1 \input texinfo @c -*-texinfo-*-
2 @comment %**start of header
3 @setfilename ../../info/eintr
4 @c setfilename emacs-lisp-intro.info
5 @c sethtmlfilename emacs-lisp-intro.html
6 @settitle Programming in Emacs Lisp
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 >>>>
20 @set print-postscript-figures
22 @c clear print-postscript-figures
25 @comment %**end of header
27 @c per rms and peterb, use 10pt fonts for the main text, mostly to
28 @c save on paper cost.
29 @c Do this inside @tex for now, so current makeinfo does not complain.
35 \global\hbadness=6666 % don't worry about not-too-underfull boxes
38 @set edition-number 3.10
39 @set update-date 28 October 2009
42 ## Summary of shell commands to create various output formats:
44 pushd /usr/local/src/emacs/lispintro/
48 makeinfo --paragraph-indent=0 --verbose emacs-lisp-intro.texi
50 ## ;; (progn (when (bufferp (get-buffer "*info*")) (kill-buffer "*info*")) (info "/usr/local/src/emacs/info/eintr"))
53 texi2dvi emacs-lisp-intro.texi
55 ## xdvi -margins 24pt -topmargin 4pt -offsets 24pt -geometry 760x1140 -s 5 -useTeXpages -mousemode 1 emacs-lisp-intro.dvi &
58 makeinfo --html --no-split --verbose emacs-lisp-intro.texi
60 ## galeon emacs-lisp-intro.html
63 makeinfo --fill-column=70 --no-split --paragraph-indent=0 \
64 --verbose --no-headers --output=emacs-lisp-intro.txt emacs-lisp-intro.texi
70 mtusb # mount -v -t ext3 /dev/sda /mnt
71 cp -v /u/intro/emacs-lisp-intro.texi /mnt/backup/intro/emacs-lisp-intro.texi
72 umtusb # umount -v /mnt
75 ## Other shell commands
77 pushd /usr/local/src/emacs/lispintro/
81 texi2dvi --pdf emacs-lisp-intro.texi
82 # xpdf emacs-lisp-intro.pdf &
84 ## DocBook -- note file extension
85 makeinfo --docbook --no-split --paragraph-indent=0 \
86 --verbose --output=emacs-lisp-intro.docbook emacs-lisp-intro.texi
88 ## XML with a Texinfo DTD -- note file extension
89 makeinfo --xml --no-split --paragraph-indent=0 \
90 --verbose --output=emacs-lisp-intro.texinfoxml emacs-lisp-intro.texi
92 ## PostScript (needs DVI)
93 # gv emacs-lisp-intro.ps &
94 # Create DVI if we lack it
95 # texi2dvi emacs-lisp-intro.texi
96 dvips emacs-lisp-intro.dvi -o emacs-lisp-intro.ps
99 # Use OpenOffice to view RTF
100 # Create HTML if we lack it
101 # makeinfo --no-split --html emacs-lisp-intro.texi
102 /usr/local/src/html2rtf.pl emacs-lisp-intro.html
105 /usr/bin/rtf2latex emacs-lisp-intro.rtf
111 @c ================ Included Figures ================
113 @c Set print-postscript-figures if you print PostScript figures.
114 @c If you clear this, the ten figures will be printed as ASCII diagrams.
115 @c (This is not relevant to Info, since Info only handles ASCII.)
116 @c Your site may require editing changes to print PostScript; in this
117 @c case, search for `print-postscript-figures' and make appropriate changes.
119 @c ================ How to Create an Info file ================
121 @c If you have `makeinfo' installed, run the following command
123 @c makeinfo emacs-lisp-intro.texi
125 @c or, if you want a single, large Info file, and no paragraph indents:
126 @c makeinfo --no-split --paragraph-indent=0 --verbose emacs-lisp-intro.texi
128 @c After creating the Info file, edit your Info `dir' file, if the
129 @c `dircategory' section below does not enable your system to
130 @c install the manual automatically.
131 @c (The `dir' file is often in the `/usr/local/share/info/' directory.)
133 @c ================ How to Create an HTML file ================
135 @c To convert to HTML format
136 @c makeinfo --html --no-split --verbose emacs-lisp-intro.texi
138 @c ================ How to Print a Book in Various Sizes ================
140 @c This book can be printed in any of three different sizes.
141 @c In the above header, set @-commands appropriately.
151 @c European A4 size paper:
156 @c ================ How to Typeset and Print ================
158 @c If you do not include PostScript figures, run either of the
159 @c following command sequences, or similar commands suited to your
162 @c texi2dvi emacs-lisp-intro.texi
163 @c lpr -d emacs-lisp-intro.dvi
167 @c tex emacs-lisp-intro.texi
168 @c texindex emacs-lisp-intro.??
169 @c tex emacs-lisp-intro.texi
170 @c lpr -d emacs-lisp-intro.dvi
172 @c If you include the PostScript figures, and you have old software,
173 @c you may need to convert the .dvi file to a .ps file before
174 @c printing. Run either of the following command sequences, or one
177 @c dvips -f < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
181 @c postscript -p < emacs-lisp-intro.dvi > emacs-lisp-intro.ps
184 @c (Note: if you edit the book so as to change the length of the
185 @c table of contents, you may have to change the value of `pageno' below.)
187 @c ================ End of Formatting Sections ================
189 @c For next or subsequent edition:
190 @c create function using with-output-to-temp-buffer
191 @c create a major mode, with keymaps
192 @c run an asynchronous process, like grep or diff
194 @c For 8.5 by 11 inch format: do not use such a small amount of
195 @c whitespace between paragraphs as smallbook format
198 \global\parskip 6pt plus 1pt
202 @c For all sized formats: print within-book cross
203 @c reference with ``...'' rather than [...]
205 @c This works with the texinfo.tex file, version 2003-05-04.08,
206 @c in the Texinfo version 4.6 of the 2003 Jun 13 distribution.
209 \if \xrefprintnodename
210 \global\def\xrefprintnodename#1{\unskip, ``#1''}
212 \global\def\xrefprintnodename#1{ ``#1''}
214 % \global\def\xrefprintnodename#1{, ``#1''}
217 @c ----------------------------------------------------
219 @dircategory GNU Emacs Lisp
221 * Emacs Lisp Intro: (eintr).
222 A simple introduction to Emacs Lisp programming.
226 This is an @cite{Introduction to Programming in Emacs Lisp}, for
227 people who are not programmers.
229 Edition @value{edition-number}, @value{update-date}
232 <p>The homepage for GNU Emacs is at
233 <a href="http://www.gnu.org/software/emacs/">http://www.gnu.org/software/emacs/</a>.
234 <br>To view this manual in other formats, click
235 <a href="/software/emacs/emacs-lisp-intro/emacs-lisp-intro.html">here</a>.
239 Copyright @copyright{} 1990--1995, 1997, 2001--2013 Free Software
246 GNU Press, @hfill @uref{http://www.fsf.org/campaigns/gnu-press/}@*
247 a division of the @hfill email: @email{sales@@fsf.org}@*
248 Free Software Foundation, Inc. @hfill Tel: +1 (617) 542-5942@*
249 51 Franklin Street, Fifth Floor @hfill Fax: +1 (617) 542-2652@*
250 Boston, MA 02110-1301 USA
257 GNU Press, http://www.fsf.org/campaigns/gnu-press/
258 a division of the email: sales@@fsf.org
259 Free Software Foundation, Inc. Tel: +1 (617) 542-5942
260 51 Franklin Street, Fifth Floor Fax: +1 (617) 542-2652
261 Boston, MA 02110-1301 USA
266 @c Printed copies are available from @uref{http://shop.fsf.org/} for $35 each.@*
269 Permission is granted to copy, distribute and/or modify this document
270 under the terms of the GNU Free Documentation License, Version 1.3 or
271 any later version published by the Free Software Foundation; there
272 being no Invariant Section, with the Front-Cover Texts being ``A GNU
273 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
274 the license is included in the section entitled ``GNU Free
275 Documentation License''.
277 (a) The FSF's Back-Cover Text is: ``You have the freedom to
278 copy and modify this GNU manual. Buying copies from the FSF
279 supports it in developing GNU and promoting software freedom.''
282 @c half title; two lines here, so do not use `shorttitlepage'
285 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
287 {\begingroup\hbox{}\vskip 0.25in \chaprm%
288 \centerline{Programming in Emacs Lisp}%
289 \endgroup\page\hbox{}\page}
294 @center @titlefont{An Introduction to}
296 @center @titlefont{Programming in Emacs Lisp}
298 @center Revised Third Edition
300 @center by Robert J. Chassell
303 @vskip 0pt plus 1filll
309 @evenheading @thispage @| @| @thischapter
310 @oddheading @thissection @| @| @thispage
314 @c Keep T.O.C. short by tightening up for largebook
317 \global\parskip 2pt plus 1pt
318 \global\advance\baselineskip by -1pt
328 @top An Introduction to Programming in Emacs Lisp
332 This master menu first lists each chapter and index; then it lists
333 every node in every chapter.
336 @c >>>> Set pageno appropriately <<<<
338 @c The first page of the Preface is a roman numeral; it is the first
339 @c right handed page after the Table of Contents; hence the following
340 @c setting must be for an odd negative number.
343 @c global@pageno = -11
346 @set COUNT-WORDS count-words-example
347 @c Length of variable name chosen so that things still line up when expanded.
350 * Preface:: What to look for.
351 * List Processing:: What is Lisp?
352 * Practicing Evaluation:: Running several programs.
353 * Writing Defuns:: How to write function definitions.
354 * Buffer Walk Through:: Exploring a few buffer-related functions.
355 * More Complex:: A few, even more complex functions.
356 * Narrowing & Widening:: Restricting your and Emacs attention to
358 * car cdr & cons:: Fundamental functions in Lisp.
359 * Cutting & Storing Text:: Removing text and saving it.
360 * List Implementation:: How lists are implemented in the computer.
361 * Yanking:: Pasting stored text.
362 * Loops & Recursion:: How to repeat a process.
363 * Regexp Search:: Regular expression searches.
364 * Counting Words:: A review of repetition and regexps.
365 * Words in a defun:: Counting words in a @code{defun}.
366 * Readying a Graph:: A prototype graph printing function.
367 * Emacs Initialization:: How to write a @file{.emacs} file.
368 * Debugging:: How to run the Emacs Lisp debuggers.
369 * Conclusion:: Now you have the basics.
370 * the-the:: An appendix: how to find reduplicated words.
371 * Kill Ring:: An appendix: how the kill ring works.
372 * Full Graph:: How to create a graph with labeled axes.
373 * Free Software and Free Manuals::
374 * GNU Free Documentation License::
379 --- The Detailed Node Listing ---
383 * Why:: Why learn Emacs Lisp?
384 * On Reading this Text:: Read, gain familiarity, pick up habits....
385 * Who You Are:: For whom this is written.
387 * Note for Novices:: You can read this as a novice.
392 * Lisp Lists:: What are lists?
393 * Run a Program:: Any list in Lisp is a program ready to run.
394 * Making Errors:: Generating an error message.
395 * Names & Definitions:: Names of symbols and function definitions.
396 * Lisp Interpreter:: What the Lisp interpreter does.
397 * Evaluation:: Running a program.
398 * Variables:: Returning a value from a variable.
399 * Arguments:: Passing information to a function.
400 * set & setq:: Setting the value of a variable.
401 * Summary:: The major points.
402 * Error Message Exercises::
406 * Numbers Lists:: List have numbers, other lists, in them.
407 * Lisp Atoms:: Elemental entities.
408 * Whitespace in Lists:: Formatting lists to be readable.
409 * Typing Lists:: How GNU Emacs helps you type lists.
413 * Complications:: Variables, Special forms, Lists within.
414 * Byte Compiling:: Specially processing code for speed.
418 * How the Interpreter Acts:: Returns and Side Effects...
419 * Evaluating Inner Lists:: Lists within lists...
423 * fill-column Example::
424 * Void Function:: The error message for a symbol
426 * Void Variable:: The error message for a symbol without a value.
430 * Data types:: Types of data passed to a function.
431 * Args as Variable or List:: An argument can be the value
432 of a variable or list.
433 * Variable Number of Arguments:: Some functions may take a
434 variable number of arguments.
435 * Wrong Type of Argument:: Passing an argument of the wrong type
437 * message:: A useful function for sending messages.
439 Setting the Value of a Variable
441 * Using set:: Setting values.
442 * Using setq:: Setting a quoted value.
443 * Counting:: Using @code{setq} to count.
445 Practicing Evaluation
447 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
449 * Buffer Names:: Buffers and files are different.
450 * Getting Buffers:: Getting a buffer itself, not merely its name.
451 * Switching Buffers:: How to change to another buffer.
452 * Buffer Size & Locations:: Where point is located and the size of
454 * Evaluation Exercise::
456 How To Write Function Definitions
458 * Primitive Functions::
459 * defun:: The @code{defun} special form.
460 * Install:: Install a function definition.
461 * Interactive:: Making a function interactive.
462 * Interactive Options:: Different options for @code{interactive}.
463 * Permanent Installation:: Installing code permanently.
464 * let:: Creating and initializing local variables.
466 * else:: If--then--else expressions.
467 * Truth & Falsehood:: What Lisp considers false and true.
468 * save-excursion:: Keeping track of point, mark, and buffer.
472 Install a Function Definition
474 * Effect of installation::
475 * Change a defun:: How to change a function definition.
477 Make a Function Interactive
479 * Interactive multiply-by-seven:: An overview.
480 * multiply-by-seven in detail:: The interactive version.
484 * Prevent confusion::
485 * Parts of let Expression::
486 * Sample let Expression::
487 * Uninitialized let Variables::
489 The @code{if} Special Form
491 * if in more detail::
492 * type-of-animal in detail:: An example of an @code{if} expression.
494 Truth and Falsehood in Emacs Lisp
496 * nil explained:: @code{nil} has two meanings.
498 @code{save-excursion}
500 * Point and mark:: A review of various locations.
501 * Template for save-excursion::
503 A Few Buffer--Related Functions
505 * Finding More:: How to find more information.
506 * simplified-beginning-of-buffer:: Shows @code{goto-char},
507 @code{point-min}, and @code{push-mark}.
508 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
509 * append-to-buffer:: Uses @code{save-excursion} and
510 @code{insert-buffer-substring}.
511 * Buffer Related Review:: Review.
514 The Definition of @code{mark-whole-buffer}
516 * mark-whole-buffer overview::
517 * Body of mark-whole-buffer:: Only three lines of code.
519 The Definition of @code{append-to-buffer}
521 * append-to-buffer overview::
522 * append interactive:: A two part interactive expression.
523 * append-to-buffer body:: Incorporates a @code{let} expression.
524 * append save-excursion:: How the @code{save-excursion} works.
526 A Few More Complex Functions
528 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
529 * insert-buffer:: Read-only, and with @code{or}.
530 * beginning-of-buffer:: Shows @code{goto-char},
531 @code{point-min}, and @code{push-mark}.
532 * Second Buffer Related Review::
533 * optional Exercise::
535 The Definition of @code{insert-buffer}
537 * insert-buffer code::
538 * insert-buffer interactive:: When you can read, but not write.
539 * insert-buffer body:: The body has an @code{or} and a @code{let}.
540 * if & or:: Using an @code{if} instead of an @code{or}.
541 * Insert or:: How the @code{or} expression works.
542 * Insert let:: Two @code{save-excursion} expressions.
543 * New insert-buffer::
545 The Interactive Expression in @code{insert-buffer}
547 * Read-only buffer:: When a buffer cannot be modified.
548 * b for interactive:: An existing buffer or else its name.
550 Complete Definition of @code{beginning-of-buffer}
552 * Optional Arguments::
553 * beginning-of-buffer opt arg:: Example with optional argument.
554 * beginning-of-buffer complete::
556 @code{beginning-of-buffer} with an Argument
558 * Disentangle beginning-of-buffer::
559 * Large buffer case::
560 * Small buffer case::
562 Narrowing and Widening
564 * Narrowing advantages:: The advantages of narrowing
565 * save-restriction:: The @code{save-restriction} special form.
566 * what-line:: The number of the line that point is on.
569 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
571 * Strange Names:: An historical aside: why the strange names?
572 * car & cdr:: Functions for extracting part of a list.
573 * cons:: Constructing a list.
574 * nthcdr:: Calling @code{cdr} repeatedly.
576 * setcar:: Changing the first element of a list.
577 * setcdr:: Changing the rest of a list.
583 * length:: How to find the length of a list.
585 Cutting and Storing Text
587 * Storing Text:: Text is stored in a list.
588 * zap-to-char:: Cutting out text up to a character.
589 * kill-region:: Cutting text out of a region.
590 * copy-region-as-kill:: A definition for copying text.
591 * Digression into C:: Minor note on C programming language macros.
592 * defvar:: How to give a variable an initial value.
593 * cons & search-fwd Review::
598 * Complete zap-to-char:: The complete implementation.
599 * zap-to-char interactive:: A three part interactive expression.
600 * zap-to-char body:: A short overview.
601 * search-forward:: How to search for a string.
602 * progn:: The @code{progn} special form.
603 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
607 * Complete kill-region:: The function definition.
608 * condition-case:: Dealing with a problem.
611 @code{copy-region-as-kill}
613 * Complete copy-region-as-kill:: The complete function definition.
614 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
616 The Body of @code{copy-region-as-kill}
618 * last-command & this-command::
619 * kill-append function::
620 * kill-new function::
622 Initializing a Variable with @code{defvar}
624 * See variable current value::
625 * defvar and asterisk::
627 How Lists are Implemented
630 * Symbols as Chest:: Exploring a powerful metaphor.
635 * Kill Ring Overview::
636 * kill-ring-yank-pointer:: The kill ring is a list.
637 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
641 * while:: Causing a stretch of code to repeat.
643 * Recursion:: Causing a function to call itself.
648 * Looping with while:: Repeat so long as test returns true.
649 * Loop Example:: A @code{while} loop that uses a list.
650 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
651 * Incrementing Loop:: A loop with an incrementing counter.
652 * Incrementing Loop Details::
653 * Decrementing Loop:: A loop with a decrementing counter.
655 Details of an Incrementing Loop
657 * Incrementing Example:: Counting pebbles in a triangle.
658 * Inc Example parts:: The parts of the function definition.
659 * Inc Example altogether:: Putting the function definition together.
661 Loop with a Decrementing Counter
663 * Decrementing Example:: More pebbles on the beach.
664 * Dec Example parts:: The parts of the function definition.
665 * Dec Example altogether:: Putting the function definition together.
667 Save your time: @code{dolist} and @code{dotimes}
674 * Building Robots:: Same model, different serial number ...
675 * Recursive Definition Parts:: Walk until you stop ...
676 * Recursion with list:: Using a list as the test whether to recurse.
677 * Recursive triangle function::
678 * Recursion with cond::
679 * Recursive Patterns:: Often used templates.
680 * No Deferment:: Don't store up work ...
681 * No deferment solution::
683 Recursion in Place of a Counter
685 * Recursive Example arg of 1 or 2::
686 * Recursive Example arg of 3 or 4::
694 Regular Expression Searches
696 * sentence-end:: The regular expression for @code{sentence-end}.
697 * re-search-forward:: Very similar to @code{search-forward}.
698 * forward-sentence:: A straightforward example of regexp search.
699 * forward-paragraph:: A somewhat complex example.
700 * etags:: How to create your own @file{TAGS} table.
702 * re-search Exercises::
704 @code{forward-sentence}
706 * Complete forward-sentence::
707 * fwd-sentence while loops:: Two @code{while} loops.
708 * fwd-sentence re-search:: A regular expression search.
710 @code{forward-paragraph}: a Goldmine of Functions
712 * forward-paragraph in brief:: Key parts of the function definition.
713 * fwd-para let:: The @code{let*} expression.
714 * fwd-para while:: The forward motion @code{while} loop.
716 Counting: Repetition and Regexps
719 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
720 * recursive-count-words:: Start with case of no words in region.
721 * Counting Exercise::
723 The @code{@value{COUNT-WORDS}} Function
725 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
726 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
728 Counting Words in a @code{defun}
730 * Divide and Conquer::
731 * Words and Symbols:: What to count?
732 * Syntax:: What constitutes a word or symbol?
733 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
734 * Several defuns:: Counting several defuns in a file.
735 * Find a File:: Do you want to look at a file?
736 * lengths-list-file:: A list of the lengths of many definitions.
737 * Several files:: Counting in definitions in different files.
738 * Several files recursively:: Recursively counting in different files.
739 * Prepare the data:: Prepare the data for display in a graph.
741 Count Words in @code{defuns} in Different Files
743 * lengths-list-many-files:: Return a list of the lengths of defuns.
744 * append:: Attach one list to another.
746 Prepare the Data for Display in a Graph
748 * Data for Display in Detail::
749 * Sorting:: Sorting lists.
750 * Files List:: Making a list of files.
751 * Counting function definitions::
755 * Columns of a graph::
756 * graph-body-print:: How to print the body of a graph.
757 * recursive-graph-body-print::
759 * Line Graph Exercise::
761 Your @file{.emacs} File
763 * Default Configuration::
764 * Site-wide Init:: You can write site-wide init files.
765 * defcustom:: Emacs will write code for you.
766 * Beginning a .emacs File:: How to write a @code{.emacs file}.
767 * Text and Auto-fill:: Automatically wrap lines.
768 * Mail Aliases:: Use abbreviations for email addresses.
769 * Indent Tabs Mode:: Don't use tabs with @TeX{}
770 * Keybindings:: Create some personal keybindings.
771 * Keymaps:: More about key binding.
772 * Loading Files:: Load (i.e., evaluate) files automatically.
773 * Autoload:: Make functions available.
774 * Simple Extension:: Define a function; bind it to a key.
775 * X11 Colors:: Colors in X.
777 * Mode Line:: How to customize your mode line.
781 * debug:: How to use the built-in debugger.
782 * debug-on-entry:: Start debugging when you call a function.
783 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
784 * edebug:: How to use Edebug, a source level debugger.
785 * Debugging Exercises::
787 Handling the Kill Ring
789 * What the Kill Ring Does::
791 * yank:: Paste a copy of a clipped element.
792 * yank-pop:: Insert element pointed to.
795 The @code{current-kill} Function
797 * Code for current-kill::
798 * Understanding current-kill::
800 @code{current-kill} in Outline
802 * Body of current-kill::
803 * Digression concerning error:: How to mislead humans, but not computers.
804 * Determining the Element::
806 A Graph with Labeled Axes
809 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
810 * print-Y-axis:: Print a label for the vertical axis.
811 * print-X-axis:: Print a horizontal label.
812 * Print Whole Graph:: The function to print a complete graph.
814 The @code{print-Y-axis} Function
816 * print-Y-axis in Detail::
817 * Height of label:: What height for the Y axis?
818 * Compute a Remainder:: How to compute the remainder of a division.
819 * Y Axis Element:: Construct a line for the Y axis.
820 * Y-axis-column:: Generate a list of Y axis labels.
821 * print-Y-axis Penultimate:: A not quite final version.
823 The @code{print-X-axis} Function
825 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
826 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
828 Printing the Whole Graph
830 * The final version:: A few changes.
831 * Test print-graph:: Run a short test.
832 * Graphing words in defuns:: Executing the final code.
833 * lambda:: How to write an anonymous function.
834 * mapcar:: Apply a function to elements of a list.
835 * Another Bug:: Yet another bug @dots{} most insidious.
836 * Final printed graph:: The graph itself!
844 Most of the GNU Emacs integrated environment is written in the programming
845 language called Emacs Lisp. The code written in this programming
846 language is the software---the sets of instructions---that tell the
847 computer what to do when you give it commands. Emacs is designed so
848 that you can write new code in Emacs Lisp and easily install it as an
849 extension to the editor.
851 (GNU Emacs is sometimes called an ``extensible editor'', but it does
852 much more than provide editing capabilities. It is better to refer to
853 Emacs as an ``extensible computing environment''. However, that
854 phrase is quite a mouthful. It is easier to refer to Emacs simply as
855 an editor. Moreover, everything you do in Emacs---find the Mayan date
856 and phases of the moon, simplify polynomials, debug code, manage
857 files, read letters, write books---all these activities are kinds of
858 editing in the most general sense of the word.)
861 * Why:: Why learn Emacs Lisp?
862 * On Reading this Text:: Read, gain familiarity, pick up habits....
863 * Who You Are:: For whom this is written.
865 * Note for Novices:: You can read this as a novice.
871 @unnumberedsec Why Study Emacs Lisp?
874 Although Emacs Lisp is usually thought of in association only with Emacs,
875 it is a full computer programming language. You can use Emacs Lisp as
876 you would any other programming language.
878 Perhaps you want to understand programming; perhaps you want to extend
879 Emacs; or perhaps you want to become a programmer. This introduction to
880 Emacs Lisp is designed to get you started: to guide you in learning the
881 fundamentals of programming, and more importantly, to show you how you
882 can teach yourself to go further.
884 @node On Reading this Text
885 @unnumberedsec On Reading this Text
887 All through this document, you will see little sample programs you can
888 run inside of Emacs. If you read this document in Info inside of GNU
889 Emacs, you can run the programs as they appear. (This is easy to do and
890 is explained when the examples are presented.) Alternatively, you can
891 read this introduction as a printed book while sitting beside a computer
892 running Emacs. (This is what I like to do; I like printed books.) If
893 you don't have a running Emacs beside you, you can still read this book,
894 but in this case, it is best to treat it as a novel or as a travel guide
895 to a country not yet visited: interesting, but not the same as being
898 Much of this introduction is dedicated to walkthroughs or guided tours
899 of code used in GNU Emacs. These tours are designed for two purposes:
900 first, to give you familiarity with real, working code (code you use
901 every day); and, second, to give you familiarity with the way Emacs
902 works. It is interesting to see how a working environment is
905 hope that you will pick up the habit of browsing through source code.
906 You can learn from it and mine it for ideas. Having GNU Emacs is like
907 having a dragon's cave of treasures.
909 In addition to learning about Emacs as an editor and Emacs Lisp as a
910 programming language, the examples and guided tours will give you an
911 opportunity to get acquainted with Emacs as a Lisp programming
912 environment. GNU Emacs supports programming and provides tools that
913 you will want to become comfortable using, such as @kbd{M-.} (the key
914 which invokes the @code{find-tag} command). You will also learn about
915 buffers and other objects that are part of the environment.
916 Learning about these features of Emacs is like learning new routes
917 around your home town.
920 In addition, I have written several programs as extended examples.
921 Although these are examples, the programs are real. I use them.
922 Other people use them. You may use them. Beyond the fragments of
923 programs used for illustrations, there is very little in here that is
924 `just for teaching purposes'; what you see is used. This is a great
925 advantage of Emacs Lisp: it is easy to learn to use it for work.
928 Finally, I hope to convey some of the skills for using Emacs to
929 learn aspects of programming that you don't know. You can often use
930 Emacs to help you understand what puzzles you or to find out how to do
931 something new. This self-reliance is not only a pleasure, but an
935 @unnumberedsec For Whom This is Written
937 This text is written as an elementary introduction for people who are
938 not programmers. If you are a programmer, you may not be satisfied with
939 this primer. The reason is that you may have become expert at reading
940 reference manuals and be put off by the way this text is organized.
942 An expert programmer who reviewed this text said to me:
945 @i{I prefer to learn from reference manuals. I ``dive into'' each
946 paragraph, and ``come up for air'' between paragraphs.}
948 @i{When I get to the end of a paragraph, I assume that that subject is
949 done, finished, that I know everything I need (with the
950 possible exception of the case when the next paragraph starts talking
951 about it in more detail). I expect that a well written reference manual
952 will not have a lot of redundancy, and that it will have excellent
953 pointers to the (one) place where the information I want is.}
956 This introduction is not written for this person!
958 Firstly, I try to say everything at least three times: first, to
959 introduce it; second, to show it in context; and third, to show it in a
960 different context, or to review it.
962 Secondly, I hardly ever put all the information about a subject in one
963 place, much less in one paragraph. To my way of thinking, that imposes
964 too heavy a burden on the reader. Instead I try to explain only what
965 you need to know at the time. (Sometimes I include a little extra
966 information so you won't be surprised later when the additional
967 information is formally introduced.)
969 When you read this text, you are not expected to learn everything the
970 first time. Frequently, you need only make, as it were, a `nodding
971 acquaintance' with some of the items mentioned. My hope is that I have
972 structured the text and given you enough hints that you will be alert to
973 what is important, and concentrate on it.
975 You will need to ``dive into'' some paragraphs; there is no other way
976 to read them. But I have tried to keep down the number of such
977 paragraphs. This book is intended as an approachable hill, rather than
978 as a daunting mountain.
980 This introduction to @cite{Programming in Emacs Lisp} has a companion
983 @cite{The GNU Emacs Lisp Reference Manual}.
986 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
987 Emacs Lisp Reference Manual}.
989 The reference manual has more detail than this introduction. In the
990 reference manual, all the information about one topic is concentrated
991 in one place. You should turn to it if you are like the programmer
992 quoted above. And, of course, after you have read this
993 @cite{Introduction}, you will find the @cite{Reference Manual} useful
994 when you are writing your own programs.
997 @unnumberedsec Lisp History
1000 Lisp was first developed in the late 1950s at the Massachusetts
1001 Institute of Technology for research in artificial intelligence. The
1002 great power of the Lisp language makes it superior for other purposes as
1003 well, such as writing editor commands and integrated environments.
1007 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
1008 in the 1960s. It is somewhat inspired by Common Lisp, which became a
1009 standard in the 1980s. However, Emacs Lisp is much simpler than Common
1010 Lisp. (The standard Emacs distribution contains an optional extensions
1011 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
1013 @node Note for Novices
1014 @unnumberedsec A Note for Novices
1016 If you don't know GNU Emacs, you can still read this document
1017 profitably. However, I recommend you learn Emacs, if only to learn to
1018 move around your computer screen. You can teach yourself how to use
1019 Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This
1020 means you press and release the @key{CTRL} key and the @kbd{h} at the
1021 same time, and then press and release @kbd{t}.)
1023 Also, I often refer to one of Emacs's standard commands by listing the
1024 keys which you press to invoke the command and then giving the name of
1025 the command in parentheses, like this: @kbd{M-C-\}
1026 (@code{indent-region}). What this means is that the
1027 @code{indent-region} command is customarily invoked by typing
1028 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
1029 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
1030 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
1031 @key{META} key, @key{CTRL} key and @key{\} key all at the same time.
1032 (On many modern keyboards the @key{META} key is labeled
1034 Sometimes a combination like this is called a keychord, since it is
1035 similar to the way you play a chord on a piano. If your keyboard does
1036 not have a @key{META} key, the @key{ESC} key prefix is used in place
1037 of it. In this case, @kbd{M-C-\} means that you press and release your
1038 @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at
1039 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
1040 along with the key that is labeled @key{ALT} and, at the same time,
1041 press the @key{\} key.
1043 In addition to typing a lone keychord, you can prefix what you type
1044 with @kbd{C-u}, which is called the `universal argument'. The
1045 @kbd{C-u} keychord passes an argument to the subsequent command.
1046 Thus, to indent a region of plain text by 6 spaces, mark the region,
1047 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
1048 Emacs either passes the number 4 to the command or otherwise runs the
1049 command differently than it would otherwise.) @xref{Arguments, ,
1050 Numeric Arguments, emacs, The GNU Emacs Manual}.
1052 If you are reading this in Info using GNU Emacs, you can read through
1053 this whole document just by pressing the space bar, @key{SPC}.
1054 (To learn about Info, type @kbd{C-h i} and then select Info.)
1056 A note on terminology: when I use the word Lisp alone, I often am
1057 referring to the various dialects of Lisp in general, but when I speak
1058 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
1061 @unnumberedsec Thank You
1063 My thanks to all who helped me with this book. My especial thanks to
1064 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
1065 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.
1066 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
1067 @w{Philip Johnson} and @w{David Stampe} for their patient
1068 encouragement. My mistakes are my own.
1072 @email{bob@@gnu.org}
1075 @c ================ Beginning of main text ================
1077 @c Start main text on right-hand (verso) page
1080 \par\vfill\supereject
1083 \par\vfill\supereject
1085 \par\vfill\supereject
1087 \par\vfill\supereject
1091 @c Note: this resetting of the page number back to 1 causes TeX to gripe
1092 @c about already having seen page numbers 1-4 before (in the preface):
1093 @c pdfTeX warning (ext4): destination with the same identifier (name{1})
1094 @c has been already used, duplicate ignored
1095 @c I guess that is harmless (what happens if a later part of the text
1096 @c makes a link to something in the first 4 pages though?).
1097 @c E.g., note that the Emacs manual has a preface, but does not bother
1098 @c resetting the page numbers back to 1 after that.
1101 @evenheading @thispage @| @| @thischapter
1102 @oddheading @thissection @| @| @thispage
1106 @node List Processing
1107 @chapter List Processing
1109 To the untutored eye, Lisp is a strange programming language. In Lisp
1110 code there are parentheses everywhere. Some people even claim that
1111 the name stands for `Lots of Isolated Silly Parentheses'. But the
1112 claim is unwarranted. Lisp stands for LISt Processing, and the
1113 programming language handles @emph{lists} (and lists of lists) by
1114 putting them between parentheses. The parentheses mark the boundaries
1115 of the list. Sometimes a list is preceded by a single apostrophe or
1116 quotation mark, @samp{'}@footnote{The single apostrophe or quotation
1117 mark is an abbreviation for the function @code{quote}; you need not
1118 think about functions now; functions are defined in @ref{Making
1119 Errors, , Generate an Error Message}.} Lists are the basis of Lisp.
1122 * Lisp Lists:: What are lists?
1123 * Run a Program:: Any list in Lisp is a program ready to run.
1124 * Making Errors:: Generating an error message.
1125 * Names & Definitions:: Names of symbols and function definitions.
1126 * Lisp Interpreter:: What the Lisp interpreter does.
1127 * Evaluation:: Running a program.
1128 * Variables:: Returning a value from a variable.
1129 * Arguments:: Passing information to a function.
1130 * set & setq:: Setting the value of a variable.
1131 * Summary:: The major points.
1132 * Error Message Exercises::
1139 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1140 This list is preceded by a single apostrophe. It could just as well be
1141 written as follows, which looks more like the kind of list you are likely
1142 to be familiar with:
1154 The elements of this list are the names of the four different flowers,
1155 separated from each other by whitespace and surrounded by parentheses,
1156 like flowers in a field with a stone wall around them.
1157 @cindex Flowers in a field
1160 * Numbers Lists:: List have numbers, other lists, in them.
1161 * Lisp Atoms:: Elemental entities.
1162 * Whitespace in Lists:: Formatting lists to be readable.
1163 * Typing Lists:: How GNU Emacs helps you type lists.
1168 @unnumberedsubsec Numbers, Lists inside of Lists
1171 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1172 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1173 separated by whitespace.
1175 In Lisp, both data and programs are represented the same way; that is,
1176 they are both lists of words, numbers, or other lists, separated by
1177 whitespace and surrounded by parentheses. (Since a program looks like
1178 data, one program may easily serve as data for another; this is a very
1179 powerful feature of Lisp.) (Incidentally, these two parenthetical
1180 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1181 @samp{.} as punctuation marks.)
1184 Here is another list, this time with a list inside of it:
1187 '(this list has (a list inside of it))
1190 The components of this list are the words @samp{this}, @samp{list},
1191 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1192 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1193 @samp{of}, @samp{it}.
1196 @subsection Lisp Atoms
1199 In Lisp, what we have been calling words are called @dfn{atoms}. This
1200 term comes from the historical meaning of the word atom, which means
1201 `indivisible'. As far as Lisp is concerned, the words we have been
1202 using in the lists cannot be divided into any smaller parts and still
1203 mean the same thing as part of a program; likewise with numbers and
1204 single character symbols like @samp{+}. On the other hand, unlike an
1205 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1206 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1208 In a list, atoms are separated from each other by whitespace. They can be
1209 right next to a parenthesis.
1211 @cindex @samp{empty list} defined
1212 Technically speaking, a list in Lisp consists of parentheses surrounding
1213 atoms separated by whitespace or surrounding other lists or surrounding
1214 both atoms and other lists. A list can have just one atom in it or
1215 have nothing in it at all. A list with nothing in it looks like this:
1216 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1217 empty list is considered both an atom and a list at the same time.
1219 @cindex Symbolic expressions, introduced
1220 @cindex @samp{expression} defined
1221 @cindex @samp{form} defined
1222 The printed representation of both atoms and lists are called
1223 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1224 The word @dfn{expression} by itself can refer to either the printed
1225 representation, or to the atom or list as it is held internally in the
1226 computer. Often, people use the term @dfn{expression}
1227 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1228 as a synonym for expression.)
1230 Incidentally, the atoms that make up our universe were named such when
1231 they were thought to be indivisible; but it has been found that physical
1232 atoms are not indivisible. Parts can split off an atom or it can
1233 fission into two parts of roughly equal size. Physical atoms were named
1234 prematurely, before their truer nature was found. In Lisp, certain
1235 kinds of atom, such as an array, can be separated into parts; but the
1236 mechanism for doing this is different from the mechanism for splitting a
1237 list. As far as list operations are concerned, the atoms of a list are
1240 As in English, the meanings of the component letters of a Lisp atom
1241 are different from the meaning the letters make as a word. For
1242 example, the word for the South American sloth, the @samp{ai}, is
1243 completely different from the two words, @samp{a}, and @samp{i}.
1245 There are many kinds of atom in nature but only a few in Lisp: for
1246 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1247 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1248 listed in the examples above are all symbols. In everyday Lisp
1249 conversation, the word ``atom'' is not often used, because programmers
1250 usually try to be more specific about what kind of atom they are dealing
1251 with. Lisp programming is mostly about symbols (and sometimes numbers)
1252 within lists. (Incidentally, the preceding three word parenthetical
1253 remark is a proper list in Lisp, since it consists of atoms, which in
1254 this case are symbols, separated by whitespace and enclosed by
1255 parentheses, without any non-Lisp punctuation.)
1258 Text between double quotation marks---even sentences or
1259 paragraphs---is also an atom. Here is an example:
1260 @cindex Text between double quotation marks
1263 '(this list includes "text between quotation marks.")
1266 @cindex @samp{string} defined
1268 In Lisp, all of the quoted text including the punctuation mark and the
1269 blank spaces is a single atom. This kind of atom is called a
1270 @dfn{string} (for `string of characters') and is the sort of thing that
1271 is used for messages that a computer can print for a human to read.
1272 Strings are a different kind of atom than numbers or symbols and are
1275 @node Whitespace in Lists
1276 @subsection Whitespace in Lists
1277 @cindex Whitespace in lists
1280 The amount of whitespace in a list does not matter. From the point of view
1281 of the Lisp language,
1292 is exactly the same as this:
1295 '(this list looks like this)
1298 Both examples show what to Lisp is the same list, the list made up of
1299 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1300 @samp{this} in that order.
1302 Extra whitespace and newlines are designed to make a list more readable
1303 by humans. When Lisp reads the expression, it gets rid of all the extra
1304 whitespace (but it needs to have at least one space between atoms in
1305 order to tell them apart.)
1307 Odd as it seems, the examples we have seen cover almost all of what Lisp
1308 lists look like! Every other list in Lisp looks more or less like one
1309 of these examples, except that the list may be longer and more complex.
1310 In brief, a list is between parentheses, a string is between quotation
1311 marks, a symbol looks like a word, and a number looks like a number.
1312 (For certain situations, square brackets, dots and a few other special
1313 characters may be used; however, we will go quite far without them.)
1316 @subsection GNU Emacs Helps You Type Lists
1317 @cindex Help typing lists
1318 @cindex Formatting help
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 belong to which
1327 list---elements of a sub-list are indented more than the elements of
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's modes.)
1338 @section Run a Program
1339 @cindex Run a program
1340 @cindex Program, running one
1342 @cindex @samp{evaluate} defined
1343 A list in Lisp---any list---is a program ready to run. If you run it
1344 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1345 of three things: do nothing except return to you the list itself; send
1346 you an error message; or, treat the first symbol in the list as a
1347 command to do something. (Usually, of course, it is the last of these
1348 three things that you really want!)
1350 @c use code for the single apostrophe, not samp.
1351 The single apostrophe, @code{'}, that I put in front of some of the
1352 example lists in preceding sections is called a @dfn{quote}; when it
1353 precedes a list, it tells Lisp to do nothing with the list, other than
1354 take it as it is written. But if there is no quote preceding a list,
1355 the first item of the list is special: it is a command for the computer
1356 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1357 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1358 understands that the @code{+} is an instruction to do something with the
1359 rest of the list: add the numbers that follow.
1362 If you are reading this inside of GNU Emacs in Info, here is how you can
1363 evaluate such a list: place your cursor immediately after the right
1364 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1370 @c use code for the number four, not samp.
1372 You will see the number @code{4} appear in the echo area. (In the
1373 jargon, what you have just done is ``evaluate the list.'' The echo area
1374 is the line at the bottom of the screen that displays or ``echoes''
1375 text.) Now try the same thing with a quoted list: place the cursor
1376 right after the following list and type @kbd{C-x C-e}:
1379 '(this is a quoted list)
1383 You will see @code{(this is a quoted list)} appear in the echo area.
1385 @cindex Lisp interpreter, explained
1386 @cindex Interpreter, Lisp, explained
1387 In both cases, what you are doing is giving a command to the program
1388 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1389 interpreter a command to evaluate the expression. The name of the Lisp
1390 interpreter comes from the word for the task done by a human who comes
1391 up with the meaning of an expression---who ``interprets'' it.
1393 You can also evaluate an atom that is not part of a list---one that is
1394 not surrounded by parentheses; again, the Lisp interpreter translates
1395 from the humanly readable expression to the language of the computer.
1396 But before discussing this (@pxref{Variables}), we will discuss what the
1397 Lisp interpreter does when you make an error.
1400 @section Generate an Error Message
1401 @cindex Generate an error message
1402 @cindex Error message generation
1404 Partly so you won't worry if you do it accidentally, we will now give
1405 a command to the Lisp interpreter that generates an error message.
1406 This is a harmless activity; and indeed, we will often try to generate
1407 error messages intentionally. Once you understand the jargon, error
1408 messages can be informative. Instead of being called ``error''
1409 messages, they should be called ``help'' messages. They are like
1410 signposts to a traveler in a strange country; deciphering them can be
1411 hard, but once understood, they can point the way.
1413 The error message is generated by a built-in GNU Emacs debugger. We
1414 will `enter the debugger'. You get out of the debugger by typing @code{q}.
1416 What we will do is evaluate a list that is not quoted and does not
1417 have a meaningful command as its first element. Here is a list almost
1418 exactly the same as the one we just used, but without the single-quote
1419 in front of it. Position the cursor right after it and type @kbd{C-x
1423 (this is an unquoted list)
1428 What you see depends on which version of Emacs you are running. GNU
1429 Emacs version 22 provides more information than version 20 and before.
1430 First, the more recent result of generating an error; then the
1431 earlier, version 20 result.
1435 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1436 you will see the following in it:
1439 A @file{*Backtrace*} window will open up and you should see the
1444 ---------- Buffer: *Backtrace* ----------
1445 Debugger entered--Lisp error: (void-function this)
1446 (this is an unquoted list)
1447 eval((this is an unquoted list))
1448 eval-last-sexp-1(nil)
1450 call-interactively(eval-last-sexp)
1451 ---------- Buffer: *Backtrace* ----------
1457 Your cursor will be in this window (you may have to wait a few seconds
1458 before it becomes visible). To quit the debugger and make the
1459 debugger window go away, type:
1466 Please type @kbd{q} right now, so you become confident that you can
1467 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1470 @cindex @samp{function} defined
1471 Based on what we already know, we can almost read this error message.
1473 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1474 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1475 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1476 an abbreviation for `evaluate' and @code{sexp} is an abbreviation for
1477 `symbolic expression'. The command means `evaluate last symbolic
1478 expression', which is the expression just before your cursor.
1480 Each line above tells you what the Lisp interpreter evaluated next.
1481 The most recent action is at the top. The buffer is called the
1482 @file{*Backtrace*} buffer because it enables you to track Emacs
1486 At the top of the @file{*Backtrace*} buffer, you see the line:
1489 Debugger entered--Lisp error: (void-function this)
1493 The Lisp interpreter tried to evaluate the first atom of the list, the
1494 word @samp{this}. It is this action that generated the error message
1495 @samp{void-function this}.
1497 The message contains the words @samp{void-function} and @samp{this}.
1499 @cindex @samp{function} defined
1500 The word @samp{function} was mentioned once before. It is a very
1501 important word. For our purposes, we can define it by saying that a
1502 @dfn{function} is a set of instructions to the computer that tell the
1503 computer to do something.
1505 Now we can begin to understand the error message: @samp{void-function
1506 this}. The function (that is, the word @samp{this}) does not have a
1507 definition of any set of instructions for the computer to carry out.
1509 The slightly odd word, @samp{void-function}, is designed to cover the
1510 way Emacs Lisp is implemented, which is that when a symbol does not
1511 have a function definition attached to it, the place that should
1512 contain the instructions is `void'.
1514 On the other hand, since we were able to add 2 plus 2 successfully, by
1515 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1516 have a set of instructions for the computer to obey and those
1517 instructions must be to add the numbers that follow the @code{+}.
1519 It is possible to prevent Emacs entering the debugger in cases like
1520 this. We do not explain how to do that here, but we will mention what
1521 the result looks like, because you may encounter a similar situation
1522 if there is a bug in some Emacs code that you are using. In such
1523 cases, you will see only one line of error message; it will appear in
1524 the echo area and look like this:
1527 Symbol's function definition is void:@: this
1532 (Also, your terminal may beep at you---some do, some don't; and others
1533 blink. This is just a device to get your attention.)
1535 The message goes away as soon as you type a key, even just to
1538 We know the meaning of the word @samp{Symbol}. It refers to the first
1539 atom of the list, the word @samp{this}. The word @samp{function}
1540 refers to the instructions that tell the computer what to do.
1541 (Technically, the symbol tells the computer where to find the
1542 instructions, but this is a complication we can ignore for the
1545 The error message can be understood: @samp{Symbol's function
1546 definition is void:@: this}. The symbol (that is, the word
1547 @samp{this}) lacks instructions for the computer to carry out.
1549 @node Names & Definitions
1550 @section Symbol Names and Function Definitions
1551 @cindex Symbol names
1553 We can articulate another characteristic of Lisp based on what we have
1554 discussed so far---an important characteristic: a symbol, like
1555 @code{+}, is not itself the set of instructions for the computer to
1556 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1557 of locating the definition or set of instructions. What we see is the
1558 name through which the instructions can be found. Names of people
1559 work the same way. I can be referred to as @samp{Bob}; however, I am
1560 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1561 consciousness consistently associated with a particular life-form.
1562 The name is not me, but it can be used to refer to me.
1564 In Lisp, one set of instructions can be attached to several names.
1565 For example, the computer instructions for adding numbers can be
1566 linked to the symbol @code{plus} as well as to the symbol @code{+}
1567 (and are in some dialects of Lisp). Among humans, I can be referred
1568 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1570 On the other hand, a symbol can have only one function definition
1571 attached to it at a time. Otherwise, the computer would be confused as
1572 to which definition to use. If this were the case among people, only
1573 one person in the world could be named @samp{Bob}. However, the function
1574 definition to which the name refers can be changed readily.
1575 (@xref{Install, , Install a Function Definition}.)
1577 Since Emacs Lisp is large, it is customary to name symbols in a way
1578 that identifies the part of Emacs to which the function belongs.
1579 Thus, all the names for functions that deal with Texinfo start with
1580 @samp{texinfo-} and those for functions that deal with reading mail
1581 start with @samp{rmail-}.
1583 @node Lisp Interpreter
1584 @section The Lisp Interpreter
1585 @cindex Lisp interpreter, what it does
1586 @cindex Interpreter, what it does
1588 Based on what we have seen, we can now start to figure out what the
1589 Lisp interpreter does when we command it to evaluate a list.
1590 First, it looks to see whether there is a quote before the list; if
1591 there is, the interpreter just gives us the list. On the other
1592 hand, if there is no quote, the interpreter looks at the first element
1593 in the list and sees whether it has a function definition. If it does,
1594 the interpreter carries out the instructions in the function definition.
1595 Otherwise, the interpreter prints an error message.
1597 This is how Lisp works. Simple. There are added complications which we
1598 will get to in a minute, but these are the fundamentals. Of course, to
1599 write Lisp programs, you need to know how to write function definitions
1600 and attach them to names, and how to do this without confusing either
1601 yourself or the computer.
1604 * Complications:: Variables, Special forms, Lists within.
1605 * Byte Compiling:: Specially processing code for speed.
1610 @unnumberedsubsec Complications
1613 Now, for the first complication. In addition to lists, the Lisp
1614 interpreter can evaluate a symbol that is not quoted and does not have
1615 parentheses around it. The Lisp interpreter will attempt to determine
1616 the symbol's value as a @dfn{variable}. This situation is described
1617 in the section on variables. (@xref{Variables}.)
1619 @cindex Special form
1620 The second complication occurs because some functions are unusual and do
1621 not work in the usual manner. Those that don't are called @dfn{special
1622 forms}. They are used for special jobs, like defining a function, and
1623 there are not many of them. In the next few chapters, you will be
1624 introduced to several of the more important special forms.
1626 The third and final complication is this: if the function that the
1627 Lisp interpreter is looking at is not a special form, and if it is part
1628 of a list, the Lisp interpreter looks to see whether the list has a list
1629 inside of it. If there is an inner list, the Lisp interpreter first
1630 figures out what it should do with the inside list, and then it works on
1631 the outside list. If there is yet another list embedded inside the
1632 inner list, it works on that one first, and so on. It always works on
1633 the innermost list first. The interpreter works on the innermost list
1634 first, to evaluate the result of that list. The result may be
1635 used by the enclosing expression.
1637 Otherwise, the interpreter works left to right, from one expression to
1640 @node Byte Compiling
1641 @subsection Byte Compiling
1642 @cindex Byte compiling
1644 One other aspect of interpreting: the Lisp interpreter is able to
1645 interpret two kinds of entity: humanly readable code, on which we will
1646 focus exclusively, and specially processed code, called @dfn{byte
1647 compiled} code, which is not humanly readable. Byte compiled code
1648 runs faster than humanly readable code.
1650 You can transform humanly readable code into byte compiled code by
1651 running one of the compile commands such as @code{byte-compile-file}.
1652 Byte compiled code is usually stored in a file that ends with a
1653 @file{.elc} extension rather than a @file{.el} extension. You will
1654 see both kinds of file in the @file{emacs/lisp} directory; the files
1655 to read are those with @file{.el} extensions.
1657 As a practical matter, for most things you might do to customize or
1658 extend Emacs, you do not need to byte compile; and I will not discuss
1659 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1660 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1667 When the Lisp interpreter works on an expression, the term for the
1668 activity is called @dfn{evaluation}. We say that the interpreter
1669 `evaluates the expression'. I've used this term several times before.
1670 The word comes from its use in everyday language, `to ascertain the
1671 value or amount of; to appraise', according to @cite{Webster's New
1672 Collegiate Dictionary}.
1675 * How the Interpreter Acts:: Returns and Side Effects...
1676 * Evaluating Inner Lists:: Lists within lists...
1680 @node How the Interpreter Acts
1681 @unnumberedsubsec How the Lisp Interpreter Acts
1684 @cindex @samp{returned value} explained
1685 After evaluating an expression, the Lisp interpreter will most likely
1686 @dfn{return} the value that the computer produces by carrying out the
1687 instructions it found in the function definition, or perhaps it will
1688 give up on that function and produce an error message. (The interpreter
1689 may also find itself tossed, so to speak, to a different function or it
1690 may attempt to repeat continually what it is doing for ever and ever in
1691 what is called an `infinite loop'. These actions are less common; and
1692 we can ignore them.) Most frequently, the interpreter returns a value.
1694 @cindex @samp{side effect} defined
1695 At the same time the interpreter returns a value, it may do something
1696 else as well, such as move a cursor or copy a file; this other kind of
1697 action is called a @dfn{side effect}. Actions that we humans think are
1698 important, such as printing results, are often ``side effects'' to the
1699 Lisp interpreter. The jargon can sound peculiar, but it turns out that
1700 it is fairly easy to learn to use side effects.
1702 In summary, evaluating a symbolic expression most commonly causes the
1703 Lisp interpreter to return a value and perhaps carry out a side effect;
1704 or else produce an error.
1706 @node Evaluating Inner Lists
1707 @subsection Evaluating Inner Lists
1708 @cindex Inner list evaluation
1709 @cindex Evaluating inner lists
1711 If evaluation applies to a list that is inside another list, the outer
1712 list may use the value returned by the first evaluation as information
1713 when the outer list is evaluated. This explains why inner expressions
1714 are evaluated first: the values they return are used by the outer
1718 We can investigate this process by evaluating another addition example.
1719 Place your cursor after the following expression and type @kbd{C-x C-e}:
1726 The number 8 will appear in the echo area.
1728 What happens is that the Lisp interpreter first evaluates the inner
1729 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1730 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1731 returns the value 8. Since there are no more enclosing expressions to
1732 evaluate, the interpreter prints that value in the echo area.
1734 Now it is easy to understand the name of the command invoked by the
1735 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1736 letters @code{sexp} are an abbreviation for `symbolic expression', and
1737 @code{eval} is an abbreviation for `evaluate'. The command means
1738 `evaluate last symbolic expression'.
1740 As an experiment, you can try evaluating the expression by putting the
1741 cursor at the beginning of the next line immediately following the
1742 expression, or inside the expression.
1745 Here is another copy of the expression:
1752 If you place the cursor at the beginning of the blank line that
1753 immediately follows the expression and type @kbd{C-x C-e}, you will
1754 still get the value 8 printed in the echo area. Now try putting the
1755 cursor inside the expression. If you put it right after the next to
1756 last parenthesis (so it appears to sit on top of the last parenthesis),
1757 you will get a 6 printed in the echo area! This is because the command
1758 evaluates the expression @code{(+ 3 3)}.
1760 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1761 you will get the number itself. In Lisp, if you evaluate a number, you
1762 get the number itself---this is how numbers differ from symbols. If you
1763 evaluate a list starting with a symbol like @code{+}, you will get a
1764 value returned that is the result of the computer carrying out the
1765 instructions in the function definition attached to that name. If a
1766 symbol by itself is evaluated, something different happens, as we will
1767 see in the next section.
1773 In Emacs Lisp, a symbol can have a value attached to it just as it can
1774 have a function definition attached to it. The two are different.
1775 The function definition is a set of instructions that a computer will
1776 obey. A value, on the other hand, is something, such as number or a
1777 name, that can vary (which is why such a symbol is called a variable).
1778 The value of a symbol can be any expression in Lisp, such as a symbol,
1779 number, list, or string. A symbol that has a value is often called a
1782 A symbol can have both a function definition and a value attached to
1783 it at the same time. Or it can have just one or the other.
1784 The two are separate. This is somewhat similar
1785 to the way the name Cambridge can refer to the city in Massachusetts
1786 and have some information attached to the name as well, such as
1787 ``great programming center''.
1790 (Incidentally, in Emacs Lisp, a symbol can have two
1791 other things attached to it, too: a property list and a documentation
1792 string; these are discussed later.)
1795 Another way to think about this is to imagine a symbol as being a chest
1796 of drawers. The function definition is put in one drawer, the value in
1797 another, and so on. What is put in the drawer holding the value can be
1798 changed without affecting the contents of the drawer holding the
1799 function definition, and vice-verse.
1802 * fill-column Example::
1803 * Void Function:: The error message for a symbol
1805 * Void Variable:: The error message for a symbol without a value.
1809 @node fill-column Example
1810 @unnumberedsubsec @code{fill-column}, an Example Variable
1813 @findex fill-column, @r{an example variable}
1814 @cindex Example variable, @code{fill-column}
1815 @cindex Variable, example of, @code{fill-column}
1816 The variable @code{fill-column} illustrates a symbol with a value
1817 attached to it: in every GNU Emacs buffer, this symbol is set to some
1818 value, usually 72 or 70, but sometimes to some other value. To find the
1819 value of this symbol, evaluate it by itself. If you are reading this in
1820 Info inside of GNU Emacs, you can do this by putting the cursor after
1821 the symbol and typing @kbd{C-x C-e}:
1828 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1829 area. This is the value for which @code{fill-column} is set for me as I
1830 write this. It may be different for you in your Info buffer. Notice
1831 that the value returned as a variable is printed in exactly the same way
1832 as the value returned by a function carrying out its instructions. From
1833 the point of view of the Lisp interpreter, a value returned is a value
1834 returned. What kind of expression it came from ceases to matter once
1837 A symbol can have any value attached to it or, to use the jargon, we can
1838 @dfn{bind} the variable to a value: to a number, such as 72; to a
1839 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1840 oak)}; we can even bind a variable to a function definition.
1842 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1843 Setting the Value of a Variable}, for information about one way to do
1847 @subsection Error Message for a Symbol Without a Function
1848 @cindex Symbol without function error
1849 @cindex Error for symbol without function
1851 When we evaluated @code{fill-column} to find its value as a variable,
1852 we did not place parentheses around the word. This is because we did
1853 not intend to use it as a function name.
1855 If @code{fill-column} were the first or only element of a list, the
1856 Lisp interpreter would attempt to find the function definition
1857 attached to it. But @code{fill-column} has no function definition.
1858 Try evaluating this:
1866 You will create a @file{*Backtrace*} buffer that says:
1870 ---------- Buffer: *Backtrace* ----------
1871 Debugger entered--Lisp error: (void-function fill-column)
1874 eval-last-sexp-1(nil)
1876 call-interactively(eval-last-sexp)
1877 ---------- Buffer: *Backtrace* ----------
1882 (Remember, to quit the debugger and make the debugger window go away,
1883 type @kbd{q} in the @file{*Backtrace*} buffer.)
1887 In GNU Emacs 20 and before, you will produce an error message that says:
1890 Symbol's function definition is void:@: fill-column
1894 (The message will go away as soon as you move the cursor or type
1899 @subsection Error Message for a Symbol Without a Value
1900 @cindex Symbol without value error
1901 @cindex Error for symbol without value
1903 If you attempt to evaluate a symbol that does not have a value bound to
1904 it, you will receive an error message. You can see this by
1905 experimenting with our 2 plus 2 addition. In the following expression,
1906 put your cursor right after the @code{+}, before the first number 2,
1915 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1920 ---------- Buffer: *Backtrace* ----------
1921 Debugger entered--Lisp error: (void-variable +)
1923 eval-last-sexp-1(nil)
1925 call-interactively(eval-last-sexp)
1926 ---------- Buffer: *Backtrace* ----------
1931 (Again, you can quit the debugger by
1932 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1934 This backtrace is different from the very first error message we saw,
1935 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1936 In this case, the function does not have a value as a variable; while
1937 in the other error message, the function (the word `this') did not
1940 In this experiment with the @code{+}, what we did was cause the Lisp
1941 interpreter to evaluate the @code{+} and look for the value of the
1942 variable instead of the function definition. We did this by placing the
1943 cursor right after the symbol rather than after the parenthesis of the
1944 enclosing list as we did before. As a consequence, the Lisp interpreter
1945 evaluated the preceding s-expression, which in this case was
1948 Since @code{+} does not have a value bound to it, just the function
1949 definition, the error message reported that the symbol's value as a
1954 In GNU Emacs version 20 and before, your error message will say:
1957 Symbol's value as variable is void:@: +
1961 The meaning is the same as in GNU Emacs 22.
1967 @cindex Passing information to functions
1969 To see how information is passed to functions, let's look again at
1970 our old standby, the addition of two plus two. In Lisp, this is written
1977 If you evaluate this expression, the number 4 will appear in your echo
1978 area. What the Lisp interpreter does is add the numbers that follow
1981 @cindex @samp{argument} defined
1982 The numbers added by @code{+} are called the @dfn{arguments} of the
1983 function @code{+}. These numbers are the information that is given to
1984 or @dfn{passed} to the function.
1986 The word `argument' comes from the way it is used in mathematics and
1987 does not refer to a disputation between two people; instead it refers to
1988 the information presented to the function, in this case, to the
1989 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1990 that follow the function. The values returned by the evaluation of
1991 these atoms or lists are passed to the function. Different functions
1992 require different numbers of arguments; some functions require none at
1993 all.@footnote{It is curious to track the path by which the word `argument'
1994 came to have two different meanings, one in mathematics and the other in
1995 everyday English. According to the @cite{Oxford English Dictionary},
1996 the word derives from the Latin for @samp{to make clear, prove}; thus it
1997 came to mean, by one thread of derivation, `the evidence offered as
1998 proof', which is to say, `the information offered', which led to its
1999 meaning in Lisp. But in the other thread of derivation, it came to mean
2000 `to assert in a manner against which others may make counter
2001 assertions', which led to the meaning of the word as a disputation.
2002 (Note here that the English word has two different definitions attached
2003 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
2004 have two different function definitions at the same time.)}
2007 * Data types:: Types of data passed to a function.
2008 * Args as Variable or List:: An argument can be the value
2009 of a variable or list.
2010 * Variable Number of Arguments:: Some functions may take a
2011 variable number of arguments.
2012 * Wrong Type of Argument:: Passing an argument of the wrong type
2014 * message:: A useful function for sending messages.
2018 @subsection Arguments' Data Types
2020 @cindex Types of data
2021 @cindex Arguments' data types
2023 The type of data that should be passed to a function depends on what
2024 kind of information it uses. The arguments to a function such as
2025 @code{+} must have values that are numbers, since @code{+} adds numbers.
2026 Other functions use different kinds of data for their arguments.
2030 For example, the @code{concat} function links together or unites two or
2031 more strings of text to produce a string. The arguments are strings.
2032 Concatenating the two character strings @code{abc}, @code{def} produces
2033 the single string @code{abcdef}. This can be seen by evaluating the
2037 (concat "abc" "def")
2041 The value produced by evaluating this expression is @code{"abcdef"}.
2043 A function such as @code{substring} uses both a string and numbers as
2044 arguments. The function returns a part of the string, a substring of
2045 the first argument. This function takes three arguments. Its first
2046 argument is the string of characters, the second and third arguments are
2047 numbers that indicate the beginning and end of the substring. The
2048 numbers are a count of the number of characters (including spaces and
2049 punctuation) from the beginning of the string.
2052 For example, if you evaluate the following:
2055 (substring "The quick brown fox jumped." 16 19)
2059 you will see @code{"fox"} appear in the echo area. The arguments are the
2060 string and the two numbers.
2062 Note that the string passed to @code{substring} is a single atom even
2063 though it is made up of several words separated by spaces. Lisp counts
2064 everything between the two quotation marks as part of the string,
2065 including the spaces. You can think of the @code{substring} function as
2066 a kind of `atom smasher' since it takes an otherwise indivisible atom
2067 and extracts a part. However, @code{substring} is only able to extract
2068 a substring from an argument that is a string, not from another type of
2069 atom such as a number or symbol.
2071 @node Args as Variable or List
2072 @subsection An Argument as the Value of a Variable or List
2074 An argument can be a symbol that returns a value when it is evaluated.
2075 For example, when the symbol @code{fill-column} by itself is evaluated,
2076 it returns a number. This number can be used in an addition.
2079 Position the cursor after the following expression and type @kbd{C-x
2087 The value will be a number two more than what you get by evaluating
2088 @code{fill-column} alone. For me, this is 74, because my value of
2089 @code{fill-column} is 72.
2091 As we have just seen, an argument can be a symbol that returns a value
2092 when evaluated. In addition, an argument can be a list that returns a
2093 value when it is evaluated. For example, in the following expression,
2094 the arguments to the function @code{concat} are the strings
2095 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2096 @code{(number-to-string (+ 2 fill-column))}.
2098 @c For GNU Emacs 22, need number-to-string
2100 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2104 If you evaluate this expression---and if, as with my Emacs,
2105 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2106 appear in the echo area. (Note that you must put spaces after the
2107 word @samp{The} and before the word @samp{red} so they will appear in
2108 the final string. The function @code{number-to-string} converts the
2109 integer that the addition function returns to a string.
2110 @code{number-to-string} is also known as @code{int-to-string}.)
2112 @node Variable Number of Arguments
2113 @subsection Variable Number of Arguments
2114 @cindex Variable number of arguments
2115 @cindex Arguments, variable number of
2117 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2118 number of arguments. (The @code{*} is the symbol for multiplication.)
2119 This can be seen by evaluating each of the following expressions in
2120 the usual way. What you will see in the echo area is printed in this
2121 text after @samp{@result{}}, which you may read as `evaluates to'.
2124 In the first set, the functions have no arguments:
2135 In this set, the functions have one argument each:
2146 In this set, the functions have three arguments each:
2150 (+ 3 4 5) @result{} 12
2152 (* 3 4 5) @result{} 60
2156 @node Wrong Type of Argument
2157 @subsection Using the Wrong Type Object as an Argument
2158 @cindex Wrong type of argument
2159 @cindex Argument, wrong type of
2161 When a function is passed an argument of the wrong type, the Lisp
2162 interpreter produces an error message. For example, the @code{+}
2163 function expects the values of its arguments to be numbers. As an
2164 experiment we can pass it the quoted symbol @code{hello} instead of a
2165 number. Position the cursor after the following expression and type
2173 When you do this you will generate an error message. What has happened
2174 is that @code{+} has tried to add the 2 to the value returned by
2175 @code{'hello}, but the value returned by @code{'hello} is the symbol
2176 @code{hello}, not a number. Only numbers can be added. So @code{+}
2177 could not carry out its addition.
2180 You will create and enter a @file{*Backtrace*} buffer that says:
2185 ---------- Buffer: *Backtrace* ----------
2186 Debugger entered--Lisp error:
2187 (wrong-type-argument number-or-marker-p hello)
2189 eval((+ 2 (quote hello)))
2190 eval-last-sexp-1(nil)
2192 call-interactively(eval-last-sexp)
2193 ---------- Buffer: *Backtrace* ----------
2198 As usual, the error message tries to be helpful and makes sense after you
2199 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2200 the abbreviation @code{'hello}.}
2202 The first part of the error message is straightforward; it says
2203 @samp{wrong type argument}. Next comes the mysterious jargon word
2204 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2205 kind of argument the @code{+} expected.
2207 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2208 trying to determine whether the information presented it (the value of
2209 the argument) is a number or a marker (a special object representing a
2210 buffer position). What it does is test to see whether the @code{+} is
2211 being given numbers to add. It also tests to see whether the
2212 argument is something called a marker, which is a specific feature of
2213 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2214 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2215 its position is kept as a marker. The mark can be considered a
2216 number---the number of characters the location is from the beginning
2217 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2218 numeric value of marker positions as numbers.
2220 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2221 practice started in the early days of Lisp programming. The @samp{p}
2222 stands for `predicate'. In the jargon used by the early Lisp
2223 researchers, a predicate refers to a function to determine whether some
2224 property is true or false. So the @samp{p} tells us that
2225 @code{number-or-marker-p} is the name of a function that determines
2226 whether it is true or false that the argument supplied is a number or
2227 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2228 a function that tests whether its argument has the value of zero, and
2229 @code{listp}, a function that tests whether its argument is a list.
2231 Finally, the last part of the error message is the symbol @code{hello}.
2232 This is the value of the argument that was passed to @code{+}. If the
2233 addition had been passed the correct type of object, the value passed
2234 would have been a number, such as 37, rather than a symbol like
2235 @code{hello}. But then you would not have got the error message.
2239 In GNU Emacs version 20 and before, the echo area displays an error
2243 Wrong type argument:@: number-or-marker-p, hello
2246 This says, in different words, the same as the top line of the
2247 @file{*Backtrace*} buffer.
2251 @subsection The @code{message} Function
2254 Like @code{+}, the @code{message} function takes a variable number of
2255 arguments. It is used to send messages to the user and is so useful
2256 that we will describe it here.
2259 A message is printed in the echo area. For example, you can print a
2260 message in your echo area by evaluating the following list:
2263 (message "This message appears in the echo area!")
2266 The whole string between double quotation marks is a single argument
2267 and is printed @i{in toto}. (Note that in this example, the message
2268 itself will appear in the echo area within double quotes; that is
2269 because you see the value returned by the @code{message} function. In
2270 most uses of @code{message} in programs that you write, the text will
2271 be printed in the echo area as a side-effect, without the quotes.
2272 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2273 detail}, for an example of this.)
2275 However, if there is a @samp{%s} in the quoted string of characters, the
2276 @code{message} function does not print the @samp{%s} as such, but looks
2277 to the argument that follows the string. It evaluates the second
2278 argument and prints the value at the location in the string where the
2282 You can see this by positioning the cursor after the following
2283 expression and typing @kbd{C-x C-e}:
2286 (message "The name of this buffer is: %s." (buffer-name))
2290 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2291 echo area. The function @code{buffer-name} returns the name of the
2292 buffer as a string, which the @code{message} function inserts in place
2295 To print a value as an integer, use @samp{%d} in the same way as
2296 @samp{%s}. For example, to print a message in the echo area that
2297 states the value of the @code{fill-column}, evaluate the following:
2300 (message "The value of fill-column is %d." fill-column)
2304 On my system, when I evaluate this list, @code{"The value of
2305 fill-column is 72."} appears in my echo area@footnote{Actually, you
2306 can use @code{%s} to print a number. It is non-specific. @code{%d}
2307 prints only the part of a number left of a decimal point, and not
2308 anything that is not a number.}.
2310 If there is more than one @samp{%s} in the quoted string, the value of
2311 the first argument following the quoted string is printed at the
2312 location of the first @samp{%s} and the value of the second argument is
2313 printed at the location of the second @samp{%s}, and so on.
2316 For example, if you evaluate the following,
2320 (message "There are %d %s in the office!"
2321 (- fill-column 14) "pink elephants")
2326 a rather whimsical message will appear in your echo area. On my system
2327 it says, @code{"There are 58 pink elephants in the office!"}.
2329 The expression @code{(- fill-column 14)} is evaluated and the resulting
2330 number is inserted in place of the @samp{%d}; and the string in double
2331 quotes, @code{"pink elephants"}, is treated as a single argument and
2332 inserted in place of the @samp{%s}. (That is to say, a string between
2333 double quotes evaluates to itself, like a number.)
2335 Finally, here is a somewhat complex example that not only illustrates
2336 the computation of a number, but also shows how you can use an
2337 expression within an expression to generate the text that is substituted
2342 (message "He saw %d %s"
2346 "The quick brown foxes jumped." 16 21)
2351 In this example, @code{message} has three arguments: the string,
2352 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2353 the expression beginning with the function @code{concat}. The value
2354 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2355 in place of the @samp{%d}; and the value returned by the expression
2356 beginning with @code{concat} is inserted in place of the @samp{%s}.
2358 When your fill column is 70 and you evaluate the expression, the
2359 message @code{"He saw 38 red foxes leaping."} appears in your echo
2363 @section Setting the Value of a Variable
2364 @cindex Variable, setting value
2365 @cindex Setting value of variable
2367 @cindex @samp{bind} defined
2368 There are several ways by which a variable can be given a value. One of
2369 the ways is to use either the function @code{set} or the function
2370 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2371 jargon for this process is to @dfn{bind} a variable to a value.)
2373 The following sections not only describe how @code{set} and @code{setq}
2374 work but also illustrate how arguments are passed.
2377 * Using set:: Setting values.
2378 * Using setq:: Setting a quoted value.
2379 * Counting:: Using @code{setq} to count.
2383 @subsection Using @code{set}
2386 To set the value of the symbol @code{flowers} to the list @code{'(rose
2387 violet daisy buttercup)}, evaluate the following expression by
2388 positioning the cursor after the expression and typing @kbd{C-x C-e}.
2391 (set 'flowers '(rose violet daisy buttercup))
2395 The list @code{(rose violet daisy buttercup)} will appear in the echo
2396 area. This is what is @emph{returned} by the @code{set} function. As a
2397 side effect, the symbol @code{flowers} is bound to the list; that is,
2398 the symbol @code{flowers}, which can be viewed as a variable, is given
2399 the list as its value. (This process, by the way, illustrates how a
2400 side effect to the Lisp interpreter, setting the value, can be the
2401 primary effect that we humans are interested in. This is because every
2402 Lisp function must return a value if it does not get an error, but it
2403 will only have a side effect if it is designed to have one.)
2405 After evaluating the @code{set} expression, you can evaluate the symbol
2406 @code{flowers} and it will return the value you just set. Here is the
2407 symbol. Place your cursor after it and type @kbd{C-x C-e}.
2414 When you evaluate @code{flowers}, the list
2415 @code{(rose violet daisy buttercup)} appears in the echo area.
2417 Incidentally, if you evaluate @code{'flowers}, the variable with a quote
2418 in front of it, what you will see in the echo area is the symbol itself,
2419 @code{flowers}. Here is the quoted symbol, so you can try this:
2425 Note also, that when you use @code{set}, you need to quote both
2426 arguments to @code{set}, unless you want them evaluated. Since we do
2427 not want either argument evaluated, neither the variable
2428 @code{flowers} nor the list @code{(rose violet daisy buttercup)}, both
2429 are quoted. (When you use @code{set} without quoting its first
2430 argument, the first argument is evaluated before anything else is
2431 done. If you did this and @code{flowers} did not have a value
2432 already, you would get an error message that the @samp{Symbol's value
2433 as variable is void}; on the other hand, if @code{flowers} did return
2434 a value after it was evaluated, the @code{set} would attempt to set
2435 the value that was returned. There are situations where this is the
2436 right thing for the function to do; but such situations are rare.)
2439 @subsection Using @code{setq}
2442 As a practical matter, you almost always quote the first argument to
2443 @code{set}. The combination of @code{set} and a quoted first argument
2444 is so common that it has its own name: the special form @code{setq}.
2445 This special form is just like @code{set} except that the first argument
2446 is quoted automatically, so you don't need to type the quote mark
2447 yourself. Also, as an added convenience, @code{setq} permits you to set
2448 several different variables to different values, all in one expression.
2450 To set the value of the variable @code{carnivores} to the list
2451 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2455 (setq carnivores '(lion tiger leopard))
2459 This is exactly the same as using @code{set} except the first argument
2460 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2461 means @code{quote}.)
2464 With @code{set}, the expression would look like this:
2467 (set 'carnivores '(lion tiger leopard))
2470 Also, @code{setq} can be used to assign different values to
2471 different variables. The first argument is bound to the value
2472 of the second argument, the third argument is bound to the value of the
2473 fourth argument, and so on. For example, you could use the following to
2474 assign a list of trees to the symbol @code{trees} and a list of herbivores
2475 to the symbol @code{herbivores}:
2479 (setq trees '(pine fir oak maple)
2480 herbivores '(gazelle antelope zebra))
2485 (The expression could just as well have been on one line, but it might
2486 not have fit on a page; and humans find it easier to read nicely
2489 Although I have been using the term `assign', there is another way of
2490 thinking about the workings of @code{set} and @code{setq}; and that is to
2491 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2492 list. This latter way of thinking is very common and in forthcoming
2493 chapters we shall come upon at least one symbol that has `pointer' as
2494 part of its name. The name is chosen because the symbol has a value,
2495 specifically a list, attached to it; or, expressed another way,
2496 the symbol is set to ``point'' to the list.
2499 @subsection Counting
2502 Here is an example that shows how to use @code{setq} in a counter. You
2503 might use this to count how many times a part of your program repeats
2504 itself. First set a variable to zero; then add one to the number each
2505 time the program repeats itself. To do this, you need a variable that
2506 serves as a counter, and two expressions: an initial @code{setq}
2507 expression that sets the counter variable to zero; and a second
2508 @code{setq} expression that increments the counter each time it is
2513 (setq counter 0) ; @r{Let's call this the initializer.}
2515 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2517 counter ; @r{This is the counter.}
2522 (The text following the @samp{;} are comments. @xref{Change a
2523 defun, , Change a Function Definition}.)
2525 If you evaluate the first of these expressions, the initializer,
2526 @code{(setq counter 0)}, and then evaluate the third expression,
2527 @code{counter}, the number @code{0} will appear in the echo area. If
2528 you then evaluate the second expression, the incrementer, @code{(setq
2529 counter (+ counter 1))}, the counter will get the value 1. So if you
2530 again evaluate @code{counter}, the number @code{1} will appear in the
2531 echo area. Each time you evaluate the second expression, the value of
2532 the counter will be incremented.
2534 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2535 the Lisp interpreter first evaluates the innermost list; this is the
2536 addition. In order to evaluate this list, it must evaluate the variable
2537 @code{counter} and the number @code{1}. When it evaluates the variable
2538 @code{counter}, it receives its current value. It passes this value and
2539 the number @code{1} to the @code{+} which adds them together. The sum
2540 is then returned as the value of the inner list and passed to the
2541 @code{setq} which sets the variable @code{counter} to this new value.
2542 Thus, the value of the variable, @code{counter}, is changed.
2547 Learning Lisp is like climbing a hill in which the first part is the
2548 steepest. You have now climbed the most difficult part; what remains
2549 becomes easier as you progress onwards.
2557 Lisp programs are made up of expressions, which are lists or single atoms.
2560 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2561 surrounded by parentheses. A list can be empty.
2564 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2565 character symbols like @code{+}, strings of characters between double
2566 quotation marks, or numbers.
2569 A number evaluates to itself.
2572 A string between double quotes also evaluates to itself.
2575 When you evaluate a symbol by itself, its value is returned.
2578 When you evaluate a list, the Lisp interpreter looks at the first symbol
2579 in the list and then at the function definition bound to that symbol.
2580 Then the instructions in the function definition are carried out.
2583 A single quotation mark,
2590 , tells the Lisp interpreter that it should
2591 return the following expression as written, and not evaluate it as it
2592 would if the quote were not there.
2595 Arguments are the information passed to a function. The arguments to a
2596 function are computed by evaluating the rest of the elements of the list
2597 of which the function is the first element.
2600 A function always returns a value when it is evaluated (unless it gets
2601 an error); in addition, it may also carry out some action called a
2602 ``side effect''. In many cases, a function's primary purpose is to
2603 create a side effect.
2606 @node Error Message Exercises
2609 A few simple exercises:
2613 Generate an error message by evaluating an appropriate symbol that is
2614 not within parentheses.
2617 Generate an error message by evaluating an appropriate symbol that is
2618 between parentheses.
2621 Create a counter that increments by two rather than one.
2624 Write an expression that prints a message in the echo area when
2628 @node Practicing Evaluation
2629 @chapter Practicing Evaluation
2630 @cindex Practicing evaluation
2631 @cindex Evaluation practice
2633 Before learning how to write a function definition in Emacs Lisp, it is
2634 useful to spend a little time evaluating various expressions that have
2635 already been written. These expressions will be lists with the
2636 functions as their first (and often only) element. Since some of the
2637 functions associated with buffers are both simple and interesting, we
2638 will start with those. In this section, we will evaluate a few of
2639 these. In another section, we will study the code of several other
2640 buffer-related functions, to see how they were written.
2643 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2645 * Buffer Names:: Buffers and files are different.
2646 * Getting Buffers:: Getting a buffer itself, not merely its name.
2647 * Switching Buffers:: How to change to another buffer.
2648 * Buffer Size & Locations:: Where point is located and the size of
2650 * Evaluation Exercise::
2654 @node How to Evaluate
2655 @unnumberedsec How to Evaluate
2658 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2659 command to move the cursor or to scroll the screen, @i{you are evaluating
2660 an expression,} the first element of which is a function. @i{This is
2663 @cindex @samp{interactive function} defined
2664 @cindex @samp{command} defined
2665 When you type keys, you cause the Lisp interpreter to evaluate an
2666 expression and that is how you get your results. Even typing plain text
2667 involves evaluating an Emacs Lisp function, in this case, one that uses
2668 @code{self-insert-command}, which simply inserts the character you
2669 typed. The functions you evaluate by typing keystrokes are called
2670 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2671 interactive will be illustrated in the chapter on how to write function
2672 definitions. @xref{Interactive, , Making a Function Interactive}.
2674 In addition to typing keyboard commands, we have seen a second way to
2675 evaluate an expression: by positioning the cursor after a list and
2676 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2677 section. There are other ways to evaluate an expression as well; these
2678 will be described as we come to them.
2680 Besides being used for practicing evaluation, the functions shown in the
2681 next few sections are important in their own right. A study of these
2682 functions makes clear the distinction between buffers and files, how to
2683 switch to a buffer, and how to determine a location within it.
2686 @section Buffer Names
2688 @findex buffer-file-name
2690 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2691 the difference between a file and a buffer. When you evaluate the
2692 following expression, @code{(buffer-name)}, the name of the buffer
2693 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2694 the name of the file to which the buffer refers appears in the echo
2695 area. Usually, the name returned by @code{(buffer-name)} is the same as
2696 the name of the file to which it refers, and the name returned by
2697 @code{(buffer-file-name)} is the full path-name of the file.
2699 A file and a buffer are two different entities. A file is information
2700 recorded permanently in the computer (unless you delete it). A buffer,
2701 on the other hand, is information inside of Emacs that will vanish at
2702 the end of the editing session (or when you kill the buffer). Usually,
2703 a buffer contains information that you have copied from a file; we say
2704 the buffer is @dfn{visiting} that file. This copy is what you work on
2705 and modify. Changes to the buffer do not change the file, until you
2706 save the buffer. When you save the buffer, the buffer is copied to the file
2707 and is thus saved permanently.
2710 If you are reading this in Info inside of GNU Emacs, you can evaluate
2711 each of the following expressions by positioning the cursor after it and
2712 typing @kbd{C-x C-e}.
2723 When I do this in Info, the value returned by evaluating
2724 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2725 evaluating @code{(buffer-file-name)} is @file{nil}.
2727 On the other hand, while I am writing this document, the value
2728 returned by evaluating @code{(buffer-name)} is
2729 @file{"introduction.texinfo"}, and the value returned by evaluating
2730 @code{(buffer-file-name)} is
2731 @file{"/gnu/work/intro/introduction.texinfo"}.
2733 @cindex @code{nil}, history of word
2734 The former is the name of the buffer and the latter is the name of the
2735 file. In Info, the buffer name is @file{"*info*"}. Info does not
2736 point to any file, so the result of evaluating
2737 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2738 from the Latin word for `nothing'; in this case, it means that the
2739 buffer is not associated with any file. (In Lisp, @code{nil} is also
2740 used to mean `false' and is a synonym for the empty list, @code{()}.)
2742 When I am writing, the name of my buffer is
2743 @file{"introduction.texinfo"}. The name of the file to which it
2744 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2746 (In the expressions, the parentheses tell the Lisp interpreter to
2747 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2748 functions; without the parentheses, the interpreter would attempt to
2749 evaluate the symbols as variables. @xref{Variables}.)
2751 In spite of the distinction between files and buffers, you will often
2752 find that people refer to a file when they mean a buffer and vice-verse.
2753 Indeed, most people say, ``I am editing a file,'' rather than saying,
2754 ``I am editing a buffer which I will soon save to a file.'' It is
2755 almost always clear from context what people mean. When dealing with
2756 computer programs, however, it is important to keep the distinction in mind,
2757 since the computer is not as smart as a person.
2759 @cindex Buffer, history of word
2760 The word `buffer', by the way, comes from the meaning of the word as a
2761 cushion that deadens the force of a collision. In early computers, a
2762 buffer cushioned the interaction between files and the computer's
2763 central processing unit. The drums or tapes that held a file and the
2764 central processing unit were pieces of equipment that were very
2765 different from each other, working at their own speeds, in spurts. The
2766 buffer made it possible for them to work together effectively.
2767 Eventually, the buffer grew from being an intermediary, a temporary
2768 holding place, to being the place where work is done. This
2769 transformation is rather like that of a small seaport that grew into a
2770 great city: once it was merely the place where cargo was warehoused
2771 temporarily before being loaded onto ships; then it became a business
2772 and cultural center in its own right.
2774 Not all buffers are associated with files. For example, a
2775 @file{*scratch*} buffer does not visit any file. Similarly, a
2776 @file{*Help*} buffer is not associated with any file.
2778 In the old days, when you lacked a @file{~/.emacs} file and started an
2779 Emacs session by typing the command @code{emacs} alone, without naming
2780 any files, Emacs started with the @file{*scratch*} buffer visible.
2781 Nowadays, you will see a splash screen. You can follow one of the
2782 commands suggested on the splash screen, visit a file, or press the
2783 spacebar to reach the @file{*scratch*} buffer.
2785 If you switch to the @file{*scratch*} buffer, type
2786 @code{(buffer-name)}, position the cursor after it, and then type
2787 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2788 will be returned and will appear in the echo area. @code{"*scratch*"}
2789 is the name of the buffer. When you type @code{(buffer-file-name)} in
2790 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2791 in the echo area, just as it does when you evaluate
2792 @code{(buffer-file-name)} in Info.
2794 Incidentally, if you are in the @file{*scratch*} buffer and want the
2795 value returned by an expression to appear in the @file{*scratch*}
2796 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2797 instead of @kbd{C-x C-e}. This causes the value returned to appear
2798 after the expression. The buffer will look like this:
2801 (buffer-name)"*scratch*"
2805 You cannot do this in Info since Info is read-only and it will not allow
2806 you to change the contents of the buffer. But you can do this in any
2807 buffer you can edit; and when you write code or documentation (such as
2808 this book), this feature is very useful.
2810 @node Getting Buffers
2811 @section Getting Buffers
2812 @findex current-buffer
2813 @findex other-buffer
2814 @cindex Getting a buffer
2816 The @code{buffer-name} function returns the @emph{name} of the buffer;
2817 to get the buffer @emph{itself}, a different function is needed: the
2818 @code{current-buffer} function. If you use this function in code, what
2819 you get is the buffer itself.
2821 A name and the object or entity to which the name refers are different
2822 from each other. You are not your name. You are a person to whom
2823 others refer by name. If you ask to speak to George and someone hands you
2824 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2825 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2826 not be satisfied. You do not want to speak to the name, but to the
2827 person to whom the name refers. A buffer is similar: the name of the
2828 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2829 get a buffer itself, you need to use a function such as
2830 @code{current-buffer}.
2832 However, there is a slight complication: if you evaluate
2833 @code{current-buffer} in an expression on its own, as we will do here,
2834 what you see is a printed representation of the name of the buffer
2835 without the contents of the buffer. Emacs works this way for two
2836 reasons: the buffer may be thousands of lines long---too long to be
2837 conveniently displayed; and, another buffer may have the same contents
2838 but a different name, and it is important to distinguish between them.
2841 Here is an expression containing the function:
2848 If you evaluate this expression in Info in Emacs in the usual way,
2849 @file{#<buffer *info*>} will appear in the echo area. The special
2850 format indicates that the buffer itself is being returned, rather than
2853 Incidentally, while you can type a number or symbol into a program, you
2854 cannot do that with the printed representation of a buffer: the only way
2855 to get a buffer itself is with a function such as @code{current-buffer}.
2857 A related function is @code{other-buffer}. This returns the most
2858 recently selected buffer other than the one you are in currently, not
2859 a printed representation of its name. If you have recently switched
2860 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2861 will return that buffer.
2864 You can see this by evaluating the expression:
2871 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2872 the name of whatever other buffer you switched back from most
2873 recently@footnote{Actually, by default, if the buffer from which you
2874 just switched is visible to you in another window, @code{other-buffer}
2875 will choose the most recent buffer that you cannot see; this is a
2876 subtlety that I often forget.}.
2878 @node Switching Buffers
2879 @section Switching Buffers
2880 @findex switch-to-buffer
2882 @cindex Switching to a buffer
2884 The @code{other-buffer} function actually provides a buffer when it is
2885 used as an argument to a function that requires one. We can see this
2886 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2889 But first, a brief introduction to the @code{switch-to-buffer}
2890 function. When you switched back and forth from Info to the
2891 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2892 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2893 rather, to save typing, you probably only typed @kbd{RET} if the
2894 default buffer was @file{*scratch*}, or if it was different, then you
2895 typed just part of the name, such as @code{*sc}, pressed your
2896 @kbd{TAB} key to cause it to expand to the full name, and then typed
2897 @kbd{RET}.} when prompted in the minibuffer for the name of
2898 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2899 b}, cause the Lisp interpreter to evaluate the interactive function
2900 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2901 different keystrokes call or run different functions. For example,
2902 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2903 @code{forward-sentence}, and so on.
2905 By writing @code{switch-to-buffer} in an expression, and giving it a
2906 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2910 (switch-to-buffer (other-buffer))
2914 The symbol @code{switch-to-buffer} is the first element of the list,
2915 so the Lisp interpreter will treat it as a function and carry out the
2916 instructions that are attached to it. But before doing that, the
2917 interpreter will note that @code{other-buffer} is inside parentheses
2918 and work on that symbol first. @code{other-buffer} is the first (and
2919 in this case, the only) element of this list, so the Lisp interpreter
2920 calls or runs the function. It returns another buffer. Next, the
2921 interpreter runs @code{switch-to-buffer}, passing to it, as an
2922 argument, the other buffer, which is what Emacs will switch to. If
2923 you are reading this in Info, try this now. Evaluate the expression.
2924 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2925 expression will move you to your most recent other buffer that you
2926 cannot see. If you really want to go to your most recently selected
2927 buffer, even if you can still see it, you need to evaluate the
2928 following more complex expression:
2931 (switch-to-buffer (other-buffer (current-buffer) t))
2935 In this case, the first argument to @code{other-buffer} tells it which
2936 buffer to skip---the current one---and the second argument tells
2937 @code{other-buffer} it is OK to switch to a visible buffer.
2938 In regular use, @code{switch-to-buffer} takes you to an invisible
2939 window since you would most likely use @kbd{C-x o} (@code{other-window})
2940 to go to another visible buffer.}
2942 In the programming examples in later sections of this document, you will
2943 see the function @code{set-buffer} more often than
2944 @code{switch-to-buffer}. This is because of a difference between
2945 computer programs and humans: humans have eyes and expect to see the
2946 buffer on which they are working on their computer terminals. This is
2947 so obvious, it almost goes without saying. However, programs do not
2948 have eyes. When a computer program works on a buffer, that buffer does
2949 not need to be visible on the screen.
2951 @code{switch-to-buffer} is designed for humans and does two different
2952 things: it switches the buffer to which Emacs's attention is directed; and
2953 it switches the buffer displayed in the window to the new buffer.
2954 @code{set-buffer}, on the other hand, does only one thing: it switches
2955 the attention of the computer program to a different buffer. The buffer
2956 on the screen remains unchanged (of course, normally nothing happens
2957 there until the command finishes running).
2959 @cindex @samp{call} defined
2960 Also, we have just introduced another jargon term, the word @dfn{call}.
2961 When you evaluate a list in which the first symbol is a function, you
2962 are calling that function. The use of the term comes from the notion of
2963 the function as an entity that can do something for you if you `call'
2964 it---just as a plumber is an entity who can fix a leak if you call him
2967 @node Buffer Size & Locations
2968 @section Buffer Size and the Location of Point
2969 @cindex Size of buffer
2971 @cindex Point location
2972 @cindex Location of point
2974 Finally, let's look at several rather simple functions,
2975 @code{buffer-size}, @code{point}, @code{point-min}, and
2976 @code{point-max}. These give information about the size of a buffer and
2977 the location of point within it.
2979 The function @code{buffer-size} tells you the size of the current
2980 buffer; that is, the function returns a count of the number of
2981 characters in the buffer.
2988 You can evaluate this in the usual way, by positioning the
2989 cursor after the expression and typing @kbd{C-x C-e}.
2991 @cindex @samp{point} defined
2992 In Emacs, the current position of the cursor is called @dfn{point}.
2993 The expression @code{(point)} returns a number that tells you where the
2994 cursor is located as a count of the number of characters from the
2995 beginning of the buffer up to point.
2998 You can see the character count for point in this buffer by evaluating
2999 the following expression in the usual way:
3006 As I write this, the value of @code{point} is 65724. The @code{point}
3007 function is frequently used in some of the examples later in this
3011 The value of point depends, of course, on its location within the
3012 buffer. If you evaluate point in this spot, the number will be larger:
3019 For me, the value of point in this location is 66043, which means that
3020 there are 319 characters (including spaces) between the two
3021 expressions. (Doubtless, you will see different numbers, since I will
3022 have edited this since I first evaluated point.)
3024 @cindex @samp{narrowing} defined
3025 The function @code{point-min} is somewhat similar to @code{point}, but
3026 it returns the value of the minimum permissible value of point in the
3027 current buffer. This is the number 1 unless @dfn{narrowing} is in
3028 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3029 or a program, to operations on just a part of a buffer.
3030 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3031 function @code{point-max} returns the value of the maximum permissible
3032 value of point in the current buffer.
3034 @node Evaluation Exercise
3037 Find a file with which you are working and move towards its middle.
3038 Find its buffer name, file name, length, and your position in the file.
3040 @node Writing Defuns
3041 @chapter How To Write Function Definitions
3042 @cindex Definition writing
3043 @cindex Function definition writing
3044 @cindex Writing a function definition
3046 When the Lisp interpreter evaluates a list, it looks to see whether the
3047 first symbol on the list has a function definition attached to it; or,
3048 put another way, whether the symbol points to a function definition. If
3049 it does, the computer carries out the instructions in the definition. A
3050 symbol that has a function definition is called, simply, a function
3051 (although, properly speaking, the definition is the function and the
3052 symbol refers to it.)
3055 * Primitive Functions::
3056 * defun:: The @code{defun} special form.
3057 * Install:: Install a function definition.
3058 * Interactive:: Making a function interactive.
3059 * Interactive Options:: Different options for @code{interactive}.
3060 * Permanent Installation:: Installing code permanently.
3061 * let:: Creating and initializing local variables.
3063 * else:: If--then--else expressions.
3064 * Truth & Falsehood:: What Lisp considers false and true.
3065 * save-excursion:: Keeping track of point, mark, and buffer.
3071 @node Primitive Functions
3072 @unnumberedsec An Aside about Primitive Functions
3074 @cindex Primitive functions
3075 @cindex Functions, primitive
3077 @cindex C language primitives
3078 @cindex Primitives written in C
3079 All functions are defined in terms of other functions, except for a few
3080 @dfn{primitive} functions that are written in the C programming
3081 language. When you write functions' definitions, you will write them in
3082 Emacs Lisp and use other functions as your building blocks. Some of the
3083 functions you will use will themselves be written in Emacs Lisp (perhaps
3084 by you) and some will be primitives written in C@. The primitive
3085 functions are used exactly like those written in Emacs Lisp and behave
3086 like them. They are written in C so we can easily run GNU Emacs on any
3087 computer that has sufficient power and can run C.
3089 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3090 distinguish between the use of functions written in C and the use of
3091 functions written in Emacs Lisp. The difference is irrelevant. I
3092 mention the distinction only because it is interesting to know. Indeed,
3093 unless you investigate, you won't know whether an already-written
3094 function is written in Emacs Lisp or C.
3097 @section The @code{defun} Special Form
3099 @cindex Special form of @code{defun}
3101 @cindex @samp{function definition} defined
3102 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3103 it that tells the computer what to do when the function is called.
3104 This code is called the @dfn{function definition} and is created by
3105 evaluating a Lisp expression that starts with the symbol @code{defun}
3106 (which is an abbreviation for @emph{define function}). Because
3107 @code{defun} does not evaluate its arguments in the usual way, it is
3108 called a @dfn{special form}.
3110 In subsequent sections, we will look at function definitions from the
3111 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3112 we will describe a simple function definition so you can see how it
3113 looks. This function definition uses arithmetic because it makes for a
3114 simple example. Some people dislike examples using arithmetic; however,
3115 if you are such a person, do not despair. Hardly any of the code we
3116 will study in the remainder of this introduction involves arithmetic or
3117 mathematics. The examples mostly involve text in one way or another.
3119 A function definition has up to five parts following the word
3124 The name of the symbol to which the function definition should be
3128 A list of the arguments that will be passed to the function. If no
3129 arguments will be passed to the function, this is an empty list,
3133 Documentation describing the function. (Technically optional, but
3134 strongly recommended.)
3137 Optionally, an expression to make the function interactive so you can
3138 use it by typing @kbd{M-x} and then the name of the function; or by
3139 typing an appropriate key or keychord.
3141 @cindex @samp{body} defined
3143 The code that instructs the computer what to do: the @dfn{body} of the
3144 function definition.
3147 It is helpful to think of the five parts of a function definition as
3148 being organized in a template, with slots for each part:
3152 (defun @var{function-name} (@var{arguments}@dots{})
3153 "@var{optional-documentation}@dots{}"
3154 (interactive @var{argument-passing-info}) ; @r{optional}
3159 As an example, here is the code for a function that multiplies its
3160 argument by 7. (This example is not interactive. @xref{Interactive,
3161 , Making a Function Interactive}, for that information.)
3165 (defun multiply-by-seven (number)
3166 "Multiply NUMBER by seven."
3171 This definition begins with a parenthesis and the symbol @code{defun},
3172 followed by the name of the function.
3174 @cindex @samp{argument list} defined
3175 The name of the function is followed by a list that contains the
3176 arguments that will be passed to the function. This list is called
3177 the @dfn{argument list}. In this example, the list has only one
3178 element, the symbol, @code{number}. When the function is used, the
3179 symbol will be bound to the value that is used as the argument to the
3182 Instead of choosing the word @code{number} for the name of the argument,
3183 I could have picked any other name. For example, I could have chosen
3184 the word @code{multiplicand}. I picked the word `number' because it
3185 tells what kind of value is intended for this slot; but I could just as
3186 well have chosen the word `multiplicand' to indicate the role that the
3187 value placed in this slot will play in the workings of the function. I
3188 could have called it @code{foogle}, but that would have been a bad
3189 choice because it would not tell humans what it means. The choice of
3190 name is up to the programmer and should be chosen to make the meaning of
3193 Indeed, you can choose any name you wish for a symbol in an argument
3194 list, even the name of a symbol used in some other function: the name
3195 you use in an argument list is private to that particular definition.
3196 In that definition, the name refers to a different entity than any use
3197 of the same name outside the function definition. Suppose you have a
3198 nick-name `Shorty' in your family; when your family members refer to
3199 `Shorty', they mean you. But outside your family, in a movie, for
3200 example, the name `Shorty' refers to someone else. Because a name in an
3201 argument list is private to the function definition, you can change the
3202 value of such a symbol inside the body of a function without changing
3203 its value outside the function. The effect is similar to that produced
3204 by a @code{let} expression. (@xref{let, , @code{let}}.)
3207 Note also that we discuss the word `number' in two different ways: as a
3208 symbol that appears in the code, and as the name of something that will
3209 be replaced by a something else during the evaluation of the function.
3210 In the first case, @code{number} is a symbol, not a number; it happens
3211 that within the function, it is a variable who value is the number in
3212 question, but our primary interest in it is as a symbol. On the other
3213 hand, when we are talking about the function, our interest is that we
3214 will substitute a number for the word @var{number}. To keep this
3215 distinction clear, we use different typography for the two
3216 circumstances. When we talk about this function, or about how it works,
3217 we refer to this number by writing @var{number}. In the function
3218 itself, we refer to it by writing @code{number}.
3221 The argument list is followed by the documentation string that
3222 describes the function. This is what you see when you type
3223 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3224 write a documentation string like this, you should make the first line
3225 a complete sentence since some commands, such as @code{apropos}, print
3226 only the first line of a multi-line documentation string. Also, you
3227 should not indent the second line of a documentation string, if you
3228 have one, because that looks odd when you use @kbd{C-h f}
3229 (@code{describe-function}). The documentation string is optional, but
3230 it is so useful, it should be included in almost every function you
3233 @findex * @r{(multiplication)}
3234 The third line of the example consists of the body of the function
3235 definition. (Most functions' definitions, of course, are longer than
3236 this.) In this function, the body is the list, @code{(* 7 number)}, which
3237 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3238 @code{*} is the function for multiplication, just as @code{+} is the
3239 function for addition.)
3241 When you use the @code{multiply-by-seven} function, the argument
3242 @code{number} evaluates to the actual number you want used. Here is an
3243 example that shows how @code{multiply-by-seven} is used; but don't try
3244 to evaluate this yet!
3247 (multiply-by-seven 3)
3251 The symbol @code{number}, specified in the function definition in the
3252 next section, is given or ``bound to'' the value 3 in the actual use of
3253 the function. Note that although @code{number} was inside parentheses
3254 in the function definition, the argument passed to the
3255 @code{multiply-by-seven} function is not in parentheses. The
3256 parentheses are written in the function definition so the computer can
3257 figure out where the argument list ends and the rest of the function
3260 If you evaluate this example, you are likely to get an error message.
3261 (Go ahead, try it!) This is because we have written the function
3262 definition, but not yet told the computer about the definition---we have
3263 not yet installed (or `loaded') the function definition in Emacs.
3264 Installing a function is the process that tells the Lisp interpreter the
3265 definition of the function. Installation is described in the next
3269 @section Install a Function Definition
3270 @cindex Install a Function Definition
3271 @cindex Definition installation
3272 @cindex Function definition installation
3274 If you are reading this inside of Info in Emacs, you can try out the
3275 @code{multiply-by-seven} function by first evaluating the function
3276 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3277 the function definition follows. Place the cursor after the last
3278 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3279 do this, @code{multiply-by-seven} will appear in the echo area. (What
3280 this means is that when a function definition is evaluated, the value it
3281 returns is the name of the defined function.) At the same time, this
3282 action installs the function definition.
3286 (defun multiply-by-seven (number)
3287 "Multiply NUMBER by seven."
3293 By evaluating this @code{defun}, you have just installed
3294 @code{multiply-by-seven} in Emacs. The function is now just as much a
3295 part of Emacs as @code{forward-word} or any other editing function you
3296 use. (@code{multiply-by-seven} will stay installed until you quit
3297 Emacs. To reload code automatically whenever you start Emacs, see
3298 @ref{Permanent Installation, , Installing Code Permanently}.)
3301 * Effect of installation::
3302 * Change a defun:: How to change a function definition.
3306 @node Effect of installation
3307 @unnumberedsubsec The effect of installation
3310 You can see the effect of installing @code{multiply-by-seven} by
3311 evaluating the following sample. Place the cursor after the following
3312 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3316 (multiply-by-seven 3)
3319 If you wish, you can read the documentation for the function by typing
3320 @kbd{C-h f} (@code{describe-function}) and then the name of the
3321 function, @code{multiply-by-seven}. When you do this, a
3322 @file{*Help*} window will appear on your screen that says:
3326 multiply-by-seven is a Lisp function.
3327 (multiply-by-seven NUMBER)
3329 Multiply NUMBER by seven.
3334 (To return to a single window on your screen, type @kbd{C-x 1}.)
3336 @node Change a defun
3337 @subsection Change a Function Definition
3338 @cindex Changing a function definition
3339 @cindex Function definition, how to change
3340 @cindex Definition, how to change
3342 If you want to change the code in @code{multiply-by-seven}, just rewrite
3343 it. To install the new version in place of the old one, evaluate the
3344 function definition again. This is how you modify code in Emacs. It is
3347 As an example, you can change the @code{multiply-by-seven} function to
3348 add the number to itself seven times instead of multiplying the number
3349 by seven. It produces the same answer, but by a different path. At
3350 the same time, we will add a comment to the code; a comment is text
3351 that the Lisp interpreter ignores, but that a human reader may find
3352 useful or enlightening. The comment is that this is the ``second
3357 (defun multiply-by-seven (number) ; @r{Second version.}
3358 "Multiply NUMBER by seven."
3359 (+ number number number number number number number))
3363 @cindex Comments in Lisp code
3364 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3365 line that follows a semicolon is a comment. The end of the line is the
3366 end of the comment. To stretch a comment over two or more lines, begin
3367 each line with a semicolon.
3369 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3370 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3371 Reference Manual}, for more about comments.
3373 You can install this version of the @code{multiply-by-seven} function by
3374 evaluating it in the same way you evaluated the first function: place
3375 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3377 In summary, this is how you write code in Emacs Lisp: you write a
3378 function; install it; test it; and then make fixes or enhancements and
3382 @section Make a Function Interactive
3383 @cindex Interactive functions
3386 You make a function interactive by placing a list that begins with
3387 the special form @code{interactive} immediately after the
3388 documentation. A user can invoke an interactive function by typing
3389 @kbd{M-x} and then the name of the function; or by typing the keys to
3390 which it is bound, for example, by typing @kbd{C-n} for
3391 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3393 Interestingly, when you call an interactive function interactively,
3394 the value returned is not automatically displayed in the echo area.
3395 This is because you often call an interactive function for its side
3396 effects, such as moving forward by a word or line, and not for the
3397 value returned. If the returned value were displayed in the echo area
3398 each time you typed a key, it would be very distracting.
3401 * Interactive multiply-by-seven:: An overview.
3402 * multiply-by-seven in detail:: The interactive version.
3406 @node Interactive multiply-by-seven
3407 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3410 Both the use of the special form @code{interactive} and one way to
3411 display a value in the echo area can be illustrated by creating an
3412 interactive version of @code{multiply-by-seven}.
3419 (defun multiply-by-seven (number) ; @r{Interactive version.}
3420 "Multiply NUMBER by seven."
3422 (message "The result is %d" (* 7 number)))
3427 You can install this code by placing your cursor after it and typing
3428 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3429 Then, you can use this code by typing @kbd{C-u} and a number and then
3430 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3431 @samp{The result is @dots{}} followed by the product will appear in the
3434 Speaking more generally, you invoke a function like this in either of two
3439 By typing a prefix argument that contains the number to be passed, and
3440 then typing @kbd{M-x} and the name of the function, as with
3441 @kbd{C-u 3 M-x forward-sentence}; or,
3444 By typing whatever key or keychord the function is bound to, as with
3449 Both the examples just mentioned work identically to move point forward
3450 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3451 it could not be used as an example of key binding.)
3453 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3456 A prefix argument is passed to an interactive function by typing the
3457 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3458 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3459 type @kbd{C-u} without a number, it defaults to 4).
3461 @node multiply-by-seven in detail
3462 @subsection An Interactive @code{multiply-by-seven}
3464 Let's look at the use of the special form @code{interactive} and then at
3465 the function @code{message} in the interactive version of
3466 @code{multiply-by-seven}. You will recall that the function definition
3471 (defun multiply-by-seven (number) ; @r{Interactive version.}
3472 "Multiply NUMBER by seven."
3474 (message "The result is %d" (* 7 number)))
3478 In this function, the expression, @code{(interactive "p")}, is a list of
3479 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3480 the function and use its value for the argument of the function.
3483 The argument will be a number. This means that the symbol
3484 @code{number} will be bound to a number in the line:
3487 (message "The result is %d" (* 7 number))
3492 For example, if your prefix argument is 5, the Lisp interpreter will
3493 evaluate the line as if it were:
3496 (message "The result is %d" (* 7 5))
3500 (If you are reading this in GNU Emacs, you can evaluate this expression
3501 yourself.) First, the interpreter will evaluate the inner list, which
3502 is @code{(* 7 5)}. This returns a value of 35. Next, it
3503 will evaluate the outer list, passing the values of the second and
3504 subsequent elements of the list to the function @code{message}.
3506 As we have seen, @code{message} is an Emacs Lisp function especially
3507 designed for sending a one line message to a user. (@xref{message, ,
3508 The @code{message} function}.) In summary, the @code{message}
3509 function prints its first argument in the echo area as is, except for
3510 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3511 which we have not mentioned). When it sees a control sequence, the
3512 function looks to the second or subsequent arguments and prints the
3513 value of the argument in the location in the string where the control
3514 sequence is located.
3516 In the interactive @code{multiply-by-seven} function, the control string
3517 is @samp{%d}, which requires a number, and the value returned by
3518 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3519 is printed in place of the @samp{%d} and the message is @samp{The result
3522 (Note that when you call the function @code{multiply-by-seven}, the
3523 message is printed without quotes, but when you call @code{message}, the
3524 text is printed in double quotes. This is because the value returned by
3525 @code{message} is what appears in the echo area when you evaluate an
3526 expression whose first element is @code{message}; but when embedded in a
3527 function, @code{message} prints the text as a side effect without
3530 @node Interactive Options
3531 @section Different Options for @code{interactive}
3532 @cindex Options for @code{interactive}
3533 @cindex Interactive options
3535 In the example, @code{multiply-by-seven} used @code{"p"} as the
3536 argument to @code{interactive}. This argument told Emacs to interpret
3537 your typing either @kbd{C-u} followed by a number or @key{META}
3538 followed by a number as a command to pass that number to the function
3539 as its argument. Emacs has more than twenty characters predefined for
3540 use with @code{interactive}. In almost every case, one of these
3541 options will enable you to pass the right information interactively to
3542 a function. (@xref{Interactive Codes, , Code Characters for
3543 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3546 Consider the function @code{zap-to-char}. Its interactive expression
3550 (interactive "p\ncZap to char: ")
3553 The first part of the argument to @code{interactive} is @samp{p}, with
3554 which you are already familiar. This argument tells Emacs to
3555 interpret a `prefix', as a number to be passed to the function. You
3556 can specify a prefix either by typing @kbd{C-u} followed by a number
3557 or by typing @key{META} followed by a number. The prefix is the
3558 number of specified characters. Thus, if your prefix is three and the
3559 specified character is @samp{x}, then you will delete all the text up
3560 to and including the third next @samp{x}. If you do not set a prefix,
3561 then you delete all the text up to and including the specified
3562 character, but no more.
3564 The @samp{c} tells the function the name of the character to which to delete.
3566 More formally, a function with two or more arguments can have
3567 information passed to each argument by adding parts to the string that
3568 follows @code{interactive}. When you do this, the information is
3569 passed to each argument in the same order it is specified in the
3570 @code{interactive} list. In the string, each part is separated from
3571 the next part by a @samp{\n}, which is a newline. For example, you
3572 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3573 This causes Emacs to pass the value of the prefix argument (if there
3574 is one) and the character.
3576 In this case, the function definition looks like the following, where
3577 @code{arg} and @code{char} are the symbols to which @code{interactive}
3578 binds the prefix argument and the specified character:
3582 (defun @var{name-of-function} (arg char)
3583 "@var{documentation}@dots{}"
3584 (interactive "p\ncZap to char: ")
3585 @var{body-of-function}@dots{})
3590 (The space after the colon in the prompt makes it look better when you
3591 are prompted. @xref{copy-to-buffer, , The Definition of
3592 @code{copy-to-buffer}}, for an example.)
3594 When a function does not take arguments, @code{interactive} does not
3595 require any. Such a function contains the simple expression
3596 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3599 Alternatively, if the special letter-codes are not right for your
3600 application, you can pass your own arguments to @code{interactive} as
3603 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3604 for an example. @xref{Using Interactive, , Using @code{Interactive},
3605 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3606 explanation about this technique.
3608 @node Permanent Installation
3609 @section Install Code Permanently
3610 @cindex Install code permanently
3611 @cindex Permanent code installation
3612 @cindex Code installation
3614 When you install a function definition by evaluating it, it will stay
3615 installed until you quit Emacs. The next time you start a new session
3616 of Emacs, the function will not be installed unless you evaluate the
3617 function definition again.
3619 At some point, you may want to have code installed automatically
3620 whenever you start a new session of Emacs. There are several ways of
3625 If you have code that is just for yourself, you can put the code for the
3626 function definition in your @file{.emacs} initialization file. When you
3627 start Emacs, your @file{.emacs} file is automatically evaluated and all
3628 the function definitions within it are installed.
3629 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3632 Alternatively, you can put the function definitions that you want
3633 installed in one or more files of their own and use the @code{load}
3634 function to cause Emacs to evaluate and thereby install each of the
3635 functions in the files.
3636 @xref{Loading Files, , Loading Files}.
3639 Thirdly, if you have code that your whole site will use, it is usual
3640 to put it in a file called @file{site-init.el} that is loaded when
3641 Emacs is built. This makes the code available to everyone who uses
3642 your machine. (See the @file{INSTALL} file that is part of the Emacs
3646 Finally, if you have code that everyone who uses Emacs may want, you
3647 can post it on a computer network or send a copy to the Free Software
3648 Foundation. (When you do this, please license the code and its
3649 documentation under a license that permits other people to run, copy,
3650 study, modify, and redistribute the code and which protects you from
3651 having your work taken from you.) If you send a copy of your code to
3652 the Free Software Foundation, and properly protect yourself and
3653 others, it may be included in the next release of Emacs. In large
3654 part, this is how Emacs has grown over the past years, by donations.
3660 The @code{let} expression is a special form in Lisp that you will need
3661 to use in most function definitions.
3663 @code{let} is used to attach or bind a symbol to a value in such a way
3664 that the Lisp interpreter will not confuse the variable with a
3665 variable of the same name that is not part of the function.
3667 To understand why the @code{let} special form is necessary, consider
3668 the situation in which you own a home that you generally refer to as
3669 `the house', as in the sentence, ``The house needs painting.'' If you
3670 are visiting a friend and your host refers to `the house', he is
3671 likely to be referring to @emph{his} house, not yours, that is, to a
3674 If your friend is referring to his house and you think he is referring
3675 to your house, you may be in for some confusion. The same thing could
3676 happen in Lisp if a variable that is used inside of one function has
3677 the same name as a variable that is used inside of another function,
3678 and the two are not intended to refer to the same value. The
3679 @code{let} special form prevents this kind of confusion.
3682 * Prevent confusion::
3683 * Parts of let Expression::
3684 * Sample let Expression::
3685 * Uninitialized let Variables::
3689 @node Prevent confusion
3690 @unnumberedsubsec @code{let} Prevents Confusion
3693 @cindex @samp{local variable} defined
3694 @cindex @samp{variable, local}, defined
3695 The @code{let} special form prevents confusion. @code{let} creates a
3696 name for a @dfn{local variable} that overshadows any use of the same
3697 name outside the @code{let} expression. This is like understanding
3698 that whenever your host refers to `the house', he means his house, not
3699 yours. (Symbols used in argument lists work the same way.
3700 @xref{defun, , The @code{defun} Special Form}.)
3702 Local variables created by a @code{let} expression retain their value
3703 @emph{only} within the @code{let} expression itself (and within
3704 expressions called within the @code{let} expression); the local
3705 variables have no effect outside the @code{let} expression.
3707 Another way to think about @code{let} is that it is like a @code{setq}
3708 that is temporary and local. The values set by @code{let} are
3709 automatically undone when the @code{let} is finished. The setting
3710 only affects expressions that are inside the bounds of the @code{let}
3711 expression. In computer science jargon, we would say ``the binding of
3712 a symbol is visible only in functions called in the @code{let} form;
3713 in Emacs Lisp, scoping is dynamic, not lexical.''
3715 @code{let} can create more than one variable at once. Also,
3716 @code{let} gives each variable it creates an initial value, either a
3717 value specified by you, or @code{nil}. (In the jargon, this is called
3718 `binding the variable to the value'.) After @code{let} has created
3719 and bound the variables, it executes the code in the body of the
3720 @code{let}, and returns the value of the last expression in the body,
3721 as the value of the whole @code{let} expression. (`Execute' is a jargon
3722 term that means to evaluate a list; it comes from the use of the word
3723 meaning `to give practical effect to' (@cite{Oxford English
3724 Dictionary}). Since you evaluate an expression to perform an action,
3725 `execute' has evolved as a synonym to `evaluate'.)
3727 @node Parts of let Expression
3728 @subsection The Parts of a @code{let} Expression
3729 @cindex @code{let} expression, parts of
3730 @cindex Parts of @code{let} expression
3732 @cindex @samp{varlist} defined
3733 A @code{let} expression is a list of three parts. The first part is
3734 the symbol @code{let}. The second part is a list, called a
3735 @dfn{varlist}, each element of which is either a symbol by itself or a
3736 two-element list, the first element of which is a symbol. The third
3737 part of the @code{let} expression is the body of the @code{let}. The
3738 body usually consists of one or more lists.
3741 A template for a @code{let} expression looks like this:
3744 (let @var{varlist} @var{body}@dots{})
3748 The symbols in the varlist are the variables that are given initial
3749 values by the @code{let} special form. Symbols by themselves are given
3750 the initial value of @code{nil}; and each symbol that is the first
3751 element of a two-element list is bound to the value that is returned
3752 when the Lisp interpreter evaluates the second element.
3754 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3755 this case, in a @code{let} expression, Emacs binds the symbol
3756 @code{thread} to an initial value of @code{nil}, and binds the symbol
3757 @code{needles} to an initial value of 3.
3759 When you write a @code{let} expression, what you do is put the
3760 appropriate expressions in the slots of the @code{let} expression
3763 If the varlist is composed of two-element lists, as is often the case,
3764 the template for the @code{let} expression looks like this:
3768 (let ((@var{variable} @var{value})
3769 (@var{variable} @var{value})
3775 @node Sample let Expression
3776 @subsection Sample @code{let} Expression
3777 @cindex Sample @code{let} expression
3778 @cindex @code{let} expression sample
3780 The following expression creates and gives initial values
3781 to the two variables @code{zebra} and @code{tiger}. The body of the
3782 @code{let} expression is a list which calls the @code{message} function.
3786 (let ((zebra 'stripes)
3788 (message "One kind of animal has %s and another is %s."
3793 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3795 The two variables are @code{zebra} and @code{tiger}. Each variable is
3796 the first element of a two-element list and each value is the second
3797 element of its two-element list. In the varlist, Emacs binds the
3798 variable @code{zebra} to the value @code{stripes}@footnote{According
3799 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3800 become impossibly dangerous as they grow older'' but the claim here is
3801 that they do not become fierce like a tiger. (1997, W. W. Norton and
3802 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3803 variable @code{tiger} to the value @code{fierce}. In this example,
3804 both values are symbols preceded by a quote. The values could just as
3805 well have been another list or a string. The body of the @code{let}
3806 follows after the list holding the variables. In this example, the
3807 body is a list that uses the @code{message} function to print a string
3811 You may evaluate the example in the usual fashion, by placing the
3812 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3813 this, the following will appear in the echo area:
3816 "One kind of animal has stripes and another is fierce."
3819 As we have seen before, the @code{message} function prints its first
3820 argument, except for @samp{%s}. In this example, the value of the variable
3821 @code{zebra} is printed at the location of the first @samp{%s} and the
3822 value of the variable @code{tiger} is printed at the location of the
3825 @node Uninitialized let Variables
3826 @subsection Uninitialized Variables in a @code{let} Statement
3827 @cindex Uninitialized @code{let} variables
3828 @cindex @code{let} variables uninitialized
3830 If you do not bind the variables in a @code{let} statement to specific
3831 initial values, they will automatically be bound to an initial value of
3832 @code{nil}, as in the following expression:
3841 "Here are %d variables with %s, %s, and %s value."
3842 birch pine fir oak))
3847 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3850 If you evaluate this expression in the usual way, the following will
3851 appear in your echo area:
3854 "Here are 3 variables with nil, nil, and some value."
3858 In this example, Emacs binds the symbol @code{birch} to the number 3,
3859 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3860 the symbol @code{oak} to the value @code{some}.
3862 Note that in the first part of the @code{let}, the variables @code{pine}
3863 and @code{fir} stand alone as atoms that are not surrounded by
3864 parentheses; this is because they are being bound to @code{nil}, the
3865 empty list. But @code{oak} is bound to @code{some} and so is a part of
3866 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3867 number 3 and so is in a list with that number. (Since a number
3868 evaluates to itself, the number does not need to be quoted. Also, the
3869 number is printed in the message using a @samp{%d} rather than a
3870 @samp{%s}.) The four variables as a group are put into a list to
3871 delimit them from the body of the @code{let}.
3874 @section The @code{if} Special Form
3876 @cindex Conditional with @code{if}
3878 A third special form, in addition to @code{defun} and @code{let}, is the
3879 conditional @code{if}. This form is used to instruct the computer to
3880 make decisions. You can write function definitions without using
3881 @code{if}, but it is used often enough, and is important enough, to be
3882 included here. It is used, for example, in the code for the
3883 function @code{beginning-of-buffer}.
3885 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3886 @emph{then} an expression is evaluated.'' If the test is not true, the
3887 expression is not evaluated. For example, you might make a decision
3888 such as, ``if it is warm and sunny, then go to the beach!''
3891 * if in more detail::
3892 * type-of-animal in detail:: An example of an @code{if} expression.
3896 @node if in more detail
3897 @unnumberedsubsec @code{if} in more detail
3900 @cindex @samp{if-part} defined
3901 @cindex @samp{then-part} defined
3902 An @code{if} expression written in Lisp does not use the word `then';
3903 the test and the action are the second and third elements of the list
3904 whose first element is @code{if}. Nonetheless, the test part of an
3905 @code{if} expression is often called the @dfn{if-part} and the second
3906 argument is often called the @dfn{then-part}.
3908 Also, when an @code{if} expression is written, the true-or-false-test
3909 is usually written on the same line as the symbol @code{if}, but the
3910 action to carry out if the test is true, the ``then-part'', is written
3911 on the second and subsequent lines. This makes the @code{if}
3912 expression easier to read.
3916 (if @var{true-or-false-test}
3917 @var{action-to-carry-out-if-test-is-true})
3922 The true-or-false-test will be an expression that
3923 is evaluated by the Lisp interpreter.
3925 Here is an example that you can evaluate in the usual manner. The test
3926 is whether the number 5 is greater than the number 4. Since it is, the
3927 message @samp{5 is greater than 4!} will be printed.
3931 (if (> 5 4) ; @r{if-part}
3932 (message "5 is greater than 4!")) ; @r{then-part}
3937 (The function @code{>} tests whether its first argument is greater than
3938 its second argument and returns true if it is.)
3939 @findex > (greater than)
3941 Of course, in actual use, the test in an @code{if} expression will not
3942 be fixed for all time as it is by the expression @code{(> 5 4)}.
3943 Instead, at least one of the variables used in the test will be bound to
3944 a value that is not known ahead of time. (If the value were known ahead
3945 of time, we would not need to run the test!)
3947 For example, the value may be bound to an argument of a function
3948 definition. In the following function definition, the character of the
3949 animal is a value that is passed to the function. If the value bound to
3950 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3951 tiger!} will be printed; otherwise, @code{nil} will be returned.
3955 (defun type-of-animal (characteristic)
3956 "Print message in echo area depending on CHARACTERISTIC.
3957 If the CHARACTERISTIC is the symbol `fierce',
3958 then warn of a tiger."
3959 (if (equal characteristic 'fierce)
3960 (message "It's a tiger!")))
3966 If you are reading this inside of GNU Emacs, you can evaluate the
3967 function definition in the usual way to install it in Emacs, and then you
3968 can evaluate the following two expressions to see the results:
3972 (type-of-animal 'fierce)
3974 (type-of-animal 'zebra)
3979 @c Following sentences rewritten to prevent overfull hbox.
3981 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
3982 following message printed in the echo area: @code{"It's a tiger!"}; and
3983 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
3984 printed in the echo area.
3986 @node type-of-animal in detail
3987 @subsection The @code{type-of-animal} Function in Detail
3989 Let's look at the @code{type-of-animal} function in detail.
3991 The function definition for @code{type-of-animal} was written by filling
3992 the slots of two templates, one for a function definition as a whole, and
3993 a second for an @code{if} expression.
3996 The template for every function that is not interactive is:
4000 (defun @var{name-of-function} (@var{argument-list})
4001 "@var{documentation}@dots{}"
4007 The parts of the function that match this template look like this:
4011 (defun type-of-animal (characteristic)
4012 "Print message in echo area depending on CHARACTERISTIC.
4013 If the CHARACTERISTIC is the symbol `fierce',
4014 then warn of a tiger."
4015 @var{body: the} @code{if} @var{expression})
4019 The name of function is @code{type-of-animal}; it is passed the value
4020 of one argument. The argument list is followed by a multi-line
4021 documentation string. The documentation string is included in the
4022 example because it is a good habit to write documentation string for
4023 every function definition. The body of the function definition
4024 consists of the @code{if} expression.
4027 The template for an @code{if} expression looks like this:
4031 (if @var{true-or-false-test}
4032 @var{action-to-carry-out-if-the-test-returns-true})
4037 In the @code{type-of-animal} function, the code for the @code{if}
4042 (if (equal characteristic 'fierce)
4043 (message "It's a tiger!")))
4048 Here, the true-or-false-test is the expression:
4051 (equal characteristic 'fierce)
4055 In Lisp, @code{equal} is a function that determines whether its first
4056 argument is equal to its second argument. The second argument is the
4057 quoted symbol @code{'fierce} and the first argument is the value of the
4058 symbol @code{characteristic}---in other words, the argument passed to
4061 In the first exercise of @code{type-of-animal}, the argument
4062 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4063 is equal to @code{fierce}, the expression, @code{(equal characteristic
4064 'fierce)}, returns a value of true. When this happens, the @code{if}
4065 evaluates the second argument or then-part of the @code{if}:
4066 @code{(message "It's tiger!")}.
4068 On the other hand, in the second exercise of @code{type-of-animal}, the
4069 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4070 is not equal to @code{fierce}, so the then-part is not evaluated and
4071 @code{nil} is returned by the @code{if} expression.
4074 @section If--then--else Expressions
4077 An @code{if} expression may have an optional third argument, called
4078 the @dfn{else-part}, for the case when the true-or-false-test returns
4079 false. When this happens, the second argument or then-part of the
4080 overall @code{if} expression is @emph{not} evaluated, but the third or
4081 else-part @emph{is} evaluated. You might think of this as the cloudy
4082 day alternative for the decision ``if it is warm and sunny, then go to
4083 the beach, else read a book!''.
4085 The word ``else'' is not written in the Lisp code; the else-part of an
4086 @code{if} expression comes after the then-part. In the written Lisp, the
4087 else-part is usually written to start on a line of its own and is
4088 indented less than the then-part:
4092 (if @var{true-or-false-test}
4093 @var{action-to-carry-out-if-the-test-returns-true}
4094 @var{action-to-carry-out-if-the-test-returns-false})
4098 For example, the following @code{if} expression prints the message @samp{4
4099 is not greater than 5!} when you evaluate it in the usual way:
4103 (if (> 4 5) ; @r{if-part}
4104 (message "4 falsely greater than 5!") ; @r{then-part}
4105 (message "4 is not greater than 5!")) ; @r{else-part}
4110 Note that the different levels of indentation make it easy to
4111 distinguish the then-part from the else-part. (GNU Emacs has several
4112 commands that automatically indent @code{if} expressions correctly.
4113 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4115 We can extend the @code{type-of-animal} function to include an
4116 else-part by simply incorporating an additional part to the @code{if}
4120 You can see the consequences of doing this if you evaluate the following
4121 version of the @code{type-of-animal} function definition to install it
4122 and then evaluate the two subsequent expressions to pass different
4123 arguments to the function.
4127 (defun type-of-animal (characteristic) ; @r{Second version.}
4128 "Print message in echo area depending on CHARACTERISTIC.
4129 If the CHARACTERISTIC is the symbol `fierce',
4130 then warn of a tiger;
4131 else say it's not fierce."
4132 (if (equal characteristic 'fierce)
4133 (message "It's a tiger!")
4134 (message "It's not fierce!")))
4141 (type-of-animal 'fierce)
4143 (type-of-animal 'zebra)
4148 @c Following sentence rewritten to prevent overfull hbox.
4150 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4151 following message printed in the echo area: @code{"It's a tiger!"}; but
4152 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4153 @code{"It's not fierce!"}.
4155 (Of course, if the @var{characteristic} were @code{ferocious}, the
4156 message @code{"It's not fierce!"} would be printed; and it would be
4157 misleading! When you write code, you need to take into account the
4158 possibility that some such argument will be tested by the @code{if}
4159 and write your program accordingly.)
4161 @node Truth & Falsehood
4162 @section Truth and Falsehood in Emacs Lisp
4163 @cindex Truth and falsehood in Emacs Lisp
4164 @cindex Falsehood and truth in Emacs Lisp
4167 There is an important aspect to the truth test in an @code{if}
4168 expression. So far, we have spoken of `true' and `false' as values of
4169 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4170 `false' is just our old friend @code{nil}. Anything else---anything
4173 The expression that tests for truth is interpreted as @dfn{true}
4174 if the result of evaluating it is a value that is not @code{nil}. In
4175 other words, the result of the test is considered true if the value
4176 returned is a number such as 47, a string such as @code{"hello"}, or a
4177 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4178 long as it is not empty), or even a buffer!
4181 * nil explained:: @code{nil} has two meanings.
4186 @unnumberedsubsec An explanation of @code{nil}
4189 Before illustrating a test for truth, we need an explanation of @code{nil}.
4191 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4192 empty list. Second, it means false and is the value returned when a
4193 true-or-false-test tests false. @code{nil} can be written as an empty
4194 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4195 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4196 to use @code{nil} for false and @code{()} for the empty list.
4198 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4199 list---is considered true. This means that if an evaluation returns
4200 something that is not an empty list, an @code{if} expression will test
4201 true. For example, if a number is put in the slot for the test, it
4202 will be evaluated and will return itself, since that is what numbers
4203 do when evaluated. In this conditional, the @code{if} expression will
4204 test true. The expression tests false only when @code{nil}, an empty
4205 list, is returned by evaluating the expression.
4207 You can see this by evaluating the two expressions in the following examples.
4209 In the first example, the number 4 is evaluated as the test in the
4210 @code{if} expression and returns itself; consequently, the then-part
4211 of the expression is evaluated and returned: @samp{true} appears in
4212 the echo area. In the second example, the @code{nil} indicates false;
4213 consequently, the else-part of the expression is evaluated and
4214 returned: @samp{false} appears in the echo area.
4231 Incidentally, if some other useful value is not available for a test that
4232 returns true, then the Lisp interpreter will return the symbol @code{t}
4233 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4234 when evaluated, as you can see by evaluating it in the usual way:
4242 On the other hand, this function returns @code{nil} if the test is false.
4248 @node save-excursion
4249 @section @code{save-excursion}
4250 @findex save-excursion
4251 @cindex Region, what it is
4252 @cindex Preserving point, mark, and buffer
4253 @cindex Point, mark, buffer preservation
4257 The @code{save-excursion} function is the fourth and final special form
4258 that we will discuss in this chapter.
4260 In Emacs Lisp programs used for editing, the @code{save-excursion}
4261 function is very common. It saves the location of point and mark,
4262 executes the body of the function, and then restores point and mark to
4263 their previous positions if their locations were changed. Its primary
4264 purpose is to keep the user from being surprised and disturbed by
4265 unexpected movement of point or mark.
4268 * Point and mark:: A review of various locations.
4269 * Template for save-excursion::
4273 @node Point and mark
4274 @unnumberedsubsec Point and Mark
4277 Before discussing @code{save-excursion}, however, it may be useful
4278 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4279 the current location of the cursor. Wherever the cursor
4280 is, that is point. More precisely, on terminals where the cursor
4281 appears to be on top of a character, point is immediately before the
4282 character. In Emacs Lisp, point is an integer. The first character in
4283 a buffer is number one, the second is number two, and so on. The
4284 function @code{point} returns the current position of the cursor as a
4285 number. Each buffer has its own value for point.
4287 The @dfn{mark} is another position in the buffer; its value can be set
4288 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4289 a mark has been set, you can use the command @kbd{C-x C-x}
4290 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4291 and set the mark to be the previous position of point. In addition, if
4292 you set another mark, the position of the previous mark is saved in the
4293 mark ring. Many mark positions can be saved this way. You can jump the
4294 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4297 The part of the buffer between point and mark is called @dfn{the
4298 region}. Numerous commands work on the region, including
4299 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4300 @code{print-region}.
4302 The @code{save-excursion} special form saves the locations of point and
4303 mark and restores those positions after the code within the body of the
4304 special form is evaluated by the Lisp interpreter. Thus, if point were
4305 in the beginning of a piece of text and some code moved point to the end
4306 of the buffer, the @code{save-excursion} would put point back to where
4307 it was before, after the expressions in the body of the function were
4310 In Emacs, a function frequently moves point as part of its internal
4311 workings even though a user would not expect this. For example,
4312 @code{count-lines-region} moves point. To prevent the user from being
4313 bothered by jumps that are both unexpected and (from the user's point of
4314 view) unnecessary, @code{save-excursion} is often used to keep point and
4315 mark in the location expected by the user. The use of
4316 @code{save-excursion} is good housekeeping.
4318 To make sure the house stays clean, @code{save-excursion} restores the
4319 values of point and mark even if something goes wrong in the code inside
4320 of it (or, to be more precise and to use the proper jargon, ``in case of
4321 abnormal exit''). This feature is very helpful.
4323 In addition to recording the values of point and mark,
4324 @code{save-excursion} keeps track of the current buffer, and restores
4325 it, too. This means you can write code that will change the buffer and
4326 have @code{save-excursion} switch you back to the original buffer.
4327 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4328 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4330 @node Template for save-excursion
4331 @subsection Template for a @code{save-excursion} Expression
4334 The template for code using @code{save-excursion} is simple:
4344 The body of the function is one or more expressions that will be
4345 evaluated in sequence by the Lisp interpreter. If there is more than
4346 one expression in the body, the value of the last one will be returned
4347 as the value of the @code{save-excursion} function. The other
4348 expressions in the body are evaluated only for their side effects; and
4349 @code{save-excursion} itself is used only for its side effect (which
4350 is restoring the positions of point and mark).
4353 In more detail, the template for a @code{save-excursion} expression
4359 @var{first-expression-in-body}
4360 @var{second-expression-in-body}
4361 @var{third-expression-in-body}
4363 @var{last-expression-in-body})
4368 An expression, of course, may be a symbol on its own or a list.
4370 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4371 within the body of a @code{let} expression. It looks like this:
4384 In the last few chapters we have introduced a fair number of functions
4385 and special forms. Here they are described in brief, along with a few
4386 similar functions that have not been mentioned yet.
4389 @item eval-last-sexp
4390 Evaluate the last symbolic expression before the current location of
4391 point. The value is printed in the echo area unless the function is
4392 invoked with an argument; in that case, the output is printed in the
4393 current buffer. This command is normally bound to @kbd{C-x C-e}.
4396 Define function. This special form has up to five parts: the name,
4397 a template for the arguments that will be passed to the function,
4398 documentation, an optional interactive declaration, and the body of the
4402 For example, in an early version of Emacs, the function definition was
4403 as follows. (It is slightly more complex now that it seeks the first
4404 non-whitespace character rather than the first visible character.)
4408 (defun back-to-indentation ()
4409 "Move point to first visible character on line."
4411 (beginning-of-line 1)
4412 (skip-chars-forward " \t"))
4419 (defun backward-to-indentation (&optional arg)
4420 "Move backward ARG lines and position at first nonblank character."
4422 (forward-line (- (or arg 1)))
4423 (skip-chars-forward " \t"))
4425 (defun back-to-indentation ()
4426 "Move point to the first non-whitespace character on this line."
4428 (beginning-of-line 1)
4429 (skip-syntax-forward " " (line-end-position))
4430 ;; Move back over chars that have whitespace syntax but have the p flag.
4431 (backward-prefix-chars))
4435 Declare to the interpreter that the function can be used
4436 interactively. This special form may be followed by a string with one
4437 or more parts that pass the information to the arguments of the
4438 function, in sequence. These parts may also tell the interpreter to
4439 prompt for information. Parts of the string are separated by
4440 newlines, @samp{\n}.
4443 Common code characters are:
4447 The name of an existing buffer.
4450 The name of an existing file.
4453 The numeric prefix argument. (Note that this `p' is lower case.)
4456 Point and the mark, as two numeric arguments, smallest first. This
4457 is the only code letter that specifies two successive arguments
4461 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4462 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4466 Declare that a list of variables is for use within the body of the
4467 @code{let} and give them an initial value, either @code{nil} or a
4468 specified value; then evaluate the rest of the expressions in the body
4469 of the @code{let} and return the value of the last one. Inside the
4470 body of the @code{let}, the Lisp interpreter does not see the values of
4471 the variables of the same names that are bound outside of the
4479 (let ((foo (buffer-name))
4480 (bar (buffer-size)))
4482 "This buffer is %s and has %d characters."
4487 @item save-excursion
4488 Record the values of point and mark and the current buffer before
4489 evaluating the body of this special form. Restore the values of point
4490 and mark and buffer afterward.
4497 (message "We are %d characters into this buffer."
4500 (goto-char (point-min)) (point))))
4505 Evaluate the first argument to the function; if it is true, evaluate
4506 the second argument; else evaluate the third argument, if there is one.
4508 The @code{if} special form is called a @dfn{conditional}. There are
4509 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4517 (if (= 22 emacs-major-version)
4518 (message "This is version 22 Emacs")
4519 (message "This is not version 22 Emacs"))
4528 The @code{<} function tests whether its first argument is smaller than
4529 its second argument. A corresponding function, @code{>}, tests whether
4530 the first argument is greater than the second. Likewise, @code{<=}
4531 tests whether the first argument is less than or equal to the second and
4532 @code{>=} tests whether the first argument is greater than or equal to
4533 the second. In all cases, both arguments must be numbers or markers
4534 (markers indicate positions in buffers).
4538 The @code{=} function tests whether two arguments, both numbers or
4544 Test whether two objects are the same. @code{equal} uses one meaning
4545 of the word `same' and @code{eq} uses another: @code{equal} returns
4546 true if the two objects have a similar structure and contents, such as
4547 two copies of the same book. On the other hand, @code{eq}, returns
4548 true if both arguments are actually the same object.
4557 The @code{string-lessp} function tests whether its first argument is
4558 smaller than the second argument. A shorter, alternative name for the
4559 same function (a @code{defalias}) is @code{string<}.
4561 The arguments to @code{string-lessp} must be strings or symbols; the
4562 ordering is lexicographic, so case is significant. The print names of
4563 symbols are used instead of the symbols themselves.
4565 @cindex @samp{empty string} defined
4566 An empty string, @samp{""}, a string with no characters in it, is
4567 smaller than any string of characters.
4569 @code{string-equal} provides the corresponding test for equality. Its
4570 shorter, alternative name is @code{string=}. There are no string test
4571 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4574 Print a message in the echo area. The first argument is a string that
4575 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4576 arguments that follow the string. The argument used by @samp{%s} must
4577 be a string or a symbol; the argument used by @samp{%d} must be a
4578 number. The argument used by @samp{%c} must be an @sc{ascii} code
4579 number; it will be printed as the character with that @sc{ascii} code.
4580 (Various other %-sequences have not been mentioned.)
4584 The @code{setq} function sets the value of its first argument to the
4585 value of the second argument. The first argument is automatically
4586 quoted by @code{setq}. It does the same for succeeding pairs of
4587 arguments. Another function, @code{set}, takes only two arguments and
4588 evaluates both of them before setting the value returned by its first
4589 argument to the value returned by its second argument.
4592 Without an argument, return the name of the buffer, as a string.
4594 @item buffer-file-name
4595 Without an argument, return the name of the file the buffer is
4598 @item current-buffer
4599 Return the buffer in which Emacs is active; it may not be
4600 the buffer that is visible on the screen.
4603 Return the most recently selected buffer (other than the buffer passed
4604 to @code{other-buffer} as an argument and other than the current
4607 @item switch-to-buffer
4608 Select a buffer for Emacs to be active in and display it in the current
4609 window so users can look at it. Usually bound to @kbd{C-x b}.
4612 Switch Emacs's attention to a buffer on which programs will run. Don't
4613 alter what the window is showing.
4616 Return the number of characters in the current buffer.
4619 Return the value of the current position of the cursor, as an
4620 integer counting the number of characters from the beginning of the
4624 Return the minimum permissible value of point in
4625 the current buffer. This is 1, unless narrowing is in effect.
4628 Return the value of the maximum permissible value of point in the
4629 current buffer. This is the end of the buffer, unless narrowing is in
4634 @node defun Exercises
4639 Write a non-interactive function that doubles the value of its
4640 argument, a number. Make that function interactive.
4643 Write a function that tests whether the current value of
4644 @code{fill-column} is greater than the argument passed to the function,
4645 and if so, prints an appropriate message.
4648 @node Buffer Walk Through
4649 @chapter A Few Buffer--Related Functions
4651 In this chapter we study in detail several of the functions used in GNU
4652 Emacs. This is called a ``walk-through''. These functions are used as
4653 examples of Lisp code, but are not imaginary examples; with the
4654 exception of the first, simplified function definition, these functions
4655 show the actual code used in GNU Emacs. You can learn a great deal from
4656 these definitions. The functions described here are all related to
4657 buffers. Later, we will study other functions.
4660 * Finding More:: How to find more information.
4661 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4662 @code{point-min}, and @code{push-mark}.
4663 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4664 * append-to-buffer:: Uses @code{save-excursion} and
4665 @code{insert-buffer-substring}.
4666 * Buffer Related Review:: Review.
4667 * Buffer Exercises::
4671 @section Finding More Information
4673 @findex describe-function, @r{introduced}
4674 @cindex Find function documentation
4675 In this walk-through, I will describe each new function as we come to
4676 it, sometimes in detail and sometimes briefly. If you are interested,
4677 you can get the full documentation of any Emacs Lisp function at any
4678 time by typing @kbd{C-h f} and then the name of the function (and then
4679 @key{RET}). Similarly, you can get the full documentation for a
4680 variable by typing @kbd{C-h v} and then the name of the variable (and
4683 @cindex Find source of function
4684 @c In version 22, tells location both of C and of Emacs Lisp
4685 Also, @code{describe-function} will tell you the location of the
4686 function definition.
4688 Put point into the name of the file that contains the function and
4689 press the @key{RET} key. In this case, @key{RET} means
4690 @code{push-button} rather than `return' or `enter'. Emacs will take
4691 you directly to the function definition.
4696 If you move point over the file name and press
4697 the @key{RET} key, which in this case means @code{help-follow} rather
4698 than `return' or `enter', Emacs will take you directly to the function
4702 More generally, if you want to see a function in its original source
4703 file, you can use the @code{find-tag} function to jump to it.
4704 @code{find-tag} works with a wide variety of languages, not just
4705 Lisp, and C, and it works with non-programming text as well. For
4706 example, @code{find-tag} will jump to the various nodes in the
4707 Texinfo source file of this document.
4708 The @code{find-tag} function depends on `tags tables' that record
4709 the locations of the functions, variables, and other items to which
4710 @code{find-tag} jumps.
4712 To use the @code{find-tag} command, type @kbd{M-.} (i.e., press the
4713 period key while holding down the @key{META} key, or else type the
4714 @key{ESC} key and then type the period key), and then, at the prompt,
4715 type in the name of the function whose source code you want to see,
4716 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4717 switch buffers and display the source code for the function on your
4718 screen. To switch back to your current buffer, type @kbd{C-x b
4719 @key{RET}}. (On some keyboards, the @key{META} key is labeled
4722 @c !!! 22.1.1 tags table location in this paragraph
4723 @cindex TAGS table, specifying
4725 Depending on how the initial default values of your copy of Emacs are
4726 set, you may also need to specify the location of your `tags table',
4727 which is a file called @file{TAGS}. For example, if you are
4728 interested in Emacs sources, the tags table you will most likely want,
4729 if it has already been created for you, will be in a subdirectory of
4730 the @file{/usr/local/share/emacs/} directory; thus you would use the
4731 @code{M-x visit-tags-table} command and specify a pathname such as
4732 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4733 has not already been created, you will have to create it yourself. It
4734 will be in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4737 To create a @file{TAGS} file in a specific directory, switch to that
4738 directory in Emacs using @kbd{M-x cd} command, or list the directory
4739 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4740 @w{@code{etags *.el}} as the command to execute:
4743 M-x compile RET etags *.el RET
4746 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4748 After you become more familiar with Emacs Lisp, you will find that you will
4749 frequently use @code{find-tag} to navigate your way around source code;
4750 and you will create your own @file{TAGS} tables.
4752 @cindex Library, as term for `file'
4753 Incidentally, the files that contain Lisp code are conventionally
4754 called @dfn{libraries}. The metaphor is derived from that of a
4755 specialized library, such as a law library or an engineering library,
4756 rather than a general library. Each library, or file, contains
4757 functions that relate to a particular topic or activity, such as
4758 @file{abbrev.el} for handling abbreviations and other typing
4759 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4760 libraries provide code for a single activity, as the various
4761 @file{rmail@dots{}} files provide code for reading electronic mail.)
4762 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4763 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4764 by topic keywords.''
4766 @node simplified-beginning-of-buffer
4767 @section A Simplified @code{beginning-of-buffer} Definition
4768 @findex simplified-beginning-of-buffer
4770 The @code{beginning-of-buffer} command is a good function to start with
4771 since you are likely to be familiar with it and it is easy to
4772 understand. Used as an interactive command, @code{beginning-of-buffer}
4773 moves the cursor to the beginning of the buffer, leaving the mark at the
4774 previous position. It is generally bound to @kbd{M-<}.
4776 In this section, we will discuss a shortened version of the function
4777 that shows how it is most frequently used. This shortened function
4778 works as written, but it does not contain the code for a complex option.
4779 In another section, we will describe the entire function.
4780 (@xref{beginning-of-buffer, , Complete Definition of
4781 @code{beginning-of-buffer}}.)
4783 Before looking at the code, let's consider what the function
4784 definition has to contain: it must include an expression that makes
4785 the function interactive so it can be called by typing @kbd{M-x
4786 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4787 must include code to leave a mark at the original position in the
4788 buffer; and it must include code to move the cursor to the beginning
4792 Here is the complete text of the shortened version of the function:
4796 (defun simplified-beginning-of-buffer ()
4797 "Move point to the beginning of the buffer;
4798 leave mark at previous position."
4801 (goto-char (point-min)))
4805 Like all function definitions, this definition has five parts following
4806 the special form @code{defun}:
4810 The name: in this example, @code{simplified-beginning-of-buffer}.
4813 A list of the arguments: in this example, an empty list, @code{()},
4816 The documentation string.
4819 The interactive expression.
4826 In this function definition, the argument list is empty; this means that
4827 this function does not require any arguments. (When we look at the
4828 definition for the complete function, we will see that it may be passed
4829 an optional argument.)
4831 The interactive expression tells Emacs that the function is intended to
4832 be used interactively. In this example, @code{interactive} does not have
4833 an argument because @code{simplified-beginning-of-buffer} does not
4837 The body of the function consists of the two lines:
4842 (goto-char (point-min))
4846 The first of these lines is the expression, @code{(push-mark)}. When
4847 this expression is evaluated by the Lisp interpreter, it sets a mark at
4848 the current position of the cursor, wherever that may be. The position
4849 of this mark is saved in the mark ring.
4851 The next line is @code{(goto-char (point-min))}. This expression
4852 jumps the cursor to the minimum point in the buffer, that is, to the
4853 beginning of the buffer (or to the beginning of the accessible portion
4854 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4855 Narrowing and Widening}.)
4857 The @code{push-mark} command sets a mark at the place where the cursor
4858 was located before it was moved to the beginning of the buffer by the
4859 @code{(goto-char (point-min))} expression. Consequently, you can, if
4860 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4862 That is all there is to the function definition!
4864 @findex describe-function
4865 When you are reading code such as this and come upon an unfamiliar
4866 function, such as @code{goto-char}, you can find out what it does by
4867 using the @code{describe-function} command. To use this command, type
4868 @kbd{C-h f} and then type in the name of the function and press
4869 @key{RET}. The @code{describe-function} command will print the
4870 function's documentation string in a @file{*Help*} window. For
4871 example, the documentation for @code{goto-char} is:
4875 Set point to POSITION, a number or marker.
4876 Beginning of buffer is position (point-min), end is (point-max).
4881 The function's one argument is the desired position.
4884 (The prompt for @code{describe-function} will offer you the symbol
4885 under or preceding the cursor, so you can save typing by positioning
4886 the cursor right over or after the function and then typing @kbd{C-h f
4889 The @code{end-of-buffer} function definition is written in the same way as
4890 the @code{beginning-of-buffer} definition except that the body of the
4891 function contains the expression @code{(goto-char (point-max))} in place
4892 of @code{(goto-char (point-min))}.
4894 @node mark-whole-buffer
4895 @section The Definition of @code{mark-whole-buffer}
4896 @findex mark-whole-buffer
4898 The @code{mark-whole-buffer} function is no harder to understand than the
4899 @code{simplified-beginning-of-buffer} function. In this case, however,
4900 we will look at the complete function, not a shortened version.
4902 The @code{mark-whole-buffer} function is not as commonly used as the
4903 @code{beginning-of-buffer} function, but is useful nonetheless: it
4904 marks a whole buffer as a region by putting point at the beginning and
4905 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4909 * mark-whole-buffer overview::
4910 * Body of mark-whole-buffer:: Only three lines of code.
4914 @node mark-whole-buffer overview
4915 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4919 In GNU Emacs 22, the code for the complete function looks like this:
4923 (defun mark-whole-buffer ()
4924 "Put point at beginning and mark at end of buffer.
4925 You probably should not use this function in Lisp programs;
4926 it is usually a mistake for a Lisp function to use any subroutine
4927 that uses or sets the mark."
4930 (push-mark (point-max) nil t)
4931 (goto-char (point-min)))
4936 Like all other functions, the @code{mark-whole-buffer} function fits
4937 into the template for a function definition. The template looks like
4942 (defun @var{name-of-function} (@var{argument-list})
4943 "@var{documentation}@dots{}"
4944 (@var{interactive-expression}@dots{})
4949 Here is how the function works: the name of the function is
4950 @code{mark-whole-buffer}; it is followed by an empty argument list,
4951 @samp{()}, which means that the function does not require arguments.
4952 The documentation comes next.
4954 The next line is an @code{(interactive)} expression that tells Emacs
4955 that the function will be used interactively. These details are similar
4956 to the @code{simplified-beginning-of-buffer} function described in the
4960 @node Body of mark-whole-buffer
4961 @subsection Body of @code{mark-whole-buffer}
4963 The body of the @code{mark-whole-buffer} function consists of three
4970 (push-mark (point-max) nil t)
4971 (goto-char (point-min))
4975 The first of these lines is the expression, @code{(push-mark (point))}.
4977 This line does exactly the same job as the first line of the body of
4978 the @code{simplified-beginning-of-buffer} function, which is written
4979 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
4980 at the current position of the cursor.
4982 I don't know why the expression in @code{mark-whole-buffer} is written
4983 @code{(push-mark (point))} and the expression in
4984 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
4985 whoever wrote the code did not know that the arguments for
4986 @code{push-mark} are optional and that if @code{push-mark} is not
4987 passed an argument, the function automatically sets mark at the
4988 location of point by default. Or perhaps the expression was written
4989 so as to parallel the structure of the next line. In any case, the
4990 line causes Emacs to determine the position of point and set a mark
4993 In earlier versions of GNU Emacs, the next line of
4994 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
4995 expression sets a mark at the point in the buffer that has the highest
4996 number. This will be the end of the buffer (or, if the buffer is
4997 narrowed, the end of the accessible portion of the buffer.
4998 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
4999 narrowing.) After this mark has been set, the previous mark, the one
5000 set at point, is no longer set, but Emacs remembers its position, just
5001 as all other recent marks are always remembered. This means that you
5002 can, if you wish, go back to that position by typing @kbd{C-u
5006 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5010 (push-mark (point-max) nil t)
5014 The expression works nearly the same as before. It sets a mark at the
5015 highest numbered place in the buffer that it can. However, in this
5016 version, @code{push-mark} has two additional arguments. The second
5017 argument to @code{push-mark} is @code{nil}. This tells the function
5018 it @emph{should} display a message that says `Mark set' when it pushes
5019 the mark. The third argument is @code{t}. This tells
5020 @code{push-mark} to activate the mark when Transient Mark mode is
5021 turned on. Transient Mark mode highlights the currently active
5022 region. It is often turned off.
5024 Finally, the last line of the function is @code{(goto-char
5025 (point-min)))}. This is written exactly the same way as it is written
5026 in @code{beginning-of-buffer}. The expression moves the cursor to
5027 the minimum point in the buffer, that is, to the beginning of the buffer
5028 (or to the beginning of the accessible portion of the buffer). As a
5029 result of this, point is placed at the beginning of the buffer and mark
5030 is set at the end of the buffer. The whole buffer is, therefore, the
5033 @node append-to-buffer
5034 @section The Definition of @code{append-to-buffer}
5035 @findex append-to-buffer
5037 The @code{append-to-buffer} command is more complex than the
5038 @code{mark-whole-buffer} command. What it does is copy the region
5039 (that is, the part of the buffer between point and mark) from the
5040 current buffer to a specified buffer.
5043 * append-to-buffer overview::
5044 * append interactive:: A two part interactive expression.
5045 * append-to-buffer body:: Incorporates a @code{let} expression.
5046 * append save-excursion:: How the @code{save-excursion} works.
5050 @node append-to-buffer overview
5051 @unnumberedsubsec An Overview of @code{append-to-buffer}
5054 @findex insert-buffer-substring
5055 The @code{append-to-buffer} command uses the
5056 @code{insert-buffer-substring} function to copy the region.
5057 @code{insert-buffer-substring} is described by its name: it takes a
5058 string of characters from part of a buffer, a ``substring'', and
5059 inserts them into another buffer.
5061 Most of @code{append-to-buffer} is
5062 concerned with setting up the conditions for
5063 @code{insert-buffer-substring} to work: the code must specify both the
5064 buffer to which the text will go, the window it comes from and goes
5065 to, and the region that will be copied.
5068 Here is the complete text of the function:
5072 (defun append-to-buffer (buffer start end)
5073 "Append to specified buffer the text of the region.
5074 It is inserted into that buffer before its point.
5078 When calling from a program, give three arguments:
5079 BUFFER (or buffer name), START and END.
5080 START and END specify the portion of the current buffer to be copied."
5082 (list (read-buffer "Append to buffer: " (other-buffer
5083 (current-buffer) t))
5084 (region-beginning) (region-end)))
5087 (let ((oldbuf (current-buffer)))
5089 (let* ((append-to (get-buffer-create buffer))
5090 (windows (get-buffer-window-list append-to t t))
5092 (set-buffer append-to)
5093 (setq point (point))
5094 (barf-if-buffer-read-only)
5095 (insert-buffer-substring oldbuf start end)
5096 (dolist (window windows)
5097 (when (= (window-point window) point)
5098 (set-window-point window (point))))))))
5102 The function can be understood by looking at it as a series of
5103 filled-in templates.
5105 The outermost template is for the function definition. In this
5106 function, it looks like this (with several slots filled in):
5110 (defun append-to-buffer (buffer start end)
5111 "@var{documentation}@dots{}"
5112 (interactive @dots{})
5117 The first line of the function includes its name and three arguments.
5118 The arguments are the @code{buffer} to which the text will be copied, and
5119 the @code{start} and @code{end} of the region in the current buffer that
5122 The next part of the function is the documentation, which is clear and
5123 complete. As is conventional, the three arguments are written in
5124 upper case so you will notice them easily. Even better, they are
5125 described in the same order as in the argument list.
5127 Note that the documentation distinguishes between a buffer and its
5128 name. (The function can handle either.)
5130 @node append interactive
5131 @subsection The @code{append-to-buffer} Interactive Expression
5133 Since the @code{append-to-buffer} function will be used interactively,
5134 the function must have an @code{interactive} expression. (For a
5135 review of @code{interactive}, see @ref{Interactive, , Making a
5136 Function Interactive}.) The expression reads as follows:
5142 "Append to buffer: "
5143 (other-buffer (current-buffer) t))
5150 This expression is not one with letters standing for parts, as
5151 described earlier. Instead, it starts a list with these parts:
5153 The first part of the list is an expression to read the name of a
5154 buffer and return it as a string. That is @code{read-buffer}. The
5155 function requires a prompt as its first argument, @samp{"Append to
5156 buffer: "}. Its second argument tells the command what value to
5157 provide if you don't specify anything.
5159 In this case that second argument is an expression containing the
5160 function @code{other-buffer}, an exception, and a @samp{t}, standing
5163 The first argument to @code{other-buffer}, the exception, is yet
5164 another function, @code{current-buffer}. That is not going to be
5165 returned. The second argument is the symbol for true, @code{t}. that
5166 tells @code{other-buffer} that it may show visible buffers (except in
5167 this case, it will not show the current buffer, which makes sense).
5170 The expression looks like this:
5173 (other-buffer (current-buffer) t)
5176 The second and third arguments to the @code{list} expression are
5177 @code{(region-beginning)} and @code{(region-end)}. These two
5178 functions specify the beginning and end of the text to be appended.
5181 Originally, the command used the letters @samp{B} and @samp{r}.
5182 The whole @code{interactive} expression looked like this:
5185 (interactive "BAppend to buffer:@: \nr")
5189 But when that was done, the default value of the buffer switched to
5190 was invisible. That was not wanted.
5192 (The prompt was separated from the second argument with a newline,
5193 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5194 two arguments that follow the symbol @code{buffer} in the function's
5195 argument list (that is, @code{start} and @code{end}) to the values of
5196 point and mark. That argument worked fine.)
5198 @node append-to-buffer body
5199 @subsection The Body of @code{append-to-buffer}
5202 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5204 (defun append-to-buffer (buffer start end)
5205 "Append to specified buffer the text of the region.
5206 It is inserted into that buffer before its point.
5208 When calling from a program, give three arguments:
5209 BUFFER (or buffer name), START and END.
5210 START and END specify the portion of the current buffer to be copied."
5212 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5213 (region-beginning) (region-end)))
5214 (let ((oldbuf (current-buffer)))
5216 (let* ((append-to (get-buffer-create buffer))
5217 (windows (get-buffer-window-list append-to t t))
5219 (set-buffer append-to)
5220 (setq point (point))
5221 (barf-if-buffer-read-only)
5222 (insert-buffer-substring oldbuf start end)
5223 (dolist (window windows)
5224 (when (= (window-point window) point)
5225 (set-window-point window (point))))))))
5228 The body of the @code{append-to-buffer} function begins with @code{let}.
5230 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5231 @code{let} expression is to create and give initial values to one or
5232 more variables that will only be used within the body of the
5233 @code{let}. This means that such a variable will not be confused with
5234 any variable of the same name outside the @code{let} expression.
5236 We can see how the @code{let} expression fits into the function as a
5237 whole by showing a template for @code{append-to-buffer} with the
5238 @code{let} expression in outline:
5242 (defun append-to-buffer (buffer start end)
5243 "@var{documentation}@dots{}"
5244 (interactive @dots{})
5245 (let ((@var{variable} @var{value}))
5250 The @code{let} expression has three elements:
5254 The symbol @code{let};
5257 A varlist containing, in this case, a single two-element list,
5258 @code{(@var{variable} @var{value})};
5261 The body of the @code{let} expression.
5265 In the @code{append-to-buffer} function, the varlist looks like this:
5268 (oldbuf (current-buffer))
5272 In this part of the @code{let} expression, the one variable,
5273 @code{oldbuf}, is bound to the value returned by the
5274 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5275 used to keep track of the buffer in which you are working and from
5276 which you will copy.
5278 The element or elements of a varlist are surrounded by a set of
5279 parentheses so the Lisp interpreter can distinguish the varlist from
5280 the body of the @code{let}. As a consequence, the two-element list
5281 within the varlist is surrounded by a circumscribing set of parentheses.
5282 The line looks like this:
5286 (let ((oldbuf (current-buffer)))
5292 The two parentheses before @code{oldbuf} might surprise you if you did
5293 not realize that the first parenthesis before @code{oldbuf} marks the
5294 boundary of the varlist and the second parenthesis marks the beginning
5295 of the two-element list, @code{(oldbuf (current-buffer))}.
5297 @node append save-excursion
5298 @subsection @code{save-excursion} in @code{append-to-buffer}
5300 The body of the @code{let} expression in @code{append-to-buffer}
5301 consists of a @code{save-excursion} expression.
5303 The @code{save-excursion} function saves the locations of point and
5304 mark, and restores them to those positions after the expressions in the
5305 body of the @code{save-excursion} complete execution. In addition,
5306 @code{save-excursion} keeps track of the original buffer, and
5307 restores it. This is how @code{save-excursion} is used in
5308 @code{append-to-buffer}.
5311 @cindex Indentation for formatting
5312 @cindex Formatting convention
5313 Incidentally, it is worth noting here that a Lisp function is normally
5314 formatted so that everything that is enclosed in a multi-line spread is
5315 indented more to the right than the first symbol. In this function
5316 definition, the @code{let} is indented more than the @code{defun}, and
5317 the @code{save-excursion} is indented more than the @code{let}, like
5333 This formatting convention makes it easy to see that the lines in
5334 the body of the @code{save-excursion} are enclosed by the parentheses
5335 associated with @code{save-excursion}, just as the
5336 @code{save-excursion} itself is enclosed by the parentheses associated
5337 with the @code{let}:
5341 (let ((oldbuf (current-buffer)))
5344 (set-buffer @dots{})
5345 (insert-buffer-substring oldbuf start end)
5351 The use of the @code{save-excursion} function can be viewed as a process
5352 of filling in the slots of a template:
5357 @var{first-expression-in-body}
5358 @var{second-expression-in-body}
5360 @var{last-expression-in-body})
5366 In this function, the body of the @code{save-excursion} contains only
5367 one expression, the @code{let*} expression. You know about a
5368 @code{let} function. The @code{let*} function is different. It has a
5369 @samp{*} in its name. It enables Emacs to set each variable in its
5370 varlist in sequence, one after another.
5372 Its critical feature is that variables later in the varlist can make
5373 use of the values to which Emacs set variables earlier in the varlist.
5374 @xref{fwd-para let, , The @code{let*} expression}.
5376 We will skip functions like @code{let*} and focus on two: the
5377 @code{set-buffer} function and the @code{insert-buffer-substring}
5381 In the old days, the @code{set-buffer} expression was simply
5384 (set-buffer (get-buffer-create buffer))
5392 (set-buffer append-to)
5396 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5397 on in the @code{let*} expression. That extra binding would not be
5398 necessary except for that @code{append-to} is used later in the
5399 varlist as an argument to @code{get-buffer-window-list}.
5404 (let ((oldbuf (current-buffer)))
5406 (let* ((append-to (get-buffer-create buffer))
5407 (windows (get-buffer-window-list append-to t t))
5409 (set-buffer append-to)
5410 (setq point (point))
5411 (barf-if-buffer-read-only)
5412 (insert-buffer-substring oldbuf start end)
5413 (dolist (window windows)
5414 (when (= (window-point window) point)
5415 (set-window-point window (point))))))))
5418 The @code{append-to-buffer} function definition inserts text from the
5419 buffer in which you are currently to a named buffer. It happens that
5420 @code{insert-buffer-substring} copies text from another buffer to the
5421 current buffer, just the reverse---that is why the
5422 @code{append-to-buffer} definition starts out with a @code{let} that
5423 binds the local symbol @code{oldbuf} to the value returned by
5424 @code{current-buffer}.
5427 The @code{insert-buffer-substring} expression looks like this:
5430 (insert-buffer-substring oldbuf start end)
5434 The @code{insert-buffer-substring} function copies a string
5435 @emph{from} the buffer specified as its first argument and inserts the
5436 string into the present buffer. In this case, the argument to
5437 @code{insert-buffer-substring} is the value of the variable created
5438 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5439 was the current buffer when you gave the @code{append-to-buffer}
5442 After @code{insert-buffer-substring} has done its work,
5443 @code{save-excursion} will restore the action to the original buffer
5444 and @code{append-to-buffer} will have done its job.
5447 Written in skeletal form, the workings of the body look like this:
5451 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5452 (save-excursion ; @r{Keep track of buffer.}
5454 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5456 @var{change-back-to-original-buffer-when-finished}
5457 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5461 In summary, @code{append-to-buffer} works as follows: it saves the
5462 value of the current buffer in the variable called @code{oldbuf}. It
5463 gets the new buffer (creating one if need be) and switches Emacs's
5464 attention to it. Using the value of @code{oldbuf}, it inserts the
5465 region of text from the old buffer into the new buffer; and then using
5466 @code{save-excursion}, it brings you back to your original buffer.
5468 In looking at @code{append-to-buffer}, you have explored a fairly
5469 complex function. It shows how to use @code{let} and
5470 @code{save-excursion}, and how to change to and come back from another
5471 buffer. Many function definitions use @code{let},
5472 @code{save-excursion}, and @code{set-buffer} this way.
5474 @node Buffer Related Review
5477 Here is a brief summary of the various functions discussed in this chapter.
5480 @item describe-function
5481 @itemx describe-variable
5482 Print the documentation for a function or variable.
5483 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5486 Find the file containing the source for a function or variable and
5487 switch buffers to it, positioning point at the beginning of the item.
5488 Conventionally bound to @kbd{M-.} (that's a period following the
5491 @item save-excursion
5492 Save the location of point and mark and restore their values after the
5493 arguments to @code{save-excursion} have been evaluated. Also, remember
5494 the current buffer and return to it.
5497 Set mark at a location and record the value of the previous mark on the
5498 mark ring. The mark is a location in the buffer that will keep its
5499 relative position even if text is added to or removed from the buffer.
5502 Set point to the location specified by the value of the argument, which
5503 can be a number, a marker, or an expression that returns the number of
5504 a position, such as @code{(point-min)}.
5506 @item insert-buffer-substring
5507 Copy a region of text from a buffer that is passed to the function as
5508 an argument and insert the region into the current buffer.
5510 @item mark-whole-buffer
5511 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5514 Switch the attention of Emacs to another buffer, but do not change the
5515 window being displayed. Used when the program rather than a human is
5516 to work on a different buffer.
5518 @item get-buffer-create
5520 Find a named buffer or create one if a buffer of that name does not
5521 exist. The @code{get-buffer} function returns @code{nil} if the named
5522 buffer does not exist.
5526 @node Buffer Exercises
5531 Write your own @code{simplified-end-of-buffer} function definition;
5532 then test it to see whether it works.
5535 Use @code{if} and @code{get-buffer} to write a function that prints a
5536 message telling you whether a buffer exists.
5539 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5544 @chapter A Few More Complex Functions
5546 In this chapter, we build on what we have learned in previous chapters
5547 by looking at more complex functions. The @code{copy-to-buffer}
5548 function illustrates use of two @code{save-excursion} expressions in
5549 one definition, while the @code{insert-buffer} function illustrates
5550 use of an asterisk in an @code{interactive} expression, use of
5551 @code{or}, and the important distinction between a name and the object
5552 to which the name refers.
5555 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5556 * insert-buffer:: Read-only, and with @code{or}.
5557 * beginning-of-buffer:: Shows @code{goto-char},
5558 @code{point-min}, and @code{push-mark}.
5559 * Second Buffer Related Review::
5560 * optional Exercise::
5563 @node copy-to-buffer
5564 @section The Definition of @code{copy-to-buffer}
5565 @findex copy-to-buffer
5567 After understanding how @code{append-to-buffer} works, it is easy to
5568 understand @code{copy-to-buffer}. This function copies text into a
5569 buffer, but instead of adding to the second buffer, it replaces all the
5570 previous text in the second buffer.
5573 The body of @code{copy-to-buffer} looks like this,
5578 (interactive "BCopy to buffer: \nr")
5579 (let ((oldbuf (current-buffer)))
5580 (with-current-buffer (get-buffer-create buffer)
5581 (barf-if-buffer-read-only)
5584 (insert-buffer-substring oldbuf start end)))))
5588 The @code{copy-to-buffer} function has a simpler @code{interactive}
5589 expression than @code{append-to-buffer}.
5592 The definition then says
5595 (with-current-buffer (get-buffer-create buffer) @dots{}
5598 First, look at the earliest inner expression; that is evaluated first.
5599 That expression starts with @code{get-buffer-create buffer}. The
5600 function tells the computer to use the buffer with the name specified
5601 as the one to which you are copying, or if such a buffer does not
5602 exist, to create it. Then, the @code{with-current-buffer} function
5603 evaluates its body with that buffer temporarily current.
5605 (This demonstrates another way to shift the computer's attention but
5606 not the user's. The @code{append-to-buffer} function showed how to do
5607 the same with @code{save-excursion} and @code{set-buffer}.
5608 @code{with-current-buffer} is a newer, and arguably easier,
5611 The @code{barf-if-buffer-read-only} function sends you an error
5612 message saying the buffer is read-only if you cannot modify it.
5614 The next line has the @code{erase-buffer} function as its sole
5615 contents. That function erases the buffer.
5617 Finally, the last two lines contain the @code{save-excursion}
5618 expression with @code{insert-buffer-substring} as its body.
5619 The @code{insert-buffer-substring} expression copies the text from
5620 the buffer you are in (and you have not seen the computer shift its
5621 attention, so you don't know that that buffer is now called
5624 Incidentally, this is what is meant by `replacement'. To replace text,
5625 Emacs erases the previous text and then inserts new text.
5628 In outline, the body of @code{copy-to-buffer} looks like this:
5632 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5633 (@var{with-the-buffer-you-are-copying-to}
5634 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5637 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5642 @section The Definition of @code{insert-buffer}
5643 @findex insert-buffer
5645 @code{insert-buffer} is yet another buffer-related function. This
5646 command copies another buffer @emph{into} the current buffer. It is the
5647 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5648 copy a region of text @emph{from} the current buffer to another buffer.
5650 Here is a discussion based on the original code. The code was
5651 simplified in 2003 and is harder to understand.
5653 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5654 a discussion of the new body.)
5656 In addition, this code illustrates the use of @code{interactive} with a
5657 buffer that might be @dfn{read-only} and the important distinction
5658 between the name of an object and the object actually referred to.
5661 * insert-buffer code::
5662 * insert-buffer interactive:: When you can read, but not write.
5663 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5664 * if & or:: Using an @code{if} instead of an @code{or}.
5665 * Insert or:: How the @code{or} expression works.
5666 * Insert let:: Two @code{save-excursion} expressions.
5667 * New insert-buffer::
5671 @node insert-buffer code
5672 @unnumberedsubsec The Code for @code{insert-buffer}
5676 Here is the earlier code:
5680 (defun insert-buffer (buffer)
5681 "Insert after point the contents of BUFFER.
5682 Puts mark after the inserted text.
5683 BUFFER may be a buffer or a buffer name."
5684 (interactive "*bInsert buffer:@: ")
5687 (or (bufferp buffer)
5688 (setq buffer (get-buffer buffer)))
5689 (let (start end newmark)
5693 (setq start (point-min) end (point-max)))
5696 (insert-buffer-substring buffer start end)
5697 (setq newmark (point)))
5698 (push-mark newmark)))
5703 As with other function definitions, you can use a template to see an
5704 outline of the function:
5708 (defun insert-buffer (buffer)
5709 "@var{documentation}@dots{}"
5710 (interactive "*bInsert buffer:@: ")
5715 @node insert-buffer interactive
5716 @subsection The Interactive Expression in @code{insert-buffer}
5717 @findex interactive, @r{example use of}
5719 In @code{insert-buffer}, the argument to the @code{interactive}
5720 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5724 * Read-only buffer:: When a buffer cannot be modified.
5725 * b for interactive:: An existing buffer or else its name.
5728 @node Read-only buffer
5729 @unnumberedsubsubsec A Read-only Buffer
5730 @cindex Read-only buffer
5731 @cindex Asterisk for read-only buffer
5732 @findex * @r{for read-only buffer}
5734 The asterisk is for the situation when the current buffer is a
5735 read-only buffer---a buffer that cannot be modified. If
5736 @code{insert-buffer} is called when the current buffer is read-only, a
5737 message to this effect is printed in the echo area and the terminal
5738 may beep or blink at you; you will not be permitted to insert anything
5739 into current buffer. The asterisk does not need to be followed by a
5740 newline to separate it from the next argument.
5742 @node b for interactive
5743 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5745 The next argument in the interactive expression starts with a lower
5746 case @samp{b}. (This is different from the code for
5747 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5748 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5749 The lower-case @samp{b} tells the Lisp interpreter that the argument
5750 for @code{insert-buffer} should be an existing buffer or else its
5751 name. (The upper-case @samp{B} option provides for the possibility
5752 that the buffer does not exist.) Emacs will prompt you for the name
5753 of the buffer, offering you a default buffer, with name completion
5754 enabled. If the buffer does not exist, you receive a message that
5755 says ``No match''; your terminal may beep at you as well.
5757 The new and simplified code generates a list for @code{interactive}.
5758 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5759 functions with which we are already familiar and the @code{progn}
5760 special form with which we are not. (It will be described later.)
5762 @node insert-buffer body
5763 @subsection The Body of the @code{insert-buffer} Function
5765 The body of the @code{insert-buffer} function has two major parts: an
5766 @code{or} expression and a @code{let} expression. The purpose of the
5767 @code{or} expression is to ensure that the argument @code{buffer} is
5768 bound to a buffer and not just the name of a buffer. The body of the
5769 @code{let} expression contains the code which copies the other buffer
5770 into the current buffer.
5773 In outline, the two expressions fit into the @code{insert-buffer}
5778 (defun insert-buffer (buffer)
5779 "@var{documentation}@dots{}"
5780 (interactive "*bInsert buffer:@: ")
5785 (let (@var{varlist})
5786 @var{body-of-}@code{let}@dots{} )
5790 To understand how the @code{or} expression ensures that the argument
5791 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5792 is first necessary to understand the @code{or} function.
5794 Before doing this, let me rewrite this part of the function using
5795 @code{if} so that you can see what is done in a manner that will be familiar.
5798 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5800 The job to be done is to make sure the value of @code{buffer} is a
5801 buffer itself and not the name of a buffer. If the value is the name,
5802 then the buffer itself must be got.
5804 You can imagine yourself at a conference where an usher is wandering
5805 around holding a list with your name on it and looking for you: the
5806 usher is ``bound'' to your name, not to you; but when the usher finds
5807 you and takes your arm, the usher becomes ``bound'' to you.
5810 In Lisp, you might describe this situation like this:
5814 (if (not (holding-on-to-guest))
5815 (find-and-take-arm-of-guest))
5819 We want to do the same thing with a buffer---if we do not have the
5820 buffer itself, we want to get it.
5823 Using a predicate called @code{bufferp} that tells us whether we have a
5824 buffer (rather than its name), we can write the code like this:
5828 (if (not (bufferp buffer)) ; @r{if-part}
5829 (setq buffer (get-buffer buffer))) ; @r{then-part}
5834 Here, the true-or-false-test of the @code{if} expression is
5835 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5836 @w{@code{(setq buffer (get-buffer buffer))}}.
5838 In the test, the function @code{bufferp} returns true if its argument is
5839 a buffer---but false if its argument is the name of the buffer. (The
5840 last character of the function name @code{bufferp} is the character
5841 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5842 indicates that the function is a predicate, which is a term that means
5843 that the function will determine whether some property is true or false.
5844 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5848 The function @code{not} precedes the expression @code{(bufferp buffer)},
5849 so the true-or-false-test looks like this:
5852 (not (bufferp buffer))
5856 @code{not} is a function that returns true if its argument is false
5857 and false if its argument is true. So if @code{(bufferp buffer)}
5858 returns true, the @code{not} expression returns false and vice-verse:
5859 what is ``not true'' is false and what is ``not false'' is true.
5861 Using this test, the @code{if} expression works as follows: when the
5862 value of the variable @code{buffer} is actually a buffer rather than
5863 its name, the true-or-false-test returns false and the @code{if}
5864 expression does not evaluate the then-part. This is fine, since we do
5865 not need to do anything to the variable @code{buffer} if it really is
5868 On the other hand, when the value of @code{buffer} is not a buffer
5869 itself, but the name of a buffer, the true-or-false-test returns true
5870 and the then-part of the expression is evaluated. In this case, the
5871 then-part is @code{(setq buffer (get-buffer buffer))}. This
5872 expression uses the @code{get-buffer} function to return an actual
5873 buffer itself, given its name. The @code{setq} then sets the variable
5874 @code{buffer} to the value of the buffer itself, replacing its previous
5875 value (which was the name of the buffer).
5878 @subsection The @code{or} in the Body
5880 The purpose of the @code{or} expression in the @code{insert-buffer}
5881 function is to ensure that the argument @code{buffer} is bound to a
5882 buffer and not just to the name of a buffer. The previous section shows
5883 how the job could have been done using an @code{if} expression.
5884 However, the @code{insert-buffer} function actually uses @code{or}.
5885 To understand this, it is necessary to understand how @code{or} works.
5888 An @code{or} function can have any number of arguments. It evaluates
5889 each argument in turn and returns the value of the first of its
5890 arguments that is not @code{nil}. Also, and this is a crucial feature
5891 of @code{or}, it does not evaluate any subsequent arguments after
5892 returning the first non-@code{nil} value.
5895 The @code{or} expression looks like this:
5899 (or (bufferp buffer)
5900 (setq buffer (get-buffer buffer)))
5905 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5906 This expression returns true (a non-@code{nil} value) if the buffer is
5907 actually a buffer, and not just the name of a buffer. In the @code{or}
5908 expression, if this is the case, the @code{or} expression returns this
5909 true value and does not evaluate the next expression---and this is fine
5910 with us, since we do not want to do anything to the value of
5911 @code{buffer} if it really is a buffer.
5913 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5914 which it will be if the value of @code{buffer} is the name of a buffer,
5915 the Lisp interpreter evaluates the next element of the @code{or}
5916 expression. This is the expression @code{(setq buffer (get-buffer
5917 buffer))}. This expression returns a non-@code{nil} value, which
5918 is the value to which it sets the variable @code{buffer}---and this
5919 value is a buffer itself, not the name of a buffer.
5921 The result of all this is that the symbol @code{buffer} is always
5922 bound to a buffer itself rather than to the name of a buffer. All
5923 this is necessary because the @code{set-buffer} function in a
5924 following line only works with a buffer itself, not with the name to a
5928 Incidentally, using @code{or}, the situation with the usher would be
5932 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5936 @subsection The @code{let} Expression in @code{insert-buffer}
5938 After ensuring that the variable @code{buffer} refers to a buffer itself
5939 and not just to the name of a buffer, the @code{insert-buffer function}
5940 continues with a @code{let} expression. This specifies three local
5941 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5942 to the initial value @code{nil}. These variables are used inside the
5943 remainder of the @code{let} and temporarily hide any other occurrence of
5944 variables of the same name in Emacs until the end of the @code{let}.
5947 The body of the @code{let} contains two @code{save-excursion}
5948 expressions. First, we will look at the inner @code{save-excursion}
5949 expression in detail. The expression looks like this:
5955 (setq start (point-min) end (point-max)))
5960 The expression @code{(set-buffer buffer)} changes Emacs's attention
5961 from the current buffer to the one from which the text will copied.
5962 In that buffer, the variables @code{start} and @code{end} are set to
5963 the beginning and end of the buffer, using the commands
5964 @code{point-min} and @code{point-max}. Note that we have here an
5965 illustration of how @code{setq} is able to set two variables in the
5966 same expression. The first argument of @code{setq} is set to the
5967 value of its second, and its third argument is set to the value of its
5970 After the body of the inner @code{save-excursion} is evaluated, the
5971 @code{save-excursion} restores the original buffer, but @code{start} and
5972 @code{end} remain set to the values of the beginning and end of the
5973 buffer from which the text will be copied.
5976 The outer @code{save-excursion} expression looks like this:
5981 (@var{inner-}@code{save-excursion}@var{-expression}
5982 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
5983 (insert-buffer-substring buffer start end)
5984 (setq newmark (point)))
5989 The @code{insert-buffer-substring} function copies the text
5990 @emph{into} the current buffer @emph{from} the region indicated by
5991 @code{start} and @code{end} in @code{buffer}. Since the whole of the
5992 second buffer lies between @code{start} and @code{end}, the whole of
5993 the second buffer is copied into the buffer you are editing. Next,
5994 the value of point, which will be at the end of the inserted text, is
5995 recorded in the variable @code{newmark}.
5997 After the body of the outer @code{save-excursion} is evaluated, point
5998 and mark are relocated to their original places.
6000 However, it is convenient to locate a mark at the end of the newly
6001 inserted text and locate point at its beginning. The @code{newmark}
6002 variable records the end of the inserted text. In the last line of
6003 the @code{let} expression, the @code{(push-mark newmark)} expression
6004 function sets a mark to this location. (The previous location of the
6005 mark is still accessible; it is recorded on the mark ring and you can
6006 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6007 located at the beginning of the inserted text, which is where it was
6008 before you called the insert function, the position of which was saved
6009 by the first @code{save-excursion}.
6012 The whole @code{let} expression looks like this:
6016 (let (start end newmark)
6020 (setq start (point-min) end (point-max)))
6021 (insert-buffer-substring buffer start end)
6022 (setq newmark (point)))
6023 (push-mark newmark))
6027 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6028 function uses @code{let}, @code{save-excursion}, and
6029 @code{set-buffer}. In addition, the function illustrates one way to
6030 use @code{or}. All these functions are building blocks that we will
6031 find and use again and again.
6033 @node New insert-buffer
6034 @subsection New Body for @code{insert-buffer}
6035 @findex insert-buffer, new version body
6036 @findex new version body for insert-buffer
6038 The body in the GNU Emacs 22 version is more confusing than the original.
6041 It consists of two expressions,
6047 (insert-buffer-substring (get-buffer buffer))
6055 except, and this is what confuses novices, very important work is done
6056 inside the @code{push-mark} expression.
6058 The @code{get-buffer} function returns a buffer with the name
6059 provided. You will note that the function is @emph{not} called
6060 @code{get-buffer-create}; it does not create a buffer if one does not
6061 already exist. The buffer returned by @code{get-buffer}, an existing
6062 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6063 whole of the buffer (since you did not specify anything else).
6065 The location into which the buffer is inserted is recorded by
6066 @code{push-mark}. Then the function returns @code{nil}, the value of
6067 its last command. Put another way, the @code{insert-buffer} function
6068 exists only to produce a side effect, inserting another buffer, not to
6071 @node beginning-of-buffer
6072 @section Complete Definition of @code{beginning-of-buffer}
6073 @findex beginning-of-buffer
6075 The basic structure of the @code{beginning-of-buffer} function has
6076 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6077 Simplified @code{beginning-of-buffer} Definition}.)
6078 This section describes the complex part of the definition.
6080 As previously described, when invoked without an argument,
6081 @code{beginning-of-buffer} moves the cursor to the beginning of the
6082 buffer (in truth, the beginning of the accessible portion of the
6083 buffer), leaving the mark at the previous position. However, when the
6084 command is invoked with a number between one and ten, the function
6085 considers that number to be a fraction of the length of the buffer,
6086 measured in tenths, and Emacs moves the cursor that fraction of the
6087 way from the beginning of the buffer. Thus, you can either call this
6088 function with the key command @kbd{M-<}, which will move the cursor to
6089 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6090 M-<} which will move the cursor to a point 70% of the way through the
6091 buffer. If a number bigger than ten is used for the argument, it
6092 moves to the end of the buffer.
6094 The @code{beginning-of-buffer} function can be called with or without an
6095 argument. The use of the argument is optional.
6098 * Optional Arguments::
6099 * beginning-of-buffer opt arg:: Example with optional argument.
6100 * beginning-of-buffer complete::
6103 @node Optional Arguments
6104 @subsection Optional Arguments
6106 Unless told otherwise, Lisp expects that a function with an argument in
6107 its function definition will be called with a value for that argument.
6108 If that does not happen, you get an error and a message that says
6109 @samp{Wrong number of arguments}.
6111 @cindex Optional arguments
6114 However, optional arguments are a feature of Lisp: a particular
6115 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6116 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6117 @samp{optional} is part of the keyword.) In a function definition, if
6118 an argument follows the keyword @code{&optional}, no value need be
6119 passed to that argument when the function is called.
6122 The first line of the function definition of @code{beginning-of-buffer}
6123 therefore looks like this:
6126 (defun beginning-of-buffer (&optional arg)
6130 In outline, the whole function looks like this:
6134 (defun beginning-of-buffer (&optional arg)
6135 "@var{documentation}@dots{}"
6137 (or (@var{is-the-argument-a-cons-cell} arg)
6138 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6140 (let (@var{determine-size-and-set-it})
6142 (@var{if-there-is-an-argument}
6143 @var{figure-out-where-to-go}
6150 The function is similar to the @code{simplified-beginning-of-buffer}
6151 function except that the @code{interactive} expression has @code{"P"}
6152 as an argument and the @code{goto-char} function is followed by an
6153 if-then-else expression that figures out where to put the cursor if
6154 there is an argument that is not a cons cell.
6156 (Since I do not explain a cons cell for many more chapters, please
6157 consider ignoring the function @code{consp}. @xref{List
6158 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6159 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6162 The @code{"P"} in the @code{interactive} expression tells Emacs to
6163 pass a prefix argument, if there is one, to the function in raw form.
6164 A prefix argument is made by typing the @key{META} key followed by a
6165 number, or by typing @kbd{C-u} and then a number. (If you don't type
6166 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6167 @code{"p"} in the @code{interactive} expression causes the function to
6168 convert a prefix arg to a number.)
6170 The true-or-false-test of the @code{if} expression looks complex, but
6171 it is not: it checks whether @code{arg} has a value that is not
6172 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6173 does; it checks whether its argument is a cons cell.) If @code{arg}
6174 has a value that is not @code{nil} (and is not a cons cell), which
6175 will be the case if @code{beginning-of-buffer} is called with a
6176 numeric argument, then this true-or-false-test will return true and
6177 the then-part of the @code{if} expression will be evaluated. On the
6178 other hand, if @code{beginning-of-buffer} is not called with an
6179 argument, the value of @code{arg} will be @code{nil} and the else-part
6180 of the @code{if} expression will be evaluated. The else-part is
6181 simply @code{point-min}, and when this is the outcome, the whole
6182 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6183 is how we saw the @code{beginning-of-buffer} function in its
6186 @node beginning-of-buffer opt arg
6187 @subsection @code{beginning-of-buffer} with an Argument
6189 When @code{beginning-of-buffer} is called with an argument, an
6190 expression is evaluated which calculates what value to pass to
6191 @code{goto-char}. This expression is rather complicated at first sight.
6192 It includes an inner @code{if} expression and much arithmetic. It looks
6197 (if (> (buffer-size) 10000)
6198 ;; @r{Avoid overflow for large buffer sizes!}
6199 (* (prefix-numeric-value arg)
6204 size (prefix-numeric-value arg))) 10)))
6209 * Disentangle beginning-of-buffer::
6210 * Large buffer case::
6211 * Small buffer case::
6215 @node Disentangle beginning-of-buffer
6216 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6219 Like other complex-looking expressions, the conditional expression
6220 within @code{beginning-of-buffer} can be disentangled by looking at it
6221 as parts of a template, in this case, the template for an if-then-else
6222 expression. In skeletal form, the expression looks like this:
6226 (if (@var{buffer-is-large}
6227 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6228 @var{else-use-alternate-calculation}
6232 The true-or-false-test of this inner @code{if} expression checks the
6233 size of the buffer. The reason for this is that the old version 18
6234 Emacs used numbers that are no bigger than eight million or so and in
6235 the computation that followed, the programmer feared that Emacs might
6236 try to use over-large numbers if the buffer were large. The term
6237 `overflow', mentioned in the comment, means numbers that are over
6238 large. More recent versions of Emacs use larger numbers, but this
6239 code has not been touched, if only because people now look at buffers
6240 that are far, far larger than ever before.
6242 There are two cases: if the buffer is large and if it is not.
6244 @node Large buffer case
6245 @unnumberedsubsubsec What happens in a large buffer
6247 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6248 whether the size of the buffer is greater than 10,000 characters. To do
6249 this, it uses the @code{>} function and the computation of @code{size}
6250 that comes from the let expression.
6252 In the old days, the function @code{buffer-size} was used. Not only
6253 was that function called several times, it gave the size of the whole
6254 buffer, not the accessible part. The computation makes much more
6255 sense when it handles just the accessible part. (@xref{Narrowing &
6256 Widening, , Narrowing and Widening}, for more information on focusing
6257 attention to an `accessible' part.)
6260 The line looks like this:
6268 When the buffer is large, the then-part of the @code{if} expression is
6269 evaluated. It reads like this (after formatting for easy reading):
6274 (prefix-numeric-value arg)
6280 This expression is a multiplication, with two arguments to the function
6283 The first argument is @code{(prefix-numeric-value arg)}. When
6284 @code{"P"} is used as the argument for @code{interactive}, the value
6285 passed to the function as its argument is passed a ``raw prefix
6286 argument'', and not a number. (It is a number in a list.) To perform
6287 the arithmetic, a conversion is necessary, and
6288 @code{prefix-numeric-value} does the job.
6290 @findex / @r{(division)}
6292 The second argument is @code{(/ size 10)}. This expression divides
6293 the numeric value by ten---the numeric value of the size of the
6294 accessible portion of the buffer. This produces a number that tells
6295 how many characters make up one tenth of the buffer size. (In Lisp,
6296 @code{/} is used for division, just as @code{*} is used for
6300 In the multiplication expression as a whole, this amount is multiplied
6301 by the value of the prefix argument---the multiplication looks like this:
6305 (* @var{numeric-value-of-prefix-arg}
6306 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6311 If, for example, the prefix argument is @samp{7}, the one-tenth value
6312 will be multiplied by 7 to give a position 70% of the way through.
6315 The result of all this is that if the accessible portion of the buffer
6316 is large, the @code{goto-char} expression reads like this:
6320 (goto-char (* (prefix-numeric-value arg)
6325 This puts the cursor where we want it.
6327 @node Small buffer case
6328 @unnumberedsubsubsec What happens in a small buffer
6330 If the buffer contains fewer than 10,000 characters, a slightly
6331 different computation is performed. You might think this is not
6332 necessary, since the first computation could do the job. However, in
6333 a small buffer, the first method may not put the cursor on exactly the
6334 desired line; the second method does a better job.
6337 The code looks like this:
6339 @c Keep this on one line.
6341 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6346 This is code in which you figure out what happens by discovering how the
6347 functions are embedded in parentheses. It is easier to read if you
6348 reformat it with each expression indented more deeply than its
6349 enclosing expression:
6357 (prefix-numeric-value arg)))
6364 Looking at parentheses, we see that the innermost operation is
6365 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6366 a number. In the following expression, this number is multiplied by
6367 the size of the accessible portion of the buffer:
6370 (* size (prefix-numeric-value arg))
6374 This multiplication creates a number that may be larger than the size of
6375 the buffer---seven times larger if the argument is 7, for example. Ten
6376 is then added to this number and finally the large number is divided by
6377 ten to provide a value that is one character larger than the percentage
6378 position in the buffer.
6380 The number that results from all this is passed to @code{goto-char} and
6381 the cursor is moved to that point.
6384 @node beginning-of-buffer complete
6385 @subsection The Complete @code{beginning-of-buffer}
6388 Here is the complete text of the @code{beginning-of-buffer} function:
6394 (defun beginning-of-buffer (&optional arg)
6395 "Move point to the beginning of the buffer;
6396 leave mark at previous position.
6397 With \\[universal-argument] prefix,
6398 do not set mark at previous position.
6400 put point N/10 of the way from the beginning.
6402 If the buffer is narrowed,
6403 this command uses the beginning and size
6404 of the accessible part of the buffer.
6408 Don't use this command in Lisp programs!
6409 \(goto-char (point-min)) is faster
6410 and avoids clobbering the mark."
6413 (and transient-mark-mode mark-active)
6417 (let ((size (- (point-max) (point-min))))
6418 (goto-char (if (and arg (not (consp arg)))
6421 ;; Avoid overflow for large buffer sizes!
6422 (* (prefix-numeric-value arg)
6424 (/ (+ 10 (* size (prefix-numeric-value arg)))
6427 (if arg (forward-line 1)))
6432 From before GNU Emacs 22
6435 (defun beginning-of-buffer (&optional arg)
6436 "Move point to the beginning of the buffer;
6437 leave mark at previous position.
6438 With arg N, put point N/10 of the way
6439 from the true beginning.
6442 Don't use this in Lisp programs!
6443 \(goto-char (point-min)) is faster
6444 and does not set the mark."
6451 (if (> (buffer-size) 10000)
6452 ;; @r{Avoid overflow for large buffer sizes!}
6453 (* (prefix-numeric-value arg)
6454 (/ (buffer-size) 10))
6457 (/ (+ 10 (* (buffer-size)
6458 (prefix-numeric-value arg)))
6461 (if arg (forward-line 1)))
6467 Except for two small points, the previous discussion shows how this
6468 function works. The first point deals with a detail in the
6469 documentation string, and the second point concerns the last line of
6473 In the documentation string, there is reference to an expression:
6476 \\[universal-argument]
6480 A @samp{\\} is used before the first square bracket of this
6481 expression. This @samp{\\} tells the Lisp interpreter to substitute
6482 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6483 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6484 be different. (@xref{Documentation Tips, , Tips for Documentation
6485 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6489 Finally, the last line of the @code{beginning-of-buffer} command says
6490 to move point to the beginning of the next line if the command is
6491 invoked with an argument:
6494 (if arg (forward-line 1)))
6498 This puts the cursor at the beginning of the first line after the
6499 appropriate tenths position in the buffer. This is a flourish that
6500 means that the cursor is always located @emph{at least} the requested
6501 tenths of the way through the buffer, which is a nicety that is,
6502 perhaps, not necessary, but which, if it did not occur, would be sure
6505 On the other hand, it also means that if you specify the command with
6506 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6507 argument' is simply a cons cell, then the command puts you at the
6508 beginning of the second line @dots{} I don't know whether this is
6509 intended or whether no one has dealt with the code to avoid this
6512 @node Second Buffer Related Review
6515 Here is a brief summary of some of the topics covered in this chapter.
6519 Evaluate each argument in sequence, and return the value of the first
6520 argument that is not @code{nil}; if none return a value that is not
6521 @code{nil}, return @code{nil}. In brief, return the first true value
6522 of the arguments; return a true value if one @emph{or} any of the
6526 Evaluate each argument in sequence, and if any are @code{nil}, return
6527 @code{nil}; if none are @code{nil}, return the value of the last
6528 argument. In brief, return a true value only if all the arguments are
6529 true; return a true value if one @emph{and} each of the others is
6533 A keyword used to indicate that an argument to a function definition
6534 is optional; this means that the function can be evaluated without the
6535 argument, if desired.
6537 @item prefix-numeric-value
6538 Convert the `raw prefix argument' produced by @code{(interactive
6539 "P")} to a numeric value.
6542 Move point forward to the beginning of the next line, or if the argument
6543 is greater than one, forward that many lines. If it can't move as far
6544 forward as it is supposed to, @code{forward-line} goes forward as far as
6545 it can and then returns a count of the number of additional lines it was
6546 supposed to move but couldn't.
6549 Delete the entire contents of the current buffer.
6552 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6555 @node optional Exercise
6556 @section @code{optional} Argument Exercise
6558 Write an interactive function with an optional argument that tests
6559 whether its argument, a number, is greater than or equal to, or else,
6560 less than the value of @code{fill-column}, and tells you which, in a
6561 message. However, if you do not pass an argument to the function, use
6562 56 as a default value.
6564 @node Narrowing & Widening
6565 @chapter Narrowing and Widening
6566 @cindex Focusing attention (narrowing)
6570 Narrowing is a feature of Emacs that makes it possible for you to focus
6571 on a specific part of a buffer, and work without accidentally changing
6572 other parts. Narrowing is normally disabled since it can confuse
6576 * Narrowing advantages:: The advantages of narrowing
6577 * save-restriction:: The @code{save-restriction} special form.
6578 * what-line:: The number of the line that point is on.
6583 @node Narrowing advantages
6584 @unnumberedsec The Advantages of Narrowing
6587 With narrowing, the rest of a buffer is made invisible, as if it weren't
6588 there. This is an advantage if, for example, you want to replace a word
6589 in one part of a buffer but not in another: you narrow to the part you want
6590 and the replacement is carried out only in that section, not in the rest
6591 of the buffer. Searches will only work within a narrowed region, not
6592 outside of one, so if you are fixing a part of a document, you can keep
6593 yourself from accidentally finding parts you do not need to fix by
6594 narrowing just to the region you want.
6595 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6597 However, narrowing does make the rest of the buffer invisible, which
6598 can scare people who inadvertently invoke narrowing and think they
6599 have deleted a part of their file. Moreover, the @code{undo} command
6600 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6601 (nor should it), so people can become quite desperate if they do not
6602 know that they can return the rest of a buffer to visibility with the
6603 @code{widen} command.
6604 (The key binding for @code{widen} is @kbd{C-x n w}.)
6606 Narrowing is just as useful to the Lisp interpreter as to a human.
6607 Often, an Emacs Lisp function is designed to work on just part of a
6608 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6609 buffer that has been narrowed. The @code{what-line} function, for
6610 example, removes the narrowing from a buffer, if it has any narrowing
6611 and when it has finished its job, restores the narrowing to what it was.
6612 On the other hand, the @code{count-lines} function
6613 uses narrowing to restrict itself to just that portion
6614 of the buffer in which it is interested and then restores the previous
6617 @node save-restriction
6618 @section The @code{save-restriction} Special Form
6619 @findex save-restriction
6621 In Emacs Lisp, you can use the @code{save-restriction} special form to
6622 keep track of whatever narrowing is in effect, if any. When the Lisp
6623 interpreter meets with @code{save-restriction}, it executes the code
6624 in the body of the @code{save-restriction} expression, and then undoes
6625 any changes to narrowing that the code caused. If, for example, the
6626 buffer is narrowed and the code that follows @code{save-restriction}
6627 gets rid of the narrowing, @code{save-restriction} returns the buffer
6628 to its narrowed region afterwards. In the @code{what-line} command,
6629 any narrowing the buffer may have is undone by the @code{widen}
6630 command that immediately follows the @code{save-restriction} command.
6631 Any original narrowing is restored just before the completion of the
6635 The template for a @code{save-restriction} expression is simple:
6645 The body of the @code{save-restriction} is one or more expressions that
6646 will be evaluated in sequence by the Lisp interpreter.
6648 Finally, a point to note: when you use both @code{save-excursion} and
6649 @code{save-restriction}, one right after the other, you should use
6650 @code{save-excursion} outermost. If you write them in reverse order,
6651 you may fail to record narrowing in the buffer to which Emacs switches
6652 after calling @code{save-excursion}. Thus, when written together,
6653 @code{save-excursion} and @code{save-restriction} should be written
6664 In other circumstances, when not written together, the
6665 @code{save-excursion} and @code{save-restriction} special forms must
6666 be written in the order appropriate to the function.
6682 /usr/local/src/emacs/lisp/simple.el
6685 "Print the current buffer line number and narrowed line number of point."
6687 (let ((start (point-min))
6688 (n (line-number-at-pos)))
6690 (message "Line %d" n)
6694 (message "line %d (narrowed line %d)"
6695 (+ n (line-number-at-pos start) -1) n))))))
6697 (defun line-number-at-pos (&optional pos)
6698 "Return (narrowed) buffer line number at position POS.
6699 If POS is nil, use current buffer location.
6700 Counting starts at (point-min), so the value refers
6701 to the contents of the accessible portion of the buffer."
6702 (let ((opoint (or pos (point))) start)
6704 (goto-char (point-min))
6705 (setq start (point))
6708 (1+ (count-lines start (point))))))
6710 (defun count-lines (start end)
6711 "Return number of lines between START and END.
6712 This is usually the number of newlines between them,
6713 but can be one more if START is not equal to END
6714 and the greater of them is not at the start of a line."
6717 (narrow-to-region start end)
6718 (goto-char (point-min))
6719 (if (eq selective-display t)
6722 (while (re-search-forward "[\n\C-m]" nil t 40)
6723 (setq done (+ 40 done)))
6724 (while (re-search-forward "[\n\C-m]" nil t 1)
6725 (setq done (+ 1 done)))
6726 (goto-char (point-max))
6727 (if (and (/= start end)
6731 (- (buffer-size) (forward-line (buffer-size)))))))
6735 @section @code{what-line}
6737 @cindex Widening, example of
6739 The @code{what-line} command tells you the number of the line in which
6740 the cursor is located. The function illustrates the use of the
6741 @code{save-restriction} and @code{save-excursion} commands. Here is the
6742 original text of the function:
6747 "Print the current line number (in the buffer) of point."
6754 (1+ (count-lines 1 (point)))))))
6758 (In recent versions of GNU Emacs, the @code{what-line} function has
6759 been expanded to tell you your line number in a narrowed buffer as
6760 well as your line number in a widened buffer. The recent version is
6761 more complex than the version shown here. If you feel adventurous,
6762 you might want to look at it after figuring out how this version
6763 works. You will probably need to use @kbd{C-h f}
6764 (@code{describe-function}). The newer version uses a conditional to
6765 determine whether the buffer has been narrowed.
6767 (Also, it uses @code{line-number-at-pos}, which among other simple
6768 expressions, such as @code{(goto-char (point-min))}, moves point to
6769 the beginning of the current line with @code{(forward-line 0)} rather
6770 than @code{beginning-of-line}.)
6772 The @code{what-line} function as shown here has a documentation line
6773 and is interactive, as you would expect. The next two lines use the
6774 functions @code{save-restriction} and @code{widen}.
6776 The @code{save-restriction} special form notes whatever narrowing is in
6777 effect, if any, in the current buffer and restores that narrowing after
6778 the code in the body of the @code{save-restriction} has been evaluated.
6780 The @code{save-restriction} special form is followed by @code{widen}.
6781 This function undoes any narrowing the current buffer may have had
6782 when @code{what-line} was called. (The narrowing that was there is
6783 the narrowing that @code{save-restriction} remembers.) This widening
6784 makes it possible for the line counting commands to count from the
6785 beginning of the buffer. Otherwise, they would have been limited to
6786 counting within the accessible region. Any original narrowing is
6787 restored just before the completion of the function by the
6788 @code{save-restriction} special form.
6790 The call to @code{widen} is followed by @code{save-excursion}, which
6791 saves the location of the cursor (i.e., of point) and of the mark, and
6792 restores them after the code in the body of the @code{save-excursion}
6793 uses the @code{beginning-of-line} function to move point.
6795 (Note that the @code{(widen)} expression comes between the
6796 @code{save-restriction} and @code{save-excursion} special forms. When
6797 you write the two @code{save- @dots{}} expressions in sequence, write
6798 @code{save-excursion} outermost.)
6801 The last two lines of the @code{what-line} function are functions to
6802 count the number of lines in the buffer and then print the number in the
6808 (1+ (count-lines 1 (point)))))))
6812 The @code{message} function prints a one-line message at the bottom of
6813 the Emacs screen. The first argument is inside of quotation marks and
6814 is printed as a string of characters. However, it may contain a
6815 @samp{%d} expression to print a following argument. @samp{%d} prints
6816 the argument as a decimal, so the message will say something such as
6820 The number that is printed in place of the @samp{%d} is computed by the
6821 last line of the function:
6824 (1+ (count-lines 1 (point)))
6830 (defun count-lines (start end)
6831 "Return number of lines between START and END.
6832 This is usually the number of newlines between them,
6833 but can be one more if START is not equal to END
6834 and the greater of them is not at the start of a line."
6837 (narrow-to-region start end)
6838 (goto-char (point-min))
6839 (if (eq selective-display t)
6842 (while (re-search-forward "[\n\C-m]" nil t 40)
6843 (setq done (+ 40 done)))
6844 (while (re-search-forward "[\n\C-m]" nil t 1)
6845 (setq done (+ 1 done)))
6846 (goto-char (point-max))
6847 (if (and (/= start end)
6851 (- (buffer-size) (forward-line (buffer-size)))))))
6855 What this does is count the lines from the first position of the
6856 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6857 one to that number. (The @code{1+} function adds one to its
6858 argument.) We add one to it because line 2 has only one line before
6859 it, and @code{count-lines} counts only the lines @emph{before} the
6862 After @code{count-lines} has done its job, and the message has been
6863 printed in the echo area, the @code{save-excursion} restores point and
6864 mark to their original positions; and @code{save-restriction} restores
6865 the original narrowing, if any.
6867 @node narrow Exercise
6868 @section Exercise with Narrowing
6870 Write a function that will display the first 60 characters of the
6871 current buffer, even if you have narrowed the buffer to its latter
6872 half so that the first line is inaccessible. Restore point, mark, and
6873 narrowing. For this exercise, you need to use a whole potpourri of
6874 functions, including @code{save-restriction}, @code{widen},
6875 @code{goto-char}, @code{point-min}, @code{message}, and
6876 @code{buffer-substring}.
6878 @cindex Properties, mention of @code{buffer-substring-no-properties}
6879 (@code{buffer-substring} is a previously unmentioned function you will
6880 have to investigate yourself; or perhaps you will have to use
6881 @code{buffer-substring-no-properties} or
6882 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6883 properties are a feature otherwise not discussed here. @xref{Text
6884 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6887 Additionally, do you really need @code{goto-char} or @code{point-min}?
6888 Or can you write the function without them?
6890 @node car cdr & cons
6891 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6892 @findex car, @r{introduced}
6893 @findex cdr, @r{introduced}
6895 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6896 functions. The @code{cons} function is used to construct lists, and
6897 the @code{car} and @code{cdr} functions are used to take them apart.
6899 In the walk through of the @code{copy-region-as-kill} function, we
6900 will see @code{cons} as well as two variants on @code{cdr},
6901 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6904 * Strange Names:: An historical aside: why the strange names?
6905 * car & cdr:: Functions for extracting part of a list.
6906 * cons:: Constructing a list.
6907 * nthcdr:: Calling @code{cdr} repeatedly.
6909 * setcar:: Changing the first element of a list.
6910 * setcdr:: Changing the rest of a list.
6916 @unnumberedsec Strange Names
6919 The name of the @code{cons} function is not unreasonable: it is an
6920 abbreviation of the word `construct'. The origins of the names for
6921 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6922 is an acronym from the phrase `Contents of the Address part of the
6923 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6924 the phrase `Contents of the Decrement part of the Register'. These
6925 phrases refer to specific pieces of hardware on the very early
6926 computer on which the original Lisp was developed. Besides being
6927 obsolete, the phrases have been completely irrelevant for more than 25
6928 years to anyone thinking about Lisp. Nonetheless, although a few
6929 brave scholars have begun to use more reasonable names for these
6930 functions, the old terms are still in use. In particular, since the
6931 terms are used in the Emacs Lisp source code, we will use them in this
6935 @section @code{car} and @code{cdr}
6937 The @sc{car} of a list is, quite simply, the first item in the list.
6938 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
6942 If you are reading this in Info in GNU Emacs, you can see this by
6943 evaluating the following:
6946 (car '(rose violet daisy buttercup))
6950 After evaluating the expression, @code{rose} will appear in the echo
6953 Clearly, a more reasonable name for the @code{car} function would be
6954 @code{first} and this is often suggested.
6956 @code{car} does not remove the first item from the list; it only reports
6957 what it is. After @code{car} has been applied to a list, the list is
6958 still the same as it was. In the jargon, @code{car} is
6959 `non-destructive'. This feature turns out to be important.
6961 The @sc{cdr} of a list is the rest of the list, that is, the
6962 @code{cdr} function returns the part of the list that follows the
6963 first item. Thus, while the @sc{car} of the list @code{'(rose violet
6964 daisy buttercup)} is @code{rose}, the rest of the list, the value
6965 returned by the @code{cdr} function, is @code{(violet daisy
6969 You can see this by evaluating the following in the usual way:
6972 (cdr '(rose violet daisy buttercup))
6976 When you evaluate this, @code{(violet daisy buttercup)} will appear in
6979 Like @code{car}, @code{cdr} does not remove any elements from the
6980 list---it just returns a report of what the second and subsequent
6983 Incidentally, in the example, the list of flowers is quoted. If it were
6984 not, the Lisp interpreter would try to evaluate the list by calling
6985 @code{rose} as a function. In this example, we do not want to do that.
6987 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
6989 (There is a lesson here: when you name new functions, consider very
6990 carefully what you are doing, since you may be stuck with the names
6991 for far longer than you expect. The reason this document perpetuates
6992 these names is that the Emacs Lisp source code uses them, and if I did
6993 not use them, you would have a hard time reading the code; but do,
6994 please, try to avoid using these terms yourself. The people who come
6995 after you will be grateful to you.)
6997 When @code{car} and @code{cdr} are applied to a list made up of symbols,
6998 such as the list @code{(pine fir oak maple)}, the element of the list
6999 returned by the function @code{car} is the symbol @code{pine} without
7000 any parentheses around it. @code{pine} is the first element in the
7001 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7002 oak maple)}, as you can see by evaluating the following expressions in
7007 (car '(pine fir oak maple))
7009 (cdr '(pine fir oak maple))
7013 On the other hand, in a list of lists, the first element is itself a
7014 list. @code{car} returns this first element as a list. For example,
7015 the following list contains three sub-lists, a list of carnivores, a
7016 list of herbivores and a list of sea mammals:
7020 (car '((lion tiger cheetah)
7021 (gazelle antelope zebra)
7022 (whale dolphin seal)))
7027 In this example, the first element or @sc{car} of the list is the list of
7028 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7029 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7033 (cdr '((lion tiger cheetah)
7034 (gazelle antelope zebra)
7035 (whale dolphin seal)))
7039 It is worth saying again that @code{car} and @code{cdr} are
7040 non-destructive---that is, they do not modify or change lists to which
7041 they are applied. This is very important for how they are used.
7043 Also, in the first chapter, in the discussion about atoms, I said that
7044 in Lisp, ``certain kinds of atom, such as an array, can be separated
7045 into parts; but the mechanism for doing this is different from the
7046 mechanism for splitting a list. As far as Lisp is concerned, the
7047 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7048 @code{car} and @code{cdr} functions are used for splitting lists and
7049 are considered fundamental to Lisp. Since they cannot split or gain
7050 access to the parts of an array, an array is considered an atom.
7051 Conversely, the other fundamental function, @code{cons}, can put
7052 together or construct a list, but not an array. (Arrays are handled
7053 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7054 Emacs Lisp Reference Manual}.)
7057 @section @code{cons}
7058 @findex cons, @r{introduced}
7060 The @code{cons} function constructs lists; it is the inverse of
7061 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7062 a four element list from the three element list, @code{(fir oak maple)}:
7065 (cons 'pine '(fir oak maple))
7070 After evaluating this list, you will see
7073 (pine fir oak maple)
7077 appear in the echo area. @code{cons} causes the creation of a new
7078 list in which the element is followed by the elements of the original
7081 We often say that `@code{cons} puts a new element at the beginning of
7082 a list; it attaches or pushes elements onto the list', but this
7083 phrasing can be misleading, since @code{cons} does not change an
7084 existing list, but creates a new one.
7086 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7090 * length:: How to find the length of a list.
7095 @unnumberedsubsec Build a list
7098 @code{cons} must have a list to attach to.@footnote{Actually, you can
7099 @code{cons} an element to an atom to produce a dotted pair. Dotted
7100 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7101 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7102 cannot start from absolutely nothing. If you are building a list, you
7103 need to provide at least an empty list at the beginning. Here is a
7104 series of @code{cons} expressions that build up a list of flowers. If
7105 you are reading this in Info in GNU Emacs, you can evaluate each of
7106 the expressions in the usual way; the value is printed in this text
7107 after @samp{@result{}}, which you may read as `evaluates to'.
7111 (cons 'buttercup ())
7112 @result{} (buttercup)
7116 (cons 'daisy '(buttercup))
7117 @result{} (daisy buttercup)
7121 (cons 'violet '(daisy buttercup))
7122 @result{} (violet daisy buttercup)
7126 (cons 'rose '(violet daisy buttercup))
7127 @result{} (rose violet daisy buttercup)
7132 In the first example, the empty list is shown as @code{()} and a list
7133 made up of @code{buttercup} followed by the empty list is constructed.
7134 As you can see, the empty list is not shown in the list that was
7135 constructed. All that you see is @code{(buttercup)}. The empty list is
7136 not counted as an element of a list because there is nothing in an empty
7137 list. Generally speaking, an empty list is invisible.
7139 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7140 two element list by putting @code{daisy} in front of @code{buttercup};
7141 and the third example constructs a three element list by putting
7142 @code{violet} in front of @code{daisy} and @code{buttercup}.
7145 @subsection Find the Length of a List: @code{length}
7148 You can find out how many elements there are in a list by using the Lisp
7149 function @code{length}, as in the following examples:
7153 (length '(buttercup))
7158 (length '(daisy buttercup))
7163 (length (cons 'violet '(daisy buttercup)))
7169 In the third example, the @code{cons} function is used to construct a
7170 three element list which is then passed to the @code{length} function as
7174 We can also use @code{length} to count the number of elements in an
7185 As you would expect, the number of elements in an empty list is zero.
7187 An interesting experiment is to find out what happens if you try to find
7188 the length of no list at all; that is, if you try to call @code{length}
7189 without giving it an argument, not even an empty list:
7197 What you see, if you evaluate this, is the error message
7200 Lisp error: (wrong-number-of-arguments length 0)
7204 This means that the function receives the wrong number of
7205 arguments, zero, when it expects some other number of arguments. In
7206 this case, one argument is expected, the argument being a list whose
7207 length the function is measuring. (Note that @emph{one} list is
7208 @emph{one} argument, even if the list has many elements inside it.)
7210 The part of the error message that says @samp{length} is the name of
7214 @code{length} is still a subroutine, but you need C-h f to discover that.
7216 In an earlier version:
7217 This is written with a special notation, @samp{#<subr},
7218 that indicates that the function @code{length} is one of the primitive
7219 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7220 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7221 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7226 @section @code{nthcdr}
7229 The @code{nthcdr} function is associated with the @code{cdr} function.
7230 What it does is take the @sc{cdr} of a list repeatedly.
7232 If you take the @sc{cdr} of the list @code{(pine fir
7233 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7234 repeat this on what was returned, you will be returned the list
7235 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7236 list will just give you the original @sc{cdr} since the function does
7237 not change the list. You need to evaluate the @sc{cdr} of the
7238 @sc{cdr} and so on.) If you continue this, eventually you will be
7239 returned an empty list, which in this case, instead of being shown as
7240 @code{()} is shown as @code{nil}.
7243 For review, here is a series of repeated @sc{cdr}s, the text following
7244 the @samp{@result{}} shows what is returned.
7248 (cdr '(pine fir oak maple))
7249 @result{}(fir oak maple)
7253 (cdr '(fir oak maple))
7254 @result{} (oak maple)
7279 You can also do several @sc{cdr}s without printing the values in
7284 (cdr (cdr '(pine fir oak maple)))
7285 @result{} (oak maple)
7290 In this example, the Lisp interpreter evaluates the innermost list first.
7291 The innermost list is quoted, so it just passes the list as it is to the
7292 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7293 second and subsequent elements of the list to the outermost @code{cdr},
7294 which produces a list composed of the third and subsequent elements of
7295 the original list. In this example, the @code{cdr} function is repeated
7296 and returns a list that consists of the original list without its
7299 The @code{nthcdr} function does the same as repeating the call to
7300 @code{cdr}. In the following example, the argument 2 is passed to the
7301 function @code{nthcdr}, along with the list, and the value returned is
7302 the list without its first two items, which is exactly the same
7303 as repeating @code{cdr} twice on the list:
7307 (nthcdr 2 '(pine fir oak maple))
7308 @result{} (oak maple)
7313 Using the original four element list, we can see what happens when
7314 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7319 ;; @r{Leave the list as it was.}
7320 (nthcdr 0 '(pine fir oak maple))
7321 @result{} (pine fir oak maple)
7325 ;; @r{Return a copy without the first element.}
7326 (nthcdr 1 '(pine fir oak maple))
7327 @result{} (fir oak maple)
7331 ;; @r{Return a copy of the list without three elements.}
7332 (nthcdr 3 '(pine fir oak maple))
7337 ;; @r{Return a copy lacking all four elements.}
7338 (nthcdr 4 '(pine fir oak maple))
7343 ;; @r{Return a copy lacking all elements.}
7344 (nthcdr 5 '(pine fir oak maple))
7353 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7354 The @code{nth} function takes the @sc{car} of the result returned by
7355 @code{nthcdr}. It returns the Nth element of the list.
7358 Thus, if it were not defined in C for speed, the definition of
7359 @code{nth} would be:
7364 "Returns the Nth element of LIST.
7365 N counts from zero. If LIST is not that long, nil is returned."
7366 (car (nthcdr n list)))
7371 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7372 but its definition was redone in C in the 1980s.)
7374 The @code{nth} function returns a single element of a list.
7375 This can be very convenient.
7377 Note that the elements are numbered from zero, not one. That is to
7378 say, the first element of a list, its @sc{car} is the zeroth element.
7379 This is called `zero-based' counting and often bothers people who
7380 are accustomed to the first element in a list being number one, which
7388 (nth 0 '("one" "two" "three"))
7391 (nth 1 '("one" "two" "three"))
7396 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7397 @code{cdr}, does not change the original list---the function is
7398 non-destructive. This is in sharp contrast to the @code{setcar} and
7399 @code{setcdr} functions.
7402 @section @code{setcar}
7405 As you might guess from their names, the @code{setcar} and @code{setcdr}
7406 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7407 They actually change the original list, unlike @code{car} and @code{cdr}
7408 which leave the original list as it was. One way to find out how this
7409 works is to experiment. We will start with the @code{setcar} function.
7412 First, we can make a list and then set the value of a variable to the
7413 list, using the @code{setq} function. Here is a list of animals:
7416 (setq animals '(antelope giraffe lion tiger))
7420 If you are reading this in Info inside of GNU Emacs, you can evaluate
7421 this expression in the usual fashion, by positioning the cursor after
7422 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7423 as I write this. This is one of the advantages of having the
7424 interpreter built into the computing environment. Incidentally, when
7425 there is nothing on the line after the final parentheses, such as a
7426 comment, point can be on the next line. Thus, if your cursor is in
7427 the first column of the next line, you do not need to move it.
7428 Indeed, Emacs permits any amount of white space after the final
7432 When we evaluate the variable @code{animals}, we see that it is bound to
7433 the list @code{(antelope giraffe lion tiger)}:
7438 @result{} (antelope giraffe lion tiger)
7443 Put another way, the variable @code{animals} points to the list
7444 @code{(antelope giraffe lion tiger)}.
7446 Next, evaluate the function @code{setcar} while passing it two
7447 arguments, the variable @code{animals} and the quoted symbol
7448 @code{hippopotamus}; this is done by writing the three element list
7449 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7453 (setcar animals 'hippopotamus)
7458 After evaluating this expression, evaluate the variable @code{animals}
7459 again. You will see that the list of animals has changed:
7464 @result{} (hippopotamus giraffe lion tiger)
7469 The first element on the list, @code{antelope} is replaced by
7470 @code{hippopotamus}.
7472 So we can see that @code{setcar} did not add a new element to the list
7473 as @code{cons} would have; it replaced @code{antelope} with
7474 @code{hippopotamus}; it @emph{changed} the list.
7477 @section @code{setcdr}
7480 The @code{setcdr} function is similar to the @code{setcar} function,
7481 except that the function replaces the second and subsequent elements of
7482 a list rather than the first element.
7484 (To see how to change the last element of a list, look ahead to
7485 @ref{kill-new function, , The @code{kill-new} function}, which uses
7486 the @code{nthcdr} and @code{setcdr} functions.)
7489 To see how this works, set the value of the variable to a list of
7490 domesticated animals by evaluating the following expression:
7493 (setq domesticated-animals '(horse cow sheep goat))
7498 If you now evaluate the list, you will be returned the list
7499 @code{(horse cow sheep goat)}:
7503 domesticated-animals
7504 @result{} (horse cow sheep goat)
7509 Next, evaluate @code{setcdr} with two arguments, the name of the
7510 variable which has a list as its value, and the list to which the
7511 @sc{cdr} of the first list will be set;
7514 (setcdr domesticated-animals '(cat dog))
7518 If you evaluate this expression, the list @code{(cat dog)} will appear
7519 in the echo area. This is the value returned by the function. The
7520 result we are interested in is the ``side effect'', which we can see by
7521 evaluating the variable @code{domesticated-animals}:
7525 domesticated-animals
7526 @result{} (horse cat dog)
7531 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7532 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7533 @code{(cow sheep goat)} to @code{(cat dog)}.
7538 Construct a list of four birds by evaluating several expressions with
7539 @code{cons}. Find out what happens when you @code{cons} a list onto
7540 itself. Replace the first element of the list of four birds with a
7541 fish. Replace the rest of that list with a list of other fish.
7543 @node Cutting & Storing Text
7544 @chapter Cutting and Storing Text
7545 @cindex Cutting and storing text
7546 @cindex Storing and cutting text
7547 @cindex Killing text
7548 @cindex Clipping text
7549 @cindex Erasing text
7550 @cindex Deleting text
7552 Whenever you cut or clip text out of a buffer with a `kill' command in
7553 GNU Emacs, it is stored in a list and you can bring it back with a
7556 (The use of the word `kill' in Emacs for processes which specifically
7557 @emph{do not} destroy the values of the entities is an unfortunate
7558 historical accident. A much more appropriate word would be `clip' since
7559 that is what the kill commands do; they clip text out of a buffer and
7560 put it into storage from which it can be brought back. I have often
7561 been tempted to replace globally all occurrences of `kill' in the Emacs
7562 sources with `clip' and all occurrences of `killed' with `clipped'.)
7565 * Storing Text:: Text is stored in a list.
7566 * zap-to-char:: Cutting out text up to a character.
7567 * kill-region:: Cutting text out of a region.
7568 * copy-region-as-kill:: A definition for copying text.
7569 * Digression into C:: Minor note on C programming language macros.
7570 * defvar:: How to give a variable an initial value.
7571 * cons & search-fwd Review::
7572 * search Exercises::
7577 @unnumberedsec Storing Text in a List
7580 When text is cut out of a buffer, it is stored on a list. Successive
7581 pieces of text are stored on the list successively, so the list might
7585 ("a piece of text" "previous piece")
7590 The function @code{cons} can be used to create a new list from a piece
7591 of text (an `atom', to use the jargon) and an existing list, like
7596 (cons "another piece"
7597 '("a piece of text" "previous piece"))
7603 If you evaluate this expression, a list of three elements will appear in
7607 ("another piece" "a piece of text" "previous piece")
7610 With the @code{car} and @code{nthcdr} functions, you can retrieve
7611 whichever piece of text you want. For example, in the following code,
7612 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7613 and the @code{car} returns the first element of that remainder---the
7614 second element of the original list:
7618 (car (nthcdr 1 '("another piece"
7621 @result{} "a piece of text"
7625 The actual functions in Emacs are more complex than this, of course.
7626 The code for cutting and retrieving text has to be written so that
7627 Emacs can figure out which element in the list you want---the first,
7628 second, third, or whatever. In addition, when you get to the end of
7629 the list, Emacs should give you the first element of the list, rather
7630 than nothing at all.
7632 The list that holds the pieces of text is called the @dfn{kill ring}.
7633 This chapter leads up to a description of the kill ring and how it is
7634 used by first tracing how the @code{zap-to-char} function works. This
7635 function uses (or `calls') a function that invokes a function that
7636 manipulates the kill ring. Thus, before reaching the mountains, we
7637 climb the foothills.
7639 A subsequent chapter describes how text that is cut from the buffer is
7640 retrieved. @xref{Yanking, , Yanking Text Back}.
7643 @section @code{zap-to-char}
7646 @c FIXME remove obsolete stuff
7647 The @code{zap-to-char} function changed little between GNU Emacs
7648 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7649 calls another function, @code{kill-region}, which enjoyed a major
7652 The @code{kill-region} function in Emacs 19 is complex, but does not
7653 use code that is important at this time. We will skip it.
7655 The @code{kill-region} function in Emacs 22 is easier to read than the
7656 same function in Emacs 19 and introduces a very important concept,
7657 that of error handling. We will walk through the function.
7659 But first, let us look at the interactive @code{zap-to-char} function.
7662 * Complete zap-to-char:: The complete implementation.
7663 * zap-to-char interactive:: A three part interactive expression.
7664 * zap-to-char body:: A short overview.
7665 * search-forward:: How to search for a string.
7666 * progn:: The @code{progn} special form.
7667 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7671 @node Complete zap-to-char
7672 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7675 The @code{zap-to-char} function removes the text in the region between
7676 the location of the cursor (i.e., of point) up to and including the
7677 next occurrence of a specified character. The text that
7678 @code{zap-to-char} removes is put in the kill ring; and it can be
7679 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7680 the command is given an argument, it removes text through that number
7681 of occurrences. Thus, if the cursor were at the beginning of this
7682 sentence and the character were @samp{s}, @samp{Thus} would be
7683 removed. If the argument were two, @samp{Thus, if the curs} would be
7684 removed, up to and including the @samp{s} in @samp{cursor}.
7686 If the specified character is not found, @code{zap-to-char} will say
7687 ``Search failed'', tell you the character you typed, and not remove
7690 In order to determine how much text to remove, @code{zap-to-char} uses
7691 a search function. Searches are used extensively in code that
7692 manipulates text, and we will focus attention on them as well as on the
7696 @c GNU Emacs version 19
7697 (defun zap-to-char (arg char) ; version 19 implementation
7698 "Kill up to and including ARG'th occurrence of CHAR.
7699 Goes backward if ARG is negative; error if CHAR not found."
7700 (interactive "*p\ncZap to char: ")
7701 (kill-region (point)
7704 (char-to-string char) nil nil arg)
7709 Here is the complete text of the version 22 implementation of the function:
7714 (defun zap-to-char (arg char)
7715 "Kill up to and including ARG'th occurrence of CHAR.
7716 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7717 Goes backward if ARG is negative; error if CHAR not found."
7718 (interactive "p\ncZap to char: ")
7719 (if (char-table-p translation-table-for-input)
7720 (setq char (or (aref translation-table-for-input char) char)))
7721 (kill-region (point) (progn
7722 (search-forward (char-to-string char)
7728 The documentation is thorough. You do need to know the jargon meaning
7731 @node zap-to-char interactive
7732 @subsection The @code{interactive} Expression
7735 The interactive expression in the @code{zap-to-char} command looks like
7739 (interactive "p\ncZap to char: ")
7742 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7743 two different things. First, and most simply, is the @samp{p}.
7744 This part is separated from the next part by a newline, @samp{\n}.
7745 The @samp{p} means that the first argument to the function will be
7746 passed the value of a `processed prefix'. The prefix argument is
7747 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7748 the function is called interactively without a prefix, 1 is passed to
7751 The second part of @code{"p\ncZap to char:@: "} is
7752 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7753 indicates that @code{interactive} expects a prompt and that the
7754 argument will be a character. The prompt follows the @samp{c} and is
7755 the string @samp{Zap to char:@: } (with a space after the colon to
7758 What all this does is prepare the arguments to @code{zap-to-char} so they
7759 are of the right type, and give the user a prompt.
7761 In a read-only buffer, the @code{zap-to-char} function copies the text
7762 to the kill ring, but does not remove it. The echo area displays a
7763 message saying that the buffer is read-only. Also, the terminal may
7764 beep or blink at you.
7766 @node zap-to-char body
7767 @subsection The Body of @code{zap-to-char}
7769 The body of the @code{zap-to-char} function contains the code that
7770 kills (that is, removes) the text in the region from the current
7771 position of the cursor up to and including the specified character.
7773 The first part of the code looks like this:
7776 (if (char-table-p translation-table-for-input)
7777 (setq char (or (aref translation-table-for-input char) char)))
7778 (kill-region (point) (progn
7779 (search-forward (char-to-string char) nil nil arg)
7784 @code{char-table-p} is an hitherto unseen function. It determines
7785 whether its argument is a character table. When it is, it sets the
7786 character passed to @code{zap-to-char} to one of them, if that
7787 character exists, or to the character itself. (This becomes important
7788 for certain characters in non-European languages. The @code{aref}
7789 function extracts an element from an array. It is an array-specific
7790 function that is not described in this document. @xref{Arrays, ,
7791 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7794 @code{(point)} is the current position of the cursor.
7796 The next part of the code is an expression using @code{progn}. The body
7797 of the @code{progn} consists of calls to @code{search-forward} and
7800 It is easier to understand how @code{progn} works after learning about
7801 @code{search-forward}, so we will look at @code{search-forward} and
7802 then at @code{progn}.
7804 @node search-forward
7805 @subsection The @code{search-forward} Function
7806 @findex search-forward
7808 The @code{search-forward} function is used to locate the
7809 zapped-for-character in @code{zap-to-char}. If the search is
7810 successful, @code{search-forward} leaves point immediately after the
7811 last character in the target string. (In @code{zap-to-char}, the
7812 target string is just one character long. @code{zap-to-char} uses the
7813 function @code{char-to-string} to ensure that the computer treats that
7814 character as a string.) If the search is backwards,
7815 @code{search-forward} leaves point just before the first character in
7816 the target. Also, @code{search-forward} returns @code{t} for true.
7817 (Moving point is therefore a `side effect'.)
7820 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7823 (search-forward (char-to-string char) nil nil arg)
7826 The @code{search-forward} function takes four arguments:
7830 The first argument is the target, what is searched for. This must be a
7831 string, such as @samp{"z"}.
7833 As it happens, the argument passed to @code{zap-to-char} is a single
7834 character. Because of the way computers are built, the Lisp
7835 interpreter may treat a single character as being different from a
7836 string of characters. Inside the computer, a single character has a
7837 different electronic format than a string of one character. (A single
7838 character can often be recorded in the computer using exactly one
7839 byte; but a string may be longer, and the computer needs to be ready
7840 for this.) Since the @code{search-forward} function searches for a
7841 string, the character that the @code{zap-to-char} function receives as
7842 its argument must be converted inside the computer from one format to
7843 the other; otherwise the @code{search-forward} function will fail.
7844 The @code{char-to-string} function is used to make this conversion.
7847 The second argument bounds the search; it is specified as a position in
7848 the buffer. In this case, the search can go to the end of the buffer,
7849 so no bound is set and the second argument is @code{nil}.
7852 The third argument tells the function what it should do if the search
7853 fails---it can signal an error (and print a message) or it can return
7854 @code{nil}. A @code{nil} as the third argument causes the function to
7855 signal an error when the search fails.
7858 The fourth argument to @code{search-forward} is the repeat count---how
7859 many occurrences of the string to look for. This argument is optional
7860 and if the function is called without a repeat count, this argument is
7861 passed the value 1. If this argument is negative, the search goes
7866 In template form, a @code{search-forward} expression looks like this:
7870 (search-forward "@var{target-string}"
7871 @var{limit-of-search}
7872 @var{what-to-do-if-search-fails}
7877 We will look at @code{progn} next.
7880 @subsection The @code{progn} Special Form
7883 @code{progn} is a special form that causes each of its arguments to be
7884 evaluated in sequence and then returns the value of the last one. The
7885 preceding expressions are evaluated only for the side effects they
7886 perform. The values produced by them are discarded.
7889 The template for a @code{progn} expression is very simple:
7898 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7899 put point in exactly the right position; and return the location of
7900 point so that @code{kill-region} will know how far to kill to.
7902 The first argument to the @code{progn} is @code{search-forward}. When
7903 @code{search-forward} finds the string, the function leaves point
7904 immediately after the last character in the target string. (In this
7905 case the target string is just one character long.) If the search is
7906 backwards, @code{search-forward} leaves point just before the first
7907 character in the target. The movement of point is a side effect.
7909 The second and last argument to @code{progn} is the expression
7910 @code{(point)}. This expression returns the value of point, which in
7911 this case will be the location to which it has been moved by
7912 @code{search-forward}. (In the source, a line that tells the function
7913 to go to the previous character, if it is going forward, was commented
7914 out in 1999; I don't remember whether that feature or mis-feature was
7915 ever a part of the distributed source.) The value of @code{point} is
7916 returned by the @code{progn} expression and is passed to
7917 @code{kill-region} as @code{kill-region}'s second argument.
7919 @node Summing up zap-to-char
7920 @subsection Summing up @code{zap-to-char}
7922 Now that we have seen how @code{search-forward} and @code{progn} work,
7923 we can see how the @code{zap-to-char} function works as a whole.
7925 The first argument to @code{kill-region} is the position of the cursor
7926 when the @code{zap-to-char} command is given---the value of point at
7927 that time. Within the @code{progn}, the search function then moves
7928 point to just after the zapped-to-character and @code{point} returns the
7929 value of this location. The @code{kill-region} function puts together
7930 these two values of point, the first one as the beginning of the region
7931 and the second one as the end of the region, and removes the region.
7933 The @code{progn} special form is necessary because the
7934 @code{kill-region} command takes two arguments; and it would fail if
7935 @code{search-forward} and @code{point} expressions were written in
7936 sequence as two additional arguments. The @code{progn} expression is
7937 a single argument to @code{kill-region} and returns the one value that
7938 @code{kill-region} needs for its second argument.
7941 @section @code{kill-region}
7944 The @code{zap-to-char} function uses the @code{kill-region} function.
7945 This function clips text from a region and copies that text to
7946 the kill ring, from which it may be retrieved.
7951 (defun kill-region (beg end &optional yank-handler)
7952 "Kill (\"cut\") text between point and mark.
7953 This deletes the text from the buffer and saves it in the kill ring.
7954 The command \\[yank] can retrieve it from there.
7955 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
7957 If you want to append the killed region to the last killed text,
7958 use \\[append-next-kill] before \\[kill-region].
7960 If the buffer is read-only, Emacs will beep and refrain from deleting
7961 the text, but put the text in the kill ring anyway. This means that
7962 you can use the killing commands to copy text from a read-only buffer.
7964 This is the primitive for programs to kill text (as opposed to deleting it).
7965 Supply two arguments, character positions indicating the stretch of text
7967 Any command that calls this function is a \"kill command\".
7968 If the previous command was also a kill command,
7969 the text killed this time appends to the text killed last time
7970 to make one entry in the kill ring.
7972 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
7973 specifies the yank-handler text property to be set on the killed
7974 text. See `insert-for-yank'."
7975 ;; Pass point first, then mark, because the order matters
7976 ;; when calling kill-append.
7977 (interactive (list (point) (mark)))
7978 (unless (and beg end)
7979 (error "The mark is not set now, so there is no region"))
7981 (let ((string (filter-buffer-substring beg end t)))
7982 (when string ;STRING is nil if BEG = END
7983 ;; Add that string to the kill ring, one way or another.
7984 (if (eq last-command 'kill-region)
7985 (kill-append string (< end beg) yank-handler)
7986 (kill-new string nil yank-handler)))
7987 (when (or string (eq last-command 'kill-region))
7988 (setq this-command 'kill-region))
7990 ((buffer-read-only text-read-only)
7991 ;; The code above failed because the buffer, or some of the characters
7992 ;; in the region, are read-only.
7993 ;; We should beep, in case the user just isn't aware of this.
7994 ;; However, there's no harm in putting
7995 ;; the region's text in the kill ring, anyway.
7996 (copy-region-as-kill beg end)
7997 ;; Set this-command now, so it will be set even if we get an error.
7998 (setq this-command 'kill-region)
7999 ;; This should barf, if appropriate, and give us the correct error.
8000 (if kill-read-only-ok
8001 (progn (message "Read only text copied to kill ring") nil)
8002 ;; Signal an error if the buffer is read-only.
8003 (barf-if-buffer-read-only)
8004 ;; If the buffer isn't read-only, the text is.
8005 (signal 'text-read-only (list (current-buffer)))))))
8008 The Emacs 22 version of that function uses @code{condition-case} and
8009 @code{copy-region-as-kill}, both of which we will explain.
8010 @code{condition-case} is an important special form.
8012 In essence, the @code{kill-region} function calls
8013 @code{condition-case}, which takes three arguments. In this function,
8014 the first argument does nothing. The second argument contains the
8015 code that does the work when all goes well. The third argument
8016 contains the code that is called in the event of an error.
8019 * Complete kill-region:: The function definition.
8020 * condition-case:: Dealing with a problem.
8025 @node Complete kill-region
8026 @unnumberedsubsec The Complete @code{kill-region} Definition
8030 We will go through the @code{condition-case} code in a moment. First,
8031 let us look at the definition of @code{kill-region}, with comments
8037 (defun kill-region (beg end)
8038 "Kill (\"cut\") text between point and mark.
8039 This deletes the text from the buffer and saves it in the kill ring.
8040 The command \\[yank] can retrieve it from there. @dots{} "
8044 ;; @bullet{} Since order matters, pass point first.
8045 (interactive (list (point) (mark)))
8046 ;; @bullet{} And tell us if we cannot cut the text.
8047 ;; `unless' is an `if' without a then-part.
8048 (unless (and beg end)
8049 (error "The mark is not set now, so there is no region"))
8053 ;; @bullet{} `condition-case' takes three arguments.
8054 ;; If the first argument is nil, as it is here,
8055 ;; information about the error signal is not
8056 ;; stored for use by another function.
8061 ;; @bullet{} The second argument to `condition-case' tells the
8062 ;; Lisp interpreter what to do when all goes well.
8066 ;; It starts with a `let' function that extracts the string
8067 ;; and tests whether it exists. If so (that is what the
8068 ;; `when' checks), it calls an `if' function that determines
8069 ;; whether the previous command was another call to
8070 ;; `kill-region'; if it was, then the new text is appended to
8071 ;; the previous text; if not, then a different function,
8072 ;; `kill-new', is called.
8076 ;; The `kill-append' function concatenates the new string and
8077 ;; the old. The `kill-new' function inserts text into a new
8078 ;; item in the kill ring.
8082 ;; `when' is an `if' without an else-part. The second `when'
8083 ;; again checks whether the current string exists; in
8084 ;; addition, it checks whether the previous command was
8085 ;; another call to `kill-region'. If one or the other
8086 ;; condition is true, then it sets the current command to
8087 ;; be `kill-region'.
8090 (let ((string (filter-buffer-substring beg end t)))
8091 (when string ;STRING is nil if BEG = END
8092 ;; Add that string to the kill ring, one way or another.
8093 (if (eq last-command 'kill-region)
8096 ;; @minus{} `yank-handler' is an optional argument to
8097 ;; `kill-region' that tells the `kill-append' and
8098 ;; `kill-new' functions how deal with properties
8099 ;; added to the text, such as `bold' or `italics'.
8100 (kill-append string (< end beg) yank-handler)
8101 (kill-new string nil yank-handler)))
8102 (when (or string (eq last-command 'kill-region))
8103 (setq this-command 'kill-region))
8108 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8109 ;; what to do with an error.
8112 ;; The third argument has a conditions part and a body part.
8113 ;; If the conditions are met (in this case,
8114 ;; if text or buffer are read-only)
8115 ;; then the body is executed.
8118 ;; The first part of the third argument is the following:
8119 ((buffer-read-only text-read-only) ;; the if-part
8120 ;; @dots{} the then-part
8121 (copy-region-as-kill beg end)
8124 ;; Next, also as part of the then-part, set this-command, so
8125 ;; it will be set in an error
8126 (setq this-command 'kill-region)
8127 ;; Finally, in the then-part, send a message if you may copy
8128 ;; the text to the kill ring without signaling an error, but
8129 ;; don't if you may not.
8132 (if kill-read-only-ok
8133 (progn (message "Read only text copied to kill ring") nil)
8134 (barf-if-buffer-read-only)
8135 ;; If the buffer isn't read-only, the text is.
8136 (signal 'text-read-only (list (current-buffer)))))
8144 (defun kill-region (beg end)
8145 "Kill between point and mark.
8146 The text is deleted but saved in the kill ring."
8151 ;; 1. `condition-case' takes three arguments.
8152 ;; If the first argument is nil, as it is here,
8153 ;; information about the error signal is not
8154 ;; stored for use by another function.
8159 ;; 2. The second argument to `condition-case'
8160 ;; tells the Lisp interpreter what to do when all goes well.
8164 ;; The `delete-and-extract-region' function usually does the
8165 ;; work. If the beginning and ending of the region are both
8166 ;; the same, then the variable `string' will be empty, or nil
8167 (let ((string (delete-and-extract-region beg end)))
8171 ;; `when' is an `if' clause that cannot take an `else-part'.
8172 ;; Emacs normally sets the value of `last-command' to the
8173 ;; previous command.
8176 ;; `kill-append' concatenates the new string and the old.
8177 ;; `kill-new' inserts text into a new item in the kill ring.
8179 (if (eq last-command 'kill-region)
8180 ;; if true, prepend string
8181 (kill-append string (< end beg))
8183 (setq this-command 'kill-region))
8187 ;; 3. The third argument to `condition-case' tells the interpreter
8188 ;; what to do with an error.
8191 ;; The third argument has a conditions part and a body part.
8192 ;; If the conditions are met (in this case,
8193 ;; if text or buffer are read-only)
8194 ;; then the body is executed.
8197 ((buffer-read-only text-read-only) ;; this is the if-part
8199 (copy-region-as-kill beg end)
8202 (if kill-read-only-ok ;; usually this variable is nil
8203 (message "Read only text copied to kill ring")
8204 ;; or else, signal an error if the buffer is read-only;
8205 (barf-if-buffer-read-only)
8206 ;; and, in any case, signal that the text is read-only.
8207 (signal 'text-read-only (list (current-buffer)))))))
8212 @node condition-case
8213 @subsection @code{condition-case}
8214 @findex condition-case
8216 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8217 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8218 expression, it provides you with help; in the jargon, this is called
8219 ``signaling an error''. Usually, the computer stops the program and
8220 shows you a message.
8222 However, some programs undertake complicated actions. They should not
8223 simply stop on an error. In the @code{kill-region} function, the most
8224 likely error is that you will try to kill text that is read-only and
8225 cannot be removed. So the @code{kill-region} function contains code
8226 to handle this circumstance. This code, which makes up the body of
8227 the @code{kill-region} function, is inside of a @code{condition-case}
8231 The template for @code{condition-case} looks like this:
8238 @var{error-handler}@dots{})
8242 The second argument, @var{bodyform}, is straightforward. The
8243 @code{condition-case} special form causes the Lisp interpreter to
8244 evaluate the code in @var{bodyform}. If no error occurs, the special
8245 form returns the code's value and produces the side-effects, if any.
8247 In short, the @var{bodyform} part of a @code{condition-case}
8248 expression determines what should happen when everything works
8251 However, if an error occurs, among its other actions, the function
8252 generating the error signal will define one or more error condition
8255 An error handler is the third argument to @code{condition case}.
8256 An error handler has two parts, a @var{condition-name} and a
8257 @var{body}. If the @var{condition-name} part of an error handler
8258 matches a condition name generated by an error, then the @var{body}
8259 part of the error handler is run.
8261 As you will expect, the @var{condition-name} part of an error handler
8262 may be either a single condition name or a list of condition names.
8264 Also, a complete @code{condition-case} expression may contain more
8265 than one error handler. When an error occurs, the first applicable
8268 Lastly, the first argument to the @code{condition-case} expression,
8269 the @var{var} argument, is sometimes bound to a variable that
8270 contains information about the error. However, if that argument is
8271 nil, as is the case in @code{kill-region}, that information is
8275 In brief, in the @code{kill-region} function, the code
8276 @code{condition-case} works like this:
8280 @var{If no errors}, @var{run only this code}
8281 @var{but}, @var{if errors}, @var{run this other code}.
8288 copy-region-as-kill is short, 12 lines, and uses
8289 filter-buffer-substring, which is longer, 39 lines
8290 and has delete-and-extract-region in it.
8291 delete-and-extract-region is written in C.
8293 see Initializing a Variable with @code{defvar}
8295 Initializing a Variable with @code{defvar} includes line 8350
8299 @subsection Lisp macro
8303 The part of the @code{condition-case} expression that is evaluated in
8304 the expectation that all goes well has a @code{when}. The code uses
8305 @code{when} to determine whether the @code{string} variable points to
8308 A @code{when} expression is simply a programmers' convenience. It is
8309 an @code{if} without the possibility of an else clause. In your mind,
8310 you can replace @code{when} with @code{if} and understand what goes
8311 on. That is what the Lisp interpreter does.
8313 Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro}
8314 enables you to define new control constructs and other language
8315 features. It tells the interpreter how to compute another Lisp
8316 expression which will in turn compute the value. In this case, the
8317 `other expression' is an @code{if} expression.
8319 The @code{kill-region} function definition also has an @code{unless}
8320 macro; it is the converse of @code{when}. The @code{unless} macro is
8321 an @code{if} without a then clause
8323 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8324 Emacs Lisp Reference Manual}. The C programming language also
8325 provides macros. These are different, but also useful.
8328 We will briefly look at C macros in
8329 @ref{Digression into C}.
8333 Regarding the @code{when} macro, in the @code{condition-case}
8334 expression, when the string has content, then another conditional
8335 expression is executed. This is an @code{if} with both a then-part
8340 (if (eq last-command 'kill-region)
8341 (kill-append string (< end beg) yank-handler)
8342 (kill-new string nil yank-handler))
8346 The then-part is evaluated if the previous command was another call to
8347 @code{kill-region}; if not, the else-part is evaluated.
8349 @code{yank-handler} is an optional argument to @code{kill-region} that
8350 tells the @code{kill-append} and @code{kill-new} functions how deal
8351 with properties added to the text, such as `bold' or `italics'.
8353 @code{last-command} is a variable that comes with Emacs that we have
8354 not seen before. Normally, whenever a function is executed, Emacs
8355 sets the value of @code{last-command} to the previous command.
8358 In this segment of the definition, the @code{if} expression checks
8359 whether the previous command was @code{kill-region}. If it was,
8362 (kill-append string (< end beg) yank-handler)
8366 concatenates a copy of the newly clipped text to the just previously
8367 clipped text in the kill ring.
8369 @node copy-region-as-kill
8370 @section @code{copy-region-as-kill}
8371 @findex copy-region-as-kill
8374 The @code{copy-region-as-kill} function copies a region of text from a
8375 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8376 in the @code{kill-ring}.
8378 If you call @code{copy-region-as-kill} immediately after a
8379 @code{kill-region} command, Emacs appends the newly copied text to the
8380 previously copied text. This means that if you yank back the text, you
8381 get it all, from both this and the previous operation. On the other
8382 hand, if some other command precedes the @code{copy-region-as-kill},
8383 the function copies the text into a separate entry in the kill ring.
8386 * Complete copy-region-as-kill:: The complete function definition.
8387 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8391 @node Complete copy-region-as-kill
8392 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8396 Here is the complete text of the version 22 @code{copy-region-as-kill}
8401 (defun copy-region-as-kill (beg end)
8402 "Save the region as if killed, but don't kill it.
8403 In Transient Mark mode, deactivate the mark.
8404 If `interprogram-cut-function' is non-nil, also save the text for a window
8405 system cut and paste."
8409 (if (eq last-command 'kill-region)
8410 (kill-append (filter-buffer-substring beg end) (< end beg))
8411 (kill-new (filter-buffer-substring beg end)))
8414 (if transient-mark-mode
8415 (setq deactivate-mark t))
8421 As usual, this function can be divided into its component parts:
8425 (defun copy-region-as-kill (@var{argument-list})
8426 "@var{documentation}@dots{}"
8432 The arguments are @code{beg} and @code{end} and the function is
8433 interactive with @code{"r"}, so the two arguments must refer to the
8434 beginning and end of the region. If you have been reading though this
8435 document from the beginning, understanding these parts of a function is
8436 almost becoming routine.
8438 The documentation is somewhat confusing unless you remember that the
8439 word `kill' has a meaning different from usual. The `Transient Mark'
8440 and @code{interprogram-cut-function} comments explain certain
8443 After you once set a mark, a buffer always contains a region. If you
8444 wish, you can use Transient Mark mode to highlight the region
8445 temporarily. (No one wants to highlight the region all the time, so
8446 Transient Mark mode highlights it only at appropriate times. Many
8447 people turn off Transient Mark mode, so the region is never
8450 Also, a windowing system allows you to copy, cut, and paste among
8451 different programs. In the X windowing system, for example, the
8452 @code{interprogram-cut-function} function is @code{x-select-text},
8453 which works with the windowing system's equivalent of the Emacs kill
8456 The body of the @code{copy-region-as-kill} function starts with an
8457 @code{if} clause. What this clause does is distinguish between two
8458 different situations: whether or not this command is executed
8459 immediately after a previous @code{kill-region} command. In the first
8460 case, the new region is appended to the previously copied text.
8461 Otherwise, it is inserted into the beginning of the kill ring as a
8462 separate piece of text from the previous piece.
8464 The last two lines of the function prevent the region from lighting up
8465 if Transient Mark mode is turned on.
8467 The body of @code{copy-region-as-kill} merits discussion in detail.
8469 @node copy-region-as-kill body
8470 @subsection The Body of @code{copy-region-as-kill}
8472 The @code{copy-region-as-kill} function works in much the same way as
8473 the @code{kill-region} function. Both are written so that two or more
8474 kills in a row combine their text into a single entry. If you yank
8475 back the text from the kill ring, you get it all in one piece.
8476 Moreover, kills that kill forward from the current position of the
8477 cursor are added to the end of the previously copied text and commands
8478 that copy text backwards add it to the beginning of the previously
8479 copied text. This way, the words in the text stay in the proper
8482 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8483 use of the @code{last-command} variable that keeps track of the
8484 previous Emacs command.
8487 * last-command & this-command::
8488 * kill-append function::
8489 * kill-new function::
8493 @node last-command & this-command
8494 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8497 Normally, whenever a function is executed, Emacs sets the value of
8498 @code{this-command} to the function being executed (which in this case
8499 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8500 the value of @code{last-command} to the previous value of
8501 @code{this-command}.
8503 In the first part of the body of the @code{copy-region-as-kill}
8504 function, an @code{if} expression determines whether the value of
8505 @code{last-command} is @code{kill-region}. If so, the then-part of
8506 the @code{if} expression is evaluated; it uses the @code{kill-append}
8507 function to concatenate the text copied at this call to the function
8508 with the text already in the first element (the @sc{car}) of the kill
8509 ring. On the other hand, if the value of @code{last-command} is not
8510 @code{kill-region}, then the @code{copy-region-as-kill} function
8511 attaches a new element to the kill ring using the @code{kill-new}
8515 The @code{if} expression reads as follows; it uses @code{eq}:
8519 (if (eq last-command 'kill-region)
8521 (kill-append (filter-buffer-substring beg end) (< end beg))
8523 (kill-new (filter-buffer-substring beg end)))
8527 @findex filter-buffer-substring
8528 (The @code{filter-buffer-substring} function returns a filtered
8529 substring of the buffer, if any. Optionally---the arguments are not
8530 here, so neither is done---the function may delete the initial text or
8531 return the text without its properties; this function is a replacement
8532 for the older @code{buffer-substring} function, which came before text
8533 properties were implemented.)
8535 @findex eq @r{(example of use)}
8537 The @code{eq} function tests whether its first argument is the same Lisp
8538 object as its second argument. The @code{eq} function is similar to the
8539 @code{equal} function in that it is used to test for equality, but
8540 differs in that it determines whether two representations are actually
8541 the same object inside the computer, but with different names.
8542 @code{equal} determines whether the structure and contents of two
8543 expressions are the same.
8545 If the previous command was @code{kill-region}, then the Emacs Lisp
8546 interpreter calls the @code{kill-append} function
8548 @node kill-append function
8549 @unnumberedsubsubsec The @code{kill-append} function
8553 The @code{kill-append} function looks like this:
8558 (defun kill-append (string before-p &optional yank-handler)
8559 "Append STRING to the end of the latest kill in the kill ring.
8560 If BEFORE-P is non-nil, prepend STRING to the kill.
8562 (let* ((cur (car kill-ring)))
8563 (kill-new (if before-p (concat string cur) (concat cur string))
8564 (or (= (length cur) 0)
8566 (get-text-property 0 'yank-handler cur)))
8573 (defun kill-append (string before-p)
8574 "Append STRING to the end of the latest kill in the kill ring.
8575 If BEFORE-P is non-nil, prepend STRING to the kill.
8576 If `interprogram-cut-function' is set, pass the resulting kill to
8578 (kill-new (if before-p
8579 (concat string (car kill-ring))
8580 (concat (car kill-ring) string))
8585 The @code{kill-append} function is fairly straightforward. It uses
8586 the @code{kill-new} function, which we will discuss in more detail in
8589 (Also, the function provides an optional argument called
8590 @code{yank-handler}; when invoked, this argument tells the function
8591 how to deal with properties added to the text, such as `bold' or
8594 @c !!! bug in GNU Emacs 22 version of kill-append ?
8595 It has a @code{let*} function to set the value of the first element of
8596 the kill ring to @code{cur}. (I do not know why the function does not
8597 use @code{let} instead; only one value is set in the expression.
8598 Perhaps this is a bug that produces no problems?)
8600 Consider the conditional that is one of the two arguments to
8601 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8602 the @sc{car} of the kill ring. Whether it prepends or appends the
8603 text depends on the results of an @code{if} expression:
8607 (if before-p ; @r{if-part}
8608 (concat string cur) ; @r{then-part}
8609 (concat cur string)) ; @r{else-part}
8614 If the region being killed is before the region that was killed in the
8615 last command, then it should be prepended before the material that was
8616 saved in the previous kill; and conversely, if the killed text follows
8617 what was just killed, it should be appended after the previous text.
8618 The @code{if} expression depends on the predicate @code{before-p} to
8619 decide whether the newly saved text should be put before or after the
8620 previously saved text.
8622 The symbol @code{before-p} is the name of one of the arguments to
8623 @code{kill-append}. When the @code{kill-append} function is
8624 evaluated, it is bound to the value returned by evaluating the actual
8625 argument. In this case, this is the expression @code{(< end beg)}.
8626 This expression does not directly determine whether the killed text in
8627 this command is located before or after the kill text of the last
8628 command; what it does is determine whether the value of the variable
8629 @code{end} is less than the value of the variable @code{beg}. If it
8630 is, it means that the user is most likely heading towards the
8631 beginning of the buffer. Also, the result of evaluating the predicate
8632 expression, @code{(< end beg)}, will be true and the text will be
8633 prepended before the previous text. On the other hand, if the value of
8634 the variable @code{end} is greater than the value of the variable
8635 @code{beg}, the text will be appended after the previous text.
8638 When the newly saved text will be prepended, then the string with the new
8639 text will be concatenated before the old text:
8647 But if the text will be appended, it will be concatenated
8651 (concat cur string))
8654 To understand how this works, we first need to review the
8655 @code{concat} function. The @code{concat} function links together or
8656 unites two strings of text. The result is a string. For example:
8660 (concat "abc" "def")
8666 (car '("first element" "second element")))
8667 @result{} "new first element"
8670 '("first element" "second element")) " modified")
8671 @result{} "first element modified"
8675 We can now make sense of @code{kill-append}: it modifies the contents
8676 of the kill ring. The kill ring is a list, each element of which is
8677 saved text. The @code{kill-append} function uses the @code{kill-new}
8678 function which in turn uses the @code{setcar} function.
8680 @node kill-new function
8681 @unnumberedsubsubsec The @code{kill-new} function
8684 @c in GNU Emacs 22, additional documentation to kill-new:
8686 Optional third arguments YANK-HANDLER controls how the STRING is later
8687 inserted into a buffer; see `insert-for-yank' for details.
8688 When a yank handler is specified, STRING must be non-empty (the yank
8689 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8691 When the yank handler has a non-nil PARAM element, the original STRING
8692 argument is not used by `insert-for-yank'. However, since Lisp code
8693 may access and use elements from the kill ring directly, the STRING
8694 argument should still be a \"useful\" string for such uses."
8697 The @code{kill-new} function looks like this:
8701 (defun kill-new (string &optional replace yank-handler)
8702 "Make STRING the latest kill in the kill ring.
8703 Set `kill-ring-yank-pointer' to point to it.
8705 If `interprogram-cut-function' is non-nil, apply it to STRING.
8706 Optional second argument REPLACE non-nil means that STRING will replace
8707 the front of the kill ring, rather than being added to the list.
8711 (if (> (length string) 0)
8713 (put-text-property 0 (length string)
8714 'yank-handler yank-handler string))
8716 (signal 'args-out-of-range
8717 (list string "yank-handler specified for empty string"))))
8720 (if (fboundp 'menu-bar-update-yank-menu)
8721 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8724 (if (and replace kill-ring)
8725 (setcar kill-ring string)
8726 (push string kill-ring)
8727 (if (> (length kill-ring) kill-ring-max)
8728 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8731 (setq kill-ring-yank-pointer kill-ring)
8732 (if interprogram-cut-function
8733 (funcall interprogram-cut-function string (not replace))))
8738 (defun kill-new (string &optional replace)
8739 "Make STRING the latest kill in the kill ring.
8740 Set the kill-ring-yank pointer to point to it.
8741 If `interprogram-cut-function' is non-nil, apply it to STRING.
8742 Optional second argument REPLACE non-nil means that STRING will replace
8743 the front of the kill ring, rather than being added to the list."
8744 (and (fboundp 'menu-bar-update-yank-menu)
8745 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8746 (if (and replace kill-ring)
8747 (setcar kill-ring string)
8748 (setq kill-ring (cons string kill-ring))
8749 (if (> (length kill-ring) kill-ring-max)
8750 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8751 (setq kill-ring-yank-pointer kill-ring)
8752 (if interprogram-cut-function
8753 (funcall interprogram-cut-function string (not replace))))
8756 (Notice that the function is not interactive.)
8758 As usual, we can look at this function in parts.
8760 The function definition has an optional @code{yank-handler} argument,
8761 which when invoked tells the function how to deal with properties
8762 added to the text, such as `bold' or `italics'. We will skip that.
8765 The first line of the documentation makes sense:
8768 Make STRING the latest kill in the kill ring.
8772 Let's skip over the rest of the documentation for the moment.
8775 Also, let's skip over the initial @code{if} expression and those lines
8776 of code involving @code{menu-bar-update-yank-menu}. We will explain
8780 The critical lines are these:
8784 (if (and replace kill-ring)
8786 (setcar kill-ring string)
8790 (push string kill-ring)
8793 (setq kill-ring (cons string kill-ring))
8794 (if (> (length kill-ring) kill-ring-max)
8795 ;; @r{avoid overly long kill ring}
8796 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8799 (setq kill-ring-yank-pointer kill-ring)
8800 (if interprogram-cut-function
8801 (funcall interprogram-cut-function string (not replace))))
8805 The conditional test is @w{@code{(and replace kill-ring)}}.
8806 This will be true when two conditions are met: the kill ring has
8807 something in it, and the @code{replace} variable is true.
8810 When the @code{kill-append} function sets @code{replace} to be true
8811 and when the kill ring has at least one item in it, the @code{setcar}
8812 expression is executed:
8815 (setcar kill-ring string)
8818 The @code{setcar} function actually changes the first element of the
8819 @code{kill-ring} list to the value of @code{string}. It replaces the
8823 On the other hand, if the kill ring is empty, or replace is false, the
8824 else-part of the condition is executed:
8827 (push string kill-ring)
8832 @code{push} puts its first argument onto the second. It is similar to
8836 (setq kill-ring (cons string kill-ring))
8844 (add-to-list kill-ring string)
8848 When it is false, the expression first constructs a new version of the
8849 kill ring by prepending @code{string} to the existing kill ring as a
8850 new element (that is what the @code{push} does). Then it executes a
8851 second @code{if} clause. This second @code{if} clause keeps the kill
8852 ring from growing too long.
8854 Let's look at these two expressions in order.
8856 The @code{push} line of the else-part sets the new value of the kill
8857 ring to what results from adding the string being killed to the old
8860 We can see how this works with an example.
8866 (setq example-list '("here is a clause" "another clause"))
8871 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8872 @code{example-list} and see what it returns:
8877 @result{} ("here is a clause" "another clause")
8883 Now, we can add a new element on to this list by evaluating the
8884 following expression:
8885 @findex push, @r{example}
8888 (push "a third clause" example-list)
8893 When we evaluate @code{example-list}, we find its value is:
8898 @result{} ("a third clause" "here is a clause" "another clause")
8903 Thus, the third clause is added to the list by @code{push}.
8906 Now for the second part of the @code{if} clause. This expression
8907 keeps the kill ring from growing too long. It looks like this:
8911 (if (> (length kill-ring) kill-ring-max)
8912 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8916 The code checks whether the length of the kill ring is greater than
8917 the maximum permitted length. This is the value of
8918 @code{kill-ring-max} (which is 60, by default). If the length of the
8919 kill ring is too long, then this code sets the last element of the
8920 kill ring to @code{nil}. It does this by using two functions,
8921 @code{nthcdr} and @code{setcdr}.
8923 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
8924 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
8925 @sc{car} of a list. In this case, however, @code{setcdr} will not be
8926 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
8927 function is used to cause it to set the @sc{cdr} of the next to last
8928 element of the kill ring---this means that since the @sc{cdr} of the
8929 next to last element is the last element of the kill ring, it will set
8930 the last element of the kill ring.
8932 @findex nthcdr, @r{example}
8933 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
8934 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
8935 @dots{} It does this @var{N} times and returns the results.
8936 (@xref{nthcdr, , @code{nthcdr}}.)
8938 @findex setcdr, @r{example}
8939 Thus, if we had a four element list that was supposed to be three
8940 elements long, we could set the @sc{cdr} of the next to last element
8941 to @code{nil}, and thereby shorten the list. (If you set the last
8942 element to some other value than @code{nil}, which you could do, then
8943 you would not have shortened the list. @xref{setcdr, ,
8946 You can see shortening by evaluating the following three expressions
8947 in turn. First set the value of @code{trees} to @code{(maple oak pine
8948 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
8949 and then find the value of @code{trees}:
8953 (setq trees '(maple oak pine birch))
8954 @result{} (maple oak pine birch)
8958 (setcdr (nthcdr 2 trees) nil)
8962 @result{} (maple oak pine)
8967 (The value returned by the @code{setcdr} expression is @code{nil} since
8968 that is what the @sc{cdr} is set to.)
8970 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
8971 @sc{cdr} a number of times that is one less than the maximum permitted
8972 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
8973 element (which will be the rest of the elements in the kill ring) to
8974 @code{nil}. This prevents the kill ring from growing too long.
8977 The next to last expression in the @code{kill-new} function is
8980 (setq kill-ring-yank-pointer kill-ring)
8983 The @code{kill-ring-yank-pointer} is a global variable that is set to be
8984 the @code{kill-ring}.
8986 Even though the @code{kill-ring-yank-pointer} is called a
8987 @samp{pointer}, it is a variable just like the kill ring. However, the
8988 name has been chosen to help humans understand how the variable is used.
8991 Now, to return to an early expression in the body of the function:
8995 (if (fboundp 'menu-bar-update-yank-menu)
8996 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9001 It starts with an @code{if} expression
9003 In this case, the expression tests first to see whether
9004 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9005 calls it. The @code{fboundp} function returns true if the symbol it
9006 is testing has a function definition that `is not void'. If the
9007 symbol's function definition were void, we would receive an error
9008 message, as we did when we created errors intentionally (@pxref{Making
9009 Errors, , Generate an Error Message}).
9012 The then-part contains an expression whose first element is the
9013 function @code{and}.
9016 The @code{and} special form evaluates each of its arguments until one
9017 of the arguments returns a value of @code{nil}, in which case the
9018 @code{and} expression returns @code{nil}; however, if none of the
9019 arguments returns a value of @code{nil}, the value resulting from
9020 evaluating the last argument is returned. (Since such a value is not
9021 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9022 @code{and} expression returns a true value only if all its arguments
9023 are true. (@xref{Second Buffer Related Review}.)
9025 The expression determines whether the second argument to
9026 @code{menu-bar-update-yank-menu} is true or not.
9028 ;; If we're supposed to be extending an existing string, and that
9029 ;; string really is at the front of the menu, then update it in place.
9032 @code{menu-bar-update-yank-menu} is one of the functions that make it
9033 possible to use the `Select and Paste' menu in the Edit item of a menu
9034 bar; using a mouse, you can look at the various pieces of text you
9035 have saved and select one piece to paste.
9037 The last expression in the @code{kill-new} function adds the newly
9038 copied string to whatever facility exists for copying and pasting
9039 among different programs running in a windowing system. In the X
9040 Windowing system, for example, the @code{x-select-text} function takes
9041 the string and stores it in memory operated by X@. You can paste the
9042 string in another program, such as an Xterm.
9045 The expression looks like this:
9049 (if interprogram-cut-function
9050 (funcall interprogram-cut-function string (not replace))))
9054 If an @code{interprogram-cut-function} exists, then Emacs executes
9055 @code{funcall}, which in turn calls its first argument as a function
9056 and passes the remaining arguments to it. (Incidentally, as far as I
9057 can see, this @code{if} expression could be replaced by an @code{and}
9058 expression similar to the one in the first part of the function.)
9060 We are not going to discuss windowing systems and other programs
9061 further, but merely note that this is a mechanism that enables GNU
9062 Emacs to work easily and well with other programs.
9064 This code for placing text in the kill ring, either concatenated with
9065 an existing element or as a new element, leads us to the code for
9066 bringing back text that has been cut out of the buffer---the yank
9067 commands. However, before discussing the yank commands, it is better
9068 to learn how lists are implemented in a computer. This will make
9069 clear such mysteries as the use of the term `pointer'. But before
9070 that, we will digress into C.
9073 @c is this true in Emacs 22? Does not seems to be
9075 (If the @w{@code{(< end beg))}}
9076 expression is true, @code{kill-append} prepends the string to the just
9077 previously clipped text. For a detailed discussion, see
9078 @ref{kill-append function, , The @code{kill-append} function}.)
9080 If you then yank back the text, i.e., `paste' it, you get both
9081 pieces of text at once. That way, if you delete two words in a row,
9082 and then yank them back, you get both words, in their proper order,
9083 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9086 On the other hand, if the previous command is not @code{kill-region},
9087 then the @code{kill-new} function is called, which adds the text to
9088 the kill ring as the latest item, and sets the
9089 @code{kill-ring-yank-pointer} variable to point to it.
9093 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9094 @c use the delete-and-extract-region function
9096 2006 Oct 26, the Digression into C is now OK but should come after
9097 copy-region-as-kill and filter-buffer-substring
9101 copy-region-as-kill is short, 12 lines, and uses
9102 filter-buffer-substring, which is longer, 39 lines
9103 and has delete-and-extract-region in it.
9104 delete-and-extract-region is written in C.
9106 see Initializing a Variable with @code{defvar}
9109 @node Digression into C
9110 @section Digression into C
9111 @findex delete-and-extract-region
9112 @cindex C, a digression into
9113 @cindex Digression into C
9115 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9116 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9117 function, which in turn uses the @code{delete-and-extract-region}
9118 function. It removes the contents of a region and you cannot get them
9121 Unlike the other code discussed here, the
9122 @code{delete-and-extract-region} function is not written in Emacs
9123 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9124 system. Since it is very simple, I will digress briefly from Lisp and
9127 @c GNU Emacs 24 in src/editfns.c
9128 @c the DEFUN for delete-and-extract-region
9131 Like many of the other Emacs primitives,
9132 @code{delete-and-extract-region} is written as an instance of a C
9133 macro, a macro being a template for code. The complete macro looks
9138 DEFUN ("delete-and-extract-region", Fdelete_and_extract_region,
9139 Sdelete_and_extract_region, 2, 2, 0,
9140 doc: /* Delete the text between START and END and return it. */)
9141 (Lisp_Object start, Lisp_Object end)
9143 validate_region (&start, &end);
9144 if (XINT (start) == XINT (end))
9145 return empty_unibyte_string;
9146 return del_range_1 (XINT (start), XINT (end), 1, 1);
9151 Without going into the details of the macro writing process, let me
9152 point out that this macro starts with the word @code{DEFUN}. The word
9153 @code{DEFUN} was chosen since the code serves the same purpose as
9154 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9155 @file{emacs/src/lisp.h}.)
9157 The word @code{DEFUN} is followed by seven parts inside of
9162 The first part is the name given to the function in Lisp,
9163 @code{delete-and-extract-region}.
9166 The second part is the name of the function in C,
9167 @code{Fdelete_and_extract_region}. By convention, it starts with
9168 @samp{F}. Since C does not use hyphens in names, underscores are used
9172 The third part is the name for the C constant structure that records
9173 information on this function for internal use. It is the name of the
9174 function in C but begins with an @samp{S} instead of an @samp{F}.
9177 The fourth and fifth parts specify the minimum and maximum number of
9178 arguments the function can have. This function demands exactly 2
9182 The sixth part is nearly like the argument that follows the
9183 @code{interactive} declaration in a function written in Lisp: a letter
9184 followed, perhaps, by a prompt. The only difference from the Lisp is
9185 when the macro is called with no arguments. Then you write a @code{0}
9186 (which is a `null string'), as in this macro.
9188 If you were to specify arguments, you would place them between
9189 quotation marks. The C macro for @code{goto-char} includes
9190 @code{"NGoto char: "} in this position to indicate that the function
9191 expects a raw prefix, in this case, a numerical location in a buffer,
9192 and provides a prompt.
9195 The seventh part is a documentation string, just like the one for a
9196 function written in Emacs Lisp. This is written as a C comment. (When
9197 you build Emacs, the program @command{lib-src/make-docfile} extracts
9198 these comments and uses them to make the ``real'' documentation.)
9202 In a C macro, the formal parameters come next, with a statement of
9203 what kind of object they are, followed by what might be called the `body'
9204 of the macro. For @code{delete-and-extract-region} the `body'
9205 consists of the following four lines:
9209 validate_region (&start, &end);
9210 if (XINT (start) == XINT (end))
9211 return empty_unibyte_string;
9212 return del_range_1 (XINT (start), XINT (end), 1, 1);
9216 The @code{validate_region} function checks whether the values
9217 passed as the beginning and end of the region are the proper type and
9218 are within range. If the beginning and end positions are the same,
9219 then return an empty string.
9221 The @code{del_range_1} function actually deletes the text. It is a
9222 complex function we will not look into. It updates the buffer and
9223 does other things. However, it is worth looking at the two arguments
9224 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9225 @w{@code{XINT (end)}}.
9227 As far as the C language is concerned, @code{start} and @code{end} are
9228 two integers that mark the beginning and end of the region to be
9229 deleted@footnote{More precisely, and requiring more expert knowledge
9230 to understand, the two integers are of type `Lisp_Object', which can
9231 also be a C union instead of an integer type.}.
9233 In early versions of Emacs, these two numbers were thirty-two bits
9234 long, but the code is slowly being generalized to handle other
9235 lengths. Three of the available bits are used to specify the type of
9236 information; the remaining bits are used as `content'.
9238 @samp{XINT} is a C macro that extracts the relevant number from the
9239 longer collection of bits; the three other bits are discarded.
9242 The command in @code{delete-and-extract-region} looks like this:
9245 del_range_1 (XINT (start), XINT (end), 1, 1);
9249 It deletes the region between the beginning position, @code{start},
9250 and the ending position, @code{end}.
9252 From the point of view of the person writing Lisp, Emacs is all very
9253 simple; but hidden underneath is a great deal of complexity to make it
9257 @section Initializing a Variable with @code{defvar}
9259 @cindex Initializing a variable
9260 @cindex Variable initialization
9265 copy-region-as-kill is short, 12 lines, and uses
9266 filter-buffer-substring, which is longer, 39 lines
9267 and has delete-and-extract-region in it.
9268 delete-and-extract-region is written in C.
9270 see Initializing a Variable with @code{defvar}
9274 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9275 functions within it, @code{kill-append} and @code{kill-new}, copy a
9276 region in a buffer and save it in a variable called the
9277 @code{kill-ring}. This section describes how the @code{kill-ring}
9278 variable is created and initialized using the @code{defvar} special
9281 (Again we note that the term @code{kill-ring} is a misnomer. The text
9282 that is clipped out of the buffer can be brought back; it is not a ring
9283 of corpses, but a ring of resurrectable text.)
9285 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9286 given an initial value by using the @code{defvar} special form. The
9287 name comes from ``define variable''.
9289 The @code{defvar} special form is similar to @code{setq} in that it sets
9290 the value of a variable. It is unlike @code{setq} in two ways: first,
9291 it only sets the value of the variable if the variable does not already
9292 have a value. If the variable already has a value, @code{defvar} does
9293 not override the existing value. Second, @code{defvar} has a
9294 documentation string.
9296 (Another special form, @code{defcustom}, is designed for variables
9297 that people customize. It has more features than @code{defvar}.
9298 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9301 * See variable current value::
9302 * defvar and asterisk::
9306 @node See variable current value
9307 @unnumberedsubsec Seeing the Current Value of a Variable
9310 You can see the current value of a variable, any variable, by using
9311 the @code{describe-variable} function, which is usually invoked by
9312 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9313 (followed by @key{RET}) when prompted, you will see what is in your
9314 current kill ring---this may be quite a lot! Conversely, if you have
9315 been doing nothing this Emacs session except read this document, you
9316 may have nothing in it. Also, you will see the documentation for
9322 List of killed text sequences.
9323 Since the kill ring is supposed to interact nicely with cut-and-paste
9324 facilities offered by window systems, use of this variable should
9327 interact nicely with `interprogram-cut-function' and
9328 `interprogram-paste-function'. The functions `kill-new',
9329 `kill-append', and `current-kill' are supposed to implement this
9330 interaction; you may want to use them instead of manipulating the kill
9336 The kill ring is defined by a @code{defvar} in the following way:
9340 (defvar kill-ring nil
9341 "List of killed text sequences.
9347 In this variable definition, the variable is given an initial value of
9348 @code{nil}, which makes sense, since if you have saved nothing, you want
9349 nothing back if you give a @code{yank} command. The documentation
9350 string is written just like the documentation string of a @code{defun}.
9351 As with the documentation string of the @code{defun}, the first line of
9352 the documentation should be a complete sentence, since some commands,
9353 like @code{apropos}, print only the first line of documentation.
9354 Succeeding lines should not be indented; otherwise they look odd when
9355 you use @kbd{C-h v} (@code{describe-variable}).
9357 @node defvar and asterisk
9358 @subsection @code{defvar} and an asterisk
9359 @findex defvar @r{for a user customizable variable}
9360 @findex defvar @r{with an asterisk}
9362 In the past, Emacs used the @code{defvar} special form both for
9363 internal variables that you would not expect a user to change and for
9364 variables that you do expect a user to change. Although you can still
9365 use @code{defvar} for user customizable variables, please use
9366 @code{defcustom} instead, since that special form provides a path into
9367 the Customization commands. (@xref{defcustom, , Specifying Variables
9368 using @code{defcustom}}.)
9370 When you specified a variable using the @code{defvar} special form,
9371 you could distinguish a variable that a user might want to change from
9372 others by typing an asterisk, @samp{*}, in the first column of its
9373 documentation string. For example:
9377 (defvar shell-command-default-error-buffer nil
9378 "*Buffer name for `shell-command' @dots{} error output.
9383 @findex set-variable
9385 You could (and still can) use the @code{set-variable} command to
9386 change the value of @code{shell-command-default-error-buffer}
9387 temporarily. However, options set using @code{set-variable} are set
9388 only for the duration of your editing session. The new values are not
9389 saved between sessions. Each time Emacs starts, it reads the original
9390 value, unless you change the value within your @file{.emacs} file,
9391 either by setting it manually or by using @code{customize}.
9392 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9394 For me, the major use of the @code{set-variable} command is to suggest
9395 variables that I might want to set in my @file{.emacs} file. There
9396 are now more than 700 such variables, far too many to remember
9397 readily. Fortunately, you can press @key{TAB} after calling the
9398 @code{M-x set-variable} command to see the list of variables.
9399 (@xref{Examining, , Examining and Setting Variables, emacs,
9400 The GNU Emacs Manual}.)
9403 @node cons & search-fwd Review
9406 Here is a brief summary of some recently introduced functions.
9411 @code{car} returns the first element of a list; @code{cdr} returns the
9412 second and subsequent elements of a list.
9419 (car '(1 2 3 4 5 6 7))
9421 (cdr '(1 2 3 4 5 6 7))
9422 @result{} (2 3 4 5 6 7)
9427 @code{cons} constructs a list by prepending its first argument to its
9441 @code{funcall} evaluates its first argument as a function. It passes
9442 its remaining arguments to its first argument.
9445 Return the result of taking @sc{cdr} `n' times on a list.
9453 The `rest of the rest', as it were.
9460 (nthcdr 3 '(1 2 3 4 5 6 7))
9467 @code{setcar} changes the first element of a list; @code{setcdr}
9468 changes the second and subsequent elements of a list.
9475 (setq triple '(1 2 3))
9482 (setcdr triple '("foo" "bar"))
9485 @result{} (37 "foo" "bar")
9490 Evaluate each argument in sequence and then return the value of the
9503 @item save-restriction
9504 Record whatever narrowing is in effect in the current buffer, if any,
9505 and restore that narrowing after evaluating the arguments.
9507 @item search-forward
9508 Search for a string, and if the string is found, move point. With a
9509 regular expression, use the similar @code{re-search-forward}.
9510 (@xref{Regexp Search, , Regular Expression Searches}, for an
9511 explanation of regular expression patterns and searches.)
9515 @code{search-forward} and @code{re-search-forward} take four
9520 The string or regular expression to search for.
9523 Optionally, the limit of the search.
9526 Optionally, what to do if the search fails, return @code{nil} or an
9530 Optionally, how many times to repeat the search; if negative, the
9531 search goes backwards.
9535 @itemx delete-and-extract-region
9536 @itemx copy-region-as-kill
9538 @code{kill-region} cuts the text between point and mark from the
9539 buffer and stores that text in the kill ring, so you can get it back
9542 @code{copy-region-as-kill} copies the text between point and mark into
9543 the kill ring, from which you can get it by yanking. The function
9544 does not cut or remove the text from the buffer.
9547 @code{delete-and-extract-region} removes the text between point and
9548 mark from the buffer and throws it away. You cannot get it back.
9549 (This is not an interactive command.)
9552 @node search Exercises
9553 @section Searching Exercises
9557 Write an interactive function that searches for a string. If the
9558 search finds the string, leave point after it and display a message
9559 that says ``Found!''. (Do not use @code{search-forward} for the name
9560 of this function; if you do, you will overwrite the existing version of
9561 @code{search-forward} that comes with Emacs. Use a name such as
9562 @code{test-search} instead.)
9565 Write a function that prints the third element of the kill ring in the
9566 echo area, if any; if the kill ring does not contain a third element,
9567 print an appropriate message.
9570 @node List Implementation
9571 @chapter How Lists are Implemented
9572 @cindex Lists in a computer
9574 In Lisp, atoms are recorded in a straightforward fashion; if the
9575 implementation is not straightforward in practice, it is, nonetheless,
9576 straightforward in theory. The atom @samp{rose}, for example, is
9577 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9578 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9579 is equally simple, but it takes a moment to get used to the idea. A
9580 list is kept using a series of pairs of pointers. In the series, the
9581 first pointer in each pair points to an atom or to another list, and the
9582 second pointer in each pair points to the next pair, or to the symbol
9583 @code{nil}, which marks the end of the list.
9585 A pointer itself is quite simply the electronic address of what is
9586 pointed to. Hence, a list is kept as a series of electronic addresses.
9589 * Lists diagrammed::
9590 * Symbols as Chest:: Exploring a powerful metaphor.
9595 @node Lists diagrammed
9596 @unnumberedsec Lists diagrammed
9599 For example, the list @code{(rose violet buttercup)} has three elements,
9600 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9601 electronic address of @samp{rose} is recorded in a segment of computer
9602 memory along with the address that gives the electronic address of where
9603 the atom @samp{violet} is located; and that address (the one that tells
9604 where @samp{violet} is located) is kept along with an address that tells
9605 where the address for the atom @samp{buttercup} is located.
9608 This sounds more complicated than it is and is easier seen in a diagram:
9610 @c clear print-postscript-figures
9611 @c !!! cons-cell-diagram #1
9615 ___ ___ ___ ___ ___ ___
9616 |___|___|--> |___|___|--> |___|___|--> nil
9619 --> rose --> violet --> buttercup
9623 @ifset print-postscript-figures
9626 @center @image{cons-1}
9627 %%%% old method of including an image
9628 % \input /usr/local/lib/tex/inputs/psfig.tex
9629 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9634 @ifclear print-postscript-figures
9638 ___ ___ ___ ___ ___ ___
9639 |___|___|--> |___|___|--> |___|___|--> nil
9642 --> rose --> violet --> buttercup
9649 In the diagram, each box represents a word of computer memory that
9650 holds a Lisp object, usually in the form of a memory address. The boxes,
9651 i.e., the addresses, are in pairs. Each arrow points to what the address
9652 is the address of, either an atom or another pair of addresses. The
9653 first box is the electronic address of @samp{rose} and the arrow points
9654 to @samp{rose}; the second box is the address of the next pair of boxes,
9655 the first part of which is the address of @samp{violet} and the second
9656 part of which is the address of the next pair. The very last box
9657 points to the symbol @code{nil}, which marks the end of the list.
9660 When a variable is set to a list with a function such as @code{setq},
9661 it stores the address of the first box in the variable. Thus,
9662 evaluation of the expression
9665 (setq bouquet '(rose violet buttercup))
9670 creates a situation like this:
9672 @c cons-cell-diagram #2
9678 | ___ ___ ___ ___ ___ ___
9679 --> |___|___|--> |___|___|--> |___|___|--> nil
9682 --> rose --> violet --> buttercup
9686 @ifset print-postscript-figures
9689 @center @image{cons-2}
9690 %%%% old method of including an image
9691 % \input /usr/local/lib/tex/inputs/psfig.tex
9692 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9697 @ifclear print-postscript-figures
9703 | ___ ___ ___ ___ ___ ___
9704 --> |___|___|--> |___|___|--> |___|___|--> nil
9707 --> rose --> violet --> buttercup
9714 In this example, the symbol @code{bouquet} holds the address of the first
9718 This same list can be illustrated in a different sort of box notation
9721 @c cons-cell-diagram #2a
9727 | -------------- --------------- ----------------
9728 | | car | cdr | | car | cdr | | car | cdr |
9729 -->| rose | o------->| violet | o------->| butter- | nil |
9730 | | | | | | | cup | |
9731 -------------- --------------- ----------------
9735 @ifset print-postscript-figures
9738 @center @image{cons-2a}
9739 %%%% old method of including an image
9740 % \input /usr/local/lib/tex/inputs/psfig.tex
9741 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9746 @ifclear print-postscript-figures
9752 | -------------- --------------- ----------------
9753 | | car | cdr | | car | cdr | | car | cdr |
9754 -->| rose | o------->| violet | o------->| butter- | nil |
9755 | | | | | | | cup | |
9756 -------------- --------------- ----------------
9762 (Symbols consist of more than pairs of addresses, but the structure of
9763 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9764 consists of a group of address-boxes, one of which is the address of
9765 the printed word @samp{bouquet}, a second of which is the address of a
9766 function definition attached to the symbol, if any, a third of which
9767 is the address of the first pair of address-boxes for the list
9768 @code{(rose violet buttercup)}, and so on. Here we are showing that
9769 the symbol's third address-box points to the first pair of
9770 address-boxes for the list.)
9772 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9773 changed; the symbol simply has an address further down the list. (In
9774 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9775 evaluation of the following expression
9778 (setq flowers (cdr bouquet))
9785 @c cons-cell-diagram #3
9792 | ___ ___ | ___ ___ ___ ___
9793 --> | | | --> | | | | | |
9794 |___|___|----> |___|___|--> |___|___|--> nil
9797 --> rose --> violet --> buttercup
9802 @ifset print-postscript-figures
9805 @center @image{cons-3}
9806 %%%% old method of including an image
9807 % \input /usr/local/lib/tex/inputs/psfig.tex
9808 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9813 @ifclear print-postscript-figures
9820 | ___ ___ | ___ ___ ___ ___
9821 --> | | | --> | | | | | |
9822 |___|___|----> |___|___|--> |___|___|--> nil
9825 --> rose --> violet --> buttercup
9833 The value of @code{flowers} is @code{(violet buttercup)}, which is
9834 to say, the symbol @code{flowers} holds the address of the pair of
9835 address-boxes, the first of which holds the address of @code{violet},
9836 and the second of which holds the address of @code{buttercup}.
9838 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9839 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9840 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9841 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9842 information about cons cells and dotted pairs.
9845 The function @code{cons} adds a new pair of addresses to the front of
9846 a series of addresses like that shown above. For example, evaluating
9850 (setq bouquet (cons 'lily bouquet))
9857 @c cons-cell-diagram #4
9864 | ___ ___ ___ ___ | ___ ___ ___ ___
9865 --> | | | | | | --> | | | | | |
9866 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9869 --> lily --> rose --> violet --> buttercup
9874 @ifset print-postscript-figures
9877 @center @image{cons-4}
9878 %%%% old method of including an image
9879 % \input /usr/local/lib/tex/inputs/psfig.tex
9880 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9885 @ifclear print-postscript-figures
9892 | ___ ___ ___ ___ | ___ ___ ___ ___
9893 --> | | | | | | --> | | | | | |
9894 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9897 --> lily --> rose --> violet --> buttercup
9906 However, this does not change the value of the symbol
9907 @code{flowers}, as you can see by evaluating the following,
9910 (eq (cdr (cdr bouquet)) flowers)
9914 which returns @code{t} for true.
9916 Until it is reset, @code{flowers} still has the value
9917 @code{(violet buttercup)}; that is, it has the address of the cons
9918 cell whose first address is of @code{violet}. Also, this does not
9919 alter any of the pre-existing cons cells; they are all still there.
9921 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
9922 of the next cons cell in the series; to get the @sc{car} of a list,
9923 you get the address of the first element of the list; to @code{cons} a
9924 new element on a list, you add a new cons cell to the front of the list.
9925 That is all there is to it! The underlying structure of Lisp is
9928 And what does the last address in a series of cons cells refer to? It
9929 is the address of the empty list, of @code{nil}.
9931 In summary, when a Lisp variable is set to a value, it is provided with
9932 the address of the list to which the variable refers.
9934 @node Symbols as Chest
9935 @section Symbols as a Chest of Drawers
9936 @cindex Symbols as a Chest of Drawers
9937 @cindex Chest of Drawers, metaphor for a symbol
9938 @cindex Drawers, Chest of, metaphor for a symbol
9940 In an earlier section, I suggested that you might imagine a symbol as
9941 being a chest of drawers. The function definition is put in one
9942 drawer, the value in another, and so on. What is put in the drawer
9943 holding the value can be changed without affecting the contents of the
9944 drawer holding the function definition, and vice-verse.
9946 Actually, what is put in each drawer is the address of the value or
9947 function definition. It is as if you found an old chest in the attic,
9948 and in one of its drawers you found a map giving you directions to
9949 where the buried treasure lies.
9951 (In addition to its name, symbol definition, and variable value, a
9952 symbol has a `drawer' for a @dfn{property list} which can be used to
9953 record other information. Property lists are not discussed here; see
9954 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
9958 Here is a fanciful representation:
9960 @c chest-of-drawers diagram
9965 Chest of Drawers Contents of Drawers
9969 ---------------------
9970 | directions to | [map to]
9971 | symbol name | bouquet
9973 +---------------------+
9975 | symbol definition | [none]
9977 +---------------------+
9978 | directions to | [map to]
9979 | variable value | (rose violet buttercup)
9981 +---------------------+
9983 | property list | [not described here]
9985 +---------------------+
9991 @ifset print-postscript-figures
9994 @center @image{drawers}
9995 %%%% old method of including an image
9996 % \input /usr/local/lib/tex/inputs/psfig.tex
9997 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10002 @ifclear print-postscript-figures
10007 Chest of Drawers Contents of Drawers
10011 ---------------------
10012 | directions to | [map to]
10013 | symbol name | bouquet
10015 +---------------------+
10017 | symbol definition | [none]
10019 +---------------------+
10020 | directions to | [map to]
10021 | variable value | (rose violet buttercup)
10023 +---------------------+
10025 | property list | [not described here]
10027 +---------------------+
10035 @node List Exercise
10038 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10039 more flowers on to this list and set this new list to
10040 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10041 What does the @code{more-flowers} list now contain?
10044 @chapter Yanking Text Back
10046 @cindex Text retrieval
10047 @cindex Retrieving text
10048 @cindex Pasting text
10050 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10051 you can bring it back with a `yank' command. The text that is cut out of
10052 the buffer is put in the kill ring and the yank commands insert the
10053 appropriate contents of the kill ring back into a buffer (not necessarily
10054 the original buffer).
10056 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10057 the kill ring into the current buffer. If the @kbd{C-y} command is
10058 followed immediately by @kbd{M-y}, the first element is replaced by
10059 the second element. Successive @kbd{M-y} commands replace the second
10060 element with the third, fourth, or fifth element, and so on. When the
10061 last element in the kill ring is reached, it is replaced by the first
10062 element and the cycle is repeated. (Thus the kill ring is called a
10063 `ring' rather than just a `list'. However, the actual data structure
10064 that holds the text is a list.
10065 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10066 list is handled as a ring.)
10069 * Kill Ring Overview::
10070 * kill-ring-yank-pointer:: The kill ring is a list.
10071 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
10074 @node Kill Ring Overview
10075 @section Kill Ring Overview
10076 @cindex Kill ring overview
10078 The kill ring is a list of textual strings. This is what it looks like:
10081 ("some text" "a different piece of text" "yet more text")
10084 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10085 string of characters saying @samp{some text} would be inserted in this
10086 buffer where my cursor is located.
10088 The @code{yank} command is also used for duplicating text by copying it.
10089 The copied text is not cut from the buffer, but a copy of it is put on the
10090 kill ring and is inserted by yanking it back.
10092 Three functions are used for bringing text back from the kill ring:
10093 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10094 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10095 which is used by the two other functions.
10097 These functions refer to the kill ring through a variable called the
10098 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10099 @code{yank} and @code{yank-pop} functions is:
10102 (insert (car kill-ring-yank-pointer))
10106 (Well, no more. In GNU Emacs 22, the function has been replaced by
10107 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10108 repetitively for each @code{yank-handler} segment. In turn,
10109 @code{insert-for-yank-1} strips text properties from the inserted text
10110 according to @code{yank-excluded-properties}. Otherwise, it is just
10111 like @code{insert}. We will stick with plain @code{insert} since it
10112 is easier to understand.)
10114 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10115 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10117 @node kill-ring-yank-pointer
10118 @section The @code{kill-ring-yank-pointer} Variable
10120 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10121 a variable. It points to something by being bound to the value of what
10122 it points to, like any other Lisp variable.
10125 Thus, if the value of the kill ring is:
10128 ("some text" "a different piece of text" "yet more text")
10133 and the @code{kill-ring-yank-pointer} points to the second clause, the
10134 value of @code{kill-ring-yank-pointer} is:
10137 ("a different piece of text" "yet more text")
10140 As explained in the previous chapter (@pxref{List Implementation}), the
10141 computer does not keep two different copies of the text being pointed to
10142 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10143 words ``a different piece of text'' and ``yet more text'' are not
10144 duplicated. Instead, the two Lisp variables point to the same pieces of
10145 text. Here is a diagram:
10147 @c cons-cell-diagram #5
10151 kill-ring kill-ring-yank-pointer
10153 | ___ ___ | ___ ___ ___ ___
10154 ---> | | | --> | | | | | |
10155 |___|___|----> |___|___|--> |___|___|--> nil
10158 | | --> "yet more text"
10160 | --> "a different piece of text"
10167 @ifset print-postscript-figures
10170 @center @image{cons-5}
10171 %%%% old method of including an image
10172 % \input /usr/local/lib/tex/inputs/psfig.tex
10173 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10178 @ifclear print-postscript-figures
10182 kill-ring kill-ring-yank-pointer
10184 | ___ ___ | ___ ___ ___ ___
10185 ---> | | | --> | | | | | |
10186 |___|___|----> |___|___|--> |___|___|--> nil
10189 | | --> "yet more text"
10191 | --> "a different piece of text
10200 Both the variable @code{kill-ring} and the variable
10201 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10202 usually described as if it were actually what it is composed of. The
10203 @code{kill-ring} is spoken of as if it were the list rather than that it
10204 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10205 spoken of as pointing to a list.
10207 These two ways of talking about the same thing sound confusing at first but
10208 make sense on reflection. The kill ring is generally thought of as the
10209 complete structure of data that holds the information of what has recently
10210 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10211 on the other hand, serves to indicate---that is, to `point to'---that part
10212 of the kill ring of which the first element (the @sc{car}) will be
10216 In GNU Emacs 22, the @code{kill-new} function calls
10218 @code{(setq kill-ring-yank-pointer kill-ring)}
10220 (defun rotate-yank-pointer (arg)
10221 "Rotate the yanking point in the kill ring.
10222 With argument, rotate that many kills forward (or backward, if negative)."
10224 (current-kill arg))
10226 (defun current-kill (n &optional do-not-move)
10227 "Rotate the yanking point by N places, and then return that kill.
10228 If N is zero, `interprogram-paste-function' is set, and calling it
10229 returns a string, then that string is added to the front of the
10230 kill ring and returned as the latest kill.
10231 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10232 yanking point; just return the Nth kill forward."
10233 (let ((interprogram-paste (and (= n 0)
10234 interprogram-paste-function
10235 (funcall interprogram-paste-function))))
10236 (if interprogram-paste
10238 ;; Disable the interprogram cut function when we add the new
10239 ;; text to the kill ring, so Emacs doesn't try to own the
10240 ;; selection, with identical text.
10241 (let ((interprogram-cut-function nil))
10242 (kill-new interprogram-paste))
10243 interprogram-paste)
10244 (or kill-ring (error "Kill ring is empty"))
10245 (let ((ARGth-kill-element
10246 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10247 (length kill-ring))
10250 (setq kill-ring-yank-pointer ARGth-kill-element))
10251 (car ARGth-kill-element)))))
10256 @node yank nthcdr Exercises
10257 @section Exercises with @code{yank} and @code{nthcdr}
10261 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10262 your kill ring. Add several items to your kill ring; look at its
10263 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10264 around the kill ring. How many items were in your kill ring? Find
10265 the value of @code{kill-ring-max}. Was your kill ring full, or could
10266 you have kept more blocks of text within it?
10269 Using @code{nthcdr} and @code{car}, construct a series of expressions
10270 to return the first, second, third, and fourth elements of a list.
10273 @node Loops & Recursion
10274 @chapter Loops and Recursion
10275 @cindex Loops and recursion
10276 @cindex Recursion and loops
10277 @cindex Repetition (loops)
10279 Emacs Lisp has two primary ways to cause an expression, or a series of
10280 expressions, to be evaluated repeatedly: one uses a @code{while}
10281 loop, and the other uses @dfn{recursion}.
10283 Repetition can be very valuable. For example, to move forward four
10284 sentences, you need only write a program that will move forward one
10285 sentence and then repeat the process four times. Since a computer does
10286 not get bored or tired, such repetitive action does not have the
10287 deleterious effects that excessive or the wrong kinds of repetition can
10290 People mostly write Emacs Lisp functions using @code{while} loops and
10291 their kin; but you can use recursion, which provides a very powerful
10292 way to think about and then to solve problems@footnote{You can write
10293 recursive functions to be frugal or wasteful of mental or computer
10294 resources; as it happens, methods that people find easy---that are
10295 frugal of `mental resources'---sometimes use considerable computer
10296 resources. Emacs was designed to run on machines that we now consider
10297 limited and its default settings are conservative. You may want to
10298 increase the values of @code{max-specpdl-size} and
10299 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10300 15 and 30 times their default value.}.
10303 * while:: Causing a stretch of code to repeat.
10305 * Recursion:: Causing a function to call itself.
10306 * Looping exercise::
10310 @section @code{while}
10314 The @code{while} special form tests whether the value returned by
10315 evaluating its first argument is true or false. This is similar to what
10316 the Lisp interpreter does with an @code{if}; what the interpreter does
10317 next, however, is different.
10319 In a @code{while} expression, if the value returned by evaluating the
10320 first argument is false, the Lisp interpreter skips the rest of the
10321 expression (the @dfn{body} of the expression) and does not evaluate it.
10322 However, if the value is true, the Lisp interpreter evaluates the body
10323 of the expression and then again tests whether the first argument to
10324 @code{while} is true or false. If the value returned by evaluating the
10325 first argument is again true, the Lisp interpreter again evaluates the
10326 body of the expression.
10329 The template for a @code{while} expression looks like this:
10333 (while @var{true-or-false-test}
10339 * Looping with while:: Repeat so long as test returns true.
10340 * Loop Example:: A @code{while} loop that uses a list.
10341 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10342 * Incrementing Loop:: A loop with an incrementing counter.
10343 * Incrementing Loop Details::
10344 * Decrementing Loop:: A loop with a decrementing counter.
10348 @node Looping with while
10349 @unnumberedsubsec Looping with @code{while}
10352 So long as the true-or-false-test of the @code{while} expression
10353 returns a true value when it is evaluated, the body is repeatedly
10354 evaluated. This process is called a loop since the Lisp interpreter
10355 repeats the same thing again and again, like an airplane doing a loop.
10356 When the result of evaluating the true-or-false-test is false, the
10357 Lisp interpreter does not evaluate the rest of the @code{while}
10358 expression and `exits the loop'.
10360 Clearly, if the value returned by evaluating the first argument to
10361 @code{while} is always true, the body following will be evaluated
10362 again and again @dots{} and again @dots{} forever. Conversely, if the
10363 value returned is never true, the expressions in the body will never
10364 be evaluated. The craft of writing a @code{while} loop consists of
10365 choosing a mechanism such that the true-or-false-test returns true
10366 just the number of times that you want the subsequent expressions to
10367 be evaluated, and then have the test return false.
10369 The value returned by evaluating a @code{while} is the value of the
10370 true-or-false-test. An interesting consequence of this is that a
10371 @code{while} loop that evaluates without error will return @code{nil}
10372 or false regardless of whether it has looped 1 or 100 times or none at
10373 all. A @code{while} expression that evaluates successfully never
10374 returns a true value! What this means is that @code{while} is always
10375 evaluated for its side effects, which is to say, the consequences of
10376 evaluating the expressions within the body of the @code{while} loop.
10377 This makes sense. It is not the mere act of looping that is desired,
10378 but the consequences of what happens when the expressions in the loop
10379 are repeatedly evaluated.
10382 @subsection A @code{while} Loop and a List
10384 A common way to control a @code{while} loop is to test whether a list
10385 has any elements. If it does, the loop is repeated; but if it does not,
10386 the repetition is ended. Since this is an important technique, we will
10387 create a short example to illustrate it.
10389 A simple way to test whether a list has elements is to evaluate the
10390 list: if it has no elements, it is an empty list and will return the
10391 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10392 the other hand, a list with elements will return those elements when it
10393 is evaluated. Since Emacs Lisp considers as true any value that is not
10394 @code{nil}, a list that returns elements will test true in a
10398 For example, you can set the variable @code{empty-list} to @code{nil} by
10399 evaluating the following @code{setq} expression:
10402 (setq empty-list ())
10406 After evaluating the @code{setq} expression, you can evaluate the
10407 variable @code{empty-list} in the usual way, by placing the cursor after
10408 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10415 On the other hand, if you set a variable to be a list with elements, the
10416 list will appear when you evaluate the variable, as you can see by
10417 evaluating the following two expressions:
10421 (setq animals '(gazelle giraffe lion tiger))
10427 Thus, to create a @code{while} loop that tests whether there are any
10428 items in the list @code{animals}, the first part of the loop will be
10439 When the @code{while} tests its first argument, the variable
10440 @code{animals} is evaluated. It returns a list. So long as the list
10441 has elements, the @code{while} considers the results of the test to be
10442 true; but when the list is empty, it considers the results of the test
10445 To prevent the @code{while} loop from running forever, some mechanism
10446 needs to be provided to empty the list eventually. An oft-used
10447 technique is to have one of the subsequent forms in the @code{while}
10448 expression set the value of the list to be the @sc{cdr} of the list.
10449 Each time the @code{cdr} function is evaluated, the list will be made
10450 shorter, until eventually only the empty list will be left. At this
10451 point, the test of the @code{while} loop will return false, and the
10452 arguments to the @code{while} will no longer be evaluated.
10454 For example, the list of animals bound to the variable @code{animals}
10455 can be set to be the @sc{cdr} of the original list with the
10456 following expression:
10459 (setq animals (cdr animals))
10463 If you have evaluated the previous expressions and then evaluate this
10464 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10465 area. If you evaluate the expression again, @code{(lion tiger)} will
10466 appear in the echo area. If you evaluate it again and yet again,
10467 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10469 A template for a @code{while} loop that uses the @code{cdr} function
10470 repeatedly to cause the true-or-false-test eventually to test false
10475 (while @var{test-whether-list-is-empty}
10477 @var{set-list-to-cdr-of-list})
10481 This test and use of @code{cdr} can be put together in a function that
10482 goes through a list and prints each element of the list on a line of its
10485 @node print-elements-of-list
10486 @subsection An Example: @code{print-elements-of-list}
10487 @findex print-elements-of-list
10489 The @code{print-elements-of-list} function illustrates a @code{while}
10492 @cindex @file{*scratch*} buffer
10493 The function requires several lines for its output. If you are
10494 reading this in a recent instance of GNU Emacs,
10495 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10496 you can evaluate the following expression inside of Info, as usual.
10498 If you are using an earlier version of Emacs, you need to copy the
10499 necessary expressions to your @file{*scratch*} buffer and evaluate
10500 them there. This is because the echo area had only one line in the
10503 You can copy the expressions by marking the beginning of the region
10504 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10505 the end of the region and then copying the region using @kbd{M-w}
10506 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10507 then provides visual feedback). In the @file{*scratch*}
10508 buffer, you can yank the expressions back by typing @kbd{C-y}
10511 After you have copied the expressions to the @file{*scratch*} buffer,
10512 evaluate each expression in turn. Be sure to evaluate the last
10513 expression, @code{(print-elements-of-list animals)}, by typing
10514 @kbd{C-u C-x C-e}, that is, by giving an argument to
10515 @code{eval-last-sexp}. This will cause the result of the evaluation
10516 to be printed in the @file{*scratch*} buffer instead of being printed
10517 in the echo area. (Otherwise you will see something like this in your
10518 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10519 each @samp{^J} stands for a `newline'.)
10522 In a recent instance of GNU Emacs, you can evaluate these expressions
10523 directly in the Info buffer, and the echo area will grow to show the
10528 (setq animals '(gazelle giraffe lion tiger))
10530 (defun print-elements-of-list (list)
10531 "Print each element of LIST on a line of its own."
10534 (setq list (cdr list))))
10536 (print-elements-of-list animals)
10542 When you evaluate the three expressions in sequence, you will see
10558 Each element of the list is printed on a line of its own (that is what
10559 the function @code{print} does) and then the value returned by the
10560 function is printed. Since the last expression in the function is the
10561 @code{while} loop, and since @code{while} loops always return
10562 @code{nil}, a @code{nil} is printed after the last element of the list.
10564 @node Incrementing Loop
10565 @subsection A Loop with an Incrementing Counter
10567 A loop is not useful unless it stops when it ought. Besides
10568 controlling a loop with a list, a common way of stopping a loop is to
10569 write the first argument as a test that returns false when the correct
10570 number of repetitions are complete. This means that the loop must
10571 have a counter---an expression that counts how many times the loop
10575 @node Incrementing Loop Details
10576 @unnumberedsubsec Details of an Incrementing Loop
10579 The test for a loop with an incrementing counter can be an expression
10580 such as @code{(< count desired-number)} which returns @code{t} for
10581 true if the value of @code{count} is less than the
10582 @code{desired-number} of repetitions and @code{nil} for false if the
10583 value of @code{count} is equal to or is greater than the
10584 @code{desired-number}. The expression that increments the count can
10585 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10586 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10587 argument. (The expression @w{@code{(1+ count)}} has the same result
10588 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10591 The template for a @code{while} loop controlled by an incrementing
10592 counter looks like this:
10596 @var{set-count-to-initial-value}
10597 (while (< count desired-number) ; @r{true-or-false-test}
10599 (setq count (1+ count))) ; @r{incrementer}
10604 Note that you need to set the initial value of @code{count}; usually it
10608 * Incrementing Example:: Counting pebbles in a triangle.
10609 * Inc Example parts:: The parts of the function definition.
10610 * Inc Example altogether:: Putting the function definition together.
10613 @node Incrementing Example
10614 @unnumberedsubsubsec Example with incrementing counter
10616 Suppose you are playing on the beach and decide to make a triangle of
10617 pebbles, putting one pebble in the first row, two in the second row,
10618 three in the third row and so on, like this:
10636 @bullet{} @bullet{}
10637 @bullet{} @bullet{} @bullet{}
10638 @bullet{} @bullet{} @bullet{} @bullet{}
10645 (About 2500 years ago, Pythagoras and others developed the beginnings of
10646 number theory by considering questions such as this.)
10648 Suppose you want to know how many pebbles you will need to make a
10649 triangle with 7 rows?
10651 Clearly, what you need to do is add up the numbers from 1 to 7. There
10652 are two ways to do this; start with the smallest number, one, and add up
10653 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10654 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10655 mechanisms illustrate common ways of writing @code{while} loops, we will
10656 create two examples, one counting up and the other counting down. In
10657 this first example, we will start with 1 and add 2, 3, 4 and so on.
10659 If you are just adding up a short list of numbers, the easiest way to do
10660 it is to add up all the numbers at once. However, if you do not know
10661 ahead of time how many numbers your list will have, or if you want to be
10662 prepared for a very long list, then you need to design your addition so
10663 that what you do is repeat a simple process many times instead of doing
10664 a more complex process once.
10666 For example, instead of adding up all the pebbles all at once, what you
10667 can do is add the number of pebbles in the first row, 1, to the number
10668 in the second row, 2, and then add the total of those two rows to the
10669 third row, 3. Then you can add the number in the fourth row, 4, to the
10670 total of the first three rows; and so on.
10672 The critical characteristic of the process is that each repetitive
10673 action is simple. In this case, at each step we add only two numbers,
10674 the number of pebbles in the row and the total already found. This
10675 process of adding two numbers is repeated again and again until the last
10676 row has been added to the total of all the preceding rows. In a more
10677 complex loop the repetitive action might not be so simple, but it will
10678 be simpler than doing everything all at once.
10680 @node Inc Example parts
10681 @unnumberedsubsubsec The parts of the function definition
10683 The preceding analysis gives us the bones of our function definition:
10684 first, we will need a variable that we can call @code{total} that will
10685 be the total number of pebbles. This will be the value returned by
10688 Second, we know that the function will require an argument: this
10689 argument will be the total number of rows in the triangle. It can be
10690 called @code{number-of-rows}.
10692 Finally, we need a variable to use as a counter. We could call this
10693 variable @code{counter}, but a better name is @code{row-number}. That
10694 is because what the counter does in this function is count rows, and a
10695 program should be written to be as understandable as possible.
10697 When the Lisp interpreter first starts evaluating the expressions in the
10698 function, the value of @code{total} should be set to zero, since we have
10699 not added anything to it. Then the function should add the number of
10700 pebbles in the first row to the total, and then add the number of
10701 pebbles in the second to the total, and then add the number of
10702 pebbles in the third row to the total, and so on, until there are no
10703 more rows left to add.
10705 Both @code{total} and @code{row-number} are used only inside the
10706 function, so they can be declared as local variables with @code{let}
10707 and given initial values. Clearly, the initial value for @code{total}
10708 should be 0. The initial value of @code{row-number} should be 1,
10709 since we start with the first row. This means that the @code{let}
10710 statement will look like this:
10720 After the internal variables are declared and bound to their initial
10721 values, we can begin the @code{while} loop. The expression that serves
10722 as the test should return a value of @code{t} for true so long as the
10723 @code{row-number} is less than or equal to the @code{number-of-rows}.
10724 (If the expression tests true only so long as the row number is less
10725 than the number of rows in the triangle, the last row will never be
10726 added to the total; hence the row number has to be either less than or
10727 equal to the number of rows.)
10730 @findex <= @r{(less than or equal)}
10731 Lisp provides the @code{<=} function that returns true if the value of
10732 its first argument is less than or equal to the value of its second
10733 argument and false otherwise. So the expression that the @code{while}
10734 will evaluate as its test should look like this:
10737 (<= row-number number-of-rows)
10740 The total number of pebbles can be found by repeatedly adding the number
10741 of pebbles in a row to the total already found. Since the number of
10742 pebbles in the row is equal to the row number, the total can be found by
10743 adding the row number to the total. (Clearly, in a more complex
10744 situation, the number of pebbles in the row might be related to the row
10745 number in a more complicated way; if this were the case, the row number
10746 would be replaced by the appropriate expression.)
10749 (setq total (+ total row-number))
10753 What this does is set the new value of @code{total} to be equal to the
10754 sum of adding the number of pebbles in the row to the previous total.
10756 After setting the value of @code{total}, the conditions need to be
10757 established for the next repetition of the loop, if there is one. This
10758 is done by incrementing the value of the @code{row-number} variable,
10759 which serves as a counter. After the @code{row-number} variable has
10760 been incremented, the true-or-false-test at the beginning of the
10761 @code{while} loop tests whether its value is still less than or equal to
10762 the value of the @code{number-of-rows} and if it is, adds the new value
10763 of the @code{row-number} variable to the @code{total} of the previous
10764 repetition of the loop.
10767 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10768 @code{row-number} variable can be incremented with this expression:
10771 (setq row-number (1+ row-number))
10774 @node Inc Example altogether
10775 @unnumberedsubsubsec Putting the function definition together
10777 We have created the parts for the function definition; now we need to
10781 First, the contents of the @code{while} expression:
10785 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10786 (setq total (+ total row-number))
10787 (setq row-number (1+ row-number))) ; @r{incrementer}
10791 Along with the @code{let} expression varlist, this very nearly
10792 completes the body of the function definition. However, it requires
10793 one final element, the need for which is somewhat subtle.
10795 The final touch is to place the variable @code{total} on a line by
10796 itself after the @code{while} expression. Otherwise, the value returned
10797 by the whole function is the value of the last expression that is
10798 evaluated in the body of the @code{let}, and this is the value
10799 returned by the @code{while}, which is always @code{nil}.
10801 This may not be evident at first sight. It almost looks as if the
10802 incrementing expression is the last expression of the whole function.
10803 But that expression is part of the body of the @code{while}; it is the
10804 last element of the list that starts with the symbol @code{while}.
10805 Moreover, the whole of the @code{while} loop is a list within the body
10809 In outline, the function will look like this:
10813 (defun @var{name-of-function} (@var{argument-list})
10814 "@var{documentation}@dots{}"
10815 (let (@var{varlist})
10816 (while (@var{true-or-false-test})
10817 @var{body-of-while}@dots{} )
10818 @dots{} )) ; @r{Need final expression here.}
10822 The result of evaluating the @code{let} is what is going to be returned
10823 by the @code{defun} since the @code{let} is not embedded within any
10824 containing list, except for the @code{defun} as a whole. However, if
10825 the @code{while} is the last element of the @code{let} expression, the
10826 function will always return @code{nil}. This is not what we want!
10827 Instead, what we want is the value of the variable @code{total}. This
10828 is returned by simply placing the symbol as the last element of the list
10829 starting with @code{let}. It gets evaluated after the preceding
10830 elements of the list are evaluated, which means it gets evaluated after
10831 it has been assigned the correct value for the total.
10833 It may be easier to see this by printing the list starting with
10834 @code{let} all on one line. This format makes it evident that the
10835 @var{varlist} and @code{while} expressions are the second and third
10836 elements of the list starting with @code{let}, and the @code{total} is
10841 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10846 Putting everything together, the @code{triangle} function definition
10851 (defun triangle (number-of-rows) ; @r{Version with}
10852 ; @r{ incrementing counter.}
10853 "Add up the number of pebbles in a triangle.
10854 The first row has one pebble, the second row two pebbles,
10855 the third row three pebbles, and so on.
10856 The argument is NUMBER-OF-ROWS."
10861 (while (<= row-number number-of-rows)
10862 (setq total (+ total row-number))
10863 (setq row-number (1+ row-number)))
10869 After you have installed @code{triangle} by evaluating the function, you
10870 can try it out. Here are two examples:
10881 The sum of the first four numbers is 10 and the sum of the first seven
10884 @node Decrementing Loop
10885 @subsection Loop with a Decrementing Counter
10887 Another common way to write a @code{while} loop is to write the test
10888 so that it determines whether a counter is greater than zero. So long
10889 as the counter is greater than zero, the loop is repeated. But when
10890 the counter is equal to or less than zero, the loop is stopped. For
10891 this to work, the counter has to start out greater than zero and then
10892 be made smaller and smaller by a form that is evaluated
10895 The test will be an expression such as @code{(> counter 0)} which
10896 returns @code{t} for true if the value of @code{counter} is greater
10897 than zero, and @code{nil} for false if the value of @code{counter} is
10898 equal to or less than zero. The expression that makes the number
10899 smaller and smaller can be a simple @code{setq} such as @code{(setq
10900 counter (1- counter))}, where @code{1-} is a built-in function in
10901 Emacs Lisp that subtracts 1 from its argument.
10904 The template for a decrementing @code{while} loop looks like this:
10908 (while (> counter 0) ; @r{true-or-false-test}
10910 (setq counter (1- counter))) ; @r{decrementer}
10915 * Decrementing Example:: More pebbles on the beach.
10916 * Dec Example parts:: The parts of the function definition.
10917 * Dec Example altogether:: Putting the function definition together.
10920 @node Decrementing Example
10921 @unnumberedsubsubsec Example with decrementing counter
10923 To illustrate a loop with a decrementing counter, we will rewrite the
10924 @code{triangle} function so the counter decreases to zero.
10926 This is the reverse of the earlier version of the function. In this
10927 case, to find out how many pebbles are needed to make a triangle with
10928 3 rows, add the number of pebbles in the third row, 3, to the number
10929 in the preceding row, 2, and then add the total of those two rows to
10930 the row that precedes them, which is 1.
10932 Likewise, to find the number of pebbles in a triangle with 7 rows, add
10933 the number of pebbles in the seventh row, 7, to the number in the
10934 preceding row, which is 6, and then add the total of those two rows to
10935 the row that precedes them, which is 5, and so on. As in the previous
10936 example, each addition only involves adding two numbers, the total of
10937 the rows already added up and the number of pebbles in the row that is
10938 being added to the total. This process of adding two numbers is
10939 repeated again and again until there are no more pebbles to add.
10941 We know how many pebbles to start with: the number of pebbles in the
10942 last row is equal to the number of rows. If the triangle has seven
10943 rows, the number of pebbles in the last row is 7. Likewise, we know how
10944 many pebbles are in the preceding row: it is one less than the number in
10947 @node Dec Example parts
10948 @unnumberedsubsubsec The parts of the function definition
10950 We start with three variables: the total number of rows in the
10951 triangle; the number of pebbles in a row; and the total number of
10952 pebbles, which is what we want to calculate. These variables can be
10953 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
10954 @code{total}, respectively.
10956 Both @code{total} and @code{number-of-pebbles-in-row} are used only
10957 inside the function and are declared with @code{let}. The initial
10958 value of @code{total} should, of course, be zero. However, the
10959 initial value of @code{number-of-pebbles-in-row} should be equal to
10960 the number of rows in the triangle, since the addition will start with
10964 This means that the beginning of the @code{let} expression will look
10970 (number-of-pebbles-in-row number-of-rows))
10975 The total number of pebbles can be found by repeatedly adding the number
10976 of pebbles in a row to the total already found, that is, by repeatedly
10977 evaluating the following expression:
10980 (setq total (+ total number-of-pebbles-in-row))
10984 After the @code{number-of-pebbles-in-row} is added to the @code{total},
10985 the @code{number-of-pebbles-in-row} should be decremented by one, since
10986 the next time the loop repeats, the preceding row will be
10987 added to the total.
10989 The number of pebbles in a preceding row is one less than the number of
10990 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
10991 used to compute the number of pebbles in the preceding row. This can be
10992 done with the following expression:
10996 (setq number-of-pebbles-in-row
10997 (1- number-of-pebbles-in-row))
11001 Finally, we know that the @code{while} loop should stop making repeated
11002 additions when there are no pebbles in a row. So the test for
11003 the @code{while} loop is simply:
11006 (while (> number-of-pebbles-in-row 0)
11009 @node Dec Example altogether
11010 @unnumberedsubsubsec Putting the function definition together
11012 We can put these expressions together to create a function definition
11013 that works. However, on examination, we find that one of the local
11014 variables is unneeded!
11017 The function definition looks like this:
11021 ;;; @r{First subtractive version.}
11022 (defun triangle (number-of-rows)
11023 "Add up the number of pebbles in a triangle."
11025 (number-of-pebbles-in-row number-of-rows))
11026 (while (> number-of-pebbles-in-row 0)
11027 (setq total (+ total number-of-pebbles-in-row))
11028 (setq number-of-pebbles-in-row
11029 (1- number-of-pebbles-in-row)))
11034 As written, this function works.
11036 However, we do not need @code{number-of-pebbles-in-row}.
11038 @cindex Argument as local variable
11039 When the @code{triangle} function is evaluated, the symbol
11040 @code{number-of-rows} will be bound to a number, giving it an initial
11041 value. That number can be changed in the body of the function as if
11042 it were a local variable, without any fear that such a change will
11043 effect the value of the variable outside of the function. This is a
11044 very useful characteristic of Lisp; it means that the variable
11045 @code{number-of-rows} can be used anywhere in the function where
11046 @code{number-of-pebbles-in-row} is used.
11049 Here is a second version of the function written a bit more cleanly:
11053 (defun triangle (number) ; @r{Second version.}
11054 "Return sum of numbers 1 through NUMBER inclusive."
11056 (while (> number 0)
11057 (setq total (+ total number))
11058 (setq number (1- number)))
11063 In brief, a properly written @code{while} loop will consist of three parts:
11067 A test that will return false after the loop has repeated itself the
11068 correct number of times.
11071 An expression the evaluation of which will return the value desired
11072 after being repeatedly evaluated.
11075 An expression to change the value passed to the true-or-false-test so
11076 that the test returns false after the loop has repeated itself the right
11080 @node dolist dotimes
11081 @section Save your time: @code{dolist} and @code{dotimes}
11083 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11084 provide for looping. Sometimes these are quicker to write than the
11085 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11086 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11088 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11089 list': @code{dolist} automatically shortens the list each time it
11090 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11091 each shorter version of the list to the first of its arguments.
11093 @code{dotimes} loops a specific number of times: you specify the number.
11101 @unnumberedsubsec The @code{dolist} Macro
11104 Suppose, for example, you want to reverse a list, so that
11105 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11108 In practice, you would use the @code{reverse} function, like this:
11112 (setq animals '(gazelle giraffe lion tiger))
11120 Here is how you could reverse the list using a @code{while} loop:
11124 (setq animals '(gazelle giraffe lion tiger))
11126 (defun reverse-list-with-while (list)
11127 "Using while, reverse the order of LIST."
11128 (let (value) ; make sure list starts empty
11130 (setq value (cons (car list) value))
11131 (setq list (cdr list)))
11134 (reverse-list-with-while animals)
11140 And here is how you could use the @code{dolist} macro:
11144 (setq animals '(gazelle giraffe lion tiger))
11146 (defun reverse-list-with-dolist (list)
11147 "Using dolist, reverse the order of LIST."
11148 (let (value) ; make sure list starts empty
11149 (dolist (element list value)
11150 (setq value (cons element value)))))
11152 (reverse-list-with-dolist animals)
11158 In Info, you can place your cursor after the closing parenthesis of
11159 each expression and type @kbd{C-x C-e}; in each case, you should see
11162 (tiger lion giraffe gazelle)
11168 For this example, the existing @code{reverse} function is obviously best.
11169 The @code{while} loop is just like our first example (@pxref{Loop
11170 Example, , A @code{while} Loop and a List}). The @code{while} first
11171 checks whether the list has elements; if so, it constructs a new list
11172 by adding the first element of the list to the existing list (which in
11173 the first iteration of the loop is @code{nil}). Since the second
11174 element is prepended in front of the first element, and the third
11175 element is prepended in front of the second element, the list is reversed.
11177 In the expression using a @code{while} loop,
11178 the @w{@code{(setq list (cdr list))}}
11179 expression shortens the list, so the @code{while} loop eventually
11180 stops. In addition, it provides the @code{cons} expression with a new
11181 first element by creating a new and shorter list at each repetition of
11184 The @code{dolist} expression does very much the same as the
11185 @code{while} expression, except that the @code{dolist} macro does some
11186 of the work you have to do when writing a @code{while} expression.
11188 Like a @code{while} loop, a @code{dolist} loops. What is different is
11189 that it automatically shortens the list each time it loops---it
11190 `@sc{cdr}s down the list' on its own---and it automatically binds
11191 the @sc{car} of each shorter version of the list to the first of its
11194 In the example, the @sc{car} of each shorter version of the list is
11195 referred to using the symbol @samp{element}, the list itself is called
11196 @samp{list}, and the value returned is called @samp{value}. The
11197 remainder of the @code{dolist} expression is the body.
11199 The @code{dolist} expression binds the @sc{car} of each shorter
11200 version of the list to @code{element} and then evaluates the body of
11201 the expression; and repeats the loop. The result is returned in
11205 @unnumberedsubsec The @code{dotimes} Macro
11208 The @code{dotimes} macro is similar to @code{dolist}, except that it
11209 loops a specific number of times.
11211 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11212 and so forth each time around the loop, and the value of the third
11213 argument is returned. You need to provide the value of the second
11214 argument, which is how many times the macro loops.
11217 For example, the following binds the numbers from 0 up to, but not
11218 including, the number 3 to the first argument, @var{number}, and then
11219 constructs a list of the three numbers. (The first number is 0, the
11220 second number is 1, and the third number is 2; this makes a total of
11221 three numbers in all, starting with zero as the first number.)
11225 (let (value) ; otherwise a value is a void variable
11226 (dotimes (number 3 value)
11227 (setq value (cons number value))))
11234 @code{dotimes} returns @code{value}, so the way to use
11235 @code{dotimes} is to operate on some expression @var{number} number of
11236 times and then return the result, either as a list or an atom.
11239 Here is an example of a @code{defun} that uses @code{dotimes} to add
11240 up the number of pebbles in a triangle.
11244 (defun triangle-using-dotimes (number-of-rows)
11245 "Using dotimes, add up the number of pebbles in a triangle."
11246 (let ((total 0)) ; otherwise a total is a void variable
11247 (dotimes (number number-of-rows total)
11248 (setq total (+ total (1+ number))))))
11250 (triangle-using-dotimes 4)
11258 A recursive function contains code that tells the Lisp interpreter to
11259 call a program that runs exactly like itself, but with slightly
11260 different arguments. The code runs exactly the same because it has
11261 the same name. However, even though the program has the same name, it
11262 is not the same entity. It is different. In the jargon, it is a
11263 different `instance'.
11265 Eventually, if the program is written correctly, the `slightly
11266 different arguments' will become sufficiently different from the first
11267 arguments that the final instance will stop.
11270 * Building Robots:: Same model, different serial number ...
11271 * Recursive Definition Parts:: Walk until you stop ...
11272 * Recursion with list:: Using a list as the test whether to recurse.
11273 * Recursive triangle function::
11274 * Recursion with cond::
11275 * Recursive Patterns:: Often used templates.
11276 * No Deferment:: Don't store up work ...
11277 * No deferment solution::
11280 @node Building Robots
11281 @subsection Building Robots: Extending the Metaphor
11282 @cindex Building robots
11283 @cindex Robots, building
11285 It is sometimes helpful to think of a running program as a robot that
11286 does a job. In doing its job, a recursive function calls on a second
11287 robot to help it. The second robot is identical to the first in every
11288 way, except that the second robot helps the first and has been
11289 passed different arguments than the first.
11291 In a recursive function, the second robot may call a third; and the
11292 third may call a fourth, and so on. Each of these is a different
11293 entity; but all are clones.
11295 Since each robot has slightly different instructions---the arguments
11296 will differ from one robot to the next---the last robot should know
11299 Let's expand on the metaphor in which a computer program is a robot.
11301 A function definition provides the blueprints for a robot. When you
11302 install a function definition, that is, when you evaluate a
11303 @code{defun} special form, you install the necessary equipment to
11304 build robots. It is as if you were in a factory, setting up an
11305 assembly line. Robots with the same name are built according to the
11306 same blueprints. So they have, as it were, the same `model number',
11307 but a different `serial number'.
11309 We often say that a recursive function `calls itself'. What we mean
11310 is that the instructions in a recursive function cause the Lisp
11311 interpreter to run a different function that has the same name and
11312 does the same job as the first, but with different arguments.
11314 It is important that the arguments differ from one instance to the
11315 next; otherwise, the process will never stop.
11317 @node Recursive Definition Parts
11318 @subsection The Parts of a Recursive Definition
11319 @cindex Parts of a Recursive Definition
11320 @cindex Recursive Definition Parts
11322 A recursive function typically contains a conditional expression which
11327 A true-or-false-test that determines whether the function is called
11328 again, here called the @dfn{do-again-test}.
11331 The name of the function. When this name is called, a new instance of
11332 the function---a new robot, as it were---is created and told what to do.
11335 An expression that returns a different value each time the function is
11336 called, here called the @dfn{next-step-expression}. Consequently, the
11337 argument (or arguments) passed to the new instance of the function
11338 will be different from that passed to the previous instance. This
11339 causes the conditional expression, the @dfn{do-again-test}, to test
11340 false after the correct number of repetitions.
11343 Recursive functions can be much simpler than any other kind of
11344 function. Indeed, when people first start to use them, they often look
11345 so mysteriously simple as to be incomprehensible. Like riding a
11346 bicycle, reading a recursive function definition takes a certain knack
11347 which is hard at first but then seems simple.
11350 There are several different common recursive patterns. A very simple
11351 pattern looks like this:
11355 (defun @var{name-of-recursive-function} (@var{argument-list})
11356 "@var{documentation}@dots{}"
11357 (if @var{do-again-test}
11359 (@var{name-of-recursive-function}
11360 @var{next-step-expression})))
11364 Each time a recursive function is evaluated, a new instance of it is
11365 created and told what to do. The arguments tell the instance what to do.
11367 An argument is bound to the value of the next-step-expression. Each
11368 instance runs with a different value of the next-step-expression.
11370 The value in the next-step-expression is used in the do-again-test.
11372 The value returned by the next-step-expression is passed to the new
11373 instance of the function, which evaluates it (or some
11374 transmogrification of it) to determine whether to continue or stop.
11375 The next-step-expression is designed so that the do-again-test returns
11376 false when the function should no longer be repeated.
11378 The do-again-test is sometimes called the @dfn{stop condition},
11379 since it stops the repetitions when it tests false.
11381 @node Recursion with list
11382 @subsection Recursion with a List
11384 The example of a @code{while} loop that printed the elements of a list
11385 of numbers can be written recursively. Here is the code, including
11386 an expression to set the value of the variable @code{animals} to a list.
11388 If you are reading this in Info in Emacs, you can evaluate this
11389 expression directly in Info. Otherwise, you must copy the example
11390 to the @file{*scratch*} buffer and evaluate each expression there.
11391 Use @kbd{C-u C-x C-e} to evaluate the
11392 @code{(print-elements-recursively animals)} expression so that the
11393 results are printed in the buffer; otherwise the Lisp interpreter will
11394 try to squeeze the results into the one line of the echo area.
11396 Also, place your cursor immediately after the last closing parenthesis
11397 of the @code{print-elements-recursively} function, before the comment.
11398 Otherwise, the Lisp interpreter will try to evaluate the comment.
11400 @findex print-elements-recursively
11403 (setq animals '(gazelle giraffe lion tiger))
11405 (defun print-elements-recursively (list)
11406 "Print each element of LIST on a line of its own.
11408 (when list ; @r{do-again-test}
11409 (print (car list)) ; @r{body}
11410 (print-elements-recursively ; @r{recursive call}
11411 (cdr list)))) ; @r{next-step-expression}
11413 (print-elements-recursively animals)
11417 The @code{print-elements-recursively} function first tests whether
11418 there is any content in the list; if there is, the function prints the
11419 first element of the list, the @sc{car} of the list. Then the
11420 function `invokes itself', but gives itself as its argument, not the
11421 whole list, but the second and subsequent elements of the list, the
11422 @sc{cdr} of the list.
11424 Put another way, if the list is not empty, the function invokes
11425 another instance of code that is similar to the initial code, but is a
11426 different thread of execution, with different arguments than the first
11429 Put in yet another way, if the list is not empty, the first robot
11430 assembles a second robot and tells it what to do; the second robot is
11431 a different individual from the first, but is the same model.
11433 When the second evaluation occurs, the @code{when} expression is
11434 evaluated and if true, prints the first element of the list it
11435 receives as its argument (which is the second element of the original
11436 list). Then the function `calls itself' with the @sc{cdr} of the list
11437 it is invoked with, which (the second time around) is the @sc{cdr} of
11438 the @sc{cdr} of the original list.
11440 Note that although we say that the function `calls itself', what we
11441 mean is that the Lisp interpreter assembles and instructs a new
11442 instance of the program. The new instance is a clone of the first,
11443 but is a separate individual.
11445 Each time the function `invokes itself', it invokes itself on a
11446 shorter version of the original list. It creates a new instance that
11447 works on a shorter list.
11449 Eventually, the function invokes itself on an empty list. It creates
11450 a new instance whose argument is @code{nil}. The conditional expression
11451 tests the value of @code{list}. Since the value of @code{list} is
11452 @code{nil}, the @code{when} expression tests false so the then-part is
11453 not evaluated. The function as a whole then returns @code{nil}.
11456 When you evaluate the expression @code{(print-elements-recursively
11457 animals)} in the @file{*scratch*} buffer, you see this result:
11473 @node Recursive triangle function
11474 @subsection Recursion in Place of a Counter
11475 @findex triangle-recursively
11478 The @code{triangle} function described in a previous section can also
11479 be written recursively. It looks like this:
11483 (defun triangle-recursively (number)
11484 "Return the sum of the numbers 1 through NUMBER inclusive.
11486 (if (= number 1) ; @r{do-again-test}
11488 (+ number ; @r{else-part}
11489 (triangle-recursively ; @r{recursive call}
11490 (1- number))))) ; @r{next-step-expression}
11492 (triangle-recursively 7)
11497 You can install this function by evaluating it and then try it by
11498 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11499 cursor immediately after the last parenthesis of the function
11500 definition, before the comment.) The function evaluates to 28.
11502 To understand how this function works, let's consider what happens in the
11503 various cases when the function is passed 1, 2, 3, or 4 as the value of
11507 * Recursive Example arg of 1 or 2::
11508 * Recursive Example arg of 3 or 4::
11512 @node Recursive Example arg of 1 or 2
11513 @unnumberedsubsubsec An argument of 1 or 2
11516 First, what happens if the value of the argument is 1?
11518 The function has an @code{if} expression after the documentation
11519 string. It tests whether the value of @code{number} is equal to 1; if
11520 so, Emacs evaluates the then-part of the @code{if} expression, which
11521 returns the number 1 as the value of the function. (A triangle with
11522 one row has one pebble in it.)
11524 Suppose, however, that the value of the argument is 2. In this case,
11525 Emacs evaluates the else-part of the @code{if} expression.
11528 The else-part consists of an addition, the recursive call to
11529 @code{triangle-recursively} and a decrementing action; and it looks like
11533 (+ number (triangle-recursively (1- number)))
11536 When Emacs evaluates this expression, the innermost expression is
11537 evaluated first; then the other parts in sequence. Here are the steps
11541 @item Step 1 @w{ } Evaluate the innermost expression.
11543 The innermost expression is @code{(1- number)} so Emacs decrements the
11544 value of @code{number} from 2 to 1.
11546 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11548 The Lisp interpreter creates an individual instance of
11549 @code{triangle-recursively}. It does not matter that this function is
11550 contained within itself. Emacs passes the result Step 1 as the
11551 argument used by this instance of the @code{triangle-recursively}
11554 In this case, Emacs evaluates @code{triangle-recursively} with an
11555 argument of 1. This means that this evaluation of
11556 @code{triangle-recursively} returns 1.
11558 @item Step 3 @w{ } Evaluate the value of @code{number}.
11560 The variable @code{number} is the second element of the list that
11561 starts with @code{+}; its value is 2.
11563 @item Step 4 @w{ } Evaluate the @code{+} expression.
11565 The @code{+} expression receives two arguments, the first
11566 from the evaluation of @code{number} (Step 3) and the second from the
11567 evaluation of @code{triangle-recursively} (Step 2).
11569 The result of the addition is the sum of 2 plus 1, and the number 3 is
11570 returned, which is correct. A triangle with two rows has three
11574 @node Recursive Example arg of 3 or 4
11575 @unnumberedsubsubsec An argument of 3 or 4
11577 Suppose that @code{triangle-recursively} is called with an argument of
11581 @item Step 1 @w{ } Evaluate the do-again-test.
11583 The @code{if} expression is evaluated first. This is the do-again
11584 test and returns false, so the else-part of the @code{if} expression
11585 is evaluated. (Note that in this example, the do-again-test causes
11586 the function to call itself when it tests false, not when it tests
11589 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11591 The innermost expression of the else-part is evaluated, which decrements
11592 3 to 2. This is the next-step-expression.
11594 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11596 The number 2 is passed to the @code{triangle-recursively} function.
11598 We already know what happens when Emacs evaluates @code{triangle-recursively} with
11599 an argument of 2. After going through the sequence of actions described
11600 earlier, it returns a value of 3. So that is what will happen here.
11602 @item Step 4 @w{ } Evaluate the addition.
11604 3 will be passed as an argument to the addition and will be added to the
11605 number with which the function was called, which is 3.
11609 The value returned by the function as a whole will be 6.
11611 Now that we know what will happen when @code{triangle-recursively} is
11612 called with an argument of 3, it is evident what will happen if it is
11613 called with an argument of 4:
11617 In the recursive call, the evaluation of
11620 (triangle-recursively (1- 4))
11625 will return the value of evaluating
11628 (triangle-recursively 3)
11632 which is 6 and this value will be added to 4 by the addition in the
11637 The value returned by the function as a whole will be 10.
11639 Each time @code{triangle-recursively} is evaluated, it evaluates a
11640 version of itself---a different instance of itself---with a smaller
11641 argument, until the argument is small enough so that it does not
11644 Note that this particular design for a recursive function
11645 requires that operations be deferred.
11647 Before @code{(triangle-recursively 7)} can calculate its answer, it
11648 must call @code{(triangle-recursively 6)}; and before
11649 @code{(triangle-recursively 6)} can calculate its answer, it must call
11650 @code{(triangle-recursively 5)}; and so on. That is to say, the
11651 calculation that @code{(triangle-recursively 7)} makes must be
11652 deferred until @code{(triangle-recursively 6)} makes its calculation;
11653 and @code{(triangle-recursively 6)} must defer until
11654 @code{(triangle-recursively 5)} completes; and so on.
11656 If each of these instances of @code{triangle-recursively} are thought
11657 of as different robots, the first robot must wait for the second to
11658 complete its job, which must wait until the third completes, and so
11661 There is a way around this kind of waiting, which we will discuss in
11662 @ref{No Deferment, , Recursion without Deferments}.
11664 @node Recursion with cond
11665 @subsection Recursion Example Using @code{cond}
11668 The version of @code{triangle-recursively} described earlier is written
11669 with the @code{if} special form. It can also be written using another
11670 special form called @code{cond}. The name of the special form
11671 @code{cond} is an abbreviation of the word @samp{conditional}.
11673 Although the @code{cond} special form is not used as often in the
11674 Emacs Lisp sources as @code{if}, it is used often enough to justify
11678 The template for a @code{cond} expression looks like this:
11688 where the @var{body} is a series of lists.
11691 Written out more fully, the template looks like this:
11696 (@var{first-true-or-false-test} @var{first-consequent})
11697 (@var{second-true-or-false-test} @var{second-consequent})
11698 (@var{third-true-or-false-test} @var{third-consequent})
11703 When the Lisp interpreter evaluates the @code{cond} expression, it
11704 evaluates the first element (the @sc{car} or true-or-false-test) of
11705 the first expression in a series of expressions within the body of the
11708 If the true-or-false-test returns @code{nil} the rest of that
11709 expression, the consequent, is skipped and the true-or-false-test of the
11710 next expression is evaluated. When an expression is found whose
11711 true-or-false-test returns a value that is not @code{nil}, the
11712 consequent of that expression is evaluated. The consequent can be one
11713 or more expressions. If the consequent consists of more than one
11714 expression, the expressions are evaluated in sequence and the value of
11715 the last one is returned. If the expression does not have a consequent,
11716 the value of the true-or-false-test is returned.
11718 If none of the true-or-false-tests test true, the @code{cond} expression
11719 returns @code{nil}.
11722 Written using @code{cond}, the @code{triangle} function looks like this:
11726 (defun triangle-using-cond (number)
11727 (cond ((<= number 0) 0)
11730 (+ number (triangle-using-cond (1- number))))))
11735 In this example, the @code{cond} returns 0 if the number is less than or
11736 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11737 number (triangle-using-cond (1- number)))} if the number is greater than
11740 @node Recursive Patterns
11741 @subsection Recursive Patterns
11742 @cindex Recursive Patterns
11744 Here are three common recursive patterns. Each involves a list.
11745 Recursion does not need to involve lists, but Lisp is designed for lists
11746 and this provides a sense of its primal capabilities.
11755 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11756 @cindex Every, type of recursive pattern
11757 @cindex Recursive pattern: every
11759 In the @code{every} recursive pattern, an action is performed on every
11763 The basic pattern is:
11767 If a list be empty, return @code{nil}.
11769 Else, act on the beginning of the list (the @sc{car} of the list)
11772 through a recursive call by the function on the rest (the
11773 @sc{cdr}) of the list,
11775 and, optionally, combine the acted-on element, using @code{cons},
11776 with the results of acting on the rest.
11785 (defun square-each (numbers-list)
11786 "Square each of a NUMBERS LIST, recursively."
11787 (if (not numbers-list) ; do-again-test
11790 (* (car numbers-list) (car numbers-list))
11791 (square-each (cdr numbers-list))))) ; next-step-expression
11795 (square-each '(1 2 3))
11802 If @code{numbers-list} is empty, do nothing. But if it has content,
11803 construct a list combining the square of the first number in the list
11804 with the result of the recursive call.
11806 (The example follows the pattern exactly: @code{nil} is returned if
11807 the numbers' list is empty. In practice, you would write the
11808 conditional so it carries out the action when the numbers' list is not
11811 The @code{print-elements-recursively} function (@pxref{Recursion with
11812 list, , Recursion with a List}) is another example of an @code{every}
11813 pattern, except in this case, rather than bring the results together
11814 using @code{cons}, we print each element of output.
11817 The @code{print-elements-recursively} function looks like this:
11821 (setq animals '(gazelle giraffe lion tiger))
11825 (defun print-elements-recursively (list)
11826 "Print each element of LIST on a line of its own.
11828 (when list ; @r{do-again-test}
11829 (print (car list)) ; @r{body}
11830 (print-elements-recursively ; @r{recursive call}
11831 (cdr list)))) ; @r{next-step-expression}
11833 (print-elements-recursively animals)
11838 The pattern for @code{print-elements-recursively} is:
11842 When the list is empty, do nothing.
11844 But when the list has at least one element,
11847 act on the beginning of the list (the @sc{car} of the list),
11849 and make a recursive call on the rest (the @sc{cdr}) of the list.
11854 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11855 @cindex Accumulate, type of recursive pattern
11856 @cindex Recursive pattern: accumulate
11858 Another recursive pattern is called the @code{accumulate} pattern. In
11859 the @code{accumulate} recursive pattern, an action is performed on
11860 every element of a list and the result of that action is accumulated
11861 with the results of performing the action on the other elements.
11863 This is very like the `every' pattern using @code{cons}, except that
11864 @code{cons} is not used, but some other combiner.
11871 If a list be empty, return zero or some other constant.
11873 Else, act on the beginning of the list (the @sc{car} of the list),
11876 and combine that acted-on element, using @code{+} or
11877 some other combining function, with
11879 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11884 Here is an example:
11888 (defun add-elements (numbers-list)
11889 "Add the elements of NUMBERS-LIST together."
11890 (if (not numbers-list)
11892 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
11896 (add-elements '(1 2 3 4))
11901 @xref{Files List, , Making a List of Files}, for an example of the
11902 accumulate pattern.
11905 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
11906 @cindex Keep, type of recursive pattern
11907 @cindex Recursive pattern: keep
11909 A third recursive pattern is called the @code{keep} pattern.
11910 In the @code{keep} recursive pattern, each element of a list is tested;
11911 the element is acted on and the results are kept only if the element
11914 Again, this is very like the `every' pattern, except the element is
11915 skipped unless it meets a criterion.
11918 The pattern has three parts:
11922 If a list be empty, return @code{nil}.
11924 Else, if the beginning of the list (the @sc{car} of the list) passes
11928 act on that element and combine it, using @code{cons} with
11930 a recursive call by the function on the rest (the @sc{cdr}) of the list.
11933 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
11937 skip on that element,
11939 and, recursively call the function on the rest (the @sc{cdr}) of the list.
11944 Here is an example that uses @code{cond}:
11948 (defun keep-three-letter-words (word-list)
11949 "Keep three letter words in WORD-LIST."
11951 ;; First do-again-test: stop-condition
11952 ((not word-list) nil)
11954 ;; Second do-again-test: when to act
11955 ((eq 3 (length (symbol-name (car word-list))))
11956 ;; combine acted-on element with recursive call on shorter list
11957 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
11959 ;; Third do-again-test: when to skip element;
11960 ;; recursively call shorter list with next-step expression
11961 (t (keep-three-letter-words (cdr word-list)))))
11965 (keep-three-letter-words '(one two three four five six))
11966 @result{} (one two six)
11970 It goes without saying that you need not use @code{nil} as the test for
11971 when to stop; and you can, of course, combine these patterns.
11974 @subsection Recursion without Deferments
11975 @cindex Deferment in recursion
11976 @cindex Recursion without Deferments
11978 Let's consider again what happens with the @code{triangle-recursively}
11979 function. We will find that the intermediate calculations are
11980 deferred until all can be done.
11983 Here is the function definition:
11987 (defun triangle-recursively (number)
11988 "Return the sum of the numbers 1 through NUMBER inclusive.
11990 (if (= number 1) ; @r{do-again-test}
11992 (+ number ; @r{else-part}
11993 (triangle-recursively ; @r{recursive call}
11994 (1- number))))) ; @r{next-step-expression}
11998 What happens when we call this function with a argument of 7?
12000 The first instance of the @code{triangle-recursively} function adds
12001 the number 7 to the value returned by a second instance of
12002 @code{triangle-recursively}, an instance that has been passed an
12003 argument of 6. That is to say, the first calculation is:
12006 (+ 7 (triangle-recursively 6))
12010 The first instance of @code{triangle-recursively}---you may want to
12011 think of it as a little robot---cannot complete its job. It must hand
12012 off the calculation for @code{(triangle-recursively 6)} to a second
12013 instance of the program, to a second robot. This second individual is
12014 completely different from the first one; it is, in the jargon, a
12015 `different instantiation'. Or, put another way, it is a different
12016 robot. It is the same model as the first; it calculates triangle
12017 numbers recursively; but it has a different serial number.
12019 And what does @code{(triangle-recursively 6)} return? It returns the
12020 number 6 added to the value returned by evaluating
12021 @code{triangle-recursively} with an argument of 5. Using the robot
12022 metaphor, it asks yet another robot to help it.
12028 (+ 7 6 (triangle-recursively 5))
12032 And what happens next?
12035 (+ 7 6 5 (triangle-recursively 4))
12038 Each time @code{triangle-recursively} is called, except for the last
12039 time, it creates another instance of the program---another robot---and
12040 asks it to make a calculation.
12043 Eventually, the full addition is set up and performed:
12049 This design for the function defers the calculation of the first step
12050 until the second can be done, and defers that until the third can be
12051 done, and so on. Each deferment means the computer must remember what
12052 is being waited on. This is not a problem when there are only a few
12053 steps, as in this example. But it can be a problem when there are
12056 @node No deferment solution
12057 @subsection No Deferment Solution
12058 @cindex No deferment solution
12059 @cindex Defermentless solution
12060 @cindex Solution without deferment
12062 The solution to the problem of deferred operations is to write in a
12063 manner that does not defer operations@footnote{The phrase @dfn{tail
12064 recursive} is used to describe such a process, one that uses
12065 `constant space'.}. This requires
12066 writing to a different pattern, often one that involves writing two
12067 function definitions, an `initialization' function and a `helper'
12070 The `initialization' function sets up the job; the `helper' function
12074 Here are the two function definitions for adding up numbers. They are
12075 so simple, I find them hard to understand.
12079 (defun triangle-initialization (number)
12080 "Return the sum of the numbers 1 through NUMBER inclusive.
12081 This is the `initialization' component of a two function
12082 duo that uses recursion."
12083 (triangle-recursive-helper 0 0 number))
12089 (defun triangle-recursive-helper (sum counter number)
12090 "Return SUM, using COUNTER, through NUMBER inclusive.
12091 This is the `helper' component of a two function duo
12092 that uses recursion."
12093 (if (> counter number)
12095 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12096 (1+ counter) ; @r{counter}
12097 number))) ; @r{number}
12102 Install both function definitions by evaluating them, then call
12103 @code{triangle-initialization} with 2 rows:
12107 (triangle-initialization 2)
12112 The `initialization' function calls the first instance of the `helper'
12113 function with three arguments: zero, zero, and a number which is the
12114 number of rows in the triangle.
12116 The first two arguments passed to the `helper' function are
12117 initialization values. These values are changed when
12118 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12119 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12120 process that is iterative in a procedure that is recursive. The
12121 process is called iterative because the computer need only record the
12122 three values, @code{sum}, @code{counter}, and @code{number}; the
12123 procedure is recursive because the function `calls itself'. On the
12124 other hand, both the process and the procedure used by
12125 @code{triangle-recursively} are called recursive. The word
12126 `recursive' has different meanings in the two contexts.}
12128 Let's see what happens when we have a triangle that has one row. (This
12129 triangle will have one pebble in it!)
12132 @code{triangle-initialization} will call its helper with
12133 the arguments @w{@code{0 0 1}}. That function will run the conditional
12134 test whether @code{(> counter number)}:
12142 and find that the result is false, so it will invoke
12143 the else-part of the @code{if} clause:
12147 (triangle-recursive-helper
12148 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12149 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12150 number) ; @r{number stays the same}
12156 which will first compute:
12160 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12161 (1+ 0) ; @r{counter}
12165 (triangle-recursive-helper 0 1 1)
12169 Again, @code{(> counter number)} will be false, so again, the Lisp
12170 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12171 new instance with new arguments.
12174 This new instance will be;
12178 (triangle-recursive-helper
12179 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12180 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12181 number) ; @r{number stays the same}
12185 (triangle-recursive-helper 1 2 1)
12189 In this case, the @code{(> counter number)} test will be true! So the
12190 instance will return the value of the sum, which will be 1, as
12193 Now, let's pass @code{triangle-initialization} an argument
12194 of 2, to find out how many pebbles there are in a triangle with two rows.
12196 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12199 In stages, the instances called will be:
12203 @r{sum counter number}
12204 (triangle-recursive-helper 0 1 2)
12206 (triangle-recursive-helper 1 2 2)
12208 (triangle-recursive-helper 3 3 2)
12212 When the last instance is called, the @code{(> counter number)} test
12213 will be true, so the instance will return the value of @code{sum},
12216 This kind of pattern helps when you are writing functions that can use
12217 many resources in a computer.
12220 @node Looping exercise
12221 @section Looping Exercise
12225 Write a function similar to @code{triangle} in which each row has a
12226 value which is the square of the row number. Use a @code{while} loop.
12229 Write a function similar to @code{triangle} that multiplies instead of
12233 Rewrite these two functions recursively. Rewrite these functions
12236 @c comma in printed title causes problem in Info cross reference
12238 Write a function for Texinfo mode that creates an index entry at the
12239 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12240 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12241 written in Texinfo.)
12243 Many of the functions you will need are described in two of the
12244 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12245 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12246 @code{forward-paragraph} to put the index entry at the beginning of
12247 the paragraph, you will have to use @w{@kbd{C-h f}}
12248 (@code{describe-function}) to find out how to make the command go
12251 For more information, see
12253 @ref{Indicating, , Indicating Definitions, texinfo}.
12256 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12257 a Texinfo manual in the current directory. Or, if you are on the
12259 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12262 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12263 Documentation Format}.
12267 @node Regexp Search
12268 @chapter Regular Expression Searches
12269 @cindex Searches, illustrating
12270 @cindex Regular expression searches
12271 @cindex Patterns, searching for
12272 @cindex Motion by sentence and paragraph
12273 @cindex Sentences, movement by
12274 @cindex Paragraphs, movement by
12276 Regular expression searches are used extensively in GNU Emacs. The
12277 two functions, @code{forward-sentence} and @code{forward-paragraph},
12278 illustrate these searches well. They use regular expressions to find
12279 where to move point. The phrase `regular expression' is often written
12282 Regular expression searches are described in @ref{Regexp Search, ,
12283 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12284 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12285 Manual}. In writing this chapter, I am presuming that you have at
12286 least a mild acquaintance with them. The major point to remember is
12287 that regular expressions permit you to search for patterns as well as
12288 for literal strings of characters. For example, the code in
12289 @code{forward-sentence} searches for the pattern of possible
12290 characters that could mark the end of a sentence, and moves point to
12293 Before looking at the code for the @code{forward-sentence} function, it
12294 is worth considering what the pattern that marks the end of a sentence
12295 must be. The pattern is discussed in the next section; following that
12296 is a description of the regular expression search function,
12297 @code{re-search-forward}. The @code{forward-sentence} function
12298 is described in the section following. Finally, the
12299 @code{forward-paragraph} function is described in the last section of
12300 this chapter. @code{forward-paragraph} is a complex function that
12301 introduces several new features.
12304 * sentence-end:: The regular expression for @code{sentence-end}.
12305 * re-search-forward:: Very similar to @code{search-forward}.
12306 * forward-sentence:: A straightforward example of regexp search.
12307 * forward-paragraph:: A somewhat complex example.
12308 * etags:: How to create your own @file{TAGS} table.
12310 * re-search Exercises::
12314 @section The Regular Expression for @code{sentence-end}
12315 @findex sentence-end
12317 The symbol @code{sentence-end} is bound to the pattern that marks the
12318 end of a sentence. What should this regular expression be?
12320 Clearly, a sentence may be ended by a period, a question mark, or an
12321 exclamation mark. Indeed, in English, only clauses that end with one
12322 of those three characters should be considered the end of a sentence.
12323 This means that the pattern should include the character set:
12329 However, we do not want @code{forward-sentence} merely to jump to a
12330 period, a question mark, or an exclamation mark, because such a character
12331 might be used in the middle of a sentence. A period, for example, is
12332 used after abbreviations. So other information is needed.
12334 According to convention, you type two spaces after every sentence, but
12335 only one space after a period, a question mark, or an exclamation mark in
12336 the body of a sentence. So a period, a question mark, or an exclamation
12337 mark followed by two spaces is a good indicator of an end of sentence.
12338 However, in a file, the two spaces may instead be a tab or the end of a
12339 line. This means that the regular expression should include these three
12340 items as alternatives.
12343 This group of alternatives will look like this:
12354 Here, @samp{$} indicates the end of the line, and I have pointed out
12355 where the tab and two spaces are inserted in the expression. Both are
12356 inserted by putting the actual characters into the expression.
12358 Two backslashes, @samp{\\}, are required before the parentheses and
12359 vertical bars: the first backslash quotes the following backslash in
12360 Emacs; and the second indicates that the following character, the
12361 parenthesis or the vertical bar, is special.
12364 Also, a sentence may be followed by one or more carriage returns, like
12375 Like tabs and spaces, a carriage return is inserted into a regular
12376 expression by inserting it literally. The asterisk indicates that the
12377 @key{RET} is repeated zero or more times.
12379 But a sentence end does not consist only of a period, a question mark or
12380 an exclamation mark followed by appropriate space: a closing quotation
12381 mark or a closing brace of some kind may precede the space. Indeed more
12382 than one such mark or brace may precede the space. These require a
12383 expression that looks like this:
12389 In this expression, the first @samp{]} is the first character in the
12390 expression; the second character is @samp{"}, which is preceded by a
12391 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12392 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12394 All this suggests what the regular expression pattern for matching the
12395 end of a sentence should be; and, indeed, if we evaluate
12396 @code{sentence-end} we find that it returns the following value:
12401 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12407 (Well, not in GNU Emacs 22; that is because of an effort to make the
12408 process simpler and to handle more glyphs and languages. When the
12409 value of @code{sentence-end} is @code{nil}, then use the value defined
12410 by the function @code{sentence-end}. (Here is a use of the difference
12411 between a value and a function in Emacs Lisp.) The function returns a
12412 value constructed from the variables @code{sentence-end-base},
12413 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12414 and @code{sentence-end-without-space}. The critical variable is
12415 @code{sentence-end-base}; its global value is similar to the one
12416 described above but it also contains two additional quotation marks.
12417 These have differing degrees of curliness. The
12418 @code{sentence-end-without-period} variable, when true, tells Emacs
12419 that a sentence may end without a period, such as text in Thai.)
12423 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12424 literally in the pattern.)
12426 This regular expression can be deciphered as follows:
12430 The first part of the pattern is the three characters, a period, a question
12431 mark and an exclamation mark, within square brackets. The pattern must
12432 begin with one or other of these characters.
12435 The second part of the pattern is the group of closing braces and
12436 quotation marks, which can appear zero or more times. These may follow
12437 the period, question mark or exclamation mark. In a regular expression,
12438 the backslash, @samp{\}, followed by the double quotation mark,
12439 @samp{"}, indicates the class of string-quote characters. Usually, the
12440 double quotation mark is the only character in this class. The
12441 asterisk, @samp{*}, indicates that the items in the previous group (the
12442 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12445 @item \\($\\| \\| \\)
12446 The third part of the pattern is one or other of: either the end of a
12447 line, or two blank spaces, or a tab. The double back-slashes are used
12448 to prevent Emacs from reading the parentheses and vertical bars as part
12449 of the search pattern; the parentheses are used to mark the group and
12450 the vertical bars are used to indicated that the patterns to either side
12451 of them are alternatives. The dollar sign is used to indicate the end
12452 of a line and both the two spaces and the tab are each inserted as is to
12453 indicate what they are.
12456 Finally, the last part of the pattern indicates that the end of the line
12457 or the whitespace following the period, question mark or exclamation
12458 mark may, but need not, be followed by one or more carriage returns. In
12459 the pattern, the carriage return is inserted as an actual carriage
12460 return between square brackets but here it is shown as @key{RET}.
12464 @node re-search-forward
12465 @section The @code{re-search-forward} Function
12466 @findex re-search-forward
12468 The @code{re-search-forward} function is very like the
12469 @code{search-forward} function. (@xref{search-forward, , The
12470 @code{search-forward} Function}.)
12472 @code{re-search-forward} searches for a regular expression. If the
12473 search is successful, it leaves point immediately after the last
12474 character in the target. If the search is backwards, it leaves point
12475 just before the first character in the target. You may tell
12476 @code{re-search-forward} to return @code{t} for true. (Moving point
12477 is therefore a `side effect'.)
12479 Like @code{search-forward}, the @code{re-search-forward} function takes
12484 The first argument is the regular expression that the function searches
12485 for. The regular expression will be a string between quotation marks.
12488 The optional second argument limits how far the function will search; it is a
12489 bound, which is specified as a position in the buffer.
12492 The optional third argument specifies how the function responds to
12493 failure: @code{nil} as the third argument causes the function to
12494 signal an error (and print a message) when the search fails; any other
12495 value causes it to return @code{nil} if the search fails and @code{t}
12496 if the search succeeds.
12499 The optional fourth argument is the repeat count. A negative repeat
12500 count causes @code{re-search-forward} to search backwards.
12504 The template for @code{re-search-forward} looks like this:
12508 (re-search-forward "@var{regular-expression}"
12509 @var{limit-of-search}
12510 @var{what-to-do-if-search-fails}
12511 @var{repeat-count})
12515 The second, third, and fourth arguments are optional. However, if you
12516 want to pass a value to either or both of the last two arguments, you
12517 must also pass a value to all the preceding arguments. Otherwise, the
12518 Lisp interpreter will mistake which argument you are passing the value
12522 In the @code{forward-sentence} function, the regular expression will be
12523 the value of the variable @code{sentence-end}. In simple form, that is:
12527 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12533 The limit of the search will be the end of the paragraph (since a
12534 sentence cannot go beyond a paragraph). If the search fails, the
12535 function will return @code{nil}; and the repeat count will be provided
12536 by the argument to the @code{forward-sentence} function.
12538 @node forward-sentence
12539 @section @code{forward-sentence}
12540 @findex forward-sentence
12542 The command to move the cursor forward a sentence is a straightforward
12543 illustration of how to use regular expression searches in Emacs Lisp.
12544 Indeed, the function looks longer and more complicated than it is; this
12545 is because the function is designed to go backwards as well as forwards;
12546 and, optionally, over more than one sentence. The function is usually
12547 bound to the key command @kbd{M-e}.
12550 * Complete forward-sentence::
12551 * fwd-sentence while loops:: Two @code{while} loops.
12552 * fwd-sentence re-search:: A regular expression search.
12556 @node Complete forward-sentence
12557 @unnumberedsubsec Complete @code{forward-sentence} function definition
12561 Here is the code for @code{forward-sentence}:
12566 (defun forward-sentence (&optional arg)
12567 "Move forward to next `sentence-end'. With argument, repeat.
12568 With negative argument, move backward repeatedly to `sentence-beginning'.
12570 The variable `sentence-end' is a regular expression that matches ends of
12571 sentences. Also, every paragraph boundary terminates sentences as well."
12575 (or arg (setq arg 1))
12576 (let ((opoint (point))
12577 (sentence-end (sentence-end)))
12579 (let ((pos (point))
12580 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12581 (if (and (re-search-backward sentence-end par-beg t)
12582 (or (< (match-end 0) pos)
12583 (re-search-backward sentence-end par-beg t)))
12584 (goto-char (match-end 0))
12585 (goto-char par-beg)))
12586 (setq arg (1+ arg)))
12590 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12591 (if (re-search-forward sentence-end par-end t)
12592 (skip-chars-backward " \t\n")
12593 (goto-char par-end)))
12594 (setq arg (1- arg)))
12595 (constrain-to-field nil opoint t)))
12603 (defun forward-sentence (&optional arg)
12604 "Move forward to next sentence-end. With argument, repeat.
12605 With negative argument, move backward repeatedly to sentence-beginning.
12606 Sentence ends are identified by the value of sentence-end
12607 treated as a regular expression. Also, every paragraph boundary
12608 terminates sentences as well."
12612 (or arg (setq arg 1))
12615 (save-excursion (start-of-paragraph-text) (point))))
12616 (if (re-search-backward
12617 (concat sentence-end "[^ \t\n]") par-beg t)
12618 (goto-char (1- (match-end 0)))
12619 (goto-char par-beg)))
12620 (setq arg (1+ arg)))
12623 (save-excursion (end-of-paragraph-text) (point))))
12624 (if (re-search-forward sentence-end par-end t)
12625 (skip-chars-backward " \t\n")
12626 (goto-char par-end)))
12627 (setq arg (1- arg))))
12632 The function looks long at first sight and it is best to look at its
12633 skeleton first, and then its muscle. The way to see the skeleton is to
12634 look at the expressions that start in the left-most columns:
12638 (defun forward-sentence (&optional arg)
12639 "@var{documentation}@dots{}"
12641 (or arg (setq arg 1))
12642 (let ((opoint (point)) (sentence-end (sentence-end)))
12644 (let ((pos (point))
12645 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12646 @var{rest-of-body-of-while-loop-when-going-backwards}
12648 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12649 @var{rest-of-body-of-while-loop-when-going-forwards}
12650 @var{handle-forms-and-equivalent}
12654 This looks much simpler! The function definition consists of
12655 documentation, an @code{interactive} expression, an @code{or}
12656 expression, a @code{let} expression, and @code{while} loops.
12658 Let's look at each of these parts in turn.
12660 We note that the documentation is thorough and understandable.
12662 The function has an @code{interactive "p"} declaration. This means
12663 that the processed prefix argument, if any, is passed to the
12664 function as its argument. (This will be a number.) If the function
12665 is not passed an argument (it is optional) then the argument
12666 @code{arg} will be bound to 1.
12668 When @code{forward-sentence} is called non-interactively without an
12669 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12670 handles this. What it does is either leave the value of @code{arg} as
12671 it is, but only if @code{arg} is bound to a value; or it sets the
12672 value of @code{arg} to 1, in the case when @code{arg} is bound to
12675 Next is a @code{let}. That specifies the values of two local
12676 variables, @code{point} and @code{sentence-end}. The local value of
12677 point, from before the search, is used in the
12678 @code{constrain-to-field} function which handles forms and
12679 equivalents. The @code{sentence-end} variable is set by the
12680 @code{sentence-end} function.
12682 @node fwd-sentence while loops
12683 @unnumberedsubsec The @code{while} loops
12685 Two @code{while} loops follow. The first @code{while} has a
12686 true-or-false-test that tests true if the prefix argument for
12687 @code{forward-sentence} is a negative number. This is for going
12688 backwards. The body of this loop is similar to the body of the second
12689 @code{while} clause, but it is not exactly the same. We will skip
12690 this @code{while} loop and concentrate on the second @code{while}
12694 The second @code{while} loop is for moving point forward. Its skeleton
12699 (while (> arg 0) ; @r{true-or-false-test}
12701 (if (@var{true-or-false-test})
12704 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12708 The @code{while} loop is of the decrementing kind.
12709 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12710 has a true-or-false-test that tests true so long as the counter (in
12711 this case, the variable @code{arg}) is greater than zero; and it has a
12712 decrementer that subtracts 1 from the value of the counter every time
12715 If no prefix argument is given to @code{forward-sentence}, which is
12716 the most common way the command is used, this @code{while} loop will
12717 run once, since the value of @code{arg} will be 1.
12719 The body of the @code{while} loop consists of a @code{let} expression,
12720 which creates and binds a local variable, and has, as its body, an
12721 @code{if} expression.
12724 The body of the @code{while} loop looks like this:
12729 (save-excursion (end-of-paragraph-text) (point))))
12730 (if (re-search-forward sentence-end par-end t)
12731 (skip-chars-backward " \t\n")
12732 (goto-char par-end)))
12736 The @code{let} expression creates and binds the local variable
12737 @code{par-end}. As we shall see, this local variable is designed to
12738 provide a bound or limit to the regular expression search. If the
12739 search fails to find a proper sentence ending in the paragraph, it will
12740 stop on reaching the end of the paragraph.
12742 But first, let us examine how @code{par-end} is bound to the value of
12743 the end of the paragraph. What happens is that the @code{let} sets the
12744 value of @code{par-end} to the value returned when the Lisp interpreter
12745 evaluates the expression
12749 (save-excursion (end-of-paragraph-text) (point))
12754 In this expression, @code{(end-of-paragraph-text)} moves point to the
12755 end of the paragraph, @code{(point)} returns the value of point, and then
12756 @code{save-excursion} restores point to its original position. Thus,
12757 the @code{let} binds @code{par-end} to the value returned by the
12758 @code{save-excursion} expression, which is the position of the end of
12759 the paragraph. (The @code{end-of-paragraph-text} function uses
12760 @code{forward-paragraph}, which we will discuss shortly.)
12763 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12764 expression that looks like this:
12768 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12769 (skip-chars-backward " \t\n") ; @r{then-part}
12770 (goto-char par-end))) ; @r{else-part}
12774 The @code{if} tests whether its first argument is true and if so,
12775 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12776 evaluates the else-part. The true-or-false-test of the @code{if}
12777 expression is the regular expression search.
12779 It may seem odd to have what looks like the `real work' of
12780 the @code{forward-sentence} function buried here, but this is a common
12781 way this kind of operation is carried out in Lisp.
12783 @node fwd-sentence re-search
12784 @unnumberedsubsec The regular expression search
12786 The @code{re-search-forward} function searches for the end of the
12787 sentence, that is, for the pattern defined by the @code{sentence-end}
12788 regular expression. If the pattern is found---if the end of the sentence is
12789 found---then the @code{re-search-forward} function does two things:
12793 The @code{re-search-forward} function carries out a side effect, which
12794 is to move point to the end of the occurrence found.
12797 The @code{re-search-forward} function returns a value of true. This is
12798 the value received by the @code{if}, and means that the search was
12803 The side effect, the movement of point, is completed before the
12804 @code{if} function is handed the value returned by the successful
12805 conclusion of the search.
12807 When the @code{if} function receives the value of true from a successful
12808 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12809 which is the expression @code{(skip-chars-backward " \t\n")}. This
12810 expression moves backwards over any blank spaces, tabs or carriage
12811 returns until a printed character is found and then leaves point after
12812 the character. Since point has already been moved to the end of the
12813 pattern that marks the end of the sentence, this action leaves point
12814 right after the closing printed character of the sentence, which is
12817 On the other hand, if the @code{re-search-forward} function fails to
12818 find a pattern marking the end of the sentence, the function returns
12819 false. The false then causes the @code{if} to evaluate its third
12820 argument, which is @code{(goto-char par-end)}: it moves point to the
12821 end of the paragraph.
12823 (And if the text is in a form or equivalent, and point may not move
12824 fully, then the @code{constrain-to-field} function comes into play.)
12826 Regular expression searches are exceptionally useful and the pattern
12827 illustrated by @code{re-search-forward}, in which the search is the
12828 test of an @code{if} expression, is handy. You will see or write code
12829 incorporating this pattern often.
12831 @node forward-paragraph
12832 @section @code{forward-paragraph}: a Goldmine of Functions
12833 @findex forward-paragraph
12837 (defun forward-paragraph (&optional arg)
12838 "Move forward to end of paragraph.
12839 With argument ARG, do it ARG times;
12840 a negative argument ARG = -N means move backward N paragraphs.
12842 A line which `paragraph-start' matches either separates paragraphs
12843 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12844 A paragraph end is the beginning of a line which is not part of the paragraph
12845 to which the end of the previous line belongs, or the end of the buffer.
12846 Returns the count of paragraphs left to move."
12848 (or arg (setq arg 1))
12849 (let* ((opoint (point))
12850 (fill-prefix-regexp
12851 (and fill-prefix (not (equal fill-prefix ""))
12852 (not paragraph-ignore-fill-prefix)
12853 (regexp-quote fill-prefix)))
12854 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12855 ;; These regexps shouldn't be anchored, because we look for them
12856 ;; starting at the left-margin. This allows paragraph commands to
12857 ;; work normally with indented text.
12858 ;; This hack will not find problem cases like "whatever\\|^something".
12859 (parstart (if (and (not (equal "" paragraph-start))
12860 (equal ?^ (aref paragraph-start 0)))
12861 (substring paragraph-start 1)
12863 (parsep (if (and (not (equal "" paragraph-separate))
12864 (equal ?^ (aref paragraph-separate 0)))
12865 (substring paragraph-separate 1)
12866 paragraph-separate))
12868 (if fill-prefix-regexp
12869 (concat parsep "\\|"
12870 fill-prefix-regexp "[ \t]*$")
12872 ;; This is used for searching.
12873 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
12875 (while (and (< arg 0) (not (bobp)))
12876 (if (and (not (looking-at parsep))
12877 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
12878 (looking-at parsep))
12879 (setq arg (1+ arg))
12880 (setq start (point))
12881 ;; Move back over paragraph-separating lines.
12882 (forward-char -1) (beginning-of-line)
12883 (while (and (not (bobp))
12884 (progn (move-to-left-margin)
12885 (looking-at parsep)))
12889 (setq arg (1+ arg))
12890 ;; Go to end of the previous (non-separating) line.
12892 ;; Search back for line that starts or separates paragraphs.
12893 (if (if fill-prefix-regexp
12894 ;; There is a fill prefix; it overrides parstart.
12895 (let (multiple-lines)
12896 (while (and (progn (beginning-of-line) (not (bobp)))
12897 (progn (move-to-left-margin)
12898 (not (looking-at parsep)))
12899 (looking-at fill-prefix-regexp))
12900 (unless (= (point) start)
12901 (setq multiple-lines t))
12903 (move-to-left-margin)
12904 ;; This deleted code caused a long hanging-indent line
12905 ;; not to be filled together with the following lines.
12906 ;; ;; Don't move back over a line before the paragraph
12907 ;; ;; which doesn't start with fill-prefix
12908 ;; ;; unless that is the only line we've moved over.
12909 ;; (and (not (looking-at fill-prefix-regexp))
12911 ;; (forward-line 1))
12913 (while (and (re-search-backward sp-parstart nil 1)
12914 (setq found-start t)
12915 ;; Found a candidate, but need to check if it is a
12917 (progn (setq start (point))
12918 (move-to-left-margin)
12919 (not (looking-at parsep)))
12920 (not (and (looking-at parstart)
12921 (or (not use-hard-newlines)
12924 (1- start) 'hard)))))
12925 (setq found-start nil)
12930 ;; Move forward over paragraph separators.
12931 ;; We know this cannot reach the place we started
12932 ;; because we know we moved back over a non-separator.
12933 (while (and (not (eobp))
12934 (progn (move-to-left-margin)
12935 (looking-at parsep)))
12937 ;; If line before paragraph is just margin, back up to there.
12939 (if (> (current-column) (current-left-margin))
12941 (skip-chars-backward " \t")
12943 (forward-line 1))))
12944 ;; No starter or separator line => use buffer beg.
12945 (goto-char (point-min))))))
12947 (while (and (> arg 0) (not (eobp)))
12948 ;; Move forward over separator lines...
12949 (while (and (not (eobp))
12950 (progn (move-to-left-margin) (not (eobp)))
12951 (looking-at parsep))
12953 (unless (eobp) (setq arg (1- arg)))
12954 ;; ... and one more line.
12956 (if fill-prefix-regexp
12957 ;; There is a fill prefix; it overrides parstart.
12958 (while (and (not (eobp))
12959 (progn (move-to-left-margin) (not (eobp)))
12960 (not (looking-at parsep))
12961 (looking-at fill-prefix-regexp))
12963 (while (and (re-search-forward sp-parstart nil 1)
12964 (progn (setq start (match-beginning 0))
12967 (progn (move-to-left-margin)
12968 (not (looking-at parsep)))
12969 (or (not (looking-at parstart))
12970 (and use-hard-newlines
12971 (not (get-text-property (1- start) 'hard)))))
12973 (if (< (point) (point-max))
12974 (goto-char start))))
12975 (constrain-to-field nil opoint t)
12976 ;; Return the number of steps that could not be done.
12980 The @code{forward-paragraph} function moves point forward to the end
12981 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
12982 number of functions that are important in themselves, including
12983 @code{let*}, @code{match-beginning}, and @code{looking-at}.
12985 The function definition for @code{forward-paragraph} is considerably
12986 longer than the function definition for @code{forward-sentence}
12987 because it works with a paragraph, each line of which may begin with a
12990 A fill prefix consists of a string of characters that are repeated at
12991 the beginning of each line. For example, in Lisp code, it is a
12992 convention to start each line of a paragraph-long comment with
12993 @samp{;;; }. In Text mode, four blank spaces make up another common
12994 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
12995 emacs, The GNU Emacs Manual}, for more information about fill
12998 The existence of a fill prefix means that in addition to being able to
12999 find the end of a paragraph whose lines begin on the left-most
13000 column, the @code{forward-paragraph} function must be able to find the
13001 end of a paragraph when all or many of the lines in the buffer begin
13002 with the fill prefix.
13004 Moreover, it is sometimes practical to ignore a fill prefix that
13005 exists, especially when blank lines separate paragraphs.
13006 This is an added complication.
13009 * forward-paragraph in brief:: Key parts of the function definition.
13010 * fwd-para let:: The @code{let*} expression.
13011 * fwd-para while:: The forward motion @code{while} loop.
13015 @node forward-paragraph in brief
13016 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13019 Rather than print all of the @code{forward-paragraph} function, we
13020 will only print parts of it. Read without preparation, the function
13024 In outline, the function looks like this:
13028 (defun forward-paragraph (&optional arg)
13029 "@var{documentation}@dots{}"
13031 (or arg (setq arg 1))
13034 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13036 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13041 The first parts of the function are routine: the function's argument
13042 list consists of one optional argument. Documentation follows.
13044 The lower case @samp{p} in the @code{interactive} declaration means
13045 that the processed prefix argument, if any, is passed to the function.
13046 This will be a number, and is the repeat count of how many paragraphs
13047 point will move. The @code{or} expression in the next line handles
13048 the common case when no argument is passed to the function, which occurs
13049 if the function is called from other code rather than interactively.
13050 This case was described earlier. (@xref{forward-sentence, The
13051 @code{forward-sentence} function}.) Now we reach the end of the
13052 familiar part of this function.
13055 @unnumberedsubsec The @code{let*} expression
13057 The next line of the @code{forward-paragraph} function begins a
13058 @code{let*} expression. This is a different than @code{let}. The
13059 symbol is @code{let*} not @code{let}.
13061 The @code{let*} special form is like @code{let} except that Emacs sets
13062 each variable in sequence, one after another, and variables in the
13063 latter part of the varlist can make use of the values to which Emacs
13064 set variables in the earlier part of the varlist.
13067 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13070 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13072 In the @code{let*} expression in this function, Emacs binds a total of
13073 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13074 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13075 @code{found-start}.
13077 The variable @code{parsep} appears twice, first, to remove instances
13078 of @samp{^}, and second, to handle fill prefixes.
13080 The variable @code{opoint} is just the value of @code{point}. As you
13081 can guess, it is used in a @code{constrain-to-field} expression, just
13082 as in @code{forward-sentence}.
13084 The variable @code{fill-prefix-regexp} is set to the value returned by
13085 evaluating the following list:
13090 (not (equal fill-prefix ""))
13091 (not paragraph-ignore-fill-prefix)
13092 (regexp-quote fill-prefix))
13097 This is an expression whose first element is the @code{and} special form.
13099 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13100 function}), the @code{and} special form evaluates each of its
13101 arguments until one of the arguments returns a value of @code{nil}, in
13102 which case the @code{and} expression returns @code{nil}; however, if
13103 none of the arguments returns a value of @code{nil}, the value
13104 resulting from evaluating the last argument is returned. (Since such
13105 a value is not @code{nil}, it is considered true in Lisp.) In other
13106 words, an @code{and} expression returns a true value only if all its
13107 arguments are true.
13110 In this case, the variable @code{fill-prefix-regexp} is bound to a
13111 non-@code{nil} value only if the following four expressions produce a
13112 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13113 @code{fill-prefix-regexp} is bound to @code{nil}.
13117 When this variable is evaluated, the value of the fill prefix, if any,
13118 is returned. If there is no fill prefix, this variable returns
13121 @item (not (equal fill-prefix "")
13122 This expression checks whether an existing fill prefix is an empty
13123 string, that is, a string with no characters in it. An empty string is
13124 not a useful fill prefix.
13126 @item (not paragraph-ignore-fill-prefix)
13127 This expression returns @code{nil} if the variable
13128 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13129 true value such as @code{t}.
13131 @item (regexp-quote fill-prefix)
13132 This is the last argument to the @code{and} special form. If all the
13133 arguments to the @code{and} are true, the value resulting from
13134 evaluating this expression will be returned by the @code{and} expression
13135 and bound to the variable @code{fill-prefix-regexp},
13138 @findex regexp-quote
13140 The result of evaluating this @code{and} expression successfully is that
13141 @code{fill-prefix-regexp} will be bound to the value of
13142 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13143 What @code{regexp-quote} does is read a string and return a regular
13144 expression that will exactly match the string and match nothing else.
13145 This means that @code{fill-prefix-regexp} will be set to a value that
13146 will exactly match the fill prefix if the fill prefix exists.
13147 Otherwise, the variable will be set to @code{nil}.
13149 The next two local variables in the @code{let*} expression are
13150 designed to remove instances of @samp{^} from @code{parstart} and
13151 @code{parsep}, the local variables which indicate the paragraph start
13152 and the paragraph separator. The next expression sets @code{parsep}
13153 again. That is to handle fill prefixes.
13155 This is the setting that requires the definition call @code{let*}
13156 rather than @code{let}. The true-or-false-test for the @code{if}
13157 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13158 @code{nil} or some other value.
13160 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13161 the else-part of the @code{if} expression and binds @code{parsep} to
13162 its local value. (@code{parsep} is a regular expression that matches
13163 what separates paragraphs.)
13165 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13166 the then-part of the @code{if} expression and binds @code{parsep} to a
13167 regular expression that includes the @code{fill-prefix-regexp} as part
13170 Specifically, @code{parsep} is set to the original value of the
13171 paragraph separate regular expression concatenated with an alternative
13172 expression that consists of the @code{fill-prefix-regexp} followed by
13173 optional whitespace to the end of the line. The whitespace is defined
13174 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13175 regexp as an alternative to @code{parsep}.
13177 According to a comment in the code, the next local variable,
13178 @code{sp-parstart}, is used for searching, and then the final two,
13179 @code{start} and @code{found-start}, are set to @code{nil}.
13181 Now we get into the body of the @code{let*}. The first part of the body
13182 of the @code{let*} deals with the case when the function is given a
13183 negative argument and is therefore moving backwards. We will skip this
13186 @node fwd-para while
13187 @unnumberedsubsec The forward motion @code{while} loop
13189 The second part of the body of the @code{let*} deals with forward
13190 motion. It is a @code{while} loop that repeats itself so long as the
13191 value of @code{arg} is greater than zero. In the most common use of
13192 the function, the value of the argument is 1, so the body of the
13193 @code{while} loop is evaluated exactly once, and the cursor moves
13194 forward one paragraph.
13197 (while (and (> arg 0) (not (eobp)))
13199 ;; Move forward over separator lines...
13200 (while (and (not (eobp))
13201 (progn (move-to-left-margin) (not (eobp)))
13202 (looking-at parsep))
13204 (unless (eobp) (setq arg (1- arg)))
13205 ;; ... and one more line.
13208 (if fill-prefix-regexp
13209 ;; There is a fill prefix; it overrides parstart.
13210 (while (and (not (eobp))
13211 (progn (move-to-left-margin) (not (eobp)))
13212 (not (looking-at parsep))
13213 (looking-at fill-prefix-regexp))
13216 (while (and (re-search-forward sp-parstart nil 1)
13217 (progn (setq start (match-beginning 0))
13220 (progn (move-to-left-margin)
13221 (not (looking-at parsep)))
13222 (or (not (looking-at parstart))
13223 (and use-hard-newlines
13224 (not (get-text-property (1- start) 'hard)))))
13227 (if (< (point) (point-max))
13228 (goto-char start))))
13231 This part handles three situations: when point is between paragraphs,
13232 when there is a fill prefix and when there is no fill prefix.
13235 The @code{while} loop looks like this:
13239 ;; @r{going forwards and not at the end of the buffer}
13240 (while (and (> arg 0) (not (eobp)))
13242 ;; @r{between paragraphs}
13243 ;; Move forward over separator lines...
13244 (while (and (not (eobp))
13245 (progn (move-to-left-margin) (not (eobp)))
13246 (looking-at parsep))
13248 ;; @r{This decrements the loop}
13249 (unless (eobp) (setq arg (1- arg)))
13250 ;; ... and one more line.
13255 (if fill-prefix-regexp
13256 ;; There is a fill prefix; it overrides parstart;
13257 ;; we go forward line by line
13258 (while (and (not (eobp))
13259 (progn (move-to-left-margin) (not (eobp)))
13260 (not (looking-at parsep))
13261 (looking-at fill-prefix-regexp))
13266 ;; There is no fill prefix;
13267 ;; we go forward character by character
13268 (while (and (re-search-forward sp-parstart nil 1)
13269 (progn (setq start (match-beginning 0))
13272 (progn (move-to-left-margin)
13273 (not (looking-at parsep)))
13274 (or (not (looking-at parstart))
13275 (and use-hard-newlines
13276 (not (get-text-property (1- start) 'hard)))))
13281 ;; and if there is no fill prefix and if we are not at the end,
13282 ;; go to whatever was found in the regular expression search
13284 (if (< (point) (point-max))
13285 (goto-char start))))
13290 We can see that this is a decrementing counter @code{while} loop,
13291 using the expression @code{(setq arg (1- arg))} as the decrementer.
13292 That expression is not far from the @code{while}, but is hidden in
13293 another Lisp macro, an @code{unless} macro. Unless we are at the end
13294 of the buffer---that is what the @code{eobp} function determines; it
13295 is an abbreviation of @samp{End Of Buffer P}---we decrease the value
13296 of @code{arg} by one.
13298 (If we are at the end of the buffer, we cannot go forward any more and
13299 the next loop of the @code{while} expression will test false since the
13300 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13301 function means exactly as you expect; it is another name for
13302 @code{null}, a function that returns true when its argument is false.)
13304 Interestingly, the loop count is not decremented until we leave the
13305 space between paragraphs, unless we come to the end of buffer or stop
13306 seeing the local value of the paragraph separator.
13308 That second @code{while} also has a @code{(move-to-left-margin)}
13309 expression. The function is self-explanatory. It is inside a
13310 @code{progn} expression and not the last element of its body, so it is
13311 only invoked for its side effect, which is to move point to the left
13312 margin of the current line.
13315 The @code{looking-at} function is also self-explanatory; it returns
13316 true if the text after point matches the regular expression given as
13319 The rest of the body of the loop looks difficult at first, but makes
13320 sense as you come to understand it.
13323 First consider what happens if there is a fill prefix:
13327 (if fill-prefix-regexp
13328 ;; There is a fill prefix; it overrides parstart;
13329 ;; we go forward line by line
13330 (while (and (not (eobp))
13331 (progn (move-to-left-margin) (not (eobp)))
13332 (not (looking-at parsep))
13333 (looking-at fill-prefix-regexp))
13339 This expression moves point forward line by line so long
13340 as four conditions are true:
13344 Point is not at the end of the buffer.
13347 We can move to the left margin of the text and are
13348 not at the end of the buffer.
13351 The text following point does not separate paragraphs.
13354 The pattern following point is the fill prefix regular expression.
13357 The last condition may be puzzling, until you remember that point was
13358 moved to the beginning of the line early in the @code{forward-paragraph}
13359 function. This means that if the text has a fill prefix, the
13360 @code{looking-at} function will see it.
13363 Consider what happens when there is no fill prefix.
13367 (while (and (re-search-forward sp-parstart nil 1)
13368 (progn (setq start (match-beginning 0))
13371 (progn (move-to-left-margin)
13372 (not (looking-at parsep)))
13373 (or (not (looking-at parstart))
13374 (and use-hard-newlines
13375 (not (get-text-property (1- start) 'hard)))))
13381 This @code{while} loop has us searching forward for
13382 @code{sp-parstart}, which is the combination of possible whitespace
13383 with a the local value of the start of a paragraph or of a paragraph
13384 separator. (The latter two are within an expression starting
13385 @code{\(?:} so that they are not referenced by the
13386 @code{match-beginning} function.)
13389 The two expressions,
13393 (setq start (match-beginning 0))
13399 mean go to the start of the text matched by the regular expression
13402 The @code{(match-beginning 0)} expression is new. It returns a number
13403 specifying the location of the start of the text that was matched by
13406 The @code{match-beginning} function is used here because of a
13407 characteristic of a forward search: a successful forward search,
13408 regardless of whether it is a plain search or a regular expression
13409 search, moves point to the end of the text that is found. In this
13410 case, a successful search moves point to the end of the pattern for
13411 @code{sp-parstart}.
13413 However, we want to put point at the end of the current paragraph, not
13414 somewhere else. Indeed, since the search possibly includes the
13415 paragraph separator, point may end up at the beginning of the next one
13416 unless we use an expression that includes @code{match-beginning}.
13418 @findex match-beginning
13419 When given an argument of 0, @code{match-beginning} returns the
13420 position that is the start of the text matched by the most recent
13421 search. In this case, the most recent search looks for
13422 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13423 the beginning position of that pattern, rather than the end position
13426 (Incidentally, when passed a positive number as an argument, the
13427 @code{match-beginning} function returns the location of point at that
13428 parenthesized expression in the last search unless that parenthesized
13429 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13430 appears here since the argument is 0.)
13433 The last expression when there is no fill prefix is
13437 (if (< (point) (point-max))
13438 (goto-char start))))
13443 This says that if there is no fill prefix and if we are not at the
13444 end, point should move to the beginning of whatever was found by the
13445 regular expression search for @code{sp-parstart}.
13447 The full definition for the @code{forward-paragraph} function not only
13448 includes code for going forwards, but also code for going backwards.
13450 If you are reading this inside of GNU Emacs and you want to see the
13451 whole function, you can type @kbd{C-h f} (@code{describe-function})
13452 and the name of the function. This gives you the function
13453 documentation and the name of the library containing the function's
13454 source. Place point over the name of the library and press the RET
13455 key; you will be taken directly to the source. (Be sure to install
13456 your sources! Without them, you are like a person who tries to drive
13457 a car with his eyes shut!)
13460 @section Create Your Own @file{TAGS} File
13462 @cindex @file{TAGS} file, create own
13464 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13465 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13466 name of the function when prompted for it. This is a good habit to
13467 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13468 to the source for a function, variable, or node. The function depends
13469 on tags tables to tell it where to go.
13471 If the @code{find-tag} function first asks you for the name of a
13472 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13473 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13474 @file{TAGS} file depends on how your copy of Emacs was installed. I
13475 just told you the location that provides both my C and my Emacs Lisp
13478 You can also create your own @file{TAGS} file for directories that
13481 You often need to build and install tags tables yourself. They are
13482 not built automatically. A tags table is called a @file{TAGS} file;
13483 the name is in upper case letters.
13485 You can create a @file{TAGS} file by calling the @code{etags} program
13486 that comes as a part of the Emacs distribution. Usually, @code{etags}
13487 is compiled and installed when Emacs is built. (@code{etags} is not
13488 an Emacs Lisp function or a part of Emacs; it is a C program.)
13491 To create a @file{TAGS} file, first switch to the directory in which
13492 you want to create the file. In Emacs you can do this with the
13493 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13494 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13495 compile command, with @w{@code{etags *.el}} as the command to execute
13498 M-x compile RET etags *.el RET
13502 to create a @file{TAGS} file for Emacs Lisp.
13504 For example, if you have a large number of files in your
13505 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13506 of which I load 12---you can create a @file{TAGS} file for the Emacs
13507 Lisp files in that directory.
13510 The @code{etags} program takes all the usual shell `wildcards'. For
13511 example, if you have two directories for which you want a single
13512 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13513 @file{../elisp/} is the second directory:
13516 M-x compile RET etags *.el ../elisp/*.el RET
13523 M-x compile RET etags --help RET
13527 to see a list of the options accepted by @code{etags} as well as a
13528 list of supported languages.
13530 The @code{etags} program handles more than 20 languages, including
13531 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13532 LaTeX, Pascal, Perl, PostScript, Python, TeX, Texinfo, makefiles, and
13533 most assemblers. The program has no switches for specifying the
13534 language; it recognizes the language in an input file according to its
13535 file name and contents.
13537 @file{etags} is very helpful when you are writing code yourself and
13538 want to refer back to functions you have already written. Just run
13539 @code{etags} again at intervals as you write new functions, so they
13540 become part of the @file{TAGS} file.
13542 If you think an appropriate @file{TAGS} file already exists for what
13543 you want, but do not know where it is, you can use the @code{locate}
13544 program to attempt to find it.
13546 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13547 for you the full path names of all your @file{TAGS} files. On my
13548 system, this command lists 34 @file{TAGS} files. On the other hand, a
13549 `plain vanilla' system I recently installed did not contain any
13552 If the tags table you want has been created, you can use the @code{M-x
13553 visit-tags-table} command to specify it. Otherwise, you will need to
13554 create the tag table yourself and then use @code{M-x
13557 @subsubheading Building Tags in the Emacs sources
13558 @cindex Building Tags in the Emacs sources
13559 @cindex Tags in the Emacs sources
13562 The GNU Emacs sources come with a @file{Makefile} that contains a
13563 sophisticated @code{etags} command that creates, collects, and merges
13564 tags tables from all over the Emacs sources and puts the information
13565 into one @file{TAGS} file in the @file{src/} directory. (The
13566 @file{src/} directory is below the top level of your Emacs directory.)
13569 To build this @file{TAGS} file, go to the top level of your Emacs
13570 source directory and run the compile command @code{make tags}:
13573 M-x compile RET make tags RET
13577 (The @code{make tags} command works well with the GNU Emacs sources,
13578 as well as with some other source packages.)
13580 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13583 @node Regexp Review
13586 Here is a brief summary of some recently introduced functions.
13590 Repeatedly evaluate the body of the expression so long as the first
13591 element of the body tests true. Then return @code{nil}. (The
13592 expression is evaluated only for its side effects.)
13601 (insert (format "foo is %d.\n" foo))
13602 (setq foo (1- foo))))
13604 @result{} foo is 2.
13611 (The @code{insert} function inserts its arguments at point; the
13612 @code{format} function returns a string formatted from its arguments
13613 the way @code{message} formats its arguments; @code{\n} produces a new
13616 @item re-search-forward
13617 Search for a pattern, and if the pattern is found, move point to rest
13621 Takes four arguments, like @code{search-forward}:
13625 A regular expression that specifies the pattern to search for.
13626 (Remember to put quotation marks around this argument!)
13629 Optionally, the limit of the search.
13632 Optionally, what to do if the search fails, return @code{nil} or an
13636 Optionally, how many times to repeat the search; if negative, the
13637 search goes backwards.
13641 Bind some variables locally to particular values,
13642 and then evaluate the remaining arguments, returning the value of the
13643 last one. While binding the local variables, use the local values of
13644 variables bound earlier, if any.
13653 (message "`bar' is %d." bar))
13654 @result{} `bar' is 21.
13658 @item match-beginning
13659 Return the position of the start of the text found by the last regular
13663 Return @code{t} for true if the text after point matches the argument,
13664 which should be a regular expression.
13667 Return @code{t} for true if point is at the end of the accessible part
13668 of a buffer. The end of the accessible part is the end of the buffer
13669 if the buffer is not narrowed; it is the end of the narrowed part if
13670 the buffer is narrowed.
13674 @node re-search Exercises
13675 @section Exercises with @code{re-search-forward}
13679 Write a function to search for a regular expression that matches two
13680 or more blank lines in sequence.
13683 Write a function to search for duplicated words, such as `the the'.
13684 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13685 Manual}, for information on how to write a regexp (a regular
13686 expression) to match a string that is composed of two identical
13687 halves. You can devise several regexps; some are better than others.
13688 The function I use is described in an appendix, along with several
13689 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13692 @node Counting Words
13693 @chapter Counting: Repetition and Regexps
13694 @cindex Repetition for word counting
13695 @cindex Regular expressions for word counting
13697 Repetition and regular expression searches are powerful tools that you
13698 often use when you write code in Emacs Lisp. This chapter illustrates
13699 the use of regular expression searches through the construction of
13700 word count commands using @code{while} loops and recursion.
13703 * Why Count Words::
13704 * @value{COUNT-WORDS}:: Use a regexp, but find a problem.
13705 * recursive-count-words:: Start with case of no words in region.
13706 * Counting Exercise::
13710 @node Why Count Words
13711 @unnumberedsec Counting words
13714 The standard Emacs distribution contains functions for counting the
13715 number of lines and words within a region.
13717 Certain types of writing ask you to count words. Thus, if you write
13718 an essay, you may be limited to 800 words; if you write a novel, you
13719 may discipline yourself to write 1000 words a day. It seems odd, but
13720 for a long time, Emacs lacked a word count command. Perhaps people used
13721 Emacs mostly for code or types of documentation that did not require
13722 word counts; or perhaps they restricted themselves to the operating
13723 system word count command, @code{wc}. Alternatively, people may have
13724 followed the publishers' convention and computed a word count by
13725 dividing the number of characters in a document by five.
13727 There are many ways to implement a command to count words. Here are
13728 some examples, which you may wish to compare with the standard Emacs
13729 command, @code{count-words-region}.
13731 @node @value{COUNT-WORDS}
13732 @section The @code{@value{COUNT-WORDS}} Function
13733 @findex @value{COUNT-WORDS}
13735 A word count command could count words in a line, paragraph, region,
13736 or buffer. What should the command cover? You could design the
13737 command to count the number of words in a complete buffer. However,
13738 the Emacs tradition encourages flexibility---you may want to count
13739 words in just a section, rather than all of a buffer. So it makes
13740 more sense to design the command to count the number of words in a
13741 region. Once you have a command to count words in a region, you can,
13742 if you wish, count words in a whole buffer by marking it with
13743 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13745 Clearly, counting words is a repetitive act: starting from the
13746 beginning of the region, you count the first word, then the second
13747 word, then the third word, and so on, until you reach the end of the
13748 region. This means that word counting is ideally suited to recursion
13749 or to a @code{while} loop.
13752 * Design @value{COUNT-WORDS}:: The definition using a @code{while} loop.
13753 * Whitespace Bug:: The Whitespace Bug in @code{@value{COUNT-WORDS}}.
13757 @node Design @value{COUNT-WORDS}
13758 @unnumberedsubsec Designing @code{@value{COUNT-WORDS}}
13761 First, we will implement the word count command with a @code{while}
13762 loop, then with recursion. The command will, of course, be
13766 The template for an interactive function definition is, as always:
13770 (defun @var{name-of-function} (@var{argument-list})
13771 "@var{documentation}@dots{}"
13772 (@var{interactive-expression}@dots{})
13777 What we need to do is fill in the slots.
13779 The name of the function should be self-explanatory and similar to the
13780 existing @code{count-lines-region} name. This makes the name easier
13781 to remember. @code{count-words-region} is the obvious choice. Since
13782 that name is now used for the standard Emacs command to count words, we
13783 will name our implementation @code{@value{COUNT-WORDS}}.
13785 The function counts words within a region. This means that the
13786 argument list must contain symbols that are bound to the two
13787 positions, the beginning and end of the region. These two positions
13788 can be called @samp{beginning} and @samp{end} respectively. The first
13789 line of the documentation should be a single sentence, since that is
13790 all that is printed as documentation by a command such as
13791 @code{apropos}. The interactive expression will be of the form
13792 @samp{(interactive "r")}, since that will cause Emacs to pass the
13793 beginning and end of the region to the function's argument list. All
13796 The body of the function needs to be written to do three tasks:
13797 first, to set up conditions under which the @code{while} loop can
13798 count words, second, to run the @code{while} loop, and third, to send
13799 a message to the user.
13801 When a user calls @code{@value{COUNT-WORDS}}, point may be at the
13802 beginning or the end of the region. However, the counting process
13803 must start at the beginning of the region. This means we will want
13804 to put point there if it is not already there. Executing
13805 @code{(goto-char beginning)} ensures this. Of course, we will want to
13806 return point to its expected position when the function finishes its
13807 work. For this reason, the body must be enclosed in a
13808 @code{save-excursion} expression.
13810 The central part of the body of the function consists of a
13811 @code{while} loop in which one expression jumps point forward word by
13812 word, and another expression counts those jumps. The true-or-false-test
13813 of the @code{while} loop should test true so long as point should jump
13814 forward, and false when point is at the end of the region.
13816 We could use @code{(forward-word 1)} as the expression for moving point
13817 forward word by word, but it is easier to see what Emacs identifies as a
13818 `word' if we use a regular expression search.
13820 A regular expression search that finds the pattern for which it is
13821 searching leaves point after the last character matched. This means
13822 that a succession of successful word searches will move point forward
13825 As a practical matter, we want the regular expression search to jump
13826 over whitespace and punctuation between words as well as over the
13827 words themselves. A regexp that refuses to jump over interword
13828 whitespace would never jump more than one word! This means that
13829 the regexp should include the whitespace and punctuation that follows
13830 a word, if any, as well as the word itself. (A word may end a buffer
13831 and not have any following whitespace or punctuation, so that part of
13832 the regexp must be optional.)
13834 Thus, what we want for the regexp is a pattern defining one or more
13835 word constituent characters followed, optionally, by one or more
13836 characters that are not word constituents. The regular expression for
13844 The buffer's syntax table determines which characters are and are not
13845 word constituents. For more information about syntax,
13846 @pxref{Syntax Tables, , Syntax Tables, elisp, The GNU Emacs Lisp
13850 The search expression looks like this:
13853 (re-search-forward "\\w+\\W*")
13857 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13858 single backslash has special meaning to the Emacs Lisp interpreter.
13859 It indicates that the following character is interpreted differently
13860 than usual. For example, the two characters, @samp{\n}, stand for
13861 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13862 backslashes in a row stand for an ordinary, `unspecial' backslash, so
13863 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13864 letter. So it discovers the letter is special.)
13866 We need a counter to count how many words there are; this variable
13867 must first be set to 0 and then incremented each time Emacs goes
13868 around the @code{while} loop. The incrementing expression is simply:
13871 (setq count (1+ count))
13874 Finally, we want to tell the user how many words there are in the
13875 region. The @code{message} function is intended for presenting this
13876 kind of information to the user. The message has to be phrased so
13877 that it reads properly regardless of how many words there are in the
13878 region: we don't want to say that ``there are 1 words in the region''.
13879 The conflict between singular and plural is ungrammatical. We can
13880 solve this problem by using a conditional expression that evaluates
13881 different messages depending on the number of words in the region.
13882 There are three possibilities: no words in the region, one word in the
13883 region, and more than one word. This means that the @code{cond}
13884 special form is appropriate.
13887 All this leads to the following function definition:
13891 ;;; @r{First version; has bugs!}
13892 (defun @value{COUNT-WORDS} (beginning end)
13893 "Print number of words in the region.
13894 Words are defined as at least one word-constituent
13895 character followed by at least one character that
13896 is not a word-constituent. The buffer's syntax
13897 table determines which characters these are."
13899 (message "Counting words in region ... ")
13903 ;;; @r{1. Set up appropriate conditions.}
13905 (goto-char beginning)
13910 ;;; @r{2. Run the} while @r{loop.}
13911 (while (< (point) end)
13912 (re-search-forward "\\w+\\W*")
13913 (setq count (1+ count)))
13917 ;;; @r{3. Send a message to the user.}
13918 (cond ((zerop count)
13920 "The region does NOT have any words."))
13923 "The region has 1 word."))
13926 "The region has %d words." count))))))
13931 As written, the function works, but not in all circumstances.
13933 @node Whitespace Bug
13934 @subsection The Whitespace Bug in @code{@value{COUNT-WORDS}}
13936 The @code{@value{COUNT-WORDS}} command described in the preceding
13937 section has two bugs, or rather, one bug with two manifestations.
13938 First, if you mark a region containing only whitespace in the middle
13939 of some text, the @code{@value{COUNT-WORDS}} command tells you that the
13940 region contains one word! Second, if you mark a region containing
13941 only whitespace at the end of the buffer or the accessible portion of
13942 a narrowed buffer, the command displays an error message that looks
13946 Search failed: "\\w+\\W*"
13949 If you are reading this in Info in GNU Emacs, you can test for these
13952 First, evaluate the function in the usual manner to install it.
13954 Here is a copy of the definition. Place your cursor after the closing
13955 parenthesis and type @kbd{C-x C-e} to install it.
13959 ;; @r{First version; has bugs!}
13960 (defun @value{COUNT-WORDS} (beginning end)
13961 "Print number of words in the region.
13962 Words are defined as at least one word-constituent character followed
13963 by at least one character that is not a word-constituent. The buffer's
13964 syntax table determines which characters these are."
13968 (message "Counting words in region ... ")
13972 ;;; @r{1. Set up appropriate conditions.}
13974 (goto-char beginning)
13979 ;;; @r{2. Run the} while @r{loop.}
13980 (while (< (point) end)
13981 (re-search-forward "\\w+\\W*")
13982 (setq count (1+ count)))
13986 ;;; @r{3. Send a message to the user.}
13987 (cond ((zerop count)
13988 (message "The region does NOT have any words."))
13989 ((= 1 count) (message "The region has 1 word."))
13990 (t (message "The region has %d words." count))))))
13996 If you wish, you can also install this keybinding by evaluating it:
13999 (global-set-key "\C-c=" '@value{COUNT-WORDS})
14002 To conduct the first test, set mark and point to the beginning and end
14003 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14004 @value{COUNT-WORDS}} if you have not bound @kbd{C-c =}):
14011 Emacs will tell you, correctly, that the region has three words.
14013 Repeat the test, but place mark at the beginning of the line and place
14014 point just @emph{before} the word @samp{one}. Again type the command
14015 @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}). Emacs should tell you
14016 that the region has no words, since it is composed only of the
14017 whitespace at the beginning of the line. But instead Emacs tells you
14018 that the region has one word!
14020 For the third test, copy the sample line to the end of the
14021 @file{*scratch*} buffer and then type several spaces at the end of the
14022 line. Place mark right after the word @samp{three} and point at the
14023 end of line. (The end of the line will be the end of the buffer.)
14024 Type @kbd{C-c =} (or @kbd{M-x @value{COUNT-WORDS}}) as you did before.
14025 Again, Emacs should tell you that the region has no words, since it is
14026 composed only of the whitespace at the end of the line. Instead,
14027 Emacs displays an error message saying @samp{Search failed}.
14029 The two bugs stem from the same problem.
14031 Consider the first manifestation of the bug, in which the command
14032 tells you that the whitespace at the beginning of the line contains
14033 one word. What happens is this: The @code{M-x @value{COUNT-WORDS}}
14034 command moves point to the beginning of the region. The @code{while}
14035 tests whether the value of point is smaller than the value of
14036 @code{end}, which it is. Consequently, the regular expression search
14037 looks for and finds the first word. It leaves point after the word.
14038 @code{count} is set to one. The @code{while} loop repeats; but this
14039 time the value of point is larger than the value of @code{end}, the
14040 loop is exited; and the function displays a message saying the number
14041 of words in the region is one. In brief, the regular expression
14042 search looks for and finds the word even though it is outside
14045 In the second manifestation of the bug, the region is whitespace at
14046 the end of the buffer. Emacs says @samp{Search failed}. What happens
14047 is that the true-or-false-test in the @code{while} loop tests true, so
14048 the search expression is executed. But since there are no more words
14049 in the buffer, the search fails.
14051 In both manifestations of the bug, the search extends or attempts to
14052 extend outside of the region.
14054 The solution is to limit the search to the region---this is a fairly
14055 simple action, but as you may have come to expect, it is not quite as
14056 simple as you might think.
14058 As we have seen, the @code{re-search-forward} function takes a search
14059 pattern as its first argument. But in addition to this first,
14060 mandatory argument, it accepts three optional arguments. The optional
14061 second argument bounds the search. The optional third argument, if
14062 @code{t}, causes the function to return @code{nil} rather than signal
14063 an error if the search fails. The optional fourth argument is a
14064 repeat count. (In Emacs, you can see a function's documentation by
14065 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14067 In the @code{@value{COUNT-WORDS}} definition, the value of the end of
14068 the region is held by the variable @code{end} which is passed as an
14069 argument to the function. Thus, we can add @code{end} as an argument
14070 to the regular expression search expression:
14073 (re-search-forward "\\w+\\W*" end)
14076 However, if you make only this change to the @code{@value{COUNT-WORDS}}
14077 definition and then test the new version of the definition on a
14078 stretch of whitespace, you will receive an error message saying
14079 @samp{Search failed}.
14081 What happens is this: the search is limited to the region, and fails
14082 as you expect because there are no word-constituent characters in the
14083 region. Since it fails, we receive an error message. But we do not
14084 want to receive an error message in this case; we want to receive the
14085 message that "The region does NOT have any words."
14087 The solution to this problem is to provide @code{re-search-forward}
14088 with a third argument of @code{t}, which causes the function to return
14089 @code{nil} rather than signal an error if the search fails.
14091 However, if you make this change and try it, you will see the message
14092 ``Counting words in region ... '' and @dots{} you will keep on seeing
14093 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14095 Here is what happens: the search is limited to the region, as before,
14096 and it fails because there are no word-constituent characters in the
14097 region, as expected. Consequently, the @code{re-search-forward}
14098 expression returns @code{nil}. It does nothing else. In particular,
14099 it does not move point, which it does as a side effect if it finds the
14100 search target. After the @code{re-search-forward} expression returns
14101 @code{nil}, the next expression in the @code{while} loop is evaluated.
14102 This expression increments the count. Then the loop repeats. The
14103 true-or-false-test tests true because the value of point is still less
14104 than the value of end, since the @code{re-search-forward} expression
14105 did not move point. @dots{} and the cycle repeats @dots{}
14107 The @code{@value{COUNT-WORDS}} definition requires yet another
14108 modification, to cause the true-or-false-test of the @code{while} loop
14109 to test false if the search fails. Put another way, there are two
14110 conditions that must be satisfied in the true-or-false-test before the
14111 word count variable is incremented: point must still be within the
14112 region and the search expression must have found a word to count.
14114 Since both the first condition and the second condition must be true
14115 together, the two expressions, the region test and the search
14116 expression, can be joined with an @code{and} special form and embedded in
14117 the @code{while} loop as the true-or-false-test, like this:
14120 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14123 @c colon in printed section title causes problem in Info cross reference
14124 @c also trouble with an overfull hbox
14127 (For information about @code{and}, see
14128 @ref{kill-new function, , The @code{kill-new} function}.)
14132 (@xref{kill-new function, , The @code{kill-new} function}, for
14133 information about @code{and}.)
14136 The @code{re-search-forward} expression returns @code{t} if the search
14137 succeeds and as a side effect moves point. Consequently, as words are
14138 found, point is moved through the region. When the search expression
14139 fails to find another word, or when point reaches the end of the
14140 region, the true-or-false-test tests false, the @code{while} loop
14141 exits, and the @code{@value{COUNT-WORDS}} function displays one or
14142 other of its messages.
14144 After incorporating these final changes, the @code{@value{COUNT-WORDS}}
14145 works without bugs (or at least, without bugs that I have found!).
14146 Here is what it looks like:
14150 ;;; @r{Final version:} @code{while}
14151 (defun @value{COUNT-WORDS} (beginning end)
14152 "Print number of words in the region."
14154 (message "Counting words in region ... ")
14158 ;;; @r{1. Set up appropriate conditions.}
14161 (goto-char beginning)
14165 ;;; @r{2. Run the} while @r{loop.}
14166 (while (and (< (point) end)
14167 (re-search-forward "\\w+\\W*" end t))
14168 (setq count (1+ count)))
14172 ;;; @r{3. Send a message to the user.}
14173 (cond ((zerop count)
14175 "The region does NOT have any words."))
14178 "The region has 1 word."))
14181 "The region has %d words." count))))))
14185 @node recursive-count-words
14186 @section Count Words Recursively
14187 @cindex Count words recursively
14188 @cindex Recursively counting words
14189 @cindex Words, counted recursively
14191 You can write the function for counting words recursively as well as
14192 with a @code{while} loop. Let's see how this is done.
14194 First, we need to recognize that the @code{@value{COUNT-WORDS}}
14195 function has three jobs: it sets up the appropriate conditions for
14196 counting to occur; it counts the words in the region; and it sends a
14197 message to the user telling how many words there are.
14199 If we write a single recursive function to do everything, we will
14200 receive a message for every recursive call. If the region contains 13
14201 words, we will receive thirteen messages, one right after the other.
14202 We don't want this! Instead, we must write two functions to do the
14203 job, one of which (the recursive function) will be used inside of the
14204 other. One function will set up the conditions and display the
14205 message; the other will return the word count.
14207 Let us start with the function that causes the message to be displayed.
14208 We can continue to call this @code{@value{COUNT-WORDS}}.
14210 This is the function that the user will call. It will be interactive.
14211 Indeed, it will be similar to our previous versions of this
14212 function, except that it will call @code{recursive-count-words} to
14213 determine how many words are in the region.
14216 We can readily construct a template for this function, based on our
14221 ;; @r{Recursive version; uses regular expression search}
14222 (defun @value{COUNT-WORDS} (beginning end)
14223 "@var{documentation}@dots{}"
14224 (@var{interactive-expression}@dots{})
14228 ;;; @r{1. Set up appropriate conditions.}
14229 (@var{explanatory message})
14230 (@var{set-up functions}@dots{}
14234 ;;; @r{2. Count the words.}
14235 @var{recursive call}
14239 ;;; @r{3. Send a message to the user.}
14240 @var{message providing word count}))
14244 The definition looks straightforward, except that somehow the count
14245 returned by the recursive call must be passed to the message
14246 displaying the word count. A little thought suggests that this can be
14247 done by making use of a @code{let} expression: we can bind a variable
14248 in the varlist of a @code{let} expression to the number of words in
14249 the region, as returned by the recursive call; and then the
14250 @code{cond} expression, using binding, can display the value to the
14253 Often, one thinks of the binding within a @code{let} expression as
14254 somehow secondary to the `primary' work of a function. But in this
14255 case, what you might consider the `primary' job of the function,
14256 counting words, is done within the @code{let} expression.
14259 Using @code{let}, the function definition looks like this:
14263 (defun @value{COUNT-WORDS} (beginning end)
14264 "Print number of words in the region."
14269 ;;; @r{1. Set up appropriate conditions.}
14270 (message "Counting words in region ... ")
14272 (goto-char beginning)
14276 ;;; @r{2. Count the words.}
14277 (let ((count (recursive-count-words end)))
14281 ;;; @r{3. Send a message to the user.}
14282 (cond ((zerop count)
14284 "The region does NOT have any words."))
14287 "The region has 1 word."))
14290 "The region has %d words." count))))))
14294 Next, we need to write the recursive counting function.
14296 A recursive function has at least three parts: the `do-again-test', the
14297 `next-step-expression', and the recursive call.
14299 The do-again-test determines whether the function will or will not be
14300 called again. Since we are counting words in a region and can use a
14301 function that moves point forward for every word, the do-again-test
14302 can check whether point is still within the region. The do-again-test
14303 should find the value of point and determine whether point is before,
14304 at, or after the value of the end of the region. We can use the
14305 @code{point} function to locate point. Clearly, we must pass the
14306 value of the end of the region to the recursive counting function as an
14309 In addition, the do-again-test should also test whether the search finds a
14310 word. If it does not, the function should not call itself again.
14312 The next-step-expression changes a value so that when the recursive
14313 function is supposed to stop calling itself, it stops. More
14314 precisely, the next-step-expression changes a value so that at the
14315 right time, the do-again-test stops the recursive function from
14316 calling itself again. In this case, the next-step-expression can be
14317 the expression that moves point forward, word by word.
14319 The third part of a recursive function is the recursive call.
14321 Somewhere, also, we also need a part that does the `work' of the
14322 function, a part that does the counting. A vital part!
14325 But already, we have an outline of the recursive counting function:
14329 (defun recursive-count-words (region-end)
14330 "@var{documentation}@dots{}"
14331 @var{do-again-test}
14332 @var{next-step-expression}
14333 @var{recursive call})
14337 Now we need to fill in the slots. Let's start with the simplest cases
14338 first: if point is at or beyond the end of the region, there cannot
14339 be any words in the region, so the function should return zero.
14340 Likewise, if the search fails, there are no words to count, so the
14341 function should return zero.
14343 On the other hand, if point is within the region and the search
14344 succeeds, the function should call itself again.
14347 Thus, the do-again-test should look like this:
14351 (and (< (point) region-end)
14352 (re-search-forward "\\w+\\W*" region-end t))
14356 Note that the search expression is part of the do-again-test---the
14357 function returns @code{t} if its search succeeds and @code{nil} if it
14358 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14359 @code{@value{COUNT-WORDS}}}, for an explanation of how
14360 @code{re-search-forward} works.)
14362 The do-again-test is the true-or-false test of an @code{if} clause.
14363 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14364 clause should call the function again; but if it fails, the else-part
14365 should return zero since either point is outside the region or the
14366 search failed because there were no words to find.
14368 But before considering the recursive call, we need to consider the
14369 next-step-expression. What is it? Interestingly, it is the search
14370 part of the do-again-test.
14372 In addition to returning @code{t} or @code{nil} for the
14373 do-again-test, @code{re-search-forward} moves point forward as a side
14374 effect of a successful search. This is the action that changes the
14375 value of point so that the recursive function stops calling itself
14376 when point completes its movement through the region. Consequently,
14377 the @code{re-search-forward} expression is the next-step-expression.
14380 In outline, then, the body of the @code{recursive-count-words}
14381 function looks like this:
14385 (if @var{do-again-test-and-next-step-combined}
14387 @var{recursive-call-returning-count}
14393 How to incorporate the mechanism that counts?
14395 If you are not used to writing recursive functions, a question like
14396 this can be troublesome. But it can and should be approached
14399 We know that the counting mechanism should be associated in some way
14400 with the recursive call. Indeed, since the next-step-expression moves
14401 point forward by one word, and since a recursive call is made for
14402 each word, the counting mechanism must be an expression that adds one
14403 to the value returned by a call to @code{recursive-count-words}.
14406 Consider several cases:
14410 If there are two words in the region, the function should return
14411 a value resulting from adding one to the value returned when it counts
14412 the first word, plus the number returned when it counts the remaining
14413 words in the region, which in this case is one.
14416 If there is one word in the region, the function should return
14417 a value resulting from adding one to the value returned when it counts
14418 that word, plus the number returned when it counts the remaining
14419 words in the region, which in this case is zero.
14422 If there are no words in the region, the function should return zero.
14425 From the sketch we can see that the else-part of the @code{if} returns
14426 zero for the case of no words. This means that the then-part of the
14427 @code{if} must return a value resulting from adding one to the value
14428 returned from a count of the remaining words.
14431 The expression will look like this, where @code{1+} is a function that
14432 adds one to its argument.
14435 (1+ (recursive-count-words region-end))
14439 The whole @code{recursive-count-words} function will then look like
14444 (defun recursive-count-words (region-end)
14445 "@var{documentation}@dots{}"
14447 ;;; @r{1. do-again-test}
14448 (if (and (< (point) region-end)
14449 (re-search-forward "\\w+\\W*" region-end t))
14453 ;;; @r{2. then-part: the recursive call}
14454 (1+ (recursive-count-words region-end))
14456 ;;; @r{3. else-part}
14462 Let's examine how this works:
14464 If there are no words in the region, the else part of the @code{if}
14465 expression is evaluated and consequently the function returns zero.
14467 If there is one word in the region, the value of point is less than
14468 the value of @code{region-end} and the search succeeds. In this case,
14469 the true-or-false-test of the @code{if} expression tests true, and the
14470 then-part of the @code{if} expression is evaluated. The counting
14471 expression is evaluated. This expression returns a value (which will
14472 be the value returned by the whole function) that is the sum of one
14473 added to the value returned by a recursive call.
14475 Meanwhile, the next-step-expression has caused point to jump over the
14476 first (and in this case only) word in the region. This means that
14477 when @code{(recursive-count-words region-end)} is evaluated a second
14478 time, as a result of the recursive call, the value of point will be
14479 equal to or greater than the value of region end. So this time,
14480 @code{recursive-count-words} will return zero. The zero will be added
14481 to one, and the original evaluation of @code{recursive-count-words}
14482 will return one plus zero, which is one, which is the correct amount.
14484 Clearly, if there are two words in the region, the first call to
14485 @code{recursive-count-words} returns one added to the value returned
14486 by calling @code{recursive-count-words} on a region containing the
14487 remaining word---that is, it adds one to one, producing two, which is
14488 the correct amount.
14490 Similarly, if there are three words in the region, the first call to
14491 @code{recursive-count-words} returns one added to the value returned
14492 by calling @code{recursive-count-words} on a region containing the
14493 remaining two words---and so on and so on.
14497 With full documentation the two functions look like this:
14501 The recursive function:
14503 @findex recursive-count-words
14506 (defun recursive-count-words (region-end)
14507 "Number of words between point and REGION-END."
14511 ;;; @r{1. do-again-test}
14512 (if (and (< (point) region-end)
14513 (re-search-forward "\\w+\\W*" region-end t))
14517 ;;; @r{2. then-part: the recursive call}
14518 (1+ (recursive-count-words region-end))
14520 ;;; @r{3. else-part}
14531 ;;; @r{Recursive version}
14532 (defun @value{COUNT-WORDS} (beginning end)
14533 "Print number of words in the region.
14537 Words are defined as at least one word-constituent
14538 character followed by at least one character that is
14539 not a word-constituent. The buffer's syntax table
14540 determines which characters these are."
14544 (message "Counting words in region ... ")
14546 (goto-char beginning)
14547 (let ((count (recursive-count-words end)))
14550 (cond ((zerop count)
14552 "The region does NOT have any words."))
14556 (message "The region has 1 word."))
14559 "The region has %d words." count))))))
14563 @node Counting Exercise
14564 @section Exercise: Counting Punctuation
14566 Using a @code{while} loop, write a function to count the number of
14567 punctuation marks in a region---period, comma, semicolon, colon,
14568 exclamation mark, and question mark. Do the same using recursion.
14570 @node Words in a defun
14571 @chapter Counting Words in a @code{defun}
14572 @cindex Counting words in a @code{defun}
14573 @cindex Word counting in a @code{defun}
14575 Our next project is to count the number of words in a function
14576 definition. Clearly, this can be done using some variant of
14577 @code{@value{COUNT-WORDS}}. @xref{Counting Words, , Counting Words:
14578 Repetition and Regexps}. If we are just going to count the words in
14579 one definition, it is easy enough to mark the definition with the
14580 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14581 @code{@value{COUNT-WORDS}}.
14583 However, I am more ambitious: I want to count the words and symbols in
14584 every definition in the Emacs sources and then print a graph that
14585 shows how many functions there are of each length: how many contain 40
14586 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14587 and so on. I have often been curious how long a typical function is,
14588 and this will tell.
14591 * Divide and Conquer::
14592 * Words and Symbols:: What to count?
14593 * Syntax:: What constitutes a word or symbol?
14594 * count-words-in-defun:: Very like @code{@value{COUNT-WORDS}}.
14595 * Several defuns:: Counting several defuns in a file.
14596 * Find a File:: Do you want to look at a file?
14597 * lengths-list-file:: A list of the lengths of many definitions.
14598 * Several files:: Counting in definitions in different files.
14599 * Several files recursively:: Recursively counting in different files.
14600 * Prepare the data:: Prepare the data for display in a graph.
14604 @node Divide and Conquer
14605 @unnumberedsec Divide and Conquer
14608 Described in one phrase, the histogram project is daunting; but
14609 divided into numerous small steps, each of which we can take one at a
14610 time, the project becomes less fearsome. Let us consider what the
14615 First, write a function to count the words in one definition. This
14616 includes the problem of handling symbols as well as words.
14619 Second, write a function to list the numbers of words in each function
14620 in a file. This function can use the @code{count-words-in-defun}
14624 Third, write a function to list the numbers of words in each function
14625 in each of several files. This entails automatically finding the
14626 various files, switching to them, and counting the words in the
14627 definitions within them.
14630 Fourth, write a function to convert the list of numbers that we
14631 created in step three to a form that will be suitable for printing as
14635 Fifth, write a function to print the results as a graph.
14638 This is quite a project! But if we take each step slowly, it will not
14641 @node Words and Symbols
14642 @section What to Count?
14643 @cindex Words and symbols in defun
14645 When we first start thinking about how to count the words in a
14646 function definition, the first question is (or ought to be) what are
14647 we going to count? When we speak of `words' with respect to a Lisp
14648 function definition, we are actually speaking, in large part, of
14649 `symbols'. For example, the following @code{multiply-by-seven}
14650 function contains the five symbols @code{defun},
14651 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14652 addition, in the documentation string, it contains the four words
14653 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14654 symbol @samp{number} is repeated, so the definition contains a total
14655 of ten words and symbols.
14659 (defun multiply-by-seven (number)
14660 "Multiply NUMBER by seven."
14666 However, if we mark the @code{multiply-by-seven} definition with
14667 @kbd{C-M-h} (@code{mark-defun}), and then call
14668 @code{@value{COUNT-WORDS}} on it, we will find that
14669 @code{@value{COUNT-WORDS}} claims the definition has eleven words, not
14670 ten! Something is wrong!
14672 The problem is twofold: @code{@value{COUNT-WORDS}} does not count the
14673 @samp{*} as a word, and it counts the single symbol,
14674 @code{multiply-by-seven}, as containing three words. The hyphens are
14675 treated as if they were interword spaces rather than intraword
14676 connectors: @samp{multiply-by-seven} is counted as if it were written
14677 @samp{multiply by seven}.
14679 The cause of this confusion is the regular expression search within
14680 the @code{@value{COUNT-WORDS}} definition that moves point forward word
14681 by word. In the canonical version of @code{@value{COUNT-WORDS}}, the
14689 This regular expression is a pattern defining one or more word
14690 constituent characters possibly followed by one or more characters
14691 that are not word constituents. What is meant by `word constituent
14692 characters' brings us to the issue of syntax, which is worth a section
14696 @section What Constitutes a Word or Symbol?
14697 @cindex Syntax categories and tables
14699 Emacs treats different characters as belonging to different
14700 @dfn{syntax categories}. For example, the regular expression,
14701 @samp{\\w+}, is a pattern specifying one or more @emph{word
14702 constituent} characters. Word constituent characters are members of
14703 one syntax category. Other syntax categories include the class of
14704 punctuation characters, such as the period and the comma, and the
14705 class of whitespace characters, such as the blank space and the tab
14706 character. (For more information, @pxref{Syntax Tables, , Syntax
14707 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14709 Syntax tables specify which characters belong to which categories.
14710 Usually, a hyphen is not specified as a `word constituent character'.
14711 Instead, it is specified as being in the `class of characters that are
14712 part of symbol names but not words.' This means that the
14713 @code{@value{COUNT-WORDS}} function treats it in the same way it treats
14714 an interword white space, which is why @code{@value{COUNT-WORDS}}
14715 counts @samp{multiply-by-seven} as three words.
14717 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14718 one symbol: modify the syntax table or modify the regular expression.
14720 We could redefine a hyphen as a word constituent character by
14721 modifying the syntax table that Emacs keeps for each mode. This
14722 action would serve our purpose, except that a hyphen is merely the
14723 most common character within symbols that is not typically a word
14724 constituent character; there are others, too.
14726 Alternatively, we can redefine the regexp used in the
14727 @code{@value{COUNT-WORDS}} definition so as to include symbols. This
14728 procedure has the merit of clarity, but the task is a little tricky.
14731 The first part is simple enough: the pattern must match ``at least one
14732 character that is a word or symbol constituent''. Thus:
14735 "\\(\\w\\|\\s_\\)+"
14739 The @samp{\\(} is the first part of the grouping construct that
14740 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14741 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14742 character and the @samp{\\s_} matches any character that is part of a
14743 symbol name but not a word-constituent character. The @samp{+}
14744 following the group indicates that the word or symbol constituent
14745 characters must be matched at least once.
14747 However, the second part of the regexp is more difficult to design.
14748 What we want is to follow the first part with ``optionally one or more
14749 characters that are not constituents of a word or symbol''. At first,
14750 I thought I could define this with the following:
14753 "\\(\\W\\|\\S_\\)*"
14757 The upper case @samp{W} and @samp{S} match characters that are
14758 @emph{not} word or symbol constituents. Unfortunately, this
14759 expression matches any character that is either not a word constituent
14760 or not a symbol constituent. This matches any character!
14762 I then noticed that every word or symbol in my test region was
14763 followed by white space (blank space, tab, or newline). So I tried
14764 placing a pattern to match one or more blank spaces after the pattern
14765 for one or more word or symbol constituents. This failed, too. Words
14766 and symbols are often separated by whitespace, but in actual code
14767 parentheses may follow symbols and punctuation may follow words. So
14768 finally, I designed a pattern in which the word or symbol constituents
14769 are followed optionally by characters that are not white space and
14770 then followed optionally by white space.
14773 Here is the full regular expression:
14776 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14779 @node count-words-in-defun
14780 @section The @code{count-words-in-defun} Function
14781 @cindex Counting words in a @code{defun}
14783 We have seen that there are several ways to write a
14784 @code{count-words-region} function. To write a
14785 @code{count-words-in-defun}, we need merely adapt one of these
14788 The version that uses a @code{while} loop is easy to understand, so I
14789 am going to adapt that. Because @code{count-words-in-defun} will be
14790 part of a more complex program, it need not be interactive and it need
14791 not display a message but just return the count. These considerations
14792 simplify the definition a little.
14794 On the other hand, @code{count-words-in-defun} will be used within a
14795 buffer that contains function definitions. Consequently, it is
14796 reasonable to ask that the function determine whether it is called
14797 when point is within a function definition, and if it is, to return
14798 the count for that definition. This adds complexity to the
14799 definition, but saves us from needing to pass arguments to the
14803 These considerations lead us to prepare the following template:
14807 (defun count-words-in-defun ()
14808 "@var{documentation}@dots{}"
14809 (@var{set up}@dots{}
14810 (@var{while loop}@dots{})
14811 @var{return count})
14816 As usual, our job is to fill in the slots.
14820 We are presuming that this function will be called within a buffer
14821 containing function definitions. Point will either be within a
14822 function definition or not. For @code{count-words-in-defun} to work,
14823 point must move to the beginning of the definition, a counter must
14824 start at zero, and the counting loop must stop when point reaches the
14825 end of the definition.
14827 The @code{beginning-of-defun} function searches backwards for an
14828 opening delimiter such as a @samp{(} at the beginning of a line, and
14829 moves point to that position, or else to the limit of the search. In
14830 practice, this means that @code{beginning-of-defun} moves point to the
14831 beginning of an enclosing or preceding function definition, or else to
14832 the beginning of the buffer. We can use @code{beginning-of-defun} to
14833 place point where we wish to start.
14835 The @code{while} loop requires a counter to keep track of the words or
14836 symbols being counted. A @code{let} expression can be used to create
14837 a local variable for this purpose, and bind it to an initial value of zero.
14839 The @code{end-of-defun} function works like @code{beginning-of-defun}
14840 except that it moves point to the end of the definition.
14841 @code{end-of-defun} can be used as part of an expression that
14842 determines the position of the end of the definition.
14844 The set up for @code{count-words-in-defun} takes shape rapidly: first
14845 we move point to the beginning of the definition, then we create a
14846 local variable to hold the count, and finally, we record the position
14847 of the end of the definition so the @code{while} loop will know when to stop
14851 The code looks like this:
14855 (beginning-of-defun)
14857 (end (save-excursion (end-of-defun) (point))))
14862 The code is simple. The only slight complication is likely to concern
14863 @code{end}: it is bound to the position of the end of the definition
14864 by a @code{save-excursion} expression that returns the value of point
14865 after @code{end-of-defun} temporarily moves it to the end of the
14868 The second part of the @code{count-words-in-defun}, after the set up,
14869 is the @code{while} loop.
14871 The loop must contain an expression that jumps point forward word by
14872 word and symbol by symbol, and another expression that counts the
14873 jumps. The true-or-false-test for the @code{while} loop should test
14874 true so long as point should jump forward, and false when point is at
14875 the end of the definition. We have already redefined the regular
14876 expression for this, so the loop is straightforward:
14880 (while (and (< (point) end)
14882 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t))
14883 (setq count (1+ count)))
14887 The third part of the function definition returns the count of words
14888 and symbols. This part is the last expression within the body of the
14889 @code{let} expression, and can be, very simply, the local variable
14890 @code{count}, which when evaluated returns the count.
14893 Put together, the @code{count-words-in-defun} definition looks like this:
14895 @findex count-words-in-defun
14898 (defun count-words-in-defun ()
14899 "Return the number of words and symbols in a defun."
14900 (beginning-of-defun)
14902 (end (save-excursion (end-of-defun) (point))))
14906 (and (< (point) end)
14908 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14910 (setq count (1+ count)))
14915 How to test this? The function is not interactive, but it is easy to
14916 put a wrapper around the function to make it interactive; we can use
14917 almost the same code as for the recursive version of
14918 @code{@value{COUNT-WORDS}}:
14922 ;;; @r{Interactive version.}
14923 (defun count-words-defun ()
14924 "Number of words and symbols in a function definition."
14927 "Counting words and symbols in function definition ... ")
14930 (let ((count (count-words-in-defun)))
14934 "The definition does NOT have any words or symbols."))
14939 "The definition has 1 word or symbol."))
14942 "The definition has %d words or symbols." count)))))
14948 Let's re-use @kbd{C-c =} as a convenient keybinding:
14951 (global-set-key "\C-c=" 'count-words-defun)
14954 Now we can try out @code{count-words-defun}: install both
14955 @code{count-words-in-defun} and @code{count-words-defun}, and set the
14956 keybinding, and then place the cursor within the following definition:
14960 (defun multiply-by-seven (number)
14961 "Multiply NUMBER by seven."
14968 Success! The definition has 10 words and symbols.
14970 The next problem is to count the numbers of words and symbols in
14971 several definitions within a single file.
14973 @node Several defuns
14974 @section Count Several @code{defuns} Within a File
14976 A file such as @file{simple.el} may have a hundred or more function
14977 definitions within it. Our long term goal is to collect statistics on
14978 many files, but as a first step, our immediate goal is to collect
14979 statistics on one file.
14981 The information will be a series of numbers, each number being the
14982 length of a function definition. We can store the numbers in a list.
14984 We know that we will want to incorporate the information regarding one
14985 file with information about many other files; this means that the
14986 function for counting definition lengths within one file need only
14987 return the list of lengths. It need not and should not display any
14990 The word count commands contain one expression to jump point forward
14991 word by word and another expression to count the jumps. The function
14992 to return the lengths of definitions can be designed to work the same
14993 way, with one expression to jump point forward definition by
14994 definition and another expression to construct the lengths' list.
14996 This statement of the problem makes it elementary to write the
14997 function definition. Clearly, we will start the count at the
14998 beginning of the file, so the first command will be @code{(goto-char
14999 (point-min))}. Next, we start the @code{while} loop; and the
15000 true-or-false test of the loop can be a regular expression search for
15001 the next function definition---so long as the search succeeds, point
15002 is moved forward and then the body of the loop is evaluated. The body
15003 needs an expression that constructs the lengths' list. @code{cons},
15004 the list construction command, can be used to create the list. That
15005 is almost all there is to it.
15008 Here is what this fragment of code looks like:
15012 (goto-char (point-min))
15013 (while (re-search-forward "^(defun" nil t)
15015 (cons (count-words-in-defun) lengths-list)))
15019 What we have left out is the mechanism for finding the file that
15020 contains the function definitions.
15022 In previous examples, we either used this, the Info file, or we
15023 switched back and forth to some other buffer, such as the
15024 @file{*scratch*} buffer.
15026 Finding a file is a new process that we have not yet discussed.
15029 @section Find a File
15030 @cindex Find a File
15032 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15033 command. This command is almost, but not quite right for the lengths
15037 Let's look at the source for @code{find-file}:
15041 (defun find-file (filename)
15042 "Edit file FILENAME.
15043 Switch to a buffer visiting file FILENAME,
15044 creating one if none already exists."
15045 (interactive "FFind file: ")
15046 (switch-to-buffer (find-file-noselect filename)))
15051 (The most recent version of the @code{find-file} function definition
15052 permits you to specify optional wildcards to visit multiple files; that
15053 makes the definition more complex and we will not discuss it here,
15054 since it is not relevant. You can see its source using either
15055 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15059 (defun find-file (filename &optional wildcards)
15060 "Edit file FILENAME.
15061 Switch to a buffer visiting file FILENAME,
15062 creating one if none already exists.
15063 Interactively, the default if you just type RET is the current directory,
15064 but the visited file name is available through the minibuffer history:
15065 type M-n to pull it into the minibuffer.
15067 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15068 expand wildcards (if any) and visit multiple files. You can
15069 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15071 To visit a file without any kind of conversion and without
15072 automatically choosing a major mode, use \\[find-file-literally]."
15073 (interactive (find-file-read-args "Find file: " nil))
15074 (let ((value (find-file-noselect filename nil nil wildcards)))
15076 (mapcar 'switch-to-buffer (nreverse value))
15077 (switch-to-buffer value))))
15080 The definition I am showing possesses short but complete documentation
15081 and an interactive specification that prompts you for a file name when
15082 you use the command interactively. The body of the definition
15083 contains two functions, @code{find-file-noselect} and
15084 @code{switch-to-buffer}.
15086 According to its documentation as shown by @kbd{C-h f} (the
15087 @code{describe-function} command), the @code{find-file-noselect}
15088 function reads the named file into a buffer and returns the buffer.
15089 (Its most recent version includes an optional wildcards argument,
15090 too, as well as another to read a file literally and an other you
15091 suppress warning messages. These optional arguments are irrelevant.)
15093 However, the @code{find-file-noselect} function does not select the
15094 buffer in which it puts the file. Emacs does not switch its attention
15095 (or yours if you are using @code{find-file-noselect}) to the selected
15096 buffer. That is what @code{switch-to-buffer} does: it switches the
15097 buffer to which Emacs attention is directed; and it switches the
15098 buffer displayed in the window to the new buffer. We have discussed
15099 buffer switching elsewhere. (@xref{Switching Buffers}.)
15101 In this histogram project, we do not need to display each file on the
15102 screen as the program determines the length of each definition within
15103 it. Instead of employing @code{switch-to-buffer}, we can work with
15104 @code{set-buffer}, which redirects the attention of the computer
15105 program to a different buffer but does not redisplay it on the screen.
15106 So instead of calling on @code{find-file} to do the job, we must write
15107 our own expression.
15109 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15111 @node lengths-list-file
15112 @section @code{lengths-list-file} in Detail
15114 The core of the @code{lengths-list-file} function is a @code{while}
15115 loop containing a function to move point forward `defun by defun' and
15116 a function to count the number of words and symbols in each defun.
15117 This core must be surrounded by functions that do various other tasks,
15118 including finding the file, and ensuring that point starts out at the
15119 beginning of the file. The function definition looks like this:
15120 @findex lengths-list-file
15124 (defun lengths-list-file (filename)
15125 "Return list of definitions' lengths within FILE.
15126 The returned list is a list of numbers.
15127 Each number is the number of words or
15128 symbols in one function definition."
15131 (message "Working on `%s' ... " filename)
15133 (let ((buffer (find-file-noselect filename))
15135 (set-buffer buffer)
15136 (setq buffer-read-only t)
15138 (goto-char (point-min))
15139 (while (re-search-forward "^(defun" nil t)
15141 (cons (count-words-in-defun) lengths-list)))
15142 (kill-buffer buffer)
15148 The function is passed one argument, the name of the file on which it
15149 will work. It has four lines of documentation, but no interactive
15150 specification. Since people worry that a computer is broken if they
15151 don't see anything going on, the first line of the body is a
15154 The next line contains a @code{save-excursion} that returns Emacs's
15155 attention to the current buffer when the function completes. This is
15156 useful in case you embed this function in another function that
15157 presumes point is restored to the original buffer.
15159 In the varlist of the @code{let} expression, Emacs finds the file and
15160 binds the local variable @code{buffer} to the buffer containing the
15161 file. At the same time, Emacs creates @code{lengths-list} as a local
15164 Next, Emacs switches its attention to the buffer.
15166 In the following line, Emacs makes the buffer read-only. Ideally,
15167 this line is not necessary. None of the functions for counting words
15168 and symbols in a function definition should change the buffer.
15169 Besides, the buffer is not going to be saved, even if it were changed.
15170 This line is entirely the consequence of great, perhaps excessive,
15171 caution. The reason for the caution is that this function and those
15172 it calls work on the sources for Emacs and it is inconvenient if they
15173 are inadvertently modified. It goes without saying that I did not
15174 realize a need for this line until an experiment went awry and started
15175 to modify my Emacs source files @dots{}
15177 Next comes a call to widen the buffer if it is narrowed. This
15178 function is usually not needed---Emacs creates a fresh buffer if none
15179 already exists; but if a buffer visiting the file already exists Emacs
15180 returns that one. In this case, the buffer may be narrowed and must
15181 be widened. If we wanted to be fully `user-friendly', we would
15182 arrange to save the restriction and the location of point, but we
15185 The @code{(goto-char (point-min))} expression moves point to the
15186 beginning of the buffer.
15188 Then comes a @code{while} loop in which the `work' of the function is
15189 carried out. In the loop, Emacs determines the length of each
15190 definition and constructs a lengths' list containing the information.
15192 Emacs kills the buffer after working through it. This is to save
15193 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15194 source files of interest; GNU Emacs 22 contains over a thousand source
15195 files. Another function will apply @code{lengths-list-file} to each
15198 Finally, the last expression within the @code{let} expression is the
15199 @code{lengths-list} variable; its value is returned as the value of
15200 the whole function.
15202 You can try this function by installing it in the usual fashion. Then
15203 place your cursor after the following expression and type @kbd{C-x
15204 C-e} (@code{eval-last-sexp}).
15206 @c !!! 22.1.1 lisp sources location here
15209 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15213 (You may need to change the pathname of the file; the one here is for
15214 GNU Emacs version 22.1.1. To change the expression, copy it to
15215 the @file{*scratch*} buffer and edit it.
15219 (Also, to see the full length of the list, rather than a truncated
15220 version, you may have to evaluate the following:
15223 (custom-set-variables '(eval-expression-print-length nil))
15227 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15228 Then evaluate the @code{lengths-list-file} expression.)
15231 The lengths' list for @file{debug.el} takes less than a second to
15232 produce and looks like this in GNU Emacs 22:
15235 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15239 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15240 took seven seconds to produce and looked like this:
15243 (75 41 80 62 20 45 44 68 45 12 34 235)
15246 (The newer version of @file{debug.el} contains more defuns than the
15247 earlier one; and my new machine is much faster than the old one.)
15249 Note that the length of the last definition in the file is first in
15252 @node Several files
15253 @section Count Words in @code{defuns} in Different Files
15255 In the previous section, we created a function that returns a list of
15256 the lengths of each definition in a file. Now, we want to define a
15257 function to return a master list of the lengths of the definitions in
15260 Working on each of a list of files is a repetitious act, so we can use
15261 either a @code{while} loop or recursion.
15264 * lengths-list-many-files:: Return a list of the lengths of defuns.
15265 * append:: Attach one list to another.
15269 @node lengths-list-many-files
15270 @unnumberedsubsec Determine the lengths of @code{defuns}
15273 The design using a @code{while} loop is routine. The argument passed
15274 the function is a list of files. As we saw earlier (@pxref{Loop
15275 Example}), you can write a @code{while} loop so that the body of the
15276 loop is evaluated if such a list contains elements, but to exit the
15277 loop if the list is empty. For this design to work, the body of the
15278 loop must contain an expression that shortens the list each time the
15279 body is evaluated, so that eventually the list is empty. The usual
15280 technique is to set the value of the list to the value of the @sc{cdr}
15281 of the list each time the body is evaluated.
15284 The template looks like this:
15288 (while @var{test-whether-list-is-empty}
15290 @var{set-list-to-cdr-of-list})
15294 Also, we remember that a @code{while} loop returns @code{nil} (the
15295 result of evaluating the true-or-false-test), not the result of any
15296 evaluation within its body. (The evaluations within the body of the
15297 loop are done for their side effects.) However, the expression that
15298 sets the lengths' list is part of the body---and that is the value
15299 that we want returned by the function as a whole. To do this, we
15300 enclose the @code{while} loop within a @code{let} expression, and
15301 arrange that the last element of the @code{let} expression contains
15302 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15303 Example with an Incrementing Counter}.)
15305 @findex lengths-list-many-files
15307 These considerations lead us directly to the function itself:
15311 ;;; @r{Use @code{while} loop.}
15312 (defun lengths-list-many-files (list-of-files)
15313 "Return list of lengths of defuns in LIST-OF-FILES."
15316 (let (lengths-list)
15318 ;;; @r{true-or-false-test}
15319 (while list-of-files
15324 ;;; @r{Generate a lengths' list.}
15326 (expand-file-name (car list-of-files)))))
15330 ;;; @r{Make files' list shorter.}
15331 (setq list-of-files (cdr list-of-files)))
15333 ;;; @r{Return final value of lengths' list.}
15338 @code{expand-file-name} is a built-in function that converts a file
15339 name to the absolute, long, path name form. The function employs the
15340 name of the directory in which the function is called.
15342 @c !!! 22.1.1 lisp sources location here
15344 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15345 Emacs is visiting the
15346 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15356 @c !!! 22.1.1 lisp sources location here
15358 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15361 The only other new element of this function definition is the as yet
15362 unstudied function @code{append}, which merits a short section for
15366 @subsection The @code{append} Function
15369 The @code{append} function attaches one list to another. Thus,
15372 (append '(1 2 3 4) '(5 6 7 8))
15383 This is exactly how we want to attach two lengths' lists produced by
15384 @code{lengths-list-file} to each other. The results contrast with
15388 (cons '(1 2 3 4) '(5 6 7 8))
15393 which constructs a new list in which the first argument to @code{cons}
15394 becomes the first element of the new list:
15397 ((1 2 3 4) 5 6 7 8)
15400 @node Several files recursively
15401 @section Recursively Count Words in Different Files
15403 Besides a @code{while} loop, you can work on each of a list of files
15404 with recursion. A recursive version of @code{lengths-list-many-files}
15405 is short and simple.
15407 The recursive function has the usual parts: the `do-again-test', the
15408 `next-step-expression', and the recursive call. The `do-again-test'
15409 determines whether the function should call itself again, which it
15410 will do if the @code{list-of-files} contains any remaining elements;
15411 the `next-step-expression' resets the @code{list-of-files} to the
15412 @sc{cdr} of itself, so eventually the list will be empty; and the
15413 recursive call calls itself on the shorter list. The complete
15414 function is shorter than this description!
15415 @findex recursive-lengths-list-many-files
15419 (defun recursive-lengths-list-many-files (list-of-files)
15420 "Return list of lengths of each defun in LIST-OF-FILES."
15421 (if list-of-files ; @r{do-again-test}
15424 (expand-file-name (car list-of-files)))
15425 (recursive-lengths-list-many-files
15426 (cdr list-of-files)))))
15431 In a sentence, the function returns the lengths' list for the first of
15432 the @code{list-of-files} appended to the result of calling itself on
15433 the rest of the @code{list-of-files}.
15435 Here is a test of @code{recursive-lengths-list-many-files}, along with
15436 the results of running @code{lengths-list-file} on each of the files
15439 Install @code{recursive-lengths-list-many-files} and
15440 @code{lengths-list-file}, if necessary, and then evaluate the
15441 following expressions. You may need to change the files' pathnames;
15442 those here work when this Info file and the Emacs sources are located
15443 in their customary places. To change the expressions, copy them to
15444 the @file{*scratch*} buffer, edit them, and then evaluate them.
15446 The results are shown after the @samp{@result{}}. (These results are
15447 for files from Emacs version 22.1.1; files from other versions of
15448 Emacs may produce different results.)
15450 @c !!! 22.1.1 lisp sources location here
15453 (cd "/usr/local/share/emacs/22.1.1/")
15455 (lengths-list-file "./lisp/macros.el")
15456 @result{} (283 263 480 90)
15460 (lengths-list-file "./lisp/mail/mailalias.el")
15461 @result{} (38 32 29 95 178 180 321 218 324)
15465 (lengths-list-file "./lisp/makesum.el")
15470 (recursive-lengths-list-many-files
15471 '("./lisp/macros.el"
15472 "./lisp/mail/mailalias.el"
15473 "./lisp/makesum.el"))
15474 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15478 The @code{recursive-lengths-list-many-files} function produces the
15481 The next step is to prepare the data in the list for display in a graph.
15483 @node Prepare the data
15484 @section Prepare the Data for Display in a Graph
15486 The @code{recursive-lengths-list-many-files} function returns a list
15487 of numbers. Each number records the length of a function definition.
15488 What we need to do now is transform this data into a list of numbers
15489 suitable for generating a graph. The new list will tell how many
15490 functions definitions contain less than 10 words and
15491 symbols, how many contain between 10 and 19 words and symbols, how
15492 many contain between 20 and 29 words and symbols, and so on.
15494 In brief, we need to go through the lengths' list produced by the
15495 @code{recursive-lengths-list-many-files} function and count the number
15496 of defuns within each range of lengths, and produce a list of those
15500 * Data for Display in Detail::
15501 * Sorting:: Sorting lists.
15502 * Files List:: Making a list of files.
15503 * Counting function definitions::
15507 @node Data for Display in Detail
15508 @unnumberedsubsec The Data for Display in Detail
15511 Based on what we have done before, we can readily foresee that it
15512 should not be too hard to write a function that `@sc{cdr}s' down the
15513 lengths' list, looks at each element, determines which length range it
15514 is in, and increments a counter for that range.
15516 However, before beginning to write such a function, we should consider
15517 the advantages of sorting the lengths' list first, so the numbers are
15518 ordered from smallest to largest. First, sorting will make it easier
15519 to count the numbers in each range, since two adjacent numbers will
15520 either be in the same length range or in adjacent ranges. Second, by
15521 inspecting a sorted list, we can discover the highest and lowest
15522 number, and thereby determine the largest and smallest length range
15526 @subsection Sorting Lists
15529 Emacs contains a function to sort lists, called (as you might guess)
15530 @code{sort}. The @code{sort} function takes two arguments, the list
15531 to be sorted, and a predicate that determines whether the first of
15532 two list elements is ``less'' than the second.
15534 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15535 Type Object as an Argument}), a predicate is a function that
15536 determines whether some property is true or false. The @code{sort}
15537 function will reorder a list according to whatever property the
15538 predicate uses; this means that @code{sort} can be used to sort
15539 non-numeric lists by non-numeric criteria---it can, for example,
15540 alphabetize a list.
15543 The @code{<} function is used when sorting a numeric list. For example,
15546 (sort '(4 8 21 17 33 7 21 7) '<)
15554 (4 7 7 8 17 21 21 33)
15558 (Note that in this example, both the arguments are quoted so that the
15559 symbols are not evaluated before being passed to @code{sort} as
15562 Sorting the list returned by the
15563 @code{recursive-lengths-list-many-files} function is straightforward;
15564 it uses the @code{<} function:
15568 In GNU Emacs 22, eval
15570 (cd "/usr/local/share/emacs/22.0.50/")
15572 (recursive-lengths-list-many-files
15573 '("./lisp/macros.el"
15574 "./lisp/mail/mailalias.el"
15575 "./lisp/makesum.el"))
15583 (recursive-lengths-list-many-files
15584 '("./lisp/macros.el"
15585 "./lisp/mailalias.el"
15586 "./lisp/makesum.el"))
15596 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15600 (Note that in this example, the first argument to @code{sort} is not
15601 quoted, since the expression must be evaluated so as to produce the
15602 list that is passed to @code{sort}.)
15605 @subsection Making a List of Files
15607 The @code{recursive-lengths-list-many-files} function requires a list
15608 of files as its argument. For our test examples, we constructed such
15609 a list by hand; but the Emacs Lisp source directory is too large for
15610 us to do for that. Instead, we will write a function to do the job
15611 for us. In this function, we will use both a @code{while} loop and a
15614 @findex directory-files
15615 We did not have to write a function like this for older versions of
15616 GNU Emacs, since they placed all the @samp{.el} files in one
15617 directory. Instead, we were able to use the @code{directory-files}
15618 function, which lists the names of files that match a specified
15619 pattern within a single directory.
15621 However, recent versions of Emacs place Emacs Lisp files in
15622 sub-directories of the top level @file{lisp} directory. This
15623 re-arrangement eases navigation. For example, all the mail related
15624 files are in a @file{lisp} sub-directory called @file{mail}. But at
15625 the same time, this arrangement forces us to create a file listing
15626 function that descends into the sub-directories.
15628 @findex files-in-below-directory
15629 We can create this function, called @code{files-in-below-directory},
15630 using familiar functions such as @code{car}, @code{nthcdr}, and
15631 @code{substring} in conjunction with an existing function called
15632 @code{directory-files-and-attributes}. This latter function not only
15633 lists all the filenames in a directory, including the names
15634 of sub-directories, but also their attributes.
15636 To restate our goal: to create a function that will enable us
15637 to feed filenames to @code{recursive-lengths-list-many-files}
15638 as a list that looks like this (but with more elements):
15642 ("./lisp/macros.el"
15643 "./lisp/mail/rmail.el"
15644 "./lisp/makesum.el")
15648 The @code{directory-files-and-attributes} function returns a list of
15649 lists. Each of the lists within the main list consists of 13
15650 elements. The first element is a string that contains the name of the
15651 file---which, in GNU/Linux, may be a `directory file', that is to
15652 say, a file with the special attributes of a directory. The second
15653 element of the list is @code{t} for a directory, a string
15654 for symbolic link (the string is the name linked to), or @code{nil}.
15656 For example, the first @samp{.el} file in the @file{lisp/} directory
15657 is @file{abbrev.el}. Its name is
15658 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15659 directory or a symbolic link.
15662 This is how @code{directory-files-and-attributes} lists that file and
15674 (20615 27034 579989 697000)
15676 (20615 26327 734791 805000)
15688 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15689 directory. The beginning of its listing looks like this:
15700 (To learn about the different attributes, look at the documentation of
15701 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15702 function does not list the filename, so its first element is
15703 @code{directory-files-and-attributes}'s second element.)
15705 We will want our new function, @code{files-in-below-directory}, to
15706 list the @samp{.el} files in the directory it is told to check, and in
15707 any directories below that directory.
15709 This gives us a hint on how to construct
15710 @code{files-in-below-directory}: within a directory, the function
15711 should add @samp{.el} filenames to a list; and if, within a directory,
15712 the function comes upon a sub-directory, it should go into that
15713 sub-directory and repeat its actions.
15715 However, we should note that every directory contains a name that
15716 refers to itself, called @file{.}, (``dot'') and a name that refers to
15717 its parent directory, called @file{..} (``double dot''). (In
15718 @file{/}, the root directory, @file{..} refers to itself, since
15719 @file{/} has no parent.) Clearly, we do not want our
15720 @code{files-in-below-directory} function to enter those directories,
15721 since they always lead us, directly or indirectly, to the current
15724 Consequently, our @code{files-in-below-directory} function must do
15729 Check to see whether it is looking at a filename that ends in
15730 @samp{.el}; and if so, add its name to a list.
15733 Check to see whether it is looking at a filename that is the name of a
15734 directory; and if so,
15738 Check to see whether it is looking at @file{.} or @file{..}; and if
15742 Or else, go into that directory and repeat the process.
15746 Let's write a function definition to do these tasks. We will use a
15747 @code{while} loop to move from one filename to another within a
15748 directory, checking what needs to be done; and we will use a recursive
15749 call to repeat the actions on each sub-directory. The recursive
15750 pattern is `accumulate'
15751 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15752 using @code{append} as the combiner.
15755 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15756 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15758 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15759 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15762 @c /usr/local/share/emacs/22.1.1/lisp/
15765 Here is the function:
15769 (defun files-in-below-directory (directory)
15770 "List the .el files in DIRECTORY and in its sub-directories."
15771 ;; Although the function will be used non-interactively,
15772 ;; it will be easier to test if we make it interactive.
15773 ;; The directory will have a name such as
15774 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15775 (interactive "DDirectory name: ")
15778 (let (el-files-list
15779 (current-directory-list
15780 (directory-files-and-attributes directory t)))
15781 ;; while we are in the current directory
15782 (while current-directory-list
15786 ;; check to see whether filename ends in `.el'
15787 ;; and if so, append its name to a list.
15788 ((equal ".el" (substring (car (car current-directory-list)) -3))
15789 (setq el-files-list
15790 (cons (car (car current-directory-list)) el-files-list)))
15793 ;; check whether filename is that of a directory
15794 ((eq t (car (cdr (car current-directory-list))))
15795 ;; decide whether to skip or recurse
15798 (substring (car (car current-directory-list)) -1))
15799 ;; then do nothing since filename is that of
15800 ;; current directory or parent, "." or ".."
15804 ;; else descend into the directory and repeat the process
15805 (setq el-files-list
15807 (files-in-below-directory
15808 (car (car current-directory-list)))
15810 ;; move to the next filename in the list; this also
15811 ;; shortens the list so the while loop eventually comes to an end
15812 (setq current-directory-list (cdr current-directory-list)))
15813 ;; return the filenames
15818 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15819 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15821 The @code{files-in-below-directory} @code{directory-files} function
15822 takes one argument, the name of a directory.
15825 Thus, on my system,
15827 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15829 @c !!! 22.1.1 lisp sources location here
15833 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15838 tells me that in and below my Lisp sources directory are 1031
15841 @code{files-in-below-directory} returns a list in reverse alphabetical
15842 order. An expression to sort the list in alphabetical order looks
15848 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15855 "Test how long it takes to find lengths of all sorted elisp defuns."
15856 (insert "\n" (current-time-string) "\n")
15859 (recursive-lengths-list-many-files
15860 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15862 (insert (format "%s" (current-time-string))))
15865 @node Counting function definitions
15866 @subsection Counting function definitions
15868 Our immediate goal is to generate a list that tells us how many
15869 function definitions contain fewer than 10 words and symbols, how many
15870 contain between 10 and 19 words and symbols, how many contain between
15871 20 and 29 words and symbols, and so on.
15873 With a sorted list of numbers, this is easy: count how many elements
15874 of the list are smaller than 10, then, after moving past the numbers
15875 just counted, count how many are smaller than 20, then, after moving
15876 past the numbers just counted, count how many are smaller than 30, and
15877 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
15878 larger than the top of that range. We can call the list of such
15879 numbers the @code{top-of-ranges} list.
15882 If we wished, we could generate this list automatically, but it is
15883 simpler to write a list manually. Here it is:
15884 @vindex top-of-ranges
15888 (defvar top-of-ranges
15891 110 120 130 140 150
15892 160 170 180 190 200
15893 210 220 230 240 250
15894 260 270 280 290 300)
15895 "List specifying ranges for `defuns-per-range'.")
15899 To change the ranges, we edit this list.
15901 Next, we need to write the function that creates the list of the
15902 number of definitions within each range. Clearly, this function must
15903 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
15906 The @code{defuns-per-range} function must do two things again and
15907 again: it must count the number of definitions within a range
15908 specified by the current top-of-range value; and it must shift to the
15909 next higher value in the @code{top-of-ranges} list after counting the
15910 number of definitions in the current range. Since each of these
15911 actions is repetitive, we can use @code{while} loops for the job.
15912 One loop counts the number of definitions in the range defined by the
15913 current top-of-range value, and the other loop selects each of the
15914 top-of-range values in turn.
15916 Several entries of the @code{sorted-lengths} list are counted for each
15917 range; this means that the loop for the @code{sorted-lengths} list
15918 will be inside the loop for the @code{top-of-ranges} list, like a
15919 small gear inside a big gear.
15921 The inner loop counts the number of definitions within the range. It
15922 is a simple counting loop of the type we have seen before.
15923 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
15924 The true-or-false test of the loop tests whether the value from the
15925 @code{sorted-lengths} list is smaller than the current value of the
15926 top of the range. If it is, the function increments the counter and
15927 tests the next value from the @code{sorted-lengths} list.
15930 The inner loop looks like this:
15934 (while @var{length-element-smaller-than-top-of-range}
15935 (setq number-within-range (1+ number-within-range))
15936 (setq sorted-lengths (cdr sorted-lengths)))
15940 The outer loop must start with the lowest value of the
15941 @code{top-of-ranges} list, and then be set to each of the succeeding
15942 higher values in turn. This can be done with a loop like this:
15946 (while top-of-ranges
15947 @var{body-of-loop}@dots{}
15948 (setq top-of-ranges (cdr top-of-ranges)))
15953 Put together, the two loops look like this:
15957 (while top-of-ranges
15959 ;; @r{Count the number of elements within the current range.}
15960 (while @var{length-element-smaller-than-top-of-range}
15961 (setq number-within-range (1+ number-within-range))
15962 (setq sorted-lengths (cdr sorted-lengths)))
15964 ;; @r{Move to next range.}
15965 (setq top-of-ranges (cdr top-of-ranges)))
15969 In addition, in each circuit of the outer loop, Emacs should record
15970 the number of definitions within that range (the value of
15971 @code{number-within-range}) in a list. We can use @code{cons} for
15972 this purpose. (@xref{cons, , @code{cons}}.)
15974 The @code{cons} function works fine, except that the list it
15975 constructs will contain the number of definitions for the highest
15976 range at its beginning and the number of definitions for the lowest
15977 range at its end. This is because @code{cons} attaches new elements
15978 of the list to the beginning of the list, and since the two loops are
15979 working their way through the lengths' list from the lower end first,
15980 the @code{defuns-per-range-list} will end up largest number first.
15981 But we will want to print our graph with smallest values first and the
15982 larger later. The solution is to reverse the order of the
15983 @code{defuns-per-range-list}. We can do this using the
15984 @code{nreverse} function, which reverses the order of a list.
15991 (nreverse '(1 2 3 4))
16002 Note that the @code{nreverse} function is ``destructive''---that is,
16003 it changes the list to which it is applied; this contrasts with the
16004 @code{car} and @code{cdr} functions, which are non-destructive. In
16005 this case, we do not want the original @code{defuns-per-range-list},
16006 so it does not matter that it is destroyed. (The @code{reverse}
16007 function provides a reversed copy of a list, leaving the original list
16012 Put all together, the @code{defuns-per-range} looks like this:
16016 (defun defuns-per-range (sorted-lengths top-of-ranges)
16017 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16018 (let ((top-of-range (car top-of-ranges))
16019 (number-within-range 0)
16020 defuns-per-range-list)
16025 (while top-of-ranges
16031 ;; @r{Need number for numeric test.}
16032 (car sorted-lengths)
16033 (< (car sorted-lengths) top-of-range))
16037 ;; @r{Count number of definitions within current range.}
16038 (setq number-within-range (1+ number-within-range))
16039 (setq sorted-lengths (cdr sorted-lengths)))
16041 ;; @r{Exit inner loop but remain within outer loop.}
16045 (setq defuns-per-range-list
16046 (cons number-within-range defuns-per-range-list))
16047 (setq number-within-range 0) ; @r{Reset count to zero.}
16051 ;; @r{Move to next range.}
16052 (setq top-of-ranges (cdr top-of-ranges))
16053 ;; @r{Specify next top of range value.}
16054 (setq top-of-range (car top-of-ranges)))
16058 ;; @r{Exit outer loop and count the number of defuns larger than}
16059 ;; @r{ the largest top-of-range value.}
16060 (setq defuns-per-range-list
16062 (length sorted-lengths)
16063 defuns-per-range-list))
16067 ;; @r{Return a list of the number of definitions within each range,}
16068 ;; @r{ smallest to largest.}
16069 (nreverse defuns-per-range-list)))
16075 The function is straightforward except for one subtle feature. The
16076 true-or-false test of the inner loop looks like this:
16080 (and (car sorted-lengths)
16081 (< (car sorted-lengths) top-of-range))
16087 instead of like this:
16090 (< (car sorted-lengths) top-of-range)
16093 The purpose of the test is to determine whether the first item in the
16094 @code{sorted-lengths} list is less than the value of the top of the
16097 The simple version of the test works fine unless the
16098 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16099 @code{(car sorted-lengths)} expression function returns
16100 @code{nil}. The @code{<} function cannot compare a number to
16101 @code{nil}, which is an empty list, so Emacs signals an error and
16102 stops the function from attempting to continue to execute.
16104 The @code{sorted-lengths} list always becomes @code{nil} when the
16105 counter reaches the end of the list. This means that any attempt to
16106 use the @code{defuns-per-range} function with the simple version of
16107 the test will fail.
16109 We solve the problem by using the @code{(car sorted-lengths)}
16110 expression in conjunction with the @code{and} expression. The
16111 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16112 value so long as the list has at least one number within it, but
16113 returns @code{nil} if the list is empty. The @code{and} expression
16114 first evaluates the @code{(car sorted-lengths)} expression, and
16115 if it is @code{nil}, returns false @emph{without} evaluating the
16116 @code{<} expression. But if the @code{(car sorted-lengths)}
16117 expression returns a non-@code{nil} value, the @code{and} expression
16118 evaluates the @code{<} expression, and returns that value as the value
16119 of the @code{and} expression.
16121 @c colon in printed section title causes problem in Info cross reference
16122 This way, we avoid an error.
16125 (For information about @code{and}, see
16126 @ref{kill-new function, , The @code{kill-new} function}.)
16130 (@xref{kill-new function, , The @code{kill-new} function}, for
16131 information about @code{and}.)
16134 Here is a short test of the @code{defuns-per-range} function. First,
16135 evaluate the expression that binds (a shortened)
16136 @code{top-of-ranges} list to the list of values, then evaluate the
16137 expression for binding the @code{sorted-lengths} list, and then
16138 evaluate the @code{defuns-per-range} function.
16142 ;; @r{(Shorter list than we will use later.)}
16143 (setq top-of-ranges
16144 '(110 120 130 140 150
16145 160 170 180 190 200))
16147 (setq sorted-lengths
16148 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16150 (defuns-per-range sorted-lengths top-of-ranges)
16156 The list returned looks like this:
16159 (2 2 2 0 0 1 0 2 0 0 4)
16163 Indeed, there are two elements of the @code{sorted-lengths} list
16164 smaller than 110, two elements between 110 and 119, two elements
16165 between 120 and 129, and so on. There are four elements with a value
16168 @c The next step is to turn this numbers' list into a graph.
16169 @node Readying a Graph
16170 @chapter Readying a Graph
16171 @cindex Readying a graph
16172 @cindex Graph prototype
16173 @cindex Prototype graph
16174 @cindex Body of graph
16176 Our goal is to construct a graph showing the numbers of function
16177 definitions of various lengths in the Emacs lisp sources.
16179 As a practical matter, if you were creating a graph, you would
16180 probably use a program such as @code{gnuplot} to do the job.
16181 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16182 however, we create one from scratch, and in the process we will
16183 re-acquaint ourselves with some of what we learned before and learn
16186 In this chapter, we will first write a simple graph printing function.
16187 This first definition will be a @dfn{prototype}, a rapidly written
16188 function that enables us to reconnoiter this unknown graph-making
16189 territory. We will discover dragons, or find that they are myth.
16190 After scouting the terrain, we will feel more confident and enhance
16191 the function to label the axes automatically.
16194 * Columns of a graph::
16195 * graph-body-print:: How to print the body of a graph.
16196 * recursive-graph-body-print::
16198 * Line Graph Exercise::
16202 @node Columns of a graph
16203 @unnumberedsec Printing the Columns of a Graph
16206 Since Emacs is designed to be flexible and work with all kinds of
16207 terminals, including character-only terminals, the graph will need to
16208 be made from one of the `typewriter' symbols. An asterisk will do; as
16209 we enhance the graph-printing function, we can make the choice of
16210 symbol a user option.
16212 We can call this function @code{graph-body-print}; it will take a
16213 @code{numbers-list} as its only argument. At this stage, we will not
16214 label the graph, but only print its body.
16216 The @code{graph-body-print} function inserts a vertical column of
16217 asterisks for each element in the @code{numbers-list}. The height of
16218 each line is determined by the value of that element of the
16219 @code{numbers-list}.
16221 Inserting columns is a repetitive act; that means that this function can
16222 be written either with a @code{while} loop or recursively.
16224 Our first challenge is to discover how to print a column of asterisks.
16225 Usually, in Emacs, we print characters onto a screen horizontally,
16226 line by line, by typing. We have two routes we can follow: write our
16227 own column-insertion function or discover whether one exists in Emacs.
16229 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16230 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16231 command, except that the latter finds only those functions that are
16232 commands. The @kbd{M-x apropos} command lists all symbols that match
16233 a regular expression, including functions that are not interactive.
16236 What we want to look for is some command that prints or inserts
16237 columns. Very likely, the name of the function will contain either
16238 the word `print' or the word `insert' or the word `column'.
16239 Therefore, we can simply type @kbd{M-x apropos RET
16240 print\|insert\|column RET} and look at the result. On my system, this
16241 command once too takes quite some time, and then produced a list of 79
16242 functions and variables. Now it does not take much time at all and
16243 produces a list of 211 functions and variables. Scanning down the
16244 list, the only function that looks as if it might do the job is
16245 @code{insert-rectangle}.
16248 Indeed, this is the function we want; its documentation says:
16253 Insert text of RECTANGLE with upper left corner at point.
16254 RECTANGLE's first line is inserted at point,
16255 its second line is inserted at a point vertically under point, etc.
16256 RECTANGLE should be a list of strings.
16257 After this command, the mark is at the upper left corner
16258 and point is at the lower right corner.
16262 We can run a quick test, to make sure it does what we expect of it.
16264 Here is the result of placing the cursor after the
16265 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16266 (@code{eval-last-sexp}). The function inserts the strings
16267 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16268 point. Also the function returns @code{nil}.
16272 (insert-rectangle '("first" "second" "third"))first
16279 Of course, we won't be inserting the text of the
16280 @code{insert-rectangle} expression itself into the buffer in which we
16281 are making the graph, but will call the function from our program. We
16282 shall, however, have to make sure that point is in the buffer at the
16283 place where the @code{insert-rectangle} function will insert its
16286 If you are reading this in Info, you can see how this works by
16287 switching to another buffer, such as the @file{*scratch*} buffer,
16288 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16289 @code{insert-rectangle} expression into the minibuffer at the prompt,
16290 and then typing @key{RET}. This causes Emacs to evaluate the
16291 expression in the minibuffer, but to use as the value of point the
16292 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16293 keybinding for @code{eval-expression}. Also, @code{nil} does not
16294 appear in the @file{*scratch*} buffer since the expression is
16295 evaluated in the minibuffer.)
16297 We find when we do this that point ends up at the end of the last
16298 inserted line---that is to say, this function moves point as a
16299 side-effect. If we were to repeat the command, with point at this
16300 position, the next insertion would be below and to the right of the
16301 previous insertion. We don't want this! If we are going to make a
16302 bar graph, the columns need to be beside each other.
16304 So we discover that each cycle of the column-inserting @code{while}
16305 loop must reposition point to the place we want it, and that place
16306 will be at the top, not the bottom, of the column. Moreover, we
16307 remember that when we print a graph, we do not expect all the columns
16308 to be the same height. This means that the top of each column may be
16309 at a different height from the previous one. We cannot simply
16310 reposition point to the same line each time, but moved over to the
16311 right---or perhaps we can@dots{}
16313 We are planning to make the columns of the bar graph out of asterisks.
16314 The number of asterisks in the column is the number specified by the
16315 current element of the @code{numbers-list}. We need to construct a
16316 list of asterisks of the right length for each call to
16317 @code{insert-rectangle}. If this list consists solely of the requisite
16318 number of asterisks, then we will have position point the right number
16319 of lines above the base for the graph to print correctly. This could
16322 Alternatively, if we can figure out some way to pass
16323 @code{insert-rectangle} a list of the same length each time, then we
16324 can place point on the same line each time, but move it over one
16325 column to the right for each new column. If we do this, however, some
16326 of the entries in the list passed to @code{insert-rectangle} must be
16327 blanks rather than asterisks. For example, if the maximum height of
16328 the graph is 5, but the height of the column is 3, then
16329 @code{insert-rectangle} requires an argument that looks like this:
16332 (" " " " "*" "*" "*")
16335 This last proposal is not so difficult, so long as we can determine
16336 the column height. There are two ways for us to specify the column
16337 height: we can arbitrarily state what it will be, which would work
16338 fine for graphs of that height; or we can search through the list of
16339 numbers and use the maximum height of the list as the maximum height
16340 of the graph. If the latter operation were difficult, then the former
16341 procedure would be easiest, but there is a function built into Emacs
16342 that determines the maximum of its arguments. We can use that
16343 function. The function is called @code{max} and it returns the
16344 largest of all its arguments, which must be numbers. Thus, for
16352 returns 7. (A corresponding function called @code{min} returns the
16353 smallest of all its arguments.)
16357 However, we cannot simply call @code{max} on the @code{numbers-list};
16358 the @code{max} function expects numbers as its argument, not a list of
16359 numbers. Thus, the following expression,
16362 (max '(3 4 6 5 7 3))
16367 produces the following error message;
16370 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16374 We need a function that passes a list of arguments to a function.
16375 This function is @code{apply}. This function `applies' its first
16376 argument (a function) to its remaining arguments, the last of which
16383 (apply 'max 3 4 7 3 '(4 8 5))
16389 (Incidentally, I don't know how you would learn of this function
16390 without a book such as this. It is possible to discover other
16391 functions, like @code{search-forward} or @code{insert-rectangle}, by
16392 guessing at a part of their names and then using @code{apropos}. Even
16393 though its base in metaphor is clear---`apply' its first argument to
16394 the rest---I doubt a novice would come up with that particular word
16395 when using @code{apropos} or other aid. Of course, I could be wrong;
16396 after all, the function was first named by someone who had to invent
16399 The second and subsequent arguments to @code{apply} are optional, so
16400 we can use @code{apply} to call a function and pass the elements of a
16401 list to it, like this, which also returns 8:
16404 (apply 'max '(4 8 5))
16407 This latter way is how we will use @code{apply}. The
16408 @code{recursive-lengths-list-many-files} function returns a numbers'
16409 list to which we can apply @code{max} (we could also apply @code{max} to
16410 the sorted numbers' list; it does not matter whether the list is
16414 Hence, the operation for finding the maximum height of the graph is this:
16417 (setq max-graph-height (apply 'max numbers-list))
16420 Now we can return to the question of how to create a list of strings
16421 for a column of the graph. Told the maximum height of the graph
16422 and the number of asterisks that should appear in the column, the
16423 function should return a list of strings for the
16424 @code{insert-rectangle} command to insert.
16426 Each column is made up of asterisks or blanks. Since the function is
16427 passed the value of the height of the column and the number of
16428 asterisks in the column, the number of blanks can be found by
16429 subtracting the number of asterisks from the height of the column.
16430 Given the number of blanks and the number of asterisks, two
16431 @code{while} loops can be used to construct the list:
16435 ;;; @r{First version.}
16436 (defun column-of-graph (max-graph-height actual-height)
16437 "Return list of strings that is one column of a graph."
16438 (let ((insert-list nil)
16439 (number-of-top-blanks
16440 (- max-graph-height actual-height)))
16444 ;; @r{Fill in asterisks.}
16445 (while (> actual-height 0)
16446 (setq insert-list (cons "*" insert-list))
16447 (setq actual-height (1- actual-height)))
16451 ;; @r{Fill in blanks.}
16452 (while (> number-of-top-blanks 0)
16453 (setq insert-list (cons " " insert-list))
16454 (setq number-of-top-blanks
16455 (1- number-of-top-blanks)))
16459 ;; @r{Return whole list.}
16464 If you install this function and then evaluate the following
16465 expression you will see that it returns the list as desired:
16468 (column-of-graph 5 3)
16476 (" " " " "*" "*" "*")
16479 As written, @code{column-of-graph} contains a major flaw: the symbols
16480 used for the blank and for the marked entries in the column are
16481 `hard-coded' as a space and asterisk. This is fine for a prototype,
16482 but you, or another user, may wish to use other symbols. For example,
16483 in testing the graph function, you many want to use a period in place
16484 of the space, to make sure the point is being repositioned properly
16485 each time the @code{insert-rectangle} function is called; or you might
16486 want to substitute a @samp{+} sign or other symbol for the asterisk.
16487 You might even want to make a graph-column that is more than one
16488 display column wide. The program should be more flexible. The way to
16489 do that is to replace the blank and the asterisk with two variables
16490 that we can call @code{graph-blank} and @code{graph-symbol} and define
16491 those variables separately.
16493 Also, the documentation is not well written. These considerations
16494 lead us to the second version of the function:
16498 (defvar graph-symbol "*"
16499 "String used as symbol in graph, usually an asterisk.")
16503 (defvar graph-blank " "
16504 "String used as blank in graph, usually a blank space.
16505 graph-blank must be the same number of columns wide
16511 (For an explanation of @code{defvar}, see
16512 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16516 ;;; @r{Second version.}
16517 (defun column-of-graph (max-graph-height actual-height)
16518 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16522 The graph-symbols are contiguous entries at the end
16524 The list will be inserted as one column of a graph.
16525 The strings are either graph-blank or graph-symbol."
16529 (let ((insert-list nil)
16530 (number-of-top-blanks
16531 (- max-graph-height actual-height)))
16535 ;; @r{Fill in @code{graph-symbols}.}
16536 (while (> actual-height 0)
16537 (setq insert-list (cons graph-symbol insert-list))
16538 (setq actual-height (1- actual-height)))
16542 ;; @r{Fill in @code{graph-blanks}.}
16543 (while (> number-of-top-blanks 0)
16544 (setq insert-list (cons graph-blank insert-list))
16545 (setq number-of-top-blanks
16546 (1- number-of-top-blanks)))
16548 ;; @r{Return whole list.}
16553 If we wished, we could rewrite @code{column-of-graph} a third time to
16554 provide optionally for a line graph as well as for a bar graph. This
16555 would not be hard to do. One way to think of a line graph is that it
16556 is no more than a bar graph in which the part of each bar that is
16557 below the top is blank. To construct a column for a line graph, the
16558 function first constructs a list of blanks that is one shorter than
16559 the value, then it uses @code{cons} to attach a graph symbol to the
16560 list; then it uses @code{cons} again to attach the `top blanks' to
16563 It is easy to see how to write such a function, but since we don't
16564 need it, we will not do it. But the job could be done, and if it were
16565 done, it would be done with @code{column-of-graph}. Even more
16566 important, it is worth noting that few changes would have to be made
16567 anywhere else. The enhancement, if we ever wish to make it, is
16570 Now, finally, we come to our first actual graph printing function.
16571 This prints the body of a graph, not the labels for the vertical and
16572 horizontal axes, so we can call this @code{graph-body-print}.
16574 @node graph-body-print
16575 @section The @code{graph-body-print} Function
16576 @findex graph-body-print
16578 After our preparation in the preceding section, the
16579 @code{graph-body-print} function is straightforward. The function
16580 will print column after column of asterisks and blanks, using the
16581 elements of a numbers' list to specify the number of asterisks in each
16582 column. This is a repetitive act, which means we can use a
16583 decrementing @code{while} loop or recursive function for the job. In
16584 this section, we will write the definition using a @code{while} loop.
16586 The @code{column-of-graph} function requires the height of the graph
16587 as an argument, so we should determine and record that as a local variable.
16589 This leads us to the following template for the @code{while} loop
16590 version of this function:
16594 (defun graph-body-print (numbers-list)
16595 "@var{documentation}@dots{}"
16596 (let ((height @dots{}
16601 (while numbers-list
16602 @var{insert-columns-and-reposition-point}
16603 (setq numbers-list (cdr numbers-list)))))
16608 We need to fill in the slots of the template.
16610 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16611 determine the height of the graph.
16613 The @code{while} loop will cycle through the @code{numbers-list} one
16614 element at a time. As it is shortened by the @code{(setq numbers-list
16615 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16616 list is the value of the argument for @code{column-of-graph}.
16618 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16619 function inserts the list returned by @code{column-of-graph}. Since
16620 the @code{insert-rectangle} function moves point to the lower right of
16621 the inserted rectangle, we need to save the location of point at the
16622 time the rectangle is inserted, move back to that position after the
16623 rectangle is inserted, and then move horizontally to the next place
16624 from which @code{insert-rectangle} is called.
16626 If the inserted columns are one character wide, as they will be if
16627 single blanks and asterisks are used, the repositioning command is
16628 simply @code{(forward-char 1)}; however, the width of a column may be
16629 greater than one. This means that the repositioning command should be
16630 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16631 itself is the length of a @code{graph-blank} and can be found using
16632 the expression @code{(length graph-blank)}. The best place to bind
16633 the @code{symbol-width} variable to the value of the width of graph
16634 column is in the varlist of the @code{let} expression.
16637 These considerations lead to the following function definition:
16641 (defun graph-body-print (numbers-list)
16642 "Print a bar graph of the NUMBERS-LIST.
16643 The numbers-list consists of the Y-axis values."
16645 (let ((height (apply 'max numbers-list))
16646 (symbol-width (length graph-blank))
16651 (while numbers-list
16652 (setq from-position (point))
16654 (column-of-graph height (car numbers-list)))
16655 (goto-char from-position)
16656 (forward-char symbol-width)
16659 ;; @r{Draw graph column by column.}
16661 (setq numbers-list (cdr numbers-list)))
16664 ;; @r{Place point for X axis labels.}
16665 (forward-line height)
16672 The one unexpected expression in this function is the
16673 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16674 expression makes the graph printing operation more interesting to
16675 watch than it would be otherwise. The expression causes Emacs to
16676 `sit' or do nothing for a zero length of time and then redraw the
16677 screen. Placed here, it causes Emacs to redraw the screen column by
16678 column. Without it, Emacs would not redraw the screen until the
16681 We can test @code{graph-body-print} with a short list of numbers.
16685 Install @code{graph-symbol}, @code{graph-blank},
16686 @code{column-of-graph}, which are in
16688 @ref{Readying a Graph, , Readying a Graph},
16691 @ref{Columns of a graph},
16693 and @code{graph-body-print}.
16697 Copy the following expression:
16700 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16704 Switch to the @file{*scratch*} buffer and place the cursor where you
16705 want the graph to start.
16708 Type @kbd{M-:} (@code{eval-expression}).
16711 Yank the @code{graph-body-print} expression into the minibuffer
16712 with @kbd{C-y} (@code{yank)}.
16715 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16719 Emacs will print a graph like this:
16733 @node recursive-graph-body-print
16734 @section The @code{recursive-graph-body-print} Function
16735 @findex recursive-graph-body-print
16737 The @code{graph-body-print} function may also be written recursively.
16738 The recursive solution is divided into two parts: an outside `wrapper'
16739 that uses a @code{let} expression to determine the values of several
16740 variables that need only be found once, such as the maximum height of
16741 the graph, and an inside function that is called recursively to print
16745 The `wrapper' is uncomplicated:
16749 (defun recursive-graph-body-print (numbers-list)
16750 "Print a bar graph of the NUMBERS-LIST.
16751 The numbers-list consists of the Y-axis values."
16752 (let ((height (apply 'max numbers-list))
16753 (symbol-width (length graph-blank))
16755 (recursive-graph-body-print-internal
16762 The recursive function is a little more difficult. It has four parts:
16763 the `do-again-test', the printing code, the recursive call, and the
16764 `next-step-expression'. The `do-again-test' is a @code{when}
16765 expression that determines whether the @code{numbers-list} contains
16766 any remaining elements; if it does, the function prints one column of
16767 the graph using the printing code and calls itself again. The
16768 function calls itself again according to the value produced by the
16769 `next-step-expression' which causes the call to act on a shorter
16770 version of the @code{numbers-list}.
16774 (defun recursive-graph-body-print-internal
16775 (numbers-list height symbol-width)
16776 "Print a bar graph.
16777 Used within recursive-graph-body-print function."
16782 (setq from-position (point))
16784 (column-of-graph height (car numbers-list)))
16787 (goto-char from-position)
16788 (forward-char symbol-width)
16789 (sit-for 0) ; @r{Draw graph column by column.}
16790 (recursive-graph-body-print-internal
16791 (cdr numbers-list) height symbol-width)))
16796 After installation, this expression can be tested; here is a sample:
16799 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16803 Here is what @code{recursive-graph-body-print} produces:
16817 Either of these two functions, @code{graph-body-print} or
16818 @code{recursive-graph-body-print}, create the body of a graph.
16821 @section Need for Printed Axes
16823 A graph needs printed axes, so you can orient yourself. For a do-once
16824 project, it may be reasonable to draw the axes by hand using Emacs's
16825 Picture mode; but a graph drawing function may be used more than once.
16827 For this reason, I have written enhancements to the basic
16828 @code{print-graph-body} function that automatically print labels for
16829 the horizontal and vertical axes. Since the label printing functions
16830 do not contain much new material, I have placed their description in
16831 an appendix. @xref{Full Graph, , A Graph with Labeled Axes}.
16833 @node Line Graph Exercise
16836 Write a line graph version of the graph printing functions.
16838 @node Emacs Initialization
16839 @chapter Your @file{.emacs} File
16840 @cindex @file{.emacs} file
16841 @cindex Customizing your @file{.emacs} file
16842 @cindex Initialization file
16844 ``You don't have to like Emacs to like it''---this seemingly
16845 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16846 the box' Emacs is a generic tool. Most people who use it, customize
16847 it to suit themselves.
16849 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16850 expressions in Emacs Lisp you can change or extend Emacs.
16853 * Default Configuration::
16854 * Site-wide Init:: You can write site-wide init files.
16855 * defcustom:: Emacs will write code for you.
16856 * Beginning a .emacs File:: How to write a @code{.emacs file}.
16857 * Text and Auto-fill:: Automatically wrap lines.
16858 * Mail Aliases:: Use abbreviations for email addresses.
16859 * Indent Tabs Mode:: Don't use tabs with @TeX{}
16860 * Keybindings:: Create some personal keybindings.
16861 * Keymaps:: More about key binding.
16862 * Loading Files:: Load (i.e., evaluate) files automatically.
16863 * Autoload:: Make functions available.
16864 * Simple Extension:: Define a function; bind it to a key.
16865 * X11 Colors:: Colors in X.
16867 * Mode Line:: How to customize your mode line.
16871 @node Default Configuration
16872 @unnumberedsec Emacs's Default Configuration
16875 There are those who appreciate Emacs's default configuration. After
16876 all, Emacs starts you in C mode when you edit a C file, starts you in
16877 Fortran mode when you edit a Fortran file, and starts you in
16878 Fundamental mode when you edit an unadorned file. This all makes
16879 sense, if you do not know who is going to use Emacs. Who knows what a
16880 person hopes to do with an unadorned file? Fundamental mode is the
16881 right default for such a file, just as C mode is the right default for
16882 editing C code. (Enough programming languages have syntaxes
16883 that enable them to share or nearly share features, so C mode is
16884 now provided by CC mode, the `C Collection'.)
16886 But when you do know who is going to use Emacs---you,
16887 yourself---then it makes sense to customize Emacs.
16889 For example, I seldom want Fundamental mode when I edit an
16890 otherwise undistinguished file; I want Text mode. This is why I
16891 customize Emacs: so it suits me.
16893 You can customize and extend Emacs by writing or adapting a
16894 @file{~/.emacs} file. This is your personal initialization file; its
16895 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
16896 may also add @file{.el} to @file{~/.emacs} and call it a
16897 @file{~/.emacs.el} file. In the past, you were forbidden to type the
16898 extra keystrokes that the name @file{~/.emacs.el} requires, but now
16899 you may. The new format is consistent with the Emacs Lisp file
16900 naming conventions; the old format saves typing.}
16902 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
16903 code yourself; or you can use Emacs's @code{customize} feature to write
16904 the code for you. You can combine your own expressions and
16905 auto-written Customize expressions in your @file{.emacs} file.
16907 (I myself prefer to write my own expressions, except for those,
16908 particularly fonts, that I find easier to manipulate using the
16909 @code{customize} command. I combine the two methods.)
16911 Most of this chapter is about writing expressions yourself. It
16912 describes a simple @file{.emacs} file; for more information, see
16913 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
16914 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
16917 @node Site-wide Init
16918 @section Site-wide Initialization Files
16920 @cindex @file{default.el} init file
16921 @cindex @file{site-init.el} init file
16922 @cindex @file{site-load.el} init file
16923 In addition to your personal initialization file, Emacs automatically
16924 loads various site-wide initialization files, if they exist. These
16925 have the same form as your @file{.emacs} file, but are loaded by
16928 Two site-wide initialization files, @file{site-load.el} and
16929 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
16930 `dumped' version of Emacs is created, as is most common. (Dumped
16931 copies of Emacs load more quickly. However, once a file is loaded and
16932 dumped, a change to it does not lead to a change in Emacs unless you
16933 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
16934 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
16935 @file{INSTALL} file.)
16937 Three other site-wide initialization files are loaded automatically
16938 each time you start Emacs, if they exist. These are
16939 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
16940 file, and @file{default.el}, and the terminal type file, which are both
16941 loaded @emph{after} your @file{.emacs} file.
16943 Settings and definitions in your @file{.emacs} file will overwrite
16944 conflicting settings and definitions in a @file{site-start.el} file,
16945 if it exists; but the settings and definitions in a @file{default.el}
16946 or terminal type file will overwrite those in your @file{.emacs} file.
16947 (You can prevent interference from a terminal type file by setting
16948 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
16949 Simple Extension}.)
16951 @c Rewritten to avoid overfull hbox.
16952 The @file{INSTALL} file that comes in the distribution contains
16953 descriptions of the @file{site-init.el} and @file{site-load.el} files.
16955 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
16956 control loading. These files are in the @file{lisp} directory of the
16957 Emacs distribution and are worth perusing.
16959 The @file{loaddefs.el} file contains a good many suggestions as to
16960 what to put into your own @file{.emacs} file, or into a site-wide
16961 initialization file.
16964 @section Specifying Variables using @code{defcustom}
16967 You can specify variables using @code{defcustom} so that you and
16968 others can then use Emacs's @code{customize} feature to set their
16969 values. (You cannot use @code{customize} to write function
16970 definitions; but you can write @code{defuns} in your @file{.emacs}
16971 file. Indeed, you can write any Lisp expression in your @file{.emacs}
16974 The @code{customize} feature depends on the @code{defcustom} special
16975 form. Although you can use @code{defvar} or @code{setq} for variables
16976 that users set, the @code{defcustom} special form is designed for the
16979 You can use your knowledge of @code{defvar} for writing the
16980 first three arguments for @code{defcustom}. The first argument to
16981 @code{defcustom} is the name of the variable. The second argument is
16982 the variable's initial value, if any; and this value is set only if
16983 the value has not already been set. The third argument is the
16986 The fourth and subsequent arguments to @code{defcustom} specify types
16987 and options; these are not featured in @code{defvar}. (These
16988 arguments are optional.)
16990 Each of these arguments consists of a keyword followed by a value.
16991 Each keyword starts with the colon character @samp{:}.
16994 For example, the customizable user option variable
16995 @code{text-mode-hook} looks like this:
16999 (defcustom text-mode-hook nil
17000 "Normal hook run when entering Text mode and many related modes."
17002 :options '(turn-on-auto-fill flyspell-mode)
17008 The name of the variable is @code{text-mode-hook}; it has no default
17009 value; and its documentation string tells you what it does.
17011 The @code{:type} keyword tells Emacs the kind of data to which
17012 @code{text-mode-hook} should be set and how to display the value in a
17013 Customization buffer.
17015 The @code{:options} keyword specifies a suggested list of values for
17016 the variable. Usually, @code{:options} applies to a hook.
17017 The list is only a suggestion; it is not exclusive; a person who sets
17018 the variable may set it to other values; the list shown following the
17019 @code{:options} keyword is intended to offer convenient choices to a
17022 Finally, the @code{:group} keyword tells the Emacs Customization
17023 command in which group the variable is located. This tells where to
17026 The @code{defcustom} function recognizes more than a dozen keywords.
17027 For more information, see @ref{Customization, , Writing Customization
17028 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17030 Consider @code{text-mode-hook} as an example.
17032 There are two ways to customize this variable. You can use the
17033 customization command or write the appropriate expressions yourself.
17036 Using the customization command, you can type:
17043 and find that the group for editing files of data is called `data'.
17044 Enter that group. Text Mode Hook is the first member. You can click
17045 on its various options, such as @code{turn-on-auto-fill}, to set the
17046 values. After you click on the button to
17049 Save for Future Sessions
17053 Emacs will write an expression into your @file{.emacs} file.
17054 It will look like this:
17058 (custom-set-variables
17059 ;; custom-set-variables was added by Custom.
17060 ;; If you edit it by hand, you could mess it up, so be careful.
17061 ;; Your init file should contain only one such instance.
17062 ;; If there is more than one, they won't work right.
17063 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17068 (The @code{text-mode-hook-identify} function tells
17069 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17070 It comes on automatically.)
17072 The @code{custom-set-variables} function works somewhat differently
17073 than a @code{setq}. While I have never learned the differences, I
17074 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17075 file by hand: I make the changes in what appears to me to be a
17076 reasonable manner and have not had any problems. Others prefer to use
17077 the Customization command and let Emacs do the work for them.
17079 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17080 This function sets the various font faces. Over time, I have set a
17081 considerable number of faces. Some of the time, I re-set them using
17082 @code{customize}; other times, I simply edit the
17083 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17085 The second way to customize your @code{text-mode-hook} is to set it
17086 yourself in your @file{.emacs} file using code that has nothing to do
17087 with the @code{custom-set-@dots{}} functions.
17090 When you do this, and later use @code{customize}, you will see a
17094 CHANGED outside Customize; operating on it here may be unreliable.
17098 This message is only a warning. If you click on the button to
17101 Save for Future Sessions
17105 Emacs will write a @code{custom-set-@dots{}} expression near the end
17106 of your @file{.emacs} file that will be evaluated after your
17107 hand-written expression. It will, therefore, overrule your
17108 hand-written expression. No harm will be done. When you do this,
17109 however, be careful to remember which expression is active; if you
17110 forget, you may confuse yourself.
17112 So long as you remember where the values are set, you will have no
17113 trouble. In any event, the values are always set in your
17114 initialization file, which is usually called @file{.emacs}.
17116 I myself use @code{customize} for hardly anything. Mostly, I write
17117 expressions myself.
17121 Incidentally, to be more complete concerning defines: @code{defsubst}
17122 defines an inline function. The syntax is just like that of
17123 @code{defun}. @code{defconst} defines a symbol as a constant. The
17124 intent is that neither programs nor users should ever change a value
17125 set by @code{defconst}. (You can change it; the value set is a
17126 variable; but please do not.)
17128 @node Beginning a .emacs File
17129 @section Beginning a @file{.emacs} File
17130 @cindex @file{.emacs} file, beginning of
17132 When you start Emacs, it loads your @file{.emacs} file unless you tell
17133 it not to by specifying @samp{-q} on the command line. (The
17134 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17136 A @file{.emacs} file contains Lisp expressions. Often, these are no
17137 more than expressions to set values; sometimes they are function
17140 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17141 Manual}, for a short description of initialization files.
17143 This chapter goes over some of the same ground, but is a walk among
17144 extracts from a complete, long-used @file{.emacs} file---my own.
17146 The first part of the file consists of comments: reminders to myself.
17147 By now, of course, I remember these things, but when I started, I did
17153 ;;;; Bob's .emacs file
17154 ; Robert J. Chassell
17155 ; 26 September 1985
17160 Look at that date! I started this file a long time ago. I have been
17161 adding to it ever since.
17165 ; Each section in this file is introduced by a
17166 ; line beginning with four semicolons; and each
17167 ; entry is introduced by a line beginning with
17168 ; three semicolons.
17173 This describes the usual conventions for comments in Emacs Lisp.
17174 Everything on a line that follows a semicolon is a comment. Two,
17175 three, and four semicolons are used as subsection and section markers.
17176 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17177 more about comments.)
17182 ; Control-h is the help key;
17183 ; after typing control-h, type a letter to
17184 ; indicate the subject about which you want help.
17185 ; For an explanation of the help facility,
17186 ; type control-h two times in a row.
17191 Just remember: type @kbd{C-h} two times for help.
17195 ; To find out about any mode, type control-h m
17196 ; while in that mode. For example, to find out
17197 ; about mail mode, enter mail mode and then type
17203 `Mode help', as I call this, is very helpful. Usually, it tells you
17204 all you need to know.
17206 Of course, you don't need to include comments like these in your
17207 @file{.emacs} file. I included them in mine because I kept forgetting
17208 about Mode help or the conventions for comments---but I was able to
17209 remember to look here to remind myself.
17211 @node Text and Auto-fill
17212 @section Text and Auto Fill Mode
17214 Now we come to the part that `turns on' Text mode and
17219 ;;; Text mode and Auto Fill mode
17220 ;; The next two lines put Emacs into Text mode
17221 ;; and Auto Fill mode, and are for writers who
17222 ;; want to start writing prose rather than code.
17223 (setq-default major-mode 'text-mode)
17224 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17228 Here is the first part of this @file{.emacs} file that does something
17229 besides remind a forgetful human!
17231 The first of the two lines in parentheses tells Emacs to turn on Text
17232 mode when you find a file, @emph{unless} that file should go into some
17233 other mode, such as C mode.
17235 @cindex Per-buffer, local variables list
17236 @cindex Local variables list, per-buffer,
17237 @cindex Automatic mode selection
17238 @cindex Mode selection, automatic
17239 When Emacs reads a file, it looks at the extension to the file name,
17240 if any. (The extension is the part that comes after a @samp{.}.) If
17241 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17242 on C mode. Also, Emacs looks at first nonblank line of the file; if
17243 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17244 possesses a list of extensions and specifications that it uses
17245 automatically. In addition, Emacs looks near the last page for a
17246 per-buffer, ``local variables list'', if any.
17249 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17252 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17256 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17257 Files'' in @cite{The GNU Emacs Manual}.
17260 Now, back to the @file{.emacs} file.
17263 Here is the line again; how does it work?
17265 @cindex Text Mode turned on
17267 (setq major-mode 'text-mode)
17271 This line is a short, but complete Emacs Lisp expression.
17273 We are already familiar with @code{setq}. It sets the following variable,
17274 @code{major-mode}, to the subsequent value, which is @code{text-mode}.
17275 The single quote mark before @code{text-mode} tells Emacs to deal directly
17276 with the @code{text-mode} symbol, not with whatever it might stand for.
17277 @xref{set & setq, , Setting the Value of a Variable},
17278 for a reminder of how @code{setq} works.
17279 The main point is that there is no difference between the procedure you
17280 use to set a value in your @file{.emacs} file and the procedure you use
17281 anywhere else in Emacs.
17284 Here is the next line:
17286 @cindex Auto Fill mode turned on
17289 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17293 In this line, the @code{add-hook} command adds
17294 @code{turn-on-auto-fill} to the variable.
17296 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17297 it!, turns on Auto Fill mode.
17299 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17300 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17301 turns on Auto Fill mode.
17303 In brief, the first line causes Emacs to enter Text mode when you edit a
17304 file, unless the file name extension, a first non-blank line, or local
17305 variables to tell Emacs otherwise.
17307 Text mode among other actions, sets the syntax table to work
17308 conveniently for writers. In Text mode, Emacs considers an apostrophe
17309 as part of a word like a letter; but Emacs does not consider a period
17310 or a space as part of a word. Thus, @kbd{M-f} moves you over
17311 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17312 the @samp{t} of @samp{it's}.
17314 The second line causes Emacs to turn on Auto Fill mode when it turns
17315 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17316 that is too wide and brings the excessively wide part of the line down
17317 to the next line. Emacs breaks lines between words, not within them.
17319 When Auto Fill mode is turned off, lines continue to the right as you
17320 type them. Depending on how you set the value of
17321 @code{truncate-lines}, the words you type either disappear off the
17322 right side of the screen, or else are shown, in a rather ugly and
17323 unreadable manner, as a continuation line on the screen.
17326 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17327 fill commands to insert two spaces after a colon:
17330 (setq colon-double-space t)
17334 @section Mail Aliases
17336 Here is a @code{setq} that `turns on' mail aliases, along with more
17342 ; To enter mail mode, type `C-x m'
17343 ; To enter RMAIL (for reading mail),
17345 (setq mail-aliases t)
17349 @cindex Mail aliases
17351 This @code{setq} command sets the value of the variable
17352 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17353 says, in effect, ``Yes, use mail aliases.''
17355 Mail aliases are convenient short names for long email addresses or
17356 for lists of email addresses. The file where you keep your `aliases'
17357 is @file{~/.mailrc}. You write an alias like this:
17360 alias geo george@@foobar.wiz.edu
17364 When you write a message to George, address it to @samp{geo}; the
17365 mailer will automatically expand @samp{geo} to the full address.
17367 @node Indent Tabs Mode
17368 @section Indent Tabs Mode
17369 @cindex Tabs, preventing
17370 @findex indent-tabs-mode
17372 By default, Emacs inserts tabs in place of multiple spaces when it
17373 formats a region. (For example, you might indent many lines of text
17374 all at once with the @code{indent-region} command.) Tabs look fine on
17375 a terminal or with ordinary printing, but they produce badly indented
17376 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17379 The following turns off Indent Tabs mode:
17383 ;;; Prevent Extraneous Tabs
17384 (setq-default indent-tabs-mode nil)
17388 Note that this line uses @code{setq-default} rather than the
17389 @code{setq} command that we have seen before. The @code{setq-default}
17390 command sets values only in buffers that do not have their own local
17391 values for the variable.
17394 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17396 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17400 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17401 Files'' in @cite{The GNU Emacs Manual}.
17406 @section Some Keybindings
17408 Now for some personal keybindings:
17412 ;;; Compare windows
17413 (global-set-key "\C-cw" 'compare-windows)
17417 @findex compare-windows
17418 @code{compare-windows} is a nifty command that compares the text in
17419 your current window with text in the next window. It makes the
17420 comparison by starting at point in each window, moving over text in
17421 each window as far as they match. I use this command all the time.
17423 This also shows how to set a key globally, for all modes.
17425 @cindex Setting a key globally
17426 @cindex Global set key
17427 @cindex Key setting globally
17428 @findex global-set-key
17429 The command is @code{global-set-key}. It is followed by the
17430 keybinding. In a @file{.emacs} file, the keybinding is written as
17431 shown: @code{\C-c} stands for `control-c', which means `press the
17432 control key and the @key{c} key at the same time'. The @code{w} means
17433 `press the @key{w} key'. The keybinding is surrounded by double
17434 quotation marks. In documentation, you would write this as
17435 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17436 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17437 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17438 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17441 The command invoked by the keys is @code{compare-windows}. Note that
17442 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17443 would first try to evaluate the symbol to determine its value.
17445 These three things, the double quotation marks, the backslash before
17446 the @samp{C}, and the single quote mark are necessary parts of
17447 keybinding that I tend to forget. Fortunately, I have come to
17448 remember that I should look at my existing @file{.emacs} file, and
17449 adapt what is there.
17451 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17452 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17453 set of keys, @kbd{C-c} followed by a single character, is strictly
17454 reserved for individuals' own use. (I call these `own' keys, since
17455 these are for my own use.) You should always be able to create such a
17456 keybinding for your own use without stomping on someone else's
17457 keybinding. If you ever write an extension to Emacs, please avoid
17458 taking any of these keys for public use. Create a key like @kbd{C-c
17459 C-w} instead. Otherwise, we will run out of `own' keys.
17462 Here is another keybinding, with a comment:
17466 ;;; Keybinding for `occur'
17467 ; I use occur a lot, so let's bind it to a key:
17468 (global-set-key "\C-co" 'occur)
17473 The @code{occur} command shows all the lines in the current buffer
17474 that contain a match for a regular expression. Matching lines are
17475 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17476 to jump to occurrences.
17478 @findex global-unset-key
17479 @cindex Unbinding key
17480 @cindex Key unbinding
17482 Here is how to unbind a key, so it does not
17488 (global-unset-key "\C-xf")
17492 There is a reason for this unbinding: I found I inadvertently typed
17493 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17494 file, as I intended, I accidentally set the width for filled text,
17495 almost always to a width I did not want. Since I hardly ever reset my
17496 default width, I simply unbound the key.
17498 @findex list-buffers, @r{rebound}
17499 @findex buffer-menu, @r{bound to key}
17501 The following rebinds an existing key:
17505 ;;; Rebind `C-x C-b' for `buffer-menu'
17506 (global-set-key "\C-x\C-b" 'buffer-menu)
17510 By default, @kbd{C-x C-b} runs the
17511 @code{list-buffers} command. This command lists
17512 your buffers in @emph{another} window. Since I
17513 almost always want to do something in that
17514 window, I prefer the @code{buffer-menu}
17515 command, which not only lists the buffers,
17516 but moves point into that window.
17521 @cindex Rebinding keys
17523 Emacs uses @dfn{keymaps} to record which keys call which commands.
17524 When you use @code{global-set-key} to set the keybinding for a single
17525 command in all parts of Emacs, you are specifying the keybinding in
17526 @code{current-global-map}.
17528 Specific modes, such as C mode or Text mode, have their own keymaps;
17529 the mode-specific keymaps override the global map that is shared by
17532 The @code{global-set-key} function binds, or rebinds, the global
17533 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17534 function @code{buffer-menu}:
17537 (global-set-key "\C-x\C-b" 'buffer-menu)
17540 Mode-specific keymaps are bound using the @code{define-key} function,
17541 which takes a specific keymap as an argument, as well as the key and
17542 the command. For example, my @file{.emacs} file contains the
17543 following expression to bind the @code{texinfo-insert-@@group} command
17544 to @kbd{C-c C-c g}:
17548 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17553 The @code{texinfo-insert-@@group} function itself is a little extension
17554 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17555 use this command all the time and prefer to type the three strokes
17556 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17557 (@samp{@@group} and its matching @samp{@@end group} are commands that
17558 keep all enclosed text together on one page; many multi-line examples
17559 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17562 Here is the @code{texinfo-insert-@@group} function definition:
17566 (defun texinfo-insert-@@group ()
17567 "Insert the string @@group in a Texinfo buffer."
17569 (beginning-of-line)
17570 (insert "@@group\n"))
17574 (Of course, I could have used Abbrev mode to save typing, rather than
17575 write a function to insert a word; but I prefer key strokes consistent
17576 with other Texinfo mode key bindings.)
17578 You will see numerous @code{define-key} expressions in
17579 @file{loaddefs.el} as well as in the various mode libraries, such as
17580 @file{cc-mode.el} and @file{lisp-mode.el}.
17582 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17583 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17584 Reference Manual}, for more information about keymaps.
17586 @node Loading Files
17587 @section Loading Files
17588 @cindex Loading files
17591 Many people in the GNU Emacs community have written extensions to
17592 Emacs. As time goes by, these extensions are often included in new
17593 releases. For example, the Calendar and Diary packages are now part
17594 of the standard GNU Emacs, as is Calc.
17596 You can use a @code{load} command to evaluate a complete file and
17597 thereby install all the functions and variables in the file into Emacs.
17600 @c (auto-compression-mode t)
17603 (load "~/emacs/slowsplit")
17606 This evaluates, i.e., loads, the @file{slowsplit.el} file or if it
17607 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17608 @file{emacs} sub-directory of your home directory. The file contains
17609 the function @code{split-window-quietly}, which John Robinson wrote in
17612 The @code{split-window-quietly} function splits a window with the
17613 minimum of redisplay. I installed it in 1989 because it worked well
17614 with the slow 1200 baud terminals I was then using. Nowadays, I only
17615 occasionally come across such a slow connection, but I continue to use
17616 the function because I like the way it leaves the bottom half of a
17617 buffer in the lower of the new windows and the top half in the upper
17621 To replace the key binding for the default
17622 @code{split-window-vertically}, you must also unset that key and bind
17623 the keys to @code{split-window-quietly}, like this:
17627 (global-unset-key "\C-x2")
17628 (global-set-key "\C-x2" 'split-window-quietly)
17633 If you load many extensions, as I do, then instead of specifying the
17634 exact location of the extension file, as shown above, you can specify
17635 that directory as part of Emacs's @code{load-path}. Then, when Emacs
17636 loads a file, it will search that directory as well as its default
17637 list of directories. (The default list is specified in @file{paths.h}
17638 when Emacs is built.)
17641 The following command adds your @file{~/emacs} directory to the
17642 existing load path:
17646 ;;; Emacs Load Path
17647 (setq load-path (cons "~/emacs" load-path))
17651 Incidentally, @code{load-library} is an interactive interface to the
17652 @code{load} function. The complete function looks like this:
17654 @findex load-library
17657 (defun load-library (library)
17658 "Load the library named LIBRARY.
17659 This is an interface to the function `load'."
17661 (list (completing-read "Load library: "
17662 (apply-partially 'locate-file-completion-table
17664 (get-load-suffixes)))))
17669 The name of the function, @code{load-library}, comes from the use of
17670 `library' as a conventional synonym for `file'. The source for the
17671 @code{load-library} command is in the @file{files.el} library.
17673 Another interactive command that does a slightly different job is
17674 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17675 Emacs, emacs, The GNU Emacs Manual}, for information on the
17676 distinction between @code{load-library} and this command.
17679 @section Autoloading
17682 Instead of installing a function by loading the file that contains it,
17683 or by evaluating the function definition, you can make the function
17684 available but not actually install it until it is first called. This
17685 is called @dfn{autoloading}.
17687 When you execute an autoloaded function, Emacs automatically evaluates
17688 the file that contains the definition, and then calls the function.
17690 Emacs starts quicker with autoloaded functions, since their libraries
17691 are not loaded right away; but you need to wait a moment when you
17692 first use such a function, while its containing file is evaluated.
17694 Rarely used functions are frequently autoloaded. The
17695 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17696 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17697 come to use a `rare' function frequently. When you do, you should
17698 load that function's file with a @code{load} expression in your
17699 @file{.emacs} file.
17701 In my @file{.emacs} file, I load 14 libraries that contain functions
17702 that would otherwise be autoloaded. (Actually, it would have been
17703 better to include these files in my `dumped' Emacs, but I forgot.
17704 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17705 Reference Manual}, and the @file{INSTALL} file for more about
17708 You may also want to include autoloaded expressions in your @file{.emacs}
17709 file. @code{autoload} is a built-in function that takes up to five
17710 arguments, the final three of which are optional. The first argument
17711 is the name of the function to be autoloaded; the second is the name
17712 of the file to be loaded. The third argument is documentation for the
17713 function, and the fourth tells whether the function can be called
17714 interactively. The fifth argument tells what type of
17715 object---@code{autoload} can handle a keymap or macro as well as a
17716 function (the default is a function).
17719 Here is a typical example:
17723 (autoload 'html-helper-mode
17724 "html-helper-mode" "Edit HTML documents" t)
17729 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17730 which is a standard part of the distribution.)
17733 This expression autoloads the @code{html-helper-mode} function. It
17734 takes it from the @file{html-helper-mode.el} file (or from the byte
17735 compiled version @file{html-helper-mode.elc}, if that exists.) The
17736 file must be located in a directory specified by @code{load-path}.
17737 The documentation says that this is a mode to help you edit documents
17738 written in the HyperText Markup Language. You can call this mode
17739 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17740 duplicate the function's regular documentation in the autoload
17741 expression because the regular function is not yet loaded, so its
17742 documentation is not available.)
17744 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17745 Manual}, for more information.
17747 @node Simple Extension
17748 @section A Simple Extension: @code{line-to-top-of-window}
17749 @findex line-to-top-of-window
17750 @cindex Simple extension in @file{.emacs} file
17752 Here is a simple extension to Emacs that moves the line point is on to
17753 the top of the window. I use this all the time, to make text easier
17756 You can put the following code into a separate file and then load it
17757 from your @file{.emacs} file, or you can include it within your
17758 @file{.emacs} file.
17761 Here is the definition:
17765 ;;; Line to top of window;
17766 ;;; replace three keystroke sequence C-u 0 C-l
17767 (defun line-to-top-of-window ()
17768 "Move the line point is on to top of window."
17775 Now for the keybinding.
17777 Nowadays, function keys as well as mouse button events and
17778 non-@sc{ascii} characters are written within square brackets, without
17779 quotation marks. (In Emacs version 18 and before, you had to write
17780 different function key bindings for each different make of terminal.)
17782 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17786 (global-set-key [f6] 'line-to-top-of-window)
17789 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17790 Your Init File, emacs, The GNU Emacs Manual}.
17792 @cindex Conditional 'twixt two versions of Emacs
17793 @cindex Version of Emacs, choosing
17794 @cindex Emacs version, choosing
17795 If you run two versions of GNU Emacs, such as versions 22 and 23, and
17796 use one @file{.emacs} file, you can select which code to evaluate with
17797 the following conditional:
17802 ((= 22 emacs-major-version)
17803 ;; evaluate version 22 code
17805 ((= 23 emacs-major-version)
17806 ;; evaluate version 23 code
17811 For example, recent versions blink
17812 their cursors by default. I hate such blinking, as well as other
17813 features, so I placed the following in my @file{.emacs}
17814 file@footnote{When I start instances of Emacs that do not load my
17815 @file{.emacs} file or any site file, I also turn off blinking:
17818 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17820 @exdent Or nowadays, using an even more sophisticated set of options,
17828 (when (>= emacs-major-version 21)
17829 (blink-cursor-mode 0)
17830 ;; Insert newline when you press `C-n' (next-line)
17831 ;; at the end of the buffer
17832 (setq next-line-add-newlines t)
17835 ;; Turn on image viewing
17836 (auto-image-file-mode t)
17839 ;; Turn on menu bar (this bar has text)
17840 ;; (Use numeric argument to turn on)
17844 ;; Turn off tool bar (this bar has icons)
17845 ;; (Use numeric argument to turn on)
17846 (tool-bar-mode nil)
17849 ;; Turn off tooltip mode for tool bar
17850 ;; (This mode causes icon explanations to pop up)
17851 ;; (Use numeric argument to turn on)
17853 ;; If tooltips turned on, make tips appear promptly
17854 (setq tooltip-delay 0.1) ; default is 0.7 second
17860 @section X11 Colors
17862 You can specify colors when you use Emacs with the MIT X Windowing
17865 I dislike the default colors and specify my own.
17868 Here are the expressions in my @file{.emacs}
17869 file that set values:
17873 ;; Set cursor color
17874 (set-cursor-color "white")
17877 (set-mouse-color "white")
17879 ;; Set foreground and background
17880 (set-foreground-color "white")
17881 (set-background-color "darkblue")
17885 ;;; Set highlighting colors for isearch and drag
17886 (set-face-foreground 'highlight "white")
17887 (set-face-background 'highlight "blue")
17891 (set-face-foreground 'region "cyan")
17892 (set-face-background 'region "blue")
17896 (set-face-foreground 'secondary-selection "skyblue")
17897 (set-face-background 'secondary-selection "darkblue")
17901 ;; Set calendar highlighting colors
17902 (setq calendar-load-hook
17904 (set-face-foreground 'diary-face "skyblue")
17905 (set-face-background 'holiday-face "slate blue")
17906 (set-face-foreground 'holiday-face "white")))
17910 The various shades of blue soothe my eye and prevent me from seeing
17911 the screen flicker.
17913 Alternatively, I could have set my specifications in various X
17914 initialization files. For example, I could set the foreground,
17915 background, cursor, and pointer (i.e., mouse) colors in my
17916 @file{~/.Xresources} file like this:
17920 Emacs*foreground: white
17921 Emacs*background: darkblue
17922 Emacs*cursorColor: white
17923 Emacs*pointerColor: white
17927 In any event, since it is not part of Emacs, I set the root color of
17928 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
17929 run more modern window managers, such as Enlightenment, Gnome, or KDE;
17930 in those cases, I often specify an image rather than a plain color.}:
17933 xsetroot -solid Navy -fg white &
17937 @node Miscellaneous
17938 @section Miscellaneous Settings for a @file{.emacs} File
17941 Here are a few miscellaneous settings:
17946 Set the shape and color of the mouse cursor:
17950 ; Cursor shapes are defined in
17951 ; `/usr/include/X11/cursorfont.h';
17952 ; for example, the `target' cursor is number 128;
17953 ; the `top_left_arrow' cursor is number 132.
17957 (let ((mpointer (x-get-resource "*mpointer"
17958 "*emacs*mpointer")))
17959 ;; If you have not set your mouse pointer
17960 ;; then set it, otherwise leave as is:
17961 (if (eq mpointer nil)
17962 (setq mpointer "132")) ; top_left_arrow
17965 (setq x-pointer-shape (string-to-int mpointer))
17966 (set-mouse-color "white"))
17971 Or you can set the values of a variety of features in an alist, like
17977 default-frame-alist
17978 '((cursor-color . "white")
17979 (mouse-color . "white")
17980 (foreground-color . "white")
17981 (background-color . "DodgerBlue4")
17982 ;; (cursor-type . bar)
17983 (cursor-type . box)
17986 (tool-bar-lines . 0)
17987 (menu-bar-lines . 1)
17991 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
17997 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
17998 into @kbd{@key{CTRL}-h}.@*
17999 (Some older keyboards needed this, although I have not seen the
18004 ;; Translate `C-h' to <DEL>.
18005 ; (keyboard-translate ?\C-h ?\C-?)
18007 ;; Translate <DEL> to `C-h'.
18008 (keyboard-translate ?\C-? ?\C-h)
18012 @item Turn off a blinking cursor!
18016 (if (fboundp 'blink-cursor-mode)
18017 (blink-cursor-mode -1))
18022 or start GNU Emacs with the command @code{emacs -nbc}.
18025 @item When using `grep'@*
18026 @samp{-i}@w{ } Ignore case distinctions@*
18027 @samp{-n}@w{ } Prefix each line of output with line number@*
18028 @samp{-H}@w{ } Print the filename for each match.@*
18029 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18032 (setq grep-command "grep -i -nH -e ")
18036 @c Evidently, no longer needed in GNU Emacs 22
18038 item Automatically uncompress compressed files when visiting them
18041 (load "uncompress")
18046 @item Find an existing buffer, even if it has a different name@*
18047 This avoids problems with symbolic links.
18050 (setq find-file-existing-other-name t)
18053 @item Set your language environment and default input method
18057 (set-language-environment "latin-1")
18058 ;; Remember you can enable or disable multilingual text input
18059 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18060 (setq default-input-method "latin-1-prefix")
18064 If you want to write with Chinese `GB' characters, set this instead:
18068 (set-language-environment "Chinese-GB")
18069 (setq default-input-method "chinese-tonepy")
18074 @subsubheading Fixing Unpleasant Key Bindings
18075 @cindex Key bindings, fixing
18076 @cindex Bindings, key, fixing unpleasant
18078 Some systems bind keys unpleasantly. Sometimes, for example, the
18079 @key{CTRL} key appears in an awkward spot rather than at the far left
18082 Usually, when people fix these sorts of keybindings, they do not
18083 change their @file{~/.emacs} file. Instead, they bind the proper keys
18084 on their consoles with the @code{loadkeys} or @code{install-keymap}
18085 commands in their boot script and then include @code{xmodmap} commands
18086 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18094 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18096 install-keymap emacs2
18102 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18103 Lock} key is at the far left of the home row:
18107 # Bind the key labeled `Caps Lock' to `Control'
18108 # (Such a broken user interface suggests that keyboard manufacturers
18109 # think that computers are typewriters from 1885.)
18111 xmodmap -e "clear Lock"
18112 xmodmap -e "add Control = Caps_Lock"
18118 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18119 key to a @key{META} key:
18123 # Some ill designed keyboards have a key labeled ALT and no Meta
18124 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18130 @section A Modified Mode Line
18131 @vindex mode-line-format
18132 @cindex Mode line format
18134 Finally, a feature I really like: a modified mode line.
18136 When I work over a network, I forget which machine I am using. Also,
18137 I tend to I lose track of where I am, and which line point is on.
18139 So I reset my mode line to look like this:
18142 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18145 I am visiting a file called @file{foo.texi}, on my machine
18146 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18147 Texinfo mode, and am at the top of the buffer.
18150 My @file{.emacs} file has a section that looks like this:
18154 ;; Set a Mode Line that tells me which machine, which directory,
18155 ;; and which line I am on, plus the other customary information.
18156 (setq-default mode-line-format
18160 "mouse-1: select window, mouse-2: delete others ..."))
18161 mode-line-mule-info
18163 mode-line-frame-identification
18167 mode-line-buffer-identification
18170 (system-name) 0 (string-match "\\..+" (system-name))))
18175 "mouse-1: select window, mouse-2: delete others ..."))
18176 (line-number-mode " Line %l ")
18182 "mouse-1: select window, mouse-2: delete others ..."))
18183 (:eval (mode-line-mode-name))
18186 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18195 Here, I redefine the default mode line. Most of the parts are from
18196 the original; but I make a few changes. I set the @emph{default} mode
18197 line format so as to permit various modes, such as Info, to override
18200 Many elements in the list are self-explanatory:
18201 @code{mode-line-modified} is a variable that tells whether the buffer
18202 has been modified, @code{mode-name} tells the name of the mode, and so
18203 on. However, the format looks complicated because of two features we
18204 have not discussed.
18206 @cindex Properties, in mode line example
18207 The first string in the mode line is a dash or hyphen, @samp{-}. In
18208 the old days, it would have been specified simply as @code{"-"}. But
18209 nowadays, Emacs can add properties to a string, such as highlighting
18210 or, as in this case, a help feature. If you place your mouse cursor
18211 over the hyphen, some help information appears (By default, you must
18212 wait seven-tenths of a second before the information appears. You can
18213 change that timing by changing the value of @code{tooltip-delay}.)
18216 The new string format has a special syntax:
18219 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18223 The @code{#(} begins a list. The first element of the list is the
18224 string itself, just one @samp{-}. The second and third
18225 elements specify the range over which the fourth element applies. A
18226 range starts @emph{after} a character, so a zero means the range
18227 starts just before the first character; a 1 means that the range ends
18228 just after the first character. The third element is the property for
18229 the range. It consists of a property list, a
18230 property name, in this case, @samp{help-echo}, followed by a value, in this
18231 case, a string. The second, third, and fourth elements of this new
18232 string format can be repeated.
18234 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18235 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18236 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18238 @code{mode-line-buffer-identification}
18239 displays the current buffer name. It is a list
18240 beginning @code{(#("%12b" 0 4 @dots{}}.
18241 The @code{#(} begins the list.
18243 The @samp{"%12b"} displays the current buffer name, using the
18244 @code{buffer-name} function with which we are familiar; the `12'
18245 specifies the maximum number of characters that will be displayed.
18246 When a name has fewer characters, whitespace is added to fill out to
18247 this number. (Buffer names can and often should be longer than 12
18248 characters; this length works well in a typical 80 column wide
18251 @code{:eval} says to evaluate the following form and use the result as
18252 a string to display. In this case, the expression displays the first
18253 component of the full system name. The end of the first component is
18254 a @samp{.} (`period'), so I use the @code{string-match} function to
18255 tell me the length of the first component. The substring from the
18256 zeroth character to that length is the name of the machine.
18259 This is the expression:
18264 (system-name) 0 (string-match "\\..+" (system-name))))
18268 @samp{%[} and @samp{%]} cause a pair of square brackets
18269 to appear for each recursive editing level. @samp{%n} says `Narrow'
18270 when narrowing is in effect. @samp{%P} tells you the percentage of
18271 the buffer that is above the bottom of the window, or `Top', `Bottom',
18272 or `All'. (A lower case @samp{p} tell you the percentage above the
18273 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18276 Remember, ``You don't have to like Emacs to like it''---your own
18277 Emacs can have different colors, different commands, and different
18278 keys than a default Emacs.
18280 On the other hand, if you want to bring up a plain `out of the box'
18281 Emacs, with no customization, type:
18288 This will start an Emacs that does @emph{not} load your
18289 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18296 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18297 first is built into the internals of Emacs and is always with you;
18298 the second requires that you instrument a function before you can use it.
18300 Both debuggers are described extensively in @ref{Debugging, ,
18301 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18302 In this chapter, I will walk through a short example of each.
18305 * debug:: How to use the built-in debugger.
18306 * debug-on-entry:: Start debugging when you call a function.
18307 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18308 * edebug:: How to use Edebug, a source level debugger.
18309 * Debugging Exercises::
18313 @section @code{debug}
18316 Suppose you have written a function definition that is intended to
18317 return the sum of the numbers 1 through a given number. (This is the
18318 @code{triangle} function discussed earlier. @xref{Decrementing
18319 Example, , Example with Decrementing Counter}, for a discussion.)
18320 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18322 However, your function definition has a bug. You have mistyped
18323 @samp{1=} for @samp{1-}. Here is the broken definition:
18325 @findex triangle-bugged
18328 (defun triangle-bugged (number)
18329 "Return sum of numbers 1 through NUMBER inclusive."
18331 (while (> number 0)
18332 (setq total (+ total number))
18333 (setq number (1= number))) ; @r{Error here.}
18338 If you are reading this in Info, you can evaluate this definition in
18339 the normal fashion. You will see @code{triangle-bugged} appear in the
18343 Now evaluate the @code{triangle-bugged} function with an
18347 (triangle-bugged 4)
18351 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18357 ---------- Buffer: *Backtrace* ----------
18358 Debugger entered--Lisp error: (void-function 1=)
18360 (setq number (1= number))
18361 (while (> number 0) (setq total (+ total number))
18362 (setq number (1= number)))
18363 (let ((total 0)) (while (> number 0) (setq total ...)
18364 (setq number ...)) total)
18368 eval((triangle-bugged 4))
18369 eval-last-sexp-1(nil)
18370 eval-last-sexp(nil)
18371 call-interactively(eval-last-sexp)
18372 ---------- Buffer: *Backtrace* ----------
18377 (I have reformatted this example slightly; the debugger does not fold
18378 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18379 the @file{*Backtrace*} buffer.)
18381 In practice, for a bug as simple as this, the `Lisp error' line will
18382 tell you what you need to know to correct the definition. The
18383 function @code{1=} is `void'.
18387 In GNU Emacs 20 and before, you will see:
18390 Symbol's function definition is void:@: 1=
18394 which has the same meaning as the @file{*Backtrace*} buffer line in
18398 However, suppose you are not quite certain what is going on?
18399 You can read the complete backtrace.
18401 In this case, you need to run a recent GNU Emacs, which automatically
18402 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18403 else, you need to start the debugger manually as described below.
18405 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18406 what Emacs did that led to the error. Emacs made an interactive call
18407 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18408 of the @code{triangle-bugged} expression. Each line above tells you
18409 what the Lisp interpreter evaluated next.
18412 The third line from the top of the buffer is
18415 (setq number (1= number))
18419 Emacs tried to evaluate this expression; in order to do so, it tried
18420 to evaluate the inner expression shown on the second line from the
18429 This is where the error occurred; as the top line says:
18432 Debugger entered--Lisp error: (void-function 1=)
18436 You can correct the mistake, re-evaluate the function definition, and
18437 then run your test again.
18439 @node debug-on-entry
18440 @section @code{debug-on-entry}
18441 @findex debug-on-entry
18443 A recent GNU Emacs starts the debugger automatically when your
18444 function has an error.
18447 GNU Emacs version 20 and before did not; it simply
18448 presented you with an error message. You had to start the debugger
18452 Incidentally, you can start the debugger manually for all versions of
18453 Emacs; the advantage is that the debugger runs even if you do not have
18454 a bug in your code. Sometimes your code will be free of bugs!
18456 You can enter the debugger when you call the function by calling
18457 @code{debug-on-entry}.
18464 M-x debug-on-entry RET triangle-bugged RET
18469 Now, evaluate the following:
18472 (triangle-bugged 5)
18476 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18477 you that it is beginning to evaluate the @code{triangle-bugged}
18482 ---------- Buffer: *Backtrace* ----------
18483 Debugger entered--entering a function:
18484 * triangle-bugged(5)
18485 eval((triangle-bugged 5))
18488 eval-last-sexp-1(nil)
18489 eval-last-sexp(nil)
18490 call-interactively(eval-last-sexp)
18491 ---------- Buffer: *Backtrace* ----------
18495 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18496 the first expression in @code{triangle-bugged}; the buffer will look
18501 ---------- Buffer: *Backtrace* ----------
18502 Debugger entered--beginning evaluation of function call form:
18503 * (let ((total 0)) (while (> number 0) (setq total ...)
18504 (setq number ...)) total)
18505 * triangle-bugged(5)
18506 eval((triangle-bugged 5))
18509 eval-last-sexp-1(nil)
18510 eval-last-sexp(nil)
18511 call-interactively(eval-last-sexp)
18512 ---------- Buffer: *Backtrace* ----------
18517 Now, type @kbd{d} again, eight times, slowly. Each time you type
18518 @kbd{d}, Emacs will evaluate another expression in the function
18522 Eventually, the buffer will look like this:
18526 ---------- Buffer: *Backtrace* ----------
18527 Debugger entered--beginning evaluation of function call form:
18528 * (setq number (1= number))
18529 * (while (> number 0) (setq total (+ total number))
18530 (setq number (1= number)))
18533 * (let ((total 0)) (while (> number 0) (setq total ...)
18534 (setq number ...)) total)
18535 * triangle-bugged(5)
18536 eval((triangle-bugged 5))
18539 eval-last-sexp-1(nil)
18540 eval-last-sexp(nil)
18541 call-interactively(eval-last-sexp)
18542 ---------- Buffer: *Backtrace* ----------
18548 Finally, after you type @kbd{d} two more times, Emacs will reach the
18549 error, and the top two lines of the @file{*Backtrace*} buffer will look
18554 ---------- Buffer: *Backtrace* ----------
18555 Debugger entered--Lisp error: (void-function 1=)
18558 ---------- Buffer: *Backtrace* ----------
18562 By typing @kbd{d}, you were able to step through the function.
18564 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18565 quits the trace, but does not cancel @code{debug-on-entry}.
18567 @findex cancel-debug-on-entry
18568 To cancel the effect of @code{debug-on-entry}, call
18569 @code{cancel-debug-on-entry} and the name of the function, like this:
18572 M-x cancel-debug-on-entry RET triangle-bugged RET
18576 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18578 @node debug-on-quit
18579 @section @code{debug-on-quit} and @code{(debug)}
18581 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18582 there are two other ways to start @code{debug}.
18584 @findex debug-on-quit
18585 You can start @code{debug} whenever you type @kbd{C-g}
18586 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18587 @code{t}. This is useful for debugging infinite loops.
18590 @cindex @code{(debug)} in code
18591 Or, you can insert a line that says @code{(debug)} into your code
18592 where you want the debugger to start, like this:
18596 (defun triangle-bugged (number)
18597 "Return sum of numbers 1 through NUMBER inclusive."
18599 (while (> number 0)
18600 (setq total (+ total number))
18601 (debug) ; @r{Start debugger.}
18602 (setq number (1= number))) ; @r{Error here.}
18607 The @code{debug} function is described in detail in @ref{Debugger, ,
18608 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18611 @section The @code{edebug} Source Level Debugger
18612 @cindex Source level debugger
18615 Edebug is a source level debugger. Edebug normally displays the
18616 source of the code you are debugging, with an arrow at the left that
18617 shows which line you are currently executing.
18619 You can walk through the execution of a function, line by line, or run
18620 quickly until reaching a @dfn{breakpoint} where execution stops.
18622 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18623 Lisp Reference Manual}.
18626 Here is a bugged function definition for @code{triangle-recursively}.
18627 @xref{Recursive triangle function, , Recursion in place of a counter},
18628 for a review of it.
18632 (defun triangle-recursively-bugged (number)
18633 "Return sum of numbers 1 through NUMBER inclusive.
18638 (triangle-recursively-bugged
18639 (1= number))))) ; @r{Error here.}
18644 Normally, you would install this definition by positioning your cursor
18645 after the function's closing parenthesis and typing @kbd{C-x C-e}
18646 (@code{eval-last-sexp}) or else by positioning your cursor within the
18647 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18648 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18652 However, to prepare this function definition for Edebug, you must
18653 first @dfn{instrument} the code using a different command. You can do
18654 this by positioning your cursor within or just after the definition
18658 M-x edebug-defun RET
18662 This will cause Emacs to load Edebug automatically if it is not
18663 already loaded, and properly instrument the function.
18665 After instrumenting the function, place your cursor after the
18666 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18669 (triangle-recursively-bugged 3)
18673 You will be jumped back to the source for
18674 @code{triangle-recursively-bugged} and the cursor positioned at the
18675 beginning of the @code{if} line of the function. Also, you will see
18676 an arrowhead at the left hand side of that line. The arrowhead marks
18677 the line where the function is executing. (In the following examples,
18678 we show the arrowhead with @samp{=>}; in a windowing system, you may
18679 see the arrowhead as a solid triangle in the window `fringe'.)
18682 =>@point{}(if (= number 1)
18687 In the example, the location of point is displayed with a star,
18688 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18691 In the example, the location of point is displayed as @samp{@point{}}
18692 (in a printed book, it is displayed with a five pointed star).
18695 If you now press @key{SPC}, point will move to the next expression to
18696 be executed; the line will look like this:
18699 =>(if @point{}(= number 1)
18703 As you continue to press @key{SPC}, point will move from expression to
18704 expression. At the same time, whenever an expression returns a value,
18705 that value will be displayed in the echo area. For example, after you
18706 move point past @code{number}, you will see the following:
18709 Result: 3 (#o3, #x3, ?\C-c)
18713 This means the value of @code{number} is 3, which is octal three,
18714 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18715 alphabet, in case you need to know this information).
18717 You can continue moving through the code until you reach the line with
18718 the error. Before evaluation, that line looks like this:
18721 => @point{}(1= number))))) ; @r{Error here.}
18726 When you press @key{SPC} once again, you will produce an error message
18730 Symbol's function definition is void:@: 1=
18736 Press @kbd{q} to quit Edebug.
18738 To remove instrumentation from a function definition, simply
18739 re-evaluate it with a command that does not instrument it.
18740 For example, you could place your cursor after the definition's
18741 closing parenthesis and type @kbd{C-x C-e}.
18743 Edebug does a great deal more than walk with you through a function.
18744 You can set it so it races through on its own, stopping only at an
18745 error or at specified stopping points; you can cause it to display the
18746 changing values of various expressions; you can find out how many
18747 times a function is called, and more.
18749 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18750 Lisp Reference Manual}.
18753 @node Debugging Exercises
18754 @section Debugging Exercises
18758 Install the @code{@value{COUNT-WORDS}} function and then cause it to
18759 enter the built-in debugger when you call it. Run the command on a
18760 region containing two words. You will need to press @kbd{d} a
18761 remarkable number of times. On your system, is a `hook' called after
18762 the command finishes? (For information on hooks, see @ref{Command
18763 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18767 Copy @code{@value{COUNT-WORDS}} into the @file{*scratch*} buffer,
18768 instrument the function for Edebug, and walk through its execution.
18769 The function does not need to have a bug, although you can introduce
18770 one if you wish. If the function lacks a bug, the walk-through
18771 completes without problems.
18774 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18775 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.,
18776 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18777 for commands made outside of the Edebug debugging buffer.)
18780 In the Edebug debugging buffer, use the @kbd{p}
18781 (@code{edebug-bounce-point}) command to see where in the region the
18782 @code{@value{COUNT-WORDS}} is working.
18785 Move point to some spot further down the function and then type the
18786 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18789 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18790 walk through the function on its own; use an upper case @kbd{T} for
18791 @code{edebug-Trace-fast-mode}.
18794 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18799 @chapter Conclusion
18801 We have now reached the end of this Introduction. You have now
18802 learned enough about programming in Emacs Lisp to set values, to write
18803 simple @file{.emacs} files for yourself and your friends, and write
18804 simple customizations and extensions to Emacs.
18806 This is a place to stop. Or, if you wish, you can now go onward, and
18809 You have learned some of the basic nuts and bolts of programming. But
18810 only some. There are a great many more brackets and hinges that are
18811 easy to use that we have not touched.
18813 A path you can follow right now lies among the sources to GNU Emacs
18816 @cite{The GNU Emacs Lisp Reference Manual}.
18819 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18820 Emacs Lisp Reference Manual}.
18823 The Emacs Lisp sources are an adventure. When you read the sources and
18824 come across a function or expression that is unfamiliar, you need to
18825 figure out or find out what it does.
18827 Go to the Reference Manual. It is a thorough, complete, and fairly
18828 easy-to-read description of Emacs Lisp. It is written not only for
18829 experts, but for people who know what you know. (The @cite{Reference
18830 Manual} comes with the standard GNU Emacs distribution. Like this
18831 introduction, it comes as a Texinfo source file, so you can read it
18832 on-line and as a typeset, printed book.)
18834 Go to the other on-line help that is part of GNU Emacs: the on-line
18835 documentation for all functions and variables, and @code{find-tag},
18836 the program that takes you to sources.
18838 Here is an example of how I explore the sources. Because of its name,
18839 @file{simple.el} is the file I looked at first, a long time ago. As
18840 it happens some of the functions in @file{simple.el} are complicated,
18841 or at least look complicated at first sight. The @code{open-line}
18842 function, for example, looks complicated.
18844 You may want to walk through this function slowly, as we did with the
18845 @code{forward-sentence} function. (@xref{forward-sentence, The
18846 @code{forward-sentence} function}.) Or you may want to skip that
18847 function and look at another, such as @code{split-line}. You don't
18848 need to read all the functions. According to
18849 @code{count-words-in-defun}, the @code{split-line} function contains
18850 102 words and symbols.
18852 Even though it is short, @code{split-line} contains expressions
18853 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18854 @code{current-column} and @code{insert-and-inherit}.
18856 Consider the @code{skip-chars-forward} function. (It is part of the
18857 function definition for @code{back-to-indentation}, which is shown in
18858 @ref{Review, , Review}.)
18860 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18861 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18862 function. This gives you the function documentation.
18864 You may be able to guess what is done by a well named function such as
18865 @code{indent-to}; or you can look it up, too. Incidentally, the
18866 @code{describe-function} function itself is in @file{help.el}; it is
18867 one of those long, but decipherable functions. You can look up
18868 @code{describe-function} using the @kbd{C-h f} command!
18870 In this instance, since the code is Lisp, the @file{*Help*} buffer
18871 contains the name of the library containing the function's source.
18872 You can put point over the name of the library and press the RET key,
18873 which in this situation is bound to @code{help-follow}, and be taken
18874 directly to the source, in the same way as @kbd{M-.}
18877 The definition for @code{describe-function} illustrates how to
18878 customize the @code{interactive} expression without using the standard
18879 character codes; and it shows how to create a temporary buffer.
18881 (The @code{indent-to} function is written in C rather than Emacs Lisp;
18882 it is a `built-in' function. @code{help-follow} takes you to its
18883 source as does @code{find-tag}, when properly set up.)
18885 You can look at a function's source using @code{find-tag}, which is
18886 bound to @kbd{M-.} Finally, you can find out what the Reference
18887 Manual has to say by visiting the manual in Info, and typing @kbd{i}
18888 (@code{Info-index}) and the name of the function, or by looking up the
18889 function in the index to a printed copy of the manual.
18891 Similarly, you can find out what is meant by
18892 @code{insert-and-inherit}.
18894 Other interesting source files include @file{paragraphs.el},
18895 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
18896 file includes short, easily understood functions as well as longer
18897 ones. The @file{loaddefs.el} file contains the many standard
18898 autoloads and many keymaps. I have never looked at it all; only at
18899 parts. @file{loadup.el} is the file that loads the standard parts of
18900 Emacs; it tells you a great deal about how Emacs is built.
18901 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
18902 Reference Manual}, for more about building.)
18904 As I said, you have learned some nuts and bolts; however, and very
18905 importantly, we have hardly touched major aspects of programming; I
18906 have said nothing about how to sort information, except to use the
18907 predefined @code{sort} function; I have said nothing about how to store
18908 information, except to use variables and lists; I have said nothing
18909 about how to write programs that write programs. These are topics for
18910 another, and different kind of book, a different kind of learning.
18912 What you have done is learn enough for much practical work with GNU
18913 Emacs. What you have done is get started. This is the end of a
18916 @c ================ Appendix ================
18919 @appendix The @code{the-the} Function
18921 @cindex Duplicated words function
18922 @cindex Words, duplicated
18924 Sometimes when you you write text, you duplicate words---as with ``you
18925 you'' near the beginning of this sentence. I find that most
18926 frequently, I duplicate ``the''; hence, I call the function for
18927 detecting duplicated words, @code{the-the}.
18930 As a first step, you could use the following regular expression to
18931 search for duplicates:
18934 \\(\\w+[ \t\n]+\\)\\1
18938 This regexp matches one or more word-constituent characters followed
18939 by one or more spaces, tabs, or newlines. However, it does not detect
18940 duplicated words on different lines, since the ending of the first
18941 word, the end of the line, is different from the ending of the second
18942 word, a space. (For more information about regular expressions, see
18943 @ref{Regexp Search, , Regular Expression Searches}, as well as
18944 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
18945 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
18946 The GNU Emacs Lisp Reference Manual}.)
18948 You might try searching just for duplicated word-constituent
18949 characters but that does not work since the pattern detects doubles
18950 such as the two occurrences of `th' in `with the'.
18952 Another possible regexp searches for word-constituent characters
18953 followed by non-word-constituent characters, reduplicated. Here,
18954 @w{@samp{\\w+}} matches one or more word-constituent characters and
18955 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
18958 \\(\\(\\w+\\)\\W*\\)\\1
18964 Here is the pattern that I use. It is not perfect, but good enough.
18965 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
18966 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
18967 any characters that are @emph{not} an @@-sign, space, newline, or tab.
18970 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
18973 One can write more complicated expressions, but I found that this
18974 expression is good enough, so I use it.
18976 Here is the @code{the-the} function, as I include it in my
18977 @file{.emacs} file, along with a handy global key binding:
18982 "Search forward for for a duplicated word."
18984 (message "Searching for for duplicated words ...")
18988 ;; This regexp is not perfect
18989 ;; but is fairly good over all:
18990 (if (re-search-forward
18991 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
18992 (message "Found duplicated word.")
18993 (message "End of buffer")))
18997 ;; Bind `the-the' to C-c \
18998 (global-set-key "\C-c\\" 'the-the)
19007 one two two three four five
19012 You can substitute the other regular expressions shown above in the
19013 function definition and try each of them on this list.
19016 @appendix Handling the Kill Ring
19017 @cindex Kill ring handling
19018 @cindex Handling the kill ring
19019 @cindex Ring, making a list like a
19021 The kill ring is a list that is transformed into a ring by the
19022 workings of the @code{current-kill} function. The @code{yank} and
19023 @code{yank-pop} commands use the @code{current-kill} function.
19025 This appendix describes the @code{current-kill} function as well as
19026 both the @code{yank} and the @code{yank-pop} commands, but first,
19027 consider the workings of the kill ring.
19030 * What the Kill Ring Does::
19032 * yank:: Paste a copy of a clipped element.
19033 * yank-pop:: Insert element pointed to.
19038 @node What the Kill Ring Does
19039 @unnumberedsec What the Kill Ring Does
19043 The kill ring has a default maximum length of sixty items; this number
19044 is too large for an explanation. Instead, set it to four. Please
19045 evaluate the following:
19049 (setq old-kill-ring-max kill-ring-max)
19050 (setq kill-ring-max 4)
19055 Then, please copy each line of the following indented example into the
19056 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19060 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19061 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19062 merely copy it to the kill ring. However, your machine may beep at
19063 you. Alternatively, for silence, you may copy the region of each line
19064 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19065 each line for this command to succeed, but it does not matter at which
19066 end you put point or mark.)
19070 Please invoke the calls in order, so that five elements attempt to
19071 fill the kill ring:
19076 second piece of text
19078 fourth line of text
19085 Then find the value of @code{kill-ring} by evaluating
19097 ("fifth bit of text" "fourth line of text"
19098 "third line" "second piece of text")
19103 The first element, @samp{first some text}, was dropped.
19106 To return to the old value for the length of the kill ring, evaluate:
19109 (setq kill-ring-max old-kill-ring-max)
19113 @appendixsec The @code{current-kill} Function
19114 @findex current-kill
19116 The @code{current-kill} function changes the element in the kill ring
19117 to which @code{kill-ring-yank-pointer} points. (Also, the
19118 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19119 to the latest element of the kill ring. The @code{kill-new}
19120 function is used directly or indirectly by @code{kill-append},
19121 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19122 and @code{kill-region}.)
19125 * Code for current-kill::
19126 * Understanding current-kill::
19130 @node Code for current-kill
19131 @unnumberedsubsec The code for @code{current-kill}
19136 The @code{current-kill} function is used by @code{yank} and by
19137 @code{yank-pop}. Here is the code for @code{current-kill}:
19141 (defun current-kill (n &optional do-not-move)
19142 "Rotate the yanking point by N places, and then return that kill.
19143 If N is zero, `interprogram-paste-function' is set, and calling it
19144 returns a string, then that string is added to the front of the
19145 kill ring and returned as the latest kill.
19148 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19149 yanking point; just return the Nth kill forward."
19150 (let ((interprogram-paste (and (= n 0)
19151 interprogram-paste-function
19152 (funcall interprogram-paste-function))))
19155 (if interprogram-paste
19157 ;; Disable the interprogram cut function when we add the new
19158 ;; text to the kill ring, so Emacs doesn't try to own the
19159 ;; selection, with identical text.
19160 (let ((interprogram-cut-function nil))
19161 (kill-new interprogram-paste))
19162 interprogram-paste)
19165 (or kill-ring (error "Kill ring is empty"))
19166 (let ((ARGth-kill-element
19167 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19168 (length kill-ring))
19171 (setq kill-ring-yank-pointer ARGth-kill-element))
19172 (car ARGth-kill-element)))))
19176 Remember also that the @code{kill-new} function sets
19177 @code{kill-ring-yank-pointer} to the latest element of the kill
19178 ring, which means that all the functions that call it set the value
19179 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19180 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19183 Here is the line in @code{kill-new}, which is explained in
19184 @ref{kill-new function, , The @code{kill-new} function}.
19187 (setq kill-ring-yank-pointer kill-ring)
19191 @node Understanding current-kill
19192 @unnumberedsubsec @code{current-kill} in Outline
19195 The @code{current-kill} function looks complex, but as usual, it can
19196 be understood by taking it apart piece by piece. First look at it in
19201 (defun current-kill (n &optional do-not-move)
19202 "Rotate the yanking point by N places, and then return that kill."
19208 This function takes two arguments, one of which is optional. It has a
19209 documentation string. It is @emph{not} interactive.
19212 * Body of current-kill::
19213 * Digression concerning error:: How to mislead humans, but not computers.
19214 * Determining the Element::
19218 @node Body of current-kill
19219 @unnumberedsubsubsec The Body of @code{current-kill}
19222 The body of the function definition is a @code{let} expression, which
19223 itself has a body as well as a @var{varlist}.
19225 The @code{let} expression declares a variable that will be only usable
19226 within the bounds of this function. This variable is called
19227 @code{interprogram-paste} and is for copying to another program. It
19228 is not for copying within this instance of GNU Emacs. Most window
19229 systems provide a facility for interprogram pasting. Sadly, that
19230 facility usually provides only for the last element. Most windowing
19231 systems have not adopted a ring of many possibilities, even though
19232 Emacs has provided it for decades.
19234 The @code{if} expression has two parts, one if there exists
19235 @code{interprogram-paste} and one if not.
19238 Let us consider the `if not' or else-part of the @code{current-kill}
19239 function. (The then-part uses the @code{kill-new} function, which
19240 we have already described. @xref{kill-new function, , The
19241 @code{kill-new} function}.)
19245 (or kill-ring (error "Kill ring is empty"))
19246 (let ((ARGth-kill-element
19247 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19248 (length kill-ring))
19251 (setq kill-ring-yank-pointer ARGth-kill-element))
19252 (car ARGth-kill-element))
19257 The code first checks whether the kill ring has content; otherwise it
19261 Note that the @code{or} expression is very similar to testing length
19268 (if (zerop (length kill-ring)) ; @r{if-part}
19269 (error "Kill ring is empty")) ; @r{then-part}
19275 If there is not anything in the kill ring, its length must be zero and
19276 an error message sent to the user: @samp{Kill ring is empty}. The
19277 @code{current-kill} function uses an @code{or} expression which is
19278 simpler. But an @code{if} expression reminds us what goes on.
19280 This @code{if} expression uses the function @code{zerop} which returns
19281 true if the value it is testing is zero. When @code{zerop} tests
19282 true, the then-part of the @code{if} is evaluated. The then-part is a
19283 list starting with the function @code{error}, which is a function that
19284 is similar to the @code{message} function
19285 (@pxref{message, , The @code{message} Function}) in that
19286 it prints a one-line message in the echo area. However, in addition
19287 to printing a message, @code{error} also stops evaluation of the
19288 function within which it is embedded. This means that the rest of the
19289 function will not be evaluated if the length of the kill ring is zero.
19291 Then the @code{current-kill} function selects the element to return.
19292 The selection depends on the number of places that @code{current-kill}
19293 rotates and on where @code{kill-ring-yank-pointer} points.
19295 Next, either the optional @code{do-not-move} argument is true or the
19296 current value of @code{kill-ring-yank-pointer} is set to point to the
19297 list. Finally, another expression returns the first element of the
19298 list even if the @code{do-not-move} argument is true.
19301 @node Digression concerning error
19302 @unnumberedsubsubsec Digression about the word `error'
19305 In my opinion, it is slightly misleading, at least to humans, to use
19306 the term `error' as the name of the @code{error} function. A better
19307 term would be `cancel'. Strictly speaking, of course, you cannot
19308 point to, much less rotate a pointer to a list that has no length, so
19309 from the point of view of the computer, the word `error' is correct.
19310 But a human expects to attempt this sort of thing, if only to find out
19311 whether the kill ring is full or empty. This is an act of
19314 From the human point of view, the act of exploration and discovery is
19315 not necessarily an error, and therefore should not be labeled as one,
19316 even in the bowels of a computer. As it is, the code in Emacs implies
19317 that a human who is acting virtuously, by exploring his or her
19318 environment, is making an error. This is bad. Even though the computer
19319 takes the same steps as it does when there is an `error', a term such as
19320 `cancel' would have a clearer connotation.
19323 @node Determining the Element
19324 @unnumberedsubsubsec Determining the Element
19327 Among other actions, the else-part of the @code{if} expression sets
19328 the value of @code{kill-ring-yank-pointer} to
19329 @code{ARGth-kill-element} when the kill ring has something in it and
19330 the value of @code{do-not-move} is @code{nil}.
19333 The code looks like this:
19337 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19338 (length kill-ring))
19343 This needs some examination. Unless it is not supposed to move the
19344 pointer, the @code{current-kill} function changes where
19345 @code{kill-ring-yank-pointer} points.
19347 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19348 expression does. Also, clearly, @code{ARGth-kill-element} is being
19349 set to be equal to some @sc{cdr} of the kill ring, using the
19350 @code{nthcdr} function that is described in an earlier section.
19351 (@xref{copy-region-as-kill}.) How does it do this?
19353 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19354 works by repeatedly taking the @sc{cdr} of a list---it takes the
19355 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19358 The two following expressions produce the same result:
19362 (setq kill-ring-yank-pointer (cdr kill-ring))
19364 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19368 However, the @code{nthcdr} expression is more complicated. It uses
19369 the @code{mod} function to determine which @sc{cdr} to select.
19371 (You will remember to look at inner functions first; indeed, we will
19372 have to go inside the @code{mod}.)
19374 The @code{mod} function returns the value of its first argument modulo
19375 the second; that is to say, it returns the remainder after dividing
19376 the first argument by the second. The value returned has the same
19377 sign as the second argument.
19385 @result{} 0 ;; @r{because there is no remainder}
19392 In this case, the first argument is often smaller than the second.
19404 We can guess what the @code{-} function does. It is like @code{+} but
19405 subtracts instead of adds; the @code{-} function subtracts its second
19406 argument from its first. Also, we already know what the @code{length}
19407 function does (@pxref{length}). It returns the length of a list.
19409 And @code{n} is the name of the required argument to the
19410 @code{current-kill} function.
19413 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19414 expression returns the whole list, as you can see by evaluating the
19419 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19420 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19421 (nthcdr (mod (- 0 4) 4)
19422 '("fourth line of text"
19424 "second piece of text"
19425 "first some text"))
19430 When the first argument to the @code{current-kill} function is one,
19431 the @code{nthcdr} expression returns the list without its first
19436 (nthcdr (mod (- 1 4) 4)
19437 '("fourth line of text"
19439 "second piece of text"
19440 "first some text"))
19444 @cindex @samp{global variable} defined
19445 @cindex @samp{variable, global}, defined
19446 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19447 are @dfn{global variables}. That means that any expression in Emacs
19448 Lisp can access them. They are not like the local variables set by
19449 @code{let} or like the symbols in an argument list.
19450 Local variables can only be accessed
19451 within the @code{let} that defines them or the function that specifies
19452 them in an argument list (and within expressions called by them).
19455 @c texi2dvi fails when the name of the section is within ifnottex ...
19456 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19457 @ref{defun, , The @code{defun} Special Form}.)
19461 @appendixsec @code{yank}
19464 After learning about @code{current-kill}, the code for the
19465 @code{yank} function is almost easy.
19467 The @code{yank} function does not use the
19468 @code{kill-ring-yank-pointer} variable directly. It calls
19469 @code{insert-for-yank} which calls @code{current-kill} which sets the
19470 @code{kill-ring-yank-pointer} variable.
19473 The code looks like this:
19478 (defun yank (&optional arg)
19479 "Reinsert (\"paste\") the last stretch of killed text.
19480 More precisely, reinsert the stretch of killed text most recently
19481 killed OR yanked. Put point at end, and set mark at beginning.
19482 With just \\[universal-argument] as argument, same but put point at
19483 beginning (and mark at end). With argument N, reinsert the Nth most
19484 recently killed stretch of killed text.
19486 When this command inserts killed text into the buffer, it honors
19487 `yank-excluded-properties' and `yank-handler' as described in the
19488 doc string for `insert-for-yank-1', which see.
19490 See also the command \\[yank-pop]."
19494 (setq yank-window-start (window-start))
19495 ;; If we don't get all the way thru, make last-command indicate that
19496 ;; for the following command.
19497 (setq this-command t)
19498 (push-mark (point))
19501 (insert-for-yank (current-kill (cond
19506 ;; This is like exchange-point-and-mark,
19507 ;; but doesn't activate the mark.
19508 ;; It is cleaner to avoid activation, even though the command
19509 ;; loop would deactivate the mark because we inserted text.
19510 (goto-char (prog1 (mark t)
19511 (set-marker (mark-marker) (point) (current-buffer)))))
19514 ;; If we do get all the way thru, make this-command indicate that.
19515 (if (eq this-command t)
19516 (setq this-command 'yank))
19521 The key expression is @code{insert-for-yank}, which inserts the string
19522 returned by @code{current-kill}, but removes some text properties from
19525 However, before getting to that expression, the function sets the value
19526 of @code{yank-window-start} to the position returned by the
19527 @code{(window-start)} expression, the position at which the display
19528 currently starts. The @code{yank} function also sets
19529 @code{this-command} and pushes the mark.
19531 After it yanks the appropriate element, if the optional argument is a
19532 @sc{cons} rather than a number or nothing, it puts point at beginning
19533 of the yanked text and mark at its end.
19535 (The @code{prog1} function is like @code{progn} but returns the value
19536 of its first argument rather than the value of its last argument. Its
19537 first argument is forced to return the buffer's mark as an integer.
19538 You can see the documentation for these functions by placing point
19539 over them in this buffer and then typing @kbd{C-h f}
19540 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19543 The last part of the function tells what to do when it succeeds.
19546 @appendixsec @code{yank-pop}
19549 After understanding @code{yank} and @code{current-kill}, you know how
19550 to approach the @code{yank-pop} function. Leaving out the
19551 documentation to save space, it looks like this:
19556 (defun yank-pop (&optional arg)
19559 (if (not (eq last-command 'yank))
19560 (error "Previous command was not a yank"))
19563 (setq this-command 'yank)
19564 (unless arg (setq arg 1))
19565 (let ((inhibit-read-only t)
19566 (before (< (point) (mark t))))
19570 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19571 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19572 (setq yank-undo-function nil)
19575 (set-marker (mark-marker) (point) (current-buffer))
19576 (insert-for-yank (current-kill arg))
19577 ;; Set the window start back where it was in the yank command,
19579 (set-window-start (selected-window) yank-window-start t)
19583 ;; This is like exchange-point-and-mark,
19584 ;; but doesn't activate the mark.
19585 ;; It is cleaner to avoid activation, even though the command
19586 ;; loop would deactivate the mark because we inserted text.
19587 (goto-char (prog1 (mark t)
19588 (set-marker (mark-marker)
19590 (current-buffer))))))
19595 The function is interactive with a small @samp{p} so the prefix
19596 argument is processed and passed to the function. The command can
19597 only be used after a previous yank; otherwise an error message is
19598 sent. This check uses the variable @code{last-command} which is set
19599 by @code{yank} and is discussed elsewhere.
19600 (@xref{copy-region-as-kill}.)
19602 The @code{let} clause sets the variable @code{before} to true or false
19603 depending whether point is before or after mark and then the region
19604 between point and mark is deleted. This is the region that was just
19605 inserted by the previous yank and it is this text that will be
19608 @code{funcall} calls its first argument as a function, passing
19609 remaining arguments to it. The first argument is whatever the
19610 @code{or} expression returns. The two remaining arguments are the
19611 positions of point and mark set by the preceding @code{yank} command.
19613 There is more, but that is the hardest part.
19616 @appendixsec The @file{ring.el} File
19617 @cindex @file{ring.el} file
19619 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19620 provides many of the features we just discussed. But functions such
19621 as @code{kill-ring-yank-pointer} do not use this library, possibly
19622 because they were written earlier.
19625 @appendix A Graph with Labeled Axes
19627 Printed axes help you understand a graph. They convey scale. In an
19628 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19629 wrote the code to print the body of a graph. Here we write the code
19630 for printing and labeling vertical and horizontal axes, along with the
19634 * Labeled Example::
19635 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19636 * print-Y-axis:: Print a label for the vertical axis.
19637 * print-X-axis:: Print a horizontal label.
19638 * Print Whole Graph:: The function to print a complete graph.
19642 @node Labeled Example
19643 @unnumberedsec Labeled Example Graph
19646 Since insertions fill a buffer to the right and below point, the new
19647 graph printing function should first print the Y or vertical axis,
19648 then the body of the graph, and finally the X or horizontal axis.
19649 This sequence lays out for us the contents of the function:
19659 Print body of graph.
19666 Here is an example of how a finished graph should look:
19679 1 - ****************
19686 In this graph, both the vertical and the horizontal axes are labeled
19687 with numbers. However, in some graphs, the horizontal axis is time
19688 and would be better labeled with months, like this:
19702 Indeed, with a little thought, we can easily come up with a variety of
19703 vertical and horizontal labeling schemes. Our task could become
19704 complicated. But complications breed confusion. Rather than permit
19705 this, it is better choose a simple labeling scheme for our first
19706 effort, and to modify or replace it later.
19709 These considerations suggest the following outline for the
19710 @code{print-graph} function:
19714 (defun print-graph (numbers-list)
19715 "@var{documentation}@dots{}"
19716 (let ((height @dots{}
19720 (print-Y-axis height @dots{} )
19721 (graph-body-print numbers-list)
19722 (print-X-axis @dots{} )))
19726 We can work on each part of the @code{print-graph} function definition
19729 @node print-graph Varlist
19730 @appendixsec The @code{print-graph} Varlist
19731 @cindex @code{print-graph} varlist
19733 In writing the @code{print-graph} function, the first task is to write
19734 the varlist in the @code{let} expression. (We will leave aside for the
19735 moment any thoughts about making the function interactive or about the
19736 contents of its documentation string.)
19738 The varlist should set several values. Clearly, the top of the label
19739 for the vertical axis must be at least the height of the graph, which
19740 means that we must obtain this information here. Note that the
19741 @code{print-graph-body} function also requires this information. There
19742 is no reason to calculate the height of the graph in two different
19743 places, so we should change @code{print-graph-body} from the way we
19744 defined it earlier to take advantage of the calculation.
19746 Similarly, both the function for printing the X axis labels and the
19747 @code{print-graph-body} function need to learn the value of the width of
19748 each symbol. We can perform the calculation here and change the
19749 definition for @code{print-graph-body} from the way we defined it in the
19752 The length of the label for the horizontal axis must be at least as long
19753 as the graph. However, this information is used only in the function
19754 that prints the horizontal axis, so it does not need to be calculated here.
19756 These thoughts lead us directly to the following form for the varlist
19757 in the @code{let} for @code{print-graph}:
19761 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19762 (symbol-width (length graph-blank)))
19767 As we shall see, this expression is not quite right.
19771 @appendixsec The @code{print-Y-axis} Function
19772 @cindex Axis, print vertical
19773 @cindex Y axis printing
19774 @cindex Vertical axis printing
19775 @cindex Print vertical axis
19777 The job of the @code{print-Y-axis} function is to print a label for
19778 the vertical axis that looks like this:
19796 The function should be passed the height of the graph, and then should
19797 construct and insert the appropriate numbers and marks.
19800 * print-Y-axis in Detail::
19801 * Height of label:: What height for the Y axis?
19802 * Compute a Remainder:: How to compute the remainder of a division.
19803 * Y Axis Element:: Construct a line for the Y axis.
19804 * Y-axis-column:: Generate a list of Y axis labels.
19805 * print-Y-axis Penultimate:: A not quite final version.
19809 @node print-Y-axis in Detail
19810 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19813 It is easy enough to see in the figure what the Y axis label should
19814 look like; but to say in words, and then to write a function
19815 definition to do the job is another matter. It is not quite true to
19816 say that we want a number and a tic every five lines: there are only
19817 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19818 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19819 and 9). It is better to say that we want a number and a tic mark on
19820 the base line (number 1) and then that we want a number and a tic on
19821 the fifth line from the bottom and on every line that is a multiple of
19825 @node Height of label
19826 @unnumberedsubsec What height should the label be?
19829 The next issue is what height the label should be? Suppose the maximum
19830 height of tallest column of the graph is seven. Should the highest
19831 label on the Y axis be @samp{5 -}, and should the graph stick up above
19832 the label? Or should the highest label be @samp{7 -}, and mark the peak
19833 of the graph? Or should the highest label be @code{10 -}, which is a
19834 multiple of five, and be higher than the topmost value of the graph?
19836 The latter form is preferred. Most graphs are drawn within rectangles
19837 whose sides are an integral number of steps long---5, 10, 15, and so
19838 on for a step distance of five. But as soon as we decide to use a
19839 step height for the vertical axis, we discover that the simple
19840 expression in the varlist for computing the height is wrong. The
19841 expression is @code{(apply 'max numbers-list)}. This returns the
19842 precise height, not the maximum height plus whatever is necessary to
19843 round up to the nearest multiple of five. A more complex expression
19846 As usual in cases like this, a complex problem becomes simpler if it is
19847 divided into several smaller problems.
19849 First, consider the case when the highest value of the graph is an
19850 integral multiple of five---when it is 5, 10, 15, or some higher
19851 multiple of five. We can use this value as the Y axis height.
19853 A fairly simply way to determine whether a number is a multiple of
19854 five is to divide it by five and see if the division results in a
19855 remainder. If there is no remainder, the number is a multiple of
19856 five. Thus, seven divided by five has a remainder of two, and seven
19857 is not an integral multiple of five. Put in slightly different
19858 language, more reminiscent of the classroom, five goes into seven
19859 once, with a remainder of two. However, five goes into ten twice,
19860 with no remainder: ten is an integral multiple of five.
19862 @node Compute a Remainder
19863 @appendixsubsec Side Trip: Compute a Remainder
19865 @findex % @r{(remainder function)}
19866 @cindex Remainder function, @code{%}
19867 In Lisp, the function for computing a remainder is @code{%}. The
19868 function returns the remainder of its first argument divided by its
19869 second argument. As it happens, @code{%} is a function in Emacs Lisp
19870 that you cannot discover using @code{apropos}: you find nothing if you
19871 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
19872 learn of the existence of @code{%} is to read about it in a book such
19873 as this or in the Emacs Lisp sources.
19875 You can try the @code{%} function by evaluating the following two
19887 The first expression returns 2 and the second expression returns 0.
19889 To test whether the returned value is zero or some other number, we
19890 can use the @code{zerop} function. This function returns @code{t} if
19891 its argument, which must be a number, is zero.
19903 Thus, the following expression will return @code{t} if the height
19904 of the graph is evenly divisible by five:
19907 (zerop (% height 5))
19911 (The value of @code{height}, of course, can be found from @code{(apply
19912 'max numbers-list)}.)
19914 On the other hand, if the value of @code{height} is not a multiple of
19915 five, we want to reset the value to the next higher multiple of five.
19916 This is straightforward arithmetic using functions with which we are
19917 already familiar. First, we divide the value of @code{height} by five
19918 to determine how many times five goes into the number. Thus, five
19919 goes into twelve twice. If we add one to this quotient and multiply by
19920 five, we will obtain the value of the next multiple of five that is
19921 larger than the height. Five goes into twelve twice. Add one to two,
19922 and multiply by five; the result is fifteen, which is the next multiple
19923 of five that is higher than twelve. The Lisp expression for this is:
19926 (* (1+ (/ height 5)) 5)
19930 For example, if you evaluate the following, the result is 15:
19933 (* (1+ (/ 12 5)) 5)
19936 All through this discussion, we have been using `five' as the value
19937 for spacing labels on the Y axis; but we may want to use some other
19938 value. For generality, we should replace `five' with a variable to
19939 which we can assign a value. The best name I can think of for this
19940 variable is @code{Y-axis-label-spacing}.
19943 Using this term, and an @code{if} expression, we produce the
19948 (if (zerop (% height Y-axis-label-spacing))
19951 (* (1+ (/ height Y-axis-label-spacing))
19952 Y-axis-label-spacing))
19957 This expression returns the value of @code{height} itself if the height
19958 is an even multiple of the value of the @code{Y-axis-label-spacing} or
19959 else it computes and returns a value of @code{height} that is equal to
19960 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
19962 We can now include this expression in the @code{let} expression of the
19963 @code{print-graph} function (after first setting the value of
19964 @code{Y-axis-label-spacing}):
19965 @vindex Y-axis-label-spacing
19969 (defvar Y-axis-label-spacing 5
19970 "Number of lines from one Y axis label to next.")
19975 (let* ((height (apply 'max numbers-list))
19976 (height-of-top-line
19977 (if (zerop (% height Y-axis-label-spacing))
19982 (* (1+ (/ height Y-axis-label-spacing))
19983 Y-axis-label-spacing)))
19984 (symbol-width (length graph-blank))))
19990 (Note use of the @code{let*} function: the initial value of height is
19991 computed once by the @code{(apply 'max numbers-list)} expression and
19992 then the resulting value of @code{height} is used to compute its
19993 final value. @xref{fwd-para let, , The @code{let*} expression}, for
19994 more about @code{let*}.)
19996 @node Y Axis Element
19997 @appendixsubsec Construct a Y Axis Element
19999 When we print the vertical axis, we want to insert strings such as
20000 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20001 Moreover, we want the numbers and dashes to line up, so shorter
20002 numbers must be padded with leading spaces. If some of the strings
20003 use two digit numbers, the strings with single digit numbers must
20004 include a leading blank space before the number.
20006 @findex number-to-string
20007 To figure out the length of the number, the @code{length} function is
20008 used. But the @code{length} function works only with a string, not with
20009 a number. So the number has to be converted from being a number to
20010 being a string. This is done with the @code{number-to-string} function.
20015 (length (number-to-string 35))
20018 (length (number-to-string 100))
20024 (@code{number-to-string} is also called @code{int-to-string}; you will
20025 see this alternative name in various sources.)
20027 In addition, in each label, each number is followed by a string such
20028 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20029 This variable is defined with @code{defvar}:
20034 (defvar Y-axis-tic " - "
20035 "String that follows number in a Y axis label.")
20039 The length of the Y label is the sum of the length of the Y axis tic
20040 mark and the length of the number of the top of the graph.
20043 (length (concat (number-to-string height) Y-axis-tic)))
20046 This value will be calculated by the @code{print-graph} function in
20047 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20048 did not think to include this in the varlist when we first proposed it.)
20050 To make a complete vertical axis label, a tic mark is concatenated
20051 with a number; and the two together may be preceded by one or more
20052 spaces depending on how long the number is. The label consists of
20053 three parts: the (optional) leading spaces, the number, and the tic
20054 mark. The function is passed the value of the number for the specific
20055 row, and the value of the width of the top line, which is calculated
20056 (just once) by @code{print-graph}.
20060 (defun Y-axis-element (number full-Y-label-width)
20061 "Construct a NUMBERed label element.
20062 A numbered element looks like this ` 5 - ',
20063 and is padded as needed so all line up with
20064 the element for the largest number."
20067 (let* ((leading-spaces
20068 (- full-Y-label-width
20070 (concat (number-to-string number)
20075 (make-string leading-spaces ? )
20076 (number-to-string number)
20081 The @code{Y-axis-element} function concatenates together the leading
20082 spaces, if any; the number, as a string; and the tic mark.
20084 To figure out how many leading spaces the label will need, the
20085 function subtracts the actual length of the label---the length of the
20086 number plus the length of the tic mark---from the desired label width.
20088 @findex make-string
20089 Blank spaces are inserted using the @code{make-string} function. This
20090 function takes two arguments: the first tells it how long the string
20091 will be and the second is a symbol for the character to insert, in a
20092 special format. The format is a question mark followed by a blank
20093 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20094 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20095 syntax for characters. (Of course, you might want to replace the
20096 blank space by some other character @dots{} You know what to do.)
20098 The @code{number-to-string} function is used in the concatenation
20099 expression, to convert the number to a string that is concatenated
20100 with the leading spaces and the tic mark.
20102 @node Y-axis-column
20103 @appendixsubsec Create a Y Axis Column
20105 The preceding functions provide all the tools needed to construct a
20106 function that generates a list of numbered and blank strings to insert
20107 as the label for the vertical axis:
20109 @findex Y-axis-column
20112 (defun Y-axis-column (height width-of-label)
20113 "Construct list of Y axis labels and blank strings.
20114 For HEIGHT of line above base and WIDTH-OF-LABEL."
20118 (while (> height 1)
20119 (if (zerop (% height Y-axis-label-spacing))
20120 ;; @r{Insert label.}
20123 (Y-axis-element height width-of-label)
20127 ;; @r{Else, insert blanks.}
20130 (make-string width-of-label ? )
20132 (setq height (1- height)))
20133 ;; @r{Insert base line.}
20135 (cons (Y-axis-element 1 width-of-label) Y-axis))
20136 (nreverse Y-axis)))
20140 In this function, we start with the value of @code{height} and
20141 repetitively subtract one from its value. After each subtraction, we
20142 test to see whether the value is an integral multiple of the
20143 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20144 using the @code{Y-axis-element} function; if not, we construct a
20145 blank label using the @code{make-string} function. The base line
20146 consists of the number one followed by a tic mark.
20149 @node print-Y-axis Penultimate
20150 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20152 The list constructed by the @code{Y-axis-column} function is passed to
20153 the @code{print-Y-axis} function, which inserts the list as a column.
20155 @findex print-Y-axis
20158 (defun print-Y-axis (height full-Y-label-width)
20159 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20160 Height must be the maximum height of the graph.
20161 Full width is the width of the highest label element."
20162 ;; Value of height and full-Y-label-width
20163 ;; are passed by `print-graph'.
20166 (let ((start (point)))
20168 (Y-axis-column height full-Y-label-width))
20169 ;; @r{Place point ready for inserting graph.}
20171 ;; @r{Move point forward by value of} full-Y-label-width
20172 (forward-char full-Y-label-width)))
20176 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20177 insert the Y axis labels created by the @code{Y-axis-column} function.
20178 In addition, it places point at the correct position for printing the body of
20181 You can test @code{print-Y-axis}:
20189 Y-axis-label-spacing
20198 Copy the following expression:
20201 (print-Y-axis 12 5)
20205 Switch to the @file{*scratch*} buffer and place the cursor where you
20206 want the axis labels to start.
20209 Type @kbd{M-:} (@code{eval-expression}).
20212 Yank the @code{graph-body-print} expression into the minibuffer
20213 with @kbd{C-y} (@code{yank)}.
20216 Press @key{RET} to evaluate the expression.
20219 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20220 }}}. (The @code{print-graph} function will pass the value of
20221 @code{height-of-top-line}, which in this case will end up as 15,
20222 thereby getting rid of what might appear as a bug.)
20226 @appendixsec The @code{print-X-axis} Function
20227 @cindex Axis, print horizontal
20228 @cindex X axis printing
20229 @cindex Print horizontal axis
20230 @cindex Horizontal axis printing
20232 X axis labels are much like Y axis labels, except that the ticks are on a
20233 line above the numbers. Labels should look like this:
20242 The first tic is under the first column of the graph and is preceded by
20243 several blank spaces. These spaces provide room in rows above for the Y
20244 axis labels. The second, third, fourth, and subsequent ticks are all
20245 spaced equally, according to the value of @code{X-axis-label-spacing}.
20247 The second row of the X axis consists of numbers, preceded by several
20248 blank spaces and also separated according to the value of the variable
20249 @code{X-axis-label-spacing}.
20251 The value of the variable @code{X-axis-label-spacing} should itself be
20252 measured in units of @code{symbol-width}, since you may want to change
20253 the width of the symbols that you are using to print the body of the
20254 graph without changing the ways the graph is labeled.
20257 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
20258 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20262 @node Similarities differences
20263 @unnumberedsubsec Similarities and differences
20266 The @code{print-X-axis} function is constructed in more or less the
20267 same fashion as the @code{print-Y-axis} function except that it has
20268 two lines: the line of tic marks and the numbers. We will write a
20269 separate function to print each line and then combine them within the
20270 @code{print-X-axis} function.
20272 This is a three step process:
20276 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20279 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20282 Write a function to print both lines, the @code{print-X-axis} function,
20283 using @code{print-X-axis-tic-line} and
20284 @code{print-X-axis-numbered-line}.
20287 @node X Axis Tic Marks
20288 @appendixsubsec X Axis Tic Marks
20290 The first function should print the X axis tic marks. We must specify
20291 the tic marks themselves and their spacing:
20295 (defvar X-axis-label-spacing
20296 (if (boundp 'graph-blank)
20297 (* 5 (length graph-blank)) 5)
20298 "Number of units from one X axis label to next.")
20303 (Note that the value of @code{graph-blank} is set by another
20304 @code{defvar}. The @code{boundp} predicate checks whether it has
20305 already been set; @code{boundp} returns @code{nil} if it has not. If
20306 @code{graph-blank} were unbound and we did not use this conditional
20307 construction, in a recent GNU Emacs, we would enter the debugger and
20308 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20309 @w{(void-variable graph-blank)}}.)
20312 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20316 (defvar X-axis-tic-symbol "|"
20317 "String to insert to point to a column in X axis.")
20322 The goal is to make a line that looks like this:
20328 The first tic is indented so that it is under the first column, which is
20329 indented to provide space for the Y axis labels.
20331 A tic element consists of the blank spaces that stretch from one tic to
20332 the next plus a tic symbol. The number of blanks is determined by the
20333 width of the tic symbol and the @code{X-axis-label-spacing}.
20336 The code looks like this:
20340 ;;; X-axis-tic-element
20344 ;; @r{Make a string of blanks.}
20345 (- (* symbol-width X-axis-label-spacing)
20346 (length X-axis-tic-symbol))
20348 ;; @r{Concatenate blanks with tic symbol.}
20354 Next, we determine how many blanks are needed to indent the first tic
20355 mark to the first column of the graph. This uses the value of
20356 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20359 The code to make @code{X-axis-leading-spaces}
20364 ;; X-axis-leading-spaces
20366 (make-string full-Y-label-width ? )
20371 We also need to determine the length of the horizontal axis, which is
20372 the length of the numbers list, and the number of ticks in the horizontal
20379 (length numbers-list)
20385 (* symbol-width X-axis-label-spacing)
20389 ;; number-of-X-ticks
20390 (if (zerop (% (X-length tic-width)))
20391 (/ (X-length tic-width))
20392 (1+ (/ (X-length tic-width))))
20397 All this leads us directly to the function for printing the X axis tic line:
20399 @findex print-X-axis-tic-line
20402 (defun print-X-axis-tic-line
20403 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20404 "Print ticks for X axis."
20405 (insert X-axis-leading-spaces)
20406 (insert X-axis-tic-symbol) ; @r{Under first column.}
20409 ;; @r{Insert second tic in the right spot.}
20412 (- (* symbol-width X-axis-label-spacing)
20413 ;; @r{Insert white space up to second tic symbol.}
20414 (* 2 (length X-axis-tic-symbol)))
20416 X-axis-tic-symbol))
20419 ;; @r{Insert remaining ticks.}
20420 (while (> number-of-X-tics 1)
20421 (insert X-axis-tic-element)
20422 (setq number-of-X-tics (1- number-of-X-tics))))
20426 The line of numbers is equally straightforward:
20429 First, we create a numbered element with blank spaces before each number:
20431 @findex X-axis-element
20434 (defun X-axis-element (number)
20435 "Construct a numbered X axis element."
20436 (let ((leading-spaces
20437 (- (* symbol-width X-axis-label-spacing)
20438 (length (number-to-string number)))))
20439 (concat (make-string leading-spaces ? )
20440 (number-to-string number))))
20444 Next, we create the function to print the numbered line, starting with
20445 the number ``1'' under the first column:
20447 @findex print-X-axis-numbered-line
20450 (defun print-X-axis-numbered-line
20451 (number-of-X-tics X-axis-leading-spaces)
20452 "Print line of X-axis numbers"
20453 (let ((number X-axis-label-spacing))
20454 (insert X-axis-leading-spaces)
20460 ;; @r{Insert white space up to next number.}
20461 (- (* symbol-width X-axis-label-spacing) 2)
20463 (number-to-string number)))
20466 ;; @r{Insert remaining numbers.}
20467 (setq number (+ number X-axis-label-spacing))
20468 (while (> number-of-X-tics 1)
20469 (insert (X-axis-element number))
20470 (setq number (+ number X-axis-label-spacing))
20471 (setq number-of-X-tics (1- number-of-X-tics)))))
20475 Finally, we need to write the @code{print-X-axis} that uses
20476 @code{print-X-axis-tic-line} and
20477 @code{print-X-axis-numbered-line}.
20479 The function must determine the local values of the variables used by both
20480 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20481 then it must call them. Also, it must print the carriage return that
20482 separates the two lines.
20484 The function consists of a varlist that specifies five local variables,
20485 and calls to each of the two line printing functions:
20487 @findex print-X-axis
20490 (defun print-X-axis (numbers-list)
20491 "Print X axis labels to length of NUMBERS-LIST."
20492 (let* ((leading-spaces
20493 (make-string full-Y-label-width ? ))
20496 ;; symbol-width @r{is provided by} graph-body-print
20497 (tic-width (* symbol-width X-axis-label-spacing))
20498 (X-length (length numbers-list))
20506 ;; @r{Make a string of blanks.}
20507 (- (* symbol-width X-axis-label-spacing)
20508 (length X-axis-tic-symbol))
20512 ;; @r{Concatenate blanks with tic symbol.}
20513 X-axis-tic-symbol))
20517 (if (zerop (% X-length tic-width))
20518 (/ X-length tic-width)
20519 (1+ (/ X-length tic-width)))))
20522 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20524 (print-X-axis-numbered-line tic-number leading-spaces)))
20529 You can test @code{print-X-axis}:
20533 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20534 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20535 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20538 Copy the following expression:
20543 (let ((full-Y-label-width 5)
20546 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20551 Switch to the @file{*scratch*} buffer and place the cursor where you
20552 want the axis labels to start.
20555 Type @kbd{M-:} (@code{eval-expression}).
20558 Yank the test expression into the minibuffer
20559 with @kbd{C-y} (@code{yank)}.
20562 Press @key{RET} to evaluate the expression.
20566 Emacs will print the horizontal axis like this:
20576 @node Print Whole Graph
20577 @appendixsec Printing the Whole Graph
20578 @cindex Printing the whole graph
20579 @cindex Whole graph printing
20580 @cindex Graph, printing all
20582 Now we are nearly ready to print the whole graph.
20584 The function to print the graph with the proper labels follows the
20585 outline we created earlier (@pxref{Full Graph, , A Graph with Labeled
20586 Axes}), but with additions.
20589 Here is the outline:
20593 (defun print-graph (numbers-list)
20594 "@var{documentation}@dots{}"
20595 (let ((height @dots{}
20599 (print-Y-axis height @dots{} )
20600 (graph-body-print numbers-list)
20601 (print-X-axis @dots{} )))
20606 * The final version:: A few changes.
20607 * Test print-graph:: Run a short test.
20608 * Graphing words in defuns:: Executing the final code.
20609 * lambda:: How to write an anonymous function.
20610 * mapcar:: Apply a function to elements of a list.
20611 * Another Bug:: Yet another bug @dots{} most insidious.
20612 * Final printed graph:: The graph itself!
20616 @node The final version
20617 @unnumberedsubsec Changes for the Final Version
20620 The final version is different from what we planned in two ways:
20621 first, it contains additional values calculated once in the varlist;
20622 second, it carries an option to specify the labels' increment per row.
20623 This latter feature turns out to be essential; otherwise, a graph may
20624 have more rows than fit on a display or on a sheet of paper.
20627 This new feature requires a change to the @code{Y-axis-column}
20628 function, to add @code{vertical-step} to it. The function looks like
20631 @findex Y-axis-column @r{Final version.}
20634 ;;; @r{Final version.}
20635 (defun Y-axis-column
20636 (height width-of-label &optional vertical-step)
20637 "Construct list of labels for Y axis.
20638 HEIGHT is maximum height of graph.
20639 WIDTH-OF-LABEL is maximum width of label.
20640 VERTICAL-STEP, an option, is a positive integer
20641 that specifies how much a Y axis label increments
20642 for each line. For example, a step of 5 means
20643 that each line is five units of the graph."
20647 (number-per-line (or vertical-step 1)))
20648 (while (> height 1)
20649 (if (zerop (% height Y-axis-label-spacing))
20652 ;; @r{Insert label.}
20656 (* height number-per-line)
20661 ;; @r{Else, insert blanks.}
20664 (make-string width-of-label ? )
20666 (setq height (1- height)))
20669 ;; @r{Insert base line.}
20670 (setq Y-axis (cons (Y-axis-element
20671 (or vertical-step 1)
20674 (nreverse Y-axis)))
20678 The values for the maximum height of graph and the width of a symbol
20679 are computed by @code{print-graph} in its @code{let} expression; so
20680 @code{graph-body-print} must be changed to accept them.
20682 @findex graph-body-print @r{Final version.}
20685 ;;; @r{Final version.}
20686 (defun graph-body-print (numbers-list height symbol-width)
20687 "Print a bar graph of the NUMBERS-LIST.
20688 The numbers-list consists of the Y-axis values.
20689 HEIGHT is maximum height of graph.
20690 SYMBOL-WIDTH is number of each column."
20693 (let (from-position)
20694 (while numbers-list
20695 (setq from-position (point))
20697 (column-of-graph height (car numbers-list)))
20698 (goto-char from-position)
20699 (forward-char symbol-width)
20702 ;; @r{Draw graph column by column.}
20704 (setq numbers-list (cdr numbers-list)))
20705 ;; @r{Place point for X axis labels.}
20706 (forward-line height)
20712 Finally, the code for the @code{print-graph} function:
20714 @findex print-graph @r{Final version.}
20717 ;;; @r{Final version.}
20719 (numbers-list &optional vertical-step)
20720 "Print labeled bar graph of the NUMBERS-LIST.
20721 The numbers-list consists of the Y-axis values.
20725 Optionally, VERTICAL-STEP, a positive integer,
20726 specifies how much a Y axis label increments for
20727 each line. For example, a step of 5 means that
20728 each row is five units."
20731 (let* ((symbol-width (length graph-blank))
20732 ;; @code{height} @r{is both the largest number}
20733 ;; @r{and the number with the most digits.}
20734 (height (apply 'max numbers-list))
20737 (height-of-top-line
20738 (if (zerop (% height Y-axis-label-spacing))
20741 (* (1+ (/ height Y-axis-label-spacing))
20742 Y-axis-label-spacing)))
20745 (vertical-step (or vertical-step 1))
20746 (full-Y-label-width
20752 (* height-of-top-line vertical-step))
20758 height-of-top-line full-Y-label-width vertical-step)
20762 numbers-list height-of-top-line symbol-width)
20763 (print-X-axis numbers-list)))
20767 @node Test print-graph
20768 @appendixsubsec Testing @code{print-graph}
20771 We can test the @code{print-graph} function with a short list of numbers:
20775 Install the final versions of @code{Y-axis-column},
20776 @code{graph-body-print}, and @code{print-graph} (in addition to the
20780 Copy the following expression:
20783 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20787 Switch to the @file{*scratch*} buffer and place the cursor where you
20788 want the axis labels to start.
20791 Type @kbd{M-:} (@code{eval-expression}).
20794 Yank the test expression into the minibuffer
20795 with @kbd{C-y} (@code{yank)}.
20798 Press @key{RET} to evaluate the expression.
20802 Emacs will print a graph that looks like this:
20823 On the other hand, if you pass @code{print-graph} a
20824 @code{vertical-step} value of 2, by evaluating this expression:
20827 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20832 The graph looks like this:
20853 (A question: is the `2' on the bottom of the vertical axis a bug or a
20854 feature? If you think it is a bug, and should be a `1' instead, (or
20855 even a `0'), you can modify the sources.)
20857 @node Graphing words in defuns
20858 @appendixsubsec Graphing Numbers of Words and Symbols
20860 Now for the graph for which all this code was written: a graph that
20861 shows how many function definitions contain fewer than 10 words and
20862 symbols, how many contain between 10 and 19 words and symbols, how
20863 many contain between 20 and 29 words and symbols, and so on.
20865 This is a multi-step process. First make sure you have loaded all the
20869 It is a good idea to reset the value of @code{top-of-ranges} in case
20870 you have set it to some different value. You can evaluate the
20875 (setq top-of-ranges
20878 110 120 130 140 150
20879 160 170 180 190 200
20880 210 220 230 240 250
20881 260 270 280 290 300)
20886 Next create a list of the number of words and symbols in each range.
20890 Evaluate the following:
20894 (setq list-for-graph
20897 (recursive-lengths-list-many-files
20898 (directory-files "/usr/local/emacs/lisp"
20906 On my old machine, this took about an hour. It looked though 303 Lisp
20907 files in my copy of Emacs version 19.23. After all that computing,
20908 the @code{list-for-graph} had this value:
20912 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
20913 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
20918 This means that my copy of Emacs had 537 function definitions with
20919 fewer than 10 words or symbols in them, 1,027 function definitions
20920 with 10 to 19 words or symbols in them, 955 function definitions with
20921 20 to 29 words or symbols in them, and so on.
20923 Clearly, just by looking at this list we can see that most function
20924 definitions contain ten to thirty words and symbols.
20926 Now for printing. We do @emph{not} want to print a graph that is
20927 1,030 lines high @dots{} Instead, we should print a graph that is
20928 fewer than twenty-five lines high. A graph that height can be
20929 displayed on almost any monitor, and easily printed on a sheet of paper.
20931 This means that each value in @code{list-for-graph} must be reduced to
20932 one-fiftieth its present value.
20934 Here is a short function to do just that, using two functions we have
20935 not yet seen, @code{mapcar} and @code{lambda}.
20939 (defun one-fiftieth (full-range)
20940 "Return list, each number one-fiftieth of previous."
20941 (mapcar (lambda (arg) (/ arg 50)) full-range))
20946 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
20947 @cindex Anonymous function
20950 @code{lambda} is the symbol for an anonymous function, a function
20951 without a name. Every time you use an anonymous function, you need to
20952 include its whole body.
20959 (lambda (arg) (/ arg 50))
20963 is a function definition that says `return the value resulting from
20964 dividing whatever is passed to me as @code{arg} by 50'.
20967 Earlier, for example, we had a function @code{multiply-by-seven}; it
20968 multiplied its argument by 7. This function is similar, except it
20969 divides its argument by 50; and, it has no name. The anonymous
20970 equivalent of @code{multiply-by-seven} is:
20973 (lambda (number) (* 7 number))
20977 (@xref{defun, , The @code{defun} Special Form}.)
20981 If we want to multiply 3 by 7, we can write:
20983 @c !!! Clear print-postscript-figures if the computer formatting this
20984 @c document is too small and cannot handle all the diagrams and figures.
20985 @c clear print-postscript-figures
20986 @c set print-postscript-figures
20987 @c lambda example diagram #1
20991 (multiply-by-seven 3)
20992 \_______________/ ^
20998 @ifset print-postscript-figures
21001 @center @image{lambda-1}
21002 %%%% old method of including an image
21003 % \input /usr/local/lib/tex/inputs/psfig.tex
21004 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21009 @ifclear print-postscript-figures
21013 (multiply-by-seven 3)
21014 \_______________/ ^
21023 This expression returns 21.
21027 Similarly, we can write:
21029 @c lambda example diagram #2
21033 ((lambda (number) (* 7 number)) 3)
21034 \____________________________/ ^
21036 anonymous function argument
21040 @ifset print-postscript-figures
21043 @center @image{lambda-2}
21044 %%%% old method of including an image
21045 % \input /usr/local/lib/tex/inputs/psfig.tex
21046 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21051 @ifclear print-postscript-figures
21055 ((lambda (number) (* 7 number)) 3)
21056 \____________________________/ ^
21058 anonymous function argument
21066 If we want to divide 100 by 50, we can write:
21068 @c lambda example diagram #3
21072 ((lambda (arg) (/ arg 50)) 100)
21073 \______________________/ \_/
21075 anonymous function argument
21079 @ifset print-postscript-figures
21082 @center @image{lambda-3}
21083 %%%% old method of including an image
21084 % \input /usr/local/lib/tex/inputs/psfig.tex
21085 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21090 @ifclear print-postscript-figures
21094 ((lambda (arg) (/ arg 50)) 100)
21095 \______________________/ \_/
21097 anonymous function argument
21104 This expression returns 2. The 100 is passed to the function, which
21105 divides that number by 50.
21107 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21108 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21109 expressions derive from the Lambda Calculus.
21112 @appendixsubsec The @code{mapcar} Function
21115 @code{mapcar} is a function that calls its first argument with each
21116 element of its second argument, in turn. The second argument must be
21119 The @samp{map} part of the name comes from the mathematical phrase,
21120 `mapping over a domain', meaning to apply a function to each of the
21121 elements in a domain. The mathematical phrase is based on the
21122 metaphor of a surveyor walking, one step at a time, over an area he is
21123 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21132 (mapcar '1+ '(2 4 6))
21138 The function @code{1+} which adds one to its argument, is executed on
21139 @emph{each} element of the list, and a new list is returned.
21141 Contrast this with @code{apply}, which applies its first argument to
21143 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21147 In the definition of @code{one-fiftieth}, the first argument is the
21148 anonymous function:
21151 (lambda (arg) (/ arg 50))
21155 and the second argument is @code{full-range}, which will be bound to
21156 @code{list-for-graph}.
21159 The whole expression looks like this:
21162 (mapcar (lambda (arg) (/ arg 50)) full-range))
21165 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21166 Lisp Reference Manual}, for more about @code{mapcar}.
21168 Using the @code{one-fiftieth} function, we can generate a list in
21169 which each element is one-fiftieth the size of the corresponding
21170 element in @code{list-for-graph}.
21174 (setq fiftieth-list-for-graph
21175 (one-fiftieth list-for-graph))
21180 The resulting list looks like this:
21184 (10 20 19 15 11 9 6 5 4 3 3 2 2
21185 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21190 This, we are almost ready to print! (We also notice the loss of
21191 information: many of the higher ranges are 0, meaning that fewer than
21192 50 defuns had that many words or symbols---but not necessarily meaning
21193 that none had that many words or symbols.)
21196 @appendixsubsec Another Bug @dots{} Most Insidious
21197 @cindex Bug, most insidious type
21198 @cindex Insidious type of bug
21200 I said `almost ready to print'! Of course, there is a bug in the
21201 @code{print-graph} function @dots{} It has a @code{vertical-step}
21202 option, but not a @code{horizontal-step} option. The
21203 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21204 @code{print-graph} function will print only by ones.
21206 This is a classic example of what some consider the most insidious
21207 type of bug, the bug of omission. This is not the kind of bug you can
21208 find by studying the code, for it is not in the code; it is an omitted
21209 feature. Your best actions are to try your program early and often;
21210 and try to arrange, as much as you can, to write code that is easy to
21211 understand and easy to change. Try to be aware, whenever you can,
21212 that whatever you have written, @emph{will} be rewritten, if not soon,
21213 eventually. A hard maxim to follow.
21215 It is the @code{print-X-axis-numbered-line} function that needs the
21216 work; and then the @code{print-X-axis} and the @code{print-graph}
21217 functions need to be adapted. Not much needs to be done; there is one
21218 nicety: the numbers ought to line up under the tic marks. This takes
21222 Here is the corrected @code{print-X-axis-numbered-line}:
21226 (defun print-X-axis-numbered-line
21227 (number-of-X-tics X-axis-leading-spaces
21228 &optional horizontal-step)
21229 "Print line of X-axis numbers"
21230 (let ((number X-axis-label-spacing)
21231 (horizontal-step (or horizontal-step 1)))
21234 (insert X-axis-leading-spaces)
21235 ;; @r{Delete extra leading spaces.}
21238 (length (number-to-string horizontal-step)))))
21243 ;; @r{Insert white space.}
21245 X-axis-label-spacing)
21248 (number-to-string horizontal-step)))
21252 (* number horizontal-step))))
21255 ;; @r{Insert remaining numbers.}
21256 (setq number (+ number X-axis-label-spacing))
21257 (while (> number-of-X-tics 1)
21258 (insert (X-axis-element
21259 (* number horizontal-step)))
21260 (setq number (+ number X-axis-label-spacing))
21261 (setq number-of-X-tics (1- number-of-X-tics)))))
21266 If you are reading this in Info, you can see the new versions of
21267 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21268 reading this in a printed book, you can see the changed lines here
21269 (the full text is too much to print).
21274 (defun print-X-axis (numbers-list horizontal-step)
21276 (print-X-axis-numbered-line
21277 tic-number leading-spaces horizontal-step))
21285 &optional vertical-step horizontal-step)
21287 (print-X-axis numbers-list horizontal-step))
21295 (defun print-X-axis (numbers-list horizontal-step)
21296 "Print X axis labels to length of NUMBERS-LIST.
21297 Optionally, HORIZONTAL-STEP, a positive integer,
21298 specifies how much an X axis label increments for
21302 ;; Value of symbol-width and full-Y-label-width
21303 ;; are passed by `print-graph'.
21304 (let* ((leading-spaces
21305 (make-string full-Y-label-width ? ))
21306 ;; symbol-width @r{is provided by} graph-body-print
21307 (tic-width (* symbol-width X-axis-label-spacing))
21308 (X-length (length numbers-list))
21314 ;; @r{Make a string of blanks.}
21315 (- (* symbol-width X-axis-label-spacing)
21316 (length X-axis-tic-symbol))
21320 ;; @r{Concatenate blanks with tic symbol.}
21321 X-axis-tic-symbol))
21323 (if (zerop (% X-length tic-width))
21324 (/ X-length tic-width)
21325 (1+ (/ X-length tic-width)))))
21329 (print-X-axis-tic-line
21330 tic-number leading-spaces X-tic)
21332 (print-X-axis-numbered-line
21333 tic-number leading-spaces horizontal-step)))
21340 (numbers-list &optional vertical-step horizontal-step)
21341 "Print labeled bar graph of the NUMBERS-LIST.
21342 The numbers-list consists of the Y-axis values.
21346 Optionally, VERTICAL-STEP, a positive integer,
21347 specifies how much a Y axis label increments for
21348 each line. For example, a step of 5 means that
21349 each row is five units.
21353 Optionally, HORIZONTAL-STEP, a positive integer,
21354 specifies how much an X axis label increments for
21356 (let* ((symbol-width (length graph-blank))
21357 ;; @code{height} @r{is both the largest number}
21358 ;; @r{and the number with the most digits.}
21359 (height (apply 'max numbers-list))
21362 (height-of-top-line
21363 (if (zerop (% height Y-axis-label-spacing))
21366 (* (1+ (/ height Y-axis-label-spacing))
21367 Y-axis-label-spacing)))
21370 (vertical-step (or vertical-step 1))
21371 (full-Y-label-width
21375 (* height-of-top-line vertical-step))
21380 height-of-top-line full-Y-label-width vertical-step)
21382 numbers-list height-of-top-line symbol-width)
21383 (print-X-axis numbers-list horizontal-step)))
21390 Graphing Definitions Re-listed
21393 Here are all the graphing definitions in their final form:
21397 (defvar top-of-ranges
21400 110 120 130 140 150
21401 160 170 180 190 200
21402 210 220 230 240 250)
21403 "List specifying ranges for `defuns-per-range'.")
21407 (defvar graph-symbol "*"
21408 "String used as symbol in graph, usually an asterisk.")
21412 (defvar graph-blank " "
21413 "String used as blank in graph, usually a blank space.
21414 graph-blank must be the same number of columns wide
21419 (defvar Y-axis-tic " - "
21420 "String that follows number in a Y axis label.")
21424 (defvar Y-axis-label-spacing 5
21425 "Number of lines from one Y axis label to next.")
21429 (defvar X-axis-tic-symbol "|"
21430 "String to insert to point to a column in X axis.")
21434 (defvar X-axis-label-spacing
21435 (if (boundp 'graph-blank)
21436 (* 5 (length graph-blank)) 5)
21437 "Number of units from one X axis label to next.")
21443 (defun count-words-in-defun ()
21444 "Return the number of words and symbols in a defun."
21445 (beginning-of-defun)
21447 (end (save-excursion (end-of-defun) (point))))
21452 (and (< (point) end)
21454 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21456 (setq count (1+ count)))
21463 (defun lengths-list-file (filename)
21464 "Return list of definitions' lengths within FILE.
21465 The returned list is a list of numbers.
21466 Each number is the number of words or
21467 symbols in one function definition."
21471 (message "Working on `%s' ... " filename)
21473 (let ((buffer (find-file-noselect filename))
21475 (set-buffer buffer)
21476 (setq buffer-read-only t)
21478 (goto-char (point-min))
21482 (while (re-search-forward "^(defun" nil t)
21484 (cons (count-words-in-defun) lengths-list)))
21485 (kill-buffer buffer)
21492 (defun lengths-list-many-files (list-of-files)
21493 "Return list of lengths of defuns in LIST-OF-FILES."
21494 (let (lengths-list)
21495 ;;; @r{true-or-false-test}
21496 (while list-of-files
21502 ;;; @r{Generate a lengths' list.}
21504 (expand-file-name (car list-of-files)))))
21505 ;;; @r{Make files' list shorter.}
21506 (setq list-of-files (cdr list-of-files)))
21507 ;;; @r{Return final value of lengths' list.}
21514 (defun defuns-per-range (sorted-lengths top-of-ranges)
21515 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21516 (let ((top-of-range (car top-of-ranges))
21517 (number-within-range 0)
21518 defuns-per-range-list)
21523 (while top-of-ranges
21527 ;; @r{Need number for numeric test.}
21528 (car sorted-lengths)
21529 (< (car sorted-lengths) top-of-range))
21531 ;; @r{Count number of definitions within current range.}
21532 (setq number-within-range (1+ number-within-range))
21533 (setq sorted-lengths (cdr sorted-lengths)))
21537 ;; @r{Exit inner loop but remain within outer loop.}
21539 (setq defuns-per-range-list
21540 (cons number-within-range defuns-per-range-list))
21541 (setq number-within-range 0) ; @r{Reset count to zero.}
21543 ;; @r{Move to next range.}
21544 (setq top-of-ranges (cdr top-of-ranges))
21545 ;; @r{Specify next top of range value.}
21546 (setq top-of-range (car top-of-ranges)))
21550 ;; @r{Exit outer loop and count the number of defuns larger than}
21551 ;; @r{ the largest top-of-range value.}
21552 (setq defuns-per-range-list
21554 (length sorted-lengths)
21555 defuns-per-range-list))
21557 ;; @r{Return a list of the number of definitions within each range,}
21558 ;; @r{ smallest to largest.}
21559 (nreverse defuns-per-range-list)))
21565 (defun column-of-graph (max-graph-height actual-height)
21566 "Return list of MAX-GRAPH-HEIGHT strings;
21567 ACTUAL-HEIGHT are graph-symbols.
21568 The graph-symbols are contiguous entries at the end
21570 The list will be inserted as one column of a graph.
21571 The strings are either graph-blank or graph-symbol."
21575 (let ((insert-list nil)
21576 (number-of-top-blanks
21577 (- max-graph-height actual-height)))
21579 ;; @r{Fill in @code{graph-symbols}.}
21580 (while (> actual-height 0)
21581 (setq insert-list (cons graph-symbol insert-list))
21582 (setq actual-height (1- actual-height)))
21586 ;; @r{Fill in @code{graph-blanks}.}
21587 (while (> number-of-top-blanks 0)
21588 (setq insert-list (cons graph-blank insert-list))
21589 (setq number-of-top-blanks
21590 (1- number-of-top-blanks)))
21592 ;; @r{Return whole list.}
21599 (defun Y-axis-element (number full-Y-label-width)
21600 "Construct a NUMBERed label element.
21601 A numbered element looks like this ` 5 - ',
21602 and is padded as needed so all line up with
21603 the element for the largest number."
21606 (let* ((leading-spaces
21607 (- full-Y-label-width
21609 (concat (number-to-string number)
21614 (make-string leading-spaces ? )
21615 (number-to-string number)
21622 (defun print-Y-axis
21623 (height full-Y-label-width &optional vertical-step)
21624 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21625 Height must be the maximum height of the graph.
21626 Full width is the width of the highest label element.
21627 Optionally, print according to VERTICAL-STEP."
21630 ;; Value of height and full-Y-label-width
21631 ;; are passed by `print-graph'.
21632 (let ((start (point)))
21634 (Y-axis-column height full-Y-label-width vertical-step))
21637 ;; @r{Place point ready for inserting graph.}
21639 ;; @r{Move point forward by value of} full-Y-label-width
21640 (forward-char full-Y-label-width)))
21646 (defun print-X-axis-tic-line
21647 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21648 "Print ticks for X axis."
21649 (insert X-axis-leading-spaces)
21650 (insert X-axis-tic-symbol) ; @r{Under first column.}
21653 ;; @r{Insert second tic in the right spot.}
21656 (- (* symbol-width X-axis-label-spacing)
21657 ;; @r{Insert white space up to second tic symbol.}
21658 (* 2 (length X-axis-tic-symbol)))
21660 X-axis-tic-symbol))
21663 ;; @r{Insert remaining ticks.}
21664 (while (> number-of-X-tics 1)
21665 (insert X-axis-tic-element)
21666 (setq number-of-X-tics (1- number-of-X-tics))))
21672 (defun X-axis-element (number)
21673 "Construct a numbered X axis element."
21674 (let ((leading-spaces
21675 (- (* symbol-width X-axis-label-spacing)
21676 (length (number-to-string number)))))
21677 (concat (make-string leading-spaces ? )
21678 (number-to-string number))))
21684 (defun graph-body-print (numbers-list height symbol-width)
21685 "Print a bar graph of the NUMBERS-LIST.
21686 The numbers-list consists of the Y-axis values.
21687 HEIGHT is maximum height of graph.
21688 SYMBOL-WIDTH is number of each column."
21691 (let (from-position)
21692 (while numbers-list
21693 (setq from-position (point))
21695 (column-of-graph height (car numbers-list)))
21696 (goto-char from-position)
21697 (forward-char symbol-width)
21700 ;; @r{Draw graph column by column.}
21702 (setq numbers-list (cdr numbers-list)))
21703 ;; @r{Place point for X axis labels.}
21704 (forward-line height)
21711 (defun Y-axis-column
21712 (height width-of-label &optional vertical-step)
21713 "Construct list of labels for Y axis.
21714 HEIGHT is maximum height of graph.
21715 WIDTH-OF-LABEL is maximum width of label.
21718 VERTICAL-STEP, an option, is a positive integer
21719 that specifies how much a Y axis label increments
21720 for each line. For example, a step of 5 means
21721 that each line is five units of the graph."
21723 (number-per-line (or vertical-step 1)))
21726 (while (> height 1)
21727 (if (zerop (% height Y-axis-label-spacing))
21728 ;; @r{Insert label.}
21732 (* height number-per-line)
21737 ;; @r{Else, insert blanks.}
21740 (make-string width-of-label ? )
21742 (setq height (1- height)))
21745 ;; @r{Insert base line.}
21746 (setq Y-axis (cons (Y-axis-element
21747 (or vertical-step 1)
21750 (nreverse Y-axis)))
21756 (defun print-X-axis-numbered-line
21757 (number-of-X-tics X-axis-leading-spaces
21758 &optional horizontal-step)
21759 "Print line of X-axis numbers"
21760 (let ((number X-axis-label-spacing)
21761 (horizontal-step (or horizontal-step 1)))
21764 (insert X-axis-leading-spaces)
21766 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21769 ;; @r{Insert white space up to next number.}
21770 (- (* symbol-width X-axis-label-spacing)
21771 (1- (length (number-to-string horizontal-step)))
21774 (number-to-string (* number horizontal-step))))
21777 ;; @r{Insert remaining numbers.}
21778 (setq number (+ number X-axis-label-spacing))
21779 (while (> number-of-X-tics 1)
21780 (insert (X-axis-element (* number horizontal-step)))
21781 (setq number (+ number X-axis-label-spacing))
21782 (setq number-of-X-tics (1- number-of-X-tics)))))
21788 (defun print-X-axis (numbers-list horizontal-step)
21789 "Print X axis labels to length of NUMBERS-LIST.
21790 Optionally, HORIZONTAL-STEP, a positive integer,
21791 specifies how much an X axis label increments for
21795 ;; Value of symbol-width and full-Y-label-width
21796 ;; are passed by `print-graph'.
21797 (let* ((leading-spaces
21798 (make-string full-Y-label-width ? ))
21799 ;; symbol-width @r{is provided by} graph-body-print
21800 (tic-width (* symbol-width X-axis-label-spacing))
21801 (X-length (length numbers-list))
21807 ;; @r{Make a string of blanks.}
21808 (- (* symbol-width X-axis-label-spacing)
21809 (length X-axis-tic-symbol))
21813 ;; @r{Concatenate blanks with tic symbol.}
21814 X-axis-tic-symbol))
21816 (if (zerop (% X-length tic-width))
21817 (/ X-length tic-width)
21818 (1+ (/ X-length tic-width)))))
21822 (print-X-axis-tic-line
21823 tic-number leading-spaces X-tic)
21825 (print-X-axis-numbered-line
21826 tic-number leading-spaces horizontal-step)))
21832 (defun one-fiftieth (full-range)
21833 "Return list, each number of which is 1/50th previous."
21834 (mapcar (lambda (arg) (/ arg 50)) full-range))
21841 (numbers-list &optional vertical-step horizontal-step)
21842 "Print labeled bar graph of the NUMBERS-LIST.
21843 The numbers-list consists of the Y-axis values.
21847 Optionally, VERTICAL-STEP, a positive integer,
21848 specifies how much a Y axis label increments for
21849 each line. For example, a step of 5 means that
21850 each row is five units.
21854 Optionally, HORIZONTAL-STEP, a positive integer,
21855 specifies how much an X axis label increments for
21857 (let* ((symbol-width (length graph-blank))
21858 ;; @code{height} @r{is both the largest number}
21859 ;; @r{and the number with the most digits.}
21860 (height (apply 'max numbers-list))
21863 (height-of-top-line
21864 (if (zerop (% height Y-axis-label-spacing))
21867 (* (1+ (/ height Y-axis-label-spacing))
21868 Y-axis-label-spacing)))
21871 (vertical-step (or vertical-step 1))
21872 (full-Y-label-width
21876 (* height-of-top-line vertical-step))
21882 height-of-top-line full-Y-label-width vertical-step)
21884 numbers-list height-of-top-line symbol-width)
21885 (print-X-axis numbers-list horizontal-step)))
21892 @node Final printed graph
21893 @appendixsubsec The Printed Graph
21895 When made and installed, you can call the @code{print-graph} command
21901 (print-graph fiftieth-list-for-graph 50 10)
21931 50 - ***************** * *
21933 10 50 100 150 200 250 300 350
21940 The largest group of functions contain 10--19 words and symbols each.
21942 @node Free Software and Free Manuals
21943 @appendix Free Software and Free Manuals
21945 @strong{by Richard M. Stallman}
21948 The biggest deficiency in free operating systems is not in the
21949 software---it is the lack of good free manuals that we can include in
21950 these systems. Many of our most important programs do not come with
21951 full manuals. Documentation is an essential part of any software
21952 package; when an important free software package does not come with a
21953 free manual, that is a major gap. We have many such gaps today.
21955 Once upon a time, many years ago, I thought I would learn Perl. I got
21956 a copy of a free manual, but I found it hard to read. When I asked
21957 Perl users about alternatives, they told me that there were better
21958 introductory manuals---but those were not free.
21960 Why was this? The authors of the good manuals had written them for
21961 O'Reilly Associates, which published them with restrictive terms---no
21962 copying, no modification, source files not available---which exclude
21963 them from the free software community.
21965 That wasn't the first time this sort of thing has happened, and (to
21966 our community's great loss) it was far from the last. Proprietary
21967 manual publishers have enticed a great many authors to restrict their
21968 manuals since then. Many times I have heard a GNU user eagerly tell me
21969 about a manual that he is writing, with which he expects to help the
21970 GNU project---and then had my hopes dashed, as he proceeded to explain
21971 that he had signed a contract with a publisher that would restrict it
21972 so that we cannot use it.
21974 Given that writing good English is a rare skill among programmers, we
21975 can ill afford to lose manuals this way.
21977 Free documentation, like free software, is a matter of freedom, not
21978 price. The problem with these manuals was not that O'Reilly Associates
21979 charged a price for printed copies---that in itself is fine. The Free
21980 Software Foundation @uref{http://shop.fsf.org, sells printed copies} of
21981 free @uref{http://www.gnu.org/doc/doc.html, GNU manuals}, too.
21982 But GNU manuals are available in source code form, while these manuals
21983 are available only on paper. GNU manuals come with permission to copy
21984 and modify; the Perl manuals do not. These restrictions are the
21987 The criterion for a free manual is pretty much the same as for free
21988 software: it is a matter of giving all users certain
21989 freedoms. Redistribution (including commercial redistribution) must be
21990 permitted, so that the manual can accompany every copy of the program,
21991 on-line or on paper. Permission for modification is crucial too.
21993 As a general rule, I don't believe that it is essential for people to
21994 have permission to modify all sorts of articles and books. The issues
21995 for writings are not necessarily the same as those for software. For
21996 example, I don't think you or I are obliged to give permission to
21997 modify articles like this one, which describe our actions and our
22000 But there is a particular reason why the freedom to modify is crucial
22001 for documentation for free software. When people exercise their right
22002 to modify the software, and add or change its features, if they are
22003 conscientious they will change the manual too---so they can provide
22004 accurate and usable documentation with the modified program. A manual
22005 which forbids programmers to be conscientious and finish the job, or
22006 more precisely requires them to write a new manual from scratch if
22007 they change the program, does not fill our community's needs.
22009 While a blanket prohibition on modification is unacceptable, some
22010 kinds of limits on the method of modification pose no problem. For
22011 example, requirements to preserve the original author's copyright
22012 notice, the distribution terms, or the list of authors, are ok. It is
22013 also no problem to require modified versions to include notice that
22014 they were modified, even to have entire sections that may not be
22015 deleted or changed, as long as these sections deal with nontechnical
22016 topics. (Some GNU manuals have them.)
22018 These kinds of restrictions are not a problem because, as a practical
22019 matter, they don't stop the conscientious programmer from adapting the
22020 manual to fit the modified program. In other words, they don't block
22021 the free software community from making full use of the manual.
22023 However, it must be possible to modify all the technical content of
22024 the manual, and then distribute the result in all the usual media,
22025 through all the usual channels; otherwise, the restrictions do block
22026 the community, the manual is not free, and so we need another manual.
22028 Unfortunately, it is often hard to find someone to write another
22029 manual when a proprietary manual exists. The obstacle is that many
22030 users think that a proprietary manual is good enough---so they don't
22031 see the need to write a free manual. They do not see that the free
22032 operating system has a gap that needs filling.
22034 Why do users think that proprietary manuals are good enough? Some have
22035 not considered the issue. I hope this article will do something to
22038 Other users consider proprietary manuals acceptable for the same
22039 reason so many people consider proprietary software acceptable: they
22040 judge in purely practical terms, not using freedom as a
22041 criterion. These people are entitled to their opinions, but since
22042 those opinions spring from values which do not include freedom, they
22043 are no guide for those of us who do value freedom.
22045 Please spread the word about this issue. We continue to lose manuals
22046 to proprietary publishing. If we spread the word that proprietary
22047 manuals are not sufficient, perhaps the next person who wants to help
22048 GNU by writing documentation will realize, before it is too late, that
22049 he must above all make it free.
22051 We can also encourage commercial publishers to sell free, copylefted
22052 manuals instead of proprietary ones. One way you can help this is to
22053 check the distribution terms of a manual before you buy it, and prefer
22054 copylefted manuals to non-copylefted ones.
22058 Note: The Free Software Foundation maintains a page on its Web site
22059 that lists free books available from other publishers:@*
22060 @uref{http://www.gnu.org/doc/other-free-books.html}
22062 @node GNU Free Documentation License
22063 @appendix GNU Free Documentation License
22065 @cindex FDL, GNU Free Documentation License
22066 @include doclicense.texi
22072 MENU ENTRY: NODE NAME.
22078 @c Place biographical information on right-hand (verso) page
22081 \par\vfill\supereject
22083 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22084 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22087 % \par\vfill\supereject
22088 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22089 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22090 %\page\hbox{}%\page
22091 %\page\hbox{}%\page
22098 @c ================ Biographical information ================
22102 @center About the Author
22107 @node About the Author
22108 @unnumbered About the Author
22112 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22113 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22114 world on software freedom. Chassell was a founding Director and
22115 Treasurer of the Free Software Foundation, Inc. He is co-author of
22116 the @cite{Texinfo} manual, and has edited more than a dozen other
22117 books. He graduated from Cambridge University, in England. He has an
22118 abiding interest in social and economic history and flies his own
22125 @c @c Prevent page number on blank verso, so eject it first.
22127 @c \par\vfill\supereject
22132 @c @evenheading @thispage @| @| @thistitle
22133 @c @oddheading @| @| @thispage