 | place point where we wish to start. | |
13997 | ||
13998 | The @code{while} loop requires a counter to keep track of the words or | |
13999 | symbols being counted. A @code{let} expression can be used to create | |
14000 | a local variable for this purpose, and bind it to an initial value of zero. | |
14001 | ||
14002 | The @code{end-of-defun} function works like @code{beginning-of-defun} | |
14003 | except that it moves point to the end of the definition. | |
14004 | @code{end-of-defun} can be used as part of an expression that | |
14005 | determines the position of the end of the definition. | |
14006 | ||
14007 | The set up for @code{count-words-in-defun} takes shape rapidly: first | |
14008 | we move point to the beginning of the definition, then we create a | |
14009 | local variable to hold the count, and finally, we record the position | |
14010 | of the end of the definition so the @code{while} loop will know when to stop | |
14011 | looping. | |
14012 | ||
14013 | @need 1250 | |
14014 | The code looks like this: | |
14015 | ||
14016 | @smallexample | |
14017 | @group | |
14018 | (beginning-of-defun) | |
14019 | (let ((count 0) | |
14020 | (end (save-excursion (end-of-defun) (point)))) | |
14021 | @end group | |
14022 | @end smallexample | |
14023 | ||
14024 | @noindent | |
14025 | The code is simple. The only slight complication is likely to concern | |
14026 | @code{end}: it is bound to the position of the end of the definition | |
14027 | by a @code{save-excursion} expression that returns the value of point | |
14028 | after @code{end-of-defun} temporarily moves it to the end of the | |
14029 | definition. | |
14030 | ||
14031 | The second part of the @code{count-words-in-defun}, after the set up, | |
14032 | is the @code{while} loop. | |
14033 | ||
14034 | The loop must contain an expression that jumps point forward word by | |
14035 | word and symbol by symbol, and another expression that counts the | |
14036 | jumps. The true-or-false-test for the @code{while} loop should test | |
14037 | true so long as point should jump forward, and false when point is at | |
14038 | the end of the definition. We have already redefined the regular | |
14039 | expression for this (@pxref{Syntax}), so the loop is straightforward: | |
14040 | ||
14041 | @smallexample | |
14042 | @group | |
14043 | (while (and (< (point) end) | |
14044 | (re-search-forward | |
14045 | "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t) | |
14046 | (setq count (1+ count))) | |
14047 | @end group | |
14048 | @end smallexample | |
14049 | ||
14050 | The third part of the function definition returns the count of words | |
14051 | and symbols. This part is the last expression within the body of the | |
14052 | @code{let} expression, and can be, very simply, the local variable | |
14053 | @code{count}, which when evaluated returns the count. | |
14054 | ||
14055 | @need 1250 | |
14056 | Put together, the @code{count-words-in-defun} definition looks like this: | |
14057 | ||
14058 | @findex count-words-in-defun | |
14059 | @smallexample | |
14060 | @group | |
14061 | (defun count-words-in-defun () | |
14062 | "Return the number of words and symbols in a defun." | |
14063 | (beginning-of-defun) | |
14064 | (let ((count 0) | |
14065 | (end (save-excursion (end-of-defun) (point)))) | |
14066 | @end group | |
14067 | @group | |
14068 | (while | |
14069 | (and (< (point) end) | |
14070 | (re-search-forward | |
14071 | "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" | |
14072 | end t)) | |
14073 | (setq count (1+ count))) | |
14074 | count)) | |
14075 | @end group | |
14076 | @end smallexample | |
14077 | ||
14078 | How to test this? The function is not interactive, but it is easy to | |
14079 | put a wrapper around the function to make it interactive; we can use | |
14080 | almost the same code as for the recursive version of | |
14081 | @code{count-words-region}: | |
14082 | ||
14083 | @smallexample | |
14084 | @group | |
14085 | ;;; @r{Interactive version.} | |
14086 | (defun count-words-defun () | |
14087 | "Number of words and symbols in a function definition." | |
14088 | (interactive) | |
14089 | (message | |
14090 | "Counting words and symbols in function definition ... ") | |
14091 | @end group | |
14092 | @group | |
14093 | (let ((count (count-words-in-defun))) | |
14094 | (cond | |
14095 | ((zerop count) | |
14096 | (message | |
14097 | "The definition does NOT have any words or symbols.")) | |
14098 | @end group | |
14099 | @group | |
14100 | ((= 1 count) | |
14101 | (message | |
14102 | "The definition has 1 word or symbol.")) | |
14103 | (t | |
14104 | (message | |
14105 | "The definition has %d words or symbols." count))))) | |
14106 | @end group | |
14107 | @end smallexample | |
14108 | ||
14109 | @need 800 | |
14110 | @noindent | |
14111 | Let's re-use @kbd{C-c =} as a convenient keybinding: | |
14112 | ||
14113 | @smallexample | |
14114 | (global-set-key "\C-c=" 'count-words-defun) | |
14115 | @end smallexample | |
14116 | ||
14117 | Now we can try out @code{count-words-defun}: install both | |
14118 | @code{count-words-in-defun} and @code{count-words-defun}, and set the | |
14119 | keybinding, and then place the cursor within the following definition: | |
14120 | ||
14121 | @smallexample | |
14122 | @group | |
14123 | (defun multiply-by-seven (number) | |
14124 | "Multiply NUMBER by seven." | |
14125 | (* 7 number)) | |
14126 | @result{} 10 | |
14127 | @end group | |
14128 | @end smallexample | |
14129 | ||
14130 | @noindent | |
14131 | Success! The definition has 10 words and symbols. | |
14132 | ||
14133 | The next problem is to count the numbers of words and symbols in | |
14134 | several definitions within a single file. | |
14135 | ||
14136 | @node Several defuns, Find a File, count-words-in-defun, Words in a defun | |
14137 | @section Count Several @code{defuns} Within a File | |
14138 | ||
14139 | A file such as @file{simple.el} may have 80 or more function | |
14140 | definitions within it. Our long term goal is to collect statistics on | |
14141 | many files, but as a first step, our immediate goal is to collect | |
14142 | statistics on one file. | |
14143 | ||
14144 | The information will be a series of numbers, each number being the | |
14145 | length of a function definition. We can store the numbers in a list. | |
14146 | ||
14147 | We know that we will want to incorporate the information regarding one | |
14148 | file with information about many other files; this means that the | |
14149 | function for counting definition lengths within one file need only | |
14150 | return the list of lengths. It need not and should not display any | |
14151 | messages. | |
14152 | ||
14153 | The word count commands contain one expression to jump point forward | |
14154 | word by word and another expression to count the jumps. The function | |
14155 | to return the lengths of definitions can be designed to work the same | |
14156 | way, with one expression to jump point forward definition by | |
14157 | definition and another expression to construct the lengths' list. | |
14158 | ||
14159 | This statement of the problem makes it elementary to write the | |
14160 | function definition. Clearly, we will start the count at the | |
14161 | beginning of the file, so the first command will be @code{(goto-char | |
14162 | (point-min))}. Next, we start the @code{while} loop; and the | |
14163 | true-or-false test of the loop can be a regular expression search for | |
14164 | the next function definition---so long as the search succeeds, point | |
14165 | is moved forward and then the body of the loop is evaluated. The body | |
14166 | needs an expression that constructs the lengths' list. @code{cons}, | |
14167 | the list construction command, can be used to create the list. That | |
14168 | is almost all there is to it. | |
14169 | ||
14170 | @need 800 | |
14171 | Here is what this fragment of code looks like: | |
14172 | ||
14173 | @smallexample | |
14174 | @group | |
14175 | (goto-char (point-min)) | |
14176 | (while (re-search-forward "^(defun" nil t) | |
14177 | (setq lengths-list | |
14178 | (cons (count-words-in-defun) lengths-list))) | |
14179 | @end group | |
14180 | @end smallexample | |
14181 | ||
14182 | What we have left out is the mechanism for finding the file that | |
14183 | contains the function definitions. | |
14184 | ||
14185 | In previous examples, we either used this, the Info file, or we | |
14186 | switched back and forth to some other buffer, such as the | |
14187 | @file{*scratch*} buffer. | |
14188 | ||
14189 | Finding a file is a new process that we have not yet discussed. | |
14190 | ||
14191 | @node Find a File, lengths-list-file, Several defuns, Words in a defun | |
14192 | @comment node-name, next, previous, up | |
14193 | @section Find a File | |
14194 | @cindex Find a File | |
14195 | ||
14196 | To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file}) | |
14197 | command. This command is almost, but not quite right for the lengths | |
14198 | problem. | |
14199 | ||
14200 | @need 1200 | |
14201 | Let's look at the source for @code{find-file} (you can use the | |
14202 | @code{find-tag} command or @kbd{C-h f} (@code{describe-function}) to | |
14203 | find the source of a function): | |
14204 | ||
14205 | @smallexample | |
14206 | @group | |
14207 | (defun find-file (filename) | |
14208 | "Edit file FILENAME. | |
14209 | Switch to a buffer visiting file FILENAME, | |
14210 | creating one if none already exists." | |
14211 | (interactive "FFind file: ") | |
14212 | (switch-to-buffer (find-file-noselect filename))) | |
14213 | @end group | |
14214 | @end smallexample | |
14215 | ||
14216 | The definition possesses short but complete documentation and an | |
14217 | interactive specification that prompts you for a file name when you | |
14218 | use the command interactively. The body of the definition contains | |
14219 | two functions, @code{find-file-noselect} and @code{switch-to-buffer}. | |
14220 | ||
14221 | According to its documentation as shown by @kbd{C-h f} (the | |
14222 | @code{describe-function} command), the @code{find-file-noselect} | |
14223 | function reads the named file into a buffer and returns the buffer. | |
14224 | However, the buffer is not selected. Emacs does not switch its | |
14225 | attention (or yours if you are using @code{find-file-noselect}) to the | |
14226 | named buffer. That is what @code{switch-to-buffer} does: it switches | |
14227 | the buffer to which Emacs attention is directed; and it switches the | |
14228 | buffer displayed in the window to the new buffer. We have discussed | |
14229 | buffer switching elsewhere. (@xref{Switching Buffers}.) | |
14230 | ||
14231 | In this histogram project, we do not need to display each file on the | |
14232 | screen as the program determines the length of each definition within | |
14233 | it. Instead of employing @code{switch-to-buffer}, we can work with | |
14234 | @code{set-buffer}, which redirects the attention of the computer | |
14235 | program to a different buffer but does not redisplay it on the screen. | |
14236 | So instead of calling on @code{find-file} to do the job, we must write | |
14237 | our own expression. | |
14238 | ||
14239 | The task is easy: use @code{find-file-noselect} and @code{set-buffer}. | |
14240 | ||
14241 | @node lengths-list-file, Several files, Find a File, Words in a defun | |
14242 | @section @code{lengths-list-file} in Detail | |
14243 | ||
14244 | The core of the @code{lengths-list-file} function is a @code{while} | |
14245 | loop containing a function to move point forward `defun by defun' and | |
14246 | a function to count the number of words and symbols in each defun. | |
14247 | This core must be surrounded by functions that do various other tasks, | |
14248 | including finding the file, and ensuring that point starts out at the | |
14249 | beginning of the file. The function definition looks like this: | |
14250 | @findex lengths-list-file | |
14251 | ||
14252 | @smallexample | |
14253 | @group | |
14254 | (defun lengths-list-file (filename) | |
14255 | "Return list of definitions' lengths within FILE. | |
14256 | The returned list is a list of numbers. | |
14257 | Each number is the number of words or | |
14258 | symbols in one function definition." | |
14259 | @end group | |
14260 | @group | |
14261 | (message "Working on `%s' ... " filename) | |
14262 | (save-excursion | |
14263 | (let ((buffer (find-file-noselect filename)) | |
14264 | (lengths-list)) | |
14265 | (set-buffer buffer) | |
14266 | (setq buffer-read-only t) | |
14267 | (widen) | |
14268 | (goto-char (point-min)) | |
14269 | (while (re-search-forward "^(defun" nil t) | |
14270 | (setq lengths-list | |
14271 | (cons (count-words-in-defun) lengths-list))) | |
14272 | (kill-buffer buffer) | |
14273 | lengths-list))) | |
14274 | @end group | |
14275 | @end smallexample | |
14276 | ||
14277 | @noindent | |
14278 | The function is passed one argument, the name of the file on which it | |
14279 | will work. It has four lines of documentation, but no interactive | |
14280 | specification. Since people worry that a computer is broken if they | |
14281 | don't see anything going on, the first line of the body is a | |
14282 | message. | |
14283 | ||
14284 | The next line contains a @code{save-excursion} that returns Emacs' | |
14285 | attention to the current buffer when the function completes. This is | |
14286 | useful in case you embed this function in another function that | |
14287 | presumes point is restored to the original buffer. | |
14288 | ||
14289 | In the varlist of the @code{let} expression, Emacs finds the file and | |
14290 | binds the local variable @code{buffer} to the buffer containing the | |
14291 | file. At the same time, Emacs creates @code{lengths-list} as a local | |
14292 | variable. | |
14293 | ||
14294 | Next, Emacs switches its attention to the buffer. | |
14295 | ||
14296 | In the following line, Emacs makes the buffer read-only. Ideally, | |
14297 | this line is not necessary. None of the functions for counting words | |
14298 | and symbols in a function definition should change the buffer. | |
14299 | Besides, the buffer is not going to be saved, even if it were changed. | |
14300 | This line is entirely the consequence of great, perhaps excessive, | |
14301 | caution. The reason for the caution is that this function and those | |
14302 | it calls work on the sources for Emacs and it is very inconvenient if | |
14303 | they are inadvertently modified. It goes without saying that I did | |
14304 | not realize a need for this line until an experiment went awry and | |
14305 | started to modify my Emacs source files @dots{} | |
14306 | ||
14307 | Next comes a call to widen the buffer if it is narrowed. This | |
14308 | function is usually not needed---Emacs creates a fresh buffer if none | |
14309 | already exists; but if a buffer visiting the file already exists Emacs | |
14310 | returns that one. In this case, the buffer may be narrowed and must | |
14311 | be widened. If we wanted to be fully `user-friendly', we would | |
14312 | arrange to save the restriction and the location of point, but we | |
14313 | won't. | |
14314 | ||
14315 | The @code{(goto-char (point-min))} expression moves point to the | |
14316 | beginning of the buffer. | |
14317 | ||
14318 | Then comes a @code{while} loop in which the `work' of the function is | |
14319 | carried out. In the loop, Emacs determines the length of each | |
14320 | definition and constructs a lengths' list containing the information. | |
14321 | ||
14322 | Emacs kills the buffer after working through it. This is to save | |
14323 | space inside of Emacs. My version of Emacs 19 contained over 300 | |
14324 | source files of interest; Emacs 21 contains over 800 source files. | |
14325 | Another function will apply @code{lengths-list-file} to each of the | |
14326 | files. | |
14327 | ||
14328 | Finally, the last expression within the @code{let} expression is the | |
14329 | @code{lengths-list} variable; its value is returned as the value of | |
14330 | the whole function. | |
14331 | ||
14332 | You can try this function by installing it in the usual fashion. Then | |
14333 | place your cursor after the following expression and type @kbd{C-x | |
14334 | C-e} (@code{eval-last-sexp}). | |
14335 | ||
14336 | @c !!! 21.0.100 lisp sources location here | |
14337 | @smallexample | |
14338 | (lengths-list-file | |
14339 | "/usr/local/share/emacs/21.0.100/lisp/emacs-lisp/debug.el") | |
14340 | @end smallexample | |
14341 | ||
14342 | @c was: (lengths-list-file "../lisp/debug.el") | |
14343 | @c !!! as of 21, Info file is in | |
14344 | @c /usr/share/info/emacs-lisp-intro.info.gz | |
14345 | @c but debug.el is in /usr/local/share/emacs/21.0.100/lisp/emacs-lisp/debug.el | |
14346 | ||
14347 | @noindent | |
14348 | (You may need to change the pathname of the file; the one here worked | |
14349 | with GNU Emacs version 21.0.100. To change the expression, copy it to | |
14350 | the @file{*scratch*} buffer and edit it. | |
14351 | ||
14352 | @need 1200 | |
14353 | @noindent | |
14354 | (Also, to see the full length of the list, rather than a truncated | |
14355 | version, you may have to evaluate the following: | |
14356 | ||
14357 | @smallexample | |
14358 | (custom-set-variables '(eval-expression-print-length nil)) | |
14359 | @end smallexample | |
14360 | ||
14361 | @noindent | |
14362 | (@xref{defcustom, , Setting Variables with @code{defcustom}}. | |
14363 | Then evaluate the @code{lengths-list-file} expression.) | |
14364 | ||
14365 | @need 1200 | |
14366 | The lengths' list for @file{debug.el} takes less than a second to | |
14367 | produce and looks like this: | |
14368 | ||
14369 | @smallexample | |
14370 | (77 95 85 87 131 89 50 25 44 44 68 35 64 45 17 34 167 457) | |
14371 | @end smallexample | |
14372 | ||
14373 | @need 1500 | |
14374 | (Using my old machine, the version 19 lengths' list for @file{debug.el} | |
14375 | took seven seconds to produce and looked like this: | |
14376 | ||
14377 | @smallexample | |
14378 | (75 41 80 62 20 45 44 68 45 12 34 235) | |
14379 | @end smallexample | |
14380 | ||
14381 | (The newer version of @file{debug.el} contains more defuns than the | |
14382 | earlier one; and my new machine is much faster than the old one.) | |
14383 | ||
14384 | Note that the length of the last definition in the file is first in | |
14385 | the list. | |
14386 | ||
14387 | @node Several files, Several files recursively, lengths-list-file, Words in a defun | |
14388 | @section Count Words in @code{defuns} in Different Files | |
14389 | ||
14390 | In the previous section, we created a function that returns a list of | |
14391 | the lengths of each definition in a file. Now, we want to define a | |
14392 | function to return a master list of the lengths of the definitions in | |
14393 | a list of files. | |
14394 | ||
14395 | Working on each of a list of files is a repetitious act, so we can use | |
14396 | either a @code{while} loop or recursion. | |
14397 | ||
14398 | @menu | |
14399 | * lengths-list-many-files:: Return a list of the lengths of defuns. | |
14400 | * append:: Attach one list to another. | |
14401 | @end menu | |
14402 | ||
14403 | @node lengths-list-many-files, append, Several files, Several files | |
14404 | @ifnottex | |
14405 | @unnumberedsubsec Determine the lengths of @code{defuns} | |
14406 | @end ifnottex | |
14407 | ||
14408 | The design using a @code{while} loop is routine. The argument passed | |
14409 | the function is a list of files. As we saw earlier (@pxref{Loop | |
14410 | Example}), you can write a @code{while} loop so that the body of the | |
14411 | loop is evaluated if such a list contains elements, but to exit the | |
14412 | loop if the list is empty. For this design to work, the body of the | |
14413 | loop must contain an expression that shortens the list each time the | |
14414 | body is evaluated, so that eventually the list is empty. The usual | |
14415 | technique is to set the value of the list to the value of the @sc{cdr} | |
14416 | of the list each time the body is evaluated. | |
14417 | ||
14418 | @need 800 | |
14419 | The template looks like this: | |
14420 | ||
14421 | @smallexample | |
14422 | @group | |
14423 | (while @var{test-whether-list-is-empty} | |
14424 | @var{body}@dots{} | |
14425 | @var{set-list-to-cdr-of-list}) | |
14426 | @end group | |
14427 | @end smallexample | |
14428 | ||
14429 | Also, we remember that a @code{while} loop returns @code{nil} (the | |
14430 | result of evaluating the true-or-false-test), not the result of any | |
14431 | evaluation within its body. (The evaluations within the body of the | |
14432 | loop are done for their side effects.) However, the expression that | |
14433 | sets the lengths' list is part of the body---and that is the value | |
14434 | that we want returned by the function as a whole. To do this, we | |
14435 | enclose the @code{while} loop within a @code{let} expression, and | |
14436 | arrange that the last element of the @code{let} expression contains | |
14437 | the value of the lengths' list. (@xref{Incrementing Example, , Loop | |
14438 | Example with an Incrementing Counter}.) | |
14439 | ||
14440 | @findex lengths-list-many-files | |
14441 | @need 1250 | |
14442 | These considerations lead us directly to the function itself: | |
14443 | ||
14444 | @smallexample | |
14445 | @group | |
14446 | ;;; @r{Use @code{while} loop.} | |
14447 | (defun lengths-list-many-files (list-of-files) | |
14448 | "Return list of lengths of defuns in LIST-OF-FILES." | |
14449 | @end group | |
14450 | @group | |
14451 | (let (lengths-list) | |
14452 | ||
14453 | ;;; @r{true-or-false-test} | |
14454 | (while list-of-files | |
14455 | (setq lengths-list | |
14456 | (append | |
14457 | lengths-list | |
14458 | ||
14459 | ;;; @r{Generate a lengths' list.} | |
14460 | (lengths-list-file | |
14461 | (expand-file-name (car list-of-files))))) | |
14462 | @end group | |
14463 | ||
14464 | @group | |
14465 | ;;; @r{Make files' list shorter.} | |
14466 | (setq list-of-files (cdr list-of-files))) | |
14467 | ||
14468 | ;;; @r{Return final value of lengths' list.} | |
14469 | lengths-list)) | |
14470 | @end group | |
14471 | @end smallexample | |
14472 | ||
14473 | @code{expand-file-name} is a built-in function that converts a file | |
14474 | name to the absolute, long, path name form of the directory in which | |
14475 | the function is called. | |
14476 | ||
14477 | @c !!! 21.0.100 lisp sources location here | |
14478 | @need 1500 | |
14479 | Thus, if @code{expand-file-name} is called on @code{debug.el} when | |
14480 | Emacs is visiting the | |
14481 | @file{/usr/local/share/emacs/21.0.100/lisp/emacs-lisp/} directory, | |
14482 | ||
14483 | @smallexample | |
14484 | debug.el | |
14485 | @end smallexample | |
14486 | ||
14487 | @need 800 | |
14488 | @noindent | |
14489 | becomes | |
14490 | ||
14491 | @c !!! 21.0.100 lisp sources location here | |
14492 | @smallexample | |
14493 | /usr/local/share/emacs/21.0.100/lisp/emacs-lisp/debug.el | |
14494 | @end smallexample | |
14495 | ||
14496 | The only other new element of this function definition is the as yet | |
14497 | unstudied function @code{append}, which merits a short section for | |
14498 | itself. | |
14499 | ||
14500 | @node append, , lengths-list-many-files, Several files | |
14501 | @subsection The @code{append} Function | |
14502 | ||
14503 | @need 800 | |
14504 | The @code{append} function attaches one list to another. Thus, | |
14505 | ||
14506 | @smallexample | |
14507 | (append '(1 2 3 4) '(5 6 7 8)) | |
14508 | @end smallexample | |
14509 | ||
14510 | @need 800 | |
14511 | @noindent | |
14512 | produces the list | |
14513 | ||
14514 | @smallexample | |
14515 | (1 2 3 4 5 6 7 8) | |
14516 | @end smallexample | |
14517 | ||
14518 | This is exactly how we want to attach two lengths' lists produced by | |
14519 | @code{lengths-list-file} to each other. The results contrast with | |
14520 | @code{cons}, | |
14521 | ||
14522 | @smallexample | |
14523 | (cons '(1 2 3 4) '(5 6 7 8)) | |
14524 | @end smallexample | |
14525 | ||
14526 | @need 1250 | |
14527 | @noindent | |
14528 | which constructs a new list in which the first argument to @code{cons} | |
14529 | becomes the first element of the new list: | |
14530 | ||
14531 | @smallexample | |
14532 | ((1 2 3 4) 5 6 7 8) | |
14533 | @end smallexample | |
14534 | ||
14535 | @node Several files recursively, Prepare the data, Several files, Words in a defun | |
14536 | @section Recursively Count Words in Different Files | |
14537 | ||
14538 | Besides a @code{while} loop, you can work on each of a list of files | |
14539 | with recursion. A recursive version of @code{lengths-list-many-files} | |
14540 | is short and simple. | |
14541 | ||
14542 | The recursive function has the usual parts: the `do-again-test', the | |
14543 | `next-step-expression', and the recursive call. The `do-again-test' | |
14544 | determines whether the function should call itself again, which it | |
14545 | will do if the @code{list-of-files} contains any remaining elements; | |
14546 | the `next-step-expression' resets the @code{list-of-files} to the | |
14547 | @sc{cdr} of itself, so eventually the list will be empty; and the | |
14548 | recursive call calls itself on the shorter list. The complete | |
14549 | function is shorter than this description! | |
14550 | @findex recursive-lengths-list-many-files | |
14551 | ||
14552 | @smallexample | |
14553 | @group | |
14554 | (defun recursive-lengths-list-many-files (list-of-files) | |
14555 | "Return list of lengths of each defun in LIST-OF-FILES." | |
14556 | (if list-of-files ; @r{do-again-test} | |
14557 | (append | |
14558 | (lengths-list-file | |
14559 | (expand-file-name (car list-of-files))) | |
14560 | (recursive-lengths-list-many-files | |
14561 | (cdr list-of-files))))) | |
14562 | @end group | |
14563 | @end smallexample | |
14564 | ||
14565 | @noindent | |
14566 | In a sentence, the function returns the lengths' list for the first of | |
14567 | the @code{list-of-files} appended to the result of calling itself on | |
14568 | the rest of the @code{list-of-files}. | |
14569 | ||
14570 | Here is a test of @code{recursive-lengths-list-many-files}, along with | |
14571 | the results of running @code{lengths-list-file} on each of the files | |
14572 | individually. | |
14573 | ||
14574 | Install @code{recursive-lengths-list-many-files} and | |
14575 | @code{lengths-list-file}, if necessary, and then evaluate the | |
14576 | following expressions. You may need to change the files' pathnames; | |
14577 | those here work when this Info file and the Emacs sources are located | |
14578 | in their customary places. To change the expressions, copy them to | |
14579 | the @file{*scratch*} buffer, edit them, and then evaluate them. | |
14580 | ||
14581 | The results are shown after the @samp{@result{}}. (These results are | |
14582 | for files from Emacs Version 21.0.100; files from other versions of | |
14583 | Emacs may produce different results.) | |
14584 | ||
14585 | @c !!! 21.0.100 lisp sources location here | |
14586 | @smallexample | |
14587 | @group | |
14588 | (cd "/usr/local/share/emacs/21.0.100/") | |
14589 | ||
14590 | (lengths-list-file "./lisp/macros.el") | |
14591 | @result{} (273 263 456 90) | |
14592 | @end group | |
14593 | ||
14594 | @group | |
14595 | (lengths-list-file "./lisp/mail/mailalias.el") | |
14596 | @result{} (38 32 26 77 174 180 321 198 324) | |
14597 | @end group | |
14598 | ||
14599 | @group | |
14600 | (lengths-list-file "./lisp/makesum.el") | |
14601 | @result{} (85 181) | |
14602 | @end group | |
14603 | ||
14604 | @group | |
14605 | (recursive-lengths-list-many-files | |
14606 | '("./lisp/macros.el" | |
14607 | "./lisp/mail/mailalias.el" | |
14608 | "./lisp/makesum.el")) | |
14609 | @result{} (273 263 456 90 38 32 26 77 174 180 321 198 324 85 181) | |
14610 | @end group | |
14611 | @end smallexample | |
14612 | ||
14613 | The @code{recursive-lengths-list-many-files} function produces the | |
14614 | output we want. | |
14615 | ||
14616 | The next step is to prepare the data in the list for display in a graph. | |
14617 | ||
14618 | @node Prepare the data, , Several files recursively, Words in a defun | |
14619 | @section Prepare the Data for Display in a Graph | |
14620 | ||
14621 | The @code{recursive-lengths-list-many-files} function returns a list | |
14622 | of numbers. Each number records the length of a function definition. | |
14623 | What we need to do now is transform this data into a list of numbers | |
14624 | suitable for generating a graph. The new list will tell how many | |
14625 | functions definitions contain less than 10 words and | |
14626 | symbols, how many contain between 10 and 19 words and symbols, how | |
14627 | many contain between 20 and 29 words and symbols, and so on. | |
14628 | ||
14629 | In brief, we need to go through the lengths' list produced by the | |
14630 | @code{recursive-lengths-list-many-files} function and count the number | |
14631 | of defuns within each range of lengths, and produce a list of those | |
14632 | numbers. | |
14633 | ||
14634 | Based on what we have done before, we can readily foresee that it | |
14635 | should not be too hard to write a function that `@sc{cdr}s' down the | |
14636 | lengths' list, looks at each element, determines which length range it | |
14637 | is in, and increments a counter for that range. | |
14638 | ||
14639 | However, before beginning to write such a function, we should consider | |
14640 | the advantages of sorting the lengths' list first, so the numbers are | |
14641 | ordered from smallest to largest. First, sorting will make it easier | |
14642 | to count the numbers in each range, since two adjacent numbers will | |
14643 | either be in the same length range or in adjacent ranges. Second, by | |
14644 | inspecting a sorted list, we can discover the highest and lowest | |
14645 | number, and thereby determine the largest and smallest length range | |
14646 | that we will need. | |
14647 | ||
14648 | @menu | |
14649 | * Sorting:: Sorting lists. | |
14650 | * Files List:: Making a list of files. | |
14651 | * Counting function definitions:: | |
14652 | @end menu | |
14653 | ||
14654 | @node Sorting, Files List, Prepare the data, Prepare the data | |
14655 | @subsection Sorting Lists | |
14656 | @findex sort | |
14657 | ||
14658 | Emacs contains a function to sort lists, called (as you might guess) | |
14659 | @code{sort}. The @code{sort} function takes two arguments, the list | |
14660 | to be sorted, and a predicate that determines whether the first of | |
14661 | two list elements is ``less'' than the second. | |
14662 | ||
14663 | As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong | |
14664 | Type Object as an Argument}), a predicate is a function that | |
14665 | determines whether some property is true or false. The @code{sort} | |
14666 | function will reorder a list according to whatever property the | |
14667 | predicate uses; this means that @code{sort} can be used to sort | |
14668 | non-numeric lists by non-numeric criteria---it can, for example, | |
14669 | alphabetize a list. | |
14670 | ||
14671 | @need 1250 | |
14672 | The @code{<} function is used when sorting a numeric list. For example, | |
14673 | ||
14674 | @smallexample | |
14675 | (sort '(4 8 21 17 33 7 21 7) '<) | |
14676 | @end smallexample | |
14677 | ||
14678 | @need 800 | |
14679 | @noindent | |
14680 | produces this: | |
14681 | ||
14682 | @smallexample | |
14683 | (4 7 7 8 17 21 21 33) | |
14684 | @end smallexample | |
14685 | ||
14686 | @noindent | |
14687 | (Note that in this example, both the arguments are quoted so that the | |
14688 | symbols are not evaluated before being passed to @code{sort} as | |
14689 | arguments.) | |
14690 | ||
14691 | Sorting the list returned by the | |
14692 | @code{recursive-lengths-list-many-files} function is straightforward; | |
14693 | it uses the @code{<} function: | |
14694 | ||
14695 | @smallexample | |
14696 | @group | |
14697 | (sort | |
14698 | (recursive-lengths-list-many-files | |
14699 | '("../lisp/macros.el" | |
14700 | "../lisp/mailalias.el" | |
14701 | "../lisp/makesum.el")) | |
14702 | '< | |
14703 | @end group | |
14704 | @end smallexample | |
14705 | ||
14706 | @need 800 | |
14707 | @noindent | |
14708 | which produces: | |
14709 | ||
14710 | @smallexample | |
14711 | (85 86 116 122 154 176 179 265) | |
14712 | @end smallexample | |
14713 | ||
14714 | @noindent | |
14715 | (Note that in this example, the first argument to @code{sort} is not | |
14716 | quoted, since the expression must be evaluated so as to produce the | |
14717 | list that is passed to @code{sort}.) | |
14718 | ||
14719 | @node Files List, Counting function definitions, Sorting, Prepare the data | |
14720 | @subsection Making a List of Files | |
14721 | ||
14722 | The @code{recursive-lengths-list-many-files} function requires a list | |
14723 | of files as its argument. For our test examples, we constructed such | |
14724 | a list by hand; but the Emacs Lisp source directory is too large for | |
14725 | us to do for that. Instead, we will write a function to do the job | |
14726 | for us. In this function, we will use both a @code{while} loop and a | |
14727 | recursive call. | |
14728 | ||
14729 | @findex directory-files | |
14730 | We did not have to write a function like this for older versions of | |
14731 | GNU Emacs, since they placed all the @samp{.el} files in one | |
14732 | directory. Instead, we were able to use the @code{directory-files} | |
14733 | function, which lists the names of files that match a specified | |
14734 | pattern within a single directory. | |
14735 | ||
14736 | However, recent versions of Emacs place Emacs Lisp files in | |
14737 | sub-directories of the top level @file{lisp} directory. This | |
14738 | re-arrangement eases navigation. For example, all the mail related | |
14739 | files are in a @file{lisp} sub-directory called @file{mail}. But at | |
14740 | the same time, this arrangement forces us to create a file listing | |
14741 | function that descends into the sub-directories. | |
14742 | ||
14743 | @findex files-in-below-directory | |
14744 | We can create this function, called @code{files-in-below-directory}, | |
14745 | using familiar functions such as @code{car}, @code{nthcdr}, and | |
14746 | @code{substring} in conjunction with an existing function called | |
14747 | @code{directory-files-and-attributes}. This latter function not only | |
14748 | lists all the filenames in a directory, including the names | |
14749 | of sub-directories, but also their attributes. | |
14750 | ||
14751 | To restate our goal: to create a function that will enable us | |
14752 | to feed filenames to @code{recursive-lengths-list-many-files} | |
14753 | as a list that looks like this (but with more elements): | |
14754 | ||
14755 | @smallexample | |
14756 | @group | |
14757 | ("../lisp/macros.el" | |
14758 | "../lisp/mail/rmail.el" | |
14759 | "../lisp/makesum.el") | |
14760 | @end group | |
14761 | @end smallexample | |
14762 | ||
14763 | The @code{directory-files-and-attributes} function returns a list of | |
14764 | lists. Each of the lists within the main list consists of 13 | |
14765 | elements. The first element is a string that contains the name of the | |
14766 | file -- which, in GNU/Linux, may be a `directory file', that is to | |
14767 | say, a file with the special attributes of a directory. The second | |
14768 | element of the list is @code{t} for a directory, a string | |
14769 | for symbolic link (the string is the name linked to), or @code{nil}. | |
14770 | ||
14771 | For example, the first @samp{.el} file in the @file{lisp/} directory | |
14772 | is @file{abbrev.el}. Its name is | |
14773 | @file{/usr/local/share/emacs/21.0.100/lisp/abbrev.el} and it is not a | |
14774 | directory or a symbolic link. | |
14775 | ||
14776 | @need 1000 | |
14777 | This is how @code{directory-files-and-attributes} lists that file and | |
14778 | its attributes: | |
14779 | ||
14780 | @smallexample | |
14781 | @group | |
14782 | ("/usr/local/share/emacs/21.0.100/lisp/abbrev.el" | |
14783 | nil | |
14784 | 1 | |
14785 | 1000 | |
14786 | 100 | |
14787 | @end group | |
14788 | @group | |
14789 | (15019 32380) | |
14790 | (14883 48041) | |
14791 | (15214 49336) | |
14792 | 11583 | |
14793 | "-rw-rw-r--" | |
14794 | @end group | |
14795 | @group | |
14796 | t | |
14797 | 341385 | |
14798 | 776) | |
14799 | @end group | |
14800 | @end smallexample | |
14801 | ||
14802 | @need 1200 | |
14803 | On the other hand, @file{mail/} is a directory within the @file{lisp/} | |
14804 | directory. The beginning of its listing looks like this: | |
14805 | ||
14806 | @smallexample | |
14807 | @group | |
14808 | ("/usr/local/share/emacs/21.0.100/lisp/mail" | |
14809 | t | |
14810 | @dots{} | |
14811 | ) | |
14812 | @end group | |
14813 | @end smallexample | |
14814 | ||
14815 | (Look at the documentation of @code{file-attributes} to learn about | |
14816 | the different attributes. Bear in mind that the | |
14817 | @code{file-attributes} function does not list the filename, so its | |
14818 | first element is @code{directory-files-and-attributes}'s second | |
14819 | element.) | |
14820 | ||
14821 | We will want our new function, @code{files-in-below-directory}, to | |
14822 | list the @samp{.el} files in the directory it is told to check, and in | |
14823 | any directories below that directory. | |
14824 | ||
14825 | This gives us a hint on how to construct | |
14826 | @code{files-in-below-directory}: within a directory, the function | |
14827 | should add @samp{.el} filenames to a list; and if, within a directory, | |
14828 | the function comes upon a sub-directory, it should go into that | |
14829 | sub-directory and repeat its actions. | |
14830 | ||
14831 | However, we should note that every directory contains a name that | |
14832 | refers to itself, called @file{.}, (``dot'') and a name that refers to | |
14833 | its parent directory, called @file{..} (``double dot''). (In | |
14834 | @file{/}, the root directory, @file{..} refers to itself, since | |
14835 | @file{/} has no parent.) Clearly, we do not want our | |
14836 | @code{files-in-below-directory} function to enter those directories, | |
14837 | since they always lead us, directly or indirectly, to the current | |
14838 | directory. | |
14839 | ||
14840 | Consequently, our @code{files-in-below-directory} function must do | |
14841 | several tasks: | |
14842 | ||
14843 | @itemize @bullet | |
14844 | @item | |
14845 | Check to see whether it is looking at a filename that ends in | |
14846 | @samp{.el}; and if so, add its name to a list. | |
14847 | ||
14848 | @item | |
14849 | Check to see whether it is looking at a filename that is the name of a | |
14850 | directory; and if so, | |
14851 | ||
14852 | @itemize @minus | |
14853 | @item | |
14854 | Check to see whether it is looking at @file{.} or @file{..}; and if | |
14855 | so skip it. | |
14856 | ||
14857 | @item | |
14858 | Or else, go into that directory and repeat the process. | |
14859 | @end itemize | |
14860 | @end itemize | |
14861 | ||
14862 | Let's write a function definition to do these tasks. We will use a | |
14863 | @code{while} loop to move from one filename to another within a | |
14864 | directory, checking what needs to be done; and we will use a recursive | |
14865 | call to repeat the actions on each sub-directory. The recursive | |
14866 | pattern is `accumulate' | |
14867 | (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}), | |
14868 | using @code{append} as the combiner. | |
14869 | ||
14870 | @ignore | |
14871 | (directory-files "/usr/local/share/emacs/21.0.100/lisp/" t "\\.el$") | |
14872 | (shell-command "find /usr/local/share/emacs/21.0.100/lisp/ -name '*.el'") | |
14873 | @end ignore | |
14874 | ||
14875 | @c /usr/local/share/emacs/21.0.100/lisp/ | |
14876 | ||
14877 | @need 800 | |
14878 | Here is the function: | |
14879 | ||
14880 | @smallexample | |
14881 | @group | |
14882 | (defun files-in-below-directory (directory) | |
14883 | "List the .el files in DIRECTORY and in its sub-directories." | |
14884 | ;; Although the function will be used non-interactively, | |
14885 | ;; it will be easier to test if we make it interactive. | |
14886 | ;; The directory will have a name such as | |
14887 | ;; "/usr/local/share/emacs/21.0.100/lisp/" | |
14888 | (interactive "DDirectory name: ") | |
14889 | @end group | |
14890 | @group | |
14891 | (let (el-files-list | |
14892 | (current-directory-list | |
14893 | (directory-files-and-attributes directory t))) | |
14894 | ;; while we are in the current directory | |
14895 | (while current-directory-list | |
14896 | @end group | |
14897 | @group | |
14898 | (cond | |
14899 | ;; check to see whether filename ends in `.el' | |
14900 | ;; and if so, append its name to a list. | |
14901 | ((equal ".el" (substring (car (car current-directory-list)) -3)) | |
14902 | (setq el-files-list | |
14903 | (cons (car (car current-directory-list)) el-files-list))) | |
14904 | @end group | |
14905 | @group | |
14906 | ;; check whether filename is that of a directory | |
14907 | ((eq t (car (cdr (car current-directory-list)))) | |
14908 | ;; decide whether to skip or recurse | |
14909 | (if | |
14910 | (equal (or "." "..") | |
14911 | (substring (car (car current-directory-list)) -1)) | |
14912 | ;; then do nothing if filename is that of | |
14913 | ;; current directory or parent | |
14914 | () | |
14915 | @end group | |
14916 | @group | |
14917 | ;; else descend into the directory and repeat the process | |
14918 | (setq el-files-list | |
14919 | (append | |
14920 | (files-in-below-directory | |
14921 | (car (car current-directory-list))) | |
14922 | el-files-list))))) | |
14923 | ;; move to the next filename in the list; this also | |
14924 | ;; shortens the list so the while loop eventually comes to an end | |
14925 | (setq current-directory-list (cdr current-directory-list))) | |
14926 | ;; return the filenames | |
14927 | el-files-list)) | |
14928 | @end group | |
14929 | @end smallexample | |
14930 | ||
14931 | @c (files-in-below-directory "/usr/local/share/emacs/21.0.100/lisp/") | |
14932 | ||
14933 | The @code{files-in-below-directory} @code{directory-files} function | |
14934 | takes one argument, the name of a directory. | |
14935 | ||
14936 | @need 1250 | |
14937 | Thus, on my system, | |
14938 | ||
14939 | @c !!! 21.0.100 lisp sources location here | |
14940 | @smallexample | |
14941 | @group | |
14942 | (length | |
14943 | (files-in-below-directory "/usr/local/share/emacs/21.0.100/lisp/")) | |
14944 | @end group | |
14945 | @end smallexample | |
14946 | ||
14947 | @noindent | |
14948 | tells me that my version 21.0.100 Lisp sources directory contains 754 | |
14949 | @samp{.el} files. | |
14950 | ||
14951 | @code{files-in-below-directory} returns a list in reverse alphabetical | |
14952 | order. An expression to sort the list in alphabetical order looks | |
14953 | like this: | |
14954 | ||
14955 | @smallexample | |
14956 | @group | |
14957 | (sort | |
14958 | (files-in-below-directory "/usr/local/share/emacs/21.0.100/lisp/") | |
14959 | 'string-lessp) | |
14960 | @end group | |
14961 | @end smallexample | |
14962 | ||
14963 | @ignore | |
14964 | (defun test () | |
14965 | "Test how long it takes to find lengths of all elisp defuns." | |
14966 | (insert "\n" (current-time-string) "\n") | |
14967 | (sit-for 0) | |
14968 | (sort | |
14969 | (recursive-lengths-list-many-files | |
14970 | '("../lisp/macros.el" | |
14971 | "../lisp/mailalias.el" | |
14972 | "../lisp/makesum.el")) | |
14973 | '<) | |
14974 | (insert (format "%s" (current-time-string)))) | |
14975 | ||
14976 | @end ignore | |
14977 | ||
14978 | @node Counting function definitions, , Files List, Prepare the data | |
14979 | @subsection Counting function definitions | |
14980 | ||
14981 | Our immediate goal is to generate a list that tells us how many | |
14982 | function definitions contain fewer than 10 words and symbols, how many | |
14983 | contain between 10 and 19 words and symbols, how many contain between | |
14984 | 20 and 29 words and symbols, and so on. | |
14985 | ||
14986 | With a sorted list of numbers, this is easy: count how many elements | |
14987 | of the list are smaller than 10, then, after moving past the numbers | |
14988 | just counted, count how many are smaller than 20, then, after moving | |
14989 | past the numbers just counted, count how many are smaller than 30, and | |
14990 | so on. Each of the numbers, 10, 20, 30, 40, and the like, is one | |
14991 | larger than the top of that range. We can call the list of such | |
14992 | numbers the @code{top-of-ranges} list. | |
14993 | ||
14994 | @need 1200 | |
14995 | If we wished, we could generate this list automatically, but it is | |
14996 | simpler to write a list manually. Here it is: | |
14997 | @vindex top-of-ranges | |
14998 | ||
14999 | @smallexample | |
15000 | @group | |
15001 | (defvar top-of-ranges | |
15002 | '(10 20 30 40 50 | |
15003 | 60 70 80 90 100 | |
15004 | 110 120 130 140 150 | |
15005 | 160 170 180 190 200 | |
15006 | 210 220 230 240 250 | |
15007 | 260 270 280 290 300) | |
15008 | "List specifying ranges for `defuns-per-range'.") | |
15009 | @end group | |
15010 | @end smallexample | |
15011 | ||
15012 | To change the ranges, we edit this list. | |
15013 | ||
15014 | Next, we need to write the function that creates the list of the | |
15015 | number of definitions within each range. Clearly, this function must | |
15016 | take the @code{sorted-lengths} and the @code{top-of-ranges} lists | |
15017 | as arguments. | |
15018 | ||
15019 | The @code{defuns-per-range} function must do two things again and | |
15020 | again: it must count the number of definitions within a range | |
15021 | specified by the current top-of-range value; and it must shift to the | |
15022 | next higher value in the @code{top-of-ranges} list after counting the | |
15023 | number of definitions in the current range. Since each of these | |
15024 | actions is repetitive, we can use @code{while} loops for the job. | |
15025 | One loop counts the number of definitions in the range defined by the | |
15026 | current top-of-range value, and the other loop selects each of the | |
15027 | top-of-range values in turn. | |
15028 | ||
15029 | Several entries of the @code{sorted-lengths} list are counted for each | |
15030 | range; this means that the loop for the @code{sorted-lengths} list | |
15031 | will be inside the loop for the @code{top-of-ranges} list, like a | |
15032 | small gear inside a big gear. | |
15033 | ||
15034 | The inner loop counts the number of definitions within the range. It | |
15035 | is a simple counting loop of the type we have seen before. | |
15036 | (@xref{Incrementing Loop, , A loop with an incrementing counter}.) | |
15037 | The true-or-false test of the loop tests whether the value from the | |
15038 | @code{sorted-lengths} list is smaller than the current value of the | |
15039 | top of the range. If it is, the function increments the counter and | |
15040 | tests the next value from the @code{sorted-lengths} list. | |
15041 | ||
15042 | @need 1250 | |
15043 | The inner loop looks like this: | |
15044 | ||
15045 | @smallexample | |
15046 | @group | |
15047 | (while @var{length-element-smaller-than-top-of-range} | |
15048 | (setq number-within-range (1+ number-within-range)) | |
15049 | (setq sorted-lengths (cdr sorted-lengths))) | |
15050 | @end group | |
15051 | @end smallexample | |
15052 | ||
15053 | The outer loop must start with the lowest value of the | |
15054 | @code{top-of-ranges} list, and then be set to each of the succeeding | |
15055 | higher values in turn. This can be done with a loop like this: | |
15056 | ||
15057 | @smallexample | |
15058 | @group | |
15059 | (while top-of-ranges | |
15060 | @var{body-of-loop}@dots{} | |
15061 | (setq top-of-ranges (cdr top-of-ranges))) | |
15062 | @end group | |
15063 | @end smallexample | |
15064 | ||
15065 | @need 1200 | |
15066 | Put together, the two loops look like this: | |
15067 | ||
15068 | @smallexample | |
15069 | @group | |
15070 | (while top-of-ranges | |
15071 | ||
15072 | ;; @r{Count the number of elements within the current range.} | |
15073 | (while @var{length-element-smaller-than-top-of-range} | |
15074 | (setq number-within-range (1+ number-within-range)) | |
15075 | (setq sorted-lengths (cdr sorted-lengths))) | |
15076 | ||
15077 | ;; @r{Move to next range.} | |
15078 | (setq top-of-ranges (cdr top-of-ranges))) | |
15079 | @end group | |
15080 | @end smallexample | |
15081 | ||
15082 | In addition, in each circuit of the outer loop, Emacs should record | |
15083 | the number of definitions within that range (the value of | |
15084 | @code{number-within-range}) in a list. We can use @code{cons} for | |
15085 | this purpose. (@xref{cons, , @code{cons}}.) | |
15086 | ||
15087 | The @code{cons} function works fine, except that the list it | |
15088 | constructs will contain the number of definitions for the highest | |
15089 | range at its beginning and the number of definitions for the lowest | |
15090 | range at its end. This is because @code{cons} attaches new elements | |
15091 | of the list to the beginning of the list, and since the two loops are | |
15092 | working their way through the lengths' list from the lower end first, | |
15093 | the @code{defuns-per-range-list} will end up largest number first. | |
15094 | But we will want to print our graph with smallest values first and the | |
15095 | larger later. The solution is to reverse the order of the | |
15096 | @code{defuns-per-range-list}. We can do this using the | |
15097 | @code{nreverse} function, which reverses the order of a list. | |
15098 | @findex nreverse | |
15099 | ||
15100 | @need 800 | |
15101 | For example, | |
15102 | ||
15103 | @smallexample | |
15104 | (nreverse '(1 2 3 4)) | |
15105 | @end smallexample | |
15106 | ||
15107 | @need 800 | |
15108 | @noindent | |
15109 | produces: | |
15110 | ||
15111 | @smallexample | |
15112 | (4 3 2 1) | |
15113 | @end smallexample | |
15114 | ||
15115 | Note that the @code{nreverse} function is ``destructive''---that is, | |
15116 | it changes the list to which it is applied; this contrasts with the | |
15117 | @code{car} and @code{cdr} functions, which are non-destructive. In | |
15118 | this case, we do not want the original @code{defuns-per-range-list}, | |
15119 | so it does not matter that it is destroyed. (The @code{reverse} | |
15120 | function provides a reversed copy of a list, leaving the original list | |
15121 | as is.) | |
15122 | @findex reverse | |
15123 | ||
15124 | @need 1250 | |
15125 | Put all together, the @code{defuns-per-range} looks like this: | |
15126 | ||
15127 | @smallexample | |
15128 | @group | |
15129 | (defun defuns-per-range (sorted-lengths top-of-ranges) | |
15130 | "SORTED-LENGTHS defuns in each TOP-OF-RANGES range." | |
15131 | (let ((top-of-range (car top-of-ranges)) | |
15132 | (number-within-range 0) | |
15133 | defuns-per-range-list) | |
15134 | @end group | |
15135 | ||
15136 | @group | |
15137 | ;; @r{Outer loop.} | |
15138 | (while top-of-ranges | |
15139 | @end group | |
15140 | ||
15141 | @group | |
15142 | ;; @r{Inner loop.} | |
15143 | (while (and | |
15144 | ;; @r{Need number for numeric test.} | |
15145 | (car sorted-lengths) | |
15146 | (< (car sorted-lengths) top-of-range)) | |
15147 | @end group | |
15148 | ||
15149 | @group | |
15150 | ;; @r{Count number of definitions within current range.} | |
15151 | (setq number-within-range (1+ number-within-range)) | |
15152 | (setq sorted-lengths (cdr sorted-lengths))) | |
15153 | ||
15154 | ;; @r{Exit inner loop but remain within outer loop.} | |
15155 | @end group | |
15156 | ||
15157 | @group | |
15158 | (setq defuns-per-range-list | |
15159 | (cons number-within-range defuns-per-range-list)) | |
15160 | (setq number-within-range 0) ; @r{Reset count to zero.} | |
15161 | @end group | |
15162 | ||
15163 | @group | |
15164 | ;; @r{Move to next range.} | |
15165 | (setq top-of-ranges (cdr top-of-ranges)) | |
15166 | ;; @r{Specify next top of range value.} | |
15167 | (setq top-of-range (car top-of-ranges))) | |
15168 | @end group | |
15169 | ||
15170 | @group | |
15171 | ;; @r{Exit outer loop and count the number of defuns larger than} | |
15172 | ;; @r{ the largest top-of-range value.} | |
15173 | (setq defuns-per-range-list | |
15174 | (cons | |
15175 | (length sorted-lengths) | |
15176 | defuns-per-range-list)) | |
15177 | @end group | |
15178 | ||
15179 | @group | |
15180 | ;; @r{Return a list of the number of definitions within each range,} | |
15181 | ;; @r{ smallest to largest.} | |
15182 | (nreverse defuns-per-range-list))) | |
15183 | @end group | |
15184 | @end smallexample | |
15185 | ||
15186 | @need 1200 | |
15187 | @noindent | |
15188 | The function is straightforward except for one subtle feature. The | |
15189 | true-or-false test of the inner loop looks like this: | |
15190 | ||
15191 | @smallexample | |
15192 | @group | |
15193 | (and (car sorted-lengths) | |
15194 | (< (car sorted-lengths) top-of-range)) | |
15195 | @end group | |
15196 | @end smallexample | |
15197 | ||
15198 | @need 800 | |
15199 | @noindent | |
15200 | instead of like this: | |
15201 | ||
15202 | @smallexample | |
15203 | (< (car sorted-lengths) top-of-range) | |
15204 | @end smallexample | |
15205 | ||
15206 | The purpose of the test is to determine whether the first item in the | |
15207 | @code{sorted-lengths} list is less than the value of the top of the | |
15208 | range. | |
15209 | ||
15210 | The simple version of the test works fine unless the | |
15211 | @code{sorted-lengths} list has a @code{nil} value. In that case, the | |
15212 | @code{(car sorted-lengths)} expression function returns | |
15213 | @code{nil}. The @code{<} function cannot compare a number to | |
15214 | @code{nil}, which is an empty list, so Emacs signals an error and | |
15215 | stops the function from attempting to continue to execute. | |
15216 | ||
15217 | The @code{sorted-lengths} list always becomes @code{nil} when the | |
15218 | counter reaches the end of the list. This means that any attempt to | |
15219 | use the @code{defuns-per-range} function with the simple version of | |
15220 | the test will fail. | |
15221 | ||
15222 | We solve the problem by using the @code{(car sorted-lengths)} | |
15223 | expression in conjunction with the @code{and} expression. The | |
15224 | @code{(car sorted-lengths)} expression returns a non-@code{nil} | |
15225 | value so long as the list has at least one number within it, but | |
15226 | returns @code{nil} if the list is empty. The @code{and} expression | |
15227 | first evaluates the @code{(car sorted-lengths)} expression, and | |
15228 | if it is @code{nil}, returns false @emph{without} evaluating the | |
15229 | @code{<} expression. But if the @code{(car sorted-lengths)} | |
15230 | expression returns a non-@code{nil} value, the @code{and} expression | |
15231 | evaluates the @code{<} expression, and returns that value as the value | |
15232 | of the @code{and} expression. | |
15233 | ||
15234 | @c colon in printed section title causes problem in Info cross reference | |
15235 | This way, we avoid an error. | |
15236 | @iftex | |
15237 | @xref{forward-paragraph, , @code{forward-paragraph}: a Goldmine of | |
15238 | Functions}, for more information about @code{and}. | |
15239 | @end iftex | |
15240 | @ifinfo | |
15241 | @xref{forward-paragraph}, for more information about @code{and}. | |
15242 | @end ifinfo | |
15243 | ||
15244 | Here is a short test of the @code{defuns-per-range} function. First, | |
15245 | evaluate the expression that binds (a shortened) | |
15246 | @code{top-of-ranges} list to the list of values, then evaluate the | |
15247 | expression for binding the @code{sorted-lengths} list, and then | |
15248 | evaluate the @code{defuns-per-range} function. | |
15249 | ||
15250 | @smallexample | |
15251 | @group | |
15252 | ;; @r{(Shorter list than we will use later.)} | |
15253 | (setq top-of-ranges | |
15254 | '(110 120 130 140 150 | |
15255 | 160 170 180 190 200)) | |
15256 | ||
15257 | (setq sorted-lengths | |
15258 | '(85 86 110 116 122 129 154 176 179 200 265 300 300)) | |
15259 | ||
15260 | (defuns-per-range sorted-lengths top-of-ranges) | |
15261 | @end group | |
15262 | @end smallexample | |
15263 | ||
15264 | @need 800 | |
15265 | @noindent | |
15266 | The list returned looks like this: | |
15267 | ||
15268 | @smallexample | |
15269 | (2 2 2 0 0 1 0 2 0 0 4) | |
15270 | @end smallexample | |
15271 | ||
15272 | @noindent | |
15273 | Indeed, there are two elements of the @code{sorted-lengths} list | |
15274 | smaller than 110, two elements between 110 and 119, two elements | |
15275 | between 120 and 129, and so on. There are four elements with a value | |
15276 | of 200 or larger. | |
15277 | ||
15278 | @c The next step is to turn this numbers' list into a graph. | |
15279 | ||
15280 | @node Readying a Graph, Emacs Initialization, Words in a defun, Top | |
15281 | @chapter Readying a Graph | |
15282 | @cindex Readying a graph | |
15283 | @cindex Graph prototype | |
15284 | @cindex Prototype graph | |
15285 | @cindex Body of graph | |
15286 | ||
15287 | Our goal is to construct a graph showing the numbers of function | |
15288 | definitions of various lengths in the Emacs lisp sources. | |
15289 | ||
15290 | As a practical matter, if you were creating a graph, you would | |
15291 | probably use a program such as @code{gnuplot} to do the job. | |
15292 | (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case, | |
15293 | however, we create one from scratch, and in the process we will | |
15294 | re-acquaint ourselves with some of what we learned before and learn | |
15295 | more. | |
15296 | ||
15297 | In this chapter, we will first write a simple graph printing function. | |
15298 | This first definition will be a @dfn{prototype}, a rapidly written | |
15299 | function that enables us to reconnoiter this unknown graph-making | |
15300 | territory. We will discover dragons, or find that they are myth. | |
15301 | After scouting the terrain, we will feel more confident and enhance | |
15302 | the function to label the axes automatically. | |
15303 | ||
15304 | @menu | |
15305 | * Columns of a graph:: | |
15306 | * graph-body-print:: How to print the body of a graph. | |
15307 | * recursive-graph-body-print:: | |
15308 | * Printed Axes:: | |
15309 | * Line Graph Exercise:: | |
15310 | @end menu | |
15311 | ||
15312 | @node Columns of a graph, graph-body-print, Readying a Graph, Readying a Graph | |
15313 | @ifnottex | |
15314 | @unnumberedsec Printing the Columns of a Graph | |
15315 | @end ifnottex | |
15316 | ||
15317 | Since Emacs is designed to be flexible and work with all kinds of | |
15318 | terminals, including character-only terminals, the graph will need to | |
15319 | be made from one of the `typewriter' symbols. An asterisk will do; as | |
15320 | we enhance the graph-printing function, we can make the choice of | |
15321 | symbol a user option. | |
15322 | ||
15323 | We can call this function @code{graph-body-print}; it will take a | |
15324 | @code{numbers-list} as its only argument. At this stage, we will not | |
15325 | label the graph, but only print its body. | |
15326 | ||
15327 | The @code{graph-body-print} function inserts a vertical column of | |
15328 | asterisks for each element in the @code{numbers-list}. The height of | |
15329 | each line is determined by the value of that element of the | |
15330 | @code{numbers-list}. | |
15331 | ||
15332 | Inserting columns is a repetitive act; that means that this function can | |
15333 | be written either with a @code{while} loop or recursively. | |
15334 | ||
15335 | Our first challenge is to discover how to print a column of asterisks. | |
15336 | Usually, in Emacs, we print characters onto a screen horizontally, | |
15337 | line by line, by typing. We have two routes we can follow: write our | |
15338 | own column-insertion function or discover whether one exists in Emacs. | |
15339 | ||
15340 | To see whether there is one in Emacs, we can use the @kbd{M-x apropos} | |
15341 | command. This command is like the @kbd{C-h a} (command-apropos) | |
15342 | command, except that the latter finds only those functions that are | |
15343 | commands. The @kbd{M-x apropos} command lists all symbols that match | |
15344 | a regular expression, including functions that are not interactive. | |
15345 | @findex apropos | |
15346 | ||
15347 | What we want to look for is some command that prints or inserts | |
15348 | columns. Very likely, the name of the function will contain either | |
15349 | the word `print' or the word `insert' or the word `column'. | |
15350 | Therefore, we can simply type @kbd{M-x apropos RET | |
15351 | print\|insert\|column RET} and look at the result. On my system, this | |
15352 | command takes quite some time, and then produces a list of 79 | |
15353 | functions and variables. Scanning down the list, the only function | |
15354 | that looks as if it might do the job is @code{insert-rectangle}. | |
15355 | ||
15356 | @need 1200 | |
15357 | Indeed, this is the function we want; its documentation says: | |
15358 | ||
15359 | @smallexample | |
15360 | @group | |
15361 | insert-rectangle: | |
15362 | Insert text of RECTANGLE with upper left corner at point. | |
15363 | RECTANGLE's first line is inserted at point, | |
15364 | its second line is inserted at a point vertically under point, etc. | |
15365 | RECTANGLE should be a list of strings. | |
15366 | @end group | |
15367 | @end smallexample | |
15368 | ||
15369 | We can run a quick test, to make sure it does what we expect of it. | |
15370 | ||
15371 | Here is the result of placing the cursor after the | |
15372 | @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e} | |
15373 | (@code{eval-last-sexp}). The function inserts the strings | |
15374 | @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below | |
15375 | point. Also the function returns @code{nil}. | |
15376 | ||
15377 | @smallexample | |
15378 | @group | |
15379 | (insert-rectangle '("first" "second" "third"))first | |
15380 | second | |
15381 | third | |
15382 | nil | |
15383 | @end group | |
15384 | @end smallexample | |
15385 | ||
15386 | @noindent | |
15387 | Of course, we won't be inserting the text of the | |
15388 | @code{insert-rectangle} expression itself into the buffer in which we | |
15389 | are making the graph, but will call the function from our program. We | |
15390 | shall, however, have to make sure that point is in the buffer at the | |
15391 | place where the @code{insert-rectangle} function will insert its | |
15392 | column of strings. | |
15393 | ||
15394 | If you are reading this in Info, you can see how this works by | |
15395 | switching to another buffer, such as the @file{*scratch*} buffer, | |
15396 | placing point somewhere in the buffer, typing @kbd{M-:}, | |
15397 | typing the @code{insert-rectangle} expression into the minibuffer at | |
15398 | the prompt, and then typing @key{RET}. This causes Emacs to evaluate | |
15399 | the expression in the minibuffer, but to use as the value of point the | |
15400 | position of point in the @file{*scratch*} buffer. (@kbd{M-:} | |
15401 | is the keybinding for @code{eval-expression}.) | |
15402 | ||
15403 | We find when we do this that point ends up at the end of the last | |
15404 | inserted line---that is to say, this function moves point as a | |
15405 | side-effect. If we were to repeat the command, with point at this | |
15406 | position, the next insertion would be below and to the right of the | |
15407 | previous insertion. We don't want this! If we are going to make a | |
15408 | bar graph, the columns need to be beside each other. | |
15409 | ||
15410 | So we discover that each cycle of the column-inserting @code{while} | |
15411 | loop must reposition point to the place we want it, and that place | |
15412 | will be at the top, not the bottom, of the column. Moreover, we | |
15413 | remember that when we print a graph, we do not expect all the columns | |
15414 | to be the same height. This means that the top of each column may be | |
15415 | at a different height from the previous one. We cannot simply | |
15416 | reposition point to the same line each time, but moved over to the | |
15417 | right---or perhaps we can@dots{} | |
15418 | ||
15419 | We are planning to make the columns of the bar graph out of asterisks. | |
15420 | The number of asterisks in the column is the number specified by the | |
15421 | current element of the @code{numbers-list}. We need to construct a | |
15422 | list of asterisks of the right length for each call to | |
15423 | @code{insert-rectangle}. If this list consists solely of the requisite | |
15424 | number of asterisks, then we will have position point the right number | |
15425 | of lines above the base for the graph to print correctly. This could | |
15426 | be difficult. | |
15427 | ||
15428 | Alternatively, if we can figure out some way to pass | |
15429 | @code{insert-rectangle} a list of the same length each time, then we | |
15430 | can place point on the same line each time, but move it over one | |
15431 | column to the right for each new column. If we do this, however, some | |
15432 | of the entries in the list passed to @code{insert-rectangle} must be | |
15433 | blanks rather than asterisks. For example, if the maximum height of | |
15434 | the graph is 5, but the height of the column is 3, then | |
15435 | @code{insert-rectangle} requires an argument that looks like this: | |
15436 | ||
15437 | @smallexample | |
15438 | (" " " " "*" "*" "*") | |
15439 | @end smallexample | |
15440 | ||
15441 | This last proposal is not so difficult, so long as we can determine | |
15442 | the column height. There are two ways for us to specify the column | |
15443 | height: we can arbitrarily state what it will be, which would work | |
15444 | fine for graphs of that height; or we can search through the list of | |
15445 | numbers and use the maximum height of the list as the maximum height | |
15446 | of the graph. If the latter operation were difficult, then the former | |
15447 | procedure would be easiest, but there is a function built into Emacs | |
15448 | that determines the maximum of its arguments. We can use that | |
15449 | function. The function is called @code{max} and it returns the | |
15450 | largest of all its arguments, which must be numbers. Thus, for | |
15451 | example, | |
15452 | ||
15453 | @smallexample | |
15454 | (max 3 4 6 5 7 3) | |
15455 | @end smallexample | |
15456 | ||
15457 | @noindent | |
15458 | returns 7. (A corresponding function called @code{min} returns the | |
15459 | smallest of all its arguments.) | |
15460 | @findex max | |
15461 | @findex min | |
15462 | ||
15463 | However, we cannot simply call @code{max} on the @code{numbers-list}; | |
15464 | the @code{max} function expects numbers as its argument, not a list of | |
15465 | numbers. Thus, the following expression, | |
15466 | ||
15467 | @smallexample | |
15468 | (max '(3 4 6 5 7 3)) | |
15469 | @end smallexample | |
15470 | ||
15471 | @need 800 | |
15472 | @noindent | |
15473 | produces the following error message; | |
15474 | ||
15475 | @smallexample | |
15476 | Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3) | |
15477 | @end smallexample | |
15478 | ||
15479 | @findex apply | |
15480 | We need a function that passes a list of arguments to a function. | |
15481 | This function is @code{apply}. This function `applies' its first | |
15482 | argument (a function) to its remaining arguments, the last of which | |
15483 | may be a list. | |
15484 | ||
15485 | @need 1250 | |
15486 | For example, | |
15487 | ||
15488 | @smallexample | |
15489 | (apply 'max 3 4 7 3 '(4 8 5)) | |
15490 | @end smallexample | |
15491 | ||
15492 | @noindent | |
15493 | returns 8. | |
15494 | ||
15495 | (Incidentally, I don't know how you would learn of this function | |
15496 | without a book such as this. It is possible to discover other | |
15497 | functions, like @code{search-forward} or @code{insert-rectangle}, by | |
15498 | guessing at a part of their names and then using @code{apropos}. Even | |
15499 | though its base in metaphor is clear---`apply' its first argument to | |
15500 | the rest---I doubt a novice would come up with that particular word | |
15501 | when using @code{apropos} or other aid. Of course, I could be wrong; | |
15502 | after all, the function was first named by someone who had to invent | |
15503 | it.) | |
15504 | ||
15505 | The second and subsequent arguments to @code{apply} are optional, so | |
15506 | we can use @code{apply} to call a function and pass the elements of a | |
15507 | list to it, like this, which also returns 8: | |
15508 | ||
15509 | @smallexample | |
15510 | (apply 'max '(4 8 5)) | |
15511 | @end smallexample | |
15512 | ||
15513 | This latter way is how we will use @code{apply}. The | |
15514 | @code{recursive-lengths-list-many-files} function returns a numbers' | |
15515 | list to which we can apply @code{max} (we could also apply @code{max} to | |
15516 | the sorted numbers' list; it does not matter whether the list is | |
15517 | sorted or not.) | |
15518 | ||
15519 | @need 800 | |
15520 | Hence, the operation for finding the maximum height of the graph is this: | |
15521 | ||
15522 | @smallexample | |
15523 | (setq max-graph-height (apply 'max numbers-list)) | |
15524 | @end smallexample | |
15525 | ||
15526 | Now we can return to the question of how to create a list of strings | |
15527 | for a column of the graph. Told the maximum height of the graph | |
15528 | and the number of asterisks that should appear in the column, the | |
15529 | function should return a list of strings for the | |
15530 | @code{insert-rectangle} command to insert. | |
15531 | ||
15532 | Each column is made up of asterisks or blanks. Since the function is | |
15533 | passed the value of the height of the column and the number of | |
15534 | asterisks in the column, the number of blanks can be found by | |
15535 | subtracting the number of asterisks from the height of the column. | |
15536 | Given the number of blanks and the number of asterisks, two | |
15537 | @code{while} loops can be used to construct the list: | |
15538 | ||
15539 | @smallexample | |
15540 | @group | |
15541 | ;;; @r{First version.} | |
15542 | (defun column-of-graph (max-graph-height actual-height) | |
15543 | "Return list of strings that is one column of a graph." | |
15544 | (let ((insert-list nil) | |
15545 | (number-of-top-blanks | |
15546 | (- max-graph-height actual-height))) | |
15547 | @end group | |
15548 | ||
15549 | @group | |
15550 | ;; @r{Fill in asterisks.} | |
15551 | (while (> actual-height 0) | |
15552 | (setq insert-list (cons "*" insert-list)) | |
15553 | (setq actual-height (1- actual-height))) | |
15554 | @end group | |
15555 | ||
15556 | @group | |
15557 | ;; @r{Fill in blanks.} | |
15558 | (while (> number-of-top-blanks 0) | |
15559 | (setq insert-list (cons " " insert-list)) | |
15560 | (setq number-of-top-blanks | |
15561 | (1- number-of-top-blanks))) | |
15562 | @end group | |
15563 | ||
15564 | @group | |
15565 | ;; @r{Return whole list.} | |
15566 | insert-list)) | |
15567 | @end group | |
15568 | @end smallexample | |
15569 | ||
15570 | If you install this function and then evaluate the following | |
15571 | expression you will see that it returns the list as desired: | |
15572 | ||
15573 | @smallexample | |
15574 | (column-of-graph 5 3) | |
15575 | @end smallexample | |
15576 | ||
15577 | @need 800 | |
15578 | @noindent | |
15579 | returns | |
15580 | ||
15581 | @smallexample | |
15582 | (" " " " "*" "*" "*") | |
15583 | @end smallexample | |
15584 | ||
15585 | As written, @code{column-of-graph} contains a major flaw: the symbols | |
15586 | used for the blank and for the marked entries in the column are | |
15587 | `hard-coded' as a space and asterisk. This is fine for a prototype, | |
15588 | but you, or another user, may wish to use other symbols. For example, | |
15589 | in testing the graph function, you many want to use a period in place | |
15590 | of the space, to make sure the point is being repositioned properly | |
15591 | each time the @code{insert-rectangle} function is called; or you might | |
15592 | want to substitute a @samp{+} sign or other symbol for the asterisk. | |
15593 | You might even want to make a graph-column that is more than one | |
15594 | display column wide. The program should be more flexible. The way to | |
15595 | do that is to replace the blank and the asterisk with two variables | |
15596 | that we can call @code{graph-blank} and @code{graph-symbol} and define | |
15597 | those variables separately. | |
15598 | ||
15599 | Also, the documentation is not well written. These considerations | |
15600 | lead us to the second version of the function: | |
15601 | ||
15602 | @smallexample | |
15603 | @group | |
15604 | (defvar graph-symbol "*" | |
15605 | "String used as symbol in graph, usually an asterisk.") | |
15606 | @end group | |
15607 | ||
15608 | @group | |
15609 | (defvar graph-blank " " | |
15610 | "String used as blank in graph, usually a blank space. | |
15611 | graph-blank must be the same number of columns wide | |
15612 | as graph-symbol.") | |
15613 | @end group | |
15614 | @end smallexample | |
15615 | ||
15616 | @noindent | |
15617 | (For an explanation of @code{defvar}, see | |
15618 | @ref{defvar, , Initializing a Variable with @code{defvar}}.) | |
15619 | ||
15620 | @smallexample | |
15621 | @group | |
15622 | ;;; @r{Second version.} | |
15623 | (defun column-of-graph (max-graph-height actual-height) | |
15624 | "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols. | |
15625 | ||
15626 | @end group | |
15627 | @group | |
15628 | The graph-symbols are contiguous entries at the end | |
15629 | of the list. | |
15630 | The list will be inserted as one column of a graph. | |
15631 | The strings are either graph-blank or graph-symbol." | |
15632 | @end group | |
15633 | ||
15634 | @group | |
15635 | (let ((insert-list nil) | |
15636 | (number-of-top-blanks | |
15637 | (- max-graph-height actual-height))) | |
15638 | @end group | |
15639 | ||
15640 | @group | |
15641 | ;; @r{Fill in @code{graph-symbols}.} | |
15642 | (while (> actual-height 0) | |
15643 | (setq insert-list (cons graph-symbol insert-list)) | |
15644 | (setq actual-height (1- actual-height))) | |
15645 | @end group | |
15646 | ||
15647 | @group | |
15648 | ;; @r{Fill in @code{graph-blanks}.} | |
15649 | (while (> number-of-top-blanks 0) | |
15650 | (setq insert-list (cons graph-blank insert-list)) | |
15651 | (setq number-of-top-blanks | |
15652 | (1- number-of-top-blanks))) | |
15653 | ||
15654 | ;; @r{Return whole list.} | |
15655 | insert-list)) | |
15656 | @end group | |
15657 | @end smallexample | |
15658 | ||
15659 | If we wished, we could rewrite @code{column-of-graph} a third time to | |
15660 | provide optionally for a line graph as well as for a bar graph. This | |
15661 | would not be hard to do. One way to think of a line graph is that it | |
15662 | is no more than a bar graph in which the part of each bar that is | |
15663 | below the top is blank. To construct a column for a line graph, the | |
15664 | function first constructs a list of blanks that is one shorter than | |
15665 | the value, then it uses @code{cons} to attach a graph symbol to the | |
15666 | list; then it uses @code{cons} again to attach the `top blanks' to | |
15667 | the list. | |
15668 | ||
15669 | It is easy to see how to write such a function, but since we don't | |
15670 | need it, we will not do it. But the job could be done, and if it were | |
15671 | done, it would be done with @code{column-of-graph}. Even more | |
15672 | important, it is worth noting that few changes would have to be made | |
15673 | anywhere else. The enhancement, if we ever wish to make it, is | |
15674 | simple. | |
15675 | ||
15676 | Now, finally, we come to our first actual graph printing function. | |
15677 | This prints the body of a graph, not the labels for the vertical and | |
15678 | horizontal axes, so we can call this @code{graph-body-print}. | |
15679 | ||
15680 | @node graph-body-print, recursive-graph-body-print, Columns of a graph, Readying a Graph | |
15681 | @section The @code{graph-body-print} Function | |
15682 | @findex graph-body-print | |
15683 | ||
15684 | After our preparation in the preceding section, the | |
15685 | @code{graph-body-print} function is straightforward. The function | |
15686 | will print column after column of asterisks and blanks, using the | |
15687 | elements of a numbers' list to specify the number of asterisks in each | |
15688 | column. This is a repetitive act, which means we can use a | |
15689 | decrementing @code{while} loop or recursive function for the job. In | |
15690 | this section, we will write the definition using a @code{while} loop. | |
15691 | ||
15692 | The @code{column-of-graph} function requires the height of the graph | |
15693 | as an argument, so we should determine and record that as a local variable. | |
15694 | ||
15695 | This leads us to the following template for the @code{while} loop | |
15696 | version of this function: | |
15697 | ||
15698 | @smallexample | |
15699 | @group | |
15700 | (defun graph-body-print (numbers-list) | |
15701 | "@var{documentation}@dots{}" | |
15702 | (let ((height @dots{} | |
15703 | @dots{})) | |
15704 | @end group | |
15705 | ||
15706 | @group | |
15707 | (while numbers-list | |
15708 | @var{insert-columns-and-reposition-point} | |
15709 | (setq numbers-list (cdr numbers-list))))) | |
15710 | @end group | |
15711 | @end smallexample | |
15712 | ||
15713 | @noindent | |
15714 | We need to fill in the slots of the template. | |
15715 | ||
15716 | Clearly, we can use the @code{(apply 'max numbers-list)} expression to | |
15717 | determine the height of the graph. | |
15718 | ||
15719 | The @code{while} loop will cycle through the @code{numbers-list} one | |
15720 | element at a time. As it is shortened by the @code{(setq numbers-list | |
15721 | (cdr numbers-list))} expression, the @sc{car} of each instance of the | |
15722 | list is the value of the argument for @code{column-of-graph}. | |
15723 | ||
15724 | At each cycle of the @code{while} loop, the @code{insert-rectangle} | |
15725 | function inserts the list returned by @code{column-of-graph}. Since | |
15726 | the @code{insert-rectangle} function moves point to the lower right of | |
15727 | the inserted rectangle, we need to save the location of point at the | |
15728 | time the rectangle is inserted, move back to that position after the | |
15729 | rectangle is inserted, and then move horizontally to the next place | |
15730 | from which @code{insert-rectangle} is called. | |
15731 | ||
15732 | If the inserted columns are one character wide, as they will be if | |
15733 | single blanks and asterisks are used, the repositioning command is | |
15734 | simply @code{(forward-char 1)}; however, the width of a column may be | |
15735 | greater than one. This means that the repositioning command should be | |
15736 | written @code{(forward-char symbol-width)}. The @code{symbol-width} | |
15737 | itself is the length of a @code{graph-blank} and can be found using | |
15738 | the expression @code{(length graph-blank)}. The best place to bind | |
15739 | the @code{symbol-width} variable to the value of the width of graph | |
15740 | column is in the varlist of the @code{let} expression. | |
15741 | ||
15742 | @need 1250 | |
15743 | These considerations lead to the following function definition: | |
15744 | ||
15745 | @smallexample | |
15746 | @group | |
15747 | (defun graph-body-print (numbers-list) | |
15748 | "Print a bar graph of the NUMBERS-LIST. | |
15749 | The numbers-list consists of the Y-axis values." | |
15750 | ||
15751 | (let ((height (apply 'max numbers-list)) | |
15752 | (symbol-width (length graph-blank)) | |
15753 | from-position) | |
15754 | @end group | |
15755 | ||
15756 | @group | |
15757 | (while numbers-list | |
15758 | (setq from-position (point)) | |
15759 | (insert-rectangle | |
15760 | (column-of-graph height (car numbers-list))) | |
15761 | (goto-char from-position) | |
15762 | (forward-char symbol-width) | |
15763 | @end group | |
15764 | @group | |
15765 | ;; @r{Draw graph column by column.} | |
15766 | (sit-for 0) | |
15767 | (setq numbers-list (cdr numbers-list))) | |
15768 | @end group | |
15769 | @group | |
15770 | ;; @r{Place point for X axis labels.} | |
15771 | (forward-line height) | |
15772 | (insert "\n") | |
15773 | )) | |
15774 | @end group | |
15775 | @end smallexample | |
15776 | ||
15777 | @noindent | |
15778 | The one unexpected expression in this function is the | |
15779 | @w{@code{(sit-for 0)}} expression in the @code{while} loop. This | |
15780 | expression makes the graph printing operation more interesting to | |
15781 | watch than it would be otherwise. The expression causes Emacs to | |
15782 | `sit' or do nothing for a zero length of time and then redraw the | |
15783 | screen. Placed here, it causes Emacs to redraw the screen column by | |
15784 | column. Without it, Emacs would not redraw the screen until the | |
15785 | function exits. | |
15786 | ||
15787 | We can test @code{graph-body-print} with a short list of numbers. | |
15788 | ||
15789 | @enumerate | |
15790 | @item | |
15791 | Install @code{graph-symbol}, @code{graph-blank}, | |
15792 | @code{column-of-graph}, which are in | |
475dc40a EZ |
15793 | @iftex |
15794 | @ref{Readying a Graph, , Readying a Graph}, | |
15795 | @end iftex | |
15796 | @ifinfo | |
15797 | @ref{Columns of a graph}, | |
15798 | @end ifinfo | |
8b096dce EZ |
15799 | and @code{graph-body-print}. |
15800 | ||
15801 | @need 800 | |
15802 | @item | |
15803 | Copy the following expression: | |
15804 | ||
15805 | @smallexample | |
15806 | (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3)) | |
15807 | @end smallexample | |
15808 | ||
15809 | @item | |
15810 | Switch to the @file{*scratch*} buffer and place the cursor where you | |
15811 | want the graph to start. | |
15812 | ||
15813 | @item | |
15814 | Type @kbd{M-:} (@code{eval-expression}). | |
15815 | ||
15816 | @item | |
15817 | Yank the @code{graph-body-print} expression into the minibuffer | |
15818 | with @kbd{C-y} (@code{yank)}. | |
15819 | ||
15820 | @item | |
15821 | Press @key{RET} to evaluate the @code{graph-body-print} expression. | |
15822 | @end enumerate | |
15823 | ||
15824 | @need 800 | |
15825 | Emacs will print a graph like this: | |
15826 | ||
15827 | @smallexample | |
15828 | @group | |
15829 | * | |
15830 | * ** | |
15831 | * **** | |
15832 | *** **** | |
15833 | ********* * | |
15834 | ************ | |
15835 | ************* | |
15836 | @end group | |
15837 | @end smallexample | |
15838 | ||
15839 | @node recursive-graph-body-print, Printed Axes, graph-body-print, Readying a Graph | |
15840 | @section The @code{recursive-graph-body-print} Function | |
15841 | @findex recursive-graph-body-print | |
15842 | ||
15843 | The @code{graph-body-print} function may also be written recursively. | |
15844 | The recursive solution is divided into two parts: an outside `wrapper' | |
15845 | that uses a @code{let} expression to determine the values of several | |
15846 | variables that need only be found once, such as the maximum height of | |
15847 | the graph, and an inside function that is called recursively to print | |
15848 | the graph. | |
15849 | ||
15850 | @need 1250 | |
15851 | The `wrapper' is uncomplicated: | |
15852 | ||
15853 | @smallexample | |
15854 | @group | |
15855 | (defun recursive-graph-body-print (numbers-list) | |
15856 | "Print a bar graph of the NUMBERS-LIST. | |
15857 | The numbers-list consists of the Y-axis values." | |
15858 | (let ((height (apply 'max numbers-list)) | |
15859 | (symbol-width (length graph-blank)) | |
15860 | from-position) | |
15861 | (recursive-graph-body-print-internal | |
15862 | numbers-list | |
15863 | height | |
15864 | symbol-width))) | |
15865 | @end group | |
15866 | @end smallexample | |
15867 | ||
15868 | The recursive function is a little more difficult. It has four parts: | |
15869 | the `do-again-test', the printing code, the recursive call, and the | |
15870 | `next-step-expression'. The `do-again-test' is an @code{if} | |
15871 | expression that determines whether the @code{numbers-list} contains | |
15872 | any remaining elements; if it does, the function prints one column of | |
15873 | the graph using the printing code and calls itself again. The | |
15874 | function calls itself again according to the value produced by the | |
15875 | `next-step-expression' which causes the call to act on a shorter | |
15876 | version of the @code{numbers-list}. | |
15877 | ||
15878 | @smallexample | |
15879 | @group | |
15880 | (defun recursive-graph-body-print-internal | |
15881 | (numbers-list height symbol-width) | |
15882 | "Print a bar graph. | |
15883 | Used within recursive-graph-body-print function." | |
15884 | @end group | |
15885 | ||
15886 | @group | |
15887 | (if numbers-list | |
15888 | (progn | |
15889 | (setq from-position (point)) | |
15890 | (insert-rectangle | |
15891 | (column-of-graph height (car numbers-list))) | |
15892 | @end group | |
15893 | @group | |
15894 | (goto-char from-position) | |
15895 | (forward-char symbol-width) | |
15896 | (sit-for 0) ; @r{Draw graph column by column.} | |
15897 | (recursive-graph-body-print-internal | |
15898 | (cdr numbers-list) height symbol-width)))) | |
15899 | @end group | |
15900 | @end smallexample | |
15901 | ||
15902 | @need 1250 | |
15903 | After installation, this expression can be tested; here is a sample: | |
15904 | ||
15905 | @smallexample | |
15906 | (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1)) | |
15907 | @end smallexample | |
15908 | ||
15909 | @need 800 | |
15910 | Here is what @code{recursive-graph-body-print} produces: | |
15911 | ||
15912 | @smallexample | |
15913 | @group | |
15914 | * | |
15915 | ** * | |
15916 | **** * | |
15917 | **** *** | |
15918 | * ********* | |
15919 | ************ | |
15920 | ************* | |
15921 | @end group | |
15922 | @end smallexample | |
15923 | ||
15924 | Either of these two functions, @code{graph-body-print} or | |
15925 | @code{recursive-graph-body-print}, create the body of a graph. | |
15926 | ||
15927 | @node Printed Axes, Line Graph Exercise, recursive-graph-body-print, Readying a Graph | |
15928 | @section Need for Printed Axes | |
15929 | ||
15930 | A graph needs printed axes, so you can orient yourself. For a do-once | |
15931 | project, it may be reasonable to draw the axes by hand using Emacs' | |
15932 | Picture mode; but a graph drawing function may be used more than once. | |
15933 | ||
15934 | For this reason, I have written enhancements to the basic | |
15935 | @code{print-graph-body} function that automatically print labels for | |
15936 | the horizontal and vertical axes. Since the label printing functions | |
15937 | do not contain much new material, I have placed their description in | |
15938 | an appendix. @xref{Full Graph, , A Graph with Labelled Axes}. | |
15939 | ||
15940 | @node Line Graph Exercise, , Printed Axes, Readying a Graph | |
15941 | @section Exercise | |
15942 | ||
15943 | Write a line graph version of the graph printing functions. | |
15944 | ||
15945 | @node Emacs Initialization, Debugging, Readying a Graph, Top | |
15946 | @chapter Your @file{.emacs} File | |
15947 | @cindex @file{.emacs} file | |
15948 | @cindex Customizing your @file{.emacs} file | |
15949 | @cindex Initialization file | |
15950 | ||
15951 | ``You don't have to like Emacs to like it'' -- this seemingly | |
15952 | paradoxical statement is the secret of GNU Emacs. The plain, `out of | |
15953 | the box' Emacs is a generic tool. Most people who use it, customize | |
15954 | it to suit themselves. | |
15955 | ||
15956 | GNU Emacs is mostly written in Emacs Lisp; this means that by writing | |
15957 | expressions in Emacs Lisp you can change or extend Emacs. | |
15958 | ||
15959 | @menu | |
15960 | * Default Configuration:: | |
15961 | * Site-wide Init:: You can write site-wide init files. | |
15962 | * defcustom:: Emacs will write code for you. | |
15963 | * Beginning a .emacs File:: How to write a @code{.emacs file}. | |
15964 | * Text and Auto-fill:: Automatically wrap lines. | |
15965 | * Mail Aliases:: Use abbreviations for email addresses. | |
15966 | * Indent Tabs Mode:: Don't use tabs with @TeX{} | |
15967 | * Keybindings:: Create some personal keybindings. | |
15968 | * Keymaps:: More about key binding. | |
15969 | * Loading Files:: Load (i.e., evaluate) files automatically. | |
15970 | * Autoload:: Make functions available. | |
15971 | * Simple Extension:: Define a function; bind it to a key. | |
15972 | * X11 Colors:: Colors in version 19 in X. | |
15973 | * Miscellaneous:: | |
15974 | * Mode Line:: How to customize your mode line. | |
15975 | @end menu | |
15976 | ||
15977 | @node Default Configuration, Site-wide Init, Emacs Initialization, Emacs Initialization | |
15978 | @ifnottex | |
15979 | @unnumberedsec Emacs' Default Configuration | |
15980 | @end ifnottex | |
15981 | ||
15982 | There are those who appreciate Emacs' default configuration. After | |
15983 | all, Emacs starts you in C mode when you edit a C file, starts you in | |
15984 | Fortran mode when you edit a Fortran file, and starts you in | |
15985 | Fundamental mode when you edit an unadorned file. This all makes | |
15986 | sense, if you do not know who is going to use Emacs. Who knows what a | |
15987 | person hopes to do with an unadorned file? Fundamental mode is the | |
15988 | right default for such a file, just as C mode is the right default for | |
15989 | editing C code. But when you do know who is going to use Emacs---you, | |
15990 | yourself---then it makes sense to customize Emacs. | |
15991 | ||
15992 | For example, I seldom want Fundamental mode when I edit an | |
15993 | otherwise undistinguished file; I want Text mode. This is why I | |
15994 | customize Emacs: so it suits me. | |
15995 | ||
15996 | You can customize and extend Emacs by writing or adapting a | |
15997 | @file{~/.emacs} file. This is your personal initialization file; its | |
15998 | contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You | |
15999 | may also add @file{.el} to @file{~/.emacs} and call it a | |
16000 | @file{~/.emacs.el} file. In the past, you were forbidden to type the | |
16001 | extra keystrokes that the name @file{~/.emacs.el} requires, but now | |
16002 | you may. The new format is consistent with the Emacs Lisp file | |
16003 | naming conventions; the old format saves typing.} | |
16004 | ||
16005 | A @file{~/.emacs} file contains Emacs Lisp code. You can write this | |
16006 | code yourself; or you can use Emacs' @code{customize} feature to write | |
16007 | the code for you. You can combine your own expressions and | |
16008 | auto-written Customize expressions in your @file{.emacs} file. | |
16009 | ||
16010 | (I myself prefer to write my own expressions, except for those, | |
16011 | particularly fonts, that I find easier to manipulate using the | |
16012 | @code{customize} command. I combine the two methods.) | |
16013 | ||
16014 | Most of this chapter is about writing expressions yourself. It | |
16015 | describes a simple @file{.emacs} file; for more information, see | |
16016 | @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and | |
16017 | @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference | |
16018 | Manual}. | |
16019 | ||
16020 | @node Site-wide Init, defcustom, Default Configuration, Emacs Initialization | |
16021 | @section Site-wide Initialization Files | |
16022 | ||
16023 | @cindex @file{default.el} init file | |
16024 | @cindex @file{site-init.el} init file | |
16025 | @cindex @file{site-load.el} init file | |
16026 | In addition to your personal initialization file, Emacs automatically | |
16027 | loads various site-wide initialization files, if they exist. These | |
16028 | have the same form as your @file{.emacs} file, but are loaded by | |
16029 | everyone. | |
16030 | ||
16031 | Two site-wide initialization files, @file{site-load.el} and | |
16032 | @file{site-init.el}, are loaded into Emacs and then `dumped' if a | |
16033 | `dumped' version of Emacs is created, as is most common. (Dumped | |
16034 | copies of Emacs load more quickly. However, once a file is loaded and | |
16035 | dumped, a change to it does not lead to a change in Emacs unless you | |
16036 | load it yourself or re-dump Emacs. @xref{Building Emacs, , Building | |
16037 | Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the | |
16038 | @file{INSTALL} file.) | |
16039 | ||
16040 | Three other site-wide initialization files are loaded automatically | |
16041 | each time you start Emacs, if they exist. These are | |
16042 | @file{site-start.el}, which is loaded @emph{before} your @file{.emacs} | |
16043 | file, and @file{default.el}, and the terminal type file, which are both | |
16044 | loaded @emph{after} your @file{.emacs} file. | |
16045 | ||
16046 | Settings and definitions in your @file{.emacs} file will overwrite | |
16047 | conflicting settings and definitions in a @file{site-start.el} file, | |
16048 | if it exists; but the settings and definitions in a @file{default.el} | |
16049 | or terminal type file will overwrite those in your @file{.emacs} file. | |
16050 | (You can prevent interference from a terminal type file by setting | |
16051 | @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A | |
16052 | Simple Extension}.) | |
16053 | ||
16054 | @c Rewritten to avoid overfull hbox. | |
16055 | The @file{INSTALL} file that comes in the distribution contains | |
16056 | descriptions of the @file{site-init.el} and @file{site-load.el} files. | |
16057 | ||
16058 | The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files | |
16059 | control loading. These files are in the @file{lisp} directory of the | |
16060 | Emacs distribution and are worth perusing. | |
16061 | ||
16062 | The @file{loaddefs.el} file contains a good many suggestions as to | |
16063 | what to put into your own @file{.emacs} file, or into a site-wide | |
16064 | initialization file. | |
16065 | ||
16066 | @node defcustom, Beginning a .emacs File, Site-wide Init, Emacs Initialization | |
16067 | @section Specifying Variables using @code{defcustom} | |
16068 | @findex defcustom | |
16069 | ||
16070 | You can specify variables using @code{defcustom} so that you and | |
16071 | others can then can use Emacs' @code{customize} feature to set their | |
16072 | values. (You cannot use @code{customize} to write function | |
16073 | definitions; but you can write @code{defuns} in your @file{.emacs} | |
16074 | file. Indeed, you can write any Lisp expression in your @file{.emacs} | |
16075 | file.) | |
16076 | ||
16077 | The @code{customize} feature depends on the @code{defcustom} special | |
16078 | form. Although you can use @code{defvar} or @code{setq} for variables | |
16079 | that users set, the @code{defcustom} special form is designed for the | |
16080 | job. | |
16081 | ||
16082 | You can use your knowledge of @code{defvar} for writing the | |
16083 | first three arguments for @code{defcustom}. The first argument to | |
16084 | @code{defcustom} is the name of the variable. The second argument is | |
16085 | the variable's initial value, if any; and this value is set only if | |
16086 | the value has not already been set. The third argument is the | |
16087 | documentation. | |
16088 | ||
16089 | The fourth and subsequent arguments to @code{defcustom} specify types | |
16090 | and options; these are not featured in @code{defvar}. (These | |
16091 | arguments are optional.) | |
16092 | ||
16093 | Each of these arguments consists of a keyword followed by a value. | |
16094 | Each keyword starts with the character @code{:}. | |
16095 | ||
16096 | @need 1250 | |
16097 | For example, the customizable user option variable | |
16098 | @code{text-mode-hook} looks like this: | |
16099 | ||
16100 | @smallexample | |
16101 | @group | |
16102 | (defcustom text-mode-hook nil | |
16103 | "Normal hook run when entering Text mode and many related modes." | |
16104 | :type 'hook | |
16105 | :options '(turn-on-auto-fill flyspell-mode) | |
16106 | :group 'data) | |
16107 | @end group | |
16108 | @end smallexample | |
16109 | ||
16110 | @noindent | |
16111 | The name of the variable is @code{text-mode-hook}; it has no default | |
16112 | value; and its documentation string tells you what it does. | |
16113 | ||
16114 | The @code{:type} keyword tells Emacs what kind of data | |
16115 | @code{text-mode-hook} should be set to and how to display the value in | |
16116 | a Customization buffer. | |
16117 | ||
16118 | The @code{:options} keyword specifies a suggested list of values for | |
16119 | the variable. Currently, you can use @code{:options} only for a hook. | |
16120 | The list is only a suggestion; it is not exclusive; a person who sets | |
16121 | the variable may set it to other values; the list shown following the | |
16122 | @code{:options} keyword is intended to offer convenient choices to a | |
16123 | user. | |
16124 | ||
16125 | Finally, the @code{:group} keyword tells the Emacs Customization | |
16126 | command in which group the variable is located. This tells where to | |
16127 | find it. | |
16128 | ||
16129 | For more information, see @ref{Customization, , Writing Customization | |
16130 | Definitions, elisp, The GNU Emacs Lisp Reference Manual}. | |
16131 | ||
16132 | Consider @code{text-mode-hook} as an example. | |
16133 | ||
16134 | There are two ways to customize this variable. You can use the | |
16135 | customization command or write the appropriate expressions yourself. | |
16136 | ||
16137 | @need 800 | |
16138 | Using the customization command, you can type: | |
16139 | ||
16140 | @smallexample | |
16141 | M-x customize | |
16142 | @end smallexample | |
16143 | ||
16144 | @noindent | |
16145 | and find that the group for editing files of data is called `data'. | |
16146 | Enter that group. Text Mode Hook is the first member. You can click | |
16147 | on its various options to set the values. After you click on the | |
16148 | button to | |
16149 | ||
16150 | @smallexample | |
16151 | Save for Future Sessions | |
16152 | @end smallexample | |
16153 | ||
16154 | @noindent | |
16155 | Emacs will write an expression into your @file{.emacs} file. | |
16156 | It will look like this: | |
16157 | ||
16158 | @smallexample | |
16159 | @group | |
16160 | (custom-set-variables | |
16161 | ;; custom-set-variables was added by Custom -- | |
16162 | ;; don't edit or cut/paste it! | |
16163 | ;; Your init file should contain only one such instance. | |
16164 | '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify)))) | |
16165 | @end group | |
16166 | @end smallexample | |
16167 | ||
16168 | @noindent | |
16169 | (The @code{text-mode-hook-identify} function tells | |
16170 | @code{toggle-text-mode-auto-fill} which buffers are in Text mode.) | |
16171 | ||
16172 | In spite of the warning, you certainly may edit, cut, and paste the | |
16173 | expression! I do all time. The purpose of the warning is to scare | |
16174 | those who do not know what they are doing, so they do not | |
16175 | inadvertently generate an error. | |
16176 | ||
16177 | The @code{custom-set-variables} works somewhat differently than a | |
16178 | @code{setq}. While I have never learned the differences, I do modify | |
16179 | the @code{custom-set-variables} expressions in my @file{.emacs} file | |
16180 | by hand: I make the changes in what appears to me to be a reasonable | |
16181 | manner and have not had any problems. Others prefer to use the | |
16182 | Customization command and let Emacs do the work for them. | |
16183 | ||
16184 | Another @code{custom-set-@dots{}} function is @code{custom-set-faces}. | |
16185 | This function sets the various font faces. Over time, I have set a | |
16186 | considerable number of faces. Some of the time, I re-set them using | |
16187 | @code{customize}; other times, I simply edit the | |
16188 | @code{custom-set-faces} expression in my @file{.emacs} file itself. | |
16189 | ||
16190 | The second way to customize your @code{text-mode-hook} is to set it | |
16191 | yourself in your @file{.emacs} file using code that has nothing to do | |
16192 | with the @code{custom-set-@dots{}} functions. | |
16193 | ||
16194 | @need 800 | |
16195 | When you do this, and later use @code{customize}, you will see a | |
16196 | message that says | |
16197 | ||
16198 | @smallexample | |
16199 | this option has been changed outside the customize buffer. | |
16200 | @end smallexample | |
16201 | ||
16202 | @need 800 | |
16203 | This message is only a warning. If you click on the button to | |
16204 | ||
16205 | @smallexample | |
16206 | Save for Future Sessions | |
16207 | @end smallexample | |
16208 | ||
16209 | @noindent | |
16210 | Emacs will write a @code{custom-set-@dots{}} expression near the end | |
16211 | of your @file{.emacs} file that will be evaluated after your | |
16212 | hand-written expression. It will, therefore, overrule your | |
16213 | hand-written expression. No harm will be done. When you do this, | |
16214 | however, be careful to remember which expression is active; if you | |
16215 | forget, you may confuse yourself. | |
16216 | ||
16217 | So long as you remember where the values are set, you will have no | |
16218 | trouble. In any event, the values are always set in your | |
16219 | initialization file, which is usually called @file{.emacs}. | |
16220 | ||
16221 | I myself use @code{customize} for hardly anything. Mostly, I write | |
16222 | expressions myself. | |
16223 | ||
16224 | @node Beginning a .emacs File, Text and Auto-fill, defcustom, Emacs Initialization | |
16225 | @section Beginning a @file{.emacs} File | |
16226 | @cindex @file{.emacs} file, beginning of | |
16227 | ||
16228 | When you start Emacs, it loads your @file{.emacs} file unless you tell | |
16229 | it not to by specifying @samp{-q} on the command line. (The | |
16230 | @code{emacs -q} command gives you a plain, out-of-the-box Emacs.) | |
16231 | ||
16232 | A @file{.emacs} file contains Lisp expressions. Often, these are no | |
16233 | more than expressions to set values; sometimes they are function | |
16234 | definitions. | |
16235 | ||
16236 | @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs | |
16237 | Manual}, for a short description of initialization files. | |
16238 | ||
16239 | This chapter goes over some of the same ground, but is a walk among | |
16240 | extracts from a complete, long-used @file{.emacs} file---my own. | |
16241 | ||
16242 | The first part of the file consists of comments: reminders to myself. | |
16243 | By now, of course, I remember these things, but when I started, I did | |
16244 | not. | |
16245 | ||
16246 | @smallexample | |
16247 | @group | |
16248 | ;;;; Bob's .emacs file | |
16249 | ; Robert J. Chassell | |
16250 | ; 26 September 1985 | |
16251 | @end group | |
16252 | @end smallexample | |
16253 | ||
16254 | @noindent | |
16255 | Look at that date! I started this file a long time ago. I have been | |
16256 | adding to it ever since. | |
16257 | ||
16258 | @smallexample | |
16259 | @group | |
16260 | ; Each section in this file is introduced by a | |
16261 | ; line beginning with four semicolons; and each | |
16262 | ; entry is introduced by a line beginning with | |
16263 | ; three semicolons. | |
16264 | @end group | |
16265 | @end smallexample | |
16266 | ||
16267 | @noindent | |
16268 | This describes the usual conventions for comments in Emacs Lisp. | |
16269 | Everything on a line that follows a semicolon is a comment. Two, | |
16270 | three, and four semicolons are used as section and subsection | |
16271 | markers. (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference | |
16272 | Manual}, for more about comments.) | |
16273 | ||
16274 | @smallexample | |
16275 | @group | |
16276 | ;;;; The Help Key | |
16277 | ; Control-h is the help key; | |
16278 | ; after typing control-h, type a letter to | |
16279 | ; indicate the subject about which you want help. | |
16280 | ; For an explanation of the help facility, | |
16281 | ; type control-h two times in a row. | |
16282 | @end group | |
16283 | @end smallexample | |
16284 | ||
16285 | @noindent | |
16286 | Just remember: type @kbd{C-h} two times for help. | |
16287 | ||
16288 | @smallexample | |
16289 | @group | |
16290 | ; To find out about any mode, type control-h m | |
16291 | ; while in that mode. For example, to find out | |
16292 | ; about mail mode, enter mail mode and then type | |
16293 | ; control-h m. | |
16294 | @end group | |
16295 | @end smallexample | |
16296 | ||
16297 | @noindent | |
16298 | `Mode help', as I call this, is very helpful. Usually, it tells you | |
16299 | all you need to know. | |
16300 | ||
16301 | Of course, you don't need to include comments like these in your | |
16302 | @file{.emacs} file. I included them in mine because I kept forgetting | |
16303 | about Mode help or the conventions for comments---but I was able to | |
16304 | remember to look here to remind myself. | |
16305 | ||
16306 | @node Text and Auto-fill, Mail Aliases, Beginning a .emacs File, Emacs Initialization | |
16307 | @section Text and Auto Fill Mode | |
16308 | ||
16309 | Now we come to the part that `turns on' Text mode and | |
16310 | Auto Fill mode. | |
16311 | ||
16312 | @smallexample | |
16313 | @group | |
16314 | ;;; Text mode and Auto Fill mode | |
16315 | ; The next three lines put Emacs into Text mode | |
16316 | ; and Auto Fill mode, and are for writers who | |
16317 | ; want to start writing prose rather than code. | |
16318 | ||
16319 | (setq default-major-mode 'text-mode) | |
16320 | (add-hook 'text-mode-hook 'text-mode-hook-identify) | |
16321 | (add-hook 'text-mode-hook 'turn-on-auto-fill) | |
16322 | @end group | |
16323 | @end smallexample | |
16324 | ||
16325 | Here is the first part of this @file{.emacs} file that does something | |
16326 | besides remind a forgetful human! | |
16327 | ||
16328 | The first of the two lines in parentheses tells Emacs to turn on Text | |
16329 | mode when you find a file, @emph{unless} that file should go into some | |
16330 | other mode, such as C mode. | |
16331 | ||
16332 | @cindex Per-buffer, local variables list | |
16333 | @cindex Local variables list, per-buffer, | |
16334 | @cindex Automatic mode selection | |
16335 | @cindex Mode selection, automatic | |
16336 | When Emacs reads a file, it looks at the extension to the file name, | |
16337 | if any. (The extension is the part that comes after a @samp{.}.) If | |
16338 | the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns | |
16339 | on C mode. Also, Emacs looks at first nonblank line of the file; if | |
16340 | the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs | |
16341 | possesses a list of extensions and specifications that it uses | |
16342 | automatically. In addition, Emacs looks near the last page for a | |
16343 | per-buffer, ``local variables list'', if any. | |
16344 | ||
16345 | @ifinfo | |
16346 | @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU | |
16347 | Emacs Manual}. | |
16348 | ||
16349 | @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs | |
16350 | Manual}. | |
16351 | @end ifinfo | |
16352 | @iftex | |
16353 | See sections ``How Major Modes are Chosen'' and ``Local Variables in | |
16354 | Files'' in @cite{The GNU Emacs Manual}. | |
16355 | @end iftex | |
16356 | ||
16357 | Now, back to the @file{.emacs} file. | |
16358 | ||
16359 | @need 800 | |
16360 | Here is the line again; how does it work? | |
16361 | ||
16362 | @cindex Text Mode turned on | |
16363 | @smallexample | |
16364 | (setq default-major-mode 'text-mode) | |
16365 | @end smallexample | |
16366 | ||
16367 | @noindent | |
16368 | This line is a short, but complete Emacs Lisp expression. | |
16369 | ||
16370 | We are already familiar with @code{setq}. It sets the following variable, | |
16371 | @code{default-major-mode}, to the subsequent value, which is | |
16372 | @code{text-mode}. The single quote mark before @code{text-mode} tells | |
16373 | Emacs to deal directly with the @code{text-mode} variable, not with | |
16374 | whatever it might stand for. @xref{set & setq, , Setting the Value of | |
16375 | a Variable}, for a reminder of how @code{setq} works. The main point | |
16376 | is that there is no difference between the procedure you use to set | |
16377 | a value in your @file{.emacs} file and the procedure you use anywhere | |
16378 | else in Emacs. | |
16379 | ||
16380 | @need 800 | |
16381 | Here are the next two lines: | |
16382 | ||
16383 | @cindex Auto Fill mode turned on | |
16384 | @findex add-hook | |
16385 | @smallexample | |
16386 | (add-hook 'text-mode-hook 'text-mode-hook-identify) | |
16387 | (add-hook 'text-mode-hook 'turn-on-auto-fill) | |
16388 | @end smallexample | |
16389 | ||
16390 | @noindent | |
16391 | In these two lines, the @code{add-hook} command first adds | |
16392 | @code{text-mode-hook-identify} to the variable called | |
16393 | @code{text-mode-hook} and then adds @code{turn-on-auto-fill} to the | |
16394 | variable. | |
16395 | ||
16396 | @code{turn-on-auto-fill} is the name of a program, that, you guessed | |
16397 | it!, turns on Auto Fill mode. @code{text-mode-hook-identify} is a | |
16398 | function that tells @code{toggle-text-mode-auto-fill} which buffers | |
16399 | are in Text mode. | |
16400 | ||
16401 | Every time Emacs turns on Text mode, Emacs runs the commands `hooked' | |
16402 | onto Text mode. So every time Emacs turns on Text mode, Emacs also | |
16403 | turns on Auto Fill mode. | |
16404 | ||
16405 | In brief, the first line causes Emacs to enter Text mode when you edit | |
16406 | a file, unless the file name extension, first non-blank line, or local | |
16407 | variables tell Emacs otherwise. | |
16408 | ||
16409 | Text mode among other actions, sets the syntax table to work | |
16410 | conveniently for writers. In Text mode, Emacs considers an apostrophe | |
16411 | as part of a word like a letter; but Emacs does not consider a period | |
16412 | or a space as part of a word. Thus, @kbd{M-f} moves you over | |
16413 | @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after | |
16414 | the @samp{t} of @samp{it's}. | |
16415 | ||
16416 | The second and third lines causes Emacs to turn on Auto Fill mode when | |
16417 | it turns on Text mode. In Auto Fill mode, Emacs automatically breaks | |
16418 | a line that is too wide and brings the excessively wide part of the | |
16419 | line down to the next line. Emacs breaks lines between words, not | |
16420 | within them. | |
16421 | ||
16422 | When Auto Fill mode is turned off, lines continue to the right as you | |
16423 | type them. Depending on how you set the value of | |
16424 | @code{truncate-lines}, the words you type either disappear off the | |
16425 | right side of the screen, or else are shown, in a rather ugly and | |
16426 | unreadable manner, as a continuation line on the screen. | |
16427 | ||
16428 | @need 1250 | |
16429 | In addition, in this part of my @file{.emacs} file, I tell the Emacs | |
16430 | fill commands to insert two spaces after a colon: | |
16431 | ||
16432 | @smallexample | |
16433 | (setq colon-double-space t) | |
16434 | @end smallexample | |
16435 | ||
16436 | @node Mail Aliases, Indent Tabs Mode, Text and Auto-fill, Emacs Initialization | |
16437 | @section Mail Aliases | |
16438 | ||
16439 | Here is a @code{setq} that `turns on' mail aliases, along with more | |
16440 | reminders. | |
16441 | ||
16442 | @smallexample | |
16443 | @group | |
16444 | ;;; Mail mode | |
16445 | ; To enter mail mode, type `C-x m' | |
16446 | ; To enter RMAIL (for reading mail), | |
16447 | ; type `M-x rmail' | |
16448 | ||
16449 | (setq mail-aliases t) | |
16450 | @end group | |
16451 | @end smallexample | |
16452 | ||
16453 | @cindex Mail aliases | |
16454 | @noindent | |
16455 | This @code{setq} command sets the value of the variable | |
16456 | @code{mail-aliases} to @code{t}. Since @code{t} means true, the line | |
16457 | says, in effect, ``Yes, use mail aliases.'' | |
16458 | ||
16459 | Mail aliases are convenient short names for long email addresses or | |
16460 | for lists of email addresses. The file where you keep your `aliases' | |
16461 | is @file{~/.mailrc}. You write an alias like this: | |
16462 | ||
16463 | @smallexample | |
16464 | alias geo george@@foobar.wiz.edu | |
16465 | @end smallexample | |
16466 | ||
16467 | @noindent | |
16468 | When you write a message to George, address it to @samp{geo}; the | |
16469 | mailer will automatically expand @samp{geo} to the full address. | |
16470 | ||
16471 | @node Indent Tabs Mode, Keybindings, Mail Aliases, Emacs Initialization | |
16472 | @section Indent Tabs Mode | |
16473 | @cindex Tabs, preventing | |
16474 | @findex indent-tabs-mode | |
16475 | ||
16476 | By default, Emacs inserts tabs in place of multiple spaces when it | |
16477 | formats a region. (For example, you might indent many lines of text | |
16478 | all at once with the @code{indent-region} command.) Tabs look fine on | |
16479 | a terminal or with ordinary printing, but they produce badly indented | |
16480 | output when you use @TeX{} or Texinfo since @TeX{} ignores tabs. | |
16481 | ||
16482 | @need 1250 | |
16483 | The following turns off Indent Tabs mode: | |
16484 | ||
16485 | @smallexample | |
16486 | @group | |
16487 | ;;; Prevent Extraneous Tabs | |
16488 | (setq-default indent-tabs-mode nil) | |
16489 | @end group | |
16490 | @end smallexample | |
16491 | ||
16492 | Note that this line uses @code{setq-default} rather than the | |
16493 | @code{setq} command that we have seen before. The @code{setq-default} | |
16494 | command sets values only in buffers that do not have their own local | |
16495 | values for the variable. | |
16496 | ||
16497 | @ifinfo | |
16498 | @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}. | |
16499 | ||
16500 | @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs | |
16501 | Manual}. | |
16502 | @end ifinfo | |
16503 | @iftex | |
16504 | See sections ``Tabs vs.@: Spaces'' and ``Local Variables in | |
16505 | Files'' in @cite{The GNU Emacs Manual}. | |
16506 | @end iftex | |
16507 | ||
16508 | @node Keybindings, Keymaps, Indent Tabs Mode, Emacs Initialization | |
16509 | @section Some Keybindings | |
16510 | ||
16511 | Now for some personal keybindings: | |
16512 | ||
16513 | @smallexample | |
16514 | @group | |
16515 | ;;; Compare windows | |
16516 | (global-set-key "\C-cw" 'compare-windows) | |
16517 | @end group | |
16518 | @end smallexample | |
16519 | ||
16520 | @findex compare-windows | |
16521 | @code{compare-windows} is a nifty command that compares the text in | |
16522 | your current window with text in the next window. It makes the | |
16523 | comparison by starting at point in each window, moving over text in | |
16524 | each window as far as they match. I use this command all the time. | |
16525 | ||
16526 | This also shows how to set a key globally, for all modes. | |
16527 | ||
16528 | @cindex Setting a key globally | |
16529 | @cindex Global set key | |
16530 | @cindex Key setting globally | |
16531 | @findex global-set-key | |
16532 | The command is @code{global-set-key}. It is followed by the | |
16533 | keybinding. In a @file{.emacs} file, the keybinding is written as | |
16534 | shown: @code{\C-c} stands for `control-c', which means `press the | |
16535 | control key and the @kbd{c} key at the same time'. The @code{w} means | |
16536 | `press the @kbd{w} key'. The keybinding is surrounded by double | |
16537 | quotation marks. In documentation, you would write this as @kbd{C-c | |
16538 | w}. (If you were binding a @key{META} key, such as @kbd{M-c}, rather | |
16539 | than a @key{CTL} key, you would write @code{\M-c}. @xref{Init | |
16540 | Rebinding, , Rebinding Keys in Your Init File, emacs, The GNU Emacs | |
16541 | Manual}, for details.) | |
16542 | ||
16543 | The command invoked by the keys is @code{compare-windows}. Note that | |
16544 | @code{compare-windows} is preceded by a single quote; otherwise, Emacs | |
16545 | would first try to evaluate the symbol to determine its value. | |
16546 | ||
16547 | These three things, the double quotation marks, the backslash before | |
16548 | the @samp{C}, and the single quote mark are necessary parts of | |
16549 | keybinding that I tend to forget. Fortunately, I have come to | |
16550 | remember that I should look at my existing @file{.emacs} file, and | |
16551 | adapt what is there. | |
16552 | ||
16553 | As for the keybinding itself: @kbd{C-c w}. This combines the prefix | |
16554 | key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This | |
16555 | set of keys, @kbd{C-c} followed by a single character, is strictly | |
16556 | reserved for individuals' own use. (I call these `own' keys, since | |
16557 | these are for my own use.) You should always be able to create such a | |
16558 | keybinding for your own use without stomping on someone else's | |
16559 | keybinding. If you ever write an extension to Emacs, please avoid | |
16560 | taking any of these keys for public use. Create a key like @kbd{C-c | |
16561 | C-w} instead. Otherwise, we will run out of `own' keys. | |
16562 | ||
16563 | @need 1250 | |
16564 | Here is another keybinding, with a comment: | |
16565 | ||
16566 | @smallexample | |
16567 | @group | |
16568 | ;;; Keybinding for `occur' | |
16569 | ; I use occur a lot, so let's bind it to a key: | |
16570 | (global-set-key "\C-co" 'occur) | |
16571 | @end group | |
16572 | @end smallexample | |
16573 | ||
16574 | @findex occur | |
16575 | The @code{occur} command shows all the lines in the current buffer | |
16576 | that contain a match for a regular expression. Matching lines are | |
16577 | shown in a buffer called @file{*Occur*}. That buffer serves as a menu | |
16578 | to jump to occurrences. | |
16579 | ||
16580 | @findex global-unset-key | |
16581 | @cindex Unbinding key | |
16582 | @cindex Key unbinding | |
16583 | @need 1250 | |
16584 | Here is how to unbind a key, so it does not | |
16585 | work: | |
16586 | ||
16587 | @smallexample | |
16588 | @group | |
16589 | ;;; Unbind `C-x f' | |
16590 | (global-unset-key "\C-xf") | |
16591 | @end group | |
16592 | @end smallexample | |
16593 | ||
16594 | There is a reason for this unbinding: I found I inadvertently typed | |
16595 | @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a | |
16596 | file, as I intended, I accidentally set the width for filled text, | |
16597 | almost always to a width I did not want. Since I hardly ever reset my | |
16598 | default width, I simply unbound the key. | |
16599 | ||
16600 | @findex list-buffers, @r{rebound} | |
16601 | @findex buffer-menu, @r{bound to key} | |
16602 | @need 1250 | |
16603 | The following rebinds an existing key: | |
16604 | ||
16605 | @smallexample | |
16606 | @group | |
16607 | ;;; Rebind `C-x C-b' for `buffer-menu' | |
16608 | (global-set-key "\C-x\C-b" 'buffer-menu) | |
16609 | @end group | |
16610 | @end smallexample | |
16611 | ||
16612 | By default, @kbd{C-x C-b} runs the | |
16613 | @code{list-buffers} command. This command lists | |
16614 | your buffers in @emph{another} window. Since I | |
16615 | almost always want to do something in that | |
16616 | window, I prefer the @code{buffer-menu} | |
16617 | command, which not only lists the buffers, | |
16618 | but moves point into that window. | |
16619 | ||
16620 | @node Keymaps, Loading Files, Keybindings, Emacs Initialization | |
16621 | @section Keymaps | |
16622 | @cindex Keymaps | |
16623 | @cindex Rebinding keys | |
16624 | ||
16625 | Emacs uses @dfn{keymaps} to record which keys call which commands. | |
16626 | When you use @code{global-set-key} to set the keybinding for a single | |
16627 | command in all parts of Emacs, you are specifying the keybinding in | |
16628 | @code{current-global-map}. | |
16629 | ||
16630 | Specific modes, such as C mode or Text mode, have their own keymaps; | |
16631 | the mode-specific keymaps override the global map that is shared by | |
16632 | all buffers. | |
16633 | ||
16634 | The @code{global-set-key} function binds, or rebinds, the global | |
16635 | keymap. For example, the following binds the key @kbd{C-x C-b} to the | |
16636 | function @code{buffer-menu}: | |
16637 | ||
16638 | @smallexample | |
16639 | (global-set-key "\C-x\C-b" 'buffer-menu) | |
16640 | @end smallexample | |
16641 | ||
16642 | Mode-specific keymaps are bound using the @code{define-key} function, | |
16643 | which takes a specific keymap as an argument, as well as the key and | |
16644 | the command. For example, my @file{.emacs} file contains the | |
16645 | following expression to bind the @code{texinfo-insert-@@group} command | |
16646 | to @kbd{C-c C-c g}: | |
16647 | ||
16648 | @smallexample | |
16649 | @group | |
16650 | (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group) | |
16651 | @end group | |
16652 | @end smallexample | |
16653 | ||
16654 | @noindent | |
16655 | The @code{texinfo-insert-@@group} function itself is a little extension | |
16656 | to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I | |
16657 | use this command all the time and prefer to type the three strokes | |
16658 | @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}. | |
16659 | (@samp{@@group} and its matching @samp{@@end group} are commands that | |
16660 | keep all enclosed text together on one page; many multi-line examples | |
16661 | in this book are surrounded by @samp{@@group @dots{} @@end group}.) | |
16662 | ||
16663 | @need 1250 | |
16664 | Here is the @code{texinfo-insert-@@group} function definition: | |
16665 | ||
16666 | @smallexample | |
16667 | @group | |
16668 | (defun texinfo-insert-@@group () | |
16669 | "Insert the string @@group in a Texinfo buffer." | |
16670 | (interactive) | |
16671 | (beginning-of-line) | |
16672 | (insert "@@group\n")) | |
16673 | @end group | |
16674 | @end smallexample | |
16675 | ||
16676 | (Of course, I could have used Abbrev mode to save typing, rather than | |
16677 | write a function to insert a word; but I prefer key strokes consistent | |
16678 | with other Texinfo mode key bindings.) | |
16679 | ||
16680 | You will see numerous @code{define-key} expressions in | |
16681 | @file{loaddefs.el} as well as in the various mode libraries, such as | |
16682 | @file{cc-mode.el} and @file{lisp-mode.el}. | |
16683 | ||
16684 | @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs | |
16685 | Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp | |
16686 | Reference Manual}, for more information about keymaps. | |
16687 | ||
16688 | @node Loading Files, Autoload, Keymaps, Emacs Initialization | |
16689 | @section Loading Files | |
16690 | @cindex Loading files | |
16691 | @c findex load | |
16692 | ||
16693 | Many people in the GNU Emacs community have written extensions to | |
16694 | Emacs. As time goes by, these extensions are often included in new | |
16695 | releases. For example, the Calendar and Diary packages are now part | |
16696 | of the standard GNU Emacs. | |
16697 | ||
16698 | (Calc, which I consider a vital part of Emacs, would be part of the | |
16699 | standard distribution except that it was so large it was packaged | |
16700 | separately and no one has changed that.) | |
16701 | ||
16702 | You can use a @code{load} command to evaluate a complete file and | |
16703 | thereby install all the functions and variables in the file into Emacs. | |
16704 | For example: | |
16705 | ||
16706 | @c (auto-compression-mode t) | |
16707 | ||
16708 | @smallexample | |
16709 | (load "~/emacs/slowsplit") | |
16710 | @end smallexample | |
16711 | ||
16712 | This evaluates, i.e.@: loads, the @file{slowsplit.el} file or if it | |
16713 | exists, the faster, byte compiled @file{slowsplit.elc} file from the | |
16714 | @file{emacs} sub-directory of your home directory. The file contains | |
16715 | the function @code{split-window-quietly}, which John Robinson wrote in | |
16716 | 1989. | |
16717 | ||
16718 | The @code{split-window-quietly} function splits a window with the | |
16719 | minimum of redisplay. I installed it in 1989 because it worked well | |
16720 | with the slow 1200 baud terminals I was then using. Nowadays, I only | |
16721 | occasionally come across such a slow connection, but I continue to use | |
16722 | the function because I like the way it leaves the bottom half of a | |
16723 | buffer in the lower of the new windows and the top half in the upper | |
16724 | window. | |
16725 | ||
16726 | @need 1250 | |
16727 | To replace the key binding for the default | |
16728 | @code{split-window-vertically}, you must also unset that key and bind | |
16729 | the keys to @code{split-window-quietly}, like this: | |
16730 | ||
16731 | @smallexample | |
16732 | @group | |
16733 | (global-unset-key "\C-x2") | |
16734 | (global-set-key "\C-x2" 'split-window-quietly) | |
16735 | @end group | |
16736 | @end smallexample | |
16737 | ||
16738 | @vindex load-path | |
16739 | If you load many extensions, as I do, then instead of specifying the | |
16740 | exact location of the extension file, as shown above, you can specify | |
16741 | that directory as part of Emacs' @code{load-path}. Then, when Emacs | |
16742 | loads a file, it will search that directory as well as its default | |
16743 | list of directories. (The default list is specified in @file{paths.h} | |
16744 | when Emacs is built.) | |
16745 | ||
16746 | @need 1250 | |
16747 | The following command adds your @file{~/emacs} directory to the | |
16748 | existing load path: | |
16749 | ||
16750 | @smallexample | |
16751 | @group | |
16752 | ;;; Emacs Load Path | |
16753 | (setq load-path (cons "~/emacs" load-path)) | |
16754 | @end group | |
16755 | @end smallexample | |
16756 | ||
16757 | Incidentally, @code{load-library} is an interactive interface to the | |
16758 | @code{load} function. The complete function looks like this: | |
16759 | ||
16760 | @findex load-library | |
16761 | @smallexample | |
16762 | @group | |
16763 | (defun load-library (library) | |
16764 | "Load the library named LIBRARY. | |
16765 | This is an interface to the function `load'." | |
16766 | (interactive "sLoad library: ") | |
16767 | (load library)) | |
16768 | @end group | |
16769 | @end smallexample | |
16770 | ||
16771 | The name of the function, @code{load-library}, comes from the use of | |
16772 | `library' as a conventional synonym for `file'. The source for the | |
16773 | @code{load-library} command is in the @file{files.el} library. | |
16774 | ||
16775 | Another interactive command that does a slightly different job is | |
16776 | @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for | |
16777 | Emacs, emacs, The GNU Emacs Manual}, for information on the | |
16778 | distinction between @code{load-library} and this command. | |
16779 | ||
16780 | @node Autoload, Simple Extension, Loading Files, Emacs Initialization | |
16781 | @section Autoloading | |
16782 | @findex autoload | |
16783 | ||
16784 | Instead of installing a function by loading the file that contains it, | |
16785 | or by evaluating the function definition, you can make the function | |
16786 | available but not actually install it until it is first called. This | |
16787 | is called @dfn{autoloading}. | |
16788 | ||
16789 | When you execute an autoloaded function, Emacs automatically evaluates | |
16790 | the file that contains the definition, and then calls the function. | |
16791 | ||
16792 | Emacs starts quicker with autoloaded functions, since their libraries | |
16793 | are not loaded right away; but you need to wait a moment when you | |
16794 | first use such a function, while its containing file is evaluated. | |
16795 | ||
16796 | Rarely used functions are frequently autoloaded. The | |
16797 | @file{loaddefs.el} library contains hundreds of autoloaded functions, | |
16798 | from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may | |
16799 | come to use a `rare' function frequently. When you do, you should | |
16800 | load that function's file with a @code{load} expression in your | |
16801 | @file{.emacs} file. | |
16802 | ||
16803 | In my @file{.emacs} file for Emacs version 21, I load 12 libraries | |
16804 | that contain functions that would otherwise be autoloaded. (Actually, | |
16805 | it would have been better to include these files in my `dumped' Emacs | |
16806 | when I built it, but I forgot. @xref{Building Emacs, , Building | |
16807 | Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the @file{INSTALL} | |
16808 | file for more about dumping.) | |
16809 | ||
16810 | You may also want to include autoloaded expressions in your @file{.emacs} | |
16811 | file. @code{autoload} is a built-in function that takes up to five | |
16812 | arguments, the final three of which are optional. The first argument | |
16813 | is the name of the function to be autoloaded; the second is the name | |
16814 | of the file to be loaded. The third argument is documentation for the | |
16815 | function, and the fourth tells whether the function can be called | |
16816 | interactively. The fifth argument tells what type of | |
16817 | object---@code{autoload} can handle a keymap or macro as well as a | |
16818 | function (the default is a function). | |
16819 | ||
16820 | @need 800 | |
16821 | Here is a typical example: | |
16822 | ||
16823 | @smallexample | |
16824 | @group | |
16825 | (autoload 'html-helper-mode | |
16826 | "html-helper-mode" "Edit HTML documents" t) | |
16827 | @end group | |
16828 | @end smallexample | |
16829 | ||
16830 | @noindent | |
16831 | (@code{html-helper-mode} is an alternative to @code{html-mode}, which | |
16832 | is a standard part of the distribution). | |
16833 | ||
16834 | @noindent | |
16835 | This expression autoloads the @code{html-helper-mode} function. It | |
16836 | takes it from the @file{html-helper-mode.el} file (or from the byte | |
16837 | compiled file @file{html-helper-mode.elc}, if it exists.) The file | |
16838 | must be located in a directory specified by @code{load-path}. The | |
16839 | documentation says that this is a mode to help you edit documents | |
16840 | written in the HyperText Markup Language. You can call this mode | |
16841 | interactively by typing @kbd{M-x html-helper-mode}. (You need to | |
16842 | duplicate the function's regular documentation in the autoload | |
16843 | expression because the regular function is not yet loaded, so its | |
16844 | documentation is not available.) | |
16845 | ||
16846 | @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference | |
16847 | Manual}, for more information. | |
16848 | ||
16849 | @node Simple Extension, X11 Colors, Autoload, Emacs Initialization | |
16850 | @section A Simple Extension: @code{line-to-top-of-window} | |
16851 | @findex line-to-top-of-window | |
16852 | @cindex Simple extension in @file{.emacs} file | |
16853 | ||
16854 | Here is a simple extension to Emacs that moves the line point is on to | |
16855 | the top of the window. I use this all the time, to make text easier | |
16856 | to read. | |
16857 | ||
16858 | You can put the following code into a separate file and then load it | |
16859 | from your @file{.emacs} file, or you can include it within your | |
16860 | @file{.emacs} file. | |
16861 | ||
16862 | @need 1250 | |
16863 | Here is the definition: | |
16864 | ||
16865 | @smallexample | |
16866 | @group | |
16867 | ;;; Line to top of window; | |
16868 | ;;; replace three keystroke sequence C-u 0 C-l | |
16869 | (defun line-to-top-of-window () | |
16870 | "Move the line point is on to top of window." | |
16871 | (interactive) | |
16872 | (recenter 0)) | |
16873 | @end group | |
16874 | @end smallexample | |
16875 | ||
16876 | @need 1250 | |
16877 | Now for the keybinding. | |
16878 | ||
16879 | Nowadays, function keys as well as mouse button events and | |
16880 | non-@sc{ascii} characters are written within square brackets, without | |
16881 | quotation marks. (In Emacs version 18 and before, you had to write | |
16882 | different function key bindings for each different make of terminal.) | |
16883 | ||
16884 | I bind @code{line-to-top-of-window} to my @key{F6} function key like | |
16885 | this: | |
16886 | ||
16887 | @smallexample | |
16888 | (global-set-key [f6] 'line-to-top-of-window) | |
16889 | @end smallexample | |
16890 | ||
16891 | For more information, see @ref{Init Rebinding, , Rebinding Keys in | |
16892 | Your Init File, emacs, The GNU Emacs Manual}. | |
16893 | ||
16894 | @cindex Conditional 'twixt two versions of Emacs | |
16895 | @cindex Version of Emacs, choosing | |
16896 | @cindex Emacs version, choosing | |
16897 | If you run two versions of GNU Emacs, such as versions 20 and 21, and | |
16898 | use one @file{.emacs} file, you can select which code to evaluate with | |
16899 | the following conditional: | |
16900 | ||
16901 | @smallexample | |
16902 | @group | |
16903 | (cond | |
16904 | ((string-equal (number-to-string 20) (substring (emacs-version) 10 12)) | |
16905 | ;; evaluate version 20 code | |
16906 | ( @dots{} )) | |
16907 | ((string-equal (number-to-string 21) (substring (emacs-version) 10 12)) | |
16908 | ;; evaluate version 21 code | |
16909 | ( @dots{} ))) | |
16910 | @end group | |
16911 | @end smallexample | |
16912 | ||
16913 | For example, in contrast to version 20, version 21 blinks its cursor | |
16914 | by default. I hate such blinking, as well as some other features in | |
16915 | version 21, so I placed the following in my @file{.emacs} | |
16916 | file@footnote{When I start instances of Emacs that do not load my | |
16917 | @file{.emacs} file or any site file, I also turn off blinking: | |
16918 | ||
16919 | @smallexample | |
16920 | emacs -q --no-site-file -eval '(blink-cursor-mode nil)' | |
16921 | @end smallexample | |
16922 | }: | |
16923 | ||
16924 | @smallexample | |
16925 | @group | |
16926 | (if (string-equal "21" (substring (emacs-version) 10 12)) | |
16927 | (progn | |
16928 | (blink-cursor-mode 0) | |
16929 | ;; Insert newline when you press `C-n' (next-line) | |
16930 | ;; at the end of the buffer | |
16931 | (setq next-line-add-newlines t) | |
16932 | @end group | |
16933 | @group | |
16934 | ;; Turn on image viewing | |
16935 | (auto-image-file-mode t) | |
16936 | @end group | |
16937 | @group | |
16938 | ;; Turn on menu bar (this bar has text) | |
16939 | ;; (Use numeric argument to turn on) | |
16940 | (menu-bar-mode 1) | |
16941 | @end group | |
16942 | @group | |
16943 | ;; Turn off tool bar (this bar has icons) | |
16944 | ;; (Use numeric argument to turn on) | |
16945 | (tool-bar-mode nil) | |
16946 | @end group | |
16947 | @group | |
16948 | ;; Turn off tooltip mode for tool bar | |
16949 | ;; (This mode causes icon explanations to pop up) | |
16950 | ;; (Use numeric argument to turn on) | |
16951 | (tooltip-mode nil) | |
16952 | ;; If tooltips turned on, make tips appear promptly | |
16953 | (setq tooltip-delay 0.1) ; default is one second | |
16954 | )) | |
16955 | @end group | |
16956 | @end smallexample | |
16957 | ||
16958 | @noindent | |
16959 | (You will note that instead of typing @code{(number-to-string 21)}, I | |
16960 | decided to save typing and wrote `21' as a string, @code{"21"}, rather | |
16961 | than convert it from an integer to a string. In this instance, this | |
16962 | expression is better than the longer, but more general | |
16963 | @code{(number-to-string 21)}. However, if you do not know ahead of | |
16964 | time what type of information will be returned, then the | |
16965 | @code{number-to-string} function will be needed.) | |
16966 | ||
16967 | @node X11 Colors, Miscellaneous, Simple Extension, Emacs Initialization | |
16968 | @section X11 Colors | |
16969 | ||
16970 | You can specify colors when you use Emacs with the MIT X Windowing | |
16971 | system. | |
16972 | ||
16973 | I dislike the default colors and specify my own. | |
16974 | ||
16975 | @need 1250 | |
16976 | Here are the expressions in my @file{.emacs} | |
16977 | file that set values: | |
16978 | ||
16979 | @smallexample | |
16980 | @group | |
16981 | ;; Set cursor color | |
16982 | (set-cursor-color "white") | |
16983 | ||
16984 | ;; Set mouse color | |
16985 | (set-mouse-color "white") | |
16986 | ||
16987 | ;; Set foreground and background | |
16988 | (set-foreground-color "white") | |
16989 | (set-background-color "darkblue") | |
16990 | @end group | |
16991 | ||
16992 | @group | |
16993 | ;;; Set highlighting colors for isearch and drag | |
16994 | (set-face-foreground 'highlight "white") | |
16995 | (set-face-background 'highlight "blue") | |
16996 | @end group | |
16997 | ||
16998 | @group | |
16999 | (set-face-foreground 'region "cyan") | |
17000 | (set-face-background 'region "blue") | |
17001 | @end group | |
17002 | ||
17003 | @group | |
17004 | (set-face-foreground 'secondary-selection "skyblue") | |
17005 | (set-face-background 'secondary-selection "darkblue") | |
17006 | @end group | |
17007 | ||
17008 | @group | |
17009 | ;; Set calendar highlighting colors | |
17010 | (setq calendar-load-hook | |
17011 | '(lambda () | |
17012 | (set-face-foreground 'diary-face "skyblue") | |
17013 | (set-face-background 'holiday-face "slate blue") | |
17014 | (set-face-foreground 'holiday-face "white"))) | |
17015 | @end group | |
17016 | @end smallexample | |
17017 | ||
17018 | The various shades of blue soothe my eye and prevent me from seeing | |
17019 | the screen flicker. | |
17020 | ||
17021 | Alternatively, I could have set my specifications in various X | |
17022 | initialization files. For example, I could set the foreground, | |
17023 | background, cursor, and pointer (i.e., mouse) colors in my | |
17024 | @file{~/.Xresources} file like this: | |
17025 | ||
17026 | @smallexample | |
17027 | @group | |
17028 | Emacs*foreground: white | |
17029 | Emacs*background: darkblue | |
17030 | Emacs*cursorColor: white | |
17031 | Emacs*pointerColor: white | |
17032 | @end group | |
17033 | @end smallexample | |
17034 | ||
17035 | In any event, since it is not part of Emacs, I set the root color of | |
17036 | my X window in my @file{~/.xinitrc} file, like this@footnote{I | |
17037 | occasionally run more modern window managers, such as Sawfish with | |
17038 | GNOME, Enlightenment, SCWM, or KDE; in those cases, I often specify an | |
17039 | image rather than a plain color.}: | |
17040 | ||
17041 | @smallexample | |
17042 | @group | |
17043 | # I use TWM for window manager. | |
17044 | xsetroot -solid Navy -fg white & | |
17045 | @end group | |
17046 | @end smallexample | |
17047 | ||
17048 | @node Miscellaneous, Mode Line, X11 Colors, Emacs Initialization | |
17049 | @section Miscellaneous Settings for a @file{.emacs} File | |
17050 | ||
17051 | Here are a few miscellaneous settings: | |
17052 | @sp 1 | |
17053 | ||
17054 | @itemize @minus | |
17055 | @item | |
17056 | Set the shape and color of the mouse cursor: | |
17057 | @smallexample | |
17058 | @group | |
17059 | ; Cursor shapes are defined in | |
17060 | ; `/usr/include/X11/cursorfont.h'; | |
17061 | ; for example, the `target' cursor is number 128; | |
17062 | ; the `top_left_arrow' cursor is number 132. | |
17063 | @end group | |
17064 | ||
17065 | @group | |
17066 | (let ((mpointer (x-get-resource "*mpointer" | |
17067 | "*emacs*mpointer"))) | |
17068 | ;; If you have not set your mouse pointer | |
17069 | ;; then set it, otherwise leave as is: | |
17070 | (if (eq mpointer nil) | |
17071 | (setq mpointer "132")) ; top_left_arrow | |
17072 | @end group | |
17073 | @group | |
17074 | (setq x-pointer-shape (string-to-int mpointer)) | |
17075 | (set-mouse-color "white")) | |
17076 | @end group | |
17077 | @end smallexample | |
17078 | @end itemize | |
17079 | ||
17080 | @node Mode Line, , Miscellaneous, Emacs Initialization | |
17081 | @section A Modified Mode Line | |
17082 | @vindex default-mode-line-format | |
17083 | @cindex Mode line format | |
17084 | ||
17085 | Finally, a feature I really like: a modified mode line. | |
17086 | ||
17087 | When I work over a network, I forget which machine I am using. Also, | |
17088 | I tend to I lose track of where I am, and which line point is on. | |
17089 | ||
17090 | So I reset my mode line to look like this: | |
17091 | ||
17092 | @smallexample | |
17093 | -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top | |
17094 | @end smallexample | |
17095 | ||
17096 | I am visiting a file called @file{foo.texi}, on my machine | |
17097 | @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in | |
17098 | Texinfo mode, and am at the top of the buffer. | |
17099 | ||
17100 | @need 1200 | |
17101 | My @file{.emacs} file has a section that looks like this: | |
17102 | ||
17103 | @smallexample | |
17104 | @group | |
17105 | ;; Set a Mode Line that tells me which machine, which directory, | |
17106 | ;; and which line I am on, plus the other customary information. | |
17107 | (setq default-mode-line-format | |
17108 | (quote | |
17109 | (#("-" 0 1 | |
17110 | (help-echo | |
17111 | "mouse-1: select window, mouse-2: delete others ...")) | |
17112 | mode-line-mule-info | |
17113 | mode-line-modified | |
17114 | mode-line-frame-identification | |
17115 | " " | |
17116 | @end group | |
17117 | @group | |
17118 | mode-line-buffer-identification | |
17119 | " " | |
17120 | (:eval (substring | |
17121 | (system-name) 0 (string-match "\\..+" (system-name)))) | |
17122 | ":" | |
17123 | default-directory | |
17124 | #(" " 0 1 | |
17125 | (help-echo | |
17126 | "mouse-1: select window, mouse-2: delete others ...")) | |
17127 | (line-number-mode " Line %l ") | |
17128 | global-mode-string | |
17129 | @end group | |
17130 | @group | |
17131 | #(" %[(" 0 6 | |
17132 | (help-echo | |
17133 | "mouse-1: select window, mouse-2: delete others ...")) | |
17134 | (:eval (mode-line-mode-name)) | |
17135 | mode-line-process | |
17136 | minor-mode-alist | |
17137 | #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...))) | |
17138 | ")%] " | |
17139 | (-3 . "%P") | |
17140 | ;; "-%-" | |
17141 | ))) | |
17142 | @end group | |
17143 | @end smallexample | |
17144 | ||
17145 | @noindent | |
17146 | Here, I redefine the default mode line. Most of the parts are from | |
17147 | the original; but I make a few changes. I set the @emph{default} mode | |
17148 | line format so as to permit various modes, such as Info, to override | |
17149 | it. | |
17150 | ||
17151 | Many elements in the list are self-explanatory: | |
17152 | @code{mode-line-modified} is a variable that tells whether the buffer | |
17153 | has been modified, @code{mode-name} tells the name of the mode, and so | |
17154 | on. However, the format looks complicated because of two features we | |
17155 | have not discussed. | |
17156 | ||
17157 | The first string in the mode line is a dash or hyphen, @samp{-}. In | |
17158 | the old days, it would have been specified simply as @code{"-"}. But | |
17159 | nowadays, Emacs can add properties to a string, such as highlighting | |
17160 | or, as in this case, a help feature. If you place your mouse cursor | |
17161 | over the hyphen, some help information appears (By default, you must | |
17162 | wait one second before the information appears. You can change that | |
17163 | timing by changing the value of @code{tooltip-delay}.) | |
17164 | ||
17165 | @need 1000 | |
17166 | The new string format has a special syntax: | |
17167 | ||
17168 | @smallexample | |
17169 | #("-" 0 1 (help-echo "mouse-1: select window, ...")) | |
17170 | @end smallexample | |
17171 | ||
17172 | @noindent | |
17173 | The @code{#(} begins a list. The first element of the list is the | |
17174 | string itself, just one @samp{-}. The second and third | |
17175 | elements specify the range over which the fourth element applies. A | |
17176 | range starts @emph{after} a character, so a zero means the range | |
17177 | starts just before the first character; a 1 means that the range ends | |
17178 | just after the first character. The third element is the property for | |
17179 | the range. It consists of a property list, a | |
17180 | property name, in this case, @samp{help-echo}, followed by a value, in this | |
17181 | case, a string. The second, third, and fourth elements of this new | |
17182 | string format can be repeated. | |
17183 | ||
17184 | @xref{Text Props and Strings, , Text Properties in String, elisp, The | |
17185 | GNU Emacs Lisp Reference Manual}, and see @ref{Mode Line Format, , Mode | |
17186 | Line Format, elisp, The GNU Emacs Lisp Reference Manual}, for more | |
17187 | information. | |
17188 | ||
17189 | @code{mode-line-buffer-identification} | |
17190 | displays the current buffer name. It is a list | |
17191 | beginning @code{(#("%12b" 0 4 @dots{}}. | |
17192 | The @code{#(} begins the list. | |
17193 | ||
17194 | The @samp{"%12b"} displays the current buffer name, using the | |
17195 | @code{buffer-name} function with which we are familiar; the `12' | |
17196 | specifies the maximum number of characters that will be displayed. | |
17197 | When a name has fewer characters, whitespace is added to fill out to | |
17198 | this number. (Buffer names can and often should be longer than 12 | |
17199 | characters; this length works well in a typical 80 column wide | |
17200 | window.) | |
17201 | ||
17202 | @code{:eval} is a new feature in GNU Emacs version 21. It says to | |
17203 | evaluate the following form and use the result as a string to display. | |
17204 | In this case, the expression displays the first component of the full | |
17205 | system name. The end of the first component is a @samp{.} (`period'), | |
17206 | so I use the @code{string-match} function to tell me the length of the | |
17207 | first component. The substring from the zeroth character to that | |
17208 | length is the name of the machine. | |
17209 | ||
17210 | @need 1250 | |
17211 | This is the expression: | |
17212 | ||
17213 | @smallexample | |
17214 | @group | |
17215 | (:eval (substring | |
17216 | (system-name) 0 (string-match "\\..+" (system-name)))) | |
17217 | @end group | |
17218 | @end smallexample | |
17219 | ||
17220 | @samp{%[} and @samp{%]} cause a pair of square brackets | |
17221 | to appear for each recursive editing level. @samp{%n} says `Narrow' | |
17222 | when narrowing is in effect. @samp{%P} tells you the percentage of | |
17223 | the buffer that is above the bottom of the window, or `Top', `Bottom', | |
17224 | or `All'. (A lower case @samp{p} tell you the percentage above the | |
17225 | @emph{top} of the window.) @samp{%-} inserts enough dashes to fill | |
17226 | out the line. | |
17227 | ||
17228 | Remember, ``You don't have to like Emacs to like it'' --- your own | |
17229 | Emacs can have different colors, different commands, and different | |
17230 | keys than a default Emacs. | |
17231 | ||
17232 | On the other hand, if you want to bring up a plain `out of the box' | |
17233 | Emacs, with no customization, type: | |
17234 | ||
17235 | @smallexample | |
17236 | emacs -q | |
17237 | @end smallexample | |
17238 | ||
17239 | @noindent | |
17240 | This will start an Emacs that does @emph{not} load your | |
17241 | @file{~/.emacs} initialization file. A plain, default Emacs. Nothing | |
17242 | more. | |
17243 | ||
17244 | @node Debugging, Conclusion, Emacs Initialization, Top | |
17245 | @chapter Debugging | |
17246 | @cindex debugging | |
17247 | ||
17248 | GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The | |
17249 | first is built into the internals of Emacs and is always with you; | |
17250 | the second requires that you instrument a function before you can use it. | |
17251 | ||
17252 | Both debuggers are described extensively in @ref{Debugging, , | |
17253 | Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}. | |
17254 | In this chapter, I will walk through a short example of each. | |
17255 | ||
17256 | @menu | |
17257 | * debug:: How to use the built-in debugger. | |
17258 | * debug-on-entry:: Start debugging when you call a function. | |
17259 | * debug-on-quit:: Start debugging when you quit with @kbd{C-g}. | |
17260 | * edebug:: How to use Edebug, a source level debugger. | |
17261 | * Debugging Exercises:: | |
17262 | @end menu | |
17263 | ||
17264 | @node debug, debug-on-entry, Debugging, Debugging | |
17265 | @section @code{debug} | |
17266 | @findex debug | |
17267 | ||
17268 | Suppose you have written a function definition that is intended to | |
17269 | return the sum of the numbers 1 through a given number. (This is the | |
17270 | @code{triangle} function discussed earlier. @xref{Decrementing | |
17271 | Example, , Example with Decrementing Counter}, for a discussion.) | |
17272 | @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.) | |
17273 | ||
17274 | However, your function definition has a bug. You have mistyped | |
17275 | @samp{1=} for @samp{1-}. Here is the broken definition: | |
17276 | ||
17277 | @findex triangle-bugged | |
17278 | @smallexample | |
17279 | @group | |
17280 | (defun triangle-bugged (number) | |
17281 | "Return sum of numbers 1 through NUMBER inclusive." | |
17282 | (let ((total 0)) | |
17283 | (while (> number 0) | |
17284 | (setq total (+ total number)) | |
17285 | (setq number (1= number))) ; @r{Error here.} | |
17286 | total)) | |
17287 | @end group | |
17288 | @end smallexample | |
17289 | ||
17290 | If you are reading this in Info, you can evaluate this definition in | |
17291 | the normal fashion. You will see @code{triangle-bugged} appear in the | |
17292 | echo area. | |
17293 | ||
17294 | @need 1250 | |
17295 | Now evaluate the @code{triangle-bugged} function with an | |
17296 | argument of 4: | |
17297 | ||
17298 | @smallexample | |
17299 | (triangle-bugged 4) | |
17300 | @end smallexample | |
17301 | ||
17302 | @noindent | |
17303 | In GNU Emacs version 21, you will create and enter a | |
17304 | @file{*Backtrace*} buffer that says: | |
17305 | ||
17306 | @noindent | |
17307 | @smallexample | |
17308 | @group | |
17309 | ---------- Buffer: *Backtrace* ---------- | |
17310 | Debugger entered--Lisp error: (void-function 1=) | |
17311 | (1= number) | |
17312 | (setq number (1= number)) | |
17313 | (while (> number 0) (setq total (+ total number)) | |
17314 | (setq number (1= number))) | |
17315 | (let ((total 0)) (while (> number 0) (setq total ...) | |
17316 | (setq number ...)) total) | |
17317 | triangle-bugged(4) | |
17318 | @end group | |
17319 | @group | |
17320 | eval((triangle-bugged 4)) | |
17321 | eval-last-sexp-1(nil) | |
17322 | eval-last-sexp(nil) | |
17323 | call-interactively(eval-last-sexp) | |
17324 | ---------- Buffer: *Backtrace* ---------- | |
17325 | @end group | |
17326 | @end smallexample | |
17327 | ||
17328 | @noindent | |
17329 | (I have reformatted this example slightly; the debugger does not fold | |
17330 | long lines. As usual, you can quit the debugger by typing @kbd{q} in | |
17331 | the @file{*Backtrace*} buffer.) | |
17332 | ||
17333 | In practice, for a bug as simple as this, the `Lisp error' line will | |
17334 | tell you what you need to know to correct the definition. The | |
17335 | function @code{1=} is `void'. | |
17336 | ||
17337 | @need 800 | |
17338 | In GNU Emacs 20 and before, you will see: | |
17339 | ||
17340 | @smallexample | |
17341 | Symbol's function definition is void:@: 1= | |
17342 | @end smallexample | |
17343 | ||
17344 | @noindent | |
17345 | which has the same meaning as the @file{*Backtrace*} buffer line in | |
17346 | version 21. | |
17347 | ||
17348 | However, suppose you are not quite certain what is going on? | |
17349 | You can read the complete backtrace. | |
17350 | ||
17351 | In this case, you need to run GNU Emacs 21, which automatically starts | |
17352 | the debugger that puts you in the @file{*Backtrace*} buffer; or else, | |
17353 | you need to start the debugger manually as described below. | |
17354 | ||
17355 | Read the @file{*Backtrace*} buffer from the bottom up; it tells you | |
17356 | what Emacs did that led to the error. Emacs made an interactive call | |
17357 | to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation | |
17358 | of the @code{triangle-bugged} expression. Each line above tells you | |
17359 | what the Lisp interpreter evaluated next. | |
17360 | ||
17361 | @need 1250 | |
17362 | The third line from the top of the buffer is | |
17363 | ||
17364 | @smallexample | |
17365 | (setq number (1= number)) | |
17366 | @end smallexample | |
17367 | ||
17368 | @noindent | |
17369 | Emacs tried to evaluate this expression; in order to do so, it tried | |
17370 | to evaluate the inner expression shown on the second line from the | |
17371 | top: | |
17372 | ||
17373 | @smallexample | |
17374 | (1= number) | |
17375 | @end smallexample | |
17376 | ||
17377 | @need 1250 | |
17378 | @noindent | |
17379 | This is where the error occurred; as the top line says: | |
17380 | ||
17381 | @smallexample | |
17382 | Debugger entered--Lisp error: (void-function 1=) | |
17383 | @end smallexample | |
17384 | ||
17385 | @noindent | |
17386 | You can correct the mistake, re-evaluate the function definition, and | |
17387 | then run your test again. | |
17388 | ||
17389 | @node debug-on-entry, debug-on-quit, debug, Debugging | |
17390 | @section @code{debug-on-entry} | |
17391 | @findex debug-on-entry | |
17392 | ||
17393 | GNU Emacs 21 starts the debugger automatically when your function has | |
17394 | an error. GNU Emacs version 20 and before did not; it simply | |
17395 | presented you with an error message. You had to start the debugger | |
17396 | manually. | |
17397 | ||
17398 | You can start the debugger manually for all versions of Emacs; the | |
17399 | advantage is that the debugger runs even if you do not have a bug in | |
17400 | your code. Sometimes your code will be free of bugs! | |
17401 | ||
17402 | You can enter the debugger when you call the function by calling | |
17403 | @code{debug-on-entry}. | |
17404 | ||
17405 | @need 1250 | |
17406 | @noindent | |
17407 | Type: | |
17408 | ||
17409 | @smallexample | |
17410 | M-x debug-on-entry RET triangle-bugged RET | |
17411 | @end smallexample | |
17412 | ||
17413 | @need 1250 | |
17414 | @noindent | |
17415 | Now, evaluate the following: | |
17416 | ||
17417 | @smallexample | |
17418 | (triangle-bugged 5) | |
17419 | @end smallexample | |
17420 | ||
17421 | @noindent | |
17422 | All versions of Emacs will create a @file{*Backtrace*} buffer and tell | |
17423 | you that it is beginning to evaluate the @code{triangle-bugged} | |
17424 | function: | |
17425 | ||
17426 | @smallexample | |
17427 | @group | |
17428 | ---------- Buffer: *Backtrace* ---------- | |
17429 | Debugger entered--entering a function: | |
17430 | * triangle-bugged(5) | |
17431 | eval((triangle-bugged 5)) | |
17432 | @end group | |
17433 | @group | |
17434 | eval-last-sexp-1(nil) | |
17435 | eval-last-sexp(nil) | |
17436 | call-interactively(eval-last-sexp) | |
17437 | ---------- Buffer: *Backtrace* ---------- | |
17438 | @end group | |
17439 | @end smallexample | |
17440 | ||
17441 | In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate | |
17442 | the first expression in @code{triangle-bugged}; the buffer will look | |
17443 | like this: | |
17444 | ||
17445 | @smallexample | |
17446 | @group | |
17447 | ---------- Buffer: *Backtrace* ---------- | |
17448 | Debugger entered--beginning evaluation of function call form: | |
17449 | * (let ((total 0)) (while (> number 0) (setq total ...) | |
17450 | (setq number ...)) total) | |
17451 | * triangle-bugged(5) | |
17452 | eval((triangle-bugged 5)) | |
17453 | @end group | |
17454 | @group | |
17455 | eval-last-sexp-1(nil) | |
17456 | eval-last-sexp(nil) | |
17457 | call-interactively(eval-last-sexp) | |
17458 | ---------- Buffer: *Backtrace* ---------- | |
17459 | @end group | |
17460 | @end smallexample | |
17461 | ||
17462 | @noindent | |
17463 | Now, type @kbd{d} again, eight times, slowly. Each time you type | |
17464 | @kbd{d}, Emacs will evaluate another expression in the function | |
17465 | definition. | |
17466 | ||
17467 | @need 1750 | |
17468 | Eventually, the buffer will look like this: | |
17469 | ||
17470 | @smallexample | |
17471 | @group | |
17472 | ---------- Buffer: *Backtrace* ---------- | |
17473 | Debugger entered--beginning evaluation of function call form: | |
17474 | * (setq number (1= number)) | |
17475 | * (while (> number 0) (setq total (+ total number)) | |
17476 | (setq number (1= number))) | |
17477 | @group | |
17478 | @end group | |
17479 | * (let ((total 0)) (while (> number 0) (setq total ...) | |
17480 | (setq number ...)) total) | |
17481 | * triangle-bugged(5) | |
17482 | eval((triangle-bugged 5)) | |
17483 | @group | |
17484 | @end group | |
17485 | eval-last-sexp-1(nil) | |
17486 | eval-last-sexp(nil) | |
17487 | call-interactively(eval-last-sexp) | |
17488 | ---------- Buffer: *Backtrace* ---------- | |
17489 | @end group | |
17490 | @end smallexample | |
17491 | ||
17492 | @noindent | |
17493 | Finally, after you type @kbd{d} two more times, Emacs will reach the | |
17494 | error, and the top two lines of the @file{*Backtrace*} buffer will look | |
17495 | like this: | |
17496 | ||
17497 | @smallexample | |
17498 | @group | |
17499 | ---------- Buffer: *Backtrace* ---------- | |
17500 | Debugger entered--Lisp error: (void-function 1=) | |
17501 | * (1= number) | |
17502 | @dots{} | |
17503 | ---------- Buffer: *Backtrace* ---------- | |
17504 | @end group | |
17505 | @end smallexample | |
17506 | ||
17507 | By typing @kbd{d}, you were able to step through the function. | |
17508 | ||
17509 | You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this | |
17510 | quits the trace, but does not cancel @code{debug-on-entry}. | |
17511 | ||
17512 | @findex cancel-debug-on-entry | |
17513 | To cancel the effect of @code{debug-on-entry}, call | |
17514 | @code{cancel-debug-on-entry} and the name of the function, like this: | |
17515 | ||
17516 | @smallexample | |
17517 | M-x cancel-debug-on-entry RET triangle-bugged RET | |
17518 | @end smallexample | |
17519 | ||
17520 | @noindent | |
17521 | (If you are reading this in Info, cancel @code{debug-on-entry} now.) | |
17522 | ||
17523 | @node debug-on-quit, edebug, debug-on-entry, Debugging | |
17524 | @section @code{debug-on-quit} and @code{(debug)} | |
17525 | ||
17526 | In addition to setting @code{debug-on-error} or calling @code{debug-on-entry}, | |
17527 | there are two other ways to start @code{debug}. | |
17528 | ||
17529 | @findex debug-on-quit | |
17530 | You can start @code{debug} whenever you type @kbd{C-g} | |
17531 | (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to | |
17532 | @code{t}. This is useful for debugging infinite loops. | |
17533 | ||
17534 | @need 1500 | |
17535 | @cindex @code{(debug)} in code | |
17536 | Or, you can insert a line that says @code{(debug)} into your code | |
17537 | where you want the debugger to start, like this: | |
17538 | ||
17539 | @smallexample | |
17540 | @group | |
17541 | (defun triangle-bugged (number) | |
17542 | "Return sum of numbers 1 through NUMBER inclusive." | |
17543 | (let ((total 0)) | |
17544 | (while (> number 0) | |
17545 | (setq total (+ total number)) | |
17546 | (debug) ; @r{Start debugger.} | |
17547 | (setq number (1= number))) ; @r{Error here.} | |
17548 | total)) | |
17549 | @end group | |
17550 | @end smallexample | |
17551 | ||
17552 | The @code{debug} function is described in detail in @ref{Debugger, , | |
17553 | The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}. | |
17554 | ||
17555 | @node edebug, Debugging Exercises, debug-on-quit, Debugging | |
17556 | @section The @code{edebug} Source Level Debugger | |
17557 | @cindex Source level debugger | |
17558 | @findex edebug | |
17559 | ||
17560 | Edebug is a source level debugger. Edebug normally displays the | |
17561 | source of the code you are debugging, with an arrow at the left that | |
17562 | shows which line you are currently executing. | |
17563 | ||
17564 | You can walk through the execution of a function, line by line, or run | |
17565 | quickly until reaching a @dfn{breakpoint} where execution stops. | |
17566 | ||
17567 | Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs | |
17568 | Lisp Reference Manual}. | |
17569 | ||
17570 | Here is a bugged function definition for @code{triangle-recursively}. | |
17571 | @xref{Recursive triangle function, , Recursion in place of a counter}, | |
17572 | for a review of it. | |
17573 | ||
17574 | @smallexample | |
17575 | @group | |
17576 | (defun triangle-recursively-bugged (number) | |
17577 | "Return sum of numbers 1 through NUMBER inclusive. | |
17578 | Uses recursion." | |
17579 | (if (= number 1) | |
17580 | 1 | |
17581 | (+ number | |
17582 | (triangle-recursively-bugged | |
17583 | (1= number))))) ; @r{Error here.} | |
17584 | @end group | |
17585 | @end smallexample | |
17586 | ||
17587 | @noindent | |
17588 | Normally, you would install this definition by positioning your cursor | |
17589 | after the function's closing parenthesis and typing @kbd{C-x C-e} | |
17590 | (@code{eval-last-sexp}) or else by positioning your cursor within the | |
17591 | definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default, | |
17592 | the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp | |
17593 | Interactive mode.) | |
17594 | ||
17595 | @need 1500 | |
17596 | However, to prepare this function definition for Edebug, you must | |
17597 | first @dfn{instrument} the code using a different command. You can do | |
17598 | this by positioning your cursor within the definition and typing | |
17599 | ||
17600 | @smallexample | |
17601 | M-x edebug-defun RET | |
17602 | @end smallexample | |
17603 | ||
17604 | @noindent | |
17605 | This will cause Emacs to load Edebug automatically if it is not | |
17606 | already loaded, and properly instrument the function. | |
17607 | ||
17608 | After instrumenting the function, place your cursor after the | |
17609 | following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}): | |
17610 | ||
17611 | @smallexample | |
17612 | (triangle-recursively-bugged 3) | |
17613 | @end smallexample | |
17614 | ||
17615 | @noindent | |
17616 | You will be jumped back to the source for | |
17617 | @code{triangle-recursively-bugged} and the cursor positioned at the | |
17618 | beginning of the @code{if} line of the function. Also, you will see | |
17619 | an arrowhead at the left hand side of that line. The arrowhead marks | |
17620 | the line where the function is executing. (In the following examples, | |
17621 | we show the arrowhead with @samp{=>}; in a windowing system, you may | |
17622 | see the arrowhead as a solid triangle in the window `fringe'.) | |
17623 | ||
17624 | @smallexample | |
17625 | =>@point{}(if (= number 1) | |
17626 | @end smallexample | |
17627 | ||
17628 | @noindent | |
17629 | @iftex | |
17630 | In the example, the location of point is displayed with a star, | |
17631 | @samp{@point{}} (in Info, it is displayed as @samp{-!-}). | |
17632 | @end iftex | |
17633 | @ifnottex | |
17634 | In the example, the location of point is displayed as @samp{@point{}} | |
17635 | (in a printed book, it is displayed with a five pointed star). | |
17636 | @end ifnottex | |
17637 | ||
17638 | If you now press @key{SPC}, point will move to the next expression to | |
17639 | be executed; the line will look like this: | |
17640 | ||
17641 | @smallexample | |
17642 | =>(if @point{}(= number 1) | |
17643 | @end smallexample | |
17644 | ||
17645 | @noindent | |
17646 | As you continue to press @key{SPC}, point will move from expression to | |
17647 | expression. At the same time, whenever an expression returns a value, | |
17648 | that value will be displayed in the echo area. For example, after you | |
17649 | move point past @code{number}, you will see the following: | |
17650 | ||
17651 | @smallexample | |
17652 | Result: 3 = C-c | |
17653 | @end smallexample | |
17654 | ||
17655 | @noindent | |
17656 | This means the value of @code{number} is 3, which is @sc{ascii} | |
17657 | `control-c' (the third letter of the alphabet). | |
17658 | ||
17659 | You can continue moving through the code until you reach the line with | |
17660 | the error. Before evaluation, that line looks like this: | |
17661 | ||
17662 | @smallexample | |
17663 | => @point{}(1= number))))) ; @r{Error here.} | |
17664 | @end smallexample | |
17665 | ||
17666 | @need 1250 | |
17667 | @noindent | |
17668 | When you press @key{SPC} once again, you will produce an error message | |
17669 | that says: | |
17670 | ||
17671 | @smallexample | |
17672 | Symbol's function definition is void:@: 1= | |
17673 | @end smallexample | |
17674 | ||
17675 | @noindent | |
17676 | This is the bug. | |
17677 | ||
17678 | Press @kbd{q} to quit Edebug. | |
17679 | ||
17680 | To remove instrumentation from a function definition, simply | |
17681 | re-evaluate it with a command that does not instrument it. | |
17682 | For example, you could place your cursor after the definition's | |
17683 | closing parenthesis and type @kbd{C-x C-e}. | |
17684 | ||
17685 | Edebug does a great deal more than walk with you through a function. | |
17686 | You can set it so it races through on its own, stopping only at an | |
17687 | error or at specified stopping points; you can cause it to display the | |
17688 | changing values of various expressions; you can find out how many | |
17689 | times a function is called, and more. | |
17690 | ||
17691 | Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs | |
17692 | Lisp Reference Manual}. | |
17693 | ||
17694 | @need 1500 | |
17695 | @node Debugging Exercises, , edebug, Debugging | |
17696 | @section Debugging Exercises | |
17697 | ||
17698 | @itemize @bullet | |
17699 | @item | |
17700 | Install the @code{count-words-region} function and then cause it to | |
17701 | enter the built-in debugger when you call it. Run the command on a | |
17702 | region containing two words. You will need to press @kbd{d} a | |
17703 | remarkable number of times. On your system, is a `hook' called after | |
17704 | the command finishes? (For information on hooks, see @ref{Command | |
17705 | Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference | |
17706 | Manual}.) | |
17707 | ||
17708 | @item | |
17709 | Copy @code{count-words-region} into the @file{*scratch*} buffer, | |
17710 | instrument the function for Edebug, and walk through its execution. | |
17711 | The function does not need to have a bug, although you can introduce | |
17712 | one if you wish. If the function lacks a bug, the walk-through | |
17713 | completes without problems. | |
17714 | ||
17715 | @item | |
17716 | While running Edebug, type @kbd{?} to see a list of all the Edebug commands. | |
17717 | (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.@: | |
17718 | @kbd{@key{CTL}-x} followed by an upper case @kbd{X}; use this prefix | |
17719 | for commands made outside of the Edebug debugging buffer.) | |
17720 | ||
17721 | @item | |
17722 | In the Edebug debugging buffer, use the @kbd{p} | |
17723 | (@code{edebug-bounce-point}) command to see where in the region the | |
17724 | @code{count-words-region} is working. | |
17725 | ||
17726 | @item | |
17727 | Move point to some spot further down function and then type the | |
17728 | @kbd{h} (@code{edebug-goto-here}) command to jump to that location. | |
17729 | ||
17730 | @item | |
17731 | Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to | |
17732 | walk through the function on its own; use an upper case @kbd{T} for | |
17733 | @code{edebug-Trace-fast-mode}. | |
17734 | ||
17735 | @item | |
17736 | Set a breakpoint, then run Edebug in Trace mode until it reaches the | |
17737 | stopping point. | |
17738 | @end itemize | |
17739 | ||
17740 | @node Conclusion, the-the, Debugging, Top | |
17741 | @chapter Conclusion | |
17742 | ||
17743 | We have now reached the end of this Introduction. You have now | |
17744 | learned enough about programming in Emacs Lisp to set values, to write | |
17745 | simple @file{.emacs} files for yourself and your friends, and write | |
17746 | simple customizations and extensions to Emacs. | |
17747 | ||
17748 | This is a place to stop. Or, if you wish, you can now go onward, and | |
17749 | teach yourself. | |
17750 | ||
17751 | You have learned some of the basic nuts and bolts of programming. But | |
17752 | only some. There are a great many more brackets and hinges that are | |
17753 | easy to use that we have not touched. | |
17754 | ||
17755 | A path you can follow right now lies among the sources to GNU Emacs | |
17756 | and in | |
17757 | @ifnotinfo | |
17758 | @cite{The GNU Emacs Lisp Reference Manual}. | |
17759 | @end ifnotinfo | |
17760 | @ifinfo | |
17761 | @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU | |
17762 | Emacs Lisp Reference Manual}. | |
17763 | @end ifinfo | |
17764 | ||
17765 | The Emacs Lisp sources are an adventure. When you read the sources and | |
17766 | come across a function or expression that is unfamiliar, you need to | |
17767 | figure out or find out what it does. | |
17768 | ||
17769 | Go to the Reference Manual. It is a thorough, complete, and fairly | |
17770 | easy-to-read description of Emacs Lisp. It is written not only for | |
17771 | experts, but for people who know what you know. (The @cite{Reference | |
17772 | Manual} comes with the standard GNU Emacs distribution. Like this | |
17773 | introduction, it comes as a Texinfo source file, so you can read it | |
17774 | on-line and as a typeset, printed book.) | |
17775 | ||
17776 | Go to the other on-line help that is part of GNU Emacs: the on-line | |
17777 | documentation for all functions, and @code{find-tags}, the program | |
17778 | that takes you to sources. | |
17779 | ||
17780 | Here is an example of how I explore the sources. Because of its name, | |
17781 | @file{simple.el} is the file I looked at first, a long time ago. As | |
17782 | it happens some of the functions in @file{simple.el} are complicated, | |
17783 | or at least look complicated at first sight. The @code{open-line} | |
17784 | function, for example, looks complicated. | |
17785 | ||
17786 | You may want to walk through this function slowly, as we did with the | |
17787 | @code{forward-sentence} function. | |
17788 | @ifnottex | |
17789 | (@xref{forward-sentence}.) | |
17790 | @end ifnottex | |
17791 | @iftex | |
17792 | (@xref{forward-sentence, , @code{forward-sentence}}.) | |
17793 | @end iftex | |
17794 | Or you may want to skip that function and look at another, such as | |
17795 | @code{split-line}. You don't need to read all the functions. | |
17796 | According to @code{count-words-in-defun}, the @code{split-line} | |
17797 | function contains 27 words and symbols. | |
17798 | ||
17799 | Even though it is short, @code{split-line} contains four expressions | |
17800 | we have not studied: @code{skip-chars-forward}, @code{indent-to}, | |
17801 | @code{current-column} and @samp{?\n}. | |
17802 | ||
17803 | Consider the @code{skip-chars-forward} function. (It is part of the | |
17804 | function definition for @code{back-to-indentation}, which is shown in | |
17805 | @ref{Review, , Review}.) | |
17806 | ||
17807 | In GNU Emacs, you can find out more about @code{skip-chars-forward} by | |
17808 | typing @kbd{C-h f} (@code{describe-function}) and the name of the | |
17809 | function. This gives you the function documentation. | |
17810 | ||
17811 | You may be able to guess what is done by a well named function such as | |
17812 | @code{indent-to}; or you can look it up, too. Incidentally, the | |
17813 | @code{describe-function} function itself is in @file{help.el}; it is | |
17814 | one of those long, but decipherable functions. You can look up | |
17815 | @code{describe-function} using the @kbd{C-h f} command! | |
17816 | ||
17817 | In this instance, since the code is Lisp, the @file{*Help*} buffer | |
17818 | contains the name of the library containing the function's source. | |
17819 | You can put point over the name of the library and press the RET key, | |
17820 | which in this situation is bound to @code{help-follow}, and be taken | |
17821 | directly to the source, in the same way as @kbd{M-.} | |
17822 | (@code{find-tag}). | |
17823 | ||
17824 | The definition for @code{describe-function} illustrates how to | |
17825 | customize the @code{interactive} expression without using the standard | |
17826 | character codes; and it shows how to create a temporary buffer. | |
17827 | ||
17828 | (The @code{indent-to} function is written in C rather than Emacs Lisp; | |
17829 | it is a `built-in' function. @code{help-follow} only provides you | |
17830 | with the documentation of a built-in function; it does not take you to | |
17831 | the source. But @code{find-tag} will take you to the source, if | |
17832 | properly set up.) | |
17833 | ||
17834 | You can look at a function's source using @code{find-tag}, which is | |
17835 | bound to @kbd{M-.} Finally, you can find out what the Reference | |
17836 | Manual has to say by visiting the manual in Info, and typing @kbd{i} | |
17837 | (@code{Info-index}) and the name of the function, or by looking up | |
17838 | @code{skip-chars-forward} in the index to a printed copy of the | |
17839 | manual. | |
17840 | ||
17841 | Similarly, you can find out what is meant by @samp{?\n}. You can try | |
17842 | using @code{Info-index} with @samp{?\n}. It turns out that this | |
17843 | action won't help; but don't give up. If you search the index for | |
17844 | @samp{\n} without the @samp{?}, you will be taken directly to the | |
17845 | relevant section of the manual. (@xref{Character Type, , Character | |
17846 | Type, elisp, The GNU Emacs Lisp Reference Manual}. @samp{?\n} stands | |
17847 | for the newline character.) | |
17848 | ||
17849 | Other interesting source files include @file{paragraphs.el}, | |
17850 | @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el} | |
17851 | file includes short, easily understood functions as well as longer | |
17852 | ones. The @file{loaddefs.el} file contains the many standard | |
17853 | autoloads and many keymaps. I have never looked at it all; only at | |
17854 | parts. @file{loadup.el} is the file that loads the standard parts of | |
17855 | Emacs; it tells you a great deal about how Emacs is built. | |
17856 | (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp | |
17857 | Reference Manual}, for more about building.) | |
17858 | ||
17859 | As I said, you have learned some nuts and bolts; however, and very | |
17860 | importantly, we have hardly touched major aspects of programming; I | |
17861 | have said nothing about how to sort information, except to use the | |
17862 | predefined @code{sort} function; I have said nothing about how to store | |
17863 | information, except to use variables and lists; I have said nothing | |
17864 | about how to write programs that write programs. These are topics for | |
17865 | another, and different kind of book, a different kind of learning. | |
17866 | ||
17867 | What you have done is learn enough for much practical work with GNU | |
17868 | Emacs. What you have done is get started. This is the end of a | |
17869 | beginning. | |
17870 | ||
17871 | @c ================ Appendix ================ | |
17872 | ||
17873 | @node the-the, Kill Ring, Conclusion, Top | |
17874 | @appendix The @code{the-the} Function | |
17875 | @findex the-the | |
17876 | @cindex Duplicated words function | |
17877 | @cindex Words, duplicated | |
17878 | ||
17879 | Sometimes when you you write text, you duplicate words---as with ``you | |
17880 | you'' near the beginning of this sentence. I find that most | |
17881 | frequently, I duplicate ``the'; hence, I call the function for | |
17882 | detecting duplicated words, @code{the-the}. | |
17883 | ||
17884 | @need 1250 | |
17885 | As a first step, you could use the following regular expression to | |
17886 | search for duplicates: | |
17887 | ||
17888 | @smallexample | |
17889 | \\(\\w+[ \t\n]+\\)\\1 | |
17890 | @end smallexample | |
17891 | ||
17892 | @noindent | |
17893 | This regexp matches one or more word-constituent characters followed | |
17894 | by one or more spaces, tabs, or newlines. However, it does not detect | |
17895 | duplicated words on different lines, since the ending of the first | |
17896 | word, the end of the line, is different from the ending of the second | |
17897 | word, a space. (For more information about regular expressions, see | |
17898 | @ref{Regexp Search, , Regular Expression Searches}, as well as | |
17899 | @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs | |
17900 | Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp, | |
17901 | The GNU Emacs Lisp Reference Manual}.) | |
17902 | ||
17903 | You might try searching just for duplicated word-constituent | |
17904 | characters but that does not work since the pattern detects doubles | |
17905 | such as the two occurrences of `th' in `with the'. | |
17906 | ||
17907 | Another possible regexp searches for word-constituent characters | |
17908 | followed by non-word-constituent characters, reduplicated. Here, | |
17909 | @w{@samp{\\w+}} matches one or more word-constituent characters and | |
17910 | @w{@samp{\\W*}} matches zero or more non-word-constituent characters. | |
17911 | ||
17912 | @smallexample | |
17913 | \\(\\(\\w+\\)\\W*\\)\\1 | |
17914 | @end smallexample | |
17915 | ||
17916 | @noindent | |
17917 | Again, not useful. | |
17918 | ||
17919 | Here is the pattern that I use. It is not perfect, but good enough. | |
17920 | @w{@samp{\\b}} matches the empty string, provided it is at the beginning | |
17921 | or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of | |
17922 | any characters that are @emph{not} an @@-sign, space, newline, or tab. | |
17923 | ||
17924 | @smallexample | |
17925 | \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b | |
17926 | @end smallexample | |
17927 | ||
17928 | One can write more complicated expressions, but I found that this | |
17929 | expression is good enough, so I use it. | |
17930 | ||
17931 | Here is the @code{the-the} function, as I include it in my | |
17932 | @file{.emacs} file, along with a handy global key binding: | |
17933 | ||
17934 | @smallexample | |
17935 | @group | |
17936 | (defun the-the () | |
17937 | "Search forward for for a duplicated word." | |
17938 | (interactive) | |
17939 | (message "Searching for for duplicated words ...") | |
17940 | (push-mark) | |
17941 | @end group | |
17942 | @group | |
17943 | ;; This regexp is not perfect | |
17944 | ;; but is fairly good over all: | |
17945 | (if (re-search-forward | |
17946 | "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move) | |
17947 | (message "Found duplicated word.") | |
17948 | (message "End of buffer"))) | |
17949 | @end group | |
17950 | ||
17951 | @group | |
17952 | ;; Bind `the-the' to C-c \ | |
17953 | (global-set-key "\C-c\\" 'the-the) | |
17954 | @end group | |
17955 | @end smallexample | |
17956 | ||
17957 | @sp 1 | |
17958 | Here is test text: | |
17959 | ||
17960 | @smallexample | |
17961 | @group | |
17962 | one two two three four five | |
17963 | five six seven | |
17964 | @end group | |
17965 | @end smallexample | |
17966 | ||
17967 | You can substitute the other regular expressions shown above in the | |
17968 | function definition and try each of them on this list. | |
17969 | ||
17970 | @node Kill Ring, Full Graph, the-the, Top | |
17971 | @appendix Handling the Kill Ring | |
17972 | @cindex Kill ring handling | |
17973 | @cindex Handling the kill ring | |
17974 | @cindex Ring, making a list like a | |
17975 | ||
17976 | The kill ring is a list that is transformed into a ring by the | |
17977 | workings of the @code{rotate-yank-pointer} function. The @code{yank} | |
17978 | and @code{yank-pop} commands use the @code{rotate-yank-pointer} | |
17979 | function. This appendix describes the @code{rotate-yank-pointer} | |
17980 | function as well as both the @code{yank} and the @code{yank-pop} | |
17981 | commands. | |
17982 | ||
17983 | @menu | |
17984 | * rotate-yank-pointer:: Move a pointer along a list and around. | |
17985 | * yank:: Paste a copy of a clipped element. | |
17986 | * yank-pop:: Insert first element pointed to. | |
17987 | @end menu | |
17988 | ||
17989 | @node rotate-yank-pointer, yank, Kill Ring, Kill Ring | |
17990 | @comment node-name, next, previous, up | |
17991 | @appendixsec The @code{rotate-yank-pointer} Function | |
17992 | @findex rotate-yank-pointer | |
17993 | ||
17994 | The @code{rotate-yank-pointer} function changes the element in the kill | |
17995 | ring to which @code{kill-ring-yank-pointer} points. For example, it can | |
17996 | change @code{kill-ring-yank-pointer} from pointing to the second | |
17997 | element to point to the third element. | |
17998 | ||
17999 | @need 800 | |
18000 | Here is the code for @code{rotate-yank-pointer}: | |
18001 | ||
18002 | @smallexample | |
18003 | @group | |
18004 | (defun rotate-yank-pointer (arg) | |
18005 | "Rotate the yanking point in the kill ring." | |
18006 | (interactive "p") | |
18007 | (let ((length (length kill-ring))) | |
18008 | @end group | |
18009 | @group | |
18010 | (if (zerop length) | |
18011 | ;; @r{then-part} | |
18012 | (error "Kill ring is empty") | |
18013 | @end group | |
18014 | @group | |
18015 | ;; @r{else-part} | |
18016 | (setq kill-ring-yank-pointer | |
18017 | (nthcdr (% (+ arg | |
18018 | (- length | |
18019 | (length | |
18020 | kill-ring-yank-pointer))) | |
18021 | length) | |
18022 | kill-ring))))) | |
18023 | @end group | |
18024 | @end smallexample | |
18025 | ||
18026 | @menu | |
18027 | * Understanding rotate-yk-ptr:: | |
18028 | * rotate-yk-ptr body:: The body of @code{rotate-yank-pointer}. | |
18029 | @end menu | |
18030 | ||
18031 | @node Understanding rotate-yk-ptr, rotate-yk-ptr body, rotate-yank-pointer, rotate-yank-pointer | |
18032 | @ifnottex | |
18033 | @unnumberedsubsec @code{rotate-yank-pointer} in Outline | |
18034 | @end ifnottex | |
18035 | ||
18036 | The @code{rotate-yank-pointer} function looks complex, but as usual, | |
18037 | it can be understood by taking it apart piece by piece. First look at | |
18038 | it in skeletal form: | |
18039 | ||
18040 | @smallexample | |
18041 | @group | |
18042 | (defun rotate-yank-pointer (arg) | |
18043 | "Rotate the yanking point in the kill ring." | |
18044 | (interactive "p") | |
18045 | (let @var{varlist} | |
18046 | @var{body}@dots{}) | |
18047 | @end group | |
18048 | @end smallexample | |
18049 | ||
18050 | This function takes one argument, called @code{arg}. It has a brief | |
18051 | documentation string; and it is interactive with a small @samp{p}, which | |
18052 | means that the argument must be a processed prefix passed to the | |
18053 | function as a number. | |
18054 | ||
18055 | The body of the function definition is a @code{let} expression, which | |
18056 | itself has a body as well as a @var{varlist}. | |
18057 | ||
18058 | The @code{let} expression declares a variable that will be only usable | |
18059 | within the bounds of this function. This variable is called | |
18060 | @code{length} and is bound to a value that is equal to the number of | |
18061 | items in the kill ring. This is done by using the function called | |
18062 | @code{length}. (Note that this function has the same name as the | |
18063 | variable called @code{length}; but one use of the word is to name the | |
18064 | function and the other is to name the variable. The two are quite | |
18065 | distinct. Similarly, an English speaker will distinguish between the | |
18066 | meanings of the word @samp{ship} when he says: "I must ship this package | |
18067 | immediately." and "I must get aboard the ship immediately.") | |
18068 | ||
18069 | The function @code{length} tells the number of items there are in a list, | |
18070 | so @code{(length kill-ring)} returns the number of items there are in the | |
18071 | kill ring. | |
18072 | ||
18073 | @node rotate-yk-ptr body, , Understanding rotate-yk-ptr, rotate-yank-pointer | |
18074 | @comment node-name, next, previous, up | |
18075 | @appendixsubsec The Body of @code{rotate-yank-pointer} | |
18076 | ||
18077 | The body of @code{rotate-yank-pointer} is a @code{let} expression and | |
18078 | the body of the @code{let} expression is an @code{if} expression. | |
18079 | ||
18080 | The purpose of the @code{if} expression is to find out whether there is | |
18081 | anything in the kill ring. If the kill ring is empty, the @code{error} | |
18082 | function stops evaluation of the function and prints a message in the | |
18083 | echo area. On the other hand, if the kill ring has something in it, the | |
18084 | work of the function is done. | |
18085 | ||
18086 | Here is the if-part and then-part of the @code{if} expression: | |
18087 | ||
18088 | @findex zerop | |
18089 | @findex error | |
18090 | @smallexample | |
18091 | @group | |
18092 | (if (zerop length) ; @r{if-part} | |
18093 | (error "Kill ring is empty") ; @r{then-part} | |
18094 | @dots{} | |
18095 | @end group | |
18096 | @end smallexample | |
18097 | ||
18098 | @noindent | |
18099 | If there is not anything in the kill ring, its length must be zero and | |
18100 | an error message sent to the user: @samp{Kill ring is empty}. The | |
18101 | @code{if} expression uses the function @code{zerop} which returns true | |
18102 | if the value it is testing is zero. When @code{zerop} tests true, the | |
18103 | then-part of the @code{if} is evaluated. The then-part is a list | |
18104 | starting with the function @code{error}, which is a function that is | |
18105 | similar to the @code{message} function (@pxref{message}), in that it | |
18106 | prints a one-line message in the echo area. However, in addition to | |
18107 | printing a message, @code{error} also stops evaluation of the function | |
18108 | within which it is embedded. This means that the rest of the function | |
18109 | will not be evaluated if the length of the kill ring is zero. | |
18110 | ||
18111 | @menu | |
18112 | * Digression concerning error:: How to mislead humans, but not computers. | |
18113 | * rotate-yk-ptr else-part:: The else-part of the @code{if} expression. | |
18114 | * Remainder Function:: The remainder, @code{%}, function. | |
18115 | * rotate-yk-ptr remainder:: Using @code{%} in @code{rotate-yank-pointer}. | |
18116 | * kill-rng-yk-ptr last elt:: Pointing to the last element. | |
18117 | @end menu | |
18118 | ||
18119 | @node Digression concerning error, rotate-yk-ptr else-part, rotate-yk-ptr body, rotate-yk-ptr body | |
18120 | @ifnottex | |
18121 | @unnumberedsubsubsec Digression about the word `error' | |
18122 | @end ifnottex | |
18123 | ||
18124 | (In my opinion, it is slightly misleading, at least to humans, to use | |
18125 | the term `error' as the name of the @code{error} function. A better | |
18126 | term would be `cancel'. Strictly speaking, of course, you cannot | |
18127 | point to, much less rotate a pointer to a list that has no length, so | |
18128 | from the point of view of the computer, the word `error' is correct. | |
18129 | But a human expects to attempt this sort of thing, if only to find out | |
18130 | whether the kill ring is full or empty. This is an act of | |
18131 | exploration. | |
18132 | ||
18133 | (From the human point of view, the act of exploration and discovery is | |
18134 | not necessarily an error, and therefore should not be labelled as one, | |
18135 | even in the bowels of a computer. As it is, the code in Emacs implies | |
18136 | that a human who is acting virtuously, by exploring his or her | |
18137 | environment, is making an error. This is bad. Even though the computer | |
18138 | takes the same steps as it does when there is an `error', a term such as | |
18139 | `cancel' would have a clearer connotation.) | |
18140 | ||
18141 | @node rotate-yk-ptr else-part, Remainder Function, Digression concerning error, rotate-yk-ptr body | |
18142 | @unnumberedsubsubsec The else-part of the @code{if} expression | |
18143 | ||
18144 | The else-part of the @code{if} expression is dedicated to setting the | |
18145 | value of @code{kill-ring-yank-pointer} when the kill ring has something | |
18146 | in it. The code looks like this: | |
18147 | ||
18148 | @smallexample | |
18149 | @group | |
18150 | (setq kill-ring-yank-pointer | |
18151 | (nthcdr (% (+ arg | |
18152 | (- length | |
18153 | (length kill-ring-yank-pointer))) | |
18154 | length) | |
18155 | kill-ring))))) | |
18156 | @end group | |
18157 | @end smallexample | |
18158 | ||
18159 | This needs some examination. Clearly, @code{kill-ring-yank-pointer} | |
18160 | is being set to be equal to some @sc{cdr} of the kill ring, using the | |
18161 | @code{nthcdr} function that is described in an earlier section. | |
18162 | (@xref{copy-region-as-kill}.) But exactly how does it do this? | |
18163 | ||
18164 | Before looking at the details of the code let's first consider the | |
18165 | purpose of the @code{rotate-yank-pointer} function. | |
18166 | ||
18167 | The @code{rotate-yank-pointer} function changes what | |
18168 | @code{kill-ring-yank-pointer} points to. If | |
18169 | @code{kill-ring-yank-pointer} starts by pointing to the first element | |
18170 | of a list, a call to @code{rotate-yank-pointer} causes it to point to | |
18171 | the second element; and if @code{kill-ring-yank-pointer} points to the | |
18172 | second element, a call to @code{rotate-yank-pointer} causes it to | |
18173 | point to the third element. (And if @code{rotate-yank-pointer} is | |
18174 | given an argument greater than 1, it jumps the pointer that many | |
18175 | elements.) | |
18176 | ||
18177 | The @code{rotate-yank-pointer} function uses @code{setq} to reset what | |
18178 | the @code{kill-ring-yank-pointer} points to. If | |
18179 | @code{kill-ring-yank-pointer} points to the first element of the kill | |
18180 | ring, then, in the simplest case, the @code{rotate-yank-pointer} | |
18181 | function must cause it to point to the second element. Put another | |
18182 | way, @code{kill-ring-yank-pointer} must be reset to have a value equal | |
18183 | to the @sc{cdr} of the kill ring. | |
18184 | ||
18185 | @need 1250 | |
18186 | That is, under these circumstances, | |
18187 | ||
18188 | @smallexample | |
18189 | @group | |
18190 | (setq kill-ring-yank-pointer | |
18191 | ("some text" "a different piece of text" "yet more text")) | |
18192 | ||
18193 | (setq kill-ring | |
18194 | ("some text" "a different piece of text" "yet more text")) | |
18195 | @end group | |
18196 | @end smallexample | |
18197 | ||
18198 | @need 800 | |
18199 | @noindent | |
18200 | the code should do this: | |
18201 | ||
18202 | @smallexample | |
18203 | (setq kill-ring-yank-pointer (cdr kill-ring)) | |
18204 | @end smallexample | |
18205 | ||
18206 | @need 1000 | |
18207 | @noindent | |
18208 | As a result, the @code{kill-ring-yank-pointer} will look like this: | |
18209 | ||
18210 | @smallexample | |
18211 | @group | |
18212 | kill-ring-yank-pointer | |
18213 | @result{} ("a different piece of text" "yet more text")) | |
18214 | @end group | |
18215 | @end smallexample | |
18216 | ||
18217 | The actual @code{setq} expression uses the @code{nthcdr} function to do | |
18218 | the job. | |
18219 | ||
18220 | As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function | |
18221 | works by repeatedly taking the @sc{cdr} of a list---it takes the | |
18222 | @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{} | |
18223 | ||
18224 | @need 800 | |
18225 | The two following expressions produce the same result: | |
18226 | ||
18227 | @smallexample | |
18228 | @group | |
18229 | (setq kill-ring-yank-pointer (cdr kill-ring)) | |
18230 | ||
18231 | (setq kill-ring-yank-pointer (nthcdr 1 kill-ring)) | |
18232 | @end group | |
18233 | @end smallexample | |
18234 | ||
18235 | In the @code{rotate-yank-pointer} function, however, the first | |
18236 | argument to @code{nthcdr} is a rather complex looking expression with | |
18237 | lots of arithmetic inside of it: | |
18238 | ||
18239 | @smallexample | |
18240 | @group | |
18241 | (% (+ arg | |
18242 | (- length | |
18243 | (length kill-ring-yank-pointer))) | |
18244 | length) | |
18245 | @end group | |
18246 | @end smallexample | |
18247 | ||
18248 | As usual, we need to look at the most deeply embedded expression first | |
18249 | and then work our way towards the light. | |
18250 | ||
18251 | The most deeply embedded expression is @code{(length | |
18252 | kill-ring-yank-pointer)}. This finds the length of the current value of | |
18253 | the @code{kill-ring-yank-pointer}. (Remember that the | |
18254 | @code{kill-ring-yank-pointer} is the name of a variable whose value is a | |
18255 | list.) | |
18256 | ||
18257 | @need 800 | |
18258 | The measurement of the length is inside the expression: | |
18259 | ||
18260 | @smallexample | |
18261 | (- length (length kill-ring-yank-pointer)) | |
18262 | @end smallexample | |
18263 | ||
18264 | @noindent | |
18265 | In this expression, the first @code{length} is the variable that was | |
18266 | assigned the length of the kill ring in the @code{let} statement at the | |
18267 | beginning of the function. (One might think this function would be | |
18268 | clearer if the variable @code{length} were named | |
18269 | @code{length-of-kill-ring} instead; but if you look at the text of the | |
18270 | whole function, you will see that it is so short that naming this | |
18271 | variable @code{length} is not a bother, unless you are pulling the | |
18272 | function apart into very tiny pieces as we are doing here.) | |
18273 | ||
18274 | So the line @code{(- length (length kill-ring-yank-pointer))} tells the | |
18275 | difference between the length of the kill ring and the length of the list | |
18276 | whose name is @code{kill-ring-yank-pointer}. | |
18277 | ||
18278 | To see how all this fits into the @code{rotate-yank-pointer} | |
18279 | function, let's begin by analyzing the case where | |
18280 | @code{kill-ring-yank-pointer} points to the first element of the kill | |
18281 | ring, just as @code{kill-ring} does, and see what happens when | |
18282 | @code{rotate-yank-pointer} is called with an argument of 1. | |
18283 | ||
18284 | The variable @code{length} and the value of the expression | |
18285 | @code{(length kill-ring-yank-pointer)} will be the same since the | |
18286 | variable @code{length} is the length of the kill ring and the | |
18287 | @code{kill-ring-yank-pointer} is pointing to the whole kill ring. | |
18288 | Consequently, the value of | |
18289 | ||
18290 | @smallexample | |
18291 | (- length (length kill-ring-yank-pointer)) | |
18292 | @end smallexample | |
18293 | ||
18294 | @noindent | |
18295 | will be zero. Since the value of @code{arg} will be 1, this will mean | |
18296 | that the value of the whole expression | |
18297 | ||
18298 | @smallexample | |
18299 | (+ arg (- length (length kill-ring-yank-pointer))) | |
18300 | @end smallexample | |
18301 | ||
18302 | @noindent | |
18303 | will be 1. | |
18304 | ||
18305 | Consequently, the argument to @code{nthcdr} will be found as the result of | |
18306 | the expression | |
18307 | ||
18308 | @smallexample | |
18309 | (% 1 length) | |
18310 | @end smallexample | |
18311 | ||
18312 | @node Remainder Function, rotate-yk-ptr remainder, rotate-yk-ptr else-part, rotate-yk-ptr body | |
18313 | @unnumberedsubsubsec The @code{%} remainder function | |
18314 | ||
18315 | To understand @code{(% 1 length)}, we need to understand @code{%}. | |
18316 | According to its documentation (which I just found by typing @kbd{C-h | |
18317 | f @kbd{%} @key{RET}}), the @code{%} function returns the remainder of | |
18318 | its first argument divided by its second argument. For example, the | |
18319 | remainder of 5 divided by 2 is 1. (2 goes into 5 twice with a | |
18320 | remainder of 1.) | |
18321 | ||
18322 | What surprises people who don't often do arithmetic is that a smaller | |
18323 | number can be divided by a larger number and have a remainder. In the | |
18324 | example we just used, 5 was divided by 2. We can reverse that and ask, | |
18325 | what is the result of dividing 2 by 5? If you can use fractions, the | |
18326 | answer is obviously 2/5 or .4; but if, as here, you can only use whole | |
18327 | numbers, the result has to be something different. Clearly, 5 can go into | |
18328 | 2 zero times, but what of the remainder? To see what the answer is, | |
18329 | consider a case that has to be familiar from childhood: | |
18330 | ||
18331 | @itemize @bullet | |
18332 | @item | |
18333 | 5 divided by 5 is 1 with a remainder of 0; | |
18334 | ||
18335 | @item | |
18336 | 6 divided by 5 is 1 with a remainder of 1; | |
18337 | ||
18338 | @item | |
18339 | 7 divided by 5 is 1 with a remainder of 2. | |
18340 | ||
18341 | @item | |
18342 | Similarly, 10 divided by 5 is 2 with a remainder of 0; | |
18343 | ||
18344 | @item | |
18345 | 11 divided by 5 is 2 with a remainder of 1; | |
18346 | ||
18347 | @item | |
18348 | 12 divided by 5 is 1 with a remainder of 2. | |
18349 | @end itemize | |
18350 | ||
18351 | @need 1250 | |
18352 | @noindent | |
18353 | By considering the cases as parallel, we can see that | |
18354 | ||
18355 | @itemize @bullet | |
18356 | @item | |
18357 | zero divided by 5 must be zero with a remainder of zero; | |
18358 | ||
18359 | @item | |
18360 | 1 divided by 5 must be zero with a remainder of 1; | |
18361 | ||
18362 | @item | |
18363 | 2 divided by 5 must be zero with a remainder of 2; | |
18364 | @end itemize | |
18365 | ||
18366 | @noindent | |
18367 | and so on. | |
18368 | ||
18369 | @need 1250 | |
18370 | So, in this code, if the value of @code{length} is 5, then the result of | |
18371 | evaluating | |
18372 | ||
18373 | @smallexample | |
18374 | (% 1 5) | |
18375 | @end smallexample | |
18376 | ||
18377 | @noindent | |
18378 | is 1. (I just checked this by placing the cursor after the expression | |
18379 | and typing @kbd{C-x C-e}. Indeed, 1 is printed in the echo area.) | |
18380 | ||
18381 | @node rotate-yk-ptr remainder, kill-rng-yk-ptr last elt, Remainder Function, rotate-yk-ptr body | |
18382 | @unnumberedsubsubsec Using @code{%} in @code{rotate-yank-pointer} | |
18383 | ||
18384 | When the @code{kill-ring-yank-pointer} points to the | |
18385 | beginning of the kill ring, and the argument passed to | |
18386 | @code{rotate-yank-pointer} is 1, the @code{%} expression returns 1: | |
18387 | ||
18388 | @smallexample | |
18389 | @group | |
18390 | (- length (length kill-ring-yank-pointer)) | |
18391 | @result{} 0 | |
18392 | @end group | |
18393 | @end smallexample | |
18394 | ||
18395 | @need 1250 | |
18396 | @noindent | |
18397 | therefore, | |
18398 | ||
18399 | @smallexample | |
18400 | @group | |
18401 | (+ arg (- length (length kill-ring-yank-pointer))) | |
18402 | @result{} 1 | |
18403 | @end group | |
18404 | @end smallexample | |
18405 | ||
18406 | @need 1250 | |
18407 | @noindent | |
18408 | and consequently: | |
18409 | ||
18410 | @smallexample | |
18411 | @group | |
18412 | (% (+ arg (- length (length kill-ring-yank-pointer))) | |
18413 | length) | |
18414 | @result{} 1 | |
18415 | @end group | |
18416 | @end smallexample | |
18417 | ||
18418 | @noindent | |
18419 | regardless of the value of @code{length}. | |
18420 | ||
18421 | @need 1250 | |
18422 | @noindent | |
18423 | As a result of this, the @code{setq kill-ring-yank-pointer} expression | |
18424 | simplifies to: | |
18425 | ||
18426 | @smallexample | |
18427 | (setq kill-ring-yank-pointer (nthcdr 1 kill-ring)) | |
18428 | @end smallexample | |
18429 | ||
18430 | @noindent | |
18431 | What it does is now easy to understand. Instead of pointing as it did | |
18432 | to the first element of the kill ring, the | |
18433 | @code{kill-ring-yank-pointer} is set to point to the second element. | |
18434 | ||
18435 | Clearly, if the argument passed to @code{rotate-yank-pointer} is two, then | |
18436 | the @code{kill-ring-yank-pointer} is set to @code{(nthcdr 2 kill-ring)}; | |
18437 | and so on for different values of the argument. | |
18438 | ||
18439 | Similarly, if the @code{kill-ring-yank-pointer} starts out pointing to | |
18440 | the second element of the kill ring, its length is shorter than the | |
18441 | length of the kill ring by 1, so the computation of the remainder is | |
18442 | based on the expression @code{(% (+ arg 1) length)}. This means that | |
18443 | the @code{kill-ring-yank-pointer} is moved from the second element of | |
18444 | the kill ring to the third element if the argument passed to | |
18445 | @code{rotate-yank-pointer} is 1. | |
18446 | ||
18447 | @node kill-rng-yk-ptr last elt, , rotate-yk-ptr remainder, rotate-yk-ptr body | |
18448 | @unnumberedsubsubsec Pointing to the last element | |
18449 | ||
18450 | The final question is, what happens if the @code{kill-ring-yank-pointer} | |
18451 | is set to the @emph{last} element of the kill ring? Will a call to | |
18452 | @code{rotate-yank-pointer} mean that nothing more can be taken from the | |
18453 | kill ring? The answer is no. What happens is different and useful. | |
18454 | The @code{kill-ring-yank-pointer} is set to point to the beginning of | |
18455 | the kill ring instead. | |
18456 | ||
18457 | Let's see how this works by looking at the code, assuming the length of the | |
18458 | kill ring is 5 and the argument passed to @code{rotate-yank-pointer} is 1. | |
18459 | When the @code{kill-ring-yank-pointer} points to the last element of | |
18460 | the kill ring, its length is 1. The code looks like this: | |
18461 | ||
18462 | @smallexample | |
18463 | (% (+ arg (- length (length kill-ring-yank-pointer))) length) | |
18464 | @end smallexample | |
18465 | ||
18466 | @need 1250 | |
18467 | When the variables are replaced by their numeric values, the expression | |
18468 | looks like this: | |
18469 | ||
18470 | @smallexample | |
18471 | (% (+ 1 (- 5 1)) 5) | |
18472 | @end smallexample | |
18473 | ||
18474 | @noindent | |
18475 | This expression can be evaluated by looking at the most embedded inner | |
18476 | expression first and working outwards: The value of @code{(- 5 1)} is 4; | |
18477 | the sum of @code{(+ 1 4)} is 5; and the remainder of dividing 5 by 5 is | |
18478 | zero. So what @code{rotate-yank-pointer} will do is | |
18479 | ||
18480 | @smallexample | |
18481 | (setq kill-ring-yank-pointer (nthcdr 0 kill-ring)) | |
18482 | @end smallexample | |
18483 | ||
18484 | @noindent | |
18485 | which will set the @code{kill-ring-yank-pointer} to point to the beginning | |
18486 | of the kill ring. | |
18487 | ||
18488 | So what happens with successive calls to @code{rotate-yank-pointer} is that | |
18489 | it moves the @code{kill-ring-yank-pointer} from element to element in the | |
18490 | kill ring until it reaches the end; then it jumps back to the beginning. | |
18491 | And this is why the kill ring is called a ring, since by jumping back to | |
18492 | the beginning, it is as if the list has no end! (And what is a ring, but | |
18493 | an entity with no end?) | |
18494 | ||
18495 | @node yank, yank-pop, rotate-yank-pointer, Kill Ring | |
18496 | @comment node-name, next, previous, up | |
18497 | @appendixsec @code{yank} | |
18498 | @findex yank | |
18499 | ||
18500 | After learning about @code{rotate-yank-pointer}, the code for the | |
18501 | @code{yank} function is almost easy. It has only one tricky part, which is | |
18502 | the computation of the argument to be passed to @code{rotate-yank-pointer}. | |
18503 | ||
18504 | @need 1250 | |
18505 | The code looks like this: | |
18506 | ||
18507 | @smallexample | |
18508 | @group | |
18509 | (defun yank (&optional arg) | |
18510 | "Reinsert the last stretch of killed text. | |
18511 | More precisely, reinsert the stretch of killed text most | |
18512 | recently killed OR yanked. | |
18513 | With just C-U as argument, same but put point in front | |
18514 | (and mark at end). With argument n, reinsert the nth | |
18515 | most recently killed stretch of killed text. | |
18516 | See also the command \\[yank-pop]." | |
18517 | @end group | |
18518 | @group | |
18519 | ||
18520 | (interactive "*P") | |
18521 | (rotate-yank-pointer (if (listp arg) 0 | |
18522 | (if (eq arg '-) -1 | |
18523 | (1- arg)))) | |
18524 | (push-mark (point)) | |
18525 | (insert (car kill-ring-yank-pointer)) | |
18526 | (if (consp arg) | |
18527 | (exchange-point-and-mark))) | |
18528 | @end group | |
18529 | @end smallexample | |
18530 | ||
18531 | Glancing over this code, we can understand the last few lines readily | |
18532 | enough. The mark is pushed, that is, remembered; then the first element | |
18533 | (the @sc{car}) of what the @code{kill-ring-yank-pointer} points to is | |
18534 | inserted; and then, if the argument passed the function is a | |
18535 | @code{cons}, point and mark are exchanged so the point is put in the | |
18536 | front of the inserted text rather than at the end. This option is | |
18537 | explained in the documentation. The function itself is interactive with | |
18538 | @code{"*P"}. This means it will not work on a read-only buffer, and that | |
18539 | the unprocessed prefix argument is passed to the function. | |
18540 | ||
18541 | @menu | |
18542 | * rotate-yk-ptr arg:: Pass the argument to @code{rotate-yank-pointer}. | |
18543 | * rotate-yk-ptr negative arg:: Pass a negative argument. | |
18544 | @end menu | |
18545 | ||
18546 | @node rotate-yk-ptr arg, rotate-yk-ptr negative arg, yank, yank | |
18547 | @unnumberedsubsubsec Passing the argument | |
18548 | ||
18549 | The hard part of @code{yank} is understanding the computation that | |
18550 | determines the value of the argument passed to | |
18551 | @code{rotate-yank-pointer}. Fortunately, it is not so difficult as it | |
18552 | looks at first sight. | |
18553 | ||
18554 | What happens is that the result of evaluating one or both of the | |
18555 | @code{if} expressions will be a number and that number will be the | |
18556 | argument passed to @code{rotate-yank-pointer}. | |
18557 | ||
18558 | @need 1250 | |
18559 | Laid out with comments, the code looks like this: | |
18560 | ||
18561 | @smallexample | |
18562 | @group | |
18563 | (if (listp arg) ; @r{if-part} | |
18564 | 0 ; @r{then-part} | |
18565 | (if (eq arg '-) ; @r{else-part, inner if} | |
18566 | -1 ; @r{inner if's then-part} | |
18567 | (1- arg)))) ; @r{inner if's else-part} | |
18568 | @end group | |
18569 | @end smallexample | |
18570 | ||
18571 | @noindent | |
18572 | This code consists of two @code{if} expression, one the else-part of | |
18573 | the other. | |
18574 | ||
18575 | The first or outer @code{if} expression tests whether the argument | |
18576 | passed to @code{yank} is a list. Oddly enough, this will be true if | |
18577 | @code{yank} is called without an argument---because then it will be | |
18578 | passed the value of @code{nil} for the optional argument and an | |
18579 | evaluation of @code{(listp nil)} returns true! So, if no argument is | |
18580 | passed to @code{yank}, the argument passed to | |
18581 | @code{rotate-yank-pointer} inside of @code{yank} is zero. This means | |
18582 | the pointer is not moved and the first element to which | |
18583 | @code{kill-ring-yank-pointer} points is inserted, as we expect. | |
18584 | Similarly, if the argument for @code{yank} is @kbd{C-u}, this will be | |
18585 | read as a list, so again, a zero will be passed to | |
18586 | @code{rotate-yank-pointer}. (@kbd{C-u} produces an unprocessed prefix | |
18587 | argument of @code{(4)}, which is a list of one element.) At the same | |
18588 | time, later in the function, this argument will be read as a | |
18589 | @code{cons} so point will be put in the front and mark at the end of | |
18590 | the insertion. (The @code{P} argument to @code{interactive} is | |
18591 | designed to provide these values for the case when an optional | |
18592 | argument is not provided or when it is @kbd{C-u}.) | |
18593 | ||
18594 | The then-part of the outer @code{if} expression handles the case when | |
18595 | there is no argument or when it is @kbd{C-u}. The else-part handles the | |
18596 | other situations. The else-part is itself another @code{if} expression. | |
18597 | ||
18598 | The inner @code{if} expression tests whether the argument is a minus | |
18599 | sign. (This is done by pressing the @key{META} and @kbd{-} keys at the | |
18600 | same time, or the @key{ESC} key and then the @kbd{-} key). In this | |
18601 | case, the @code{rotate-yank-pointer} function is passed @kbd{-1} as an | |
18602 | argument. This moves the @code{kill-ring-yank-pointer} backwards, which | |
18603 | is what is desired. | |
18604 | ||
18605 | If the true-or-false-test of the inner @code{if} expression is false | |
18606 | (that is, if the argument is not a minus sign), the else-part of the | |
18607 | expression is evaluated. This is the expression @code{(1- arg)}. | |
18608 | Because of the two @code{if} expressions, it will only occur when the | |
18609 | argument is a positive number or when it is a negative number (not | |
18610 | just a minus sign on its own). What @code{(1- arg)} does is decrement | |
18611 | the number and return it. (The @code{1-} function subtracts one from | |
18612 | its argument.) This means that if the argument to | |
18613 | @code{rotate-yank-pointer} is 1, it is reduced to zero, which means | |
18614 | the first element to which @code{kill-ring-yank-pointer} points is | |
18615 | yanked back, as you would expect. | |
18616 | ||
18617 | @node rotate-yk-ptr negative arg, , rotate-yk-ptr arg, yank | |
18618 | @unnumberedsubsubsec Passing a negative argument | |
18619 | ||
18620 | Finally, the question arises, what happens if either the remainder | |
18621 | function, @code{%}, or the @code{nthcdr} function is passed a negative | |
18622 | argument, as they quite well may? | |
18623 | ||
18624 | The answers can be found by a quick test. When @code{(% -1 5)} is | |
18625 | evaluated, a negative number is returned; and if @code{nthcdr} is | |
18626 | called with a negative number, it returns the same value as if it were | |
18627 | called with a first argument of zero. This can be seen be evaluating | |
18628 | the following code. | |
18629 | ||
18630 | Here the @samp{@result{}} points to the result of evaluating the code | |
18631 | preceding it. This was done by positioning the cursor after the code | |
18632 | and typing @kbd{C-x C-e} (@code{eval-last-sexp}) in the usual fashion. | |
18633 | You can do this if you are reading this in Info inside of GNU Emacs. | |
18634 | ||
18635 | @smallexample | |
18636 | @group | |
18637 | (% -1 5) | |
18638 | @result{} -1 | |
18639 | @end group | |
18640 | ||
18641 | @group | |
18642 | (setq animals '(cats dogs elephants)) | |
18643 | @result{} (cats dogs elephants) | |
18644 | @end group | |
18645 | ||
18646 | @group | |
18647 | (nthcdr 1 animals) | |
18648 | @result{} (dogs elephants) | |
18649 | @end group | |
18650 | ||
18651 | @group | |
18652 | (nthcdr 0 animals) | |
18653 | @result{} (cats dogs elephants) | |
18654 | @end group | |
18655 | ||
18656 | @group | |
18657 | (nthcdr -1 animals) | |
18658 | @result{} (cats dogs elephants) | |
18659 | @end group | |
18660 | @end smallexample | |
18661 | ||
18662 | So, if a minus sign or a negative number is passed to @code{yank}, the | |
18663 | @code{kill-ring-yank-point} is rotated backwards until it reaches the | |
18664 | beginning of the list. Then it stays there. Unlike the other case, | |
18665 | when it jumps from the end of the list to the beginning of the list, | |
18666 | making a ring, it stops. This makes sense. You often want to get back | |
18667 | to the most recently clipped out piece of text, but you don't usually | |
18668 | want to insert text from as many as thirty kill commands ago. So you | |
18669 | need to work through the ring to get to the end, but won't cycle around | |
18670 | it inadvertently if you are trying to come back to the beginning. | |
18671 | ||
18672 | Incidentally, any number passed to @code{yank} with a minus sign | |
18673 | preceding it will be treated as @minus{}1. This is evidently a | |
18674 | simplification for writing the program. You don't need to jump back | |
18675 | towards the beginning of the kill ring more than one place at a time | |
18676 | and doing this is easier than writing a function to determine the | |
18677 | magnitude of the number that follows the minus sign. | |
18678 | ||
18679 | @node yank-pop, , yank, Kill Ring | |
18680 | @comment node-name, next, previous, up | |
18681 | @appendixsec @code{yank-pop} | |
18682 | @findex yank-pop | |
18683 | ||
18684 | After understanding @code{yank}, the @code{yank-pop} function is easy. | |
18685 | Leaving out the documentation to save space, it looks like this: | |
18686 | ||
18687 | @smallexample | |
18688 | @group | |
18689 | (defun yank-pop (arg) | |
18690 | (interactive "*p") | |
18691 | (if (not (eq last-command 'yank)) | |
18692 | (error "Previous command was not a yank")) | |
18693 | @end group | |
18694 | @group | |
18695 | (setq this-command 'yank) | |
18696 | (let ((before (< (point) (mark)))) | |
18697 | (delete-region (point) (mark)) | |
18698 | (rotate-yank-pointer arg) | |
18699 | @end group | |
18700 | @group | |
18701 | (set-mark (point)) | |
18702 | (insert (car kill-ring-yank-pointer)) | |
18703 | (if before (exchange-point-and-mark)))) | |
18704 | @end group | |
18705 | @end smallexample | |
18706 | ||
18707 | The function is interactive with a small @samp{p} so the prefix | |
18708 | argument is processed and passed to the function. The command can | |
18709 | only be used after a previous yank; otherwise an error message is | |
18710 | sent. This check uses the variable @code{last-command} which is | |
18711 | discussed elsewhere. (@xref{copy-region-as-kill}.) | |
18712 | ||
18713 | The @code{let} clause sets the variable @code{before} to true or false | |
18714 | depending whether point is before or after mark and then the region | |
18715 | between point and mark is deleted. This is the region that was just | |
18716 | inserted by the previous yank and it is this text that will be | |
18717 | replaced. Next the @code{kill-ring-yank-pointer} is rotated so that | |
18718 | the previously inserted text is not reinserted yet again. Mark is set | |
18719 | at the beginning of the place the new text will be inserted and then | |
18720 | the first element to which @code{kill-ring-yank-pointer} points is | |
18721 | inserted. This leaves point after the new text. If in the previous | |
18722 | yank, point was left before the inserted text, point and mark are now | |
18723 | exchanged so point is again left in front of the newly inserted text. | |
18724 | That is all there is to it! | |
18725 | ||
18726 | @node Full Graph, GNU Free Documentation License, Kill Ring, Top | |
18727 | @appendix A Graph with Labelled Axes | |
18728 | ||
18729 | Printed axes help you understand a graph. They convey scale. In an | |
18730 | earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we | |
18731 | wrote the code to print the body of a graph. Here we write the code | |
18732 | for printing and labelling vertical and horizontal axes, along with the | |
18733 | body itself. | |
18734 | ||
18735 | @menu | |
18736 | * Labelled Example:: | |
18737 | * print-graph Varlist:: @code{let} expression in @code{print-graph}. | |
18738 | * print-Y-axis:: Print a label for the vertical axis. | |
18739 | * print-X-axis:: Print a horizontal label. | |
18740 | * Print Whole Graph:: The function to print a complete graph. | |
18741 | @end menu | |
18742 | ||
18743 | @node Labelled Example, print-graph Varlist, Full Graph, Full Graph | |
18744 | @ifnottex | |
18745 | @unnumberedsec Labelled Example Graph | |
18746 | @end ifnottex | |
18747 | ||
18748 | Since insertions fill a buffer to the right and below point, the new | |
18749 | graph printing function should first print the Y or vertical axis, | |
18750 | then the body of the graph, and finally the X or horizontal axis. | |
18751 | This sequence lays out for us the contents of the function: | |
18752 | ||
18753 | @enumerate | |
18754 | @item | |
18755 | Set up code. | |
18756 | ||
18757 | @item | |
18758 | Print Y axis. | |
18759 | ||
18760 | @item | |
18761 | Print body of graph. | |
18762 | ||
18763 | @item | |
18764 | Print X axis. | |
18765 | @end enumerate | |
18766 | ||
18767 | @need 800 | |
18768 | Here is an example of how a finished graph should look: | |
18769 | ||
18770 | @smallexample | |
18771 | @group | |
18772 | 10 - | |
18773 | * | |
18774 | * * | |
18775 | * ** | |
18776 | * *** | |
18777 | 5 - * ******* | |
18778 | * *** ******* | |
18779 | ************* | |
18780 | *************** | |
18781 | 1 - **************** | |
18782 | | | | | | |
18783 | 1 5 10 15 | |
18784 | @end group | |
18785 | @end smallexample | |
18786 | ||
18787 | @noindent | |
18788 | In this graph, both the vertical and the horizontal axes are labelled | |
18789 | with numbers. However, in some graphs, the horizontal axis is time | |
18790 | and would be better labelled with months, like this: | |
18791 | ||
18792 | @smallexample | |
18793 | @group | |
18794 | 5 - * | |
18795 | * ** * | |
18796 | ******* | |
18797 | ********** ** | |
18798 | 1 - ************** | |
18799 | | ^ | | |
18800 | Jan June Jan | |
18801 | @end group | |
18802 | @end smallexample | |
18803 | ||
18804 | Indeed, with a little thought, we can easily come up with a variety of | |
18805 | vertical and horizontal labelling schemes. Our task could become | |
18806 | complicated. But complications breed confusion. Rather than permit | |
18807 | this, it is better choose a simple labelling scheme for our first | |
18808 | effort, and to modify or replace it later. | |
18809 | ||
18810 | @need 1200 | |
18811 | These considerations suggest the following outline for the | |
18812 | @code{print-graph} function: | |
18813 | ||
18814 | @smallexample | |
18815 | @group | |
18816 | (defun print-graph (numbers-list) | |
18817 | "@var{documentation}@dots{}" | |
18818 | (let ((height @dots{} | |
18819 | @dots{})) | |
18820 | @end group | |
18821 | @group | |
18822 | (print-Y-axis height @dots{} ) | |
18823 | (graph-body-print numbers-list) | |
18824 | (print-X-axis @dots{} ))) | |
18825 | @end group | |
18826 | @end smallexample | |
18827 | ||
18828 | We can work on each part of the @code{print-graph} function definition | |
18829 | in turn. | |
18830 | ||
18831 | @node print-graph Varlist, print-Y-axis, Labelled Example, Full Graph | |
18832 | @comment node-name, next, previous, up | |
18833 | @appendixsec The @code{print-graph} Varlist | |
18834 | @cindex @code{print-graph} varlist | |
18835 | ||
18836 | In writing the @code{print-graph} function, the first task is to write | |
18837 | the varlist in the @code{let} expression. (We will leave aside for the | |
18838 | moment any thoughts about making the function interactive or about the | |
18839 | contents of its documentation string.) | |
18840 | ||
18841 | The varlist should set several values. Clearly, the top of the label | |
18842 | for the vertical axis must be at least the height of the graph, which | |
18843 | means that we must obtain this information here. Note that the | |
18844 | @code{print-graph-body} function also requires this information. There | |
18845 | is no reason to calculate the height of the graph in two different | |
18846 | places, so we should change @code{print-graph-body} from the way we | |
18847 | defined it earlier to take advantage of the calculation. | |
18848 | ||
18849 | Similarly, both the function for printing the X axis labels and the | |
18850 | @code{print-graph-body} function need to learn the value of the width of | |
18851 | each symbol. We can perform the calculation here and change the | |
18852 | definition for @code{print-graph-body} from the way we defined it in the | |
18853 | previous chapter. | |
18854 | ||
18855 | The length of the label for the horizontal axis must be at least as long | |
18856 | as the graph. However, this information is used only in the function | |
18857 | that prints the horizontal axis, so it does not need to be calculated here. | |
18858 | ||
18859 | These thoughts lead us directly to the following form for the varlist | |
18860 | in the @code{let} for @code{print-graph}: | |
18861 | ||
18862 | @smallexample | |
18863 | @group | |
18864 | (let ((height (apply 'max numbers-list)) ; @r{First version.} | |
18865 | (symbol-width (length graph-blank))) | |
18866 | @end group | |
18867 | @end smallexample | |
18868 | ||
18869 | @noindent | |
18870 | As we shall see, this expression is not quite right. | |
18871 | ||
18872 | @node print-Y-axis, print-X-axis, print-graph Varlist, Full Graph | |
18873 | @comment node-name, next, previous, up | |
18874 | @appendixsec The @code{print-Y-axis} Function | |
18875 | @cindex Axis, print vertical | |
18876 | @cindex Y axis printing | |
18877 | @cindex Vertical axis printing | |
18878 | @cindex Print vertical axis | |
18879 | ||
18880 | The job of the @code{print-Y-axis} function is to print a label for | |
18881 | the vertical axis that looks like this: | |
18882 | ||
18883 | @smallexample | |
18884 | @group | |
18885 | 10 - | |
18886 | ||
18887 | ||
18888 | ||
18889 | ||
18890 | 5 - | |
18891 | ||
18892 | ||
18893 | ||
18894 | 1 - | |
18895 | @end group | |
18896 | @end smallexample | |
18897 | ||
18898 | @noindent | |
18899 | The function should be passed the height of the graph, and then should | |
18900 | construct and insert the appropriate numbers and marks. | |
18901 | ||
18902 | It is easy enough to see in the figure what the Y axis label should | |
18903 | look like; but to say in words, and then to write a function | |
18904 | definition to do the job is another matter. It is not quite true to | |
18905 | say that we want a number and a tic every five lines: there are only | |
18906 | three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4), | |
18907 | but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8, | |
18908 | and 9). It is better to say that we want a number and a tic mark on | |
18909 | the base line (number 1) and then that we want a number and a tic on | |
18910 | the fifth line from the bottom and on every line that is a multiple of | |
18911 | five. | |
18912 | ||
18913 | @menu | |
18914 | * Height of label:: What height for the Y axis? | |
18915 | * Compute a Remainder:: How to compute the remainder of a division. | |
18916 | * Y Axis Element:: Construct a line for the Y axis. | |
18917 | * Y-axis-column:: Generate a list of Y axis labels. | |
18918 | * print-Y-axis Penultimate:: A not quite final version. | |
18919 | @end menu | |
18920 | ||
18921 | @node Height of label, Compute a Remainder, print-Y-axis, print-Y-axis | |
18922 | @ifnottex | |
18923 | @unnumberedsubsec What height should the label be? | |
18924 | @end ifnottex | |
18925 | ||
18926 | The next issue is what height the label should be? Suppose the maximum | |
18927 | height of tallest column of the graph is seven. Should the highest | |
18928 | label on the Y axis be @samp{5 -}, and should the graph stick up above | |
18929 | the label? Or should the highest label be @samp{7 -}, and mark the peak | |
18930 | of the graph? Or should the highest label be @code{10 -}, which is a | |
18931 | multiple of five, and be higher than the topmost value of the graph? | |
18932 | ||
18933 | The latter form is preferred. Most graphs are drawn within rectangles | |
18934 | whose sides are an integral number of steps long---5, 10, 15, and so | |
18935 | on for a step distance of five. But as soon as we decide to use a | |
18936 | step height for the vertical axis, we discover that the simple | |
18937 | expression in the varlist for computing the height is wrong. The | |
18938 | expression is @code{(apply 'max numbers-list)}. This returns the | |
18939 | precise height, not the maximum height plus whatever is necessary to | |
18940 | round up to the nearest multiple of five. A more complex expression | |
18941 | is required. | |
18942 | ||
18943 | As usual in cases like this, a complex problem becomes simpler if it is | |
18944 | divided into several smaller problems. | |
18945 | ||
18946 | First, consider the case when the highest value of the graph is an | |
18947 | integral multiple of five---when it is 5, 10, 15 ,or some higher | |
18948 | multiple of five. We can use this value as the Y axis height. | |
18949 | ||
18950 | A fairly simply way to determine whether a number is a multiple of | |
18951 | five is to divide it by five and see if the division results in a | |
18952 | remainder. If there is no remainder, the number is a multiple of | |
18953 | five. Thus, seven divided by five has a remainder of two, and seven | |
18954 | is not an integral multiple of five. Put in slightly different | |
18955 | language, more reminiscent of the classroom, five goes into seven | |
18956 | once, with a remainder of two. However, five goes into ten twice, | |
18957 | with no remainder: ten is an integral multiple of five. | |
18958 | ||
18959 | @node Compute a Remainder, Y Axis Element, Height of label, print-Y-axis | |
18960 | @appendixsubsec Side Trip: Compute a Remainder | |
18961 | ||
18962 | @findex % @r{(remainder function)} | |
18963 | @cindex Remainder function, @code{%} | |
18964 | In Lisp, the function for computing a remainder is @code{%}. The | |
18965 | function returns the remainder of its first argument divided by its | |
18966 | second argument. As it happens, @code{%} is a function in Emacs Lisp | |
18967 | that you cannot discover using @code{apropos}: you find nothing if you | |
18968 | type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to | |
18969 | learn of the existence of @code{%} is to read about it in a book such | |
18970 | as this or in the Emacs Lisp sources. The @code{%} function is used | |
18971 | in the code for @code{rotate-yank-pointer}, which is described in an | |
18972 | appendix. (@xref{rotate-yk-ptr body, , The Body of | |
18973 | @code{rotate-yank-pointer}}.) | |
18974 | ||
18975 | You can try the @code{%} function by evaluating the following two | |
18976 | expressions: | |
18977 | ||
18978 | @smallexample | |
18979 | @group | |
18980 | (% 7 5) | |
18981 | ||
18982 | (% 10 5) | |
18983 | @end group | |
18984 | @end smallexample | |
18985 | ||
18986 | @noindent | |
18987 | The first expression returns 2 and the second expression returns 0. | |
18988 | ||
18989 | To test whether the returned value is zero or some other number, we | |
18990 | can use the @code{zerop} function. This function returns @code{t} if | |
18991 | its argument, which must be a number, is zero. | |
18992 | ||
18993 | @smallexample | |
18994 | @group | |
18995 | (zerop (% 7 5)) | |
18996 | @result{} nil | |
18997 | ||
18998 | (zerop (% 10 5)) | |
18999 | @result{} t | |
19000 | @end group | |
19001 | @end smallexample | |
19002 | ||
19003 | Thus, the following expression will return @code{t} if the height | |
19004 | of the graph is evenly divisible by five: | |
19005 | ||
19006 | @smallexample | |
19007 | (zerop (% height 5)) | |
19008 | @end smallexample | |
19009 | ||
19010 | @noindent | |
19011 | (The value of @code{height}, of course, can be found from @code{(apply | |
19012 | 'max numbers-list)}.) | |
19013 | ||
19014 | On the other hand, if the value of @code{height} is not a multiple of | |
19015 | five, we want to reset the value to the next higher multiple of five. | |
19016 | This is straightforward arithmetic using functions with which we are | |
19017 | already familiar. First, we divide the value of @code{height} by five | |
19018 | to determine how many times five goes into the number. Thus, five | |
19019 | goes into twelve twice. If we add one to this quotient and multiply by | |
19020 | five, we will obtain the value of the next multiple of five that is | |
19021 | larger than the height. Five goes into twelve twice. Add one to two, | |
19022 | and multiply by five; the result is fifteen, which is the next multiple | |
19023 | of five that is higher than twelve. The Lisp expression for this is: | |
19024 | ||
19025 | @smallexample | |
19026 | (* (1+ (/ height 5)) 5) | |
19027 | @end smallexample | |
19028 | ||
19029 | @noindent | |
19030 | For example, if you evaluate the following, the result is 15: | |
19031 | ||
19032 | @smallexample | |
19033 | (* (1+ (/ 12 5)) 5) | |
19034 | @end smallexample | |
19035 | ||
19036 | All through this discussion, we have been using `five' as the value | |
19037 | for spacing labels on the Y axis; but we may want to use some other | |
19038 | value. For generality, we should replace `five' with a variable to | |
19039 | which we can assign a value. The best name I can think of for this | |
19040 | variable is @code{Y-axis-label-spacing}. | |
19041 | ||
19042 | @need 1250 | |
19043 | Using this term, and an @code{if} expression, we produce the | |
19044 | following: | |
19045 | ||
19046 | @smallexample | |
19047 | @group | |
19048 | (if (zerop (% height Y-axis-label-spacing)) | |
19049 | height | |
19050 | ;; @r{else} | |
19051 | (* (1+ (/ height Y-axis-label-spacing)) | |
19052 | Y-axis-label-spacing)) | |
19053 | @end group | |
19054 | @end smallexample | |
19055 | ||
19056 | @noindent | |
19057 | This expression returns the value of @code{height} itself if the height | |
19058 | is an even multiple of the value of the @code{Y-axis-label-spacing} or | |
19059 | else it computes and returns a value of @code{height} that is equal to | |
19060 | the next higher multiple of the value of the @code{Y-axis-label-spacing}. | |
19061 | ||
19062 | We can now include this expression in the @code{let} expression of the | |
19063 | @code{print-graph} function (after first setting the value of | |
19064 | @code{Y-axis-label-spacing}): | |
19065 | @vindex Y-axis-label-spacing | |
19066 | ||
19067 | @smallexample | |
19068 | @group | |
19069 | (defvar Y-axis-label-spacing 5 | |
19070 | "Number of lines from one Y axis label to next.") | |
19071 | @end group | |
19072 | ||
19073 | @group | |
19074 | @dots{} | |
19075 | (let* ((height (apply 'max numbers-list)) | |
19076 | (height-of-top-line | |
19077 | (if (zerop (% height Y-axis-label-spacing)) | |
19078 | height | |
19079 | @end group | |
19080 | @group | |
19081 | ;; @r{else} | |
19082 | (* (1+ (/ height Y-axis-label-spacing)) | |
19083 | Y-axis-label-spacing))) | |
19084 | (symbol-width (length graph-blank)))) | |
19085 | @dots{} | |
19086 | @end group | |
19087 | @end smallexample | |
19088 | ||
19089 | @noindent | |
19090 | (Note use of the @code{let*} function: the initial value of height is | |
19091 | computed once by the @code{(apply 'max numbers-list)} expression and | |
19092 | then the resulting value of @code{height} is used to compute its | |
19093 | final value. @xref{fwd-para let, , The @code{let*} expression}, for | |
19094 | more about @code{let*}.) | |
19095 | ||
19096 | @node Y Axis Element, Y-axis-column, Compute a Remainder, print-Y-axis | |
19097 | @appendixsubsec Construct a Y Axis Element | |
19098 | ||
19099 | When we print the vertical axis, we want to insert strings such as | |
19100 | @w{@samp{5 -}} and @w{@samp{10 - }} every five lines. | |
19101 | Moreover, we want the numbers and dashes to line up, so shorter | |
19102 | numbers must be padded with leading spaces. If some of the strings | |
19103 | use two digit numbers, the strings with single digit numbers must | |
19104 | include a leading blank space before the number. | |
19105 | ||
19106 | @findex number-to-string | |
19107 | To figure out the length of the number, the @code{length} function is | |
19108 | used. But the @code{length} function works only with a string, not with | |
19109 | a number. So the number has to be converted from being a number to | |
19110 | being a string. This is done with the @code{number-to-string} function. | |
19111 | For example, | |
19112 | ||
19113 | @smallexample | |
19114 | @group | |
19115 | (length (number-to-string 35)) | |
19116 | @result{} 2 | |
19117 | ||
19118 | (length (number-to-string 100)) | |
19119 | @result{} 3 | |
19120 | @end group | |
19121 | @end smallexample | |
19122 | ||
19123 | @noindent | |
19124 | (@code{number-to-string} is also called @code{int-to-string}; you will | |
19125 | see this alternative name in various sources.) | |
19126 | ||
19127 | In addition, in each label, each number is followed by a string such | |
19128 | as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker. | |
19129 | This variable is defined with @code{defvar}: | |
19130 | ||
19131 | @vindex Y-axis-tic | |
19132 | @smallexample | |
19133 | @group | |
19134 | (defvar Y-axis-tic " - " | |
19135 | "String that follows number in a Y axis label.") | |
19136 | @end group | |
19137 | @end smallexample | |
19138 | ||
19139 | The length of the Y label is the sum of the length of the Y axis tic | |
19140 | mark and the length of the number of the top of the graph. | |
19141 | ||
19142 | @smallexample | |
19143 | (length (concat (number-to-string height) Y-axis-tic))) | |
19144 | @end smallexample | |
19145 | ||
19146 | This value will be calculated by the @code{print-graph} function in | |
19147 | its varlist as @code{full-Y-label-width} and passed on. (Note that we | |
19148 | did not think to include this in the varlist when we first proposed it.) | |
19149 | ||
19150 | To make a complete vertical axis label, a tic mark is concatenated | |
19151 | with a number; and the two together may be preceded by one or more | |
19152 | spaces depending on how long the number is. The label consists of | |
19153 | three parts: the (optional) leading spaces, the number, and the tic | |
19154 | mark. The function is passed the value of the number for the specific | |
19155 | row, and the value of the width of the top line, which is calculated | |
19156 | (just once) by @code{print-graph}. | |
19157 | ||
19158 | @smallexample | |
19159 | @group | |
19160 | (defun Y-axis-element (number full-Y-label-width) | |
19161 | "Construct a NUMBERed label element. | |
19162 | A numbered element looks like this ` 5 - ', | |
19163 | and is padded as needed so all line up with | |
19164 | the element for the largest number." | |
19165 | @end group | |
19166 | @group | |
19167 | (let* ((leading-spaces | |
19168 | (- full-Y-label-width | |
19169 | (length | |
19170 | (concat (number-to-string number) | |
19171 | Y-axis-tic))))) | |
19172 | @end group | |
19173 | @group | |
19174 | (concat | |
19175 | (make-string leading-spaces ? ) | |
19176 | (number-to-string number) | |
19177 | Y-axis-tic))) | |
19178 | @end group | |
19179 | @end smallexample | |
19180 | ||
19181 | The @code{Y-axis-element} function concatenates together the leading | |
19182 | spaces, if any; the number, as a string; and the tic mark. | |
19183 | ||
19184 | To figure out how many leading spaces the label will need, the | |
19185 | function subtracts the actual length of the label---the length of the | |
19186 | number plus the length of the tic mark---from the desired label width. | |
19187 | ||
19188 | @findex make-string | |
19189 | Blank spaces are inserted using the @code{make-string} function. This | |
19190 | function takes two arguments: the first tells it how long the string | |
19191 | will be and the second is a symbol for the character to insert, in a | |
19192 | special format. The format is a question mark followed by a blank | |
19193 | space, like this, @samp{? }. @xref{Character Type, , Character Type, | |
19194 | elisp, The GNU Emacs Lisp Reference Manual}, for a description of the | |
19195 | syntax for characters. | |
19196 | ||
19197 | The @code{number-to-string} function is used in the concatenation | |
19198 | expression, to convert the number to a string that is concatenated | |
19199 | with the leading spaces and the tic mark. | |
19200 | ||
19201 | @node Y-axis-column, print-Y-axis Penultimate, Y Axis Element, print-Y-axis | |
19202 | @appendixsubsec Create a Y Axis Column | |
19203 | ||
19204 | The preceding functions provide all the tools needed to construct a | |
19205 | function that generates a list of numbered and blank strings to insert | |
19206 | as the label for the vertical axis: | |
19207 | ||
19208 | @findex Y-axis-column | |
19209 | @smallexample | |
19210 | @group | |
19211 | (defun Y-axis-column (height width-of-label) | |
19212 | "Construct list of Y axis labels and blank strings. | |
19213 | For HEIGHT of line above base and WIDTH-OF-LABEL." | |
19214 | (let (Y-axis) | |
19215 | @group | |
19216 | @end group | |
19217 | (while (> height 1) | |
19218 | (if (zerop (% height Y-axis-label-spacing)) | |
19219 | ;; @r{Insert label.} | |
19220 | (setq Y-axis | |
19221 | (cons | |
19222 | (Y-axis-element height width-of-label) | |
19223 | Y-axis)) | |
19224 | @group | |
19225 | @end group | |
19226 | ;; @r{Else, insert blanks.} | |
19227 | (setq Y-axis | |
19228 | (cons | |
19229 | (make-string width-of-label ? ) | |
19230 | Y-axis))) | |
19231 | (setq height (1- height))) | |
19232 | ;; @r{Insert base line.} | |
19233 | (setq Y-axis | |
19234 | (cons (Y-axis-element 1 width-of-label) Y-axis)) | |
19235 | (nreverse Y-axis))) | |
19236 | @end group | |
19237 | @end smallexample | |
19238 | ||
19239 | In this function, we start with the value of @code{height} and | |
19240 | repetitively subtract one from its value. After each subtraction, we | |
19241 | test to see whether the value is an integral multiple of the | |
19242 | @code{Y-axis-label-spacing}. If it is, we construct a numbered label | |
19243 | using the @code{Y-axis-element} function; if not, we construct a | |
19244 | blank label using the @code{make-string} function. The base line | |
19245 | consists of the number one followed by a tic mark. | |
19246 | ||
19247 | @node print-Y-axis Penultimate, , Y-axis-column, print-Y-axis | |
19248 | @appendixsubsec The Not Quite Final Version of @code{print-Y-axis} | |
19249 | ||
19250 | The list constructed by the @code{Y-axis-column} function is passed to | |
19251 | the @code{print-Y-axis} function, which inserts the list as a column. | |
19252 | ||
19253 | @findex print-Y-axis | |
19254 | @smallexample | |
19255 | @group | |
19256 | (defun print-Y-axis (height full-Y-label-width) | |
19257 | "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH. | |
19258 | Height must be the maximum height of the graph. | |
19259 | Full width is the width of the highest label element." | |
19260 | ;; Value of height and full-Y-label-width | |
19261 | ;; are passed by `print-graph'. | |
19262 | @end group | |
19263 | @group | |
19264 | (let ((start (point))) | |
19265 | (insert-rectangle | |
19266 | (Y-axis-column height full-Y-label-width)) | |
19267 | ;; @r{Place point ready for inserting graph.} | |
19268 | (goto-char start) | |
19269 | ;; @r{Move point forward by value of} full-Y-label-width | |
19270 | (forward-char full-Y-label-width))) | |
19271 | @end group | |
19272 | @end smallexample | |
19273 | ||
19274 | The @code{print-Y-axis} uses the @code{insert-rectangle} function to | |
19275 | insert the Y axis labels created by the @code{Y-axis-column} function. | |
19276 | In addition, it places point at the correct position for printing the body of | |
19277 | the graph. | |
19278 | ||
19279 | You can test @code{print-Y-axis}: | |
19280 | ||
19281 | @enumerate | |
19282 | @item | |
19283 | Install | |
19284 | ||
19285 | @smallexample | |
19286 | @group | |
19287 | Y-axis-label-spacing | |
19288 | Y-axis-tic | |
19289 | Y-axis-element | |
19290 | Y-axis-column | |
19291 | print-Y-axis | |
19292 | @end group | |
19293 | @end smallexample | |
19294 | ||
19295 | @item | |
19296 | Copy the following expression: | |
19297 | ||
19298 | @smallexample | |
19299 | (print-Y-axis 12 5) | |
19300 | @end smallexample | |
19301 | ||
19302 | @item | |
19303 | Switch to the @file{*scratch*} buffer and place the cursor where you | |
19304 | want the axis labels to start. | |
19305 | ||
19306 | @item | |
19307 | Type @kbd{M-:} (@code{eval-expression}). | |
19308 | ||
19309 | @item | |
19310 | Yank the @code{graph-body-print} expression into the minibuffer | |
19311 | with @kbd{C-y} (@code{yank)}. | |
19312 | ||
19313 | @item | |
19314 | Press @key{RET} to evaluate the expression. | |
19315 | @end enumerate | |
19316 | ||
19317 | Emacs will print labels vertically, the top one being | |
19318 | @w{@samp{10 -@w{ }}}. (The @code{print-graph} function | |
19319 | will pass the value of @code{height-of-top-line}, which | |
19320 | in this case would end up as 15.) | |
19321 | ||
19322 | @node print-X-axis, Print Whole Graph, print-Y-axis, Full Graph | |
19323 | @appendixsec The @code{print-X-axis} Function | |
19324 | @cindex Axis, print horizontal | |
19325 | @cindex X axis printing | |
19326 | @cindex Print horizontal axis | |
19327 | @cindex Horizontal axis printing | |
19328 | ||
19329 | X axis labels are much like Y axis labels, except that the tics are on a | |
19330 | line above the numbers. Labels should look like this: | |
19331 | ||
19332 | @smallexample | |
19333 | @group | |
19334 | | | | | | |
19335 | 1 5 10 15 | |
19336 | @end group | |
19337 | @end smallexample | |
19338 | ||
19339 | The first tic is under the first column of the graph and is preceded by | |
19340 | several blank spaces. These spaces provide room in rows above for the Y | |
19341 | axis labels. The second, third, fourth, and subsequent tics are all | |
19342 | spaced equally, according to the value of @code{X-axis-label-spacing}. | |
19343 | ||
19344 | The second row of the X axis consists of numbers, preceded by several | |
19345 | blank spaces and also separated according to the value of the variable | |
19346 | @code{X-axis-label-spacing}. | |
19347 | ||
19348 | The value of the variable @code{X-axis-label-spacing} should itself be | |
19349 | measured in units of @code{symbol-width}, since you may want to change | |
19350 | the width of the symbols that you are using to print the body of the | |
19351 | graph without changing the ways the graph is labelled. | |
19352 | ||
19353 | @menu | |
19354 | * Similarities differences:: Much like @code{print-Y-axis}, but not exactly. | |
19355 | * X Axis Tic Marks:: Create tic marks for the horizontal axis. | |
19356 | @end menu | |
19357 | ||
19358 | @node Similarities differences, X Axis Tic Marks, print-X-axis, print-X-axis | |
19359 | @ifnottex | |
19360 | @unnumberedsubsec Similarities and differences | |
19361 | @end ifnottex | |
19362 | ||
19363 | The @code{print-X-axis} function is constructed in more or less the | |
19364 | same fashion as the @code{print-Y-axis} function except that it has | |
19365 | two lines: the line of tic marks and the numbers. We will write a | |
19366 | separate function to print each line and then combine them within the | |
19367 | @code{print-X-axis} function. | |
19368 | ||
19369 | This is a three step process: | |
19370 | ||
19371 | @enumerate | |
19372 | @item | |
19373 | Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}. | |
19374 | ||
19375 | @item | |
19376 | Write a function to print the X numbers, @code{print-X-axis-numbered-line}. | |
19377 | ||
19378 | @item | |
19379 | Write a function to print both lines, the @code{print-X-axis} function, | |
19380 | using @code{print-X-axis-tic-line} and | |
19381 | @code{print-X-axis-numbered-line}. | |
19382 | @end enumerate | |
19383 | ||
19384 | @node X Axis Tic Marks, , Similarities differences, print-X-axis | |
19385 | @appendixsubsec X Axis Tic Marks | |
19386 | ||
19387 | The first function should print the X axis tic marks. We must specify | |
19388 | the tic marks themselves and their spacing: | |
19389 | ||
19390 | @smallexample | |
19391 | @group | |
19392 | (defvar X-axis-label-spacing | |
19393 | (if (boundp 'graph-blank) | |
19394 | (* 5 (length graph-blank)) 5) | |
19395 | "Number of units from one X axis label to next.") | |
19396 | @end group | |
19397 | @end smallexample | |
19398 | ||
19399 | @noindent | |
19400 | (Note that the value of @code{graph-blank} is set by another | |
19401 | @code{defvar}. The @code{boundp} predicate checks whether it has | |
19402 | already been set; @code{boundp} returns @code{nil} if it has not. | |
19403 | If @code{graph-blank} were unbound and we did not use this conditional | |
19404 | construction, in GNU Emacs 21, we would enter the debugger and see an | |
19405 | error message saying | |
19406 | @samp{@w{Debugger entered--Lisp error:} @w{(void-variable graph-blank)}}.) | |
19407 | ||
19408 | @need 1200 | |
19409 | Here is the @code{defvar} for @code{X-axis-tic-symbol}: | |
19410 | ||
19411 | @smallexample | |
19412 | @group | |
19413 | (defvar X-axis-tic-symbol "|" | |
19414 | "String to insert to point to a column in X axis.") | |
19415 | @end group | |
19416 | @end smallexample | |
19417 | ||
19418 | @need 1250 | |
19419 | The goal is to make a line that looks like this: | |
19420 | ||
19421 | @smallexample | |
19422 | | | | | | |
19423 | @end smallexample | |
19424 | ||
19425 | The first tic is indented so that it is under the first column, which is | |
19426 | indented to provide space for the Y axis labels. | |
19427 | ||
19428 | A tic element consists of the blank spaces that stretch from one tic to | |
19429 | the next plus a tic symbol. The number of blanks is determined by the | |
19430 | width of the tic symbol and the @code{X-axis-label-spacing}. | |
19431 | ||
19432 | @need 1250 | |
19433 | The code looks like this: | |
19434 | ||
19435 | @smallexample | |
19436 | @group | |
19437 | ;;; X-axis-tic-element | |
19438 | @dots{} | |
19439 | (concat | |
19440 | (make-string | |
19441 | ;; @r{Make a string of blanks.} | |
19442 | (- (* symbol-width X-axis-label-spacing) | |
19443 | (length X-axis-tic-symbol)) | |
19444 | ? ) | |
19445 | ;; @r{Concatenate blanks with tic symbol.} | |
19446 | X-axis-tic-symbol) | |
19447 | @dots{} | |
19448 | @end group | |
19449 | @end smallexample | |
19450 | ||
19451 | Next, we determine how many blanks are needed to indent the first tic | |
19452 | mark to the first column of the graph. This uses the value of | |
19453 | @code{full-Y-label-width} passed it by the @code{print-graph} function. | |
19454 | ||
19455 | @need 1250 | |
19456 | The code to make @code{X-axis-leading-spaces} | |
19457 | looks like this: | |
19458 | ||
19459 | @smallexample | |
19460 | @group | |
19461 | ;; X-axis-leading-spaces | |
19462 | @dots{} | |
19463 | (make-string full-Y-label-width ? ) | |
19464 | @dots{} | |
19465 | @end group | |
19466 | @end smallexample | |
19467 | ||
19468 | We also need to determine the length of the horizontal axis, which is | |
19469 | the length of the numbers list, and the number of tics in the horizontal | |
19470 | axis: | |
19471 | ||
19472 | @smallexample | |
19473 | @group | |
19474 | ;; X-length | |
19475 | @dots{} | |
19476 | (length numbers-list) | |
19477 | @end group | |
19478 | ||
19479 | @group | |
19480 | ;; tic-width | |
19481 | @dots{} | |
19482 | (* symbol-width X-axis-label-spacing) | |
19483 | @end group | |
19484 | ||
19485 | @group | |
19486 | ;; number-of-X-tics | |
19487 | (if (zerop (% (X-length tic-width))) | |
19488 | (/ (X-length tic-width)) | |
19489 | (1+ (/ (X-length tic-width)))) | |
19490 | @end group | |
19491 | @end smallexample | |
19492 | ||
19493 | @need 1250 | |
19494 | All this leads us directly to the function for printing the X axis tic line: | |
19495 | ||
19496 | @findex print-X-axis-tic-line | |
19497 | @smallexample | |
19498 | @group | |
19499 | (defun print-X-axis-tic-line | |
19500 | (number-of-X-tics X-axis-leading-spaces X-axis-tic-element) | |
19501 | "Print tics for X axis." | |
19502 | (insert X-axis-leading-spaces) | |
19503 | (insert X-axis-tic-symbol) ; @r{Under first column.} | |
19504 | @end group | |
19505 | @group | |
19506 | ;; @r{Insert second tic in the right spot.} | |
19507 | (insert (concat | |
19508 | (make-string | |
19509 | (- (* symbol-width X-axis-label-spacing) | |
19510 | ;; @r{Insert white space up to second tic symbol.} | |
19511 | (* 2 (length X-axis-tic-symbol))) | |
19512 | ? ) | |
19513 | X-axis-tic-symbol)) | |
19514 | @end group | |
19515 | @group | |
19516 | ;; @r{Insert remaining tics.} | |
19517 | (while (> number-of-X-tics 1) | |
19518 | (insert X-axis-tic-element) | |
19519 | (setq number-of-X-tics (1- number-of-X-tics)))) | |
19520 | @end group | |
19521 | @end smallexample | |
19522 | ||
19523 | The line of numbers is equally straightforward: | |
19524 | ||
19525 | @need 1250 | |
19526 | First, we create a numbered element with blank spaces before each number: | |
19527 | ||
19528 | @findex X-axis-element | |
19529 | @smallexample | |
19530 | @group | |
19531 | (defun X-axis-element (number) | |
19532 | "Construct a numbered X axis element." | |
19533 | (let ((leading-spaces | |
19534 | (- (* symbol-width X-axis-label-spacing) | |
19535 | (length (number-to-string number))))) | |
19536 | (concat (make-string leading-spaces ? ) | |
19537 | (number-to-string number)))) | |
19538 | @end group | |
19539 | @end smallexample | |
19540 | ||
19541 | Next, we create the function to print the numbered line, starting with | |
19542 | the number ``1'' under the first column: | |
19543 | ||
19544 | @findex print-X-axis-numbered-line | |
19545 | @smallexample | |
19546 | @group | |
19547 | (defun print-X-axis-numbered-line | |
19548 | (number-of-X-tics X-axis-leading-spaces) | |
19549 | "Print line of X-axis numbers" | |
19550 | (let ((number X-axis-label-spacing)) | |
19551 | (insert X-axis-leading-spaces) | |
19552 | (insert "1") | |
19553 | @end group | |
19554 | @group | |
19555 | (insert (concat | |
19556 | (make-string | |
19557 | ;; @r{Insert white space up to next number.} | |
19558 | (- (* symbol-width X-axis-label-spacing) 2) | |
19559 | ? ) | |
19560 | (number-to-string number))) | |
19561 | @end group | |
19562 | @group | |
19563 | ;; @r{Insert remaining numbers.} | |
19564 | (setq number (+ number X-axis-label-spacing)) | |
19565 | (while (> number-of-X-tics 1) | |
19566 | (insert (X-axis-element number)) | |
19567 | (setq number (+ number X-axis-label-spacing)) | |
19568 | (setq number-of-X-tics (1- number-of-X-tics))))) | |
19569 | @end group | |
19570 | @end smallexample | |
19571 | ||
19572 | Finally, we need to write the @code{print-X-axis} that uses | |
19573 | @code{print-X-axis-tic-line} and | |
19574 | @code{print-X-axis-numbered-line}. | |
19575 | ||
19576 | The function must determine the local values of the variables used by both | |
19577 | @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and | |
19578 | then it must call them. Also, it must print the carriage return that | |
19579 | separates the two lines. | |
19580 | ||
19581 | The function consists of a varlist that specifies five local variables, | |
19582 | and calls to each of the two line printing functions: | |
19583 | ||
19584 | @findex print-X-axis | |
19585 | @smallexample | |
19586 | @group | |
19587 | (defun print-X-axis (numbers-list) | |
19588 | "Print X axis labels to length of NUMBERS-LIST." | |
19589 | (let* ((leading-spaces | |
19590 | (make-string full-Y-label-width ? )) | |
19591 | @end group | |
19592 | @group | |
19593 | ;; symbol-width @r{is provided by} graph-body-print | |
19594 | (tic-width (* symbol-width X-axis-label-spacing)) | |
19595 | (X-length (length numbers-list)) | |
19596 | @end group | |
19597 | @group | |
19598 | (X-tic | |
19599 | (concat | |
19600 | (make-string | |
19601 | @end group | |
19602 | @group | |
19603 | ;; @r{Make a string of blanks.} | |
19604 | (- (* symbol-width X-axis-label-spacing) | |
19605 | (length X-axis-tic-symbol)) | |
19606 | ? ) | |
19607 | @end group | |
19608 | @group | |
19609 | ;; @r{Concatenate blanks with tic symbol.} | |
19610 | X-axis-tic-symbol)) | |
19611 | @end group | |
19612 | @group | |
19613 | (tic-number | |
19614 | (if (zerop (% X-length tic-width)) | |
19615 | (/ X-length tic-width) | |
19616 | (1+ (/ X-length tic-width))))) | |
19617 | @end group | |
19618 | @group | |
19619 | (print-X-axis-tic-line tic-number leading-spaces X-tic) | |
19620 | (insert "\n") | |
19621 | (print-X-axis-numbered-line tic-number leading-spaces))) | |
19622 | @end group | |
19623 | @end smallexample | |
19624 | ||
19625 | @need 1250 | |
19626 | You can test @code{print-X-axis}: | |
19627 | ||
19628 | @enumerate | |
19629 | @item | |
19630 | Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing}, | |
19631 | @code{print-X-axis-tic-line}, as well as @code{X-axis-element}, | |
19632 | @code{print-X-axis-numbered-line}, and @code{print-X-axis}. | |
19633 | ||
19634 | @item | |
19635 | Copy the following expression: | |
19636 | ||
19637 | @smallexample | |
19638 | @group | |
19639 | (progn | |
19640 | (let ((full-Y-label-width 5) | |
19641 | (symbol-width 1)) | |
19642 | (print-X-axis | |
19643 | '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16)))) | |
19644 | @end group | |
19645 | @end smallexample | |
19646 | ||
19647 | @item | |
19648 | Switch to the @file{*scratch*} buffer and place the cursor where you | |
19649 | want the axis labels to start. | |
19650 | ||
19651 | @item | |
19652 | Type @kbd{M-:} (@code{eval-expression}). | |
19653 | ||
19654 | @item | |
19655 | Yank the test expression into the minibuffer | |
19656 | with @kbd{C-y} (@code{yank)}. | |
19657 | ||
19658 | @item | |
19659 | Press @key{RET} to evaluate the expression. | |
19660 | @end enumerate | |
19661 | ||
19662 | @need 1250 | |
19663 | Emacs will print the horizontal axis like this: | |
19664 | ||
19665 | @smallexample | |
19666 | @group | |
19667 | | | | | | | |
19668 | 1 5 10 15 20 | |
19669 | @end group | |
19670 | @end smallexample | |
19671 | ||
19672 | @node Print Whole Graph, , print-X-axis, Full Graph | |
19673 | @appendixsec Printing the Whole Graph | |
19674 | @cindex Printing the whole graph | |
19675 | @cindex Whole graph printing | |
19676 | @cindex Graph, printing all | |
19677 | ||
19678 | Now we are nearly ready to print the whole graph. | |
19679 | ||
19680 | The function to print the graph with the proper labels follows the | |
19681 | outline we created earlier (@pxref{Full Graph, , A Graph with Labelled | |
19682 | Axes}), but with additions. | |
19683 | ||
19684 | @need 1250 | |
19685 | Here is the outline: | |
19686 | ||
19687 | @smallexample | |
19688 | @group | |
19689 | (defun print-graph (numbers-list) | |
19690 | "@var{documentation}@dots{}" | |
19691 | (let ((height @dots{} | |
19692 | @dots{})) | |
19693 | @end group | |
19694 | @group | |
19695 | (print-Y-axis height @dots{} ) | |
19696 | (graph-body-print numbers-list) | |
19697 | (print-X-axis @dots{} ))) | |
19698 | @end group | |
19699 | @end smallexample | |
19700 | ||
19701 | @menu | |
19702 | * The final version:: A few changes. | |
19703 | * Test print-graph:: Run a short test. | |
19704 | * Graphing words in defuns:: Executing the final code. | |
19705 | * lambda:: How to write an anonymous function. | |
19706 | * mapcar:: Apply a function to elements of a list. | |
19707 | * Another Bug:: Yet another bug @dots{} most insidious. | |
19708 | * Final printed graph:: The graph itself! | |
19709 | @end menu | |
19710 | ||
19711 | @node The final version, Test print-graph, Print Whole Graph, Print Whole Graph | |
19712 | @ifnottex | |
19713 | @unnumberedsubsec Changes for the Final Version | |
19714 | @end ifnottex | |
19715 | ||
19716 | The final version is different from what we planned in two ways: | |
19717 | first, it contains additional values calculated once in the varlist; | |
19718 | second, it carries an option to specify the labels' increment per row. | |
19719 | This latter feature turns out to be essential; otherwise, a graph may | |
19720 | have more rows than fit on a display or on a sheet of paper. | |
19721 | ||
19722 | @need 1500 | |
19723 | This new feature requires a change to the @code{Y-axis-column} | |
19724 | function, to add @code{vertical-step} to it. The function looks like | |
19725 | this: | |
19726 | ||
19727 | @findex Y-axis-column @r{Final version.} | |
19728 | @smallexample | |
19729 | @group | |
19730 | ;;; @r{Final version.} | |
19731 | (defun Y-axis-column | |
19732 | (height width-of-label &optional vertical-step) | |
19733 | "Construct list of labels for Y axis. | |
19734 | HEIGHT is maximum height of graph. | |
19735 | WIDTH-OF-LABEL is maximum width of label. | |
19736 | VERTICAL-STEP, an option, is a positive integer | |
19737 | that specifies how much a Y axis label increments | |
19738 | for each line. For example, a step of 5 means | |
19739 | that each line is five units of the graph." | |
19740 | @end group | |
19741 | @group | |
19742 | (let (Y-axis | |
19743 | (number-per-line (or vertical-step 1))) | |
19744 | (while (> height 1) | |
19745 | (if (zerop (% height Y-axis-label-spacing)) | |
19746 | @end group | |
19747 | @group | |
19748 | ;; @r{Insert label.} | |
19749 | (setq Y-axis | |
19750 | (cons | |
19751 | (Y-axis-element | |
19752 | (* height number-per-line) | |
19753 | width-of-label) | |
19754 | Y-axis)) | |
19755 | @end group | |
19756 | @group | |
19757 | ;; @r{Else, insert blanks.} | |
19758 | (setq Y-axis | |
19759 | (cons | |
19760 | (make-string width-of-label ? ) | |
19761 | Y-axis))) | |
19762 | (setq height (1- height))) | |
19763 | @end group | |
19764 | @group | |
19765 | ;; @r{Insert base line.} | |
19766 | (setq Y-axis (cons (Y-axis-element | |
19767 | (or vertical-step 1) | |
19768 | width-of-label) | |
19769 | Y-axis)) | |
19770 | (nreverse Y-axis))) | |
19771 | @end group | |
19772 | @end smallexample | |
19773 | ||
19774 | The values for the maximum height of graph and the width of a symbol | |
19775 | are computed by @code{print-graph} in its @code{let} expression; so | |
19776 | @code{graph-body-print} must be changed to accept them. | |
19777 | ||
19778 | @findex graph-body-print @r{Final version.} | |
19779 | @smallexample | |
19780 | @group | |
19781 | ;;; @r{Final version.} | |
19782 | (defun graph-body-print (numbers-list height symbol-width) | |
19783 | "Print a bar graph of the NUMBERS-LIST. | |
19784 | The numbers-list consists of the Y-axis values. | |
19785 | HEIGHT is maximum height of graph. | |
19786 | SYMBOL-WIDTH is number of each column." | |
19787 | @end group | |
19788 | @group | |
19789 | (let (from-position) | |
19790 | (while numbers-list | |
19791 | (setq from-position (point)) | |
19792 | (insert-rectangle | |
19793 | (column-of-graph height (car numbers-list))) | |
19794 | (goto-char from-position) | |
19795 | (forward-char symbol-width) | |
19796 | @end group | |
19797 | @group | |
19798 | ;; @r{Draw graph column by column.} | |
19799 | (sit-for 0) | |
19800 | (setq numbers-list (cdr numbers-list))) | |
19801 | ;; @r{Place point for X axis labels.} | |
19802 | (forward-line height) | |
19803 | (insert "\n"))) | |
19804 | @end group | |
19805 | @end smallexample | |
19806 | ||
19807 | @need 1250 | |
19808 | Finally, the code for the @code{print-graph} function: | |
19809 | ||
19810 | @findex print-graph @r{Final version.} | |
19811 | @smallexample | |
19812 | @group | |
19813 | ;;; @r{Final version.} | |
19814 | (defun print-graph | |
19815 | (numbers-list &optional vertical-step) | |
19816 | "Print labelled bar graph of the NUMBERS-LIST. | |
19817 | The numbers-list consists of the Y-axis values. | |
19818 | @end group | |
19819 | ||
19820 | @group | |
19821 | Optionally, VERTICAL-STEP, a positive integer, | |
19822 | specifies how much a Y axis label increments for | |
19823 | each line. For example, a step of 5 means that | |
19824 | each row is five units." | |
19825 | @end group | |
19826 | @group | |
19827 | (let* ((symbol-width (length graph-blank)) | |
19828 | ;; @code{height} @r{is both the largest number} | |
19829 | ;; @r{and the number with the most digits.} | |
19830 | (height (apply 'max numbers-list)) | |
19831 | @end group | |
19832 | @group | |
19833 | (height-of-top-line | |
19834 | (if (zerop (% height Y-axis-label-spacing)) | |
19835 | height | |
19836 | ;; @r{else} | |
19837 | (* (1+ (/ height Y-axis-label-spacing)) | |
19838 | Y-axis-label-spacing))) | |
19839 | @end group | |
19840 | @group | |
19841 | (vertical-step (or vertical-step 1)) | |
19842 | (full-Y-label-width | |
19843 | (length | |
19844 | @end group | |
19845 | @group | |
19846 | (concat | |
19847 | (number-to-string | |
19848 | (* height-of-top-line vertical-step)) | |
19849 | Y-axis-tic)))) | |
19850 | @end group | |
19851 | ||
19852 | @group | |
19853 | (print-Y-axis | |
19854 | height-of-top-line full-Y-label-width vertical-step) | |
19855 | @end group | |
19856 | @group | |
19857 | (graph-body-print | |
19858 | numbers-list height-of-top-line symbol-width) | |
19859 | (print-X-axis numbers-list))) | |
19860 | @end group | |
19861 | @end smallexample | |
19862 | ||
19863 | @node Test print-graph, Graphing words in defuns, The final version, Print Whole Graph | |
19864 | @appendixsubsec Testing @code{print-graph} | |
19865 | ||
19866 | @need 1250 | |
19867 | We can test the @code{print-graph} function with a short list of numbers: | |
19868 | ||
19869 | @enumerate | |
19870 | @item | |
19871 | Install the final versions of @code{Y-axis-column}, | |
19872 | @code{graph-body-print}, and @code{print-graph} (in addition to the | |
19873 | rest of the code.) | |
19874 | ||
19875 | @item | |
19876 | Copy the following expression: | |
19877 | ||
19878 | @smallexample | |
19879 | (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1)) | |
19880 | @end smallexample | |
19881 | ||
19882 | @item | |
19883 | Switch to the @file{*scratch*} buffer and place the cursor where you | |
19884 | want the axis labels to start. | |
19885 | ||
19886 | @item | |
19887 | Type @kbd{M-:} (@code{eval-expression}). | |
19888 | ||
19889 | @item | |
19890 | Yank the test expression into the minibuffer | |
19891 | with @kbd{C-y} (@code{yank)}. | |
19892 | ||
19893 | @item | |
19894 | Press @key{RET} to evaluate the expression. | |
19895 | @end enumerate | |
19896 | ||
19897 | @need 1250 | |
19898 | Emacs will print a graph that looks like this: | |
19899 | ||
19900 | @smallexample | |
19901 | @group | |
19902 | 10 - | |
19903 | ||
19904 | ||
19905 | * | |
19906 | ** * | |
19907 | 5 - **** * | |
19908 | **** *** | |
19909 | * ********* | |
19910 | ************ | |
19911 | 1 - ************* | |
19912 | ||
19913 | | | | | | |
19914 | 1 5 10 15 | |
19915 | @end group | |
19916 | @end smallexample | |
19917 | ||
19918 | On the other hand, if you pass @code{print-graph} a | |
19919 | @code{vertical-step} value of 2, by evaluating this expression: | |
19920 | ||
19921 | @smallexample | |
19922 | (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2) | |
19923 | @end smallexample | |
19924 | ||
19925 | @need 1250 | |
19926 | @noindent | |
19927 | The graph looks like this: | |
19928 | ||
19929 | @smallexample | |
19930 | @group | |
19931 | 20 - | |
19932 | ||
19933 | ||
19934 | * | |
19935 | ** * | |
19936 | 10 - **** * | |
19937 | **** *** | |
19938 | * ********* | |
19939 | ************ | |
19940 | 2 - ************* | |
19941 | ||
19942 | | | | | | |
19943 | 1 5 10 15 | |
19944 | @end group | |
19945 | @end smallexample | |
19946 | ||
19947 | @noindent | |
19948 | (A question: is the `2' on the bottom of the vertical axis a bug or a | |
19949 | feature? If you think it is a bug, and should be a `1' instead, (or | |
19950 | even a `0'), you can modify the sources.) | |
19951 | ||
19952 | @node Graphing words in defuns, lambda, Test print-graph, Print Whole Graph | |
19953 | @appendixsubsec Graphing Numbers of Words and Symbols | |
19954 | ||
19955 | Now for the graph for which all this code was written: a graph that | |
19956 | shows how many function definitions contain fewer than 10 words and | |
19957 | symbols, how many contain between 10 and 19 words and symbols, how | |
19958 | many contain between 20 and 29 words and symbols, and so on. | |
19959 | ||
19960 | This is a multi-step process. First make sure you have loaded all the | |
19961 | requisite code. | |
19962 | ||
19963 | @need 1500 | |
19964 | It is a good idea to reset the value of @code{top-of-ranges} in case | |
19965 | you have set it to some different value. You can evaluate the | |
19966 | following: | |
19967 | ||
19968 | @smallexample | |
19969 | @group | |
19970 | (setq top-of-ranges | |
19971 | '(10 20 30 40 50 | |
19972 | 60 70 80 90 100 | |
19973 | 110 120 130 140 150 | |
19974 | 160 170 180 190 200 | |
19975 | 210 220 230 240 250 | |
19976 | 260 270 280 290 300) | |
19977 | @end group | |
19978 | @end smallexample | |
19979 | ||
19980 | @noindent | |
19981 | Next create a list of the number of words and symbols in each range. | |
19982 | ||
19983 | @need 1500 | |
19984 | @noindent | |
19985 | Evaluate the following: | |
19986 | ||
19987 | @smallexample | |
19988 | @group | |
19989 | (setq list-for-graph | |
19990 | (defuns-per-range | |
19991 | (sort | |
19992 | (recursive-lengths-list-many-files | |
19993 | (directory-files "/usr/local/emacs/lisp" | |
19994 | t ".+el$")) | |
19995 | '<) | |
19996 | top-of-ranges)) | |
19997 | @end group | |
19998 | @end smallexample | |
19999 | ||
20000 | @noindent | |
20001 | On my machine, this takes about an hour. It looks though 303 Lisp | |
20002 | files in my copy of Emacs version 19.23. After all that computing, | |
20003 | the @code{list-for-graph} has this value: | |
20004 | ||
20005 | @smallexample | |
20006 | @group | |
20007 | (537 1027 955 785 594 483 349 292 224 199 166 120 116 99 | |
20008 | 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220) | |
20009 | @end group | |
20010 | @end smallexample | |
20011 | ||
20012 | @noindent | |
20013 | This means that my copy of Emacs has 537 function definitions with | |
20014 | fewer than 10 words or symbols in them, 1,027 function definitions | |
20015 | with 10 to 19 words or symbols in them, 955 function definitions with | |
20016 | 20 to 29 words or symbols in them, and so on. | |
20017 | ||
20018 | Clearly, just by looking at this list we can see that most function | |
20019 | definitions contain ten to thirty words and symbols. | |
20020 | ||
20021 | Now for printing. We do @emph{not} want to print a graph that is | |
20022 | 1,030 lines high @dots{} Instead, we should print a graph that is | |
20023 | fewer than twenty-five lines high. A graph that height can be | |
20024 | displayed on almost any monitor, and easily printed on a sheet of paper. | |
20025 | ||
20026 | This means that each value in @code{list-for-graph} must be reduced to | |
20027 | one-fiftieth its present value. | |
20028 | ||
20029 | Here is a short function to do just that, using two functions we have | |
20030 | not yet seen, @code{mapcar} and @code{lambda}. | |
20031 | ||
20032 | @smallexample | |
20033 | @group | |
20034 | (defun one-fiftieth (full-range) | |
20035 | "Return list, each number one-fiftieth of previous." | |
20036 | (mapcar '(lambda (arg) (/ arg 50)) full-range)) | |
20037 | @end group | |
20038 | @end smallexample | |
20039 | ||
20040 | @node lambda, mapcar, Graphing words in defuns, Print Whole Graph | |
20041 | @appendixsubsec A @code{lambda} Expression: Useful Anonymity | |
20042 | @cindex Anonymous function | |
20043 | @findex lambda | |
20044 | ||
20045 | @code{lambda} is the symbol for an anonymous function, a function | |
20046 | without a name. Every time you use an anonymous function, you need to | |
20047 | include its whole body. | |
20048 | ||
20049 | @need 1250 | |
20050 | @noindent | |
20051 | Thus, | |
20052 | ||
20053 | @smallexample | |
20054 | (lambda (arg) (/ arg 50)) | |
20055 | @end smallexample | |
20056 | ||
20057 | @noindent | |
20058 | is a function definition that says `return the value resulting from | |
20059 | dividing whatever is passed to me as @code{arg} by 50'. | |
20060 | ||
20061 | Earlier, for example, we had a function @code{multiply-by-seven}; it | |
20062 | multiplied its argument by 7. This function is similar, except it | |
20063 | divides its argument by 50; and, it has no name. The anonymous | |
20064 | equivalent of @code{multiply-by-seven} is: | |
20065 | ||
20066 | @smallexample | |
20067 | (lambda (number) (* 7 number)) | |
20068 | @end smallexample | |
20069 | ||
20070 | @noindent | |
20071 | (@xref{defun, , The @code{defun} Special Form}.) | |
20072 | ||
20073 | @need 1250 | |
20074 | @noindent | |
20075 | If we want to multiply 3 by 7, we can write: | |
20076 | ||
20077 | @c !!! Clear print-postscript-figures if the computer formatting this | |
20078 | @c document is too small and cannot handle all the diagrams and figures. | |
20079 | @c clear print-postscript-figures | |
20080 | @c set print-postscript-figures | |
20081 | @c lambda example diagram #1 | |
20082 | @ifnottex | |
20083 | @smallexample | |
20084 | @group | |
20085 | (multiply-by-seven 3) | |
20086 | \_______________/ ^ | |
20087 | | | | |
20088 | function argument | |
20089 | @end group | |
20090 | @end smallexample | |
20091 | @end ifnottex | |
20092 | @ifset print-postscript-figures | |
20093 | @sp 1 | |
20094 | @tex | |
20095 | @image{lambda-1} | |
20096 | %%%% old method of including an image | |
20097 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
20098 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}} | |
20099 | % \catcode`\@=0 % | |
20100 | @end tex | |
20101 | @sp 1 | |
20102 | @end ifset | |
20103 | @ifclear print-postscript-figures | |
20104 | @iftex | |
20105 | @smallexample | |
20106 | @group | |
20107 | (multiply-by-seven 3) | |
20108 | \_______________/ ^ | |
20109 | | | | |
20110 | function argument | |
20111 | @end group | |
20112 | @end smallexample | |
20113 | @end iftex | |
20114 | @end ifclear | |
20115 | ||
20116 | @noindent | |
20117 | This expression returns 21. | |
20118 | ||
20119 | @need 1250 | |
20120 | @noindent | |
20121 | Similarly, we can write: | |
20122 | ||
20123 | @c lambda example diagram #2 | |
20124 | @ifnottex | |
20125 | @smallexample | |
20126 | @group | |
20127 | ((lambda (number) (* 7 number)) 3) | |
20128 | \____________________________/ ^ | |
20129 | | | | |
20130 | anonymous function argument | |
20131 | @end group | |
20132 | @end smallexample | |
20133 | @end ifnottex | |
20134 | @ifset print-postscript-figures | |
20135 | @sp 1 | |
20136 | @tex | |
20137 | @image{lambda-2} | |
20138 | %%%% old method of including an image | |
20139 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
20140 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}} | |
20141 | % \catcode`\@=0 % | |
20142 | @end tex | |
20143 | @sp 1 | |
20144 | @end ifset | |
20145 | @ifclear print-postscript-figures | |
20146 | @iftex | |
20147 | @smallexample | |
20148 | @group | |
20149 | ((lambda (number) (* 7 number)) 3) | |
20150 | \____________________________/ ^ | |
20151 | | | | |
20152 | anonymous function argument | |
20153 | @end group | |
20154 | @end smallexample | |
20155 | @end iftex | |
20156 | @end ifclear | |
20157 | ||
20158 | @need 1250 | |
20159 | @noindent | |
20160 | If we want to divide 100 by 50, we can write: | |
20161 | ||
20162 | @c lambda example diagram #3 | |
20163 | @ifnottex | |
20164 | @smallexample | |
20165 | @group | |
20166 | ((lambda (arg) (/ arg 50)) 100) | |
20167 | \______________________/ \_/ | |
20168 | | | | |
20169 | anonymous function argument | |
20170 | @end group | |
20171 | @end smallexample | |
20172 | @end ifnottex | |
20173 | @ifset print-postscript-figures | |
20174 | @sp 1 | |
20175 | @tex | |
20176 | @image{lambda-3} | |
20177 | %%%% old method of including an image | |
20178 | % \input /usr/local/lib/tex/inputs/psfig.tex | |
20179 | % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}} | |
20180 | % \catcode`\@=0 % | |
20181 | @end tex | |
20182 | @sp 1 | |
20183 | @end ifset | |
20184 | @ifclear print-postscript-figures | |
20185 | @iftex | |
20186 | @smallexample | |
20187 | @group | |
20188 | ((lambda (arg) (/ arg 50)) 100) | |
20189 | \______________________/ \_/ | |
20190 | | | | |
20191 | anonymous function argument | |
20192 | @end group | |
20193 | @end smallexample | |
20194 | @end iftex | |
20195 | @end ifclear | |
20196 | ||
20197 | @noindent | |
20198 | This expression returns 2. The 100 is passed to the function, which | |
20199 | divides that number by 50. | |
20200 | ||
20201 | @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs | |
20202 | Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda | |
20203 | expressions derive from the Lambda Calculus. | |
20204 | ||
20205 | @node mapcar, Another Bug, lambda, Print Whole Graph | |
20206 | @appendixsubsec The @code{mapcar} Function | |
20207 | @findex mapcar | |
20208 | ||
20209 | @code{mapcar} is a function that calls its first argument with each | |
20210 | element of its second argument, in turn. The second argument must be | |
20211 | a sequence. | |
20212 | ||
20213 | The @samp{map} part of the name comes from the mathematical phrase, | |
20214 | `mapping over a domain', meaning to apply a function to each of the | |
20215 | elements in a domain. The mathematical phrase is based on the | |
20216 | metaphor of a surveyor walking, one step at a time, over an area he is | |
20217 | mapping. And @samp{car}, of course, comes from the Lisp notion of the | |
20218 | first of a list. | |
20219 | ||
20220 | @need 1250 | |
20221 | @noindent | |
20222 | For example, | |
20223 | ||
20224 | @smallexample | |
20225 | @group | |
20226 | (mapcar '1+ '(2 4 6)) | |
20227 | @result{} (3 5 7) | |
20228 | @end group | |
20229 | @end smallexample | |
20230 | ||
20231 | @noindent | |
20232 | The function @code{1+} which adds one to its argument, is executed on | |
20233 | @emph{each} element of the list, and a new list is returned. | |
20234 | ||
20235 | Contrast this with @code{apply}, which applies its first argument to | |
20236 | all the remaining. | |
20237 | (@xref{Readying a Graph, , Readying a Graph}, for a explanation of | |
20238 | @code{apply}.) | |
20239 | ||
20240 | @need 1250 | |
20241 | In the definition of @code{one-fiftieth}, the first argument is the | |
20242 | anonymous function: | |
20243 | ||
20244 | @smallexample | |
20245 | (lambda (arg) (/ arg 50)) | |
20246 | @end smallexample | |
20247 | ||
20248 | @noindent | |
20249 | and the second argument is @code{full-range}, which will be bound to | |
20250 | @code{list-for-graph}. | |
20251 | ||
20252 | @need 1250 | |
20253 | The whole expression looks like this: | |
20254 | ||
20255 | @smallexample | |
20256 | (mapcar '(lambda (arg) (/ arg 50)) full-range)) | |
20257 | @end smallexample | |
20258 | ||
20259 | @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs | |
20260 | Lisp Reference Manual}, for more about @code{mapcar}. | |
20261 | ||
20262 | Using the @code{one-fiftieth} function, we can generate a list in | |
20263 | which each element is one-fiftieth the size of the corresponding | |
20264 | element in @code{list-for-graph}. | |
20265 | ||
20266 | @smallexample | |
20267 | @group | |
20268 | (setq fiftieth-list-for-graph | |
20269 | (one-fiftieth list-for-graph)) | |
20270 | @end group | |
20271 | @end smallexample | |
20272 | ||
20273 | @need 1250 | |
20274 | The resulting list looks like this: | |
20275 | ||
20276 | @smallexample | |
20277 | @group | |
20278 | (10 20 19 15 11 9 6 5 4 3 3 2 2 | |
20279 | 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4) | |
20280 | @end group | |
20281 | @end smallexample | |
20282 | ||
20283 | @noindent | |
20284 | This, we are almost ready to print! (We also notice the loss of | |
20285 | information: many of the higher ranges are 0, meaning that fewer than | |
20286 | 50 defuns had that many words or symbols---but not necessarily meaning | |
20287 | that none had that many words or symbols.) | |
20288 | ||
20289 | @node Another Bug, Final printed graph, mapcar, Print Whole Graph | |
20290 | @appendixsubsec Another Bug @dots{} Most Insidious | |
20291 | @cindex Bug, most insidious type | |
20292 | @cindex Insidious type of bug | |
20293 | ||
20294 | I said `almost ready to print'! Of course, there is a bug in the | |
20295 | @code{print-graph} function @dots{} It has a @code{vertical-step} | |
20296 | option, but not a @code{horizontal-step} option. The | |
20297 | @code{top-of-range} scale goes from 10 to 300 by tens. But the | |
20298 | @code{print-graph} function will print only by ones. | |
20299 | ||
20300 | This is a classic example of what some consider the most insidious | |
20301 | type of bug, the bug of omission. This is not the kind of bug you can | |
20302 | find by studying the code, for it is not in the code; it is an omitted | |
20303 | feature. Your best actions are to try your program early and often; | |
20304 | and try to arrange, as much as you can, to write code that is easy to | |
20305 | understand and easy to change. Try to be aware, whenever you can, | |
20306 | that whatever you have written, @emph{will} be rewritten, if not soon, | |
20307 | eventually. A hard maxim to follow. | |
20308 | ||
20309 | It is the @code{print-X-axis-numbered-line} function that needs the | |
20310 | work; and then the @code{print-X-axis} and the @code{print-graph} | |
20311 | functions need to be adapted. Not much needs to be done; there is one | |
20312 | nicety: the numbers ought to line up under the tic marks. This takes | |
20313 | a little thought. | |
20314 | ||
20315 | @need 1250 | |
20316 | Here is the corrected @code{print-X-axis-numbered-line}: | |
20317 | ||
20318 | @smallexample | |
20319 | @group | |
20320 | (defun print-X-axis-numbered-line | |
20321 | (number-of-X-tics X-axis-leading-spaces | |
20322 | &optional horizontal-step) | |
20323 | "Print line of X-axis numbers" | |
20324 | (let ((number X-axis-label-spacing) | |
20325 | (horizontal-step (or horizontal-step 1))) | |
20326 | @end group | |
20327 | @group | |
20328 | (insert X-axis-leading-spaces) | |
20329 | ;; @r{Delete extra leading spaces.} | |
20330 | (delete-char | |
20331 | (- (1- | |
20332 | (length (number-to-string horizontal-step))))) | |
20333 | (insert (concat | |
20334 | (make-string | |
20335 | @end group | |
20336 | @group | |
20337 | ;; @r{Insert white space.} | |
20338 | (- (* symbol-width | |
20339 | X-axis-label-spacing) | |
20340 | (1- | |
20341 | (length | |
20342 | (number-to-string horizontal-step))) | |
20343 | 2) | |
20344 | ? ) | |
20345 | (number-to-string | |
20346 | (* number horizontal-step)))) | |
20347 | @end group | |
20348 | @group | |
20349 | ;; @r{Insert remaining numbers.} | |
20350 | (setq number (+ number X-axis-label-spacing)) | |
20351 | (while (> number-of-X-tics 1) | |
20352 | (insert (X-axis-element | |
20353 | (* number horizontal-step))) | |
20354 | (setq number (+ number X-axis-label-spacing)) | |
20355 | (setq number-of-X-tics (1- number-of-X-tics))))) | |
20356 | @end group | |
20357 | @end smallexample | |
20358 | ||
20359 | @need 1500 | |
20360 | If you are reading this in Info, you can see the new versions of | |
20361 | @code{print-X-axis} @code{print-graph} and evaluate them. If you are | |
20362 | reading this in a printed book, you can see the changed lines here | |
20363 | (the full text is too much to print). | |
20364 | ||
20365 | @iftex | |
20366 | @smallexample | |
20367 | @group | |
20368 | (defun print-X-axis (numbers-list horizontal-step) | |
20369 | @dots{} | |
20370 | (print-X-axis-numbered-line | |
20371 | tic-number leading-spaces horizontal-step)) | |
20372 | @end group | |
20373 | @end smallexample | |
20374 | ||
20375 | @smallexample | |
20376 | @group | |
20377 | (defun print-graph | |
20378 | (numbers-list | |
20379 | &optional vertical-step horizontal-step) | |
20380 | @dots{} | |
20381 | (print-X-axis numbers-list horizontal-step)) | |
20382 | @end group | |
20383 | @end smallexample | |
20384 | @end iftex | |
20385 | ||
20386 | @ifnottex | |
20387 | @smallexample | |
20388 | @group | |
20389 | (defun print-X-axis (numbers-list horizontal-step) | |
20390 | "Print X axis labels to length of NUMBERS-LIST. | |
20391 | Optionally, HORIZONTAL-STEP, a positive integer, | |
20392 | specifies how much an X axis label increments for | |
20393 | each column." | |
20394 | @end group | |
20395 | @group | |
20396 | ;; Value of symbol-width and full-Y-label-width | |
20397 | ;; are passed by `print-graph'. | |
20398 | (let* ((leading-spaces | |
20399 | (make-string full-Y-label-width ? )) | |
20400 | ;; symbol-width @r{is provided by} graph-body-print | |
20401 | (tic-width (* symbol-width X-axis-label-spacing)) | |
20402 | (X-length (length numbers-list)) | |
20403 | @end group | |
20404 | @group | |
20405 | (X-tic | |
20406 | (concat | |
20407 | (make-string | |
20408 | ;; @r{Make a string of blanks.} | |
20409 | (- (* symbol-width X-axis-label-spacing) | |
20410 | (length X-axis-tic-symbol)) | |
20411 | ? ) | |
20412 | @end group | |
20413 | @group | |
20414 | ;; @r{Concatenate blanks with tic symbol.} | |
20415 | X-axis-tic-symbol)) | |
20416 | (tic-number | |
20417 | (if (zerop (% X-length tic-width)) | |
20418 | (/ X-length tic-width) | |
20419 | (1+ (/ X-length tic-width))))) | |
20420 | @end group | |
20421 | ||
20422 | @group | |
20423 | (print-X-axis-tic-line | |
20424 | tic-number leading-spaces X-tic) | |
20425 | (insert "\n") | |
20426 | (print-X-axis-numbered-line | |
20427 | tic-number leading-spaces horizontal-step))) | |
20428 | @end group | |
20429 | @end smallexample | |
20430 | ||
20431 | @smallexample | |
20432 | @group | |
20433 | (defun print-graph | |
20434 | (numbers-list &optional vertical-step horizontal-step) | |
20435 | "Print labelled bar graph of the NUMBERS-LIST. | |
20436 | The numbers-list consists of the Y-axis values. | |
20437 | @end group | |
20438 | ||
20439 | @group | |
20440 | Optionally, VERTICAL-STEP, a positive integer, | |
20441 | specifies how much a Y axis label increments for | |
20442 | each line. For example, a step of 5 means that | |
20443 | each row is five units. | |
20444 | @end group | |
20445 | ||
20446 | @group | |
20447 | Optionally, HORIZONTAL-STEP, a positive integer, | |
20448 | specifies how much an X axis label increments for | |
20449 | each column." | |
20450 | (let* ((symbol-width (length graph-blank)) | |
20451 | ;; @code{height} @r{is both the largest number} | |
20452 | ;; @r{and the number with the most digits.} | |
20453 | (height (apply 'max numbers-list)) | |
20454 | @end group | |
20455 | @group | |
20456 | (height-of-top-line | |
20457 | (if (zerop (% height Y-axis-label-spacing)) | |
20458 | height | |
20459 | ;; @r{else} | |
20460 | (* (1+ (/ height Y-axis-label-spacing)) | |
20461 | Y-axis-label-spacing))) | |
20462 | @end group | |
20463 | @group | |
20464 | (vertical-step (or vertical-step 1)) | |
20465 | (full-Y-label-width | |
20466 | (length | |
20467 | (concat | |
20468 | (number-to-string | |
20469 | (* height-of-top-line vertical-step)) | |
20470 | Y-axis-tic)))) | |
20471 | @end group | |
20472 | @group | |
20473 | (print-Y-axis | |
20474 | height-of-top-line full-Y-label-width vertical-step) | |
20475 | (graph-body-print | |
20476 | numbers-list height-of-top-line symbol-width) | |
20477 | (print-X-axis numbers-list horizontal-step))) | |
20478 | @end group | |
20479 | @end smallexample | |
20480 | @end ifnottex | |
20481 | ||
20482 | @ignore | |
20483 | Graphing Definitions Re-listed | |
20484 | ||
20485 | @need 1250 | |
20486 | Here are all the graphing definitions in their final form: | |
20487 | ||
20488 | @smallexample | |
20489 | @group | |
20490 | (defvar top-of-ranges | |
20491 | '(10 20 30 40 50 | |
20492 | 60 70 80 90 100 | |
20493 | 110 120 130 140 150 | |
20494 | 160 170 180 190 200 | |
20495 | 210 220 230 240 250) | |
20496 | "List specifying ranges for `defuns-per-range'.") | |
20497 | @end group | |
20498 | ||
20499 | @group | |
20500 | (defvar graph-symbol "*" | |
20501 | "String used as symbol in graph, usually an asterisk.") | |
20502 | @end group | |
20503 | ||
20504 | @group | |
20505 | (defvar graph-blank " " | |
20506 | "String used as blank in graph, usually a blank space. | |
20507 | graph-blank must be the same number of columns wide | |
20508 | as graph-symbol.") | |
20509 | @end group | |
20510 | ||
20511 | @group | |
20512 | (defvar Y-axis-tic " - " | |
20513 | "String that follows number in a Y axis label.") | |
20514 | @end group | |
20515 | ||
20516 | @group | |
20517 | (defvar Y-axis-label-spacing 5 | |
20518 | "Number of lines from one Y axis label to next.") | |
20519 | @end group | |
20520 | ||
20521 | @group | |
20522 | (defvar X-axis-tic-symbol "|" | |
20523 | "String to insert to point to a column in X axis.") | |
20524 | @end group | |
20525 | ||
20526 | @group | |
20527 | (defvar X-axis-label-spacing | |
20528 | (if (boundp 'graph-blank) | |
20529 | (* 5 (length graph-blank)) 5) | |
20530 | "Number of units from one X axis label to next.") | |
20531 | @end group | |
20532 | @end smallexample | |
20533 | ||
20534 | @smallexample | |
20535 | @group | |
20536 | (defun count-words-in-defun () | |
20537 | "Return the number of words and symbols in a defun." | |
20538 | (beginning-of-defun) | |
20539 | (let ((count 0) | |
20540 | (end (save-excursion (end-of-defun) (point)))) | |
20541 | @end group | |
20542 | ||
20543 | @group | |
20544 | (while | |
20545 | (and (< (point) end) | |
20546 | (re-search-forward | |
20547 | "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" | |
20548 | end t)) | |
20549 | (setq count (1+ count))) | |
20550 | count)) | |
20551 | @end group | |
20552 | @end smallexample | |
20553 | ||
20554 | @smallexample | |
20555 | @group | |
20556 | (defun lengths-list-file (filename) | |
20557 | "Return list of definitions' lengths within FILE. | |
20558 | The returned list is a list of numbers. | |
20559 | Each number is the number of words or | |
20560 | symbols in one function definition." | |
20561 | @end group | |
20562 | ||
20563 | @group | |
20564 | (message "Working on `%s' ... " filename) | |
20565 | (save-excursion | |
20566 | (let ((buffer (find-file-noselect filename)) | |
20567 | (lengths-list)) | |
20568 | (set-buffer buffer) | |
20569 | (setq buffer-read-only t) | |
20570 | (widen) | |
20571 | (goto-char (point-min)) | |
20572 | @end group | |
20573 | ||
20574 | @group | |
20575 | (while (re-search-forward "^(defun" nil t) | |
20576 | (setq lengths-list | |
20577 | (cons (count-words-in-defun) lengths-list))) | |
20578 | (kill-buffer buffer) | |
20579 | lengths-list))) | |
20580 | @end group | |
20581 | @end smallexample | |
20582 | ||
20583 | @smallexample | |
20584 | @group | |
20585 | (defun lengths-list-many-files (list-of-files) | |
20586 | "Return list of lengths of defuns in LIST-OF-FILES." | |
20587 | (let (lengths-list) | |
20588 | ;;; @r{true-or-false-test} | |
20589 | (while list-of-files | |
20590 | (setq lengths-list | |
20591 | (append | |
20592 | lengths-list | |
20593 | @end group | |
20594 | @group | |
20595 | ;;; @r{Generate a lengths' list.} | |
20596 | (lengths-list-file | |
20597 | (expand-file-name (car list-of-files))))) | |
20598 | ;;; @r{Make files' list shorter.} | |
20599 | (setq list-of-files (cdr list-of-files))) | |
20600 | ;;; @r{Return final value of lengths' list.} | |
20601 | lengths-list)) | |
20602 | @end group | |
20603 | @end smallexample | |
20604 | ||
20605 | @smallexample | |
20606 | @group | |
20607 | (defun defuns-per-range (sorted-lengths top-of-ranges) | |
20608 | "SORTED-LENGTHS defuns in each TOP-OF-RANGES range." | |
20609 | (let ((top-of-range (car top-of-ranges)) | |
20610 | (number-within-range 0) | |
20611 | defuns-per-range-list) | |
20612 | @end group | |
20613 | ||
20614 | @group | |
20615 | ;; @r{Outer loop.} | |
20616 | (while top-of-ranges | |
20617 | ||
20618 | ;; @r{Inner loop.} | |
20619 | (while (and | |
20620 | ;; @r{Need number for numeric test.} | |
20621 | (car sorted-lengths) | |
20622 | (< (car sorted-lengths) top-of-range)) | |
20623 | ||
20624 | ;; @r{Count number of definitions within current range.} | |
20625 | (setq number-within-range (1+ number-within-range)) | |
20626 | (setq sorted-lengths (cdr sorted-lengths))) | |
20627 | @end group | |
20628 | ||
20629 | @group | |
20630 | ;; @r{Exit inner loop but remain within outer loop.} | |
20631 | ||
20632 | (setq defuns-per-range-list | |
20633 | (cons number-within-range defuns-per-range-list)) | |
20634 | (setq number-within-range 0) ; @r{Reset count to zero.} | |
20635 | ||
20636 | ;; @r{Move to next range.} | |
20637 | (setq top-of-ranges (cdr top-of-ranges)) | |
20638 | ;; @r{Specify next top of range value.} | |
20639 | (setq top-of-range (car top-of-ranges))) | |
20640 | @end group | |
20641 | ||
20642 | @group | |
20643 | ;; @r{Exit outer loop and count the number of defuns larger than} | |
20644 | ;; @r{ the largest top-of-range value.} | |
20645 | (setq defuns-per-range-list | |
20646 | (cons | |
20647 | (length sorted-lengths) | |
20648 | defuns-per-range-list)) | |
20649 | ||
20650 | ;; @r{Return a list of the number of definitions within each range,} | |
20651 | ;; @r{ smallest to largest.} | |
20652 | (nreverse defuns-per-range-list))) | |
20653 | @end group | |
20654 | @end smallexample | |
20655 | ||
20656 | @smallexample | |
20657 | @group | |
20658 | (defun column-of-graph (max-graph-height actual-height) | |
20659 | "Return list of MAX-GRAPH-HEIGHT strings; | |
20660 | ACTUAL-HEIGHT are graph-symbols. | |
20661 | The graph-symbols are contiguous entries at the end | |
20662 | of the list. | |
20663 | The list will be inserted as one column of a graph. | |
20664 | The strings are either graph-blank or graph-symbol." | |
20665 | @end group | |
20666 | ||
20667 | @group | |
20668 | (let ((insert-list nil) | |
20669 | (number-of-top-blanks | |
20670 | (- max-graph-height actual-height))) | |
20671 | ||
20672 | ;; @r{Fill in @code{graph-symbols}.} | |
20673 | (while (> actual-height 0) | |
20674 | (setq insert-list (cons graph-symbol insert-list)) | |
20675 | (setq actual-height (1- actual-height))) | |
20676 | @end group | |
20677 | ||
20678 | @group | |
20679 | ;; @r{Fill in @code{graph-blanks}.} | |
20680 | (while (> number-of-top-blanks 0) | |
20681 | (setq insert-list (cons graph-blank insert-list)) | |
20682 | (setq number-of-top-blanks | |
20683 | (1- number-of-top-blanks))) | |
20684 | ||
20685 | ;; @r{Return whole list.} | |
20686 | insert-list)) | |
20687 | @end group | |
20688 | @end smallexample | |
20689 | ||
20690 | @smallexample | |
20691 | @group | |
20692 | (defun Y-axis-element (number full-Y-label-width) | |
20693 | "Construct a NUMBERed label element. | |
20694 | A numbered element looks like this ` 5 - ', | |
20695 | and is padded as needed so all line up with | |
20696 | the element for the largest number." | |
20697 | @end group | |
20698 | @group | |
20699 | (let* ((leading-spaces | |
20700 | (- full-Y-label-width | |
20701 | (length | |
20702 | (concat (number-to-string number) | |
20703 | Y-axis-tic))))) | |
20704 | @end group | |
20705 | @group | |
20706 | (concat | |
20707 | (make-string leading-spaces ? ) | |
20708 | (number-to-string number) | |
20709 | Y-axis-tic))) | |
20710 | @end group | |
20711 | @end smallexample | |
20712 | ||
20713 | @smallexample | |
20714 | @group | |
20715 | (defun print-Y-axis | |
20716 | (height full-Y-label-width &optional vertical-step) | |
20717 | "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH. | |
20718 | Height must be the maximum height of the graph. | |
20719 | Full width is the width of the highest label element. | |
20720 | Optionally, print according to VERTICAL-STEP." | |
20721 | @end group | |
20722 | @group | |
20723 | ;; Value of height and full-Y-label-width | |
20724 | ;; are passed by `print-graph'. | |
20725 | (let ((start (point))) | |
20726 | (insert-rectangle | |
20727 | (Y-axis-column height full-Y-label-width vertical-step)) | |
20728 | @end group | |
20729 | @group | |
20730 | ;; @r{Place point ready for inserting graph.} | |
20731 | (goto-char start) | |
20732 | ;; @r{Move point forward by value of} full-Y-label-width | |
20733 | (forward-char full-Y-label-width))) | |
20734 | @end group | |
20735 | @end smallexample | |
20736 | ||
20737 | @smallexample | |
20738 | @group | |
20739 | (defun print-X-axis-tic-line | |
20740 | (number-of-X-tics X-axis-leading-spaces X-axis-tic-element) | |
20741 | "Print tics for X axis." | |
20742 | (insert X-axis-leading-spaces) | |
20743 | (insert X-axis-tic-symbol) ; @r{Under first column.} | |
20744 | @end group | |
20745 | @group | |
20746 | ;; @r{Insert second tic in the right spot.} | |
20747 | (insert (concat | |
20748 | (make-string | |
20749 | (- (* symbol-width X-axis-label-spacing) | |
20750 | ;; @r{Insert white space up to second tic symbol.} | |
20751 | (* 2 (length X-axis-tic-symbol))) | |
20752 | ? ) | |
20753 | X-axis-tic-symbol)) | |
20754 | @end group | |
20755 | @group | |
20756 | ;; @r{Insert remaining tics.} | |
20757 | (while (> number-of-X-tics 1) | |
20758 | (insert X-axis-tic-element) | |
20759 | (setq number-of-X-tics (1- number-of-X-tics)))) | |
20760 | @end group | |
20761 | @end smallexample | |
20762 | ||
20763 | @smallexample | |
20764 | @group | |
20765 | (defun X-axis-element (number) | |
20766 | "Construct a numbered X axis element." | |
20767 | (let ((leading-spaces | |
20768 | (- (* symbol-width X-axis-label-spacing) | |
20769 | (length (number-to-string number))))) | |
20770 | (concat (make-string leading-spaces ? ) | |
20771 | (number-to-string number)))) | |
20772 | @end group | |
20773 | @end smallexample | |
20774 | ||
20775 | @smallexample | |
20776 | @group | |
20777 | (defun graph-body-print (numbers-list height symbol-width) | |
20778 | "Print a bar graph of the NUMBERS-LIST. | |
20779 | The numbers-list consists of the Y-axis values. | |
20780 | HEIGHT is maximum height of graph. | |
20781 | SYMBOL-WIDTH is number of each column." | |
20782 | @end group | |
20783 | @group | |
20784 | (let (from-position) | |
20785 | (while numbers-list | |
20786 | (setq from-position (point)) | |
20787 | (insert-rectangle | |
20788 | (column-of-graph height (car numbers-list))) | |
20789 | (goto-char from-position) | |
20790 | (forward-char symbol-width) | |
20791 | @end group | |
20792 | @group | |
20793 | ;; @r{Draw graph column by column.} | |
20794 | (sit-for 0) | |
20795 | (setq numbers-list (cdr numbers-list))) | |
20796 | ;; @r{Place point for X axis labels.} | |
20797 | (forward-line height) | |
20798 | (insert "\n"))) | |
20799 | @end group | |
20800 | @end smallexample | |
20801 | ||
20802 | @smallexample | |
20803 | @group | |
20804 | (defun Y-axis-column | |
20805 | (height width-of-label &optional vertical-step) | |
20806 | "Construct list of labels for Y axis. | |
20807 | HEIGHT is maximum height of graph. | |
20808 | WIDTH-OF-LABEL is maximum width of label. | |
20809 | @end group | |
20810 | @group | |
20811 | VERTICAL-STEP, an option, is a positive integer | |
20812 | that specifies how much a Y axis label increments | |
20813 | for each line. For example, a step of 5 means | |
20814 | that each line is five units of the graph." | |
20815 | (let (Y-axis | |
20816 | (number-per-line (or vertical-step 1))) | |
20817 | @end group | |
20818 | @group | |
20819 | (while (> height 1) | |
20820 | (if (zerop (% height Y-axis-label-spacing)) | |
20821 | ;; @r{Insert label.} | |
20822 | (setq Y-axis | |
20823 | (cons | |
20824 | (Y-axis-element | |
20825 | (* height number-per-line) | |
20826 | width-of-label) | |
20827 | Y-axis)) | |
20828 | @end group | |
20829 | @group | |
20830 | ;; @r{Else, insert blanks.} | |
20831 | (setq Y-axis | |
20832 | (cons | |
20833 | (make-string width-of-label ? ) | |
20834 | Y-axis))) | |
20835 | (setq height (1- height))) | |
20836 | @end group | |
20837 | @group | |
20838 | ;; @r{Insert base line.} | |
20839 | (setq Y-axis (cons (Y-axis-element | |
20840 | (or vertical-step 1) | |
20841 | width-of-label) | |
20842 | Y-axis)) | |
20843 | (nreverse Y-axis))) | |
20844 | @end group | |
20845 | @end smallexample | |
20846 | ||
20847 | @smallexample | |
20848 | @group | |
20849 | (defun print-X-axis-numbered-line | |
20850 | (number-of-X-tics X-axis-leading-spaces | |
20851 | &optional horizontal-step) | |
20852 | "Print line of X-axis numbers" | |
20853 | (let ((number X-axis-label-spacing) | |
20854 | (horizontal-step (or horizontal-step 1))) | |
20855 | @end group | |
20856 | @group | |
20857 | (insert X-axis-leading-spaces) | |
20858 | ;; line up number | |
20859 | (delete-char (- (1- (length (number-to-string horizontal-step))))) | |
20860 | (insert (concat | |
20861 | (make-string | |
20862 | ;; @r{Insert white space up to next number.} | |
20863 | (- (* symbol-width X-axis-label-spacing) | |
20864 | (1- (length (number-to-string horizontal-step))) | |
20865 | 2) | |
20866 | ? ) | |
20867 | (number-to-string (* number horizontal-step)))) | |
20868 | @end group | |
20869 | @group | |
20870 | ;; @r{Insert remaining numbers.} | |
20871 | (setq number (+ number X-axis-label-spacing)) | |
20872 | (while (> number-of-X-tics 1) | |
20873 | (insert (X-axis-element (* number horizontal-step))) | |
20874 | (setq number (+ number X-axis-label-spacing)) | |
20875 | (setq number-of-X-tics (1- number-of-X-tics))))) | |
20876 | @end group | |
20877 | @end smallexample | |
20878 | ||
20879 | @smallexample | |
20880 | @group | |
20881 | (defun print-X-axis (numbers-list horizontal-step) | |
20882 | "Print X axis labels to length of NUMBERS-LIST. | |
20883 | Optionally, HORIZONTAL-STEP, a positive integer, | |
20884 | specifies how much an X axis label increments for | |
20885 | each column." | |
20886 | @end group | |
20887 | @group | |
20888 | ;; Value of symbol-width and full-Y-label-width | |
20889 | ;; are passed by `print-graph'. | |
20890 | (let* ((leading-spaces | |
20891 | (make-string full-Y-label-width ? )) | |
20892 | ;; symbol-width @r{is provided by} graph-body-print | |
20893 | (tic-width (* symbol-width X-axis-label-spacing)) | |
20894 | (X-length (length numbers-list)) | |
20895 | @end group | |
20896 | @group | |
20897 | (X-tic | |
20898 | (concat | |
20899 | (make-string | |
20900 | ;; @r{Make a string of blanks.} | |
20901 | (- (* symbol-width X-axis-label-spacing) | |
20902 | (length X-axis-tic-symbol)) | |
20903 | ? ) | |
20904 | @end group | |
20905 | @group | |
20906 | ;; @r{Concatenate blanks with tic symbol.} | |
20907 | X-axis-tic-symbol)) | |
20908 | (tic-number | |
20909 | (if (zerop (% X-length tic-width)) | |
20910 | (/ X-length tic-width) | |
20911 | (1+ (/ X-length tic-width))))) | |
20912 | @end group | |
20913 | ||
20914 | @group | |
20915 | (print-X-axis-tic-line | |
20916 | tic-number leading-spaces X-tic) | |
20917 | (insert "\n") | |
20918 | (print-X-axis-numbered-line | |
20919 | tic-number leading-spaces horizontal-step))) | |
20920 | @end group | |
20921 | @end smallexample | |
20922 | ||
20923 | @smallexample | |
20924 | @group | |
20925 | (defun one-fiftieth (full-range) | |
20926 | "Return list, each number of which is 1/50th previous." | |
20927 | (mapcar '(lambda (arg) (/ arg 50)) full-range)) | |
20928 | @end group | |
20929 | @end smallexample | |
20930 | ||
20931 | @smallexample | |
20932 | @group | |
20933 | (defun print-graph | |
20934 | (numbers-list &optional vertical-step horizontal-step) | |
20935 | "Print labelled bar graph of the NUMBERS-LIST. | |
20936 | The numbers-list consists of the Y-axis values. | |
20937 | @end group | |
20938 | ||
20939 | @group | |
20940 | Optionally, VERTICAL-STEP, a positive integer, | |
20941 | specifies how much a Y axis label increments for | |
20942 | each line. For example, a step of 5 means that | |
20943 | each row is five units. | |
20944 | @end group | |
20945 | ||
20946 | @group | |
20947 | Optionally, HORIZONTAL-STEP, a positive integer, | |
20948 | specifies how much an X axis label increments for | |
20949 | each column." | |
20950 | (let* ((symbol-width (length graph-blank)) | |
20951 | ;; @code{height} @r{is both the largest number} | |
20952 | ;; @r{and the number with the most digits.} | |
20953 | (height (apply 'max numbers-list)) | |
20954 | @end group | |
20955 | @group | |
20956 | (height-of-top-line | |
20957 | (if (zerop (% height Y-axis-label-spacing)) | |
20958 | height | |
20959 | ;; @r{else} | |
20960 | (* (1+ (/ height Y-axis-label-spacing)) | |
20961 | Y-axis-label-spacing))) | |
20962 | @end group | |
20963 | @group | |
20964 | (vertical-step (or vertical-step 1)) | |
20965 | (full-Y-label-width | |
20966 | (length | |
20967 | (concat | |
20968 | (number-to-string | |
20969 | (* height-of-top-line vertical-step)) | |
20970 | Y-axis-tic)))) | |
20971 | @end group | |
20972 | @group | |
20973 | ||
20974 | (print-Y-axis | |
20975 | height-of-top-line full-Y-label-width vertical-step) | |
20976 | (graph-body-print | |
20977 | numbers-list height-of-top-line symbol-width) | |
20978 | (print-X-axis numbers-list horizontal-step))) | |
20979 | @end group | |
20980 | @end smallexample | |
20981 | @end ignore | |
20982 | ||
20983 | @page | |
20984 | @node Final printed graph, , Another Bug, Print Whole Graph | |
20985 | @appendixsubsec The Printed Graph | |
20986 | ||
20987 | When made and installed, you can call the @code{print-graph} command | |
20988 | like this: | |
20989 | ||
20990 | @smallexample | |
20991 | @group | |
20992 | (print-graph fiftieth-list-for-graph 50 10) | |
20993 | @end group | |
20994 | @end smallexample | |
20995 | ||
20996 | Here is the graph: | |
20997 | ||
20998 | @sp 2 | |
20999 | ||
21000 | @smallexample | |
21001 | @group | |
21002 | 1000 - * | |
21003 | ** | |
21004 | ** | |
21005 | ** | |
21006 | ** | |
21007 | 750 - *** | |
21008 | *** | |
21009 | *** | |
21010 | *** | |
21011 | **** | |
21012 | 500 - ***** | |
21013 | ****** | |
21014 | ****** | |
21015 | ****** | |
21016 | ******* | |
21017 | 250 - ******** | |
21018 | ********* * | |
21019 | *********** * | |
21020 | ************* * | |
21021 | 50 - ***************** * * | |
21022 | | | | | | | | | | |
21023 | 10 50 100 150 200 250 300 350 | |
21024 | @end group | |
21025 | @end smallexample | |
21026 | ||
21027 | @sp 2 | |
21028 | ||
21029 | The largest group of functions contain 10 -- 19 words and symbols each. | |
21030 | ||
21031 | @node GNU Free Documentation License, Index, Full Graph, Top | |
21032 | @appendix GNU Free Documentation License | |
21033 | ||
21034 | @cindex FDL, GNU Free Documentation License | |
21035 | @center Version 1.1, March 2000 | |
21036 | ||
21037 | @display | |
21038 | Copyright @copyright{} 2000 Free Software Foundation, Inc. | |
21039 | 59 Temple Place, Suite 330, Boston, MA 02111-1307, USA | |
21040 | ||
21041 | Everyone is permitted to copy and distribute verbatim copies | |
21042 | of this license document, but changing it is not allowed. | |
21043 | @end display | |
21044 | ||
21045 | @enumerate 0 | |
21046 | @item | |
21047 | PREAMBLE | |
21048 | ||
21049 | The purpose of this License is to make a manual, textbook, or other | |
21050 | written document @dfn{free} in the sense of freedom: to assure everyone | |
21051 | the effective freedom to copy and redistribute it, with or without | |
21052 | modifying it, either commercially or noncommercially. Secondarily, | |
21053 | this License preserves for the author and publisher a way to get | |
21054 | credit for their work, while not being considered responsible for | |
21055 | modifications made by others. | |
21056 | ||
21057 | This License is a kind of ``copyleft'', which means that derivative | |
21058 | works of the document must themselves be free in the same sense. It | |
21059 | complements the GNU General Public License, which is a copyleft | |
21060 | license designed for free software. | |
21061 | ||
21062 | We have designed this License in order to use it for manuals for free | |
21063 | software, because free software needs free documentation: a free | |
21064 | program should come with manuals providing the same freedoms that the | |
21065 | software does. But this License is not limited to software manuals; | |
21066 | it can be used for any textual work, regardless of subject matter or | |
21067 | whether it is published as a printed book. We recommend this License | |
21068 | principally for works whose purpose is instruction or reference. | |
21069 | ||
21070 | @item | |
21071 | APPLICABILITY AND DEFINITIONS | |
21072 | ||
21073 | This License applies to any manual or other work that contains a | |
21074 | notice placed by the copyright holder saying it can be distributed | |
21075 | under the terms of this License. The ``Document'', below, refers to any | |
21076 | such manual or work. Any member of the public is a licensee, and is | |
21077 | addressed as ``you''. | |
21078 | ||
21079 | A ``Modified Version'' of the Document means any work containing the | |
21080 | Document or a portion of it, either copied verbatim, or with | |
21081 | modifications and/or translated into another language. | |
21082 | ||
21083 | A ``Secondary Section'' is a named appendix or a front-matter section of | |
21084 | the Document that deals exclusively with the relationship of the | |
21085 | publishers or authors of the Document to the Document's overall subject | |
21086 | (or to related matters) and contains nothing that could fall directly | |
21087 | within that overall subject. (For example, if the Document is in part a | |
21088 | textbook of mathematics, a Secondary Section may not explain any | |
21089 | mathematics.) The relationship could be a matter of historical | |
21090 | connection with the subject or with related matters, or of legal, | |
21091 | commercial, philosophical, ethical or political position regarding | |
21092 | them. | |
21093 | ||
21094 | The ``Invariant Sections'' are certain Secondary Sections whose titles | |
21095 | are designated, as being those of Invariant Sections, in the notice | |
21096 | that says that the Document is released under this License. | |
21097 | ||
21098 | The ``Cover Texts'' are certain short passages of text that are listed, | |
21099 | as Front-Cover Texts or Back-Cover Texts, in the notice that says that | |
21100 | the Document is released under this License. | |
21101 | ||
21102 | A ``Transparent'' copy of the Document means a machine-readable copy, | |
21103 | represented in a format whose specification is available to the | |
21104 | general public, whose contents can be viewed and edited directly and | |
21105 | straightforwardly with generic text editors or (for images composed of | |
21106 | pixels) generic paint programs or (for drawings) some widely available | |
21107 | drawing editor, and that is suitable for input to text formatters or | |
21108 | for automatic translation to a variety of formats suitable for input | |
21109 | to text formatters. A copy made in an otherwise Transparent file | |
21110 | format whose markup has been designed to thwart or discourage | |
21111 | subsequent modification by readers is not Transparent. A copy that is | |
21112 | not ``Transparent'' is called ``Opaque''. | |
21113 | ||
21114 | Examples of suitable formats for Transparent copies include plain | |
21115 | @sc{ascii} without markup, Texinfo input format, La@TeX{} input format, | |
21116 | @acronym{SGML} or @acronym{XML} using a publicly available | |
21117 | @acronym{DTD}, and standard-conforming simple @acronym{HTML} designed | |
21118 | for human modification. Opaque formats include PostScript, | |
21119 | @acronym{PDF}, proprietary formats that can be read and edited only by | |
21120 | proprietary word processors, @acronym{SGML} or @acronym{XML} for which | |
21121 | the @acronym{DTD} and/or processing tools are not generally available, | |
21122 | and the machine-generated @acronym{HTML} produced by some word | |
21123 | processors for output purposes only. | |
21124 | ||
21125 | The ``Title Page'' means, for a printed book, the title page itself, | |
21126 | plus such following pages as are needed to hold, legibly, the material | |
21127 | this License requires to appear in the title page. For works in | |
21128 | formats which do not have any title page as such, ``Title Page'' means | |
21129 | the text near the most prominent appearance of the work's title, | |
21130 | preceding the beginning of the body of the text. | |
21131 | ||
21132 | @item | |
21133 | VERBATIM COPYING | |
21134 | ||
21135 | You may copy and distribute the Document in any medium, either | |
21136 | commercially or noncommercially, provided that this License, the | |
21137 | copyright notices, and the license notice saying this License applies | |
21138 | to the Document are reproduced in all copies, and that you add no other | |
21139 | conditions whatsoever to those of this License. You may not use | |
21140 | technical measures to obstruct or control the reading or further | |
21141 | copying of the copies you make or distribute. However, you may accept | |
21142 | compensation in exchange for copies. If you distribute a large enough | |
21143 | number of copies you must also follow the conditions in section 3. | |
21144 | ||
21145 | You may also lend copies, under the same conditions stated above, and | |
21146 | you may publicly display copies. | |
21147 | ||
21148 | @item | |
21149 | COPYING IN QUANTITY | |
21150 | ||
21151 | If you publish printed copies of the Document numbering more than 100, | |
21152 | and the Document's license notice requires Cover Texts, you must enclose | |
21153 | the copies in covers that carry, clearly and legibly, all these Cover | |
21154 | Texts: Front-Cover Texts on the front cover, and Back-Cover Texts on | |
21155 | the back cover. Both covers must also clearly and legibly identify | |
21156 | you as the publisher of these copies. The front cover must present | |
21157 | the full title with all words of the title equally prominent and | |
21158 | visible. You may add other material on the covers in addition. | |
21159 | Copying with changes limited to the covers, as long as they preserve | |
21160 | the title of the Document and satisfy these conditions, can be treated | |
21161 | as verbatim copying in other respects. | |
21162 | ||
21163 | If the required texts for either cover are too voluminous to fit | |
21164 | legibly, you should put the first ones listed (as many as fit | |
21165 | reasonably) on the actual cover, and continue the rest onto adjacent | |
21166 | pages. | |
21167 | ||
21168 | If you publish or distribute Opaque copies of the Document numbering | |
21169 | more than 100, you must either include a machine-readable Transparent | |
21170 | copy along with each Opaque copy, or state in or with each Opaque copy | |
21171 | a publicly-accessible computer-network location containing a complete | |
21172 | Transparent copy of the Document, free of added material, which the | |
21173 | general network-using public has access to download anonymously at no | |
21174 | charge using public-standard network protocols. If you use the latter | |
21175 | option, you must take reasonably prudent steps, when you begin | |
21176 | distribution of Opaque copies in quantity, to ensure that this | |
21177 | Transparent copy will remain thus accessible at the stated location | |
21178 | until at least one year after the last time you distribute an Opaque | |
21179 | copy (directly or through your agents or retailers) of that edition to | |
21180 | the public. | |
21181 | ||
21182 | It is requested, but not required, that you contact the authors of the | |
21183 | Document well before redistributing any large number of copies, to give | |
21184 | them a chance to provide you with an updated version of the Document. | |
21185 | ||
21186 | @item | |
21187 | MODIFICATIONS | |
21188 | ||
21189 | You may copy and distribute a Modified Version of the Document under | |
21190 | the conditions of sections 2 and 3 above, provided that you release | |
21191 | the Modified Version under precisely this License, with the Modified | |
21192 | Version filling the role of the Document, thus licensing distribution | |
21193 | and modification of the Modified Version to whoever possesses a copy | |
21194 | of it. In addition, you must do these things in the Modified Version: | |
21195 | ||
21196 | @enumerate A | |
21197 | @item | |
21198 | Use in the Title Page (and on the covers, if any) a title distinct | |
21199 | from that of the Document, and from those of previous versions | |
21200 | (which should, if there were any, be listed in the History section | |
21201 | of the Document). You may use the same title as a previous version | |
21202 | if the original publisher of that version gives permission. | |
21203 | ||
21204 | @item | |
21205 | List on the Title Page, as authors, one or more persons or entities | |
21206 | responsible for authorship of the modifications in the Modified | |
21207 | Version, together with at least five of the principal authors of the | |
21208 | Document (all of its principal authors, if it has less than five). | |
21209 | ||
21210 | @item | |
21211 | State on the Title page the name of the publisher of the | |
21212 | Modified Version, as the publisher. | |
21213 | ||
21214 | @item | |
21215 | Preserve all the copyright notices of the Document. | |
21216 | ||
21217 | @item | |
21218 | Add an appropriate copyright notice for your modifications | |
21219 | adjacent to the other copyright notices. | |
21220 | ||
21221 | @item | |
21222 | Include, immediately after the copyright notices, a license notice | |
21223 | giving the public permission to use the Modified Version under the | |
21224 | terms of this License, in the form shown in the Addendum below. | |
21225 | ||
21226 | @item | |
21227 | Preserve in that license notice the full lists of Invariant Sections | |
21228 | and required Cover Texts given in the Document's license notice. | |
21229 | ||
21230 | @item | |
21231 | Include an unaltered copy of this License. | |
21232 | ||
21233 | @item | |
21234 | Preserve the section entitled ``History'', and its title, and add to | |
21235 | it an item stating at least the title, year, new authors, and | |
21236 | publisher of the Modified Version as given on the Title Page. If | |
21237 | there is no section entitled ``History'' in the Document, create one | |
21238 | stating the title, year, authors, and publisher of the Document as | |
21239 | given on its Title Page, then add an item describing the Modified | |
21240 | Version as stated in the previous sentence. | |
21241 | ||
21242 | @item | |
21243 | Preserve the network location, if any, given in the Document for | |
21244 | public access to a Transparent copy of the Document, and likewise | |
21245 | the network locations given in the Document for previous versions | |
21246 | it was based on. These may be placed in the ``History'' section. | |
21247 | You may omit a network location for a work that was published at | |
21248 | least four years before the Document itself, or if the original | |
21249 | publisher of the version it refers to gives permission. | |
21250 | ||
21251 | @item | |
21252 | In any section entitled ``Acknowledgments'' or ``Dedications'', | |
21253 | preserve the section's title, and preserve in the section all the | |
21254 | substance and tone of each of the contributor acknowledgments | |
21255 | and/or dedications given therein. | |
21256 | ||
21257 | @item | |
21258 | Preserve all the Invariant Sections of the Document, | |
21259 | unaltered in their text and in their titles. Section numbers | |
21260 | or the equivalent are not considered part of the section titles. | |
21261 | ||
21262 | @item | |
21263 | Delete any section entitled ``Endorsements''. Such a section | |
21264 | may not be included in the Modified Version. | |
21265 | ||
21266 | @item | |
21267 | Do not retitle any existing section as ``Endorsements'' | |
21268 | or to conflict in title with any Invariant Section. | |
21269 | @end enumerate | |
21270 | ||
21271 | If the Modified Version includes new front-matter sections or | |
21272 | appendices that qualify as Secondary Sections and contain no material | |
21273 | copied from the Document, you may at your option designate some or all | |
21274 | of these sections as invariant. To do this, add their titles to the | |
21275 | list of Invariant Sections in the Modified Version's license notice. | |
21276 | These titles must be distinct from any other section titles. | |
21277 | ||
21278 | You may add a section entitled ``Endorsements'', provided it contains | |
21279 | nothing but endorsements of your Modified Version by various | |
21280 | parties---for example, statements of peer review or that the text has | |
21281 | been approved by an organization as the authoritative definition of a | |
21282 | standard. | |
21283 | ||
21284 | You may add a passage of up to five words as a Front-Cover Text, and a | |
21285 | passage of up to 25 words as a Back-Cover Text, to the end of the list | |
21286 | of Cover Texts in the Modified Version. Only one passage of | |
21287 | Front-Cover Text and one of Back-Cover Text may be added by (or | |
21288 | through arrangements made by) any one entity. If the Document already | |
21289 | includes a cover text for the same cover, previously added by you or | |
21290 | by arrangement made by the same entity you are acting on behalf of, | |
21291 | you may not add another; but you may replace the old one, on explicit | |
21292 | permission from the previous publisher that added the old one. | |
21293 | ||
21294 | The author(s) and publisher(s) of the Document do not by this License | |
21295 | give permission to use their names for publicity for or to assert or | |
21296 | imply endorsement of any Modified Version. | |
21297 | ||
21298 | @item | |
21299 | COMBINING DOCUMENTS | |
21300 | ||
21301 | You may combine the Document with other documents released under this | |
21302 | License, under the terms defined in section 4 above for modified | |
21303 | versions, provided that you include in the combination all of the | |
21304 | Invariant Sections of all of the original documents, unmodified, and | |
21305 | list them all as Invariant Sections of your combined work in its | |
21306 | license notice. | |
21307 | ||
21308 | The combined work need only contain one copy of this License, and | |
21309 | multiple identical Invariant Sections may be replaced with a single | |
21310 | copy. If there are multiple Invariant Sections with the same name but | |
21311 | different contents, make the title of each such section unique by | |
21312 | adding at the end of it, in parentheses, the name of the original | |
21313 | author or publisher of that section if known, or else a unique number. | |
21314 | Make the same adjustment to the section titles in the list of | |
21315 | Invariant Sections in the license notice of the combined work. | |
21316 | ||
21317 | In the combination, you must combine any sections entitled ``History'' | |
21318 | in the various original documents, forming one section entitled | |
21319 | ``History''; likewise combine any sections entitled ``Acknowledgments'', | |
21320 | and any sections entitled ``Dedications''. You must delete all sections | |
21321 | entitled ``Endorsements.'' | |
21322 | ||
21323 | @item | |
21324 | COLLECTIONS OF DOCUMENTS | |
21325 | ||
21326 | You may make a collection consisting of the Document and other documents | |
21327 | released under this License, and replace the individual copies of this | |
21328 | License in the various documents with a single copy that is included in | |
21329 | the collection, provided that you follow the rules of this License for | |
21330 | verbatim copying of each of the documents in all other respects. | |
21331 | ||
21332 | You may extract a single document from such a collection, and distribute | |
21333 | it individually under this License, provided you insert a copy of this | |
21334 | License into the extracted document, and follow this License in all | |
21335 | other respects regarding verbatim copying of that document. | |
21336 | ||
21337 | @item | |
21338 | AGGREGATION WITH INDEPENDENT WORKS | |
21339 | ||
21340 | A compilation of the Document or its derivatives with other separate | |
21341 | and independent documents or works, in or on a volume of a storage or | |
21342 | distribution medium, does not as a whole count as a Modified Version | |
21343 | of the Document, provided no compilation copyright is claimed for the | |
21344 | compilation. Such a compilation is called an ``aggregate'', and this | |
21345 | License does not apply to the other self-contained works thus compiled | |
21346 | with the Document, on account of their being thus compiled, if they | |
21347 | are not themselves derivative works of the Document. | |
21348 | ||
21349 | If the Cover Text requirement of section 3 is applicable to these | |
21350 | copies of the Document, then if the Document is less than one quarter | |
21351 | of the entire aggregate, the Document's Cover Texts may be placed on | |
21352 | covers that surround only the Document within the aggregate. | |
21353 | Otherwise they must appear on covers around the whole aggregate. | |
21354 | ||
21355 | @item | |
21356 | TRANSLATION | |
21357 | ||
21358 | Translation is considered a kind of modification, so you may | |
21359 | distribute translations of the Document under the terms of section 4. | |
21360 | Replacing Invariant Sections with translations requires special | |
21361 | permission from their copyright holders, but you may include | |
21362 | translations of some or all Invariant Sections in addition to the | |
21363 | original versions of these Invariant Sections. You may include a | |
21364 | translation of this License provided that you also include the | |
21365 | original English version of this License. In case of a disagreement | |
21366 | between the translation and the original English version of this | |
21367 | License, the original English version will prevail. | |
21368 | ||
21369 | @item | |
21370 | TERMINATION | |
21371 | ||
21372 | You may not copy, modify, sublicense, or distribute the Document except | |
21373 | as expressly provided for under this License. Any other attempt to | |
21374 | copy, modify, sublicense or distribute the Document is void, and will | |
21375 | automatically terminate your rights under this License. However, | |
21376 | parties who have received copies, or rights, from you under this | |
21377 | License will not have their licenses terminated so long as such | |
21378 | parties remain in full compliance. | |
21379 | ||
21380 | @item | |
21381 | FUTURE REVISIONS OF THIS LICENSE | |
21382 | ||
21383 | The Free Software Foundation may publish new, revised versions | |
21384 | of the GNU Free Documentation License from time to time. Such new | |
21385 | versions will be similar in spirit to the present version, but may | |
21386 | differ in detail to address new problems or concerns. See | |
21387 | @uref{http://www.gnu.org/copyleft/}. | |
21388 | ||
21389 | Each version of the License is given a distinguishing version number. | |
21390 | If the Document specifies that a particular numbered version of this | |
21391 | License ``or any later version'' applies to it, you have the option of | |
21392 | following the terms and conditions either of that specified version or | |
21393 | of any later version that has been published (not as a draft) by the | |
21394 | Free Software Foundation. If the Document does not specify a version | |
21395 | number of this License, you may choose any version ever published (not | |
21396 | as a draft) by the Free Software Foundation. | |
21397 | @end enumerate | |
21398 | ||
21399 | @node Index, About the Author, GNU Free Documentation License, Top | |
21400 | @comment node-name, next, previous, up | |
21401 | @unnumbered Index | |
21402 | ||
d586ab6c | 21403 | @ignore |
8b096dce | 21404 | MENU ENTRY: NODE NAME. |
d586ab6c | 21405 | @end ignore |
8b096dce EZ |
21406 | |
21407 | @printindex cp | |
21408 | ||
21409 | @iftex | |
21410 | @c Place biographical information on right-hand (verso) page | |
21411 | ||
21412 | @tex | |
21413 | \ifodd\pageno | |
21414 | \par\vfill\supereject | |
21415 | \global\evenheadline={\hfil} \global\evenfootline={\hfil} | |
21416 | \global\oddheadline={\hfil} \global\oddfootline={\hfil} | |
21417 | \page\hbox{}\page | |
21418 | \else | |
21419 | \par\vfill\supereject | |
21420 | \par\vfill\supereject | |
21421 | \global\evenheadline={\hfil} \global\evenfootline={\hfil} | |
21422 | \global\oddheadline={\hfil} \global\oddfootline={\hfil} | |
21423 | \page\hbox{}\page | |
21424 | \page\hbox{}\page | |
21425 | \fi | |
21426 | @end tex | |
21427 | ||
21428 | @page | |
21429 | @w{ } | |
21430 | ||
21431 | @c ================ Biographical information ================ | |
21432 | ||
21433 | @w{ } | |
21434 | @sp 8 | |
21435 | @center About the Author | |
21436 | @sp 1 | |
21437 | @end iftex | |
21438 | ||
21439 | @ifnottex | |
21440 | @node About the Author, , Index, Top | |
21441 | @unnumbered About the Author | |
21442 | @end ifnottex | |
21443 | ||
21444 | @quotation | |
21445 | Robert J. Chassell has worked with GNU Emacs since 1985. He writes | |
21446 | and edits, teaches Emacs and Emacs Lisp, and speaks throughout the | |
21447 | world on software freedom. Chassell was a founding Director and | |
21448 | Treasurer of the Free Software Foundation, Inc. He is co-author of | |
21449 | the @cite{Texinfo} manual, and has edited more than a dozen other | |
21450 | books. He graduated from Cambridge University, in England. He has an | |
21451 | abiding interest in social and economic history and flies his own | |
21452 | airplane. | |
21453 | @end quotation | |
21454 | ||
21455 | @page | |
21456 | @w{ } | |
21457 | ||
21458 | @c Prevent page number on blank verso, so eject it first. | |
21459 | @tex | |
21460 | \par\vfill\supereject | |
21461 | @end tex | |
21462 | ||
21463 | @iftex | |
21464 | @headings off | |
21465 | @evenheading @thispage @| @| @thistitle | |
21466 | @oddheading @| @| @thispage | |
21467 | @end iftex | |
21468 | ||
8b096dce | 21469 | @bye |