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 ----------------------------------------------------
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}
231 Copyright @copyright{} 1990, 1991, 1992, 1993, 1994, 1995, 1997, 2001,
232 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 Free Software Foundation, Inc.
238 GNU Press, @hfill @uref{http://www.gnupress.org}@*
239 a division of the @hfill General: @email{press@@gnu.org}@*
240 Free Software Foundation, Inc. @hfill Orders:@w{ } @email{sales@@gnu.org}@*
241 51 Franklin Street, Fifth Floor @hfill Tel: +1 (617) 542-5942@*
242 Boston, MA 02110-1301 USA @hfill Fax: +1 (617) 542-2652@*
249 GNU Press, Website: http://www.gnupress.org
250 a division of the General: press@@gnu.org
251 Free Software Foundation, Inc. Orders: sales@@gnu.org
252 51 Franklin Street, Fifth Floor Tel: +1 (617) 542-5942
253 Boston, MA 02110-1301 USA Fax: +1 (617) 542-2652
258 @c Printed copies are available for $30 each.@*
261 Permission is granted to copy, distribute and/or modify this document
262 under the terms of the GNU Free Documentation License, Version 1.3 or
263 any later version published by the Free Software Foundation; there
264 being no Invariant Section, with the Front-Cover Texts being ``A GNU
265 Manual'', and with the Back-Cover Texts as in (a) below. A copy of
266 the license is included in the section entitled ``GNU Free
267 Documentation License''.
269 (a) The FSF's Back-Cover Text is: ``You have the freedom to
270 copy and modify this GNU manual. Buying copies from the FSF
271 supports it in developing GNU and promoting software freedom.''
274 @c half title; two lines here, so do not use `shorttitlepage'
277 \hbox{}\vskip 1.5in \chaprm \centerline{An Introduction to}%
279 {\begingroup\hbox{}\vskip 0.25in \chaprm%
280 \centerline{Programming in Emacs Lisp}%
281 \endgroup\page\hbox{}\page}
286 @center @titlefont{An Introduction to}
288 @center @titlefont{Programming in Emacs Lisp}
290 @center Revised Third Edition
292 @center by Robert J. Chassell
295 @vskip 0pt plus 1filll
301 @evenheading @thispage @| @| @thischapter
302 @oddheading @thissection @| @| @thispage
306 @c Keep T.O.C. short by tightening up for largebook
309 \global\parskip 2pt plus 1pt
310 \global\advance\baselineskip by -1pt
319 @node Top, Preface, (dir), (dir)
320 @top An Introduction to Programming in Emacs Lisp
324 This master menu first lists each chapter and index; then it lists
325 every node in every chapter.
328 @c >>>> Set pageno appropriately <<<<
330 @c The first page of the Preface is a roman numeral; it is the first
331 @c right handed page after the Table of Contents; hence the following
332 @c setting must be for an odd negative number.
335 @c global@pageno = -11
339 * Preface:: What to look for.
340 * List Processing:: What is Lisp?
341 * Practicing Evaluation:: Running several programs.
342 * Writing Defuns:: How to write function definitions.
343 * Buffer Walk Through:: Exploring a few buffer-related functions.
344 * More Complex:: A few, even more complex functions.
345 * Narrowing & Widening:: Restricting your and Emacs attention to
347 * car cdr & cons:: Fundamental functions in Lisp.
348 * Cutting & Storing Text:: Removing text and saving it.
349 * List Implementation:: How lists are implemented in the computer.
350 * Yanking:: Pasting stored text.
351 * Loops & Recursion:: How to repeat a process.
352 * Regexp Search:: Regular expression searches.
353 * Counting Words:: A review of repetition and regexps.
354 * Words in a defun:: Counting words in a @code{defun}.
355 * Readying a Graph:: A prototype graph printing function.
356 * Emacs Initialization:: How to write a @file{.emacs} file.
357 * Debugging:: How to run the Emacs Lisp debuggers.
358 * Conclusion:: Now you have the basics.
359 * the-the:: An appendix: how to find reduplicated words.
360 * Kill Ring:: An appendix: how the kill ring works.
361 * Full Graph:: How to create a graph with labelled axes.
362 * Free Software and Free Manuals::
363 * GNU Free Documentation License::
368 --- The Detailed Node Listing ---
372 * Why:: Why learn Emacs Lisp?
373 * On Reading this Text:: Read, gain familiarity, pick up habits....
374 * Who You Are:: For whom this is written.
376 * Note for Novices:: You can read this as a novice.
381 * Lisp Lists:: What are lists?
382 * Run a Program:: Any list in Lisp is a program ready to run.
383 * Making Errors:: Generating an error message.
384 * Names & Definitions:: Names of symbols and function definitions.
385 * Lisp Interpreter:: What the Lisp interpreter does.
386 * Evaluation:: Running a program.
387 * Variables:: Returning a value from a variable.
388 * Arguments:: Passing information to a function.
389 * set & setq:: Setting the value of a variable.
390 * Summary:: The major points.
391 * Error Message Exercises::
395 * Numbers Lists:: List have numbers, other lists, in them.
396 * Lisp Atoms:: Elemental entities.
397 * Whitespace in Lists:: Formatting lists to be readable.
398 * Typing Lists:: How GNU Emacs helps you type lists.
402 * Complications:: Variables, Special forms, Lists within.
403 * Byte Compiling:: Specially processing code for speed.
407 * How the Interpreter Acts:: Returns and Side Effects...
408 * Evaluating Inner Lists:: Lists within lists...
412 * fill-column Example::
413 * Void Function:: The error message for a symbol
415 * Void Variable:: The error message for a symbol without a value.
419 * Data types:: Types of data passed to a function.
420 * Args as Variable or List:: An argument can be the value
421 of a variable or list.
422 * Variable Number of Arguments:: Some functions may take a
423 variable number of arguments.
424 * Wrong Type of Argument:: Passing an argument of the wrong type
426 * message:: A useful function for sending messages.
428 Setting the Value of a Variable
430 * Using set:: Setting values.
431 * Using setq:: Setting a quoted value.
432 * Counting:: Using @code{setq} to count.
434 Practicing Evaluation
436 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
438 * Buffer Names:: Buffers and files are different.
439 * Getting Buffers:: Getting a buffer itself, not merely its name.
440 * Switching Buffers:: How to change to another buffer.
441 * Buffer Size & Locations:: Where point is located and the size of
443 * Evaluation Exercise::
445 How To Write Function Definitions
447 * Primitive Functions::
448 * defun:: The @code{defun} special form.
449 * Install:: Install a function definition.
450 * Interactive:: Making a function interactive.
451 * Interactive Options:: Different options for @code{interactive}.
452 * Permanent Installation:: Installing code permanently.
453 * let:: Creating and initializing local variables.
455 * else:: If--then--else expressions.
456 * Truth & Falsehood:: What Lisp considers false and true.
457 * save-excursion:: Keeping track of point, mark, and buffer.
461 Install a Function Definition
463 * Effect of installation::
464 * Change a defun:: How to change a function definition.
466 Make a Function Interactive
468 * Interactive multiply-by-seven:: An overview.
469 * multiply-by-seven in detail:: The interactive version.
473 * Prevent confusion::
474 * Parts of let Expression::
475 * Sample let Expression::
476 * Uninitialized let Variables::
478 The @code{if} Special Form
480 * if in more detail::
481 * type-of-animal in detail:: An example of an @code{if} expression.
483 Truth and Falsehood in Emacs Lisp
485 * nil explained:: @code{nil} has two meanings.
487 @code{save-excursion}
489 * Point and mark:: A review of various locations.
490 * Template for save-excursion::
492 A Few Buffer--Related Functions
494 * Finding More:: How to find more information.
495 * simplified-beginning-of-buffer:: Shows @code{goto-char},
496 @code{point-min}, and @code{push-mark}.
497 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
498 * append-to-buffer:: Uses @code{save-excursion} and
499 @code{insert-buffer-substring}.
500 * Buffer Related Review:: Review.
503 The Definition of @code{mark-whole-buffer}
505 * mark-whole-buffer overview::
506 * Body of mark-whole-buffer:: Only three lines of code.
508 The Definition of @code{append-to-buffer}
510 * append-to-buffer overview::
511 * append interactive:: A two part interactive expression.
512 * append-to-buffer body:: Incorporates a @code{let} expression.
513 * append save-excursion:: How the @code{save-excursion} works.
515 A Few More Complex Functions
517 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
518 * insert-buffer:: Read-only, and with @code{or}.
519 * beginning-of-buffer:: Shows @code{goto-char},
520 @code{point-min}, and @code{push-mark}.
521 * Second Buffer Related Review::
522 * optional Exercise::
524 The Definition of @code{insert-buffer}
526 * insert-buffer code::
527 * insert-buffer interactive:: When you can read, but not write.
528 * insert-buffer body:: The body has an @code{or} and a @code{let}.
529 * if & or:: Using an @code{if} instead of an @code{or}.
530 * Insert or:: How the @code{or} expression works.
531 * Insert let:: Two @code{save-excursion} expressions.
532 * New insert-buffer::
534 The Interactive Expression in @code{insert-buffer}
536 * Read-only buffer:: When a buffer cannot be modified.
537 * b for interactive:: An existing buffer or else its name.
539 Complete Definition of @code{beginning-of-buffer}
541 * Optional Arguments::
542 * beginning-of-buffer opt arg:: Example with optional argument.
543 * beginning-of-buffer complete::
545 @code{beginning-of-buffer} with an Argument
547 * Disentangle beginning-of-buffer::
548 * Large buffer case::
549 * Small buffer case::
551 Narrowing and Widening
553 * Narrowing advantages:: The advantages of narrowing
554 * save-restriction:: The @code{save-restriction} special form.
555 * what-line:: The number of the line that point is on.
558 @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
560 * Strange Names:: An historical aside: why the strange names?
561 * car & cdr:: Functions for extracting part of a list.
562 * cons:: Constructing a list.
563 * nthcdr:: Calling @code{cdr} repeatedly.
565 * setcar:: Changing the first element of a list.
566 * setcdr:: Changing the rest of a list.
572 * length:: How to find the length of a list.
574 Cutting and Storing Text
576 * Storing Text:: Text is stored in a list.
577 * zap-to-char:: Cutting out text up to a character.
578 * kill-region:: Cutting text out of a region.
579 * copy-region-as-kill:: A definition for copying text.
580 * Digression into C:: Minor note on C programming language macros.
581 * defvar:: How to give a variable an initial value.
582 * cons & search-fwd Review::
587 * Complete zap-to-char:: The complete implementation.
588 * zap-to-char interactive:: A three part interactive expression.
589 * zap-to-char body:: A short overview.
590 * search-forward:: How to search for a string.
591 * progn:: The @code{progn} special form.
592 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
596 * Complete kill-region:: The function definition.
597 * condition-case:: Dealing with a problem.
600 @code{copy-region-as-kill}
602 * Complete copy-region-as-kill:: The complete function definition.
603 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
605 The Body of @code{copy-region-as-kill}
607 * last-command & this-command::
608 * kill-append function::
609 * kill-new function::
611 Initializing a Variable with @code{defvar}
613 * See variable current value::
614 * defvar and asterisk::
616 How Lists are Implemented
619 * Symbols as Chest:: Exploring a powerful metaphor.
624 * Kill Ring Overview::
625 * kill-ring-yank-pointer:: The kill ring is a list.
626 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
630 * while:: Causing a stretch of code to repeat.
632 * Recursion:: Causing a function to call itself.
637 * Looping with while:: Repeat so long as test returns true.
638 * Loop Example:: A @code{while} loop that uses a list.
639 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
640 * Incrementing Loop:: A loop with an incrementing counter.
641 * Incrementing Loop Details::
642 * Decrementing Loop:: A loop with a decrementing counter.
644 Details of an Incrementing Loop
646 * Incrementing Example:: Counting pebbles in a triangle.
647 * Inc Example parts:: The parts of the function definition.
648 * Inc Example altogether:: Putting the function definition together.
650 Loop with a Decrementing Counter
652 * Decrementing Example:: More pebbles on the beach.
653 * Dec Example parts:: The parts of the function definition.
654 * Dec Example altogether:: Putting the function definition together.
656 Save your time: @code{dolist} and @code{dotimes}
663 * Building Robots:: Same model, different serial number ...
664 * Recursive Definition Parts:: Walk until you stop ...
665 * Recursion with list:: Using a list as the test whether to recurse.
666 * Recursive triangle function::
667 * Recursion with cond::
668 * Recursive Patterns:: Often used templates.
669 * No Deferment:: Don't store up work ...
670 * No deferment solution::
672 Recursion in Place of a Counter
674 * Recursive Example arg of 1 or 2::
675 * Recursive Example arg of 3 or 4::
683 Regular Expression Searches
685 * sentence-end:: The regular expression for @code{sentence-end}.
686 * re-search-forward:: Very similar to @code{search-forward}.
687 * forward-sentence:: A straightforward example of regexp search.
688 * forward-paragraph:: A somewhat complex example.
689 * etags:: How to create your own @file{TAGS} table.
691 * re-search Exercises::
693 @code{forward-sentence}
695 * Complete forward-sentence::
696 * fwd-sentence while loops:: Two @code{while} loops.
697 * fwd-sentence re-search:: A regular expression search.
699 @code{forward-paragraph}: a Goldmine of Functions
701 * forward-paragraph in brief:: Key parts of the function definition.
702 * fwd-para let:: The @code{let*} expression.
703 * fwd-para while:: The forward motion @code{while} loop.
705 Counting: Repetition and Regexps
708 * count-words-region:: Use a regexp, but find a problem.
709 * recursive-count-words:: Start with case of no words in region.
710 * Counting Exercise::
712 The @code{count-words-region} Function
714 * Design count-words-region:: The definition using a @code{while} loop.
715 * Whitespace Bug:: The Whitespace Bug in @code{count-words-region}.
717 Counting Words in a @code{defun}
719 * Divide and Conquer::
720 * Words and Symbols:: What to count?
721 * Syntax:: What constitutes a word or symbol?
722 * count-words-in-defun:: Very like @code{count-words}.
723 * Several defuns:: Counting several defuns in a file.
724 * Find a File:: Do you want to look at a file?
725 * lengths-list-file:: A list of the lengths of many definitions.
726 * Several files:: Counting in definitions in different files.
727 * Several files recursively:: Recursively counting in different files.
728 * Prepare the data:: Prepare the data for display in a graph.
730 Count Words in @code{defuns} in Different Files
732 * lengths-list-many-files:: Return a list of the lengths of defuns.
733 * append:: Attach one list to another.
735 Prepare the Data for Display in a Graph
737 * Data for Display in Detail::
738 * Sorting:: Sorting lists.
739 * Files List:: Making a list of files.
740 * Counting function definitions::
744 * Columns of a graph::
745 * graph-body-print:: How to print the body of a graph.
746 * recursive-graph-body-print::
748 * Line Graph Exercise::
750 Your @file{.emacs} File
752 * Default Configuration::
753 * Site-wide Init:: You can write site-wide init files.
754 * defcustom:: Emacs will write code for you.
755 * Beginning a .emacs File:: How to write a @code{.emacs file}.
756 * Text and Auto-fill:: Automatically wrap lines.
757 * Mail Aliases:: Use abbreviations for email addresses.
758 * Indent Tabs Mode:: Don't use tabs with @TeX{}
759 * Keybindings:: Create some personal keybindings.
760 * Keymaps:: More about key binding.
761 * Loading Files:: Load (i.e., evaluate) files automatically.
762 * Autoload:: Make functions available.
763 * Simple Extension:: Define a function; bind it to a key.
764 * X11 Colors:: Colors in X.
766 * Mode Line:: How to customize your mode line.
770 * debug:: How to use the built-in debugger.
771 * debug-on-entry:: Start debugging when you call a function.
772 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
773 * edebug:: How to use Edebug, a source level debugger.
774 * Debugging Exercises::
776 Handling the Kill Ring
778 * What the Kill Ring Does::
780 * yank:: Paste a copy of a clipped element.
781 * yank-pop:: Insert element pointed to.
784 The @code{current-kill} Function
786 * Code for current-kill::
787 * Understanding current-kill::
789 @code{current-kill} in Outline
791 * Body of current-kill::
792 * Digression concerning error:: How to mislead humans, but not computers.
793 * Determining the Element::
795 A Graph with Labelled Axes
798 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
799 * print-Y-axis:: Print a label for the vertical axis.
800 * print-X-axis:: Print a horizontal label.
801 * Print Whole Graph:: The function to print a complete graph.
803 The @code{print-Y-axis} Function
805 * print-Y-axis in Detail::
806 * Height of label:: What height for the Y axis?
807 * Compute a Remainder:: How to compute the remainder of a division.
808 * Y Axis Element:: Construct a line for the Y axis.
809 * Y-axis-column:: Generate a list of Y axis labels.
810 * print-Y-axis Penultimate:: A not quite final version.
812 The @code{print-X-axis} Function
814 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
815 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
817 Printing the Whole Graph
819 * The final version:: A few changes.
820 * Test print-graph:: Run a short test.
821 * Graphing words in defuns:: Executing the final code.
822 * lambda:: How to write an anonymous function.
823 * mapcar:: Apply a function to elements of a list.
824 * Another Bug:: Yet another bug @dots{} most insidious.
825 * Final printed graph:: The graph itself!
830 @node Preface, List Processing, Top, Top
831 @comment node-name, next, previous, up
834 Most of the GNU Emacs integrated environment is written in the programming
835 language called Emacs Lisp. The code written in this programming
836 language is the software---the sets of instructions---that tell the
837 computer what to do when you give it commands. Emacs is designed so
838 that you can write new code in Emacs Lisp and easily install it as an
839 extension to the editor.
841 (GNU Emacs is sometimes called an ``extensible editor'', but it does
842 much more than provide editing capabilities. It is better to refer to
843 Emacs as an ``extensible computing environment''. However, that
844 phrase is quite a mouthful. It is easier to refer to Emacs simply as
845 an editor. Moreover, everything you do in Emacs---find the Mayan date
846 and phases of the moon, simplify polynomials, debug code, manage
847 files, read letters, write books---all these activities are kinds of
848 editing in the most general sense of the word.)
851 * Why:: Why learn Emacs Lisp?
852 * On Reading this Text:: Read, gain familiarity, pick up habits....
853 * Who You Are:: For whom this is written.
855 * Note for Novices:: You can read this as a novice.
859 @node Why, On Reading this Text, Preface, Preface
861 @unnumberedsec Why Study Emacs Lisp?
864 Although Emacs Lisp is usually thought of in association only with Emacs,
865 it is a full computer programming language. You can use Emacs Lisp as
866 you would any other programming language.
868 Perhaps you want to understand programming; perhaps you want to extend
869 Emacs; or perhaps you want to become a programmer. This introduction to
870 Emacs Lisp is designed to get you started: to guide you in learning the
871 fundamentals of programming, and more importantly, to show you how you
872 can teach yourself to go further.
874 @node On Reading this Text, Who You Are, Why, Preface
875 @comment node-name, next, previous, up
876 @unnumberedsec On Reading this Text
878 All through this document, you will see little sample programs you can
879 run inside of Emacs. If you read this document in Info inside of GNU
880 Emacs, you can run the programs as they appear. (This is easy to do and
881 is explained when the examples are presented.) Alternatively, you can
882 read this introduction as a printed book while sitting beside a computer
883 running Emacs. (This is what I like to do; I like printed books.) If
884 you don't have a running Emacs beside you, you can still read this book,
885 but in this case, it is best to treat it as a novel or as a travel guide
886 to a country not yet visited: interesting, but not the same as being
889 Much of this introduction is dedicated to walk-throughs or guided tours
890 of code used in GNU Emacs. These tours are designed for two purposes:
891 first, to give you familiarity with real, working code (code you use
892 every day); and, second, to give you familiarity with the way Emacs
893 works. It is interesting to see how a working environment is
896 hope that you will pick up the habit of browsing through source code.
897 You can learn from it and mine it for ideas. Having GNU Emacs is like
898 having a dragon's cave of treasures.
900 In addition to learning about Emacs as an editor and Emacs Lisp as a
901 programming language, the examples and guided tours will give you an
902 opportunity to get acquainted with Emacs as a Lisp programming
903 environment. GNU Emacs supports programming and provides tools that
904 you will want to become comfortable using, such as @kbd{M-.} (the key
905 which invokes the @code{find-tag} command). You will also learn about
906 buffers and other objects that are part of the environment.
907 Learning about these features of Emacs is like learning new routes
908 around your home town.
911 In addition, I have written several programs as extended examples.
912 Although these are examples, the programs are real. I use them.
913 Other people use them. You may use them. Beyond the fragments of
914 programs used for illustrations, there is very little in here that is
915 `just for teaching purposes'; what you see is used. This is a great
916 advantage of Emacs Lisp: it is easy to learn to use it for work.
919 Finally, I hope to convey some of the skills for using Emacs to
920 learn aspects of programming that you don't know. You can often use
921 Emacs to help you understand what puzzles you or to find out how to do
922 something new. This self-reliance is not only a pleasure, but an
925 @node Who You Are, Lisp History, On Reading this Text, Preface
926 @comment node-name, next, previous, up
927 @unnumberedsec For Whom This is Written
929 This text is written as an elementary introduction for people who are
930 not programmers. If you are a programmer, you may not be satisfied with
931 this primer. The reason is that you may have become expert at reading
932 reference manuals and be put off by the way this text is organized.
934 An expert programmer who reviewed this text said to me:
937 @i{I prefer to learn from reference manuals. I ``dive into'' each
938 paragraph, and ``come up for air'' between paragraphs.}
940 @i{When I get to the end of a paragraph, I assume that that subject is
941 done, finished, that I know everything I need (with the
942 possible exception of the case when the next paragraph starts talking
943 about it in more detail). I expect that a well written reference manual
944 will not have a lot of redundancy, and that it will have excellent
945 pointers to the (one) place where the information I want is.}
948 This introduction is not written for this person!
950 Firstly, I try to say everything at least three times: first, to
951 introduce it; second, to show it in context; and third, to show it in a
952 different context, or to review it.
954 Secondly, I hardly ever put all the information about a subject in one
955 place, much less in one paragraph. To my way of thinking, that imposes
956 too heavy a burden on the reader. Instead I try to explain only what
957 you need to know at the time. (Sometimes I include a little extra
958 information so you won't be surprised later when the additional
959 information is formally introduced.)
961 When you read this text, you are not expected to learn everything the
962 first time. Frequently, you need only make, as it were, a `nodding
963 acquaintance' with some of the items mentioned. My hope is that I have
964 structured the text and given you enough hints that you will be alert to
965 what is important, and concentrate on it.
967 You will need to ``dive into'' some paragraphs; there is no other way
968 to read them. But I have tried to keep down the number of such
969 paragraphs. This book is intended as an approachable hill, rather than
970 as a daunting mountain.
972 This introduction to @cite{Programming in Emacs Lisp} has a companion
975 @cite{The GNU Emacs Lisp Reference Manual}.
978 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
979 Emacs Lisp Reference Manual}.
981 The reference manual has more detail than this introduction. In the
982 reference manual, all the information about one topic is concentrated
983 in one place. You should turn to it if you are like the programmer
984 quoted above. And, of course, after you have read this
985 @cite{Introduction}, you will find the @cite{Reference Manual} useful
986 when you are writing your own programs.
988 @node Lisp History, Note for Novices, Who You Are, Preface
989 @unnumberedsec Lisp History
992 Lisp was first developed in the late 1950s at the Massachusetts
993 Institute of Technology for research in artificial intelligence. The
994 great power of the Lisp language makes it superior for other purposes as
995 well, such as writing editor commands and integrated environments.
999 GNU Emacs Lisp is largely inspired by Maclisp, which was written at MIT
1000 in the 1960s. It is somewhat inspired by Common Lisp, which became a
1001 standard in the 1980s. However, Emacs Lisp is much simpler than Common
1002 Lisp. (The standard Emacs distribution contains an optional extensions
1003 file, @file{cl.el}, that adds many Common Lisp features to Emacs Lisp.)
1005 @node Note for Novices, Thank You, Lisp History, Preface
1006 @comment node-name, next, previous, up
1007 @unnumberedsec A Note for Novices
1009 If you don't know GNU Emacs, you can still read this document
1010 profitably. However, I recommend you learn Emacs, if only to learn to
1011 move around your computer screen. You can teach yourself how to use
1012 Emacs with the on-line tutorial. To use it, type @kbd{C-h t}. (This
1013 means you press and release the @key{CTRL} key and the @kbd{h} at the
1014 same time, and then press and release @kbd{t}.)
1016 Also, I often refer to one of Emacs' standard commands by listing the
1017 keys which you press to invoke the command and then giving the name of
1018 the command in parentheses, like this: @kbd{M-C-\}
1019 (@code{indent-region}). What this means is that the
1020 @code{indent-region} command is customarily invoked by typing
1021 @kbd{M-C-\}. (You can, if you wish, change the keys that are typed to
1022 invoke the command; this is called @dfn{rebinding}. @xref{Keymaps, ,
1023 Keymaps}.) The abbreviation @kbd{M-C-\} means that you type your
1024 @key{META} key, @key{CTRL} key and @key{\} key all at the same time.
1025 (On many modern keyboards the @key{META} key is labelled
1027 Sometimes a combination like this is called a keychord, since it is
1028 similar to the way you play a chord on a piano. If your keyboard does
1029 not have a @key{META} key, the @key{ESC} key prefix is used in place
1030 of it. In this case, @kbd{M-C-\} means that you press and release your
1031 @key{ESC} key and then type the @key{CTRL} key and the @key{\} key at
1032 the same time. But usually @kbd{M-C-\} means press the @key{CTRL} key
1033 along with the key that is labelled @key{ALT} and, at the same time,
1034 press the @key{\} key.
1036 In addition to typing a lone keychord, you can prefix what you type
1037 with @kbd{C-u}, which is called the `universal argument'. The
1038 @kbd{C-u} keychord passes an argument to the subsequent command.
1039 Thus, to indent a region of plain text by 6 spaces, mark the region,
1040 and then type @w{@kbd{C-u 6 M-C-\}}. (If you do not specify a number,
1041 Emacs either passes the number 4 to the command or otherwise runs the
1042 command differently than it would otherwise.) @xref{Arguments, ,
1043 Numeric Arguments, emacs, The GNU Emacs Manual}.
1045 If you are reading this in Info using GNU Emacs, you can read through
1046 this whole document just by pressing the space bar, @key{SPC}.
1047 (To learn about Info, type @kbd{C-h i} and then select Info.)
1049 A note on terminology: when I use the word Lisp alone, I often am
1050 referring to the various dialects of Lisp in general, but when I speak
1051 of Emacs Lisp, I am referring to GNU Emacs Lisp in particular.
1053 @node Thank You, , Note for Novices, Preface
1054 @comment node-name, next, previous, up
1055 @unnumberedsec Thank You
1057 My thanks to all who helped me with this book. My especial thanks to
1058 @r{Jim Blandy}, @r{Noah Friedman}, @w{Jim Kingdon}, @r{Roland
1059 McGrath}, @w{Frank Ritter}, @w{Randy Smith}, @w{Richard M.@:
1060 Stallman}, and @w{Melissa Weisshaus}. My thanks also go to both
1061 @w{Philip Johnson} and @w{David Stampe} for their patient
1062 encouragement. My mistakes are my own.
1066 @email{bob@@gnu.org}
1069 @c ================ Beginning of main text ================
1071 @c Start main text on right-hand (verso) page
1074 \par\vfill\supereject
1077 \par\vfill\supereject
1079 \par\vfill\supereject
1081 \par\vfill\supereject
1087 @evenheading @thispage @| @| @thischapter
1088 @oddheading @thissection @| @| @thispage
1092 @node List Processing, Practicing Evaluation, Preface, Top
1093 @comment node-name, next, previous, up
1094 @chapter List Processing
1096 To the untutored eye, Lisp is a strange programming language. In Lisp
1097 code there are parentheses everywhere. Some people even claim that
1098 the name stands for `Lots of Isolated Silly Parentheses'. But the
1099 claim is unwarranted. Lisp stands for LISt Processing, and the
1100 programming language handles @emph{lists} (and lists of lists) by
1101 putting them between parentheses. The parentheses mark the boundaries
1102 of the list. Sometimes a list is preceded by a single apostrophe or
1103 quotation mark, @samp{'}@footnote{The single apostrophe or quotation
1104 mark is an abbreviation for the function @code{quote}; you need not
1105 think about functions now; functions are defined in @ref{Making
1106 Errors, , Generate an Error Message}.} Lists are the basis of Lisp.
1109 * Lisp Lists:: What are lists?
1110 * Run a Program:: Any list in Lisp is a program ready to run.
1111 * Making Errors:: Generating an error message.
1112 * Names & Definitions:: Names of symbols and function definitions.
1113 * Lisp Interpreter:: What the Lisp interpreter does.
1114 * Evaluation:: Running a program.
1115 * Variables:: Returning a value from a variable.
1116 * Arguments:: Passing information to a function.
1117 * set & setq:: Setting the value of a variable.
1118 * Summary:: The major points.
1119 * Error Message Exercises::
1122 @node Lisp Lists, Run a Program, List Processing, List Processing
1123 @comment node-name, next, previous, up
1127 In Lisp, a list looks like this: @code{'(rose violet daisy buttercup)}.
1128 This list is preceded by a single apostrophe. It could just as well be
1129 written as follows, which looks more like the kind of list you are likely
1130 to be familiar with:
1142 The elements of this list are the names of the four different flowers,
1143 separated from each other by whitespace and surrounded by parentheses,
1144 like flowers in a field with a stone wall around them.
1145 @cindex Flowers in a field
1148 * Numbers Lists:: List have numbers, other lists, in them.
1149 * Lisp Atoms:: Elemental entities.
1150 * Whitespace in Lists:: Formatting lists to be readable.
1151 * Typing Lists:: How GNU Emacs helps you type lists.
1154 @node Numbers Lists, Lisp Atoms, Lisp Lists, Lisp Lists
1156 @unnumberedsubsec Numbers, Lists inside of Lists
1159 Lists can also have numbers in them, as in this list: @code{(+ 2 2)}.
1160 This list has a plus-sign, @samp{+}, followed by two @samp{2}s, each
1161 separated by whitespace.
1163 In Lisp, both data and programs are represented the same way; that is,
1164 they are both lists of words, numbers, or other lists, separated by
1165 whitespace and surrounded by parentheses. (Since a program looks like
1166 data, one program may easily serve as data for another; this is a very
1167 powerful feature of Lisp.) (Incidentally, these two parenthetical
1168 remarks are @emph{not} Lisp lists, because they contain @samp{;} and
1169 @samp{.} as punctuation marks.)
1172 Here is another list, this time with a list inside of it:
1175 '(this list has (a list inside of it))
1178 The components of this list are the words @samp{this}, @samp{list},
1179 @samp{has}, and the list @samp{(a list inside of it)}. The interior
1180 list is made up of the words @samp{a}, @samp{list}, @samp{inside},
1181 @samp{of}, @samp{it}.
1183 @node Lisp Atoms, Whitespace in Lists, Numbers Lists, Lisp Lists
1184 @comment node-name, next, previous, up
1185 @subsection Lisp Atoms
1188 In Lisp, what we have been calling words are called @dfn{atoms}. This
1189 term comes from the historical meaning of the word atom, which means
1190 `indivisible'. As far as Lisp is concerned, the words we have been
1191 using in the lists cannot be divided into any smaller parts and still
1192 mean the same thing as part of a program; likewise with numbers and
1193 single character symbols like @samp{+}. On the other hand, unlike an
1194 ancient atom, a list can be split into parts. (@xref{car cdr & cons,
1195 , @code{car} @code{cdr} & @code{cons} Fundamental Functions}.)
1197 In a list, atoms are separated from each other by whitespace. They can be
1198 right next to a parenthesis.
1200 @cindex @samp{empty list} defined
1201 Technically speaking, a list in Lisp consists of parentheses surrounding
1202 atoms separated by whitespace or surrounding other lists or surrounding
1203 both atoms and other lists. A list can have just one atom in it or
1204 have nothing in it at all. A list with nothing in it looks like this:
1205 @code{()}, and is called the @dfn{empty list}. Unlike anything else, an
1206 empty list is considered both an atom and a list at the same time.
1208 @cindex Symbolic expressions, introduced
1209 @cindex @samp{expression} defined
1210 @cindex @samp{form} defined
1211 The printed representation of both atoms and lists are called
1212 @dfn{symbolic expressions} or, more concisely, @dfn{s-expressions}.
1213 The word @dfn{expression} by itself can refer to either the printed
1214 representation, or to the atom or list as it is held internally in the
1215 computer. Often, people use the term @dfn{expression}
1216 indiscriminately. (Also, in many texts, the word @dfn{form} is used
1217 as a synonym for expression.)
1219 Incidentally, the atoms that make up our universe were named such when
1220 they were thought to be indivisible; but it has been found that physical
1221 atoms are not indivisible. Parts can split off an atom or it can
1222 fission into two parts of roughly equal size. Physical atoms were named
1223 prematurely, before their truer nature was found. In Lisp, certain
1224 kinds of atom, such as an array, can be separated into parts; but the
1225 mechanism for doing this is different from the mechanism for splitting a
1226 list. As far as list operations are concerned, the atoms of a list are
1229 As in English, the meanings of the component letters of a Lisp atom
1230 are different from the meaning the letters make as a word. For
1231 example, the word for the South American sloth, the @samp{ai}, is
1232 completely different from the two words, @samp{a}, and @samp{i}.
1234 There are many kinds of atom in nature but only a few in Lisp: for
1235 example, @dfn{numbers}, such as 37, 511, or 1729, and @dfn{symbols}, such
1236 as @samp{+}, @samp{foo}, or @samp{forward-line}. The words we have
1237 listed in the examples above are all symbols. In everyday Lisp
1238 conversation, the word ``atom'' is not often used, because programmers
1239 usually try to be more specific about what kind of atom they are dealing
1240 with. Lisp programming is mostly about symbols (and sometimes numbers)
1241 within lists. (Incidentally, the preceding three word parenthetical
1242 remark is a proper list in Lisp, since it consists of atoms, which in
1243 this case are symbols, separated by whitespace and enclosed by
1244 parentheses, without any non-Lisp punctuation.)
1247 Text between double quotation marks---even sentences or
1248 paragraphs---is also an atom. Here is an example:
1249 @cindex Text between double quotation marks
1252 '(this list includes "text between quotation marks.")
1255 @cindex @samp{string} defined
1257 In Lisp, all of the quoted text including the punctuation mark and the
1258 blank spaces is a single atom. This kind of atom is called a
1259 @dfn{string} (for `string of characters') and is the sort of thing that
1260 is used for messages that a computer can print for a human to read.
1261 Strings are a different kind of atom than numbers or symbols and are
1264 @node Whitespace in Lists, Typing Lists, Lisp Atoms, Lisp Lists
1265 @comment node-name, next, previous, up
1266 @subsection Whitespace in Lists
1267 @cindex Whitespace in lists
1270 The amount of whitespace in a list does not matter. From the point of view
1271 of the Lisp language,
1282 is exactly the same as this:
1285 '(this list looks like this)
1288 Both examples show what to Lisp is the same list, the list made up of
1289 the symbols @samp{this}, @samp{list}, @samp{looks}, @samp{like}, and
1290 @samp{this} in that order.
1292 Extra whitespace and newlines are designed to make a list more readable
1293 by humans. When Lisp reads the expression, it gets rid of all the extra
1294 whitespace (but it needs to have at least one space between atoms in
1295 order to tell them apart.)
1297 Odd as it seems, the examples we have seen cover almost all of what Lisp
1298 lists look like! Every other list in Lisp looks more or less like one
1299 of these examples, except that the list may be longer and more complex.
1300 In brief, a list is between parentheses, a string is between quotation
1301 marks, a symbol looks like a word, and a number looks like a number.
1302 (For certain situations, square brackets, dots and a few other special
1303 characters may be used; however, we will go quite far without them.)
1305 @node Typing Lists, , Whitespace in Lists, Lisp Lists
1306 @comment node-name, next, previous, up
1307 @subsection GNU Emacs Helps You Type Lists
1308 @cindex Help typing lists
1309 @cindex Formatting help
1311 When you type a Lisp expression in GNU Emacs using either Lisp
1312 Interaction mode or Emacs Lisp mode, you have available to you several
1313 commands to format the Lisp expression so it is easy to read. For
1314 example, pressing the @key{TAB} key automatically indents the line the
1315 cursor is on by the right amount. A command to properly indent the
1316 code in a region is customarily bound to @kbd{M-C-\}. Indentation is
1317 designed so that you can see which elements of a list belong to which
1318 list---elements of a sub-list are indented more than the elements of
1321 In addition, when you type a closing parenthesis, Emacs momentarily
1322 jumps the cursor back to the matching opening parenthesis, so you can
1323 see which one it is. This is very useful, since every list you type
1324 in Lisp must have its closing parenthesis match its opening
1325 parenthesis. (@xref{Major Modes, , Major Modes, emacs, The GNU Emacs
1326 Manual}, for more information about Emacs' modes.)
1328 @node Run a Program, Making Errors, Lisp Lists, List Processing
1329 @comment node-name, next, previous, up
1330 @section Run a Program
1331 @cindex Run a program
1332 @cindex Program, running one
1334 @cindex @samp{evaluate} defined
1335 A list in Lisp---any list---is a program ready to run. If you run it
1336 (for which the Lisp jargon is @dfn{evaluate}), the computer will do one
1337 of three things: do nothing except return to you the list itself; send
1338 you an error message; or, treat the first symbol in the list as a
1339 command to do something. (Usually, of course, it is the last of these
1340 three things that you really want!)
1342 @c use code for the single apostrophe, not samp.
1343 The single apostrophe, @code{'}, that I put in front of some of the
1344 example lists in preceding sections is called a @dfn{quote}; when it
1345 precedes a list, it tells Lisp to do nothing with the list, other than
1346 take it as it is written. But if there is no quote preceding a list,
1347 the first item of the list is special: it is a command for the computer
1348 to obey. (In Lisp, these commands are called @emph{functions}.) The list
1349 @code{(+ 2 2)} shown above did not have a quote in front of it, so Lisp
1350 understands that the @code{+} is an instruction to do something with the
1351 rest of the list: add the numbers that follow.
1354 If you are reading this inside of GNU Emacs in Info, here is how you can
1355 evaluate such a list: place your cursor immediately after the right
1356 hand parenthesis of the following list and then type @kbd{C-x C-e}:
1362 @c use code for the number four, not samp.
1364 You will see the number @code{4} appear in the echo area. (In the
1365 jargon, what you have just done is ``evaluate the list.'' The echo area
1366 is the line at the bottom of the screen that displays or ``echoes''
1367 text.) Now try the same thing with a quoted list: place the cursor
1368 right after the following list and type @kbd{C-x C-e}:
1371 '(this is a quoted list)
1375 You will see @code{(this is a quoted list)} appear in the echo area.
1377 @cindex Lisp interpreter, explained
1378 @cindex Interpreter, Lisp, explained
1379 In both cases, what you are doing is giving a command to the program
1380 inside of GNU Emacs called the @dfn{Lisp interpreter}---giving the
1381 interpreter a command to evaluate the expression. The name of the Lisp
1382 interpreter comes from the word for the task done by a human who comes
1383 up with the meaning of an expression---who ``interprets'' it.
1385 You can also evaluate an atom that is not part of a list---one that is
1386 not surrounded by parentheses; again, the Lisp interpreter translates
1387 from the humanly readable expression to the language of the computer.
1388 But before discussing this (@pxref{Variables}), we will discuss what the
1389 Lisp interpreter does when you make an error.
1391 @node Making Errors, Names & Definitions, Run a Program, List Processing
1392 @comment node-name, next, previous, up
1393 @section Generate an Error Message
1394 @cindex Generate an error message
1395 @cindex Error message generation
1397 Partly so you won't worry if you do it accidentally, we will now give
1398 a command to the Lisp interpreter that generates an error message.
1399 This is a harmless activity; and indeed, we will often try to generate
1400 error messages intentionally. Once you understand the jargon, error
1401 messages can be informative. Instead of being called ``error''
1402 messages, they should be called ``help'' messages. They are like
1403 signposts to a traveller in a strange country; deciphering them can be
1404 hard, but once understood, they can point the way.
1406 The error message is generated by a built-in GNU Emacs debugger. We
1407 will `enter the debugger'. You get out of the debugger by typing @code{q}.
1409 What we will do is evaluate a list that is not quoted and does not
1410 have a meaningful command as its first element. Here is a list almost
1411 exactly the same as the one we just used, but without the single-quote
1412 in front of it. Position the cursor right after it and type @kbd{C-x
1416 (this is an unquoted list)
1420 What you see depends on which version of Emacs you are running. GNU
1421 Emacs version 22 provides more information than version 20 and before.
1422 First, the more recent result of generating an error; then the
1423 earlier, version 20 result.
1427 In GNU Emacs version 22, a @file{*Backtrace*} window will open up and
1428 you will see the following in it:
1432 ---------- Buffer: *Backtrace* ----------
1433 Debugger entered--Lisp error: (void-function this)
1434 (this is an unquoted list)
1435 eval((this is an unquoted list))
1436 eval-last-sexp-1(nil)
1438 call-interactively(eval-last-sexp)
1439 ---------- Buffer: *Backtrace* ----------
1445 Your cursor will be in this window (you may have to wait a few seconds
1446 before it becomes visible). To quit the debugger and make the
1447 debugger window go away, type:
1454 Please type @kbd{q} right now, so you become confident that you can
1455 get out of the debugger. Then, type @kbd{C-x C-e} again to re-enter
1458 @cindex @samp{function} defined
1459 Based on what we already know, we can almost read this error message.
1461 You read the @file{*Backtrace*} buffer from the bottom up; it tells
1462 you what Emacs did. When you typed @kbd{C-x C-e}, you made an
1463 interactive call to the command @code{eval-last-sexp}. @code{eval} is
1464 an abbreviation for `evaluate' and @code{sexp} is an abbreviation for
1465 `symbolic expression'. The command means `evaluate last symbolic
1466 expression', which is the expression just before your cursor.
1468 Each line above tells you what the Lisp interpreter evaluated next.
1469 The most recent action is at the top. The buffer is called the
1470 @file{*Backtrace*} buffer because it enables you to track Emacs
1474 At the top of the @file{*Backtrace*} buffer, you see the line:
1477 Debugger entered--Lisp error: (void-function this)
1481 The Lisp interpreter tried to evaluate the first atom of the list, the
1482 word @samp{this}. It is this action that generated the error message
1483 @samp{void-function this}.
1485 The message contains the words @samp{void-function} and @samp{this}.
1487 @cindex @samp{function} defined
1488 The word @samp{function} was mentioned once before. It is a very
1489 important word. For our purposes, we can define it by saying that a
1490 @dfn{function} is a set of instructions to the computer that tell the
1491 computer to do something.
1493 Now we can begin to understand the error message: @samp{void-function
1494 this}. The function (that is, the word @samp{this}) does not have a
1495 definition of any set of instructions for the computer to carry out.
1497 The slightly odd word, @samp{void-function}, is designed to cover the
1498 way Emacs Lisp is implemented, which is that when a symbol does not
1499 have a function definition attached to it, the place that should
1500 contain the instructions is `void'.
1502 On the other hand, since we were able to add 2 plus 2 successfully, by
1503 evaluating @code{(+ 2 2)}, we can infer that the symbol @code{+} must
1504 have a set of instructions for the computer to obey and those
1505 instructions must be to add the numbers that follow the @code{+}.
1508 In GNU Emacs version 20, and in earlier versions, you will see only
1509 one line of error message; it will appear in the echo area and look
1513 Symbol's function definition is void:@: this
1517 (Also, your terminal may beep at you---some do, some don't; and others
1518 blink. This is just a device to get your attention.) The message goes
1519 away as soon as you type another key, even just to move the cursor.
1521 We know the meaning of the word @samp{Symbol}. It refers to the first
1522 atom of the list, the word @samp{this}. The word @samp{function}
1523 refers to the instructions that tell the computer what to do.
1524 (Technically, the symbol tells the computer where to find the
1525 instructions, but this is a complication we can ignore for the
1528 The error message can be understood: @samp{Symbol's function
1529 definition is void:@: this}. The symbol (that is, the word
1530 @samp{this}) lacks instructions for the computer to carry out.
1532 @node Names & Definitions, Lisp Interpreter, Making Errors, List Processing
1533 @comment node-name, next, previous, up
1534 @section Symbol Names and Function Definitions
1535 @cindex Symbol names
1537 We can articulate another characteristic of Lisp based on what we have
1538 discussed so far---an important characteristic: a symbol, like
1539 @code{+}, is not itself the set of instructions for the computer to
1540 carry out. Instead, the symbol is used, perhaps temporarily, as a way
1541 of locating the definition or set of instructions. What we see is the
1542 name through which the instructions can be found. Names of people
1543 work the same way. I can be referred to as @samp{Bob}; however, I am
1544 not the letters @samp{B}, @samp{o}, @samp{b} but am, or was, the
1545 consciousness consistently associated with a particular life-form.
1546 The name is not me, but it can be used to refer to me.
1548 In Lisp, one set of instructions can be attached to several names.
1549 For example, the computer instructions for adding numbers can be
1550 linked to the symbol @code{plus} as well as to the symbol @code{+}
1551 (and are in some dialects of Lisp). Among humans, I can be referred
1552 to as @samp{Robert} as well as @samp{Bob} and by other words as well.
1554 On the other hand, a symbol can have only one function definition
1555 attached to it at a time. Otherwise, the computer would be confused as
1556 to which definition to use. If this were the case among people, only
1557 one person in the world could be named @samp{Bob}. However, the function
1558 definition to which the name refers can be changed readily.
1559 (@xref{Install, , Install a Function Definition}.)
1561 Since Emacs Lisp is large, it is customary to name symbols in a way
1562 that identifies the part of Emacs to which the function belongs.
1563 Thus, all the names for functions that deal with Texinfo start with
1564 @samp{texinfo-} and those for functions that deal with reading mail
1565 start with @samp{rmail-}.
1567 @node Lisp Interpreter, Evaluation, Names & Definitions, List Processing
1568 @comment node-name, next, previous, up
1569 @section The Lisp Interpreter
1570 @cindex Lisp interpreter, what it does
1571 @cindex Interpreter, what it does
1573 Based on what we have seen, we can now start to figure out what the
1574 Lisp interpreter does when we command it to evaluate a list.
1575 First, it looks to see whether there is a quote before the list; if
1576 there is, the interpreter just gives us the list. On the other
1577 hand, if there is no quote, the interpreter looks at the first element
1578 in the list and sees whether it has a function definition. If it does,
1579 the interpreter carries out the instructions in the function definition.
1580 Otherwise, the interpreter prints an error message.
1582 This is how Lisp works. Simple. There are added complications which we
1583 will get to in a minute, but these are the fundamentals. Of course, to
1584 write Lisp programs, you need to know how to write function definitions
1585 and attach them to names, and how to do this without confusing either
1586 yourself or the computer.
1589 * Complications:: Variables, Special forms, Lists within.
1590 * Byte Compiling:: Specially processing code for speed.
1593 @node Complications, Byte Compiling, Lisp Interpreter, Lisp Interpreter
1595 @unnumberedsubsec Complications
1598 Now, for the first complication. In addition to lists, the Lisp
1599 interpreter can evaluate a symbol that is not quoted and does not have
1600 parentheses around it. The Lisp interpreter will attempt to determine
1601 the symbol's value as a @dfn{variable}. This situation is described
1602 in the section on variables. (@xref{Variables}.)
1604 @cindex Special form
1605 The second complication occurs because some functions are unusual and do
1606 not work in the usual manner. Those that don't are called @dfn{special
1607 forms}. They are used for special jobs, like defining a function, and
1608 there are not many of them. In the next few chapters, you will be
1609 introduced to several of the more important special forms.
1611 The third and final complication is this: if the function that the
1612 Lisp interpreter is looking at is not a special form, and if it is part
1613 of a list, the Lisp interpreter looks to see whether the list has a list
1614 inside of it. If there is an inner list, the Lisp interpreter first
1615 figures out what it should do with the inside list, and then it works on
1616 the outside list. If there is yet another list embedded inside the
1617 inner list, it works on that one first, and so on. It always works on
1618 the innermost list first. The interpreter works on the innermost list
1619 first, to evaluate the result of that list. The result may be
1620 used by the enclosing expression.
1622 Otherwise, the interpreter works left to right, from one expression to
1625 @node Byte Compiling, , Complications, Lisp Interpreter
1626 @subsection Byte Compiling
1627 @cindex Byte compiling
1629 One other aspect of interpreting: the Lisp interpreter is able to
1630 interpret two kinds of entity: humanly readable code, on which we will
1631 focus exclusively, and specially processed code, called @dfn{byte
1632 compiled} code, which is not humanly readable. Byte compiled code
1633 runs faster than humanly readable code.
1635 You can transform humanly readable code into byte compiled code by
1636 running one of the compile commands such as @code{byte-compile-file}.
1637 Byte compiled code is usually stored in a file that ends with a
1638 @file{.elc} extension rather than a @file{.el} extension. You will
1639 see both kinds of file in the @file{emacs/lisp} directory; the files
1640 to read are those with @file{.el} extensions.
1642 As a practical matter, for most things you might do to customize or
1643 extend Emacs, you do not need to byte compile; and I will not discuss
1644 the topic here. @xref{Byte Compilation, , Byte Compilation, elisp,
1645 The GNU Emacs Lisp Reference Manual}, for a full description of byte
1648 @node Evaluation, Variables, Lisp Interpreter, List Processing
1649 @comment node-name, next, previous, up
1653 When the Lisp interpreter works on an expression, the term for the
1654 activity is called @dfn{evaluation}. We say that the interpreter
1655 `evaluates the expression'. I've used this term several times before.
1656 The word comes from its use in everyday language, `to ascertain the
1657 value or amount of; to appraise', according to @cite{Webster's New
1658 Collegiate Dictionary}.
1661 * How the Interpreter Acts:: Returns and Side Effects...
1662 * Evaluating Inner Lists:: Lists within lists...
1665 @node How the Interpreter Acts, Evaluating Inner Lists, Evaluation, Evaluation
1667 @unnumberedsubsec How the Lisp Interpreter Acts
1670 @cindex @samp{returned value} explained
1671 After evaluating an expression, the Lisp interpreter will most likely
1672 @dfn{return} the value that the computer produces by carrying out the
1673 instructions it found in the function definition, or perhaps it will
1674 give up on that function and produce an error message. (The interpreter
1675 may also find itself tossed, so to speak, to a different function or it
1676 may attempt to repeat continually what it is doing for ever and ever in
1677 what is called an `infinite loop'. These actions are less common; and
1678 we can ignore them.) Most frequently, the interpreter returns a value.
1680 @cindex @samp{side effect} defined
1681 At the same time the interpreter returns a value, it may do something
1682 else as well, such as move a cursor or copy a file; this other kind of
1683 action is called a @dfn{side effect}. Actions that we humans think are
1684 important, such as printing results, are often ``side effects'' to the
1685 Lisp interpreter. The jargon can sound peculiar, but it turns out that
1686 it is fairly easy to learn to use side effects.
1688 In summary, evaluating a symbolic expression most commonly causes the
1689 Lisp interpreter to return a value and perhaps carry out a side effect;
1690 or else produce an error.
1692 @node Evaluating Inner Lists, , How the Interpreter Acts, Evaluation
1693 @comment node-name, next, previous, up
1694 @subsection Evaluating Inner Lists
1695 @cindex Inner list evaluation
1696 @cindex Evaluating inner lists
1698 If evaluation applies to a list that is inside another list, the outer
1699 list may use the value returned by the first evaluation as information
1700 when the outer list is evaluated. This explains why inner expressions
1701 are evaluated first: the values they return are used by the outer
1705 We can investigate this process by evaluating another addition example.
1706 Place your cursor after the following expression and type @kbd{C-x C-e}:
1713 The number 8 will appear in the echo area.
1715 What happens is that the Lisp interpreter first evaluates the inner
1716 expression, @code{(+ 3 3)}, for which the value 6 is returned; then it
1717 evaluates the outer expression as if it were written @code{(+ 2 6)}, which
1718 returns the value 8. Since there are no more enclosing expressions to
1719 evaluate, the interpreter prints that value in the echo area.
1721 Now it is easy to understand the name of the command invoked by the
1722 keystrokes @kbd{C-x C-e}: the name is @code{eval-last-sexp}. The
1723 letters @code{sexp} are an abbreviation for `symbolic expression', and
1724 @code{eval} is an abbreviation for `evaluate'. The command means
1725 `evaluate last symbolic expression'.
1727 As an experiment, you can try evaluating the expression by putting the
1728 cursor at the beginning of the next line immediately following the
1729 expression, or inside the expression.
1732 Here is another copy of the expression:
1739 If you place the cursor at the beginning of the blank line that
1740 immediately follows the expression and type @kbd{C-x C-e}, you will
1741 still get the value 8 printed in the echo area. Now try putting the
1742 cursor inside the expression. If you put it right after the next to
1743 last parenthesis (so it appears to sit on top of the last parenthesis),
1744 you will get a 6 printed in the echo area! This is because the command
1745 evaluates the expression @code{(+ 3 3)}.
1747 Now put the cursor immediately after a number. Type @kbd{C-x C-e} and
1748 you will get the number itself. In Lisp, if you evaluate a number, you
1749 get the number itself---this is how numbers differ from symbols. If you
1750 evaluate a list starting with a symbol like @code{+}, you will get a
1751 value returned that is the result of the computer carrying out the
1752 instructions in the function definition attached to that name. If a
1753 symbol by itself is evaluated, something different happens, as we will
1754 see in the next section.
1756 @node Variables, Arguments, Evaluation, List Processing
1757 @comment node-name, next, previous, up
1761 In Emacs Lisp, a symbol can have a value attached to it just as it can
1762 have a function definition attached to it. The two are different.
1763 The function definition is a set of instructions that a computer will
1764 obey. A value, on the other hand, is something, such as number or a
1765 name, that can vary (which is why such a symbol is called a variable).
1766 The value of a symbol can be any expression in Lisp, such as a symbol,
1767 number, list, or string. A symbol that has a value is often called a
1770 A symbol can have both a function definition and a value attached to
1771 it at the same time. Or it can have just one or the other.
1772 The two are separate. This is somewhat similar
1773 to the way the name Cambridge can refer to the city in Massachusetts
1774 and have some information attached to the name as well, such as
1775 ``great programming center''.
1778 (Incidentally, in Emacs Lisp, a symbol can have two
1779 other things attached to it, too: a property list and a documentation
1780 string; these are discussed later.)
1783 Another way to think about this is to imagine a symbol as being a chest
1784 of drawers. The function definition is put in one drawer, the value in
1785 another, and so on. What is put in the drawer holding the value can be
1786 changed without affecting the contents of the drawer holding the
1787 function definition, and vice-verse.
1790 * fill-column Example::
1791 * Void Function:: The error message for a symbol
1793 * Void Variable:: The error message for a symbol without a value.
1796 @node fill-column Example, Void Function, Variables, Variables
1798 @unnumberedsubsec @code{fill-column}, an Example Variable
1801 @findex fill-column, @r{an example variable}
1802 @cindex Example variable, @code{fill-column}
1803 @cindex Variable, example of, @code{fill-column}
1804 The variable @code{fill-column} illustrates a symbol with a value
1805 attached to it: in every GNU Emacs buffer, this symbol is set to some
1806 value, usually 72 or 70, but sometimes to some other value. To find the
1807 value of this symbol, evaluate it by itself. If you are reading this in
1808 Info inside of GNU Emacs, you can do this by putting the cursor after
1809 the symbol and typing @kbd{C-x C-e}:
1816 After I typed @kbd{C-x C-e}, Emacs printed the number 72 in my echo
1817 area. This is the value for which @code{fill-column} is set for me as I
1818 write this. It may be different for you in your Info buffer. Notice
1819 that the value returned as a variable is printed in exactly the same way
1820 as the value returned by a function carrying out its instructions. From
1821 the point of view of the Lisp interpreter, a value returned is a value
1822 returned. What kind of expression it came from ceases to matter once
1825 A symbol can have any value attached to it or, to use the jargon, we can
1826 @dfn{bind} the variable to a value: to a number, such as 72; to a
1827 string, @code{"such as this"}; to a list, such as @code{(spruce pine
1828 oak)}; we can even bind a variable to a function definition.
1830 A symbol can be bound to a value in several ways. @xref{set & setq, ,
1831 Setting the Value of a Variable}, for information about one way to do
1834 @node Void Function, Void Variable, fill-column Example, Variables
1835 @comment node-name, next, previous, up
1836 @subsection Error Message for a Symbol Without a Function
1837 @cindex Symbol without function error
1838 @cindex Error for symbol without function
1840 When we evaluated @code{fill-column} to find its value as a variable,
1841 we did not place parentheses around the word. This is because we did
1842 not intend to use it as a function name.
1844 If @code{fill-column} were the first or only element of a list, the
1845 Lisp interpreter would attempt to find the function definition
1846 attached to it. But @code{fill-column} has no function definition.
1847 Try evaluating this:
1855 In GNU Emacs version 22, you will create a @file{*Backtrace*} buffer
1860 ---------- Buffer: *Backtrace* ----------
1861 Debugger entered--Lisp error: (void-function fill-column)
1864 eval-last-sexp-1(nil)
1866 call-interactively(eval-last-sexp)
1867 ---------- Buffer: *Backtrace* ----------
1872 (Remember, to quit the debugger and make the debugger window go away,
1873 type @kbd{q} in the @file{*Backtrace*} buffer.)
1877 In GNU Emacs 20 and before, you will produce an error message that says:
1880 Symbol's function definition is void:@: fill-column
1884 (The message will go away as soon as you move the cursor or type
1888 @node Void Variable, , Void Function, Variables
1889 @comment node-name, next, previous, up
1890 @subsection Error Message for a Symbol Without a Value
1891 @cindex Symbol without value error
1892 @cindex Error for symbol without value
1894 If you attempt to evaluate a symbol that does not have a value bound to
1895 it, you will receive an error message. You can see this by
1896 experimenting with our 2 plus 2 addition. In the following expression,
1897 put your cursor right after the @code{+}, before the first number 2,
1906 In GNU Emacs 22, you will create a @file{*Backtrace*} buffer that
1911 ---------- Buffer: *Backtrace* ----------
1912 Debugger entered--Lisp error: (void-variable +)
1914 eval-last-sexp-1(nil)
1916 call-interactively(eval-last-sexp)
1917 ---------- Buffer: *Backtrace* ----------
1922 (As with the other times we entered the debugger, you can quit by
1923 typing @kbd{q} in the @file{*Backtrace*} buffer.)
1925 This backtrace is different from the very first error message we saw,
1926 which said, @samp{Debugger entered--Lisp error: (void-function this)}.
1927 In this case, the function does not have a value as a variable; while
1928 in the other error message, the function (the word `this') did not
1931 In this experiment with the @code{+}, what we did was cause the Lisp
1932 interpreter to evaluate the @code{+} and look for the value of the
1933 variable instead of the function definition. We did this by placing the
1934 cursor right after the symbol rather than after the parenthesis of the
1935 enclosing list as we did before. As a consequence, the Lisp interpreter
1936 evaluated the preceding s-expression, which in this case was the
1939 Since @code{+} does not have a value bound to it, just the function
1940 definition, the error message reported that the symbol's value as a
1945 In GNU Emacs version 20 and before, your error message will say:
1948 Symbol's value as variable is void:@: +
1952 The meaning is the same as in GNU Emacs 22.
1955 @node Arguments, set & setq, Variables, List Processing
1956 @comment node-name, next, previous, up
1959 @cindex Passing information to functions
1961 To see how information is passed to functions, let's look again at
1962 our old standby, the addition of two plus two. In Lisp, this is written
1969 If you evaluate this expression, the number 4 will appear in your echo
1970 area. What the Lisp interpreter does is add the numbers that follow
1973 @cindex @samp{argument} defined
1974 The numbers added by @code{+} are called the @dfn{arguments} of the
1975 function @code{+}. These numbers are the information that is given to
1976 or @dfn{passed} to the function.
1978 The word `argument' comes from the way it is used in mathematics and
1979 does not refer to a disputation between two people; instead it refers to
1980 the information presented to the function, in this case, to the
1981 @code{+}. In Lisp, the arguments to a function are the atoms or lists
1982 that follow the function. The values returned by the evaluation of
1983 these atoms or lists are passed to the function. Different functions
1984 require different numbers of arguments; some functions require none at
1985 all.@footnote{It is curious to track the path by which the word `argument'
1986 came to have two different meanings, one in mathematics and the other in
1987 everyday English. According to the @cite{Oxford English Dictionary},
1988 the word derives from the Latin for @samp{to make clear, prove}; thus it
1989 came to mean, by one thread of derivation, `the evidence offered as
1990 proof', which is to say, `the information offered', which led to its
1991 meaning in Lisp. But in the other thread of derivation, it came to mean
1992 `to assert in a manner against which others may make counter
1993 assertions', which led to the meaning of the word as a disputation.
1994 (Note here that the English word has two different definitions attached
1995 to it at the same time. By contrast, in Emacs Lisp, a symbol cannot
1996 have two different function definitions at the same time.)}
1999 * Data types:: Types of data passed to a function.
2000 * Args as Variable or List:: An argument can be the value
2001 of a variable or list.
2002 * Variable Number of Arguments:: Some functions may take a
2003 variable number of arguments.
2004 * Wrong Type of Argument:: Passing an argument of the wrong type
2006 * message:: A useful function for sending messages.
2009 @node Data types, Args as Variable or List, Arguments, Arguments
2010 @comment node-name, next, previous, up
2011 @subsection Arguments' Data Types
2013 @cindex Types of data
2014 @cindex Arguments' data types
2016 The type of data that should be passed to a function depends on what
2017 kind of information it uses. The arguments to a function such as
2018 @code{+} must have values that are numbers, since @code{+} adds numbers.
2019 Other functions use different kinds of data for their arguments.
2023 For example, the @code{concat} function links together or unites two or
2024 more strings of text to produce a string. The arguments are strings.
2025 Concatenating the two character strings @code{abc}, @code{def} produces
2026 the single string @code{abcdef}. This can be seen by evaluating the
2030 (concat "abc" "def")
2034 The value produced by evaluating this expression is @code{"abcdef"}.
2036 A function such as @code{substring} uses both a string and numbers as
2037 arguments. The function returns a part of the string, a substring of
2038 the first argument. This function takes three arguments. Its first
2039 argument is the string of characters, the second and third arguments are
2040 numbers that indicate the beginning and end of the substring. The
2041 numbers are a count of the number of characters (including spaces and
2042 punctuations) from the beginning of the string.
2045 For example, if you evaluate the following:
2048 (substring "The quick brown fox jumped." 16 19)
2052 you will see @code{"fox"} appear in the echo area. The arguments are the
2053 string and the two numbers.
2055 Note that the string passed to @code{substring} is a single atom even
2056 though it is made up of several words separated by spaces. Lisp counts
2057 everything between the two quotation marks as part of the string,
2058 including the spaces. You can think of the @code{substring} function as
2059 a kind of `atom smasher' since it takes an otherwise indivisible atom
2060 and extracts a part. However, @code{substring} is only able to extract
2061 a substring from an argument that is a string, not from another type of
2062 atom such as a number or symbol.
2064 @node Args as Variable or List, Variable Number of Arguments, Data types, Arguments
2065 @comment node-name, next, previous, up
2066 @subsection An Argument as the Value of a Variable or List
2068 An argument can be a symbol that returns a value when it is evaluated.
2069 For example, when the symbol @code{fill-column} by itself is evaluated,
2070 it returns a number. This number can be used in an addition.
2073 Position the cursor after the following expression and type @kbd{C-x
2081 The value will be a number two more than what you get by evaluating
2082 @code{fill-column} alone. For me, this is 74, because my value of
2083 @code{fill-column} is 72.
2085 As we have just seen, an argument can be a symbol that returns a value
2086 when evaluated. In addition, an argument can be a list that returns a
2087 value when it is evaluated. For example, in the following expression,
2088 the arguments to the function @code{concat} are the strings
2089 @w{@code{"The "}} and @w{@code{" red foxes."}} and the list
2090 @code{(number-to-string (+ 2 fill-column))}.
2092 @c For GNU Emacs 22, need number-to-string
2094 (concat "The " (number-to-string (+ 2 fill-column)) " red foxes.")
2098 If you evaluate this expression---and if, as with my Emacs,
2099 @code{fill-column} evaluates to 72---@code{"The 74 red foxes."} will
2100 appear in the echo area. (Note that you must put spaces after the
2101 word @samp{The} and before the word @samp{red} so they will appear in
2102 the final string. The function @code{number-to-string} converts the
2103 integer that the addition function returns to a string.
2104 @code{number-to-string} is also known as @code{int-to-string}.)
2106 @node Variable Number of Arguments, Wrong Type of Argument, Args as Variable or List, Arguments
2107 @comment node-name, next, previous, up
2108 @subsection Variable Number of Arguments
2109 @cindex Variable number of arguments
2110 @cindex Arguments, variable number of
2112 Some functions, such as @code{concat}, @code{+} or @code{*}, take any
2113 number of arguments. (The @code{*} is the symbol for multiplication.)
2114 This can be seen by evaluating each of the following expressions in
2115 the usual way. What you will see in the echo area is printed in this
2116 text after @samp{@result{}}, which you may read as `evaluates to'.
2119 In the first set, the functions have no arguments:
2130 In this set, the functions have one argument each:
2141 In this set, the functions have three arguments each:
2145 (+ 3 4 5) @result{} 12
2147 (* 3 4 5) @result{} 60
2151 @node Wrong Type of Argument, message, Variable Number of Arguments, Arguments
2152 @comment node-name, next, previous, up
2153 @subsection Using the Wrong Type Object as an Argument
2154 @cindex Wrong type of argument
2155 @cindex Argument, wrong type of
2157 When a function is passed an argument of the wrong type, the Lisp
2158 interpreter produces an error message. For example, the @code{+}
2159 function expects the values of its arguments to be numbers. As an
2160 experiment we can pass it the quoted symbol @code{hello} instead of a
2161 number. Position the cursor after the following expression and type
2169 When you do this you will generate an error message. What has happened
2170 is that @code{+} has tried to add the 2 to the value returned by
2171 @code{'hello}, but the value returned by @code{'hello} is the symbol
2172 @code{hello}, not a number. Only numbers can be added. So @code{+}
2173 could not carry out its addition.
2176 In GNU Emacs version 22, you will create and enter a
2177 @file{*Backtrace*} buffer that says:
2182 ---------- Buffer: *Backtrace* ----------
2183 Debugger entered--Lisp error:
2184 (wrong-type-argument number-or-marker-p hello)
2186 eval((+ 2 (quote hello)))
2187 eval-last-sexp-1(nil)
2189 call-interactively(eval-last-sexp)
2190 ---------- Buffer: *Backtrace* ----------
2195 As usual, the error message tries to be helpful and makes sense after you
2196 learn how to read it.@footnote{@code{(quote hello)} is an expansion of
2197 the abbreviation @code{'hello}.}
2199 The first part of the error message is straightforward; it says
2200 @samp{wrong type argument}. Next comes the mysterious jargon word
2201 @w{@samp{number-or-marker-p}}. This word is trying to tell you what
2202 kind of argument the @code{+} expected.
2204 The symbol @code{number-or-marker-p} says that the Lisp interpreter is
2205 trying to determine whether the information presented it (the value of
2206 the argument) is a number or a marker (a special object representing a
2207 buffer position). What it does is test to see whether the @code{+} is
2208 being given numbers to add. It also tests to see whether the
2209 argument is something called a marker, which is a specific feature of
2210 Emacs Lisp. (In Emacs, locations in a buffer are recorded as markers.
2211 When the mark is set with the @kbd{C-@@} or @kbd{C-@key{SPC}} command,
2212 its position is kept as a marker. The mark can be considered a
2213 number---the number of characters the location is from the beginning
2214 of the buffer.) In Emacs Lisp, @code{+} can be used to add the
2215 numeric value of marker positions as numbers.
2217 The @samp{p} of @code{number-or-marker-p} is the embodiment of a
2218 practice started in the early days of Lisp programming. The @samp{p}
2219 stands for `predicate'. In the jargon used by the early Lisp
2220 researchers, a predicate refers to a function to determine whether some
2221 property is true or false. So the @samp{p} tells us that
2222 @code{number-or-marker-p} is the name of a function that determines
2223 whether it is true or false that the argument supplied is a number or
2224 a marker. Other Lisp symbols that end in @samp{p} include @code{zerop},
2225 a function that tests whether its argument has the value of zero, and
2226 @code{listp}, a function that tests whether its argument is a list.
2228 Finally, the last part of the error message is the symbol @code{hello}.
2229 This is the value of the argument that was passed to @code{+}. If the
2230 addition had been passed the correct type of object, the value passed
2231 would have been a number, such as 37, rather than a symbol like
2232 @code{hello}. But then you would not have got the error message.
2236 In GNU Emacs version 20 and before, the echo area displays an error
2240 Wrong type argument:@: number-or-marker-p, hello
2243 This says, in different words, the same as the top line of the
2244 @file{*Backtrace*} buffer.
2247 @node message, , Wrong Type of Argument, Arguments
2248 @comment node-name, next, previous, up
2249 @subsection The @code{message} Function
2252 Like @code{+}, the @code{message} function takes a variable number of
2253 arguments. It is used to send messages to the user and is so useful
2254 that we will describe it here.
2257 A message is printed in the echo area. For example, you can print a
2258 message in your echo area by evaluating the following list:
2261 (message "This message appears in the echo area!")
2264 The whole string between double quotation marks is a single argument
2265 and is printed @i{in toto}. (Note that in this example, the message
2266 itself will appear in the echo area within double quotes; that is
2267 because you see the value returned by the @code{message} function. In
2268 most uses of @code{message} in programs that you write, the text will
2269 be printed in the echo area as a side-effect, without the quotes.
2270 @xref{multiply-by-seven in detail, , @code{multiply-by-seven} in
2271 detail}, for an example of this.)
2273 However, if there is a @samp{%s} in the quoted string of characters, the
2274 @code{message} function does not print the @samp{%s} as such, but looks
2275 to the argument that follows the string. It evaluates the second
2276 argument and prints the value at the location in the string where the
2280 You can see this by positioning the cursor after the following
2281 expression and typing @kbd{C-x C-e}:
2284 (message "The name of this buffer is: %s." (buffer-name))
2288 In Info, @code{"The name of this buffer is: *info*."} will appear in the
2289 echo area. The function @code{buffer-name} returns the name of the
2290 buffer as a string, which the @code{message} function inserts in place
2293 To print a value as an integer, use @samp{%d} in the same way as
2294 @samp{%s}. For example, to print a message in the echo area that
2295 states the value of the @code{fill-column}, evaluate the following:
2298 (message "The value of fill-column is %d." fill-column)
2302 On my system, when I evaluate this list, @code{"The value of
2303 fill-column is 72."} appears in my echo area@footnote{Actually, you
2304 can use @code{%s} to print a number. It is non-specific. @code{%d}
2305 prints only the part of a number left of a decimal point, and not
2306 anything that is not a number.}.
2308 If there is more than one @samp{%s} in the quoted string, the value of
2309 the first argument following the quoted string is printed at the
2310 location of the first @samp{%s} and the value of the second argument is
2311 printed at the location of the second @samp{%s}, and so on.
2314 For example, if you evaluate the following,
2318 (message "There are %d %s in the office!"
2319 (- fill-column 14) "pink elephants")
2324 a rather whimsical message will appear in your echo area. On my system
2325 it says, @code{"There are 58 pink elephants in the office!"}.
2327 The expression @code{(- fill-column 14)} is evaluated and the resulting
2328 number is inserted in place of the @samp{%d}; and the string in double
2329 quotes, @code{"pink elephants"}, is treated as a single argument and
2330 inserted in place of the @samp{%s}. (That is to say, a string between
2331 double quotes evaluates to itself, like a number.)
2333 Finally, here is a somewhat complex example that not only illustrates
2334 the computation of a number, but also shows how you can use an
2335 expression within an expression to generate the text that is substituted
2340 (message "He saw %d %s"
2344 "The quick brown foxes jumped." 16 21)
2349 In this example, @code{message} has three arguments: the string,
2350 @code{"He saw %d %s"}, the expression, @code{(- fill-column 32)}, and
2351 the expression beginning with the function @code{concat}. The value
2352 resulting from the evaluation of @code{(- fill-column 32)} is inserted
2353 in place of the @samp{%d}; and the value returned by the expression
2354 beginning with @code{concat} is inserted in place of the @samp{%s}.
2356 When your fill column is 70 and you evaluate the expression, the
2357 message @code{"He saw 38 red foxes leaping."} appears in your echo
2360 @node set & setq, Summary, Arguments, List Processing
2361 @comment node-name, next, previous, up
2362 @section Setting the Value of a Variable
2363 @cindex Variable, setting value
2364 @cindex Setting value of variable
2366 @cindex @samp{bind} defined
2367 There are several ways by which a variable can be given a value. One of
2368 the ways is to use either the function @code{set} or the function
2369 @code{setq}. Another way is to use @code{let} (@pxref{let}). (The
2370 jargon for this process is to @dfn{bind} a variable to a value.)
2372 The following sections not only describe how @code{set} and @code{setq}
2373 work but also illustrate how arguments are passed.
2376 * Using set:: Setting values.
2377 * Using setq:: Setting a quoted value.
2378 * Counting:: Using @code{setq} to count.
2381 @node Using set, Using setq, set & setq, set & setq
2382 @comment node-name, next, previous, up
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.)
2438 @node Using setq, Counting, Using set, set & setq
2439 @comment node-name, next, previous, up
2440 @subsection Using @code{setq}
2443 As a practical matter, you almost always quote the first argument to
2444 @code{set}. The combination of @code{set} and a quoted first argument
2445 is so common that it has its own name: the special form @code{setq}.
2446 This special form is just like @code{set} except that the first argument
2447 is quoted automatically, so you don't need to type the quote mark
2448 yourself. Also, as an added convenience, @code{setq} permits you to set
2449 several different variables to different values, all in one expression.
2451 To set the value of the variable @code{carnivores} to the list
2452 @code{'(lion tiger leopard)} using @code{setq}, the following expression
2456 (setq carnivores '(lion tiger leopard))
2460 This is exactly the same as using @code{set} except the first argument
2461 is automatically quoted by @code{setq}. (The @samp{q} in @code{setq}
2462 means @code{quote}.)
2465 With @code{set}, the expression would look like this:
2468 (set 'carnivores '(lion tiger leopard))
2471 Also, @code{setq} can be used to assign different values to
2472 different variables. The first argument is bound to the value
2473 of the second argument, the third argument is bound to the value of the
2474 fourth argument, and so on. For example, you could use the following to
2475 assign a list of trees to the symbol @code{trees} and a list of herbivores
2476 to the symbol @code{herbivores}:
2480 (setq trees '(pine fir oak maple)
2481 herbivores '(gazelle antelope zebra))
2486 (The expression could just as well have been on one line, but it might
2487 not have fit on a page; and humans find it easier to read nicely
2490 Although I have been using the term `assign', there is another way of
2491 thinking about the workings of @code{set} and @code{setq}; and that is to
2492 say that @code{set} and @code{setq} make the symbol @emph{point} to the
2493 list. This latter way of thinking is very common and in forthcoming
2494 chapters we shall come upon at least one symbol that has `pointer' as
2495 part of its name. The name is chosen because the symbol has a value,
2496 specifically a list, attached to it; or, expressed another way,
2497 the symbol is set to ``point'' to the list.
2499 @node Counting, , Using setq, set & setq
2500 @comment node-name, next, previous, up
2501 @subsection Counting
2504 Here is an example that shows how to use @code{setq} in a counter. You
2505 might use this to count how many times a part of your program repeats
2506 itself. First set a variable to zero; then add one to the number each
2507 time the program repeats itself. To do this, you need a variable that
2508 serves as a counter, and two expressions: an initial @code{setq}
2509 expression that sets the counter variable to zero; and a second
2510 @code{setq} expression that increments the counter each time it is
2515 (setq counter 0) ; @r{Let's call this the initializer.}
2517 (setq counter (+ counter 1)) ; @r{This is the incrementer.}
2519 counter ; @r{This is the counter.}
2524 (The text following the @samp{;} are comments. @xref{Change a
2525 defun, , Change a Function Definition}.)
2527 If you evaluate the first of these expressions, the initializer,
2528 @code{(setq counter 0)}, and then evaluate the third expression,
2529 @code{counter}, the number @code{0} will appear in the echo area. If
2530 you then evaluate the second expression, the incrementer, @code{(setq
2531 counter (+ counter 1))}, the counter will get the value 1. So if you
2532 again evaluate @code{counter}, the number @code{1} will appear in the
2533 echo area. Each time you evaluate the second expression, the value of
2534 the counter will be incremented.
2536 When you evaluate the incrementer, @code{(setq counter (+ counter 1))},
2537 the Lisp interpreter first evaluates the innermost list; this is the
2538 addition. In order to evaluate this list, it must evaluate the variable
2539 @code{counter} and the number @code{1}. When it evaluates the variable
2540 @code{counter}, it receives its current value. It passes this value and
2541 the number @code{1} to the @code{+} which adds them together. The sum
2542 is then returned as the value of the inner list and passed to the
2543 @code{setq} which sets the variable @code{counter} to this new value.
2544 Thus, the value of the variable, @code{counter}, is changed.
2546 @node Summary, Error Message Exercises, set & setq, List Processing
2547 @comment node-name, next, previous, up
2550 Learning Lisp is like climbing a hill in which the first part is the
2551 steepest. You have now climbed the most difficult part; what remains
2552 becomes easier as you progress onwards.
2560 Lisp programs are made up of expressions, which are lists or single atoms.
2563 Lists are made up of zero or more atoms or inner lists, separated by whitespace and
2564 surrounded by parentheses. A list can be empty.
2567 Atoms are multi-character symbols, like @code{forward-paragraph}, single
2568 character symbols like @code{+}, strings of characters between double
2569 quotation marks, or numbers.
2572 A number evaluates to itself.
2575 A string between double quotes also evaluates to itself.
2578 When you evaluate a symbol by itself, its value is returned.
2581 When you evaluate a list, the Lisp interpreter looks at the first symbol
2582 in the list and then at the function definition bound to that symbol.
2583 Then the instructions in the function definition are carried out.
2586 A single quotation mark,
2593 , tells the Lisp interpreter that it should
2594 return the following expression as written, and not evaluate it as it
2595 would if the quote were not there.
2598 Arguments are the information passed to a function. The arguments to a
2599 function are computed by evaluating the rest of the elements of the list
2600 of which the function is the first element.
2603 A function always returns a value when it is evaluated (unless it gets
2604 an error); in addition, it may also carry out some action called a
2605 ``side effect''. In many cases, a function's primary purpose is to
2606 create a side effect.
2609 @node Error Message Exercises, , Summary, List Processing
2610 @comment node-name, next, previous, up
2613 A few simple exercises:
2617 Generate an error message by evaluating an appropriate symbol that is
2618 not within parentheses.
2621 Generate an error message by evaluating an appropriate symbol that is
2622 between parentheses.
2625 Create a counter that increments by two rather than one.
2628 Write an expression that prints a message in the echo area when
2632 @node Practicing Evaluation, Writing Defuns, List Processing, Top
2633 @comment node-name, next, previous, up
2634 @chapter Practicing Evaluation
2635 @cindex Practicing evaluation
2636 @cindex Evaluation practice
2638 Before learning how to write a function definition in Emacs Lisp, it is
2639 useful to spend a little time evaluating various expressions that have
2640 already been written. These expressions will be lists with the
2641 functions as their first (and often only) element. Since some of the
2642 functions associated with buffers are both simple and interesting, we
2643 will start with those. In this section, we will evaluate a few of
2644 these. In another section, we will study the code of several other
2645 buffer-related functions, to see how they were written.
2648 * How to Evaluate:: Typing editing commands or @kbd{C-x C-e}
2650 * Buffer Names:: Buffers and files are different.
2651 * Getting Buffers:: Getting a buffer itself, not merely its name.
2652 * Switching Buffers:: How to change to another buffer.
2653 * Buffer Size & Locations:: Where point is located and the size of
2655 * Evaluation Exercise::
2658 @node How to Evaluate, Buffer Names, Practicing Evaluation, Practicing Evaluation
2660 @unnumberedsec How to Evaluate
2663 @i{Whenever you give an editing command} to Emacs Lisp, such as the
2664 command to move the cursor or to scroll the screen, @i{you are evaluating
2665 an expression,} the first element of which is a function. @i{This is
2668 @cindex @samp{interactive function} defined
2669 @cindex @samp{command} defined
2670 When you type keys, you cause the Lisp interpreter to evaluate an
2671 expression and that is how you get your results. Even typing plain text
2672 involves evaluating an Emacs Lisp function, in this case, one that uses
2673 @code{self-insert-command}, which simply inserts the character you
2674 typed. The functions you evaluate by typing keystrokes are called
2675 @dfn{interactive} functions, or @dfn{commands}; how you make a function
2676 interactive will be illustrated in the chapter on how to write function
2677 definitions. @xref{Interactive, , Making a Function Interactive}.
2679 In addition to typing keyboard commands, we have seen a second way to
2680 evaluate an expression: by positioning the cursor after a list and
2681 typing @kbd{C-x C-e}. This is what we will do in the rest of this
2682 section. There are other ways to evaluate an expression as well; these
2683 will be described as we come to them.
2685 Besides being used for practicing evaluation, the functions shown in the
2686 next few sections are important in their own right. A study of these
2687 functions makes clear the distinction between buffers and files, how to
2688 switch to a buffer, and how to determine a location within it.
2690 @node Buffer Names, Getting Buffers, How to Evaluate, Practicing Evaluation
2691 @comment node-name, next, previous, up
2692 @section Buffer Names
2694 @findex buffer-file-name
2696 The two functions, @code{buffer-name} and @code{buffer-file-name}, show
2697 the difference between a file and a buffer. When you evaluate the
2698 following expression, @code{(buffer-name)}, the name of the buffer
2699 appears in the echo area. When you evaluate @code{(buffer-file-name)},
2700 the name of the file to which the buffer refers appears in the echo
2701 area. Usually, the name returned by @code{(buffer-name)} is the same as
2702 the name of the file to which it refers, and the name returned by
2703 @code{(buffer-file-name)} is the full path-name of the file.
2705 A file and a buffer are two different entities. A file is information
2706 recorded permanently in the computer (unless you delete it). A buffer,
2707 on the other hand, is information inside of Emacs that will vanish at
2708 the end of the editing session (or when you kill the buffer). Usually,
2709 a buffer contains information that you have copied from a file; we say
2710 the buffer is @dfn{visiting} that file. This copy is what you work on
2711 and modify. Changes to the buffer do not change the file, until you
2712 save the buffer. When you save the buffer, the buffer is copied to the file
2713 and is thus saved permanently.
2716 If you are reading this in Info inside of GNU Emacs, you can evaluate
2717 each of the following expressions by positioning the cursor after it and
2718 typing @kbd{C-x C-e}.
2729 When I do this in Info, the value returned by evaluating
2730 @code{(buffer-name)} is @file{"*info*"}, and the value returned by
2731 evaluating @code{(buffer-file-name)} is @file{nil}.
2733 On the other hand, while I am writing this document, the value
2734 returned by evaluating @code{(buffer-name)} is
2735 @file{"introduction.texinfo"}, and the value returned by evaluating
2736 @code{(buffer-file-name)} is
2737 @file{"/gnu/work/intro/introduction.texinfo"}.
2739 @cindex @code{nil}, history of word
2740 The former is the name of the buffer and the latter is the name of the
2741 file. In Info, the buffer name is @file{"*info*"}. Info does not
2742 point to any file, so the result of evaluating
2743 @code{(buffer-file-name)} is @file{nil}. The symbol @code{nil} is
2744 from the Latin word for `nothing'; in this case, it means that the
2745 buffer is not associated with any file. (In Lisp, @code{nil} is also
2746 used to mean `false' and is a synonym for the empty list, @code{()}.)
2748 When I am writing, the name of my buffer is
2749 @file{"introduction.texinfo"}. The name of the file to which it
2750 points is @file{"/gnu/work/intro/introduction.texinfo"}.
2752 (In the expressions, the parentheses tell the Lisp interpreter to
2753 treat @w{@code{buffer-name}} and @w{@code{buffer-file-name}} as
2754 functions; without the parentheses, the interpreter would attempt to
2755 evaluate the symbols as variables. @xref{Variables}.)
2757 In spite of the distinction between files and buffers, you will often
2758 find that people refer to a file when they mean a buffer and vice-verse.
2759 Indeed, most people say, ``I am editing a file,'' rather than saying,
2760 ``I am editing a buffer which I will soon save to a file.'' It is
2761 almost always clear from context what people mean. When dealing with
2762 computer programs, however, it is important to keep the distinction in mind,
2763 since the computer is not as smart as a person.
2765 @cindex Buffer, history of word
2766 The word `buffer', by the way, comes from the meaning of the word as a
2767 cushion that deadens the force of a collision. In early computers, a
2768 buffer cushioned the interaction between files and the computer's
2769 central processing unit. The drums or tapes that held a file and the
2770 central processing unit were pieces of equipment that were very
2771 different from each other, working at their own speeds, in spurts. The
2772 buffer made it possible for them to work together effectively.
2773 Eventually, the buffer grew from being an intermediary, a temporary
2774 holding place, to being the place where work is done. This
2775 transformation is rather like that of a small seaport that grew into a
2776 great city: once it was merely the place where cargo was warehoused
2777 temporarily before being loaded onto ships; then it became a business
2778 and cultural center in its own right.
2780 Not all buffers are associated with files. For example, a
2781 @file{*scratch*} buffer does not visit any file. Similarly, a
2782 @file{*Help*} buffer is not associated with any file.
2784 In the old days, when you lacked a @file{~/.emacs} file and started an
2785 Emacs session by typing the command @code{emacs} alone, without naming
2786 any files, Emacs started with the @file{*scratch*} buffer visible.
2787 Nowadays, you will see a splash screen. You can follow one of the
2788 commands suggested on the splash screen, visit a file, or press the
2789 spacebar to reach the @file{*scratch*} buffer.
2791 If you switch to the @file{*scratch*} buffer, type
2792 @code{(buffer-name)}, position the cursor after it, and then type
2793 @kbd{C-x C-e} to evaluate the expression. The name @code{"*scratch*"}
2794 will be returned and will appear in the echo area. @code{"*scratch*"}
2795 is the name of the buffer. When you type @code{(buffer-file-name)} in
2796 the @file{*scratch*} buffer and evaluate that, @code{nil} will appear
2797 in the echo area, just as it does when you evaluate
2798 @code{(buffer-file-name)} in Info.
2800 Incidentally, if you are in the @file{*scratch*} buffer and want the
2801 value returned by an expression to appear in the @file{*scratch*}
2802 buffer itself rather than in the echo area, type @kbd{C-u C-x C-e}
2803 instead of @kbd{C-x C-e}. This causes the value returned to appear
2804 after the expression. The buffer will look like this:
2807 (buffer-name)"*scratch*"
2811 You cannot do this in Info since Info is read-only and it will not allow
2812 you to change the contents of the buffer. But you can do this in any
2813 buffer you can edit; and when you write code or documentation (such as
2814 this book), this feature is very useful.
2816 @node Getting Buffers, Switching Buffers, Buffer Names, Practicing Evaluation
2817 @comment node-name, next, previous, up
2818 @section Getting Buffers
2819 @findex current-buffer
2820 @findex other-buffer
2821 @cindex Getting a buffer
2823 The @code{buffer-name} function returns the @emph{name} of the buffer;
2824 to get the buffer @emph{itself}, a different function is needed: the
2825 @code{current-buffer} function. If you use this function in code, what
2826 you get is the buffer itself.
2828 A name and the object or entity to which the name refers are different
2829 from each other. You are not your name. You are a person to whom
2830 others refer by name. If you ask to speak to George and someone hands you
2831 a card with the letters @samp{G}, @samp{e}, @samp{o}, @samp{r},
2832 @samp{g}, and @samp{e} written on it, you might be amused, but you would
2833 not be satisfied. You do not want to speak to the name, but to the
2834 person to whom the name refers. A buffer is similar: the name of the
2835 scratch buffer is @file{*scratch*}, but the name is not the buffer. To
2836 get a buffer itself, you need to use a function such as
2837 @code{current-buffer}.
2839 However, there is a slight complication: if you evaluate
2840 @code{current-buffer} in an expression on its own, as we will do here,
2841 what you see is a printed representation of the name of the buffer
2842 without the contents of the buffer. Emacs works this way for two
2843 reasons: the buffer may be thousands of lines long---too long to be
2844 conveniently displayed; and, another buffer may have the same contents
2845 but a different name, and it is important to distinguish between them.
2848 Here is an expression containing the function:
2855 If you evaluate this expression in Info in Emacs in the usual way,
2856 @file{#<buffer *info*>} will appear in the echo area. The special
2857 format indicates that the buffer itself is being returned, rather than
2860 Incidentally, while you can type a number or symbol into a program, you
2861 cannot do that with the printed representation of a buffer: the only way
2862 to get a buffer itself is with a function such as @code{current-buffer}.
2864 A related function is @code{other-buffer}. This returns the most
2865 recently selected buffer other than the one you are in currently, not
2866 a printed representation of its name. If you have recently switched
2867 back and forth from the @file{*scratch*} buffer, @code{other-buffer}
2868 will return that buffer.
2871 You can see this by evaluating the expression:
2878 You should see @file{#<buffer *scratch*>} appear in the echo area, or
2879 the name of whatever other buffer you switched back from most
2880 recently@footnote{Actually, by default, if the buffer from which you
2881 just switched is visible to you in another window, @code{other-buffer}
2882 will choose the most recent buffer that you cannot see; this is a
2883 subtlety that I often forget.}.
2885 @node Switching Buffers, Buffer Size & Locations, Getting Buffers, Practicing Evaluation
2886 @comment node-name, next, previous, up
2887 @section Switching Buffers
2888 @findex switch-to-buffer
2890 @cindex Switching to a buffer
2892 The @code{other-buffer} function actually provides a buffer when it is
2893 used as an argument to a function that requires one. We can see this
2894 by using @code{other-buffer} and @code{switch-to-buffer} to switch to a
2897 But first, a brief introduction to the @code{switch-to-buffer}
2898 function. When you switched back and forth from Info to the
2899 @file{*scratch*} buffer to evaluate @code{(buffer-name)}, you most
2900 likely typed @kbd{C-x b} and then typed @file{*scratch*}@footnote{Or
2901 rather, to save typing, you probably only typed @kbd{RET} if the
2902 default buffer was @file{*scratch*}, or if it was different, then you
2903 typed just part of the name, such as @code{*sc}, pressed your
2904 @kbd{TAB} key to cause it to expand to the full name, and then typed
2905 your @kbd{RET} key.} when prompted in the minibuffer for the name of
2906 the buffer to which you wanted to switch. The keystrokes, @kbd{C-x
2907 b}, cause the Lisp interpreter to evaluate the interactive function
2908 @code{switch-to-buffer}. As we said before, this is how Emacs works:
2909 different keystrokes call or run different functions. For example,
2910 @kbd{C-f} calls @code{forward-char}, @kbd{M-e} calls
2911 @code{forward-sentence}, and so on.
2913 By writing @code{switch-to-buffer} in an expression, and giving it a
2914 buffer to switch to, we can switch buffers just the way @kbd{C-x b}
2918 Here is the Lisp expression:
2921 (switch-to-buffer (other-buffer))
2925 The symbol @code{switch-to-buffer} is the first element of the list,
2926 so the Lisp interpreter will treat it as a function and carry out the
2927 instructions that are attached to it. But before doing that, the
2928 interpreter will note that @code{other-buffer} is inside parentheses
2929 and work on that symbol first. @code{other-buffer} is the first (and
2930 in this case, the only) element of this list, so the Lisp interpreter
2931 calls or runs the function. It returns another buffer. Next, the
2932 interpreter runs @code{switch-to-buffer}, passing to it, as an
2933 argument, the other buffer, which is what Emacs will switch to. If
2934 you are reading this in Info, try this now. Evaluate the expression.
2935 (To get back, type @kbd{C-x b @key{RET}}.)@footnote{Remember, this
2936 expression will move you to your most recent other buffer that you
2937 cannot see. If you really want to go to your most recently selected
2938 buffer, even if you can still see it, you need to evaluate the
2939 following more complex expression:
2942 (switch-to-buffer (other-buffer (current-buffer) t))
2946 In this case, the first argument to @code{other-buffer} tells it which
2947 buffer to skip---the current one---and the second argument tells
2948 @code{other-buffer} it is OK to switch to a visible buffer.
2949 In regular use, @code{switch-to-buffer} takes you to an invisible
2950 window since you would most likely use @kbd{C-x o} (@code{other-window})
2951 to go to another visible buffer.}
2953 In the programming examples in later sections of this document, you will
2954 see the function @code{set-buffer} more often than
2955 @code{switch-to-buffer}. This is because of a difference between
2956 computer programs and humans: humans have eyes and expect to see the
2957 buffer on which they are working on their computer terminals. This is
2958 so obvious, it almost goes without saying. However, programs do not
2959 have eyes. When a computer program works on a buffer, that buffer does
2960 not need to be visible on the screen.
2962 @code{switch-to-buffer} is designed for humans and does two different
2963 things: it switches the buffer to which Emacs' attention is directed; and
2964 it switches the buffer displayed in the window to the new buffer.
2965 @code{set-buffer}, on the other hand, does only one thing: it switches
2966 the attention of the computer program to a different buffer. The buffer
2967 on the screen remains unchanged (of course, normally nothing happens
2968 there until the command finishes running).
2970 @cindex @samp{call} defined
2971 Also, we have just introduced another jargon term, the word @dfn{call}.
2972 When you evaluate a list in which the first symbol is a function, you
2973 are calling that function. The use of the term comes from the notion of
2974 the function as an entity that can do something for you if you `call'
2975 it---just as a plumber is an entity who can fix a leak if you call him
2978 @node Buffer Size & Locations, Evaluation Exercise, Switching Buffers, Practicing Evaluation
2979 @comment node-name, next, previous, up
2980 @section Buffer Size and the Location of Point
2981 @cindex Size of buffer
2983 @cindex Point location
2984 @cindex Location of point
2986 Finally, let's look at several rather simple functions,
2987 @code{buffer-size}, @code{point}, @code{point-min}, and
2988 @code{point-max}. These give information about the size of a buffer and
2989 the location of point within it.
2991 The function @code{buffer-size} tells you the size of the current
2992 buffer; that is, the function returns a count of the number of
2993 characters in the buffer.
3000 You can evaluate this in the usual way, by positioning the
3001 cursor after the expression and typing @kbd{C-x C-e}.
3003 @cindex @samp{point} defined
3004 In Emacs, the current position of the cursor is called @dfn{point}.
3005 The expression @code{(point)} returns a number that tells you where the
3006 cursor is located as a count of the number of characters from the
3007 beginning of the buffer up to point.
3010 You can see the character count for point in this buffer by evaluating
3011 the following expression in the usual way:
3018 As I write this, the value of @code{point} is 65724. The @code{point}
3019 function is frequently used in some of the examples later in this
3023 The value of point depends, of course, on its location within the
3024 buffer. If you evaluate point in this spot, the number will be larger:
3031 For me, the value of point in this location is 66043, which means that
3032 there are 319 characters (including spaces) between the two
3033 expressions. (Doubtless, you will see different numbers, since I will
3034 have edited this since I first evaluated point.)
3036 @cindex @samp{narrowing} defined
3037 The function @code{point-min} is somewhat similar to @code{point}, but
3038 it returns the value of the minimum permissible value of point in the
3039 current buffer. This is the number 1 unless @dfn{narrowing} is in
3040 effect. (Narrowing is a mechanism whereby you can restrict yourself,
3041 or a program, to operations on just a part of a buffer.
3042 @xref{Narrowing & Widening, , Narrowing and Widening}.) Likewise, the
3043 function @code{point-max} returns the value of the maximum permissible
3044 value of point in the current buffer.
3046 @node Evaluation Exercise, , Buffer Size & Locations, Practicing Evaluation
3049 Find a file with which you are working and move towards its middle.
3050 Find its buffer name, file name, length, and your position in the file.
3052 @node Writing Defuns, Buffer Walk Through, Practicing Evaluation, Top
3053 @comment node-name, next, previous, up
3054 @chapter How To Write Function Definitions
3055 @cindex Definition writing
3056 @cindex Function definition writing
3057 @cindex Writing a function definition
3059 When the Lisp interpreter evaluates a list, it looks to see whether the
3060 first symbol on the list has a function definition attached to it; or,
3061 put another way, whether the symbol points to a function definition. If
3062 it does, the computer carries out the instructions in the definition. A
3063 symbol that has a function definition is called, simply, a function
3064 (although, properly speaking, the definition is the function and the
3065 symbol refers to it.)
3068 * Primitive Functions::
3069 * defun:: The @code{defun} special form.
3070 * Install:: Install a function definition.
3071 * Interactive:: Making a function interactive.
3072 * Interactive Options:: Different options for @code{interactive}.
3073 * Permanent Installation:: Installing code permanently.
3074 * let:: Creating and initializing local variables.
3076 * else:: If--then--else expressions.
3077 * Truth & Falsehood:: What Lisp considers false and true.
3078 * save-excursion:: Keeping track of point, mark, and buffer.
3083 @node Primitive Functions, defun, Writing Defuns, Writing Defuns
3085 @unnumberedsec An Aside about Primitive Functions
3087 @cindex Primitive functions
3088 @cindex Functions, primitive
3090 @cindex C language primitives
3091 @cindex Primitives written in C
3092 All functions are defined in terms of other functions, except for a few
3093 @dfn{primitive} functions that are written in the C programming
3094 language. When you write functions' definitions, you will write them in
3095 Emacs Lisp and use other functions as your building blocks. Some of the
3096 functions you will use will themselves be written in Emacs Lisp (perhaps
3097 by you) and some will be primitives written in C. The primitive
3098 functions are used exactly like those written in Emacs Lisp and behave
3099 like them. They are written in C so we can easily run GNU Emacs on any
3100 computer that has sufficient power and can run C.
3102 Let me re-emphasize this: when you write code in Emacs Lisp, you do not
3103 distinguish between the use of functions written in C and the use of
3104 functions written in Emacs Lisp. The difference is irrelevant. I
3105 mention the distinction only because it is interesting to know. Indeed,
3106 unless you investigate, you won't know whether an already-written
3107 function is written in Emacs Lisp or C.
3109 @node defun, Install, Primitive Functions, Writing Defuns
3110 @comment node-name, next, previous, up
3111 @section The @code{defun} Special Form
3113 @cindex Special form of @code{defun}
3115 @cindex @samp{function definition} defined
3116 In Lisp, a symbol such as @code{mark-whole-buffer} has code attached to
3117 it that tells the computer what to do when the function is called.
3118 This code is called the @dfn{function definition} and is created by
3119 evaluating a Lisp expression that starts with the symbol @code{defun}
3120 (which is an abbreviation for @emph{define function}). Because
3121 @code{defun} does not evaluate its arguments in the usual way, it is
3122 called a @dfn{special form}.
3124 In subsequent sections, we will look at function definitions from the
3125 Emacs source code, such as @code{mark-whole-buffer}. In this section,
3126 we will describe a simple function definition so you can see how it
3127 looks. This function definition uses arithmetic because it makes for a
3128 simple example. Some people dislike examples using arithmetic; however,
3129 if you are such a person, do not despair. Hardly any of the code we
3130 will study in the remainder of this introduction involves arithmetic or
3131 mathematics. The examples mostly involve text in one way or another.
3133 A function definition has up to five parts following the word
3138 The name of the symbol to which the function definition should be
3142 A list of the arguments that will be passed to the function. If no
3143 arguments will be passed to the function, this is an empty list,
3147 Documentation describing the function. (Technically optional, but
3148 strongly recommended.)
3151 Optionally, an expression to make the function interactive so you can
3152 use it by typing @kbd{M-x} and then the name of the function; or by
3153 typing an appropriate key or keychord.
3155 @cindex @samp{body} defined
3157 The code that instructs the computer what to do: the @dfn{body} of the
3158 function definition.
3161 It is helpful to think of the five parts of a function definition as
3162 being organized in a template, with slots for each part:
3166 (defun @var{function-name} (@var{arguments}@dots{})
3167 "@var{optional-documentation}@dots{}"
3168 (interactive @var{argument-passing-info}) ; @r{optional}
3173 As an example, here is the code for a function that multiplies its
3174 argument by 7. (This example is not interactive. @xref{Interactive,
3175 , Making a Function Interactive}, for that information.)
3179 (defun multiply-by-seven (number)
3180 "Multiply NUMBER by seven."
3185 This definition begins with a parenthesis and the symbol @code{defun},
3186 followed by the name of the function.
3188 @cindex @samp{argument list} defined
3189 The name of the function is followed by a list that contains the
3190 arguments that will be passed to the function. This list is called
3191 the @dfn{argument list}. In this example, the list has only one
3192 element, the symbol, @code{number}. When the function is used, the
3193 symbol will be bound to the value that is used as the argument to the
3196 Instead of choosing the word @code{number} for the name of the argument,
3197 I could have picked any other name. For example, I could have chosen
3198 the word @code{multiplicand}. I picked the word `number' because it
3199 tells what kind of value is intended for this slot; but I could just as
3200 well have chosen the word `multiplicand' to indicate the role that the
3201 value placed in this slot will play in the workings of the function. I
3202 could have called it @code{foogle}, but that would have been a bad
3203 choice because it would not tell humans what it means. The choice of
3204 name is up to the programmer and should be chosen to make the meaning of
3207 Indeed, you can choose any name you wish for a symbol in an argument
3208 list, even the name of a symbol used in some other function: the name
3209 you use in an argument list is private to that particular definition.
3210 In that definition, the name refers to a different entity than any use
3211 of the same name outside the function definition. Suppose you have a
3212 nick-name `Shorty' in your family; when your family members refer to
3213 `Shorty', they mean you. But outside your family, in a movie, for
3214 example, the name `Shorty' refers to someone else. Because a name in an
3215 argument list is private to the function definition, you can change the
3216 value of such a symbol inside the body of a function without changing
3217 its value outside the function. The effect is similar to that produced
3218 by a @code{let} expression. (@xref{let, , @code{let}}.)
3221 Note also that we discuss the word `number' in two different ways: as a
3222 symbol that appears in the code, and as the name of something that will
3223 be replaced by a something else during the evaluation of the function.
3224 In the first case, @code{number} is a symbol, not a number; it happens
3225 that within the function, it is a variable who value is the number in
3226 question, but our primary interest in it is as a symbol. On the other
3227 hand, when we are talking about the function, our interest is that we
3228 will substitute a number for the word @var{number}. To keep this
3229 distinction clear, we use different typography for the two
3230 circumstances. When we talk about this function, or about how it works,
3231 we refer to this number by writing @var{number}. In the function
3232 itself, we refer to it by writing @code{number}.
3235 The argument list is followed by the documentation string that
3236 describes the function. This is what you see when you type
3237 @w{@kbd{C-h f}} and the name of a function. Incidentally, when you
3238 write a documentation string like this, you should make the first line
3239 a complete sentence since some commands, such as @code{apropos}, print
3240 only the first line of a multi-line documentation string. Also, you
3241 should not indent the second line of a documentation string, if you
3242 have one, because that looks odd when you use @kbd{C-h f}
3243 (@code{describe-function}). The documentation string is optional, but
3244 it is so useful, it should be included in almost every function you
3247 @findex * @r{(multiplication)}
3248 The third line of the example consists of the body of the function
3249 definition. (Most functions' definitions, of course, are longer than
3250 this.) In this function, the body is the list, @code{(* 7 number)}, which
3251 says to multiply the value of @var{number} by 7. (In Emacs Lisp,
3252 @code{*} is the function for multiplication, just as @code{+} is the
3253 function for addition.)
3255 When you use the @code{multiply-by-seven} function, the argument
3256 @code{number} evaluates to the actual number you want used. Here is an
3257 example that shows how @code{multiply-by-seven} is used; but don't try
3258 to evaluate this yet!
3261 (multiply-by-seven 3)
3265 The symbol @code{number}, specified in the function definition in the
3266 next section, is given or ``bound to'' the value 3 in the actual use of
3267 the function. Note that although @code{number} was inside parentheses
3268 in the function definition, the argument passed to the
3269 @code{multiply-by-seven} function is not in parentheses. The
3270 parentheses are written in the function definition so the computer can
3271 figure out where the argument list ends and the rest of the function
3274 If you evaluate this example, you are likely to get an error message.
3275 (Go ahead, try it!) This is because we have written the function
3276 definition, but not yet told the computer about the definition---we have
3277 not yet installed (or `loaded') the function definition in Emacs.
3278 Installing a function is the process that tells the Lisp interpreter the
3279 definition of the function. Installation is described in the next
3282 @node Install, Interactive, defun, Writing Defuns
3283 @comment node-name, next, previous, up
3284 @section Install a Function Definition
3285 @cindex Install a Function Definition
3286 @cindex Definition installation
3287 @cindex Function definition installation
3289 If you are reading this inside of Info in Emacs, you can try out the
3290 @code{multiply-by-seven} function by first evaluating the function
3291 definition and then evaluating @code{(multiply-by-seven 3)}. A copy of
3292 the function definition follows. Place the cursor after the last
3293 parenthesis of the function definition and type @kbd{C-x C-e}. When you
3294 do this, @code{multiply-by-seven} will appear in the echo area. (What
3295 this means is that when a function definition is evaluated, the value it
3296 returns is the name of the defined function.) At the same time, this
3297 action installs the function definition.
3301 (defun multiply-by-seven (number)
3302 "Multiply NUMBER by seven."
3308 By evaluating this @code{defun}, you have just installed
3309 @code{multiply-by-seven} in Emacs. The function is now just as much a
3310 part of Emacs as @code{forward-word} or any other editing function you
3311 use. (@code{multiply-by-seven} will stay installed until you quit
3312 Emacs. To reload code automatically whenever you start Emacs, see
3313 @ref{Permanent Installation, , Installing Code Permanently}.)
3316 * Effect of installation::
3317 * Change a defun:: How to change a function definition.
3320 @node Effect of installation, Change a defun, Install, Install
3322 @unnumberedsubsec The effect of installation
3325 You can see the effect of installing @code{multiply-by-seven} by
3326 evaluating the following sample. Place the cursor after the following
3327 expression and type @kbd{C-x C-e}. The number 21 will appear in the
3331 (multiply-by-seven 3)
3334 If you wish, you can read the documentation for the function by typing
3335 @kbd{C-h f} (@code{describe-function}) and then the name of the
3336 function, @code{multiply-by-seven}. When you do this, a
3337 @file{*Help*} window will appear on your screen that says:
3341 multiply-by-seven is a Lisp function.
3342 (multiply-by-seven NUMBER)
3344 Multiply NUMBER by seven.
3349 (To return to a single window on your screen, type @kbd{C-x 1}.)
3351 @node Change a defun, , Effect of installation, Install
3352 @comment node-name, next, previous, up
3353 @subsection Change a Function Definition
3354 @cindex Changing a function definition
3355 @cindex Function definition, how to change
3356 @cindex Definition, how to change
3358 If you want to change the code in @code{multiply-by-seven}, just rewrite
3359 it. To install the new version in place of the old one, evaluate the
3360 function definition again. This is how you modify code in Emacs. It is
3363 As an example, you can change the @code{multiply-by-seven} function to
3364 add the number to itself seven times instead of multiplying the number
3365 by seven. It produces the same answer, but by a different path. At
3366 the same time, we will add a comment to the code; a comment is text
3367 that the Lisp interpreter ignores, but that a human reader may find
3368 useful or enlightening. The comment is that this is the ``second
3373 (defun multiply-by-seven (number) ; @r{Second version.}
3374 "Multiply NUMBER by seven."
3375 (+ number number number number number number number))
3379 @cindex Comments in Lisp code
3380 The comment follows a semicolon, @samp{;}. In Lisp, everything on a
3381 line that follows a semicolon is a comment. The end of the line is the
3382 end of the comment. To stretch a comment over two or more lines, begin
3383 each line with a semicolon.
3385 @xref{Beginning a .emacs File, , Beginning a @file{.emacs}
3386 File}, and @ref{Comments, , Comments, elisp, The GNU Emacs Lisp
3387 Reference Manual}, for more about comments.
3389 You can install this version of the @code{multiply-by-seven} function by
3390 evaluating it in the same way you evaluated the first function: place
3391 the cursor after the last parenthesis and type @kbd{C-x C-e}.
3393 In summary, this is how you write code in Emacs Lisp: you write a
3394 function; install it; test it; and then make fixes or enhancements and
3397 @node Interactive, Interactive Options, Install, Writing Defuns
3398 @comment node-name, next, previous, up
3399 @section Make a Function Interactive
3400 @cindex Interactive functions
3403 You make a function interactive by placing a list that begins with
3404 the special form @code{interactive} immediately after the
3405 documentation. A user can invoke an interactive function by typing
3406 @kbd{M-x} and then the name of the function; or by typing the keys to
3407 which it is bound, for example, by typing @kbd{C-n} for
3408 @code{next-line} or @kbd{C-x h} for @code{mark-whole-buffer}.
3410 Interestingly, when you call an interactive function interactively,
3411 the value returned is not automatically displayed in the echo area.
3412 This is because you often call an interactive function for its side
3413 effects, such as moving forward by a word or line, and not for the
3414 value returned. If the returned value were displayed in the echo area
3415 each time you typed a key, it would be very distracting.
3418 * Interactive multiply-by-seven:: An overview.
3419 * multiply-by-seven in detail:: The interactive version.
3422 @node Interactive multiply-by-seven, multiply-by-seven in detail, Interactive, Interactive
3424 @unnumberedsubsec An Interactive @code{multiply-by-seven}, An Overview
3427 Both the use of the special form @code{interactive} and one way to
3428 display a value in the echo area can be illustrated by creating an
3429 interactive version of @code{multiply-by-seven}.
3436 (defun multiply-by-seven (number) ; @r{Interactive version.}
3437 "Multiply NUMBER by seven."
3439 (message "The result is %d" (* 7 number)))
3444 You can install this code by placing your cursor after it and typing
3445 @kbd{C-x C-e}. The name of the function will appear in your echo area.
3446 Then, you can use this code by typing @kbd{C-u} and a number and then
3447 typing @kbd{M-x multiply-by-seven} and pressing @key{RET}. The phrase
3448 @samp{The result is @dots{}} followed by the product will appear in the
3451 Speaking more generally, you invoke a function like this in either of two
3456 By typing a prefix argument that contains the number to be passed, and
3457 then typing @kbd{M-x} and the name of the function, as with
3458 @kbd{C-u 3 M-x forward-sentence}; or,
3461 By typing whatever key or keychord the function is bound to, as with
3466 Both the examples just mentioned work identically to move point forward
3467 three sentences. (Since @code{multiply-by-seven} is not bound to a key,
3468 it could not be used as an example of key binding.)
3470 (@xref{Keybindings, , Some Keybindings}, to learn how to bind a command
3473 A prefix argument is passed to an interactive function by typing the
3474 @key{META} key followed by a number, for example, @kbd{M-3 M-e}, or by
3475 typing @kbd{C-u} and then a number, for example, @kbd{C-u 3 M-e} (if you
3476 type @kbd{C-u} without a number, it defaults to 4).
3478 @node multiply-by-seven in detail, , Interactive multiply-by-seven, Interactive
3479 @comment node-name, next, previous, up
3480 @subsection An Interactive @code{multiply-by-seven}
3482 Let's look at the use of the special form @code{interactive} and then at
3483 the function @code{message} in the interactive version of
3484 @code{multiply-by-seven}. You will recall that the function definition
3489 (defun multiply-by-seven (number) ; @r{Interactive version.}
3490 "Multiply NUMBER by seven."
3492 (message "The result is %d" (* 7 number)))
3496 In this function, the expression, @code{(interactive "p")}, is a list of
3497 two elements. The @code{"p"} tells Emacs to pass the prefix argument to
3498 the function and use its value for the argument of the function.
3501 The argument will be a number. This means that the symbol
3502 @code{number} will be bound to a number in the line:
3505 (message "The result is %d" (* 7 number))
3510 For example, if your prefix argument is 5, the Lisp interpreter will
3511 evaluate the line as if it were:
3514 (message "The result is %d" (* 7 5))
3518 (If you are reading this in GNU Emacs, you can evaluate this expression
3519 yourself.) First, the interpreter will evaluate the inner list, which
3520 is @code{(* 7 5)}. This returns a value of 35. Next, it
3521 will evaluate the outer list, passing the values of the second and
3522 subsequent elements of the list to the function @code{message}.
3524 As we have seen, @code{message} is an Emacs Lisp function especially
3525 designed for sending a one line message to a user. (@xref{message, ,
3526 The @code{message} function}.) In summary, the @code{message}
3527 function prints its first argument in the echo area as is, except for
3528 occurrences of @samp{%d} or @samp{%s} (and various other %-sequences
3529 which we have not mentioned). When it sees a control sequence, the
3530 function looks to the second or subsequent arguments and prints the
3531 value of the argument in the location in the string where the control
3532 sequence is located.
3534 In the interactive @code{multiply-by-seven} function, the control string
3535 is @samp{%d}, which requires a number, and the value returned by
3536 evaluating @code{(* 7 5)} is the number 35. Consequently, the number 35
3537 is printed in place of the @samp{%d} and the message is @samp{The result
3540 (Note that when you call the function @code{multiply-by-seven}, the
3541 message is printed without quotes, but when you call @code{message}, the
3542 text is printed in double quotes. This is because the value returned by
3543 @code{message} is what appears in the echo area when you evaluate an
3544 expression whose first element is @code{message}; but when embedded in a
3545 function, @code{message} prints the text as a side effect without
3548 @node Interactive Options, Permanent Installation, Interactive, Writing Defuns
3549 @comment node-name, next, previous, up
3550 @section Different Options for @code{interactive}
3551 @cindex Options for @code{interactive}
3552 @cindex Interactive options
3554 In the example, @code{multiply-by-seven} used @code{"p"} as the
3555 argument to @code{interactive}. This argument told Emacs to interpret
3556 your typing either @kbd{C-u} followed by a number or @key{META}
3557 followed by a number as a command to pass that number to the function
3558 as its argument. Emacs has more than twenty characters predefined for
3559 use with @code{interactive}. In almost every case, one of these
3560 options will enable you to pass the right information interactively to
3561 a function. (@xref{Interactive Codes, , Code Characters for
3562 @code{interactive}, elisp, The GNU Emacs Lisp Reference Manual}.)
3565 Consider the function @code{zap-to-char}. Its interactive expression
3569 (interactive "p\ncZap to char: ")
3572 The first part of the argument to @code{interactive} is @samp{p}, with
3573 which you are already familiar. This argument tells Emacs to
3574 interpret a `prefix', as a number to be passed to the function. You
3575 can specify a prefix either by typing @kbd{C-u} followed by a number
3576 or by typing @key{META} followed by a number. The prefix is the
3577 number of specified characters. Thus, if your prefix is three and the
3578 specified character is @samp{x}, then you will delete all the text up
3579 to and including the third next @samp{x}. If you do not set a prefix,
3580 then you delete all the text up to and including the specified
3581 character, but no more.
3583 The @samp{c} tells the function the name of the character to which to delete.
3585 More formally, a function with two or more arguments can have
3586 information passed to each argument by adding parts to the string that
3587 follows @code{interactive}. When you do this, the information is
3588 passed to each argument in the same order it is specified in the
3589 @code{interactive} list. In the string, each part is separated from
3590 the next part by a @samp{\n}, which is a newline. For example, you
3591 can follow @samp{p} with a @samp{\n} and an @samp{cZap to char:@: }.
3592 This causes Emacs to pass the value of the prefix argument (if there
3593 is one) and the character.
3595 In this case, the function definition looks like the following, where
3596 @code{arg} and @code{char} are the symbols to which @code{interactive}
3597 binds the prefix argument and the specified character:
3601 (defun @var{name-of-function} (arg char)
3602 "@var{documentation}@dots{}"
3603 (interactive "p\ncZap to char: ")
3604 @var{body-of-function}@dots{})
3609 (The space after the colon in the prompt makes it look better when you
3610 are prompted. @xref{copy-to-buffer, , The Definition of
3611 @code{copy-to-buffer}}, for an example.)
3613 When a function does not take arguments, @code{interactive} does not
3614 require any. Such a function contains the simple expression
3615 @code{(interactive)}. The @code{mark-whole-buffer} function is like
3618 Alternatively, if the special letter-codes are not right for your
3619 application, you can pass your own arguments to @code{interactive} as
3622 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}},
3623 for an example. @xref{Using Interactive, , Using @code{Interactive},
3624 elisp, The GNU Emacs Lisp Reference Manual}, for a more complete
3625 explanation about this technique.
3627 @node Permanent Installation, let, Interactive Options, Writing Defuns
3628 @comment node-name, next, previous, up
3629 @section Install Code Permanently
3630 @cindex Install code permanently
3631 @cindex Permanent code installation
3632 @cindex Code installation
3634 When you install a function definition by evaluating it, it will stay
3635 installed until you quit Emacs. The next time you start a new session
3636 of Emacs, the function will not be installed unless you evaluate the
3637 function definition again.
3639 At some point, you may want to have code installed automatically
3640 whenever you start a new session of Emacs. There are several ways of
3645 If you have code that is just for yourself, you can put the code for the
3646 function definition in your @file{.emacs} initialization file. When you
3647 start Emacs, your @file{.emacs} file is automatically evaluated and all
3648 the function definitions within it are installed.
3649 @xref{Emacs Initialization, , Your @file{.emacs} File}.
3652 Alternatively, you can put the function definitions that you want
3653 installed in one or more files of their own and use the @code{load}
3654 function to cause Emacs to evaluate and thereby install each of the
3655 functions in the files.
3656 @xref{Loading Files, , Loading Files}.
3659 Thirdly, if you have code that your whole site will use, it is usual
3660 to put it in a file called @file{site-init.el} that is loaded when
3661 Emacs is built. This makes the code available to everyone who uses
3662 your machine. (See the @file{INSTALL} file that is part of the Emacs
3666 Finally, if you have code that everyone who uses Emacs may want, you
3667 can post it on a computer network or send a copy to the Free Software
3668 Foundation. (When you do this, please license the code and its
3669 documentation under a license that permits other people to run, copy,
3670 study, modify, and redistribute the code and which protects you from
3671 having your work taken from you.) If you send a copy of your code to
3672 the Free Software Foundation, and properly protect yourself and
3673 others, it may be included in the next release of Emacs. In large
3674 part, this is how Emacs has grown over the past years, by donations.
3676 @node let, if, Permanent Installation, Writing Defuns
3677 @comment node-name, next, previous, up
3681 The @code{let} expression is a special form in Lisp that you will need
3682 to use in most function definitions.
3684 @code{let} is used to attach or bind a symbol to a value in such a way
3685 that the Lisp interpreter will not confuse the variable with a
3686 variable of the same name that is not part of the function.
3688 To understand why the @code{let} special form is necessary, consider
3689 the situation in which you own a home that you generally refer to as
3690 `the house', as in the sentence, ``The house needs painting.'' If you
3691 are visiting a friend and your host refers to `the house', he is
3692 likely to be referring to @emph{his} house, not yours, that is, to a
3695 If your friend is referring to his house and you think he is referring
3696 to your house, you may be in for some confusion. The same thing could
3697 happen in Lisp if a variable that is used inside of one function has
3698 the same name as a variable that is used inside of another function,
3699 and the two are not intended to refer to the same value. The
3700 @code{let} special form prevents this kind of confusion.
3703 * Prevent confusion::
3704 * Parts of let Expression::
3705 * Sample let Expression::
3706 * Uninitialized let Variables::
3709 @node Prevent confusion, Parts of let Expression, let, let
3711 @unnumberedsubsec @code{let} Prevents Confusion
3714 @cindex @samp{local variable} defined
3715 @cindex @samp{variable, local}, defined
3716 The @code{let} special form prevents confusion. @code{let} creates a
3717 name for a @dfn{local variable} that overshadows any use of the same
3718 name outside the @code{let} expression. This is like understanding
3719 that whenever your host refers to `the house', he means his house, not
3720 yours. (Symbols used in argument lists work the same way.
3721 @xref{defun, , The @code{defun} Special Form}.)
3723 Local variables created by a @code{let} expression retain their value
3724 @emph{only} within the @code{let} expression itself (and within
3725 expressions called within the @code{let} expression); the local
3726 variables have no effect outside the @code{let} expression.
3728 Another way to think about @code{let} is that it is like a @code{setq}
3729 that is temporary and local. The values set by @code{let} are
3730 automatically undone when the @code{let} is finished. The setting
3731 only affects expressions that are inside the bounds of the @code{let}
3732 expression. In computer science jargon, we would say ``the binding of
3733 a symbol is visible only in functions called in the @code{let} form;
3734 in Emacs Lisp, scoping is dynamic, not lexical.''
3736 @code{let} can create more than one variable at once. Also,
3737 @code{let} gives each variable it creates an initial value, either a
3738 value specified by you, or @code{nil}. (In the jargon, this is called
3739 `binding the variable to the value'.) After @code{let} has created
3740 and bound the variables, it executes the code in the body of the
3741 @code{let}, and returns the value of the last expression in the body,
3742 as the value of the whole @code{let} expression. (`Execute' is a jargon
3743 term that means to evaluate a list; it comes from the use of the word
3744 meaning `to give practical effect to' (@cite{Oxford English
3745 Dictionary}). Since you evaluate an expression to perform an action,
3746 `execute' has evolved as a synonym to `evaluate'.)
3748 @node Parts of let Expression, Sample let Expression, Prevent confusion, let
3749 @comment node-name, next, previous, up
3750 @subsection The Parts of a @code{let} Expression
3751 @cindex @code{let} expression, parts of
3752 @cindex Parts of @code{let} expression
3754 @cindex @samp{varlist} defined
3755 A @code{let} expression is a list of three parts. The first part is
3756 the symbol @code{let}. The second part is a list, called a
3757 @dfn{varlist}, each element of which is either a symbol by itself or a
3758 two-element list, the first element of which is a symbol. The third
3759 part of the @code{let} expression is the body of the @code{let}. The
3760 body usually consists of one or more lists.
3763 A template for a @code{let} expression looks like this:
3766 (let @var{varlist} @var{body}@dots{})
3770 The symbols in the varlist are the variables that are given initial
3771 values by the @code{let} special form. Symbols by themselves are given
3772 the initial value of @code{nil}; and each symbol that is the first
3773 element of a two-element list is bound to the value that is returned
3774 when the Lisp interpreter evaluates the second element.
3776 Thus, a varlist might look like this: @code{(thread (needles 3))}. In
3777 this case, in a @code{let} expression, Emacs binds the symbol
3778 @code{thread} to an initial value of @code{nil}, and binds the symbol
3779 @code{needles} to an initial value of 3.
3781 When you write a @code{let} expression, what you do is put the
3782 appropriate expressions in the slots of the @code{let} expression
3785 If the varlist is composed of two-element lists, as is often the case,
3786 the template for the @code{let} expression looks like this:
3790 (let ((@var{variable} @var{value})
3791 (@var{variable} @var{value})
3797 @node Sample let Expression, Uninitialized let Variables, Parts of let Expression, let
3798 @comment node-name, next, previous, up
3799 @subsection Sample @code{let} Expression
3800 @cindex Sample @code{let} expression
3801 @cindex @code{let} expression sample
3803 The following expression creates and gives initial values
3804 to the two variables @code{zebra} and @code{tiger}. The body of the
3805 @code{let} expression is a list which calls the @code{message} function.
3809 (let ((zebra 'stripes)
3811 (message "One kind of animal has %s and another is %s."
3816 Here, the varlist is @code{((zebra 'stripes) (tiger 'fierce))}.
3818 The two variables are @code{zebra} and @code{tiger}. Each variable is
3819 the first element of a two-element list and each value is the second
3820 element of its two-element list. In the varlist, Emacs binds the
3821 variable @code{zebra} to the value @code{stripes}@footnote{According
3822 to Jared Diamond in @cite{Guns, Germs, and Steel}, ``@dots{} zebras
3823 become impossibly dangerous as they grow older'' but the claim here is
3824 that they do not become fierce like a tiger. (1997, W. W. Norton and
3825 Co., ISBN 0-393-03894-2, page 171)}, and binds the
3826 variable @code{tiger} to the value @code{fierce}. In this example,
3827 both values are symbols preceded by a quote. The values could just as
3828 well have been another list or a string. The body of the @code{let}
3829 follows after the list holding the variables. In this example, the
3830 body is a list that uses the @code{message} function to print a string
3834 You may evaluate the example in the usual fashion, by placing the
3835 cursor after the last parenthesis and typing @kbd{C-x C-e}. When you do
3836 this, the following will appear in the echo area:
3839 "One kind of animal has stripes and another is fierce."
3842 As we have seen before, the @code{message} function prints its first
3843 argument, except for @samp{%s}. In this example, the value of the variable
3844 @code{zebra} is printed at the location of the first @samp{%s} and the
3845 value of the variable @code{tiger} is printed at the location of the
3848 @node Uninitialized let Variables, , Sample let Expression, let
3849 @comment node-name, next, previous, up
3850 @subsection Uninitialized Variables in a @code{let} Statement
3851 @cindex Uninitialized @code{let} variables
3852 @cindex @code{let} variables uninitialized
3854 If you do not bind the variables in a @code{let} statement to specific
3855 initial values, they will automatically be bound to an initial value of
3856 @code{nil}, as in the following expression:
3865 "Here are %d variables with %s, %s, and %s value."
3866 birch pine fir oak))
3871 Here, the varlist is @code{((birch 3) pine fir (oak 'some))}.
3874 If you evaluate this expression in the usual way, the following will
3875 appear in your echo area:
3878 "Here are 3 variables with nil, nil, and some value."
3882 In this example, Emacs binds the symbol @code{birch} to the number 3,
3883 binds the symbols @code{pine} and @code{fir} to @code{nil}, and binds
3884 the symbol @code{oak} to the value @code{some}.
3886 Note that in the first part of the @code{let}, the variables @code{pine}
3887 and @code{fir} stand alone as atoms that are not surrounded by
3888 parentheses; this is because they are being bound to @code{nil}, the
3889 empty list. But @code{oak} is bound to @code{some} and so is a part of
3890 the list @code{(oak 'some)}. Similarly, @code{birch} is bound to the
3891 number 3 and so is in a list with that number. (Since a number
3892 evaluates to itself, the number does not need to be quoted. Also, the
3893 number is printed in the message using a @samp{%d} rather than a
3894 @samp{%s}.) The four variables as a group are put into a list to
3895 delimit them from the body of the @code{let}.
3897 @node if, else, let, Writing Defuns
3898 @comment node-name, next, previous, up
3899 @section The @code{if} Special Form
3901 @cindex Conditional with @code{if}
3903 A third special form, in addition to @code{defun} and @code{let}, is the
3904 conditional @code{if}. This form is used to instruct the computer to
3905 make decisions. You can write function definitions without using
3906 @code{if}, but it is used often enough, and is important enough, to be
3907 included here. It is used, for example, in the code for the
3908 function @code{beginning-of-buffer}.
3910 The basic idea behind an @code{if}, is that ``@emph{if} a test is true,
3911 @emph{then} an expression is evaluated.'' If the test is not true, the
3912 expression is not evaluated. For example, you might make a decision
3913 such as, ``if it is warm and sunny, then go to the beach!''
3916 * if in more detail::
3917 * type-of-animal in detail:: An example of an @code{if} expression.
3920 @node if in more detail, type-of-animal in detail, if, if
3922 @unnumberedsubsec @code{if} in more detail
3925 @cindex @samp{if-part} defined
3926 @cindex @samp{then-part} defined
3927 An @code{if} expression written in Lisp does not use the word `then';
3928 the test and the action are the second and third elements of the list
3929 whose first element is @code{if}. Nonetheless, the test part of an
3930 @code{if} expression is often called the @dfn{if-part} and the second
3931 argument is often called the @dfn{then-part}.
3933 Also, when an @code{if} expression is written, the true-or-false-test
3934 is usually written on the same line as the symbol @code{if}, but the
3935 action to carry out if the test is true, the ``then-part'', is written
3936 on the second and subsequent lines. This makes the @code{if}
3937 expression easier to read.
3941 (if @var{true-or-false-test}
3942 @var{action-to-carry-out-if-test-is-true})
3947 The true-or-false-test will be an expression that
3948 is evaluated by the Lisp interpreter.
3950 Here is an example that you can evaluate in the usual manner. The test
3951 is whether the number 5 is greater than the number 4. Since it is, the
3952 message @samp{5 is greater than 4!} will be printed.
3956 (if (> 5 4) ; @r{if-part}
3957 (message "5 is greater than 4!")) ; @r{then-part}
3962 (The function @code{>} tests whether its first argument is greater than
3963 its second argument and returns true if it is.)
3964 @findex > (greater than)
3966 Of course, in actual use, the test in an @code{if} expression will not
3967 be fixed for all time as it is by the expression @code{(> 5 4)}.
3968 Instead, at least one of the variables used in the test will be bound to
3969 a value that is not known ahead of time. (If the value were known ahead
3970 of time, we would not need to run the test!)
3972 For example, the value may be bound to an argument of a function
3973 definition. In the following function definition, the character of the
3974 animal is a value that is passed to the function. If the value bound to
3975 @code{characteristic} is @code{fierce}, then the message, @samp{It's a
3976 tiger!} will be printed; otherwise, @code{nil} will be returned.
3980 (defun type-of-animal (characteristic)
3981 "Print message in echo area depending on CHARACTERISTIC.
3982 If the CHARACTERISTIC is the symbol `fierce',
3983 then warn of a tiger."
3984 (if (equal characteristic 'fierce)
3985 (message "It's a tiger!")))
3991 If you are reading this inside of GNU Emacs, you can evaluate the
3992 function definition in the usual way to install it in Emacs, and then you
3993 can evaluate the following two expressions to see the results:
3997 (type-of-animal 'fierce)
3999 (type-of-animal 'zebra)
4004 @c Following sentences rewritten to prevent overfull hbox.
4006 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4007 following message printed in the echo area: @code{"It's a tiger!"}; and
4008 when you evaluate @code{(type-of-animal 'zebra)} you will see @code{nil}
4009 printed in the echo area.
4011 @node type-of-animal in detail, , if in more detail, if
4012 @comment node-name, next, previous, up
4013 @subsection The @code{type-of-animal} Function in Detail
4015 Let's look at the @code{type-of-animal} function in detail.
4017 The function definition for @code{type-of-animal} was written by filling
4018 the slots of two templates, one for a function definition as a whole, and
4019 a second for an @code{if} expression.
4022 The template for every function that is not interactive is:
4026 (defun @var{name-of-function} (@var{argument-list})
4027 "@var{documentation}@dots{}"
4033 The parts of the function that match this template look like this:
4037 (defun type-of-animal (characteristic)
4038 "Print message in echo area depending on CHARACTERISTIC.
4039 If the CHARACTERISTIC is the symbol `fierce',
4040 then warn of a tiger."
4041 @var{body: the} @code{if} @var{expression})
4045 The name of function is @code{type-of-animal}; it is passed the value
4046 of one argument. The argument list is followed by a multi-line
4047 documentation string. The documentation string is included in the
4048 example because it is a good habit to write documentation string for
4049 every function definition. The body of the function definition
4050 consists of the @code{if} expression.
4053 The template for an @code{if} expression looks like this:
4057 (if @var{true-or-false-test}
4058 @var{action-to-carry-out-if-the-test-returns-true})
4063 In the @code{type-of-animal} function, the code for the @code{if}
4068 (if (equal characteristic 'fierce)
4069 (message "It's a tiger!")))
4074 Here, the true-or-false-test is the expression:
4077 (equal characteristic 'fierce)
4081 In Lisp, @code{equal} is a function that determines whether its first
4082 argument is equal to its second argument. The second argument is the
4083 quoted symbol @code{'fierce} and the first argument is the value of the
4084 symbol @code{characteristic}---in other words, the argument passed to
4087 In the first exercise of @code{type-of-animal}, the argument
4088 @code{fierce} is passed to @code{type-of-animal}. Since @code{fierce}
4089 is equal to @code{fierce}, the expression, @code{(equal characteristic
4090 'fierce)}, returns a value of true. When this happens, the @code{if}
4091 evaluates the second argument or then-part of the @code{if}:
4092 @code{(message "It's tiger!")}.
4094 On the other hand, in the second exercise of @code{type-of-animal}, the
4095 argument @code{zebra} is passed to @code{type-of-animal}. @code{zebra}
4096 is not equal to @code{fierce}, so the then-part is not evaluated and
4097 @code{nil} is returned by the @code{if} expression.
4099 @node else, Truth & Falsehood, if, Writing Defuns
4100 @comment node-name, next, previous, up
4101 @section If--then--else Expressions
4104 An @code{if} expression may have an optional third argument, called
4105 the @dfn{else-part}, for the case when the true-or-false-test returns
4106 false. When this happens, the second argument or then-part of the
4107 overall @code{if} expression is @emph{not} evaluated, but the third or
4108 else-part @emph{is} evaluated. You might think of this as the cloudy
4109 day alternative for the decision ``if it is warm and sunny, then go to
4110 the beach, else read a book!''.
4112 The word ``else'' is not written in the Lisp code; the else-part of an
4113 @code{if} expression comes after the then-part. In the written Lisp, the
4114 else-part is usually written to start on a line of its own and is
4115 indented less than the then-part:
4119 (if @var{true-or-false-test}
4120 @var{action-to-carry-out-if-the-test-returns-true}
4121 @var{action-to-carry-out-if-the-test-returns-false})
4125 For example, the following @code{if} expression prints the message @samp{4
4126 is not greater than 5!} when you evaluate it in the usual way:
4130 (if (> 4 5) ; @r{if-part}
4131 (message "4 falsely greater than 5!") ; @r{then-part}
4132 (message "4 is not greater than 5!")) ; @r{else-part}
4137 Note that the different levels of indentation make it easy to
4138 distinguish the then-part from the else-part. (GNU Emacs has several
4139 commands that automatically indent @code{if} expressions correctly.
4140 @xref{Typing Lists, , GNU Emacs Helps You Type Lists}.)
4142 We can extend the @code{type-of-animal} function to include an
4143 else-part by simply incorporating an additional part to the @code{if}
4147 You can see the consequences of doing this if you evaluate the following
4148 version of the @code{type-of-animal} function definition to install it
4149 and then evaluate the two subsequent expressions to pass different
4150 arguments to the function.
4154 (defun type-of-animal (characteristic) ; @r{Second version.}
4155 "Print message in echo area depending on CHARACTERISTIC.
4156 If the CHARACTERISTIC is the symbol `fierce',
4157 then warn of a tiger;
4158 else say it's not fierce."
4159 (if (equal characteristic 'fierce)
4160 (message "It's a tiger!")
4161 (message "It's not fierce!")))
4168 (type-of-animal 'fierce)
4170 (type-of-animal 'zebra)
4175 @c Following sentence rewritten to prevent overfull hbox.
4177 When you evaluate @code{(type-of-animal 'fierce)}, you will see the
4178 following message printed in the echo area: @code{"It's a tiger!"}; but
4179 when you evaluate @code{(type-of-animal 'zebra)}, you will see
4180 @code{"It's not fierce!"}.
4182 (Of course, if the @var{characteristic} were @code{ferocious}, the
4183 message @code{"It's not fierce!"} would be printed; and it would be
4184 misleading! When you write code, you need to take into account the
4185 possibility that some such argument will be tested by the @code{if}
4186 and write your program accordingly.)
4188 @node Truth & Falsehood, save-excursion, else, Writing Defuns
4189 @comment node-name, next, previous, up
4190 @section Truth and Falsehood in Emacs Lisp
4191 @cindex Truth and falsehood in Emacs Lisp
4192 @cindex Falsehood and truth in Emacs Lisp
4195 There is an important aspect to the truth test in an @code{if}
4196 expression. So far, we have spoken of `true' and `false' as values of
4197 predicates as if they were new kinds of Emacs Lisp objects. In fact,
4198 `false' is just our old friend @code{nil}. Anything else---anything
4201 The expression that tests for truth is interpreted as @dfn{true}
4202 if the result of evaluating it is a value that is not @code{nil}. In
4203 other words, the result of the test is considered true if the value
4204 returned is a number such as 47, a string such as @code{"hello"}, or a
4205 symbol (other than @code{nil}) such as @code{flowers}, or a list (so
4206 long as it is not empty), or even a buffer!
4209 * nil explained:: @code{nil} has two meanings.
4212 @node nil explained, , Truth & Falsehood, Truth & Falsehood
4214 @unnumberedsubsec An explanation of @code{nil}
4217 Before illustrating a test for truth, we need an explanation of @code{nil}.
4219 In Emacs Lisp, the symbol @code{nil} has two meanings. First, it means the
4220 empty list. Second, it means false and is the value returned when a
4221 true-or-false-test tests false. @code{nil} can be written as an empty
4222 list, @code{()}, or as @code{nil}. As far as the Lisp interpreter is
4223 concerned, @code{()} and @code{nil} are the same. Humans, however, tend
4224 to use @code{nil} for false and @code{()} for the empty list.
4226 In Emacs Lisp, any value that is not @code{nil}---is not the empty
4227 list---is considered true. This means that if an evaluation returns
4228 something that is not an empty list, an @code{if} expression will test
4229 true. For example, if a number is put in the slot for the test, it
4230 will be evaluated and will return itself, since that is what numbers
4231 do when evaluated. In this conditional, the @code{if} expression will
4232 test true. The expression tests false only when @code{nil}, an empty
4233 list, is returned by evaluating the expression.
4235 You can see this by evaluating the two expressions in the following examples.
4237 In the first example, the number 4 is evaluated as the test in the
4238 @code{if} expression and returns itself; consequently, the then-part
4239 of the expression is evaluated and returned: @samp{true} appears in
4240 the echo area. In the second example, the @code{nil} indicates false;
4241 consequently, the else-part of the expression is evaluated and
4242 returned: @samp{false} appears in the echo area.
4259 Incidentally, if some other useful value is not available for a test that
4260 returns true, then the Lisp interpreter will return the symbol @code{t}
4261 for true. For example, the expression @code{(> 5 4)} returns @code{t}
4262 when evaluated, as you can see by evaluating it in the usual way:
4270 On the other hand, this function returns @code{nil} if the test is false.
4276 @node save-excursion, Review, Truth & Falsehood, Writing Defuns
4277 @comment node-name, next, previous, up
4278 @section @code{save-excursion}
4279 @findex save-excursion
4280 @cindex Region, what it is
4281 @cindex Preserving point, mark, and buffer
4282 @cindex Point, mark, buffer preservation
4286 The @code{save-excursion} function is the fourth and final special form
4287 that we will discuss in this chapter.
4289 In Emacs Lisp programs used for editing, the @code{save-excursion}
4290 function is very common. It saves the location of point and mark,
4291 executes the body of the function, and then restores point and mark to
4292 their previous positions if their locations were changed. Its primary
4293 purpose is to keep the user from being surprised and disturbed by
4294 unexpected movement of point or mark.
4297 * Point and mark:: A review of various locations.
4298 * Template for save-excursion::
4301 @node Point and mark, Template for save-excursion, save-excursion, save-excursion
4303 @unnumberedsubsec Point and Mark
4306 Before discussing @code{save-excursion}, however, it may be useful
4307 first to review what point and mark are in GNU Emacs. @dfn{Point} is
4308 the current location of the cursor. Wherever the cursor
4309 is, that is point. More precisely, on terminals where the cursor
4310 appears to be on top of a character, point is immediately before the
4311 character. In Emacs Lisp, point is an integer. The first character in
4312 a buffer is number one, the second is number two, and so on. The
4313 function @code{point} returns the current position of the cursor as a
4314 number. Each buffer has its own value for point.
4316 The @dfn{mark} is another position in the buffer; its value can be set
4317 with a command such as @kbd{C-@key{SPC}} (@code{set-mark-command}). If
4318 a mark has been set, you can use the command @kbd{C-x C-x}
4319 (@code{exchange-point-and-mark}) to cause the cursor to jump to the mark
4320 and set the mark to be the previous position of point. In addition, if
4321 you set another mark, the position of the previous mark is saved in the
4322 mark ring. Many mark positions can be saved this way. You can jump the
4323 cursor to a saved mark by typing @kbd{C-u C-@key{SPC}} one or more
4326 The part of the buffer between point and mark is called @dfn{the
4327 region}. Numerous commands work on the region, including
4328 @code{center-region}, @code{count-lines-region}, @code{kill-region}, and
4329 @code{print-region}.
4331 The @code{save-excursion} special form saves the locations of point and
4332 mark and restores those positions after the code within the body of the
4333 special form is evaluated by the Lisp interpreter. Thus, if point were
4334 in the beginning of a piece of text and some code moved point to the end
4335 of the buffer, the @code{save-excursion} would put point back to where
4336 it was before, after the expressions in the body of the function were
4339 In Emacs, a function frequently moves point as part of its internal
4340 workings even though a user would not expect this. For example,
4341 @code{count-lines-region} moves point. To prevent the user from being
4342 bothered by jumps that are both unexpected and (from the user's point of
4343 view) unnecessary, @code{save-excursion} is often used to keep point and
4344 mark in the location expected by the user. The use of
4345 @code{save-excursion} is good housekeeping.
4347 To make sure the house stays clean, @code{save-excursion} restores the
4348 values of point and mark even if something goes wrong in the code inside
4349 of it (or, to be more precise and to use the proper jargon, ``in case of
4350 abnormal exit''). This feature is very helpful.
4352 In addition to recording the values of point and mark,
4353 @code{save-excursion} keeps track of the current buffer, and restores
4354 it, too. This means you can write code that will change the buffer and
4355 have @code{save-excursion} switch you back to the original buffer.
4356 This is how @code{save-excursion} is used in @code{append-to-buffer}.
4357 (@xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
4359 @node Template for save-excursion, , Point and mark, save-excursion
4360 @comment node-name, next, previous, up
4361 @subsection Template for a @code{save-excursion} Expression
4364 The template for code using @code{save-excursion} is simple:
4374 The body of the function is one or more expressions that will be
4375 evaluated in sequence by the Lisp interpreter. If there is more than
4376 one expression in the body, the value of the last one will be returned
4377 as the value of the @code{save-excursion} function. The other
4378 expressions in the body are evaluated only for their side effects; and
4379 @code{save-excursion} itself is used only for its side effect (which
4380 is restoring the positions of point and mark).
4383 In more detail, the template for a @code{save-excursion} expression
4389 @var{first-expression-in-body}
4390 @var{second-expression-in-body}
4391 @var{third-expression-in-body}
4393 @var{last-expression-in-body})
4398 An expression, of course, may be a symbol on its own or a list.
4400 In Emacs Lisp code, a @code{save-excursion} expression often occurs
4401 within the body of a @code{let} expression. It looks like this:
4411 @node Review, defun Exercises, save-excursion, Writing Defuns
4412 @comment node-name, next, previous, up
4415 In the last few chapters we have introduced a fair number of functions
4416 and special forms. Here they are described in brief, along with a few
4417 similar functions that have not been mentioned yet.
4420 @item eval-last-sexp
4421 Evaluate the last symbolic expression before the current location of
4422 point. The value is printed in the echo area unless the function is
4423 invoked with an argument; in that case, the output is printed in the
4424 current buffer. This command is normally bound to @kbd{C-x C-e}.
4427 Define function. This special form has up to five parts: the name,
4428 a template for the arguments that will be passed to the function,
4429 documentation, an optional interactive declaration, and the body of the
4433 For example, in an early version of Emacs, the function definition was
4434 as follows. (It is slightly more complex now that it seeks the first
4435 non-whitespace character rather than the first visible character.)
4439 (defun back-to-indentation ()
4440 "Move point to first visible character on line."
4442 (beginning-of-line 1)
4443 (skip-chars-forward " \t"))
4450 (defun backward-to-indentation (&optional arg)
4451 "Move backward ARG lines and position at first nonblank character."
4453 (forward-line (- (or arg 1)))
4454 (skip-chars-forward " \t"))
4456 (defun back-to-indentation ()
4457 "Move point to the first non-whitespace character on this line."
4459 (beginning-of-line 1)
4460 (skip-syntax-forward " " (line-end-position))
4461 ;; Move back over chars that have whitespace syntax but have the p flag.
4462 (backward-prefix-chars))
4466 Declare to the interpreter that the function can be used
4467 interactively. This special form may be followed by a string with one
4468 or more parts that pass the information to the arguments of the
4469 function, in sequence. These parts may also tell the interpreter to
4470 prompt for information. Parts of the string are separated by
4471 newlines, @samp{\n}.
4474 Common code characters are:
4478 The name of an existing buffer.
4481 The name of an existing file.
4484 The numeric prefix argument. (Note that this `p' is lower case.)
4487 Point and the mark, as two numeric arguments, smallest first. This
4488 is the only code letter that specifies two successive arguments
4492 @xref{Interactive Codes, , Code Characters for @samp{interactive},
4493 elisp, The GNU Emacs Lisp Reference Manual}, for a complete list of
4497 Declare that a list of variables is for use within the body of the
4498 @code{let} and give them an initial value, either @code{nil} or a
4499 specified value; then evaluate the rest of the expressions in the body
4500 of the @code{let} and return the value of the last one. Inside the
4501 body of the @code{let}, the Lisp interpreter does not see the values of
4502 the variables of the same names that are bound outside of the
4510 (let ((foo (buffer-name))
4511 (bar (buffer-size)))
4513 "This buffer is %s and has %d characters."
4518 @item save-excursion
4519 Record the values of point and mark and the current buffer before
4520 evaluating the body of this special form. Restore the values of point
4521 and mark and buffer afterward.
4528 (message "We are %d characters into this buffer."
4531 (goto-char (point-min)) (point))))
4536 Evaluate the first argument to the function; if it is true, evaluate
4537 the second argument; else evaluate the third argument, if there is one.
4539 The @code{if} special form is called a @dfn{conditional}. There are
4540 other conditionals in Emacs Lisp, but @code{if} is perhaps the most
4548 (if (= 22 emacs-major-version)
4549 (message "This is version 22 Emacs")
4550 (message "This is not version 22 Emacs"))
4559 The @code{<} function tests whether its first argument is smaller than
4560 its second argument. A corresponding function, @code{>}, tests whether
4561 the first argument is greater than the second. Likewise, @code{<=}
4562 tests whether the first argument is less than or equal to the second and
4563 @code{>=} tests whether the first argument is greater than or equal to
4564 the second. In all cases, both arguments must be numbers or markers
4565 (markers indicate positions in buffers).
4569 The @code{=} function tests whether two arguments, both numbers or
4575 Test whether two objects are the same. @code{equal} uses one meaning
4576 of the word `same' and @code{eq} uses another: @code{equal} returns
4577 true if the two objects have a similar structure and contents, such as
4578 two copies of the same book. On the other hand, @code{eq}, returns
4579 true if both arguments are actually the same object.
4588 The @code{string-lessp} function tests whether its first argument is
4589 smaller than the second argument. A shorter, alternative name for the
4590 same function (a @code{defalias}) is @code{string<}.
4592 The arguments to @code{string-lessp} must be strings or symbols; the
4593 ordering is lexicographic, so case is significant. The print names of
4594 symbols are used instead of the symbols themselves.
4596 @cindex @samp{empty string} defined
4597 An empty string, @samp{""}, a string with no characters in it, is
4598 smaller than any string of characters.
4600 @code{string-equal} provides the corresponding test for equality. Its
4601 shorter, alternative name is @code{string=}. There are no string test
4602 functions that correspond to @var{>}, @code{>=}, or @code{<=}.
4605 Print a message in the echo area. The first argument is a string that
4606 can contain @samp{%s}, @samp{%d}, or @samp{%c} to print the value of
4607 arguments that follow the string. The argument used by @samp{%s} must
4608 be a string or a symbol; the argument used by @samp{%d} must be a
4609 number. The argument used by @samp{%c} must be an @sc{ascii} code
4610 number; it will be printed as the character with that @sc{ascii} code.
4611 (Various other %-sequences have not been mentioned.)
4615 The @code{setq} function sets the value of its first argument to the
4616 value of the second argument. The first argument is automatically
4617 quoted by @code{setq}. It does the same for succeeding pairs of
4618 arguments. Another function, @code{set}, takes only two arguments and
4619 evaluates both of them before setting the value returned by its first
4620 argument to the value returned by its second argument.
4623 Without an argument, return the name of the buffer, as a string.
4625 @itemx buffer-file-name
4626 Without an argument, return the name of the file the buffer is
4629 @item current-buffer
4630 Return the buffer in which Emacs is active; it may not be
4631 the buffer that is visible on the screen.
4634 Return the most recently selected buffer (other than the buffer passed
4635 to @code{other-buffer} as an argument and other than the current
4638 @item switch-to-buffer
4639 Select a buffer for Emacs to be active in and display it in the current
4640 window so users can look at it. Usually bound to @kbd{C-x b}.
4643 Switch Emacs' attention to a buffer on which programs will run. Don't
4644 alter what the window is showing.
4647 Return the number of characters in the current buffer.
4650 Return the value of the current position of the cursor, as an
4651 integer counting the number of characters from the beginning of the
4655 Return the minimum permissible value of point in
4656 the current buffer. This is 1, unless narrowing is in effect.
4659 Return the value of the maximum permissible value of point in the
4660 current buffer. This is the end of the buffer, unless narrowing is in
4665 @node defun Exercises, , Review, Writing Defuns
4670 Write a non-interactive function that doubles the value of its
4671 argument, a number. Make that function interactive.
4674 Write a function that tests whether the current value of
4675 @code{fill-column} is greater than the argument passed to the function,
4676 and if so, prints an appropriate message.
4679 @node Buffer Walk Through, More Complex, Writing Defuns, Top
4680 @comment node-name, next, previous, up
4681 @chapter A Few Buffer--Related Functions
4683 In this chapter we study in detail several of the functions used in GNU
4684 Emacs. This is called a ``walk-through''. These functions are used as
4685 examples of Lisp code, but are not imaginary examples; with the
4686 exception of the first, simplified function definition, these functions
4687 show the actual code used in GNU Emacs. You can learn a great deal from
4688 these definitions. The functions described here are all related to
4689 buffers. Later, we will study other functions.
4692 * Finding More:: How to find more information.
4693 * simplified-beginning-of-buffer:: Shows @code{goto-char},
4694 @code{point-min}, and @code{push-mark}.
4695 * mark-whole-buffer:: Almost the same as @code{beginning-of-buffer}.
4696 * append-to-buffer:: Uses @code{save-excursion} and
4697 @code{insert-buffer-substring}.
4698 * Buffer Related Review:: Review.
4699 * Buffer Exercises::
4702 @node Finding More, simplified-beginning-of-buffer, Buffer Walk Through, Buffer Walk Through
4703 @section Finding More Information
4705 @findex describe-function, @r{introduced}
4706 @cindex Find function documentation
4707 In this walk-through, I will describe each new function as we come to
4708 it, sometimes in detail and sometimes briefly. If you are interested,
4709 you can get the full documentation of any Emacs Lisp function at any
4710 time by typing @kbd{C-h f} and then the name of the function (and then
4711 @key{RET}). Similarly, you can get the full documentation for a
4712 variable by typing @kbd{C-h v} and then the name of the variable (and
4715 @cindex Find source of function
4716 @c In version 22, tells location both of C and of Emacs Lisp
4717 Also, @code{describe-function} will tell you the location of the
4718 function definition.
4720 Put point into the name of the file that contains the function and
4721 press the @key{RET} key. In this case, @key{RET} means
4722 @code{push-button} rather than `return' or `enter'. Emacs will take
4723 you directly to the function definition.
4728 If you move point over the file name and press
4729 the @key{RET} key, which in this case means @code{help-follow} rather
4730 than `return' or `enter', Emacs will take you directly to the function
4734 More generally, if you want to see a function in its original source
4735 file, you can use the @code{find-tags} function to jump to it.
4736 @code{find-tags} works with a wide variety of languages, not just
4737 Lisp, and C, and it works with non-programming text as well. For
4738 example, @code{find-tags} will jump to the various nodes in the
4739 Texinfo source file of this document.
4740 The @code{find-tags} function depends on `tags tables' that record
4741 the locations of the functions, variables, and other items to which
4742 @code{find-tags} jumps.
4744 To use the @code{find-tags} command, type @kbd{M-.} (i.e., press the
4745 period key while holding down the @key{META} key, or else type the
4746 @key{ESC} key and then type the period key), and then, at the prompt,
4747 type in the name of the function whose source code you want to see,
4748 such as @code{mark-whole-buffer}, and then type @key{RET}. Emacs will
4749 switch buffers and display the source code for the function on your
4750 screen. To switch back to your current buffer, type @kbd{C-x b
4751 @key{RET}}. (On some keyboards, the @key{META} key is labelled
4754 @c !!! 22.1.1 tags table location in this paragraph
4755 @cindex TAGS table, specifying
4757 Depending on how the initial default values of your copy of Emacs are
4758 set, you may also need to specify the location of your `tags table',
4759 which is a file called @file{TAGS}. For example, if you are
4760 interested in Emacs sources, the tags table you will most likely want,
4761 if it has already been created for you, will be in a subdirectory of
4762 the @file{/usr/local/share/emacs/} directory; thus you would use the
4763 @code{M-x visit-tags-table} command and specify a pathname such as
4764 @file{/usr/local/share/emacs/22.1.1/lisp/TAGS}. If the tags table
4765 has not already been created, you will have to create it yourself. It
4766 will in a file such as @file{/usr/local/src/emacs/src/TAGS}.
4769 To create a @file{TAGS} file in a specific directory, switch to that
4770 directory in Emacs using @kbd{M-x cd} command, or list the directory
4771 with @kbd{C-x d} (@code{dired}). Then run the compile command, with
4772 @w{@code{etags *.el}} as the command to execute:
4775 M-x compile RET etags *.el RET
4778 For more information, see @ref{etags, , Create Your Own @file{TAGS} File}.
4780 After you become more familiar with Emacs Lisp, you will find that you will
4781 frequently use @code{find-tags} to navigate your way around source code;
4782 and you will create your own @file{TAGS} tables.
4784 @cindex Library, as term for `file'
4785 Incidentally, the files that contain Lisp code are conventionally
4786 called @dfn{libraries}. The metaphor is derived from that of a
4787 specialized library, such as a law library or an engineering library,
4788 rather than a general library. Each library, or file, contains
4789 functions that relate to a particular topic or activity, such as
4790 @file{abbrev.el} for handling abbreviations and other typing
4791 shortcuts, and @file{help.el} for on-line help. (Sometimes several
4792 libraries provide code for a single activity, as the various
4793 @file{rmail@dots{}} files provide code for reading electronic mail.)
4794 In @cite{The GNU Emacs Manual}, you will see sentences such as ``The
4795 @kbd{C-h p} command lets you search the standard Emacs Lisp libraries
4796 by topic keywords.''
4798 @node simplified-beginning-of-buffer, mark-whole-buffer, Finding More, Buffer Walk Through
4799 @comment node-name, next, previous, up
4800 @section A Simplified @code{beginning-of-buffer} Definition
4801 @findex simplified-beginning-of-buffer
4803 The @code{beginning-of-buffer} command is a good function to start with
4804 since you are likely to be familiar with it and it is easy to
4805 understand. Used as an interactive command, @code{beginning-of-buffer}
4806 moves the cursor to the beginning of the buffer, leaving the mark at the
4807 previous position. It is generally bound to @kbd{M-<}.
4809 In this section, we will discuss a shortened version of the function
4810 that shows how it is most frequently used. This shortened function
4811 works as written, but it does not contain the code for a complex option.
4812 In another section, we will describe the entire function.
4813 (@xref{beginning-of-buffer, , Complete Definition of
4814 @code{beginning-of-buffer}}.)
4816 Before looking at the code, let's consider what the function
4817 definition has to contain: it must include an expression that makes
4818 the function interactive so it can be called by typing @kbd{M-x
4819 beginning-of-buffer} or by typing a keychord such as @kbd{M-<}; it
4820 must include code to leave a mark at the original position in the
4821 buffer; and it must include code to move the cursor to the beginning
4825 Here is the complete text of the shortened version of the function:
4829 (defun simplified-beginning-of-buffer ()
4830 "Move point to the beginning of the buffer;
4831 leave mark at previous position."
4834 (goto-char (point-min)))
4838 Like all function definitions, this definition has five parts following
4839 the special form @code{defun}:
4843 The name: in this example, @code{simplified-beginning-of-buffer}.
4846 A list of the arguments: in this example, an empty list, @code{()},
4849 The documentation string.
4852 The interactive expression.
4859 In this function definition, the argument list is empty; this means that
4860 this function does not require any arguments. (When we look at the
4861 definition for the complete function, we will see that it may be passed
4862 an optional argument.)
4864 The interactive expression tells Emacs that the function is intended to
4865 be used interactively. In this example, @code{interactive} does not have
4866 an argument because @code{simplified-beginning-of-buffer} does not
4870 The body of the function consists of the two lines:
4875 (goto-char (point-min))
4879 The first of these lines is the expression, @code{(push-mark)}. When
4880 this expression is evaluated by the Lisp interpreter, it sets a mark at
4881 the current position of the cursor, wherever that may be. The position
4882 of this mark is saved in the mark ring.
4884 The next line is @code{(goto-char (point-min))}. This expression
4885 jumps the cursor to the minimum point in the buffer, that is, to the
4886 beginning of the buffer (or to the beginning of the accessible portion
4887 of the buffer if it is narrowed. @xref{Narrowing & Widening, ,
4888 Narrowing and Widening}.)
4890 The @code{push-mark} command sets a mark at the place where the cursor
4891 was located before it was moved to the beginning of the buffer by the
4892 @code{(goto-char (point-min))} expression. Consequently, you can, if
4893 you wish, go back to where you were originally by typing @kbd{C-x C-x}.
4895 That is all there is to the function definition!
4897 @findex describe-function
4898 When you are reading code such as this and come upon an unfamiliar
4899 function, such as @code{goto-char}, you can find out what it does by
4900 using the @code{describe-function} command. To use this command, type
4901 @kbd{C-h f} and then type in the name of the function and press
4902 @key{RET}. The @code{describe-function} command will print the
4903 function's documentation string in a @file{*Help*} window. For
4904 example, the documentation for @code{goto-char} is:
4908 Set point to POSITION, a number or marker.
4909 Beginning of buffer is position (point-min), end is (point-max).
4914 The function's one argument is the desired position.
4917 (The prompt for @code{describe-function} will offer you the symbol
4918 under or preceding the cursor, so you can save typing by positioning
4919 the cursor right over or after the function and then typing @kbd{C-h f
4922 The @code{end-of-buffer} function definition is written in the same way as
4923 the @code{beginning-of-buffer} definition except that the body of the
4924 function contains the expression @code{(goto-char (point-max))} in place
4925 of @code{(goto-char (point-min))}.
4927 @node mark-whole-buffer, append-to-buffer, simplified-beginning-of-buffer, Buffer Walk Through
4928 @comment node-name, next, previous, up
4929 @section The Definition of @code{mark-whole-buffer}
4930 @findex mark-whole-buffer
4932 The @code{mark-whole-buffer} function is no harder to understand than the
4933 @code{simplified-beginning-of-buffer} function. In this case, however,
4934 we will look at the complete function, not a shortened version.
4936 The @code{mark-whole-buffer} function is not as commonly used as the
4937 @code{beginning-of-buffer} function, but is useful nonetheless: it
4938 marks a whole buffer as a region by putting point at the beginning and
4939 a mark at the end of the buffer. It is generally bound to @kbd{C-x
4943 * mark-whole-buffer overview::
4944 * Body of mark-whole-buffer:: Only three lines of code.
4947 @node mark-whole-buffer overview, Body of mark-whole-buffer, mark-whole-buffer, mark-whole-buffer
4949 @unnumberedsubsec An overview of @code{mark-whole-buffer}
4953 In GNU Emacs 22, the code for the complete function looks like this:
4957 (defun mark-whole-buffer ()
4958 "Put point at beginning and mark at end of buffer.
4959 You probably should not use this function in Lisp programs;
4960 it is usually a mistake for a Lisp function to use any subroutine
4961 that uses or sets the mark."
4964 (push-mark (point-max) nil t)
4965 (goto-char (point-min)))
4970 Like all other functions, the @code{mark-whole-buffer} function fits
4971 into the template for a function definition. The template looks like
4976 (defun @var{name-of-function} (@var{argument-list})
4977 "@var{documentation}@dots{}"
4978 (@var{interactive-expression}@dots{})
4983 Here is how the function works: the name of the function is
4984 @code{mark-whole-buffer}; it is followed by an empty argument list,
4985 @samp{()}, which means that the function does not require arguments.
4986 The documentation comes next.
4988 The next line is an @code{(interactive)} expression that tells Emacs
4989 that the function will be used interactively. These details are similar
4990 to the @code{simplified-beginning-of-buffer} function described in the
4994 @node Body of mark-whole-buffer, , mark-whole-buffer overview, mark-whole-buffer
4995 @comment node-name, next, previous, up
4996 @subsection Body of @code{mark-whole-buffer}
4998 The body of the @code{mark-whole-buffer} function consists of three
5005 (push-mark (point-max) nil t)
5006 (goto-char (point-min))
5010 The first of these lines is the expression, @code{(push-mark (point))}.
5012 This line does exactly the same job as the first line of the body of
5013 the @code{simplified-beginning-of-buffer} function, which is written
5014 @code{(push-mark)}. In both cases, the Lisp interpreter sets a mark
5015 at the current position of the cursor.
5017 I don't know why the expression in @code{mark-whole-buffer} is written
5018 @code{(push-mark (point))} and the expression in
5019 @code{beginning-of-buffer} is written @code{(push-mark)}. Perhaps
5020 whoever wrote the code did not know that the arguments for
5021 @code{push-mark} are optional and that if @code{push-mark} is not
5022 passed an argument, the function automatically sets mark at the
5023 location of point by default. Or perhaps the expression was written
5024 so as to parallel the structure of the next line. In any case, the
5025 line causes Emacs to determine the position of point and set a mark
5028 In earlier versions of GNU Emacs, the next line of
5029 @code{mark-whole-buffer} was @code{(push-mark (point-max))}. This
5030 expression sets a mark at the point in the buffer that has the highest
5031 number. This will be the end of the buffer (or, if the buffer is
5032 narrowed, the end of the accessible portion of the buffer.
5033 @xref{Narrowing & Widening, , Narrowing and Widening}, for more about
5034 narrowing.) After this mark has been set, the previous mark, the one
5035 set at point, is no longer set, but Emacs remembers its position, just
5036 as all other recent marks are always remembered. This means that you
5037 can, if you wish, go back to that position by typing @kbd{C-u
5041 In GNU Emacs 22, the @code{(point-max)} is slightly more complicated.
5045 (push-mark (point-max) nil t)
5049 The expression works nearly the same as before. It sets a mark at the
5050 highest numbered place in the buffer that it can. However, in this
5051 version, @code{push-mark} has two additional arguments. The second
5052 argument to @code{push-mark} is @code{nil}. This tells the function
5053 it @emph{should} display a message that says `Mark set' when it pushes
5054 the mark. The third argument is @code{t}. This tells
5055 @code{push-mark} to activate the mark when Transient Mark mode is
5056 turned on. Transient Mark mode highlights the currently active
5057 region. It is often turned off.
5059 Finally, the last line of the function is @code{(goto-char
5060 (point-min)))}. This is written exactly the same way as it is written
5061 in @code{beginning-of-buffer}. The expression moves the cursor to
5062 the minimum point in the buffer, that is, to the beginning of the buffer
5063 (or to the beginning of the accessible portion of the buffer). As a
5064 result of this, point is placed at the beginning of the buffer and mark
5065 is set at the end of the buffer. The whole buffer is, therefore, the
5068 @node append-to-buffer, Buffer Related Review, mark-whole-buffer, Buffer Walk Through
5069 @comment node-name, next, previous, up
5070 @section The Definition of @code{append-to-buffer}
5071 @findex append-to-buffer
5073 The @code{append-to-buffer} command is more complex than the
5074 @code{mark-whole-buffer} command. What it does is copy the region
5075 (that is, the part of the buffer between point and mark) from the
5076 current buffer to a specified buffer.
5079 * append-to-buffer overview::
5080 * append interactive:: A two part interactive expression.
5081 * append-to-buffer body:: Incorporates a @code{let} expression.
5082 * append save-excursion:: How the @code{save-excursion} works.
5085 @node append-to-buffer overview, append interactive, append-to-buffer, append-to-buffer
5087 @unnumberedsubsec An Overview of @code{append-to-buffer}
5090 @findex insert-buffer-substring
5091 The @code{append-to-buffer} command uses the
5092 @code{insert-buffer-substring} function to copy the region.
5093 @code{insert-buffer-substring} is described by its name: it takes a
5094 string of characters from part of a buffer, a ``substring'', and
5095 inserts them into another buffer.
5097 Most of @code{append-to-buffer} is
5098 concerned with setting up the conditions for
5099 @code{insert-buffer-substring} to work: the code must specify both the
5100 buffer to which the text will go, the window it comes from and goes
5101 to, and the region that will be copied.
5104 Here is the complete text of the function:
5108 (defun append-to-buffer (buffer start end)
5109 "Append to specified buffer the text of the region.
5110 It is inserted into that buffer before its point.
5114 When calling from a program, give three arguments:
5115 BUFFER (or buffer name), START and END.
5116 START and END specify the portion of the current buffer to be copied."
5118 (list (read-buffer "Append to buffer: " (other-buffer
5119 (current-buffer) t))
5120 (region-beginning) (region-end)))
5123 (let ((oldbuf (current-buffer)))
5125 (let* ((append-to (get-buffer-create buffer))
5126 (windows (get-buffer-window-list append-to t t))
5128 (set-buffer append-to)
5129 (setq point (point))
5130 (barf-if-buffer-read-only)
5131 (insert-buffer-substring oldbuf start end)
5132 (dolist (window windows)
5133 (when (= (window-point window) point)
5134 (set-window-point window (point))))))))
5138 The function can be understood by looking at it as a series of
5139 filled-in templates.
5141 The outermost template is for the function definition. In this
5142 function, it looks like this (with several slots filled in):
5146 (defun append-to-buffer (buffer start end)
5147 "@var{documentation}@dots{}"
5148 (interactive @dots{})
5153 The first line of the function includes its name and three arguments.
5154 The arguments are the @code{buffer} to which the text will be copied, and
5155 the @code{start} and @code{end} of the region in the current buffer that
5158 The next part of the function is the documentation, which is clear and
5159 complete. As is conventional, the three arguments are written in
5160 upper case so you will notice them easily. Even better, they are
5161 described in the same order as in the argument list.
5163 Note that the documentation distinguishes between a buffer and its
5164 name. (The function can handle either.)
5166 @node append interactive, append-to-buffer body, append-to-buffer overview, append-to-buffer
5167 @comment node-name, next, previous, up
5168 @subsection The @code{append-to-buffer} Interactive Expression
5170 Since the @code{append-to-buffer} function will be used interactively,
5171 the function must have an @code{interactive} expression. (For a
5172 review of @code{interactive}, see @ref{Interactive, , Making a
5173 Function Interactive}.) The expression reads as follows:
5179 "Append to buffer: "
5180 (other-buffer (current-buffer) t))
5187 This expression is not one with letters standing for parts, as
5188 described earlier. Instead, it starts a list with these parts:
5190 The first part of the list is an expression to read the name of a
5191 buffer and return it as a string. That is @code{read-buffer}. The
5192 function requires a prompt as its first argument, @samp{"Append to
5193 buffer: "}. Its second argument tells the command what value to
5194 provide if you don't specify anything.
5196 In this case that second argument is an expression containing the
5197 function @code{other-buffer}, an exception, and a @samp{t}, standing
5200 The first argument to @code{other-buffer}, the exception, is yet
5201 another function, @code{current-buffer}. That is not going to be
5202 returned. The second argument is the symbol for true, @code{t}. that
5203 tells @code{other-buffer} that it may show visible buffers (except in
5204 this case, it will not show the current buffer, which makes sense).
5207 The expression looks like this:
5210 (other-buffer (current-buffer) t)
5213 The second and third arguments to the @code{list} expression are
5214 @code{(region-beginning)} and @code{(region-end)}. These two
5215 functions specify the beginning and end of the text to be appended.
5218 Originally, the command used the letters @samp{B} and @samp{r}.
5219 The whole @code{interactive} expression looked like this:
5222 (interactive "BAppend to buffer:@: \nr")
5226 But when that was done, the default value of the buffer switched to
5227 was invisible. That was not wanted.
5229 (The prompt was separated from the second argument with a newline,
5230 @samp{\n}. It was followed by an @samp{r} that told Emacs to bind the
5231 two arguments that follow the symbol @code{buffer} in the function's
5232 argument list (that is, @code{start} and @code{end}) to the values of
5233 point and mark. That argument worked fine.)
5235 @node append-to-buffer body, append save-excursion, append interactive, append-to-buffer
5236 @comment node-name, next, previous, up
5237 @subsection The Body of @code{append-to-buffer}
5240 in GNU Emacs 22 in /usr/local/src/emacs/lisp/simple.el
5242 (defun append-to-buffer (buffer start end)
5243 "Append to specified buffer the text of the region.
5244 It is inserted into that buffer before its point.
5246 When calling from a program, give three arguments:
5247 BUFFER (or buffer name), START and END.
5248 START and END specify the portion of the current buffer to be copied."
5250 (list (read-buffer "Append to buffer: " (other-buffer (current-buffer) t))
5251 (region-beginning) (region-end)))
5252 (let ((oldbuf (current-buffer)))
5254 (let* ((append-to (get-buffer-create buffer))
5255 (windows (get-buffer-window-list append-to t t))
5257 (set-buffer append-to)
5258 (setq point (point))
5259 (barf-if-buffer-read-only)
5260 (insert-buffer-substring oldbuf start end)
5261 (dolist (window windows)
5262 (when (= (window-point window) point)
5263 (set-window-point window (point))))))))
5266 The body of the @code{append-to-buffer} function begins with @code{let}.
5268 As we have seen before (@pxref{let, , @code{let}}), the purpose of a
5269 @code{let} expression is to create and give initial values to one or
5270 more variables that will only be used within the body of the
5271 @code{let}. This means that such a variable will not be confused with
5272 any variable of the same name outside the @code{let} expression.
5274 We can see how the @code{let} expression fits into the function as a
5275 whole by showing a template for @code{append-to-buffer} with the
5276 @code{let} expression in outline:
5280 (defun append-to-buffer (buffer start end)
5281 "@var{documentation}@dots{}"
5282 (interactive @dots{})
5283 (let ((@var{variable} @var{value}))
5288 The @code{let} expression has three elements:
5292 The symbol @code{let};
5295 A varlist containing, in this case, a single two-element list,
5296 @code{(@var{variable} @var{value})};
5299 The body of the @code{let} expression.
5303 In the @code{append-to-buffer} function, the varlist looks like this:
5306 (oldbuf (current-buffer))
5310 In this part of the @code{let} expression, the one variable,
5311 @code{oldbuf}, is bound to the value returned by the
5312 @code{(current-buffer)} expression. The variable, @code{oldbuf}, is
5313 used to keep track of the buffer in which you are working and from
5314 which you will copy.
5316 The element or elements of a varlist are surrounded by a set of
5317 parentheses so the Lisp interpreter can distinguish the varlist from
5318 the body of the @code{let}. As a consequence, the two-element list
5319 within the varlist is surrounded by a circumscribing set of parentheses.
5320 The line looks like this:
5324 (let ((oldbuf (current-buffer)))
5330 The two parentheses before @code{oldbuf} might surprise you if you did
5331 not realize that the first parenthesis before @code{oldbuf} marks the
5332 boundary of the varlist and the second parenthesis marks the beginning
5333 of the two-element list, @code{(oldbuf (current-buffer))}.
5335 @node append save-excursion, , append-to-buffer body, append-to-buffer
5336 @comment node-name, next, previous, up
5337 @subsection @code{save-excursion} in @code{append-to-buffer}
5339 The body of the @code{let} expression in @code{append-to-buffer}
5340 consists of a @code{save-excursion} expression.
5342 The @code{save-excursion} function saves the locations of point and
5343 mark, and restores them to those positions after the expressions in the
5344 body of the @code{save-excursion} complete execution. In addition,
5345 @code{save-excursion} keeps track of the original buffer, and
5346 restores it. This is how @code{save-excursion} is used in
5347 @code{append-to-buffer}.
5350 @cindex Indentation for formatting
5351 @cindex Formatting convention
5352 Incidentally, it is worth noting here that a Lisp function is normally
5353 formatted so that everything that is enclosed in a multi-line spread is
5354 indented more to the right than the first symbol. In this function
5355 definition, the @code{let} is indented more than the @code{defun}, and
5356 the @code{save-excursion} is indented more than the @code{let}, like
5372 This formatting convention makes it easy to see that the lines in
5373 the body of the @code{save-excursion} are enclosed by the parentheses
5374 associated with @code{save-excursion}, just as the
5375 @code{save-excursion} itself is enclosed by the parentheses associated
5376 with the @code{let}:
5380 (let ((oldbuf (current-buffer)))
5383 (set-buffer @dots{})
5384 (insert-buffer-substring oldbuf start end)
5390 The use of the @code{save-excursion} function can be viewed as a process
5391 of filling in the slots of a template:
5396 @var{first-expression-in-body}
5397 @var{second-expression-in-body}
5399 @var{last-expression-in-body})
5405 In this function, the body of the @code{save-excursion} contains only
5406 one expression, the @code{let*} expression. You know about a
5407 @code{let} function. The @code{let*} function is different. It has a
5408 @samp{*} in its name. It enables Emacs to set each variable in its
5409 varlist in sequence, one after another.
5411 Its critical feature is that variables later in the varlist can make
5412 use of the values to which Emacs set variables earlier in the varlist.
5413 @xref{fwd-para let, , The @code{let*} expression}.
5415 We will skip functions like @code{let*} and focus on two: the
5416 @code{set-buffer} function and the @code{insert-buffer-substring}
5420 In the old days, the @code{set-buffer} expression was simply
5423 (set-buffer (get-buffer-create buffer))
5431 (set-buffer append-to)
5435 @code{append-to} is bound to @code{(get-buffer-create buffer)} earlier
5436 on in the @code{let*} expression. That extra binding would not be
5437 necessary except for that @code{append-to} is used later in the
5438 varlist as an argument to @code{get-buffer-window-list}.
5443 (let ((oldbuf (current-buffer)))
5445 (let* ((append-to (get-buffer-create buffer))
5446 (windows (get-buffer-window-list append-to t t))
5448 (set-buffer append-to)
5449 (setq point (point))
5450 (barf-if-buffer-read-only)
5451 (insert-buffer-substring oldbuf start end)
5452 (dolist (window windows)
5453 (when (= (window-point window) point)
5454 (set-window-point window (point))))))))
5457 The @code{append-to-buffer} function definition inserts text from the
5458 buffer in which you are currently to a named buffer. It happens that
5459 @code{insert-buffer-substring} copies text from another buffer to the
5460 current buffer, just the reverse---that is why the
5461 @code{append-to-buffer} definition starts out with a @code{let} that
5462 binds the local symbol @code{oldbuf} to the value returned by
5463 @code{current-buffer}.
5466 The @code{insert-buffer-substring} expression looks like this:
5469 (insert-buffer-substring oldbuf start end)
5473 The @code{insert-buffer-substring} function copies a string
5474 @emph{from} the buffer specified as its first argument and inserts the
5475 string into the present buffer. In this case, the argument to
5476 @code{insert-buffer-substring} is the value of the variable created
5477 and bound by the @code{let}, namely the value of @code{oldbuf}, which
5478 was the current buffer when you gave the @code{append-to-buffer}
5481 After @code{insert-buffer-substring} has done its work,
5482 @code{save-excursion} will restore the action to the original buffer
5483 and @code{append-to-buffer} will have done its job.
5486 Written in skeletal form, the workings of the body look like this:
5490 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5491 (save-excursion ; @r{Keep track of buffer.}
5493 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})
5495 @var{change-back-to-original-buffer-when-finished}
5496 @var{let-the-local-meaning-of-}@code{oldbuf}@var{-disappear-when-finished}
5500 In summary, @code{append-to-buffer} works as follows: it saves the
5501 value of the current buffer in the variable called @code{oldbuf}. It
5502 gets the new buffer (creating one if need be) and switches Emacs'
5503 attention to it. Using the value of @code{oldbuf}, it inserts the
5504 region of text from the old buffer into the new buffer; and then using
5505 @code{save-excursion}, it brings you back to your original buffer.
5507 In looking at @code{append-to-buffer}, you have explored a fairly
5508 complex function. It shows how to use @code{let} and
5509 @code{save-excursion}, and how to change to and come back from another
5510 buffer. Many function definitions use @code{let},
5511 @code{save-excursion}, and @code{set-buffer} this way.
5513 @node Buffer Related Review, Buffer Exercises, append-to-buffer, Buffer Walk Through
5514 @comment node-name, next, previous, up
5517 Here is a brief summary of the various functions discussed in this chapter.
5520 @item describe-function
5521 @itemx describe-variable
5522 Print the documentation for a function or variable.
5523 Conventionally bound to @kbd{C-h f} and @kbd{C-h v}.
5526 Find the file containing the source for a function or variable and
5527 switch buffers to it, positioning point at the beginning of the item.
5528 Conventionally bound to @kbd{M-.} (that's a period following the
5531 @item save-excursion
5532 Save the location of point and mark and restore their values after the
5533 arguments to @code{save-excursion} have been evaluated. Also, remember
5534 the current buffer and return to it.
5537 Set mark at a location and record the value of the previous mark on the
5538 mark ring. The mark is a location in the buffer that will keep its
5539 relative position even if text is added to or removed from the buffer.
5542 Set point to the location specified by the value of the argument, which
5543 can be a number, a marker, or an expression that returns the number of
5544 a position, such as @code{(point-min)}.
5546 @item insert-buffer-substring
5547 Copy a region of text from a buffer that is passed to the function as
5548 an argument and insert the region into the current buffer.
5550 @item mark-whole-buffer
5551 Mark the whole buffer as a region. Normally bound to @kbd{C-x h}.
5554 Switch the attention of Emacs to another buffer, but do not change the
5555 window being displayed. Used when the program rather than a human is
5556 to work on a different buffer.
5558 @item get-buffer-create
5560 Find a named buffer or create one if a buffer of that name does not
5561 exist. The @code{get-buffer} function returns @code{nil} if the named
5562 buffer does not exist.
5566 @node Buffer Exercises, , Buffer Related Review, Buffer Walk Through
5571 Write your own @code{simplified-end-of-buffer} function definition;
5572 then test it to see whether it works.
5575 Use @code{if} and @code{get-buffer} to write a function that prints a
5576 message telling you whether a buffer exists.
5579 Using @code{find-tag}, find the source for the @code{copy-to-buffer}
5583 @node More Complex, Narrowing & Widening, Buffer Walk Through, Top
5584 @comment node-name, next, previous, up
5585 @chapter A Few More Complex Functions
5587 In this chapter, we build on what we have learned in previous chapters
5588 by looking at more complex functions. The @code{copy-to-buffer}
5589 function illustrates use of two @code{save-excursion} expressions in
5590 one definition, while the @code{insert-buffer} function illustrates
5591 use of an asterisk in an @code{interactive} expression, use of
5592 @code{or}, and the important distinction between a name and the object
5593 to which the name refers.
5596 * copy-to-buffer:: With @code{set-buffer}, @code{get-buffer-create}.
5597 * insert-buffer:: Read-only, and with @code{or}.
5598 * beginning-of-buffer:: Shows @code{goto-char},
5599 @code{point-min}, and @code{push-mark}.
5600 * Second Buffer Related Review::
5601 * optional Exercise::
5604 @node copy-to-buffer, insert-buffer, More Complex, More Complex
5605 @comment node-name, next, previous, up
5606 @section The Definition of @code{copy-to-buffer}
5607 @findex copy-to-buffer
5609 After understanding how @code{append-to-buffer} works, it is easy to
5610 understand @code{copy-to-buffer}. This function copies text into a
5611 buffer, but instead of adding to the second buffer, it replaces all the
5612 previous text in the second buffer.
5615 The body of @code{copy-to-buffer} looks like this,
5620 (interactive "BCopy to buffer: \nr")
5621 (let ((oldbuf (current-buffer)))
5622 (with-current-buffer (get-buffer-create buffer)
5623 (barf-if-buffer-read-only)
5626 (insert-buffer-substring oldbuf start end)))))
5630 The @code{copy-to-buffer} function has a simpler @code{interactive}
5631 expression than @code{append-to-buffer}.
5634 The definition then says
5637 (with-current-buffer (get-buffer-create buffer) @dots{}
5640 First, look at the earliest inner expression; that is evaluated first.
5641 That expression starts with @code{get-buffer-create buffer}. The
5642 function tells the computer to use the buffer with the name specified
5643 as the one to which you are copying, or if such a buffer does not
5644 exist, to create it. Then, the @code{with-current-buffer} function
5645 evaluates its body with that buffer temporarily current.
5647 (This demonstrates another way to shift the computer's attention but
5648 not the user's. The @code{append-to-buffer} function showed how to do
5649 the same with @code{save-excursion} and @code{set-buffer}.
5650 @code{with-current-buffer} is a newer, and arguably easier,
5653 The @code{barf-if-buffer-read-only} function sends you an error
5654 message saying the buffer is read-only if you cannot modify it.
5656 The next line has the @code{erase-buffer} function as its sole
5657 contents. That function erases the buffer.
5659 Finally, the last two lines contain the @code{save-excursion}
5660 expression with @code{insert-buffer-substring} as its body.
5661 The @code{insert-buffer-substring} expression copies the text from
5662 the buffer you are in (and you have not seen the computer shift its
5663 attention, so you don't know that that buffer is now called
5666 Incidentally, this is what is meant by `replacement'. To replace text,
5667 Emacs erases the previous text and then inserts new text.
5670 In outline, the body of @code{copy-to-buffer} looks like this:
5674 (let (@var{bind-}@code{oldbuf}@var{-to-value-of-}@code{current-buffer})
5675 (@var{with-the-buffer-you-are-copying-to}
5676 (@var{but-do-not-erase-or-copy-to-a-read-only-buffer})
5679 @var{insert-substring-from-}@code{oldbuf}@var{-into-buffer})))
5683 @node insert-buffer, beginning-of-buffer, copy-to-buffer, More Complex
5684 @comment node-name, next, previous, up
5685 @section The Definition of @code{insert-buffer}
5686 @findex insert-buffer
5688 @code{insert-buffer} is yet another buffer-related function. This
5689 command copies another buffer @emph{into} the current buffer. It is the
5690 reverse of @code{append-to-buffer} or @code{copy-to-buffer}, since they
5691 copy a region of text @emph{from} the current buffer to another buffer.
5693 Here is a discussion based on the original code. The code was
5694 simplified in 2003 and is harder to understand.
5696 (@xref{New insert-buffer, , New Body for @code{insert-buffer}}, to see
5697 a discussion of the new body.)
5699 In addition, this code illustrates the use of @code{interactive} with a
5700 buffer that might be @dfn{read-only} and the important distinction
5701 between the name of an object and the object actually referred to.
5704 * insert-buffer code::
5705 * insert-buffer interactive:: When you can read, but not write.
5706 * insert-buffer body:: The body has an @code{or} and a @code{let}.
5707 * if & or:: Using an @code{if} instead of an @code{or}.
5708 * Insert or:: How the @code{or} expression works.
5709 * Insert let:: Two @code{save-excursion} expressions.
5710 * New insert-buffer::
5713 @node insert-buffer code, insert-buffer interactive, insert-buffer, insert-buffer
5715 @unnumberedsubsec The Code for @code{insert-buffer}
5719 Here is the earlier code:
5723 (defun insert-buffer (buffer)
5724 "Insert after point the contents of BUFFER.
5725 Puts mark after the inserted text.
5726 BUFFER may be a buffer or a buffer name."
5727 (interactive "*bInsert buffer:@: ")
5730 (or (bufferp buffer)
5731 (setq buffer (get-buffer buffer)))
5732 (let (start end newmark)
5736 (setq start (point-min) end (point-max)))
5739 (insert-buffer-substring buffer start end)
5740 (setq newmark (point)))
5741 (push-mark newmark)))
5746 As with other function definitions, you can use a template to see an
5747 outline of the function:
5751 (defun insert-buffer (buffer)
5752 "@var{documentation}@dots{}"
5753 (interactive "*bInsert buffer:@: ")
5758 @node insert-buffer interactive, insert-buffer body, insert-buffer code, insert-buffer
5759 @comment node-name, next, previous, up
5760 @subsection The Interactive Expression in @code{insert-buffer}
5761 @findex interactive, @r{example use of}
5763 In @code{insert-buffer}, the argument to the @code{interactive}
5764 declaration has two parts, an asterisk, @samp{*}, and @samp{bInsert
5768 * Read-only buffer:: When a buffer cannot be modified.
5769 * b for interactive:: An existing buffer or else its name.
5772 @node Read-only buffer, b for interactive, insert-buffer interactive, insert-buffer interactive
5773 @comment node-name, next, previous, up
5774 @unnumberedsubsubsec A Read-only Buffer
5775 @cindex Read-only buffer
5776 @cindex Asterisk for read-only buffer
5777 @findex * @r{for read-only buffer}
5779 The asterisk is for the situation when the current buffer is a
5780 read-only buffer---a buffer that cannot be modified. If
5781 @code{insert-buffer} is called when the current buffer is read-only, a
5782 message to this effect is printed in the echo area and the terminal
5783 may beep or blink at you; you will not be permitted to insert anything
5784 into current buffer. The asterisk does not need to be followed by a
5785 newline to separate it from the next argument.
5787 @node b for interactive, , Read-only buffer, insert-buffer interactive
5788 @comment node-name, next, previous, up
5789 @unnumberedsubsubsec @samp{b} in an Interactive Expression
5791 The next argument in the interactive expression starts with a lower
5792 case @samp{b}. (This is different from the code for
5793 @code{append-to-buffer}, which uses an upper-case @samp{B}.
5794 @xref{append-to-buffer, , The Definition of @code{append-to-buffer}}.)
5795 The lower-case @samp{b} tells the Lisp interpreter that the argument
5796 for @code{insert-buffer} should be an existing buffer or else its
5797 name. (The upper-case @samp{B} option provides for the possibility
5798 that the buffer does not exist.) Emacs will prompt you for the name
5799 of the buffer, offering you a default buffer, with name completion
5800 enabled. If the buffer does not exist, you receive a message that
5801 says ``No match''; your terminal may beep at you as well.
5803 The new and simplified code generates a list for @code{interactive}.
5804 It uses the @code{barf-if-buffer-read-only} and @code{read-buffer}
5805 functions with which we are already familiar and the @code{progn}
5806 special form with which we are not. (It will be described later.)
5808 @node insert-buffer body, if & or, insert-buffer interactive, insert-buffer
5809 @comment node-name, next, previous, up
5810 @subsection The Body of the @code{insert-buffer} Function
5812 The body of the @code{insert-buffer} function has two major parts: an
5813 @code{or} expression and a @code{let} expression. The purpose of the
5814 @code{or} expression is to ensure that the argument @code{buffer} is
5815 bound to a buffer and not just the name of a buffer. The body of the
5816 @code{let} expression contains the code which copies the other buffer
5817 into the current buffer.
5820 In outline, the two expressions fit into the @code{insert-buffer}
5825 (defun insert-buffer (buffer)
5826 "@var{documentation}@dots{}"
5827 (interactive "*bInsert buffer:@: ")
5832 (let (@var{varlist})
5833 @var{body-of-}@code{let}@dots{} )
5837 To understand how the @code{or} expression ensures that the argument
5838 @code{buffer} is bound to a buffer and not to the name of a buffer, it
5839 is first necessary to understand the @code{or} function.
5841 Before doing this, let me rewrite this part of the function using
5842 @code{if} so that you can see what is done in a manner that will be familiar.
5844 @node if & or, Insert or, insert-buffer body, insert-buffer
5845 @comment node-name, next, previous, up
5846 @subsection @code{insert-buffer} With an @code{if} Instead of an @code{or}
5848 The job to be done is to make sure the value of @code{buffer} is a
5849 buffer itself and not the name of a buffer. If the value is the name,
5850 then the buffer itself must be got.
5852 You can imagine yourself at a conference where an usher is wandering
5853 around holding a list with your name on it and looking for you: the
5854 usher is ``bound'' to your name, not to you; but when the usher finds
5855 you and takes your arm, the usher becomes ``bound'' to you.
5858 In Lisp, you might describe this situation like this:
5862 (if (not (holding-on-to-guest))
5863 (find-and-take-arm-of-guest))
5867 We want to do the same thing with a buffer---if we do not have the
5868 buffer itself, we want to get it.
5871 Using a predicate called @code{bufferp} that tells us whether we have a
5872 buffer (rather than its name), we can write the code like this:
5876 (if (not (bufferp buffer)) ; @r{if-part}
5877 (setq buffer (get-buffer buffer))) ; @r{then-part}
5882 Here, the true-or-false-test of the @code{if} expression is
5883 @w{@code{(not (bufferp buffer))}}; and the then-part is the expression
5884 @w{@code{(setq buffer (get-buffer buffer))}}.
5886 In the test, the function @code{bufferp} returns true if its argument is
5887 a buffer---but false if its argument is the name of the buffer. (The
5888 last character of the function name @code{bufferp} is the character
5889 @samp{p}; as we saw earlier, such use of @samp{p} is a convention that
5890 indicates that the function is a predicate, which is a term that means
5891 that the function will determine whether some property is true or false.
5892 @xref{Wrong Type of Argument, , Using the Wrong Type Object as an
5896 The function @code{not} precedes the expression @code{(bufferp buffer)},
5897 so the true-or-false-test looks like this:
5900 (not (bufferp buffer))
5904 @code{not} is a function that returns true if its argument is false
5905 and false if its argument is true. So if @code{(bufferp buffer)}
5906 returns true, the @code{not} expression returns false and vice-verse:
5907 what is ``not true'' is false and what is ``not false'' is true.
5909 Using this test, the @code{if} expression works as follows: when the
5910 value of the variable @code{buffer} is actually a buffer rather than
5911 its name, the true-or-false-test returns false and the @code{if}
5912 expression does not evaluate the then-part. This is fine, since we do
5913 not need to do anything to the variable @code{buffer} if it really is
5916 On the other hand, when the value of @code{buffer} is not a buffer
5917 itself, but the name of a buffer, the true-or-false-test returns true
5918 and the then-part of the expression is evaluated. In this case, the
5919 then-part is @code{(setq buffer (get-buffer buffer))}. This
5920 expression uses the @code{get-buffer} function to return an actual
5921 buffer itself, given its name. The @code{setq} then sets the variable
5922 @code{buffer} to the value of the buffer itself, replacing its previous
5923 value (which was the name of the buffer).
5925 @node Insert or, Insert let, if & or, insert-buffer
5926 @comment node-name, next, previous, up
5927 @subsection The @code{or} in the Body
5929 The purpose of the @code{or} expression in the @code{insert-buffer}
5930 function is to ensure that the argument @code{buffer} is bound to a
5931 buffer and not just to the name of a buffer. The previous section shows
5932 how the job could have been done using an @code{if} expression.
5933 However, the @code{insert-buffer} function actually uses @code{or}.
5934 To understand this, it is necessary to understand how @code{or} works.
5937 An @code{or} function can have any number of arguments. It evaluates
5938 each argument in turn and returns the value of the first of its
5939 arguments that is not @code{nil}. Also, and this is a crucial feature
5940 of @code{or}, it does not evaluate any subsequent arguments after
5941 returning the first non-@code{nil} value.
5944 The @code{or} expression looks like this:
5948 (or (bufferp buffer)
5949 (setq buffer (get-buffer buffer)))
5954 The first argument to @code{or} is the expression @code{(bufferp buffer)}.
5955 This expression returns true (a non-@code{nil} value) if the buffer is
5956 actually a buffer, and not just the name of a buffer. In the @code{or}
5957 expression, if this is the case, the @code{or} expression returns this
5958 true value and does not evaluate the next expression---and this is fine
5959 with us, since we do not want to do anything to the value of
5960 @code{buffer} if it really is a buffer.
5962 On the other hand, if the value of @code{(bufferp buffer)} is @code{nil},
5963 which it will be if the value of @code{buffer} is the name of a buffer,
5964 the Lisp interpreter evaluates the next element of the @code{or}
5965 expression. This is the expression @code{(setq buffer (get-buffer
5966 buffer))}. This expression returns a non-@code{nil} value, which
5967 is the value to which it sets the variable @code{buffer}---and this
5968 value is a buffer itself, not the name of a buffer.
5970 The result of all this is that the symbol @code{buffer} is always
5971 bound to a buffer itself rather than to the name of a buffer. All
5972 this is necessary because the @code{set-buffer} function in a
5973 following line only works with a buffer itself, not with the name to a
5977 Incidentally, using @code{or}, the situation with the usher would be
5981 (or (holding-on-to-guest) (find-and-take-arm-of-guest))
5984 @node Insert let, New insert-buffer, Insert or, insert-buffer
5985 @comment node-name, next, previous, up
5986 @subsection The @code{let} Expression in @code{insert-buffer}
5988 After ensuring that the variable @code{buffer} refers to a buffer itself
5989 and not just to the name of a buffer, the @code{insert-buffer function}
5990 continues with a @code{let} expression. This specifies three local
5991 variables, @code{start}, @code{end}, and @code{newmark} and binds them
5992 to the initial value @code{nil}. These variables are used inside the
5993 remainder of the @code{let} and temporarily hide any other occurrence of
5994 variables of the same name in Emacs until the end of the @code{let}.
5997 The body of the @code{let} contains two @code{save-excursion}
5998 expressions. First, we will look at the inner @code{save-excursion}
5999 expression in detail. The expression looks like this:
6005 (setq start (point-min) end (point-max)))
6010 The expression @code{(set-buffer buffer)} changes Emacs' attention
6011 from the current buffer to the one from which the text will copied.
6012 In that buffer, the variables @code{start} and @code{end} are set to
6013 the beginning and end of the buffer, using the commands
6014 @code{point-min} and @code{point-max}. Note that we have here an
6015 illustration of how @code{setq} is able to set two variables in the
6016 same expression. The first argument of @code{setq} is set to the
6017 value of its second, and its third argument is set to the value of its
6020 After the body of the inner @code{save-excursion} is evaluated, the
6021 @code{save-excursion} restores the original buffer, but @code{start} and
6022 @code{end} remain set to the values of the beginning and end of the
6023 buffer from which the text will be copied.
6026 The outer @code{save-excursion} expression looks like this:
6031 (@var{inner-}@code{save-excursion}@var{-expression}
6032 (@var{go-to-new-buffer-and-set-}@code{start}@var{-and-}@code{end})
6033 (insert-buffer-substring buffer start end)
6034 (setq newmark (point)))
6039 The @code{insert-buffer-substring} function copies the text
6040 @emph{into} the current buffer @emph{from} the region indicated by
6041 @code{start} and @code{end} in @code{buffer}. Since the whole of the
6042 second buffer lies between @code{start} and @code{end}, the whole of
6043 the second buffer is copied into the buffer you are editing. Next,
6044 the value of point, which will be at the end of the inserted text, is
6045 recorded in the variable @code{newmark}.
6047 After the body of the outer @code{save-excursion} is evaluated, point
6048 and mark are relocated to their original places.
6050 However, it is convenient to locate a mark at the end of the newly
6051 inserted text and locate point at its beginning. The @code{newmark}
6052 variable records the end of the inserted text. In the last line of
6053 the @code{let} expression, the @code{(push-mark newmark)} expression
6054 function sets a mark to this location. (The previous location of the
6055 mark is still accessible; it is recorded on the mark ring and you can
6056 go back to it with @kbd{C-u C-@key{SPC}}.) Meanwhile, point is
6057 located at the beginning of the inserted text, which is where it was
6058 before you called the insert function, the position of which was saved
6059 by the first @code{save-excursion}.
6062 The whole @code{let} expression looks like this:
6066 (let (start end newmark)
6070 (setq start (point-min) end (point-max)))
6071 (insert-buffer-substring buffer start end)
6072 (setq newmark (point)))
6073 (push-mark newmark))
6077 Like the @code{append-to-buffer} function, the @code{insert-buffer}
6078 function uses @code{let}, @code{save-excursion}, and
6079 @code{set-buffer}. In addition, the function illustrates one way to
6080 use @code{or}. All these functions are building blocks that we will
6081 find and use again and again.
6083 @node New insert-buffer, , Insert let, insert-buffer
6084 @comment node-name, next, previous, up
6085 @subsection New Body for @code{insert-buffer}
6086 @findex insert-buffer, new version body
6087 @findex new version body for insert-buffer
6089 The body in the GNU Emacs 22 version is more confusing than the original.
6092 It consists of two expressions,
6098 (insert-buffer-substring (get-buffer buffer))
6106 except, and this is what confuses novices, very important work is done
6107 inside the @code{push-mark} expression.
6109 The @code{get-buffer} function returns a buffer with the name
6110 provided. You will note that the function is @emph{not} called
6111 @code{get-buffer-create}; it does not create a buffer if one does not
6112 already exist. The buffer returned by @code{get-buffer}, an existing
6113 buffer, is passed to @code{insert-buffer-substring}, which inserts the
6114 whole of the buffer (since you did not specify anything else).
6116 The location into which the buffer is inserted is recorded by
6117 @code{push-mark}. Then the function returns @code{nil}, the value of
6118 its last command. Put another way, the @code{insert-buffer} function
6119 exists only to produce a side effect, inserting another buffer, not to
6122 @node beginning-of-buffer, Second Buffer Related Review, insert-buffer, More Complex
6123 @comment node-name, next, previous, up
6124 @section Complete Definition of @code{beginning-of-buffer}
6125 @findex beginning-of-buffer
6127 The basic structure of the @code{beginning-of-buffer} function has
6128 already been discussed. (@xref{simplified-beginning-of-buffer, , A
6129 Simplified @code{beginning-of-buffer} Definition}.)
6130 This section describes the complex part of the definition.
6132 As previously described, when invoked without an argument,
6133 @code{beginning-of-buffer} moves the cursor to the beginning of the
6134 buffer (in truth, the beginning of the accessible portion of the
6135 buffer), leaving the mark at the previous position. However, when the
6136 command is invoked with a number between one and ten, the function
6137 considers that number to be a fraction of the length of the buffer,
6138 measured in tenths, and Emacs moves the cursor that fraction of the
6139 way from the beginning of the buffer. Thus, you can either call this
6140 function with the key command @kbd{M-<}, which will move the cursor to
6141 the beginning of the buffer, or with a key command such as @kbd{C-u 7
6142 M-<} which will move the cursor to a point 70% of the way through the
6143 buffer. If a number bigger than ten is used for the argument, it
6144 moves to the end of the buffer.
6146 The @code{beginning-of-buffer} function can be called with or without an
6147 argument. The use of the argument is optional.
6150 * Optional Arguments::
6151 * beginning-of-buffer opt arg:: Example with optional argument.
6152 * beginning-of-buffer complete::
6155 @node Optional Arguments, beginning-of-buffer opt arg, beginning-of-buffer, beginning-of-buffer
6156 @subsection Optional Arguments
6158 Unless told otherwise, Lisp expects that a function with an argument in
6159 its function definition will be called with a value for that argument.
6160 If that does not happen, you get an error and a message that says
6161 @samp{Wrong number of arguments}.
6163 @cindex Optional arguments
6166 However, optional arguments are a feature of Lisp: a particular
6167 @dfn{keyword} is used to tell the Lisp interpreter that an argument is
6168 optional. The keyword is @code{&optional}. (The @samp{&} in front of
6169 @samp{optional} is part of the keyword.) In a function definition, if
6170 an argument follows the keyword @code{&optional}, no value need be
6171 passed to that argument when the function is called.
6174 The first line of the function definition of @code{beginning-of-buffer}
6175 therefore looks like this:
6178 (defun beginning-of-buffer (&optional arg)
6182 In outline, the whole function looks like this:
6186 (defun beginning-of-buffer (&optional arg)
6187 "@var{documentation}@dots{}"
6189 (or (@var{is-the-argument-a-cons-cell} arg)
6190 (and @var{are-both-transient-mark-mode-and-mark-active-true})
6192 (let (@var{determine-size-and-set-it})
6194 (@var{if-there-is-an-argument}
6195 @var{figure-out-where-to-go}
6202 The function is similar to the @code{simplified-beginning-of-buffer}
6203 function except that the @code{interactive} expression has @code{"P"}
6204 as an argument and the @code{goto-char} function is followed by an
6205 if-then-else expression that figures out where to put the cursor if
6206 there is an argument that is not a cons cell.
6208 (Since I do not explain a cons cell for many more chapters, please
6209 consider ignoring the function @code{consp}. @xref{List
6210 Implementation, , How Lists are Implemented}, and @ref{Cons Cell Type,
6211 , Cons Cell and List Types, elisp, The GNU Emacs Lisp Reference
6214 The @code{"P"} in the @code{interactive} expression tells Emacs to
6215 pass a prefix argument, if there is one, to the function in raw form.
6216 A prefix argument is made by typing the @key{META} key followed by a
6217 number, or by typing @kbd{C-u} and then a number. (If you don't type
6218 a number, @kbd{C-u} defaults to a cons cell with a 4. A lowercase
6219 @code{"p"} in the @code{interactive} expression causes the function to
6220 convert a prefix arg to a number.)
6222 The true-or-false-test of the @code{if} expression looks complex, but
6223 it is not: it checks whether @code{arg} has a value that is not
6224 @code{nil} and whether it is a cons cell. (That is what @code{consp}
6225 does; it checks whether its argument is a cons cell.) If @code{arg}
6226 has a value that is not @code{nil} (and is not a cons cell), which
6227 will be the case if @code{beginning-of-buffer} is called with a
6228 numeric argument, then this true-or-false-test will return true and
6229 the then-part of the @code{if} expression will be evaluated. On the
6230 other hand, if @code{beginning-of-buffer} is not called with an
6231 argument, the value of @code{arg} will be @code{nil} and the else-part
6232 of the @code{if} expression will be evaluated. The else-part is
6233 simply @code{point-min}, and when this is the outcome, the whole
6234 @code{goto-char} expression is @code{(goto-char (point-min))}, which
6235 is how we saw the @code{beginning-of-buffer} function in its
6238 @node beginning-of-buffer opt arg, beginning-of-buffer complete, Optional Arguments, beginning-of-buffer
6239 @subsection @code{beginning-of-buffer} with an Argument
6241 When @code{beginning-of-buffer} is called with an argument, an
6242 expression is evaluated which calculates what value to pass to
6243 @code{goto-char}. This expression is rather complicated at first sight.
6244 It includes an inner @code{if} expression and much arithmetic. It looks
6249 (if (> (buffer-size) 10000)
6250 ;; @r{Avoid overflow for large buffer sizes!}
6251 (* (prefix-numeric-value arg)
6256 size (prefix-numeric-value arg))) 10)))
6261 * Disentangle beginning-of-buffer::
6262 * Large buffer case::
6263 * Small buffer case::
6266 @node Disentangle beginning-of-buffer, Large buffer case, beginning-of-buffer opt arg, beginning-of-buffer opt arg
6268 @unnumberedsubsubsec Disentangle @code{beginning-of-buffer}
6271 Like other complex-looking expressions, the conditional expression
6272 within @code{beginning-of-buffer} can be disentangled by looking at it
6273 as parts of a template, in this case, the template for an if-then-else
6274 expression. In skeletal form, the expression looks like this:
6278 (if (@var{buffer-is-large}
6279 @var{divide-buffer-size-by-10-and-multiply-by-arg}
6280 @var{else-use-alternate-calculation}
6284 The true-or-false-test of this inner @code{if} expression checks the
6285 size of the buffer. The reason for this is that the old version 18
6286 Emacs used numbers that are no bigger than eight million or so and in
6287 the computation that followed, the programmer feared that Emacs might
6288 try to use over-large numbers if the buffer were large. The term
6289 `overflow', mentioned in the comment, means numbers that are over
6290 large. More recent versions of Emacs use larger numbers, but this
6291 code has not been touched, if only because people now look at buffers
6292 that are far, far larger than ever before.
6294 There are two cases: if the buffer is large and if it is not.
6296 @node Large buffer case, Small buffer case, Disentangle beginning-of-buffer, beginning-of-buffer opt arg
6297 @comment node-name, next, previous, up
6298 @unnumberedsubsubsec What happens in a large buffer
6300 In @code{beginning-of-buffer}, the inner @code{if} expression tests
6301 whether the size of the buffer is greater than 10,000 characters. To do
6302 this, it uses the @code{>} function and the computation of @code{size}
6303 that comes from the let expression.
6305 In the old days, the function @code{buffer-size} was used. Not only
6306 was that function called several times, it gave the size of the whole
6307 buffer, not the accessible part. The computation makes much more
6308 sense when it handles just the accessible part. (@xref{Narrowing &
6309 Widening, , Narrowing and Widening}, for more information on focusing
6310 attention to an `accessible' part.)
6313 The line looks like this:
6321 When the buffer is large, the then-part of the @code{if} expression is
6322 evaluated. It reads like this (after formatting for easy reading):
6327 (prefix-numeric-value arg)
6333 This expression is a multiplication, with two arguments to the function
6336 The first argument is @code{(prefix-numeric-value arg)}. When
6337 @code{"P"} is used as the argument for @code{interactive}, the value
6338 passed to the function as its argument is passed a ``raw prefix
6339 argument'', and not a number. (It is a number in a list.) To perform
6340 the arithmetic, a conversion is necessary, and
6341 @code{prefix-numeric-value} does the job.
6343 @findex / @r{(division)}
6345 The second argument is @code{(/ size 10)}. This expression divides
6346 the numeric value by ten --- the numeric value of the size of the
6347 accessible portion of the buffer. This produces a number that tells
6348 how many characters make up one tenth of the buffer size. (In Lisp,
6349 @code{/} is used for division, just as @code{*} is used for
6353 In the multiplication expression as a whole, this amount is multiplied
6354 by the value of the prefix argument---the multiplication looks like this:
6358 (* @var{numeric-value-of-prefix-arg}
6359 @var{number-of-characters-in-one-tenth-of-the-accessible-buffer})
6364 If, for example, the prefix argument is @samp{7}, the one-tenth value
6365 will be multiplied by 7 to give a position 70% of the way through.
6368 The result of all this is that if the accessible portion of the buffer
6369 is large, the @code{goto-char} expression reads like this:
6373 (goto-char (* (prefix-numeric-value arg)
6378 This puts the cursor where we want it.
6380 @node Small buffer case, , Large buffer case, beginning-of-buffer opt arg
6381 @comment node-name, next, previous, up
6382 @unnumberedsubsubsec What happens in a small buffer
6384 If the buffer contains fewer than 10,000 characters, a slightly
6385 different computation is performed. You might think this is not
6386 necessary, since the first computation could do the job. However, in
6387 a small buffer, the first method may not put the cursor on exactly the
6388 desired line; the second method does a better job.
6391 The code looks like this:
6393 @c Keep this on one line.
6395 (/ (+ 10 (* size (prefix-numeric-value arg))) 10))
6400 This is code in which you figure out what happens by discovering how the
6401 functions are embedded in parentheses. It is easier to read if you
6402 reformat it with each expression indented more deeply than its
6403 enclosing expression:
6411 (prefix-numeric-value arg)))
6418 Looking at parentheses, we see that the innermost operation is
6419 @code{(prefix-numeric-value arg)}, which converts the raw argument to
6420 a number. In the following expression, this number is multiplied by
6421 the size of the accessible portion of the buffer:
6424 (* size (prefix-numeric-value arg))
6428 This multiplication creates a number that may be larger than the size of
6429 the buffer---seven times larger if the argument is 7, for example. Ten
6430 is then added to this number and finally the large number is divided by
6431 ten to provide a value that is one character larger than the percentage
6432 position in the buffer.
6434 The number that results from all this is passed to @code{goto-char} and
6435 the cursor is moved to that point.
6438 @node beginning-of-buffer complete, , beginning-of-buffer opt arg, beginning-of-buffer
6439 @comment node-name, next, previous, up
6440 @subsection The Complete @code{beginning-of-buffer}
6443 Here is the complete text of the @code{beginning-of-buffer} function:
6449 (defun beginning-of-buffer (&optional arg)
6450 "Move point to the beginning of the buffer;
6451 leave mark at previous position.
6452 With \\[universal-argument] prefix,
6453 do not set mark at previous position.
6455 put point N/10 of the way from the beginning.
6457 If the buffer is narrowed,
6458 this command uses the beginning and size
6459 of the accessible part of the buffer.
6463 Don't use this command in Lisp programs!
6464 \(goto-char (point-min)) is faster
6465 and avoids clobbering the mark."
6468 (and transient-mark-mode mark-active)
6472 (let ((size (- (point-max) (point-min))))
6473 (goto-char (if (and arg (not (consp arg)))
6476 ;; Avoid overflow for large buffer sizes!
6477 (* (prefix-numeric-value arg)
6479 (/ (+ 10 (* size (prefix-numeric-value arg)))
6482 (if arg (forward-line 1)))
6487 From before GNU Emacs 22
6490 (defun beginning-of-buffer (&optional arg)
6491 "Move point to the beginning of the buffer;
6492 leave mark at previous position.
6493 With arg N, put point N/10 of the way
6494 from the true beginning.
6497 Don't use this in Lisp programs!
6498 \(goto-char (point-min)) is faster
6499 and does not set the mark."
6506 (if (> (buffer-size) 10000)
6507 ;; @r{Avoid overflow for large buffer sizes!}
6508 (* (prefix-numeric-value arg)
6509 (/ (buffer-size) 10))
6512 (/ (+ 10 (* (buffer-size)
6513 (prefix-numeric-value arg)))
6516 (if arg (forward-line 1)))
6522 Except for two small points, the previous discussion shows how this
6523 function works. The first point deals with a detail in the
6524 documentation string, and the second point concerns the last line of
6528 In the documentation string, there is reference to an expression:
6531 \\[universal-argument]
6535 A @samp{\\} is used before the first square bracket of this
6536 expression. This @samp{\\} tells the Lisp interpreter to substitute
6537 whatever key is currently bound to the @samp{[@dots{}]}. In the case
6538 of @code{universal-argument}, that is usually @kbd{C-u}, but it might
6539 be different. (@xref{Documentation Tips, , Tips for Documentation
6540 Strings, elisp, The GNU Emacs Lisp Reference Manual}, for more
6544 Finally, the last line of the @code{beginning-of-buffer} command says
6545 to move point to the beginning of the next line if the command is
6546 invoked with an argument:
6549 (if arg (forward-line 1)))
6553 This puts the cursor at the beginning of the first line after the
6554 appropriate tenths position in the buffer. This is a flourish that
6555 means that the cursor is always located @emph{at least} the requested
6556 tenths of the way through the buffer, which is a nicety that is,
6557 perhaps, not necessary, but which, if it did not occur, would be sure
6560 On the other hand, it also means that if you specify the command with
6561 a @kbd{C-u}, but without a number, that is to say, if the `raw prefix
6562 argument' is simply a cons cell, then the command puts you at the
6563 beginning of the second line @dots{} I don't know whether this is
6564 intended or whether no one has dealt with the code to avoid this
6567 @node Second Buffer Related Review, optional Exercise, beginning-of-buffer, More Complex
6568 @comment node-name, next, previous, up
6571 Here is a brief summary of some of the topics covered in this chapter.
6575 Evaluate each argument in sequence, and return the value of the first
6576 argument that is not @code{nil}; if none return a value that is not
6577 @code{nil}, return @code{nil}. In brief, return the first true value
6578 of the arguments; return a true value if one @emph{or} any of the
6582 Evaluate each argument in sequence, and if any are @code{nil}, return
6583 @code{nil}; if none are @code{nil}, return the value of the last
6584 argument. In brief, return a true value only if all the arguments are
6585 true; return a true value if one @emph{and} each of the others is
6589 A keyword used to indicate that an argument to a function definition
6590 is optional; this means that the function can be evaluated without the
6591 argument, if desired.
6593 @item prefix-numeric-value
6594 Convert the `raw prefix argument' produced by @code{(interactive
6595 "P")} to a numeric value.
6598 Move point forward to the beginning of the next line, or if the argument
6599 is greater than one, forward that many lines. If it can't move as far
6600 forward as it is supposed to, @code{forward-line} goes forward as far as
6601 it can and then returns a count of the number of additional lines it was
6602 supposed to move but couldn't.
6605 Delete the entire contents of the current buffer.
6608 Return @code{t} if its argument is a buffer; otherwise return @code{nil}.
6611 @node optional Exercise, , Second Buffer Related Review, More Complex
6612 @section @code{optional} Argument Exercise
6614 Write an interactive function with an optional argument that tests
6615 whether its argument, a number, is greater than or equal to, or else,
6616 less than the value of @code{fill-column}, and tells you which, in a
6617 message. However, if you do not pass an argument to the function, use
6618 56 as a default value.
6620 @node Narrowing & Widening, car cdr & cons, More Complex, Top
6621 @comment node-name, next, previous, up
6622 @chapter Narrowing and Widening
6623 @cindex Focusing attention (narrowing)
6627 Narrowing is a feature of Emacs that makes it possible for you to focus
6628 on a specific part of a buffer, and work without accidentally changing
6629 other parts. Narrowing is normally disabled since it can confuse
6633 * Narrowing advantages:: The advantages of narrowing
6634 * save-restriction:: The @code{save-restriction} special form.
6635 * what-line:: The number of the line that point is on.
6639 @node Narrowing advantages, save-restriction, Narrowing & Widening, Narrowing & Widening
6641 @unnumberedsec The Advantages of Narrowing
6644 With narrowing, the rest of a buffer is made invisible, as if it weren't
6645 there. This is an advantage if, for example, you want to replace a word
6646 in one part of a buffer but not in another: you narrow to the part you want
6647 and the replacement is carried out only in that section, not in the rest
6648 of the buffer. Searches will only work within a narrowed region, not
6649 outside of one, so if you are fixing a part of a document, you can keep
6650 yourself from accidentally finding parts you do not need to fix by
6651 narrowing just to the region you want.
6652 (The key binding for @code{narrow-to-region} is @kbd{C-x n n}.)
6654 However, narrowing does make the rest of the buffer invisible, which
6655 can scare people who inadvertently invoke narrowing and think they
6656 have deleted a part of their file. Moreover, the @code{undo} command
6657 (which is usually bound to @kbd{C-x u}) does not turn off narrowing
6658 (nor should it), so people can become quite desperate if they do not
6659 know that they can return the rest of a buffer to visibility with the
6660 @code{widen} command.
6661 (The key binding for @code{widen} is @kbd{C-x n w}.)
6663 Narrowing is just as useful to the Lisp interpreter as to a human.
6664 Often, an Emacs Lisp function is designed to work on just part of a
6665 buffer; or conversely, an Emacs Lisp function needs to work on all of a
6666 buffer that has been narrowed. The @code{what-line} function, for
6667 example, removes the narrowing from a buffer, if it has any narrowing
6668 and when it has finished its job, restores the narrowing to what it was.
6669 On the other hand, the @code{count-lines} function, which is called by
6670 @code{what-line}, uses narrowing to restrict itself to just that portion
6671 of the buffer in which it is interested and then restores the previous
6674 @node save-restriction, what-line, Narrowing advantages, Narrowing & Widening
6675 @comment node-name, next, previous, up
6676 @section The @code{save-restriction} Special Form
6677 @findex save-restriction
6679 In Emacs Lisp, you can use the @code{save-restriction} special form to
6680 keep track of whatever narrowing is in effect, if any. When the Lisp
6681 interpreter meets with @code{save-restriction}, it executes the code
6682 in the body of the @code{save-restriction} expression, and then undoes
6683 any changes to narrowing that the code caused. If, for example, the
6684 buffer is narrowed and the code that follows @code{save-restriction}
6685 gets rid of the narrowing, @code{save-restriction} returns the buffer
6686 to its narrowed region afterwards. In the @code{what-line} command,
6687 any narrowing the buffer may have is undone by the @code{widen}
6688 command that immediately follows the @code{save-restriction} command.
6689 Any original narrowing is restored just before the completion of the
6693 The template for a @code{save-restriction} expression is simple:
6703 The body of the @code{save-restriction} is one or more expressions that
6704 will be evaluated in sequence by the Lisp interpreter.
6706 Finally, a point to note: when you use both @code{save-excursion} and
6707 @code{save-restriction}, one right after the other, you should use
6708 @code{save-excursion} outermost. If you write them in reverse order,
6709 you may fail to record narrowing in the buffer to which Emacs switches
6710 after calling @code{save-excursion}. Thus, when written together,
6711 @code{save-excursion} and @code{save-restriction} should be written
6722 In other circumstances, when not written together, the
6723 @code{save-excursion} and @code{save-restriction} special forms must
6724 be written in the order appropriate to the function.
6740 /usr/local/src/emacs/lisp/simple.el
6743 "Print the current buffer line number and narrowed line number of point."
6745 (let ((start (point-min))
6746 (n (line-number-at-pos)))
6748 (message "Line %d" n)
6752 (message "line %d (narrowed line %d)"
6753 (+ n (line-number-at-pos start) -1) n))))))
6755 (defun line-number-at-pos (&optional pos)
6756 "Return (narrowed) buffer line number at position POS.
6757 If POS is nil, use current buffer location.
6758 Counting starts at (point-min), so the value refers
6759 to the contents of the accessible portion of the buffer."
6760 (let ((opoint (or pos (point))) start)
6762 (goto-char (point-min))
6763 (setq start (point))
6766 (1+ (count-lines start (point))))))
6768 (defun count-lines (start end)
6769 "Return number of lines between START and END.
6770 This is usually the number of newlines between them,
6771 but can be one more if START is not equal to END
6772 and the greater of them is not at the start of a line."
6775 (narrow-to-region start end)
6776 (goto-char (point-min))
6777 (if (eq selective-display t)
6780 (while (re-search-forward "[\n\C-m]" nil t 40)
6781 (setq done (+ 40 done)))
6782 (while (re-search-forward "[\n\C-m]" nil t 1)
6783 (setq done (+ 1 done)))
6784 (goto-char (point-max))
6785 (if (and (/= start end)
6789 (- (buffer-size) (forward-line (buffer-size)))))))
6792 @node what-line, narrow Exercise, save-restriction, Narrowing & Widening
6793 @comment node-name, next, previous, up
6794 @section @code{what-line}
6796 @cindex Widening, example of
6798 The @code{what-line} command tells you the number of the line in which
6799 the cursor is located. The function illustrates the use of the
6800 @code{save-restriction} and @code{save-excursion} commands. Here is the
6801 original text of the function:
6806 "Print the current line number (in the buffer) of point."
6813 (1+ (count-lines 1 (point)))))))
6817 (In recent versions of GNU Emacs, the @code{what-line} function has
6818 been expanded to tell you your line number in a narrowed buffer as
6819 well as your line number in a widened buffer. The recent version is
6820 more complex than the version shown here. If you feel adventurous,
6821 you might want to look at it after figuring out how this version
6822 works. You will probably need to use @kbd{C-h f}
6823 (@code{describe-function}). The newer version uses a conditional to
6824 determine whether the buffer has been narrowed.
6826 (Also, it uses @code{line-number-at-pos}, which among other simple
6827 expressions, such as @code{(goto-char (point-min))}, moves point to
6828 the beginning of the current line with @code{(forward-line 0)} rather
6829 than @code{beginning-of-line}.)
6831 The @code{what-line} function as shown here has a documentation line
6832 and is interactive, as you would expect. The next two lines use the
6833 functions @code{save-restriction} and @code{widen}.
6835 The @code{save-restriction} special form notes whatever narrowing is in
6836 effect, if any, in the current buffer and restores that narrowing after
6837 the code in the body of the @code{save-restriction} has been evaluated.
6839 The @code{save-restriction} special form is followed by @code{widen}.
6840 This function undoes any narrowing the current buffer may have had
6841 when @code{what-line} was called. (The narrowing that was there is
6842 the narrowing that @code{save-restriction} remembers.) This widening
6843 makes it possible for the line counting commands to count from the
6844 beginning of the buffer. Otherwise, they would have been limited to
6845 counting within the accessible region. Any original narrowing is
6846 restored just before the completion of the function by the
6847 @code{save-restriction} special form.
6849 The call to @code{widen} is followed by @code{save-excursion}, which
6850 saves the location of the cursor (i.e., of point) and of the mark, and
6851 restores them after the code in the body of the @code{save-excursion}
6852 uses the @code{beginning-of-line} function to move point.
6854 (Note that the @code{(widen)} expression comes between the
6855 @code{save-restriction} and @code{save-excursion} special forms. When
6856 you write the two @code{save- @dots{}} expressions in sequence, write
6857 @code{save-excursion} outermost.)
6860 The last two lines of the @code{what-line} function are functions to
6861 count the number of lines in the buffer and then print the number in the
6867 (1+ (count-lines 1 (point)))))))
6871 The @code{message} function prints a one-line message at the bottom of
6872 the Emacs screen. The first argument is inside of quotation marks and
6873 is printed as a string of characters. However, it may contain a
6874 @samp{%d} expression to print a following argument. @samp{%d} prints
6875 the argument as a decimal, so the message will say something such as
6879 The number that is printed in place of the @samp{%d} is computed by the
6880 last line of the function:
6883 (1+ (count-lines 1 (point)))
6889 (defun count-lines (start end)
6890 "Return number of lines between START and END.
6891 This is usually the number of newlines between them,
6892 but can be one more if START is not equal to END
6893 and the greater of them is not at the start of a line."
6896 (narrow-to-region start end)
6897 (goto-char (point-min))
6898 (if (eq selective-display t)
6901 (while (re-search-forward "[\n\C-m]" nil t 40)
6902 (setq done (+ 40 done)))
6903 (while (re-search-forward "[\n\C-m]" nil t 1)
6904 (setq done (+ 1 done)))
6905 (goto-char (point-max))
6906 (if (and (/= start end)
6910 (- (buffer-size) (forward-line (buffer-size)))))))
6914 What this does is count the lines from the first position of the
6915 buffer, indicated by the @code{1}, up to @code{(point)}, and then add
6916 one to that number. (The @code{1+} function adds one to its
6917 argument.) We add one to it because line 2 has only one line before
6918 it, and @code{count-lines} counts only the lines @emph{before} the
6921 After @code{count-lines} has done its job, and the message has been
6922 printed in the echo area, the @code{save-excursion} restores point and
6923 mark to their original positions; and @code{save-restriction} restores
6924 the original narrowing, if any.
6926 @node narrow Exercise, , what-line, Narrowing & Widening
6927 @section Exercise with Narrowing
6929 Write a function that will display the first 60 characters of the
6930 current buffer, even if you have narrowed the buffer to its latter
6931 half so that the first line is inaccessible. Restore point, mark, and
6932 narrowing. For this exercise, you need to use a whole potpourri of
6933 functions, including @code{save-restriction}, @code{widen},
6934 @code{goto-char}, @code{point-min}, @code{message}, and
6935 @code{buffer-substring}.
6937 @cindex Properties, mention of @code{buffer-substring-no-properties}
6938 (@code{buffer-substring} is a previously unmentioned function you will
6939 have to investigate yourself; or perhaps you will have to use
6940 @code{buffer-substring-no-properties} or
6941 @code{filter-buffer-substring} @dots{}, yet other functions. Text
6942 properties are a feature otherwise not discussed here. @xref{Text
6943 Properties, , Text Properties, elisp, The GNU Emacs Lisp Reference
6946 Additionally, do you really need @code{goto-char} or @code{point-min}?
6947 Or can you write the function without them?
6949 @node car cdr & cons, Cutting & Storing Text, Narrowing & Widening, Top
6950 @comment node-name, next, previous, up
6951 @chapter @code{car}, @code{cdr}, @code{cons}: Fundamental Functions
6952 @findex car, @r{introduced}
6953 @findex cdr, @r{introduced}
6955 In Lisp, @code{car}, @code{cdr}, and @code{cons} are fundamental
6956 functions. The @code{cons} function is used to construct lists, and
6957 the @code{car} and @code{cdr} functions are used to take them apart.
6959 In the walk through of the @code{copy-region-as-kill} function, we
6960 will see @code{cons} as well as two variants on @code{cdr},
6961 namely, @code{setcdr} and @code{nthcdr}. (@xref{copy-region-as-kill}.)
6964 * Strange Names:: An historical aside: why the strange names?
6965 * car & cdr:: Functions for extracting part of a list.
6966 * cons:: Constructing a list.
6967 * nthcdr:: Calling @code{cdr} repeatedly.
6969 * setcar:: Changing the first element of a list.
6970 * setcdr:: Changing the rest of a list.
6974 @node Strange Names, car & cdr, car cdr & cons, car cdr & cons
6976 @unnumberedsec Strange Names
6979 The name of the @code{cons} function is not unreasonable: it is an
6980 abbreviation of the word `construct'. The origins of the names for
6981 @code{car} and @code{cdr}, on the other hand, are esoteric: @code{car}
6982 is an acronym from the phrase `Contents of the Address part of the
6983 Register'; and @code{cdr} (pronounced `could-er') is an acronym from
6984 the phrase `Contents of the Decrement part of the Register'. These
6985 phrases refer to specific pieces of hardware on the very early
6986 computer on which the original Lisp was developed. Besides being
6987 obsolete, the phrases have been completely irrelevant for more than 25
6988 years to anyone thinking about Lisp. Nonetheless, although a few
6989 brave scholars have begun to use more reasonable names for these
6990 functions, the old terms are still in use. In particular, since the
6991 terms are used in the Emacs Lisp source code, we will use them in this
6994 @node car & cdr, cons, Strange Names, car cdr & cons
6995 @comment node-name, next, previous, up
6996 @section @code{car} and @code{cdr}
6998 The @sc{car} of a list is, quite simply, the first item in the list.
6999 Thus the @sc{car} of the list @code{(rose violet daisy buttercup)} is
7003 If you are reading this in Info in GNU Emacs, you can see this by
7004 evaluating the following:
7007 (car '(rose violet daisy buttercup))
7011 After evaluating the expression, @code{rose} will appear in the echo
7014 Clearly, a more reasonable name for the @code{car} function would be
7015 @code{first} and this is often suggested.
7017 @code{car} does not remove the first item from the list; it only reports
7018 what it is. After @code{car} has been applied to a list, the list is
7019 still the same as it was. In the jargon, @code{car} is
7020 `non-destructive'. This feature turns out to be important.
7022 The @sc{cdr} of a list is the rest of the list, that is, the
7023 @code{cdr} function returns the part of the list that follows the
7024 first item. Thus, while the @sc{car} of the list @code{'(rose violet
7025 daisy buttercup)} is @code{rose}, the rest of the list, the value
7026 returned by the @code{cdr} function, is @code{(violet daisy
7030 You can see this by evaluating the following in the usual way:
7033 (cdr '(rose violet daisy buttercup))
7037 When you evaluate this, @code{(violet daisy buttercup)} will appear in
7040 Like @code{car}, @code{cdr} does not remove any elements from the
7041 list---it just returns a report of what the second and subsequent
7044 Incidentally, in the example, the list of flowers is quoted. If it were
7045 not, the Lisp interpreter would try to evaluate the list by calling
7046 @code{rose} as a function. In this example, we do not want to do that.
7048 Clearly, a more reasonable name for @code{cdr} would be @code{rest}.
7050 (There is a lesson here: when you name new functions, consider very
7051 carefully what you are doing, since you may be stuck with the names
7052 for far longer than you expect. The reason this document perpetuates
7053 these names is that the Emacs Lisp source code uses them, and if I did
7054 not use them, you would have a hard time reading the code; but do,
7055 please, try to avoid using these terms yourself. The people who come
7056 after you will be grateful to you.)
7058 When @code{car} and @code{cdr} are applied to a list made up of symbols,
7059 such as the list @code{(pine fir oak maple)}, the element of the list
7060 returned by the function @code{car} is the symbol @code{pine} without
7061 any parentheses around it. @code{pine} is the first element in the
7062 list. However, the @sc{cdr} of the list is a list itself, @code{(fir
7063 oak maple)}, as you can see by evaluating the following expressions in
7068 (car '(pine fir oak maple))
7070 (cdr '(pine fir oak maple))
7074 On the other hand, in a list of lists, the first element is itself a
7075 list. @code{car} returns this first element as a list. For example,
7076 the following list contains three sub-lists, a list of carnivores, a
7077 list of herbivores and a list of sea mammals:
7081 (car '((lion tiger cheetah)
7082 (gazelle antelope zebra)
7083 (whale dolphin seal)))
7088 In this example, the first element or @sc{car} of the list is the list of
7089 carnivores, @code{(lion tiger cheetah)}, and the rest of the list is
7090 @code{((gazelle antelope zebra) (whale dolphin seal))}.
7094 (cdr '((lion tiger cheetah)
7095 (gazelle antelope zebra)
7096 (whale dolphin seal)))
7100 It is worth saying again that @code{car} and @code{cdr} are
7101 non-destructive---that is, they do not modify or change lists to which
7102 they are applied. This is very important for how they are used.
7104 Also, in the first chapter, in the discussion about atoms, I said that
7105 in Lisp, ``certain kinds of atom, such as an array, can be separated
7106 into parts; but the mechanism for doing this is different from the
7107 mechanism for splitting a list. As far as Lisp is concerned, the
7108 atoms of a list are unsplittable.'' (@xref{Lisp Atoms}.) The
7109 @code{car} and @code{cdr} functions are used for splitting lists and
7110 are considered fundamental to Lisp. Since they cannot split or gain
7111 access to the parts of an array, an array is considered an atom.
7112 Conversely, the other fundamental function, @code{cons}, can put
7113 together or construct a list, but not an array. (Arrays are handled
7114 by array-specific functions. @xref{Arrays, , Arrays, elisp, The GNU
7115 Emacs Lisp Reference Manual}.)
7117 @node cons, nthcdr, car & cdr, car cdr & cons
7118 @comment node-name, next, previous, up
7119 @section @code{cons}
7120 @findex cons, @r{introduced}
7122 The @code{cons} function constructs lists; it is the inverse of
7123 @code{car} and @code{cdr}. For example, @code{cons} can be used to make
7124 a four element list from the three element list, @code{(fir oak maple)}:
7127 (cons 'pine '(fir oak maple))
7132 After evaluating this list, you will see
7135 (pine fir oak maple)
7139 appear in the echo area. @code{cons} causes the creation of a new
7140 list in which the element is followed by the elements of the original
7143 We often say that `@code{cons} puts a new element at the beginning of
7144 a list; it attaches or pushes elements onto the list', but this
7145 phrasing can be misleading, since @code{cons} does not change an
7146 existing list, but creates a new one.
7148 Like @code{car} and @code{cdr}, @code{cons} is non-destructive.
7152 * length:: How to find the length of a list.
7155 @node Build a list, length, cons, cons
7157 @unnumberedsubsec Build a list
7160 @code{cons} must have a list to attach to.@footnote{Actually, you can
7161 @code{cons} an element to an atom to produce a dotted pair. Dotted
7162 pairs are not discussed here; see @ref{Dotted Pair Notation, , Dotted
7163 Pair Notation, elisp, The GNU Emacs Lisp Reference Manual}.} You
7164 cannot start from absolutely nothing. If you are building a list, you
7165 need to provide at least an empty list at the beginning. Here is a
7166 series of @code{cons} expressions that build up a list of flowers. If
7167 you are reading this in Info in GNU Emacs, you can evaluate each of
7168 the expressions in the usual way; the value is printed in this text
7169 after @samp{@result{}}, which you may read as `evaluates to'.
7173 (cons 'buttercup ())
7174 @result{} (buttercup)
7178 (cons 'daisy '(buttercup))
7179 @result{} (daisy buttercup)
7183 (cons 'violet '(daisy buttercup))
7184 @result{} (violet daisy buttercup)
7188 (cons 'rose '(violet daisy buttercup))
7189 @result{} (rose violet daisy buttercup)
7194 In the first example, the empty list is shown as @code{()} and a list
7195 made up of @code{buttercup} followed by the empty list is constructed.
7196 As you can see, the empty list is not shown in the list that was
7197 constructed. All that you see is @code{(buttercup)}. The empty list is
7198 not counted as an element of a list because there is nothing in an empty
7199 list. Generally speaking, an empty list is invisible.
7201 The second example, @code{(cons 'daisy '(buttercup))} constructs a new,
7202 two element list by putting @code{daisy} in front of @code{buttercup};
7203 and the third example constructs a three element list by putting
7204 @code{violet} in front of @code{daisy} and @code{buttercup}.
7206 @node length, , Build a list, cons
7207 @comment node-name, next, previous, up
7208 @subsection Find the Length of a List: @code{length}
7211 You can find out how many elements there are in a list by using the Lisp
7212 function @code{length}, as in the following examples:
7216 (length '(buttercup))
7221 (length '(daisy buttercup))
7226 (length (cons 'violet '(daisy buttercup)))
7232 In the third example, the @code{cons} function is used to construct a
7233 three element list which is then passed to the @code{length} function as
7237 We can also use @code{length} to count the number of elements in an
7248 As you would expect, the number of elements in an empty list is zero.
7250 An interesting experiment is to find out what happens if you try to find
7251 the length of no list at all; that is, if you try to call @code{length}
7252 without giving it an argument, not even an empty list:
7260 What you see, if you evaluate this, is the error message
7263 Lisp error: (wrong-number-of-arguments length 0)
7267 This means that the function receives the wrong number of
7268 arguments, zero, when it expects some other number of arguments. In
7269 this case, one argument is expected, the argument being a list whose
7270 length the function is measuring. (Note that @emph{one} list is
7271 @emph{one} argument, even if the list has many elements inside it.)
7273 The part of the error message that says @samp{length} is the name of
7277 @code{length} is still a subroutine, but you need C-h f to discover that.
7279 In an earlier version:
7280 This is written with a special notation, @samp{#<subr},
7281 that indicates that the function @code{length} is one of the primitive
7282 functions written in C rather than in Emacs Lisp. (@samp{subr} is an
7283 abbreviation for `subroutine'.) @xref{What Is a Function, , What Is a
7284 Function?, elisp , The GNU Emacs Lisp Reference Manual}, for more
7288 @node nthcdr, nth, cons, car cdr & cons
7289 @comment node-name, next, previous, up
7290 @section @code{nthcdr}
7293 The @code{nthcdr} function is associated with the @code{cdr} function.
7294 What it does is take the @sc{cdr} of a list repeatedly.
7296 If you take the @sc{cdr} of the list @code{(pine fir
7297 oak maple)}, you will be returned the list @code{(fir oak maple)}. If you
7298 repeat this on what was returned, you will be returned the list
7299 @code{(oak maple)}. (Of course, repeated @sc{cdr}ing on the original
7300 list will just give you the original @sc{cdr} since the function does
7301 not change the list. You need to evaluate the @sc{cdr} of the
7302 @sc{cdr} and so on.) If you continue this, eventually you will be
7303 returned an empty list, which in this case, instead of being shown as
7304 @code{()} is shown as @code{nil}.
7307 For review, here is a series of repeated @sc{cdr}s, the text following
7308 the @samp{@result{}} shows what is returned.
7312 (cdr '(pine fir oak maple))
7313 @result{}(fir oak maple)
7317 (cdr '(fir oak maple))
7318 @result{} (oak maple)
7343 You can also do several @sc{cdr}s without printing the values in
7348 (cdr (cdr '(pine fir oak maple)))
7349 @result{} (oak maple)
7354 In this example, the Lisp interpreter evaluates the innermost list first.
7355 The innermost list is quoted, so it just passes the list as it is to the
7356 innermost @code{cdr}. This @code{cdr} passes a list made up of the
7357 second and subsequent elements of the list to the outermost @code{cdr},
7358 which produces a list composed of the third and subsequent elements of
7359 the original list. In this example, the @code{cdr} function is repeated
7360 and returns a list that consists of the original list without its
7363 The @code{nthcdr} function does the same as repeating the call to
7364 @code{cdr}. In the following example, the argument 2 is passed to the
7365 function @code{nthcdr}, along with the list, and the value returned is
7366 the list without its first two items, which is exactly the same
7367 as repeating @code{cdr} twice on the list:
7371 (nthcdr 2 '(pine fir oak maple))
7372 @result{} (oak maple)
7377 Using the original four element list, we can see what happens when
7378 various numeric arguments are passed to @code{nthcdr}, including 0, 1,
7383 ;; @r{Leave the list as it was.}
7384 (nthcdr 0 '(pine fir oak maple))
7385 @result{} (pine fir oak maple)
7389 ;; @r{Return a copy without the first element.}
7390 (nthcdr 1 '(pine fir oak maple))
7391 @result{} (fir oak maple)
7395 ;; @r{Return a copy of the list without three elements.}
7396 (nthcdr 3 '(pine fir oak maple))
7401 ;; @r{Return a copy lacking all four elements.}
7402 (nthcdr 4 '(pine fir oak maple))
7407 ;; @r{Return a copy lacking all elements.}
7408 (nthcdr 5 '(pine fir oak maple))
7413 @node nth, setcar, nthcdr, car cdr & cons
7414 @comment node-name, next, previous, up
7418 The @code{nthcdr} function takes the @sc{cdr} of a list repeatedly.
7419 The @code{nth} function takes the @sc{car} of the result returned by
7420 @code{nthcdr}. It returns the Nth element of the list.
7423 Thus, if it were not defined in C for speed, the definition of
7424 @code{nth} would be:
7429 "Returns the Nth element of LIST.
7430 N counts from zero. If LIST is not that long, nil is returned."
7431 (car (nthcdr n list)))
7436 (Originally, @code{nth} was defined in Emacs Lisp in @file{subr.el},
7437 but its definition was redone in C in the 1980s.)
7439 The @code{nth} function returns a single element of a list.
7440 This can be very convenient.
7442 Note that the elements are numbered from zero, not one. That is to
7443 say, the first element of a list, its @sc{car} is the zeroth element.
7444 This is called `zero-based' counting and often bothers people who
7445 are accustomed to the first element in a list being number one, which
7453 (nth 0 '("one" "two" "three"))
7456 (nth 1 '("one" "two" "three"))
7461 It is worth mentioning that @code{nth}, like @code{nthcdr} and
7462 @code{cdr}, does not change the original list---the function is
7463 non-destructive. This is in sharp contrast to the @code{setcar} and
7464 @code{setcdr} functions.
7466 @node setcar, setcdr, nth, car cdr & cons
7467 @comment node-name, next, previous, up
7468 @section @code{setcar}
7471 As you might guess from their names, the @code{setcar} and @code{setcdr}
7472 functions set the @sc{car} or the @sc{cdr} of a list to a new value.
7473 They actually change the original list, unlike @code{car} and @code{cdr}
7474 which leave the original list as it was. One way to find out how this
7475 works is to experiment. We will start with the @code{setcar} function.
7478 First, we can make a list and then set the value of a variable to the
7479 list, using the @code{setq} function. Here is a list of animals:
7482 (setq animals '(antelope giraffe lion tiger))
7486 If you are reading this in Info inside of GNU Emacs, you can evaluate
7487 this expression in the usual fashion, by positioning the cursor after
7488 the expression and typing @kbd{C-x C-e}. (I'm doing this right here
7489 as I write this. This is one of the advantages of having the
7490 interpreter built into the computing environment. Incidentally, when
7491 there is nothing on the line after the final parentheses, such as a
7492 comment, point can be on the next line. Thus, if your cursor is in
7493 the first column of the next line, you do not need to move it.
7494 Indeed, Emacs permits any amount of white space after the final
7498 When we evaluate the variable @code{animals}, we see that it is bound to
7499 the list @code{(antelope giraffe lion tiger)}:
7504 @result{} (antelope giraffe lion tiger)
7509 Put another way, the variable @code{animals} points to the list
7510 @code{(antelope giraffe lion tiger)}.
7512 Next, evaluate the function @code{setcar} while passing it two
7513 arguments, the variable @code{animals} and the quoted symbol
7514 @code{hippopotamus}; this is done by writing the three element list
7515 @code{(setcar animals 'hippopotamus)} and then evaluating it in the
7519 (setcar animals 'hippopotamus)
7524 After evaluating this expression, evaluate the variable @code{animals}
7525 again. You will see that the list of animals has changed:
7530 @result{} (hippopotamus giraffe lion tiger)
7535 The first element on the list, @code{antelope} is replaced by
7536 @code{hippopotamus}.
7538 So we can see that @code{setcar} did not add a new element to the list
7539 as @code{cons} would have; it replaced @code{antelope} with
7540 @code{hippopotamus}; it @emph{changed} the list.
7542 @node setcdr, cons Exercise, setcar, car cdr & cons
7543 @comment node-name, next, previous, up
7544 @section @code{setcdr}
7547 The @code{setcdr} function is similar to the @code{setcar} function,
7548 except that the function replaces the second and subsequent elements of
7549 a list rather than the first element.
7551 (To see how to change the last element of a list, look ahead to
7552 @ref{kill-new function, , The @code{kill-new} function}, which uses
7553 the @code{nthcdr} and @code{setcdr} functions.)
7556 To see how this works, set the value of the variable to a list of
7557 domesticated animals by evaluating the following expression:
7560 (setq domesticated-animals '(horse cow sheep goat))
7565 If you now evaluate the list, you will be returned the list
7566 @code{(horse cow sheep goat)}:
7570 domesticated-animals
7571 @result{} (horse cow sheep goat)
7576 Next, evaluate @code{setcdr} with two arguments, the name of the
7577 variable which has a list as its value, and the list to which the
7578 @sc{cdr} of the first list will be set;
7581 (setcdr domesticated-animals '(cat dog))
7585 If you evaluate this expression, the list @code{(cat dog)} will appear
7586 in the echo area. This is the value returned by the function. The
7587 result we are interested in is the ``side effect'', which we can see by
7588 evaluating the variable @code{domesticated-animals}:
7592 domesticated-animals
7593 @result{} (horse cat dog)
7598 Indeed, the list is changed from @code{(horse cow sheep goat)} to
7599 @code{(horse cat dog)}. The @sc{cdr} of the list is changed from
7600 @code{(cow sheep goat)} to @code{(cat dog)}.
7602 @node cons Exercise, , setcdr, car cdr & cons
7605 Construct a list of four birds by evaluating several expressions with
7606 @code{cons}. Find out what happens when you @code{cons} a list onto
7607 itself. Replace the first element of the list of four birds with a
7608 fish. Replace the rest of that list with a list of other fish.
7610 @node Cutting & Storing Text, List Implementation, car cdr & cons, Top
7611 @comment node-name, next, previous, up
7612 @chapter Cutting and Storing Text
7613 @cindex Cutting and storing text
7614 @cindex Storing and cutting text
7615 @cindex Killing text
7616 @cindex Clipping text
7617 @cindex Erasing text
7618 @cindex Deleting text
7620 Whenever you cut or clip text out of a buffer with a `kill' command in
7621 GNU Emacs, it is stored in a list and you can bring it back with a
7624 (The use of the word `kill' in Emacs for processes which specifically
7625 @emph{do not} destroy the values of the entities is an unfortunate
7626 historical accident. A much more appropriate word would be `clip' since
7627 that is what the kill commands do; they clip text out of a buffer and
7628 put it into storage from which it can be brought back. I have often
7629 been tempted to replace globally all occurrences of `kill' in the Emacs
7630 sources with `clip' and all occurrences of `killed' with `clipped'.)
7633 * Storing Text:: Text is stored in a list.
7634 * zap-to-char:: Cutting out text up to a character.
7635 * kill-region:: Cutting text out of a region.
7636 * copy-region-as-kill:: A definition for copying text.
7637 * Digression into C:: Minor note on C programming language macros.
7638 * defvar:: How to give a variable an initial value.
7639 * cons & search-fwd Review::
7640 * search Exercises::
7643 @node Storing Text, zap-to-char, Cutting & Storing Text, Cutting & Storing Text
7645 @unnumberedsec Storing Text in a List
7648 When text is cut out of a buffer, it is stored on a list. Successive
7649 pieces of text are stored on the list successively, so the list might
7653 ("a piece of text" "previous piece")
7658 The function @code{cons} can be used to create a new list from a piece
7659 of text (an `atom', to use the jargon) and an existing list, like
7664 (cons "another piece"
7665 '("a piece of text" "previous piece"))
7671 If you evaluate this expression, a list of three elements will appear in
7675 ("another piece" "a piece of text" "previous piece")
7678 With the @code{car} and @code{nthcdr} functions, you can retrieve
7679 whichever piece of text you want. For example, in the following code,
7680 @code{nthcdr 1 @dots{}} returns the list with the first item removed;
7681 and the @code{car} returns the first element of that remainder---the
7682 second element of the original list:
7686 (car (nthcdr 1 '("another piece"
7689 @result{} "a piece of text"
7693 The actual functions in Emacs are more complex than this, of course.
7694 The code for cutting and retrieving text has to be written so that
7695 Emacs can figure out which element in the list you want---the first,
7696 second, third, or whatever. In addition, when you get to the end of
7697 the list, Emacs should give you the first element of the list, rather
7698 than nothing at all.
7700 The list that holds the pieces of text is called the @dfn{kill ring}.
7701 This chapter leads up to a description of the kill ring and how it is
7702 used by first tracing how the @code{zap-to-char} function works. This
7703 function uses (or `calls') a function that invokes a function that
7704 manipulates the kill ring. Thus, before reaching the mountains, we
7705 climb the foothills.
7707 A subsequent chapter describes how text that is cut from the buffer is
7708 retrieved. @xref{Yanking, , Yanking Text Back}.
7710 @node zap-to-char, kill-region, Storing Text, Cutting & Storing Text
7711 @comment node-name, next, previous, up
7712 @section @code{zap-to-char}
7715 The @code{zap-to-char} function changed little between GNU Emacs
7716 version 19 and GNU Emacs version 22. However, @code{zap-to-char}
7717 calls another function, @code{kill-region}, which enjoyed a major
7720 The @code{kill-region} function in Emacs 19 is complex, but does not
7721 use code that is important at this time. We will skip it.
7723 The @code{kill-region} function in Emacs 22 is easier to read than the
7724 same function in Emacs 19 and introduces a very important concept,
7725 that of error handling. We will walk through the function.
7727 But first, let us look at the interactive @code{zap-to-char} function.
7730 * Complete zap-to-char:: The complete implementation.
7731 * zap-to-char interactive:: A three part interactive expression.
7732 * zap-to-char body:: A short overview.
7733 * search-forward:: How to search for a string.
7734 * progn:: The @code{progn} special form.
7735 * Summing up zap-to-char:: Using @code{point} and @code{search-forward}.
7738 @node Complete zap-to-char, zap-to-char interactive, zap-to-char, zap-to-char
7740 @unnumberedsubsec The Complete @code{zap-to-char} Implementation
7743 The @code{zap-to-char} function removes the text in the region between
7744 the location of the cursor (i.e., of point) up to and including the
7745 next occurrence of a specified character. The text that
7746 @code{zap-to-char} removes is put in the kill ring; and it can be
7747 retrieved from the kill ring by typing @kbd{C-y} (@code{yank}). If
7748 the command is given an argument, it removes text through that number
7749 of occurrences. Thus, if the cursor were at the beginning of this
7750 sentence and the character were @samp{s}, @samp{Thus} would be
7751 removed. If the argument were two, @samp{Thus, if the curs} would be
7752 removed, up to and including the @samp{s} in @samp{cursor}.
7754 If the specified character is not found, @code{zap-to-char} will say
7755 ``Search failed'', tell you the character you typed, and not remove
7758 In order to determine how much text to remove, @code{zap-to-char} uses
7759 a search function. Searches are used extensively in code that
7760 manipulates text, and we will focus attention on them as well as on the
7764 @c GNU Emacs version 19
7765 (defun zap-to-char (arg char) ; version 19 implementation
7766 "Kill up to and including ARG'th occurrence of CHAR.
7767 Goes backward if ARG is negative; error if CHAR not found."
7768 (interactive "*p\ncZap to char: ")
7769 (kill-region (point)
7772 (char-to-string char) nil nil arg)
7777 Here is the complete text of the version 22 implementation of the function:
7782 (defun zap-to-char (arg char)
7783 "Kill up to and including ARG'th occurrence of CHAR.
7784 Case is ignored if `case-fold-search' is non-nil in the current buffer.
7785 Goes backward if ARG is negative; error if CHAR not found."
7786 (interactive "p\ncZap to char: ")
7787 (if (char-table-p translation-table-for-input)
7788 (setq char (or (aref translation-table-for-input char) char)))
7789 (kill-region (point) (progn
7790 (search-forward (char-to-string char)
7796 The documentation is thorough. You do need to know the jargon meaning
7799 @node zap-to-char interactive, zap-to-char body, Complete zap-to-char, zap-to-char
7800 @comment node-name, next, previous, up
7801 @subsection The @code{interactive} Expression
7804 The interactive expression in the @code{zap-to-char} command looks like
7808 (interactive "p\ncZap to char: ")
7811 The part within quotation marks, @code{"p\ncZap to char:@: "}, specifies
7812 two different things. First, and most simply, is the @samp{p}.
7813 This part is separated from the next part by a newline, @samp{\n}.
7814 The @samp{p} means that the first argument to the function will be
7815 passed the value of a `processed prefix'. The prefix argument is
7816 passed by typing @kbd{C-u} and a number, or @kbd{M-} and a number. If
7817 the function is called interactively without a prefix, 1 is passed to
7820 The second part of @code{"p\ncZap to char:@: "} is
7821 @samp{cZap to char:@: }. In this part, the lower case @samp{c}
7822 indicates that @code{interactive} expects a prompt and that the
7823 argument will be a character. The prompt follows the @samp{c} and is
7824 the string @samp{Zap to char:@: } (with a space after the colon to
7827 What all this does is prepare the arguments to @code{zap-to-char} so they
7828 are of the right type, and give the user a prompt.
7830 In a read-only buffer, the @code{zap-to-char} function copies the text
7831 to the kill ring, but does not remove it. The echo area displays a
7832 message saying that the buffer is read-only. Also, the terminal may
7833 beep or blink at you.
7835 @node zap-to-char body, search-forward, zap-to-char interactive, zap-to-char
7836 @comment node-name, next, previous, up
7837 @subsection The Body of @code{zap-to-char}
7839 The body of the @code{zap-to-char} function contains the code that
7840 kills (that is, removes) the text in the region from the current
7841 position of the cursor up to and including the specified character.
7843 The first part of the code looks like this:
7846 (if (char-table-p translation-table-for-input)
7847 (setq char (or (aref translation-table-for-input char) char)))
7848 (kill-region (point) (progn
7849 (search-forward (char-to-string char) nil nil arg)
7854 @code{char-table-p} is an hitherto unseen function. It determines
7855 whether its argument is a character table. When it is, it sets the
7856 character passed to @code{zap-to-char} to one of them, if that
7857 character exists, or to the character itself. (This becomes important
7858 for certain characters in non-European languages. The @code{aref}
7859 function extracts an element from an array. It is an array-specific
7860 function that is not described in this document. @xref{Arrays, ,
7861 Arrays, elisp, The GNU Emacs Lisp Reference Manual}.)
7864 @code{(point)} is the current position of the cursor.
7866 The next part of the code is an expression using @code{progn}. The body
7867 of the @code{progn} consists of calls to @code{search-forward} and
7870 It is easier to understand how @code{progn} works after learning about
7871 @code{search-forward}, so we will look at @code{search-forward} and
7872 then at @code{progn}.
7874 @node search-forward, progn, zap-to-char body, zap-to-char
7875 @comment node-name, next, previous, up
7876 @subsection The @code{search-forward} Function
7877 @findex search-forward
7879 The @code{search-forward} function is used to locate the
7880 zapped-for-character in @code{zap-to-char}. If the search is
7881 successful, @code{search-forward} leaves point immediately after the
7882 last character in the target string. (In @code{zap-to-char}, the
7883 target string is just one character long. @code{zap-to-char} uses the
7884 function @code{char-to-string} to ensure that the computer treats that
7885 character as a string.) If the search is backwards,
7886 @code{search-forward} leaves point just before the first character in
7887 the target. Also, @code{search-forward} returns @code{t} for true.
7888 (Moving point is therefore a `side effect'.)
7891 In @code{zap-to-char}, the @code{search-forward} function looks like this:
7894 (search-forward (char-to-string char) nil nil arg)
7897 The @code{search-forward} function takes four arguments:
7901 The first argument is the target, what is searched for. This must be a
7902 string, such as @samp{"z"}.
7904 As it happens, the argument passed to @code{zap-to-char} is a single
7905 character. Because of the way computers are built, the Lisp
7906 interpreter may treat a single character as being different from a
7907 string of characters. Inside the computer, a single character has a
7908 different electronic format than a string of one character. (A single
7909 character can often be recorded in the computer using exactly one
7910 byte; but a string may be longer, and the computer needs to be ready
7911 for this.) Since the @code{search-forward} function searches for a
7912 string, the character that the @code{zap-to-char} function receives as
7913 its argument must be converted inside the computer from one format to
7914 the other; otherwise the @code{search-forward} function will fail.
7915 The @code{char-to-string} function is used to make this conversion.
7918 The second argument bounds the search; it is specified as a position in
7919 the buffer. In this case, the search can go to the end of the buffer,
7920 so no bound is set and the second argument is @code{nil}.
7923 The third argument tells the function what it should do if the search
7924 fails---it can signal an error (and print a message) or it can return
7925 @code{nil}. A @code{nil} as the third argument causes the function to
7926 signal an error when the search fails.
7929 The fourth argument to @code{search-forward} is the repeat count---how
7930 many occurrences of the string to look for. This argument is optional
7931 and if the function is called without a repeat count, this argument is
7932 passed the value 1. If this argument is negative, the search goes
7937 In template form, a @code{search-forward} expression looks like this:
7941 (search-forward "@var{target-string}"
7942 @var{limit-of-search}
7943 @var{what-to-do-if-search-fails}
7948 We will look at @code{progn} next.
7950 @node progn, Summing up zap-to-char, search-forward, zap-to-char
7951 @comment node-name, next, previous, up
7952 @subsection The @code{progn} Special Form
7955 @code{progn} is a special form that causes each of its arguments to be
7956 evaluated in sequence and then returns the value of the last one. The
7957 preceding expressions are evaluated only for the side effects they
7958 perform. The values produced by them are discarded.
7961 The template for a @code{progn} expression is very simple:
7970 In @code{zap-to-char}, the @code{progn} expression has to do two things:
7971 put point in exactly the right position; and return the location of
7972 point so that @code{kill-region} will know how far to kill to.
7974 The first argument to the @code{progn} is @code{search-forward}. When
7975 @code{search-forward} finds the string, the function leaves point
7976 immediately after the last character in the target string. (In this
7977 case the target string is just one character long.) If the search is
7978 backwards, @code{search-forward} leaves point just before the first
7979 character in the target. The movement of point is a side effect.
7981 The second and last argument to @code{progn} is the expression
7982 @code{(point)}. This expression returns the value of point, which in
7983 this case will be the location to which it has been moved by
7984 @code{search-forward}. (In the source, a line that tells the function
7985 to go to the previous character, if it is going forward, was commented
7986 out in 1999; I don't remember whether that feature or mis-feature was
7987 ever a part of the distributed source.) The value of @code{point} is
7988 returned by the @code{progn} expression and is passed to
7989 @code{kill-region} as @code{kill-region}'s second argument.
7991 @node Summing up zap-to-char, , progn, zap-to-char
7992 @comment node-name, next, previous, up
7993 @subsection Summing up @code{zap-to-char}
7995 Now that we have seen how @code{search-forward} and @code{progn} work,
7996 we can see how the @code{zap-to-char} function works as a whole.
7998 The first argument to @code{kill-region} is the position of the cursor
7999 when the @code{zap-to-char} command is given---the value of point at
8000 that time. Within the @code{progn}, the search function then moves
8001 point to just after the zapped-to-character and @code{point} returns the
8002 value of this location. The @code{kill-region} function puts together
8003 these two values of point, the first one as the beginning of the region
8004 and the second one as the end of the region, and removes the region.
8006 The @code{progn} special form is necessary because the
8007 @code{kill-region} command takes two arguments; and it would fail if
8008 @code{search-forward} and @code{point} expressions were written in
8009 sequence as two additional arguments. The @code{progn} expression is
8010 a single argument to @code{kill-region} and returns the one value that
8011 @code{kill-region} needs for its second argument.
8013 @node kill-region, copy-region-as-kill, zap-to-char, Cutting & Storing Text
8014 @comment node-name, next, previous, up
8015 @section @code{kill-region}
8018 The @code{zap-to-char} function uses the @code{kill-region} function.
8019 This function clips text from a region and copies that text to
8020 the kill ring, from which it may be retrieved.
8025 (defun kill-region (beg end &optional yank-handler)
8026 "Kill (\"cut\") text between point and mark.
8027 This deletes the text from the buffer and saves it in the kill ring.
8028 The command \\[yank] can retrieve it from there.
8029 \(If you want to kill and then yank immediately, use \\[kill-ring-save].)
8031 If you want to append the killed region to the last killed text,
8032 use \\[append-next-kill] before \\[kill-region].
8034 If the buffer is read-only, Emacs will beep and refrain from deleting
8035 the text, but put the text in the kill ring anyway. This means that
8036 you can use the killing commands to copy text from a read-only buffer.
8038 This is the primitive for programs to kill text (as opposed to deleting it).
8039 Supply two arguments, character positions indicating the stretch of text
8041 Any command that calls this function is a \"kill command\".
8042 If the previous command was also a kill command,
8043 the text killed this time appends to the text killed last time
8044 to make one entry in the kill ring.
8046 In Lisp code, optional third arg YANK-HANDLER, if non-nil,
8047 specifies the yank-handler text property to be set on the killed
8048 text. See `insert-for-yank'."
8049 ;; Pass point first, then mark, because the order matters
8050 ;; when calling kill-append.
8051 (interactive (list (point) (mark)))
8052 (unless (and beg end)
8053 (error "The mark is not set now, so there is no region"))
8055 (let ((string (filter-buffer-substring beg end t)))
8056 (when string ;STRING is nil if BEG = END
8057 ;; Add that string to the kill ring, one way or another.
8058 (if (eq last-command 'kill-region)
8059 (kill-append string (< end beg) yank-handler)
8060 (kill-new string nil yank-handler)))
8061 (when (or string (eq last-command 'kill-region))
8062 (setq this-command 'kill-region))
8064 ((buffer-read-only text-read-only)
8065 ;; The code above failed because the buffer, or some of the characters
8066 ;; in the region, are read-only.
8067 ;; We should beep, in case the user just isn't aware of this.
8068 ;; However, there's no harm in putting
8069 ;; the region's text in the kill ring, anyway.
8070 (copy-region-as-kill beg end)
8071 ;; Set this-command now, so it will be set even if we get an error.
8072 (setq this-command 'kill-region)
8073 ;; This should barf, if appropriate, and give us the correct error.
8074 (if kill-read-only-ok
8075 (progn (message "Read only text copied to kill ring") nil)
8076 ;; Signal an error if the buffer is read-only.
8077 (barf-if-buffer-read-only)
8078 ;; If the buffer isn't read-only, the text is.
8079 (signal 'text-read-only (list (current-buffer)))))))
8082 The Emacs 22 version of that function uses @code{condition-case} and
8083 @code{copy-region-as-kill}, both of which we will explain.
8084 @code{condition-case} is an important special form.
8086 In essence, the @code{kill-region} function calls
8087 @code{condition-case}, which takes three arguments. In this function,
8088 the first argument does nothing. The second argument contains the
8089 code that does the work when all goes well. The third argument
8090 contains the code that is called in the event of an error.
8093 * Complete kill-region:: The function definition.
8094 * condition-case:: Dealing with a problem.
8098 @node Complete kill-region, condition-case, kill-region, kill-region
8100 @unnumberedsubsec The Complete @code{kill-region} Definition
8104 We will go through the @code{condition-case} code in a moment. First,
8105 let us look at the definition of @code{kill-region}, with comments
8111 (defun kill-region (beg end)
8112 "Kill (\"cut\") text between point and mark.
8113 This deletes the text from the buffer and saves it in the kill ring.
8114 The command \\[yank] can retrieve it from there. @dots{} "
8118 ;; @bullet{} Since order matters, pass point first.
8119 (interactive (list (point) (mark)))
8120 ;; @bullet{} And tell us if we cannot cut the text.
8121 ;; `unless' is an `if' without a then-part.
8122 (unless (and beg end)
8123 (error "The mark is not set now, so there is no region"))
8127 ;; @bullet{} `condition-case' takes three arguments.
8128 ;; If the first argument is nil, as it is here,
8129 ;; information about the error signal is not
8130 ;; stored for use by another function.
8135 ;; @bullet{} The second argument to `condition-case' tells the
8136 ;; Lisp interpreter what to do when all goes well.
8140 ;; It starts with a `let' function that extracts the string
8141 ;; and tests whether it exists. If so (that is what the
8142 ;; `when' checks), it calls an `if' function that determines
8143 ;; whether the previous command was another call to
8144 ;; `kill-region'; if it was, then the new text is appended to
8145 ;; the previous text; if not, then a different function,
8146 ;; `kill-new', is called.
8150 ;; The `kill-append' function concatenates the new string and
8151 ;; the old. The `kill-new' function inserts text into a new
8152 ;; item in the kill ring.
8156 ;; `when' is an `if' without an else-part. The second `when'
8157 ;; again checks whether the current string exists; in
8158 ;; addition, it checks whether the previous command was
8159 ;; another call to `kill-region'. If one or the other
8160 ;; condition is true, then it sets the current command to
8161 ;; be `kill-region'.
8164 (let ((string (filter-buffer-substring beg end t)))
8165 (when string ;STRING is nil if BEG = END
8166 ;; Add that string to the kill ring, one way or another.
8167 (if (eq last-command 'kill-region)
8170 ;; @minus{} `yank-handler' is an optional argument to
8171 ;; `kill-region' that tells the `kill-append' and
8172 ;; `kill-new' functions how deal with properties
8173 ;; added to the text, such as `bold' or `italics'.
8174 (kill-append string (< end beg) yank-handler)
8175 (kill-new string nil yank-handler)))
8176 (when (or string (eq last-command 'kill-region))
8177 (setq this-command 'kill-region))
8182 ;; @bullet{} The third argument to `condition-case' tells the interpreter
8183 ;; what to do with an error.
8186 ;; The third argument has a conditions part and a body part.
8187 ;; If the conditions are met (in this case,
8188 ;; if text or buffer are read-only)
8189 ;; then the body is executed.
8192 ;; The first part of the third argument is the following:
8193 ((buffer-read-only text-read-only) ;; the if-part
8194 ;; @dots{} the then-part
8195 (copy-region-as-kill beg end)
8198 ;; Next, also as part of the then-part, set this-command, so
8199 ;; it will be set in an error
8200 (setq this-command 'kill-region)
8201 ;; Finally, in the then-part, send a message if you may copy
8202 ;; the text to the kill ring without signally an error, but
8203 ;; don't if you may not.
8206 (if kill-read-only-ok
8207 (progn (message "Read only text copied to kill ring") nil)
8208 (barf-if-buffer-read-only)
8209 ;; If the buffer isn't read-only, the text is.
8210 (signal 'text-read-only (list (current-buffer)))))
8218 (defun kill-region (beg end)
8219 "Kill between point and mark.
8220 The text is deleted but saved in the kill ring."
8225 ;; 1. `condition-case' takes three arguments.
8226 ;; If the first argument is nil, as it is here,
8227 ;; information about the error signal is not
8228 ;; stored for use by another function.
8233 ;; 2. The second argument to `condition-case'
8234 ;; tells the Lisp interpreter what to do when all goes well.
8238 ;; The `delete-and-extract-region' function usually does the
8239 ;; work. If the beginning and ending of the region are both
8240 ;; the same, then the variable `string' will be empty, or nil
8241 (let ((string (delete-and-extract-region beg end)))
8245 ;; `when' is an `if' clause that cannot take an `else-part'.
8246 ;; Emacs normally sets the value of `last-command' to the
8247 ;; previous command.
8250 ;; `kill-append' concatenates the new string and the old.
8251 ;; `kill-new' inserts text into a new item in the kill ring.
8253 (if (eq last-command 'kill-region)
8254 ;; if true, prepend string
8255 (kill-append string (< end beg))
8257 (setq this-command 'kill-region))
8261 ;; 3. The third argument to `condition-case' tells the interpreter
8262 ;; what to do with an error.
8265 ;; The third argument has a conditions part and a body part.
8266 ;; If the conditions are met (in this case,
8267 ;; if text or buffer are read-only)
8268 ;; then the body is executed.
8271 ((buffer-read-only text-read-only) ;; this is the if-part
8273 (copy-region-as-kill beg end)
8276 (if kill-read-only-ok ;; usually this variable is nil
8277 (message "Read only text copied to kill ring")
8278 ;; or else, signal an error if the buffer is read-only;
8279 (barf-if-buffer-read-only)
8280 ;; and, in any case, signal that the text is read-only.
8281 (signal 'text-read-only (list (current-buffer)))))))
8286 @node condition-case, Lisp macro, Complete kill-region, kill-region
8287 @comment node-name, next, previous, up
8288 @subsection @code{condition-case}
8289 @findex condition-case
8291 As we have seen earlier (@pxref{Making Errors, , Generate an Error
8292 Message}), when the Emacs Lisp interpreter has trouble evaluating an
8293 expression, it provides you with help; in the jargon, this is called
8294 ``signaling an error''. Usually, the computer stops the program and
8295 shows you a message.
8297 However, some programs undertake complicated actions. They should not
8298 simply stop on an error. In the @code{kill-region} function, the most
8299 likely error is that you will try to kill text that is read-only and
8300 cannot be removed. So the @code{kill-region} function contains code
8301 to handle this circumstance. This code, which makes up the body of
8302 the @code{kill-region} function, is inside of a @code{condition-case}
8306 The template for @code{condition-case} looks like this:
8313 @var{error-handler}@dots{})
8317 The second argument, @var{bodyform}, is straightforward. The
8318 @code{condition-case} special form causes the Lisp interpreter to
8319 evaluate the code in @var{bodyform}. If no error occurs, the special
8320 form returns the code's value and produces the side-effects, if any.
8322 In short, the @var{bodyform} part of a @code{condition-case}
8323 expression determines what should happen when everything works
8326 However, if an error occurs, among its other actions, the function
8327 generating the error signal will define one or more error condition
8330 An error handler is the third argument to @code{condition case}.
8331 An error handler has two parts, a @var{condition-name} and a
8332 @var{body}. If the @var{condition-name} part of an error handler
8333 matches a condition name generated by an error, then the @var{body}
8334 part of the error handler is run.
8336 As you will expect, the @var{condition-name} part of an error handler
8337 may be either a single condition name or a list of condition names.
8339 Also, a complete @code{condition-case} expression may contain more
8340 than one error handler. When an error occurs, the first applicable
8343 Lastly, the first argument to the @code{condition-case} expression,
8344 the @var{var} argument, is sometimes bound to a variable that
8345 contains information about the error. However, if that argument is
8346 nil, as is the case in @code{kill-region}, that information is
8350 In brief, in the @code{kill-region} function, the code
8351 @code{condition-case} works like this:
8355 @var{If no errors}, @var{run only this code}
8356 @var{but}, @var{if errors}, @var{run this other code}.
8363 copy-region-as-kill is short, 12 lines, and uses
8364 filter-buffer-substring, which is longer, 39 lines
8365 and has delete-and-extract-region in it.
8366 delete-and-extract-region is written in C.
8368 see Initializing a Variable with @code{defvar}
8370 Initializing a Variable with @code{defvar} includes line 8350
8373 @node Lisp macro, , condition-case, kill-region
8374 @comment node-name, next, previous, up
8375 @subsection Lisp macro
8379 The part of the @code{condition-case} expression that is evaluated in
8380 the expectation that all goes well has a @code{when}. The code uses
8381 @code{when} to determine whether the @code{string} variable points to
8384 A @code{when} expression is simply a programmers' convenience. It is
8385 an @code{if} without the possibility of an else clause. In your mind,
8386 you can replace @code{when} with @code{if} and understand what goes
8387 on. That is what the Lisp interpreter does.
8389 Technically speaking, @code{when} is a Lisp macro. A Lisp @dfn{macro}
8390 enables you to define new control constructs and other language
8391 features. It tells the interpreter how to compute another Lisp
8392 expression which will in turn compute the value. In this case, the
8393 `other expression' is an @code{if} expression.
8395 The @code{kill-region} function definition also has an @code{unless}
8396 macro; it is the converse of @code{when}. The @code{unless} macro is
8397 an @code{if} without a then clause
8399 For more about Lisp macros, see @ref{Macros, , Macros, elisp, The GNU
8400 Emacs Lisp Reference Manual}. The C programming language also
8401 provides macros. These are different, but also useful.
8404 We will briefly look at C macros in
8405 @ref{Digression into C}.
8409 Regarding the @code{when} macro, in the @code{condition-case}
8410 expression, when the string has content, then another conditional
8411 expression is executed. This is an @code{if} with both a then-part
8416 (if (eq last-command 'kill-region)
8417 (kill-append string (< end beg) yank-handler)
8418 (kill-new string nil yank-handler))
8422 The then-part is evaluated if the previous command was another call to
8423 @code{kill-region}; if not, the else-part is evaluated.
8425 @code{yank-handler} is an optional argument to @code{kill-region} that
8426 tells the @code{kill-append} and @code{kill-new} functions how deal
8427 with properties added to the text, such as `bold' or `italics'.
8429 @code{last-command} is a variable that comes with Emacs that we have
8430 not seen before. Normally, whenever a function is executed, Emacs
8431 sets the value of @code{last-command} to the previous command.
8434 In this segment of the definition, the @code{if} expression checks
8435 whether the previous command was @code{kill-region}. If it was,
8438 (kill-append string (< end beg) yank-handler)
8442 concatenates a copy of the newly clipped text to the just previously
8443 clipped text in the kill ring.
8445 @node copy-region-as-kill, Digression into C, kill-region, Cutting & Storing Text
8446 @comment node-name, next, previous, up
8447 @section @code{copy-region-as-kill}
8448 @findex copy-region-as-kill
8451 The @code{copy-region-as-kill} function copies a region of text from a
8452 buffer and (via either @code{kill-append} or @code{kill-new}) saves it
8453 in the @code{kill-ring}.
8455 If you call @code{copy-region-as-kill} immediately after a
8456 @code{kill-region} command, Emacs appends the newly copied text to the
8457 previously copied text. This means that if you yank back the text, you
8458 get it all, from both this and the previous operation. On the other
8459 hand, if some other command precedes the @code{copy-region-as-kill},
8460 the function copies the text into a separate entry in the kill ring.
8463 * Complete copy-region-as-kill:: The complete function definition.
8464 * copy-region-as-kill body:: The body of @code{copy-region-as-kill}.
8467 @node Complete copy-region-as-kill, copy-region-as-kill body, copy-region-as-kill, copy-region-as-kill
8469 @unnumberedsubsec The complete @code{copy-region-as-kill} function definition
8473 Here is the complete text of the version 22 @code{copy-region-as-kill}
8478 (defun copy-region-as-kill (beg end)
8479 "Save the region as if killed, but don't kill it.
8480 In Transient Mark mode, deactivate the mark.
8481 If `interprogram-cut-function' is non-nil, also save the text for a window
8482 system cut and paste."
8486 (if (eq last-command 'kill-region)
8487 (kill-append (filter-buffer-substring beg end) (< end beg))
8488 (kill-new (filter-buffer-substring beg end)))
8491 (if transient-mark-mode
8492 (setq deactivate-mark t))
8498 As usual, this function can be divided into its component parts:
8502 (defun copy-region-as-kill (@var{argument-list})
8503 "@var{documentation}@dots{}"
8509 The arguments are @code{beg} and @code{end} and the function is
8510 interactive with @code{"r"}, so the two arguments must refer to the
8511 beginning and end of the region. If you have been reading though this
8512 document from the beginning, understanding these parts of a function is
8513 almost becoming routine.
8515 The documentation is somewhat confusing unless you remember that the
8516 word `kill' has a meaning different from usual. The `Transient Mark'
8517 and @code{interprogram-cut-function} comments explain certain
8520 After you once set a mark, a buffer always contains a region. If you
8521 wish, you can use Transient Mark mode to highlight the region
8522 temporarily. (No one wants to highlight the region all the time, so
8523 Transient Mark mode highlights it only at appropriate times. Many
8524 people turn off Transient Mark mode, so the region is never
8527 Also, a windowing system allows you to copy, cut, and paste among
8528 different programs. In the X windowing system, for example, the
8529 @code{interprogram-cut-function} function is @code{x-select-text},
8530 which works with the windowing system's equivalent of the Emacs kill
8533 The body of the @code{copy-region-as-kill} function starts with an
8534 @code{if} clause. What this clause does is distinguish between two
8535 different situations: whether or not this command is executed
8536 immediately after a previous @code{kill-region} command. In the first
8537 case, the new region is appended to the previously copied text.
8538 Otherwise, it is inserted into the beginning of the kill ring as a
8539 separate piece of text from the previous piece.
8541 The last two lines of the function prevent the region from lighting up
8542 if Transient Mark mode is turned on.
8544 The body of @code{copy-region-as-kill} merits discussion in detail.
8546 @node copy-region-as-kill body, , Complete copy-region-as-kill, copy-region-as-kill
8547 @comment node-name, next, previous, up
8548 @subsection The Body of @code{copy-region-as-kill}
8550 The @code{copy-region-as-kill} function works in much the same way as
8551 the @code{kill-region} function. Both are written so that two or more
8552 kills in a row combine their text into a single entry. If you yank
8553 back the text from the kill ring, you get it all in one piece.
8554 Moreover, kills that kill forward from the current position of the
8555 cursor are added to the end of the previously copied text and commands
8556 that copy text backwards add it to the beginning of the previously
8557 copied text. This way, the words in the text stay in the proper
8560 Like @code{kill-region}, the @code{copy-region-as-kill} function makes
8561 use of the @code{last-command} variable that keeps track of the
8562 previous Emacs command.
8565 * last-command & this-command::
8566 * kill-append function::
8567 * kill-new function::
8570 @node last-command & this-command, kill-append function, copy-region-as-kill body, copy-region-as-kill body
8572 @unnumberedsubsubsec @code{last-command} and @code{this-command}
8575 Normally, whenever a function is executed, Emacs sets the value of
8576 @code{this-command} to the function being executed (which in this case
8577 would be @code{copy-region-as-kill}). At the same time, Emacs sets
8578 the value of @code{last-command} to the previous value of
8579 @code{this-command}.
8581 In the first part of the body of the @code{copy-region-as-kill}
8582 function, an @code{if} expression determines whether the value of
8583 @code{last-command} is @code{kill-region}. If so, the then-part of
8584 the @code{if} expression is evaluated; it uses the @code{kill-append}
8585 function to concatenate the text copied at this call to the function
8586 with the text already in the first element (the @sc{car}) of the kill
8587 ring. On the other hand, if the value of @code{last-command} is not
8588 @code{kill-region}, then the @code{copy-region-as-kill} function
8589 attaches a new element to the kill ring using the @code{kill-new}
8593 The @code{if} expression reads as follows; it uses @code{eq}:
8597 (if (eq last-command 'kill-region)
8599 (kill-append (filter-buffer-substring beg end) (< end beg))
8601 (kill-new (filter-buffer-substring beg end)))
8605 @findex filter-buffer-substring
8606 (The @code{filter-buffer-substring} function returns a filtered
8607 substring of the buffer, if any. Optionally---the arguments are not
8608 here, so neither is done---the function may delete the initial text or
8609 return the text without its properties; this function is a replacement
8610 for the older @code{buffer-substring} function, which came before text
8611 properties were implemented.)
8613 @findex eq @r{(example of use)}
8615 The @code{eq} function tests whether its first argument is the same Lisp
8616 object as its second argument. The @code{eq} function is similar to the
8617 @code{equal} function in that it is used to test for equality, but
8618 differs in that it determines whether two representations are actually
8619 the same object inside the computer, but with different names.
8620 @code{equal} determines whether the structure and contents of two
8621 expressions are the same.
8623 If the previous command was @code{kill-region}, then the Emacs Lisp
8624 interpreter calls the @code{kill-append} function
8626 @node kill-append function, kill-new function, last-command & this-command, copy-region-as-kill body
8627 @unnumberedsubsubsec The @code{kill-append} function
8631 The @code{kill-append} function looks like this:
8636 (defun kill-append (string before-p &optional yank-handler)
8637 "Append STRING to the end of the latest kill in the kill ring.
8638 If BEFORE-P is non-nil, prepend STRING to the kill.
8640 (let* ((cur (car kill-ring)))
8641 (kill-new (if before-p (concat string cur) (concat cur string))
8642 (or (= (length cur) 0)
8644 (get-text-property 0 'yank-handler cur)))
8651 (defun kill-append (string before-p)
8652 "Append STRING to the end of the latest kill in the kill ring.
8653 If BEFORE-P is non-nil, prepend STRING to the kill.
8654 If `interprogram-cut-function' is set, pass the resulting kill to
8656 (kill-new (if before-p
8657 (concat string (car kill-ring))
8658 (concat (car kill-ring) string))
8663 The @code{kill-append} function is fairly straightforward. It uses
8664 the @code{kill-new} function, which we will discuss in more detail in
8667 (Also, the function provides an optional argument called
8668 @code{yank-handler}; when invoked, this argument tells the function
8669 how to deal with properties added to the text, such as `bold' or
8672 @c !!! bug in GNU Emacs 22 version of kill-append ?
8673 It has a @code{let*} function to set the value of the first element of
8674 the kill ring to @code{cur}. (I do not know why the function does not
8675 use @code{let} instead; only one value is set in the expression.
8676 Perhaps this is a bug that produces no problems?)
8678 Consider the conditional that is one of the two arguments to
8679 @code{kill-new}. It uses @code{concat} to concatenate the new text to
8680 the @sc{car} of the kill ring. Whether it prepends or appends the
8681 text depends on the results of an @code{if} expression:
8685 (if before-p ; @r{if-part}
8686 (concat string cur) ; @r{then-part}
8687 (concat cur string)) ; @r{else-part}
8692 If the region being killed is before the region that was killed in the
8693 last command, then it should be prepended before the material that was
8694 saved in the previous kill; and conversely, if the killed text follows
8695 what was just killed, it should be appended after the previous text.
8696 The @code{if} expression depends on the predicate @code{before-p} to
8697 decide whether the newly saved text should be put before or after the
8698 previously saved text.
8700 The symbol @code{before-p} is the name of one of the arguments to
8701 @code{kill-append}. When the @code{kill-append} function is
8702 evaluated, it is bound to the value returned by evaluating the actual
8703 argument. In this case, this is the expression @code{(< end beg)}.
8704 This expression does not directly determine whether the killed text in
8705 this command is located before or after the kill text of the last
8706 command; what it does is determine whether the value of the variable
8707 @code{end} is less than the value of the variable @code{beg}. If it
8708 is, it means that the user is most likely heading towards the
8709 beginning of the buffer. Also, the result of evaluating the predicate
8710 expression, @code{(< end beg)}, will be true and the text will be
8711 prepended before the previous text. On the other hand, if the value of
8712 the variable @code{end} is greater than the value of the variable
8713 @code{beg}, the text will be appended after the previous text.
8716 When the newly saved text will be prepended, then the string with the new
8717 text will be concatenated before the old text:
8725 But if the text will be appended, it will be concatenated
8729 (concat cur string))
8732 To understand how this works, we first need to review the
8733 @code{concat} function. The @code{concat} function links together or
8734 unites two strings of text. The result is a string. For example:
8738 (concat "abc" "def")
8744 (car '("first element" "second element")))
8745 @result{} "new first element"
8748 '("first element" "second element")) " modified")
8749 @result{} "first element modified"
8753 We can now make sense of @code{kill-append}: it modifies the contents
8754 of the kill ring. The kill ring is a list, each element of which is
8755 saved text. The @code{kill-append} function uses the @code{kill-new}
8756 function which in turn uses the @code{setcar} function.
8758 @node kill-new function, , kill-append function, copy-region-as-kill body
8759 @unnumberedsubsubsec The @code{kill-new} function
8762 @c in GNU Emacs 22, additional documentation to kill-new:
8764 Optional third arguments YANK-HANDLER controls how the STRING is later
8765 inserted into a buffer; see `insert-for-yank' for details.
8766 When a yank handler is specified, STRING must be non-empty (the yank
8767 handler, if non-nil, is stored as a `yank-handler' text property on STRING).
8769 When the yank handler has a non-nil PARAM element, the original STRING
8770 argument is not used by `insert-for-yank'. However, since Lisp code
8771 may access and use elements from the kill ring directly, the STRING
8772 argument should still be a \"useful\" string for such uses."
8775 The @code{kill-new} function looks like this:
8779 (defun kill-new (string &optional replace yank-handler)
8780 "Make STRING the latest kill in the kill ring.
8781 Set `kill-ring-yank-pointer' to point to it.
8783 If `interprogram-cut-function' is non-nil, apply it to STRING.
8784 Optional second argument REPLACE non-nil means that STRING will replace
8785 the front of the kill ring, rather than being added to the list.
8789 (if (> (length string) 0)
8791 (put-text-property 0 (length string)
8792 'yank-handler yank-handler string))
8794 (signal 'args-out-of-range
8795 (list string "yank-handler specified for empty string"))))
8798 (if (fboundp 'menu-bar-update-yank-menu)
8799 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8802 (if (and replace kill-ring)
8803 (setcar kill-ring string)
8804 (push string kill-ring)
8805 (if (> (length kill-ring) kill-ring-max)
8806 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8809 (setq kill-ring-yank-pointer kill-ring)
8810 (if interprogram-cut-function
8811 (funcall interprogram-cut-function string (not replace))))
8816 (defun kill-new (string &optional replace)
8817 "Make STRING the latest kill in the kill ring.
8818 Set the kill-ring-yank pointer to point to it.
8819 If `interprogram-cut-function' is non-nil, apply it to STRING.
8820 Optional second argument REPLACE non-nil means that STRING will replace
8821 the front of the kill ring, rather than being added to the list."
8822 (and (fboundp 'menu-bar-update-yank-menu)
8823 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
8824 (if (and replace kill-ring)
8825 (setcar kill-ring string)
8826 (setq kill-ring (cons string kill-ring))
8827 (if (> (length kill-ring) kill-ring-max)
8828 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8829 (setq kill-ring-yank-pointer kill-ring)
8830 (if interprogram-cut-function
8831 (funcall interprogram-cut-function string (not replace))))
8834 (Notice that the function is not interactive.)
8836 As usual, we can look at this function in parts.
8838 The function definition has an optional @code{yank-handler} argument,
8839 which when invoked tells the function how to deal with properties
8840 added to the text, such as `bold' or `italics'. We will skip that.
8843 The first line of the documentation makes sense:
8846 Make STRING the latest kill in the kill ring.
8850 Let's skip over the rest of the documentation for the moment.
8853 Also, let's skip over the initial @code{if} expression and those lines
8854 of code involving @code{menu-bar-update-yank-menu}. We will explain
8858 The critical lines are these:
8862 (if (and replace kill-ring)
8864 (setcar kill-ring string)
8868 (push string kill-ring)
8871 (setq kill-ring (cons string kill-ring))
8872 (if (> (length kill-ring) kill-ring-max)
8873 ;; @r{avoid overly long kill ring}
8874 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil)))
8877 (setq kill-ring-yank-pointer kill-ring)
8878 (if interprogram-cut-function
8879 (funcall interprogram-cut-function string (not replace))))
8883 The conditional test is @w{@code{(and replace kill-ring)}}.
8884 This will be true when two conditions are met: the kill ring has
8885 something in it, and the @code{replace} variable is true.
8888 When the @code{kill-append} function sets @code{replace} to be true
8889 and when the kill ring has at least one item in it, the @code{setcar}
8890 expression is executed:
8893 (setcar kill-ring string)
8896 The @code{setcar} function actually changes the first element of the
8897 @code{kill-ring} list to the value of @code{string}. It replaces the
8901 On the other hand, if the kill ring is empty, or replace is false, the
8902 else-part of the condition is executed:
8905 (push string kill-ring)
8910 @code{push} puts its first argument onto the second. It is similar to
8914 (setq kill-ring (cons string kill-ring))
8922 (add-to-list kill-ring string)
8926 When it is false, the expression first constructs a new version of the
8927 kill ring by prepending @code{string} to the existing kill ring as a
8928 new element (that is what the @code{push} does). Then it executes a
8929 second @code{if} clause. This second @code{if} clause keeps the kill
8930 ring from growing too long.
8932 Let's look at these two expressions in order.
8934 The @code{push} line of the else-part sets the new value of the kill
8935 ring to what results from adding the string being killed to the old
8938 We can see how this works with an example.
8944 (setq example-list '("here is a clause" "another clause"))
8949 After evaluating this expression with @kbd{C-x C-e}, you can evaluate
8950 @code{example-list} and see what it returns:
8955 @result{} ("here is a clause" "another clause")
8961 Now, we can add a new element on to this list by evaluating the
8962 following expression:
8963 @findex push, @r{example}
8966 (push "a third clause" example-list)
8971 When we evaluate @code{example-list}, we find its value is:
8976 @result{} ("a third clause" "here is a clause" "another clause")
8981 Thus, the third clause is added to the list by @code{push}.
8984 Now for the second part of the @code{if} clause. This expression
8985 keeps the kill ring from growing too long. It looks like this:
8989 (if (> (length kill-ring) kill-ring-max)
8990 (setcdr (nthcdr (1- kill-ring-max) kill-ring) nil))
8994 The code checks whether the length of the kill ring is greater than
8995 the maximum permitted length. This is the value of
8996 @code{kill-ring-max} (which is 60, by default). If the length of the
8997 kill ring is too long, then this code sets the last element of the
8998 kill ring to @code{nil}. It does this by using two functions,
8999 @code{nthcdr} and @code{setcdr}.
9001 We looked at @code{setcdr} earlier (@pxref{setcdr, , @code{setcdr}}).
9002 It sets the @sc{cdr} of a list, just as @code{setcar} sets the
9003 @sc{car} of a list. In this case, however, @code{setcdr} will not be
9004 setting the @sc{cdr} of the whole kill ring; the @code{nthcdr}
9005 function is used to cause it to set the @sc{cdr} of the next to last
9006 element of the kill ring---this means that since the @sc{cdr} of the
9007 next to last element is the last element of the kill ring, it will set
9008 the last element of the kill ring.
9010 @findex nthcdr, @r{example}
9011 The @code{nthcdr} function works by repeatedly taking the @sc{cdr} of a
9012 list---it takes the @sc{cdr} of the @sc{cdr} of the @sc{cdr}
9013 @dots{} It does this @var{N} times and returns the results.
9014 (@xref{nthcdr, , @code{nthcdr}}.)
9016 @findex setcdr, @r{example}
9017 Thus, if we had a four element list that was supposed to be three
9018 elements long, we could set the @sc{cdr} of the next to last element
9019 to @code{nil}, and thereby shorten the list. (If you set the last
9020 element to some other value than @code{nil}, which you could do, then
9021 you would not have shortened the list. @xref{setcdr, ,
9024 You can see shortening by evaluating the following three expressions
9025 in turn. First set the value of @code{trees} to @code{(maple oak pine
9026 birch)}, then set the @sc{cdr} of its second @sc{cdr} to @code{nil}
9027 and then find the value of @code{trees}:
9031 (setq trees '(maple oak pine birch))
9032 @result{} (maple oak pine birch)
9036 (setcdr (nthcdr 2 trees) nil)
9040 @result{} (maple oak pine)
9045 (The value returned by the @code{setcdr} expression is @code{nil} since
9046 that is what the @sc{cdr} is set to.)
9048 To repeat, in @code{kill-new}, the @code{nthcdr} function takes the
9049 @sc{cdr} a number of times that is one less than the maximum permitted
9050 size of the kill ring and @code{setcdr} sets the @sc{cdr} of that
9051 element (which will be the rest of the elements in the kill ring) to
9052 @code{nil}. This prevents the kill ring from growing too long.
9055 The next to last expression in the @code{kill-new} function is
9058 (setq kill-ring-yank-pointer kill-ring)
9061 The @code{kill-ring-yank-pointer} is a global variable that is set to be
9062 the @code{kill-ring}.
9064 Even though the @code{kill-ring-yank-pointer} is called a
9065 @samp{pointer}, it is a variable just like the kill ring. However, the
9066 name has been chosen to help humans understand how the variable is used.
9069 Now, to return to an early expression in the body of the function:
9073 (if (fboundp 'menu-bar-update-yank-menu)
9074 (menu-bar-update-yank-menu string (and replace (car kill-ring))))
9079 It starts with an @code{if} expression
9081 In this case, the expression tests first to see whether
9082 @code{menu-bar-update-yank-menu} exists as a function, and if so,
9083 calls it. The @code{fboundp} function returns true if the symbol it
9084 is testing has a function definition that `is not void'. If the
9085 symbol's function definition were void, we would receive an error
9086 message, as we did when we created errors intentionally (@pxref{Making
9087 Errors, , Generate an Error Message}).
9090 The then-part contains an expression whose first element is the
9091 function @code{and}.
9094 The @code{and} special form evaluates each of its arguments until one
9095 of the arguments returns a value of @code{nil}, in which case the
9096 @code{and} expression returns @code{nil}; however, if none of the
9097 arguments returns a value of @code{nil}, the value resulting from
9098 evaluating the last argument is returned. (Since such a value is not
9099 @code{nil}, it is considered true in Emacs Lisp.) In other words, an
9100 @code{and} expression returns a true value only if all its arguments
9101 are true. (@xref{Second Buffer Related Review}.)
9103 The expression determines whether the second argument to
9104 @code{menu-bar-update-yank-menu} is true or not.
9106 ;; If we're supposed to be extending an existing string, and that
9107 ;; string really is at the front of the menu, then update it in place.
9110 @code{menu-bar-update-yank-menu} is one of the functions that make it
9111 possible to use the `Select and Paste' menu in the Edit item of a menu
9112 bar; using a mouse, you can look at the various pieces of text you
9113 have saved and select one piece to paste.
9115 The last expression in the @code{kill-new} function adds the newly
9116 copied string to whatever facility exists for copying and pasting
9117 among different programs running in a windowing system. In the X
9118 Windowing system, for example, the @code{x-select-text} function takes
9119 the string and stores it in memory operated by X. You can paste the
9120 string in another program, such as an Xterm.
9123 The expression looks like this:
9127 (if interprogram-cut-function
9128 (funcall interprogram-cut-function string (not replace))))
9132 If an @code{interprogram-cut-function} exists, then Emacs executes
9133 @code{funcall}, which in turn calls its first argument as a function
9134 and passes the remaining arguments to it. (Incidentally, as far as I
9135 can see, this @code{if} expression could be replaced by an @code{and}
9136 expression similar to the one in the first part of the function.)
9138 We are not going to discuss windowing systems and other programs
9139 further, but merely note that this is a mechanism that enables GNU
9140 Emacs to work easily and well with other programs.
9142 This code for placing text in the kill ring, either concatenated with
9143 an existing element or as a new element, leads us to the code for
9144 bringing back text that has been cut out of the buffer---the yank
9145 commands. However, before discussing the yank commands, it is better
9146 to learn how lists are implemented in a computer. This will make
9147 clear such mysteries as the use of the term `pointer'. But before
9148 that, we will digress into C.
9151 @c is this true in Emacs 22? Does not seems to be
9153 (If the @w{@code{(< end beg))}}
9154 expression is true, @code{kill-append} prepends the string to the just
9155 previously clipped text. For a detailed discussion, see
9156 @ref{kill-append function, , The @code{kill-append} function}.)
9158 If you then yank back the text, i.e., `paste' it, you get both
9159 pieces of text at once. That way, if you delete two words in a row,
9160 and then yank them back, you get both words, in their proper order,
9161 with one yank. (The @w{@code{(< end beg))}} expression makes sure the
9164 On the other hand, if the previous command is not @code{kill-region},
9165 then the @code{kill-new} function is called, which adds the text to
9166 the kill ring as the latest item, and sets the
9167 @code{kill-ring-yank-pointer} variable to point to it.
9171 @c Evidently, changed for Emacs 22. The zap-to-char command does not
9172 @c use the delete-and-extract-region function
9174 2006 Oct 26, the Digression into C is now OK but should come after
9175 copy-region-as-kill and filter-buffer-substring
9179 copy-region-as-kill is short, 12 lines, and uses
9180 filter-buffer-substring, which is longer, 39 lines
9181 and has delete-and-extract-region in it.
9182 delete-and-extract-region is written in C.
9184 see Initializing a Variable with @code{defvar}
9187 @node Digression into C, defvar, copy-region-as-kill, Cutting & Storing Text
9188 @comment node-name, next, previous, up
9189 @section Digression into C
9190 @findex delete-and-extract-region
9191 @cindex C, a digression into
9192 @cindex Digression into C
9194 The @code{copy-region-as-kill} function (@pxref{copy-region-as-kill, ,
9195 @code{copy-region-as-kill}}) uses the @code{filter-buffer-substring}
9196 function, which in turn uses the @code{delete-and-extract-region}
9197 function. It removes the contents of a region and you cannot get them
9200 Unlike the other code discussed here, the
9201 @code{delete-and-extract-region} function is not written in Emacs
9202 Lisp; it is written in C and is one of the primitives of the GNU Emacs
9203 system. Since it is very simple, I will digress briefly from Lisp and
9206 @c GNU Emacs 22 in /usr/local/src/emacs/src/editfns.c
9207 @c the DEFUN for buffer-substring-no-properties
9210 Like many of the other Emacs primitives,
9211 @code{delete-and-extract-region} is written as an instance of a C
9212 macro, a macro being a template for code. The complete macro looks
9217 DEFUN ("buffer-substring-no-properties", Fbuffer_substring_no_properties,
9218 Sbuffer_substring_no_properties, 2, 2, 0,
9219 doc: /* Return the characters of part of the buffer,
9220 without the text properties.
9221 The two arguments START and END are character positions;
9222 they can be in either order. */)
9224 Lisp_Object start, end;
9228 validate_region (&start, &end);
9232 return make_buffer_string (b, e, 0);
9237 Without going into the details of the macro writing process, let me
9238 point out that this macro starts with the word @code{DEFUN}. The word
9239 @code{DEFUN} was chosen since the code serves the same purpose as
9240 @code{defun} does in Lisp. (The @code{DEFUN} C macro is defined in
9241 @file{emacs/src/lisp.h}.)
9243 The word @code{DEFUN} is followed by seven parts inside of
9248 The first part is the name given to the function in Lisp,
9249 @code{delete-and-extract-region}.
9252 The second part is the name of the function in C,
9253 @code{Fdelete_and_extract_region}. By convention, it starts with
9254 @samp{F}. Since C does not use hyphens in names, underscores are used
9258 The third part is the name for the C constant structure that records
9259 information on this function for internal use. It is the name of the
9260 function in C but begins with an @samp{S} instead of an @samp{F}.
9263 The fourth and fifth parts specify the minimum and maximum number of
9264 arguments the function can have. This function demands exactly 2
9268 The sixth part is nearly like the argument that follows the
9269 @code{interactive} declaration in a function written in Lisp: a letter
9270 followed, perhaps, by a prompt. The only difference from the Lisp is
9271 when the macro is called with no arguments. Then you write a @code{0}
9272 (which is a `null string'), as in this macro.
9274 If you were to specify arguments, you would place them between
9275 quotation marks. The C macro for @code{goto-char} includes
9276 @code{"NGoto char: "} in this position to indicate that the function
9277 expects a raw prefix, in this case, a numerical location in a buffer,
9278 and provides a prompt.
9281 The seventh part is a documentation string, just like the one for a
9282 function written in Emacs Lisp, except that every newline must be
9283 written explicitly as @samp{\n} followed by a backslash and carriage
9287 Thus, the first two lines of documentation for @code{goto-char} are
9292 "Set point to POSITION, a number or marker.\n\
9293 Beginning of buffer is position (point-min), end is (point-max)."
9299 In a C macro, the formal parameters come next, with a statement of
9300 what kind of object they are, followed by what might be called the `body'
9301 of the macro. For @code{delete-and-extract-region} the `body'
9302 consists of the following four lines:
9306 validate_region (&start, &end);
9307 if (XINT (start) == XINT (end))
9308 return build_string ("");
9309 return del_range_1 (XINT (start), XINT (end), 1, 1);
9313 The @code{validate_region} function checks whether the values
9314 passed as the beginning and end of the region are the proper type and
9315 are within range. If the beginning and end positions are the same,
9316 then return and empty string.
9318 The @code{del_range_1} function actually deletes the text. It is a
9319 complex function we will not look into. It updates the buffer and
9320 does other things. However, it is worth looking at the two arguments
9321 passed to @code{del_range}. These are @w{@code{XINT (start)}} and
9322 @w{@code{XINT (end)}}.
9324 As far as the C language is concerned, @code{start} and @code{end} are
9325 two integers that mark the beginning and end of the region to be
9326 deleted@footnote{More precisely, and requiring more expert knowledge
9327 to understand, the two integers are of type `Lisp_Object', which can
9328 also be a C union instead of an integer type.}.
9330 In early versions of Emacs, these two numbers were thirty-two bits
9331 long, but the code is slowly being generalized to handle other
9332 lengths. Three of the available bits are used to specify the type of
9333 information; the remaining bits are used as `content'.
9335 @samp{XINT} is a C macro that extracts the relevant number from the
9336 longer collection of bits; the three other bits are discarded.
9339 The command in @code{delete-and-extract-region} looks like this:
9342 del_range_1 (XINT (start), XINT (end), 1, 1);
9346 It deletes the region between the beginning position, @code{start},
9347 and the ending position, @code{end}.
9349 From the point of view of the person writing Lisp, Emacs is all very
9350 simple; but hidden underneath is a great deal of complexity to make it
9353 @node defvar, cons & search-fwd Review, Digression into C, Cutting & Storing Text
9354 @comment node-name, next, previous, up
9355 @section Initializing a Variable with @code{defvar}
9357 @cindex Initializing a variable
9358 @cindex Variable initialization
9363 copy-region-as-kill is short, 12 lines, and uses
9364 filter-buffer-substring, which is longer, 39 lines
9365 and has delete-and-extract-region in it.
9366 delete-and-extract-region is written in C.
9368 see Initializing a Variable with @code{defvar}
9372 The @code{copy-region-as-kill} function is written in Emacs Lisp. Two
9373 functions within it, @code{kill-append} and @code{kill-new}, copy a
9374 region in a buffer and save it in a variable called the
9375 @code{kill-ring}. This section describes how the @code{kill-ring}
9376 variable is created and initialized using the @code{defvar} special
9379 (Again we note that the term @code{kill-ring} is a misnomer. The text
9380 that is clipped out of the buffer can be brought back; it is not a ring
9381 of corpses, but a ring of resurrectable text.)
9383 In Emacs Lisp, a variable such as the @code{kill-ring} is created and
9384 given an initial value by using the @code{defvar} special form. The
9385 name comes from ``define variable''.
9387 The @code{defvar} special form is similar to @code{setq} in that it sets
9388 the value of a variable. It is unlike @code{setq} in two ways: first,
9389 it only sets the value of the variable if the variable does not already
9390 have a value. If the variable already has a value, @code{defvar} does
9391 not override the existing value. Second, @code{defvar} has a
9392 documentation string.
9394 (Another special form, @code{defcustom}, is designed for variables
9395 that people customize. It has more features than @code{defvar}.
9396 (@xref{defcustom, , Setting Variables with @code{defcustom}}.)
9399 * See variable current value::
9400 * defvar and asterisk::
9403 @node See variable current value, defvar and asterisk, defvar, defvar
9405 @unnumberedsubsec Seeing the Current Value of a Variable
9408 You can see the current value of a variable, any variable, by using
9409 the @code{describe-variable} function, which is usually invoked by
9410 typing @kbd{C-h v}. If you type @kbd{C-h v} and then @code{kill-ring}
9411 (followed by @key{RET}) when prompted, you will see what is in your
9412 current kill ring---this may be quite a lot! Conversely, if you have
9413 been doing nothing this Emacs session except read this document, you
9414 may have nothing in it. Also, you will see the documentation for
9420 List of killed text sequences.
9421 Since the kill ring is supposed to interact nicely with cut-and-paste
9422 facilities offered by window systems, use of this variable should
9425 interact nicely with `interprogram-cut-function' and
9426 `interprogram-paste-function'. The functions `kill-new',
9427 `kill-append', and `current-kill' are supposed to implement this
9428 interaction; you may want to use them instead of manipulating the kill
9434 The kill ring is defined by a @code{defvar} in the following way:
9438 (defvar kill-ring nil
9439 "List of killed text sequences.
9445 In this variable definition, the variable is given an initial value of
9446 @code{nil}, which makes sense, since if you have saved nothing, you want
9447 nothing back if you give a @code{yank} command. The documentation
9448 string is written just like the documentation string of a @code{defun}.
9449 As with the documentation string of the @code{defun}, the first line of
9450 the documentation should be a complete sentence, since some commands,
9451 like @code{apropos}, print only the first line of documentation.
9452 Succeeding lines should not be indented; otherwise they look odd when
9453 you use @kbd{C-h v} (@code{describe-variable}).
9455 @node defvar and asterisk, , See variable current value, defvar
9456 @subsection @code{defvar} and an asterisk
9457 @findex defvar @r{for a user customizable variable}
9458 @findex defvar @r{with an asterisk}
9460 In the past, Emacs used the @code{defvar} special form both for
9461 internal variables that you would not expect a user to change and for
9462 variables that you do expect a user to change. Although you can still
9463 use @code{defvar} for user customizable variables, please use
9464 @code{defcustom} instead, since that special form provides a path into
9465 the Customization commands. (@xref{defcustom, , Specifying Variables
9466 using @code{defcustom}}.)
9468 When you specified a variable using the @code{defvar} special form,
9469 you could distinguish a variable that a user might want to change from
9470 others by typing an asterisk, @samp{*}, in the first column of its
9471 documentation string. For example:
9475 (defvar shell-command-default-error-buffer nil
9476 "*Buffer name for `shell-command' @dots{} error output.
9481 @findex set-variable
9483 You could (and still can) use the @code{set-variable} command to
9484 change the value of @code{shell-command-default-error-buffer}
9485 temporarily. However, options set using @code{set-variable} are set
9486 only for the duration of your editing session. The new values are not
9487 saved between sessions. Each time Emacs starts, it reads the original
9488 value, unless you change the value within your @file{.emacs} file,
9489 either by setting it manually or by using @code{customize}.
9490 @xref{Emacs Initialization, , Your @file{.emacs} File}.
9492 For me, the major use of the @code{set-variable} command is to suggest
9493 variables that I might want to set in my @file{.emacs} file. There
9494 are now more than 700 such variables --- far too many to remember
9495 readily. Fortunately, you can press @key{TAB} after calling the
9496 @code{M-x set-variable} command to see the list of variables.
9497 (@xref{Examining, , Examining and Setting Variables, emacs,
9498 The GNU Emacs Manual}.)
9501 @node cons & search-fwd Review, search Exercises, defvar, Cutting & Storing Text
9502 @comment node-name, next, previous, up
9505 Here is a brief summary of some recently introduced functions.
9510 @code{car} returns the first element of a list; @code{cdr} returns the
9511 second and subsequent elements of a list.
9518 (car '(1 2 3 4 5 6 7))
9520 (cdr '(1 2 3 4 5 6 7))
9521 @result{} (2 3 4 5 6 7)
9526 @code{cons} constructs a list by prepending its first argument to its
9540 @code{funcall} evaluates its first argument as a function. It passes
9541 its remaining arguments to its first argument.
9544 Return the result of taking @sc{cdr} `n' times on a list.
9552 The `rest of the rest', as it were.
9559 (nthcdr 3 '(1 2 3 4 5 6 7))
9566 @code{setcar} changes the first element of a list; @code{setcdr}
9567 changes the second and subsequent elements of a list.
9574 (setq triple '(1 2 3))
9581 (setcdr triple '("foo" "bar"))
9584 @result{} (37 "foo" "bar")
9589 Evaluate each argument in sequence and then return the value of the
9602 @item save-restriction
9603 Record whatever narrowing is in effect in the current buffer, if any,
9604 and restore that narrowing after evaluating the arguments.
9606 @item search-forward
9607 Search for a string, and if the string is found, move point. With a
9608 regular expression, use the similar @code{re-search-forward}.
9609 (@xref{Regexp Search, , Regular Expression Searches}, for an
9610 explanation of regular expression patterns and searches.)
9614 @code{search-forward} and @code{re-search-forward} take four
9619 The string or regular expression to search for.
9622 Optionally, the limit of the search.
9625 Optionally, what to do if the search fails, return @code{nil} or an
9629 Optionally, how many times to repeat the search; if negative, the
9630 search goes backwards.
9634 @itemx delete-and-extract-region
9635 @itemx copy-region-as-kill
9637 @code{kill-region} cuts the text between point and mark from the
9638 buffer and stores that text in the kill ring, so you can get it back
9641 @code{copy-region-as-kill} copies the text between point and mark into
9642 the kill ring, from which you can get it by yanking. The function
9643 does not cut or remove the text from the buffer.
9646 @code{delete-and-extract-region} removes the text between point and
9647 mark from the buffer and throws it away. You cannot get it back.
9648 (This is not an interactive command.)
9651 @node search Exercises, , cons & search-fwd Review, Cutting & Storing Text
9652 @section Searching Exercises
9656 Write an interactive function that searches for a string. If the
9657 search finds the string, leave point after it and display a message
9658 that says ``Found!''. (Do not use @code{search-forward} for the name
9659 of this function; if you do, you will overwrite the existing version of
9660 @code{search-forward} that comes with Emacs. Use a name such as
9661 @code{test-search} instead.)
9664 Write a function that prints the third element of the kill ring in the
9665 echo area, if any; if the kill ring does not contain a third element,
9666 print an appropriate message.
9669 @node List Implementation, Yanking, Cutting & Storing Text, Top
9670 @comment node-name, next, previous, up
9671 @chapter How Lists are Implemented
9672 @cindex Lists in a computer
9674 In Lisp, atoms are recorded in a straightforward fashion; if the
9675 implementation is not straightforward in practice, it is, nonetheless,
9676 straightforward in theory. The atom @samp{rose}, for example, is
9677 recorded as the four contiguous letters @samp{r}, @samp{o}, @samp{s},
9678 @samp{e}. A list, on the other hand, is kept differently. The mechanism
9679 is equally simple, but it takes a moment to get used to the idea. A
9680 list is kept using a series of pairs of pointers. In the series, the
9681 first pointer in each pair points to an atom or to another list, and the
9682 second pointer in each pair points to the next pair, or to the symbol
9683 @code{nil}, which marks the end of the list.
9685 A pointer itself is quite simply the electronic address of what is
9686 pointed to. Hence, a list is kept as a series of electronic addresses.
9689 * Lists diagrammed::
9690 * Symbols as Chest:: Exploring a powerful metaphor.
9694 @node Lists diagrammed, Symbols as Chest, List Implementation, List Implementation
9696 @unnumberedsec Lists diagrammed
9699 For example, the list @code{(rose violet buttercup)} has three elements,
9700 @samp{rose}, @samp{violet}, and @samp{buttercup}. In the computer, the
9701 electronic address of @samp{rose} is recorded in a segment of computer
9702 memory along with the address that gives the electronic address of where
9703 the atom @samp{violet} is located; and that address (the one that tells
9704 where @samp{violet} is located) is kept along with an address that tells
9705 where the address for the atom @samp{buttercup} is located.
9708 This sounds more complicated than it is and is easier seen in a diagram:
9710 @c clear print-postscript-figures
9711 @c !!! cons-cell-diagram #1
9715 ___ ___ ___ ___ ___ ___
9716 |___|___|--> |___|___|--> |___|___|--> nil
9719 --> rose --> violet --> buttercup
9723 @ifset print-postscript-figures
9726 @center @image{cons-1}
9727 %%%% old method of including an image
9728 % \input /usr/local/lib/tex/inputs/psfig.tex
9729 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-1.eps}}
9734 @ifclear print-postscript-figures
9738 ___ ___ ___ ___ ___ ___
9739 |___|___|--> |___|___|--> |___|___|--> nil
9742 --> rose --> violet --> buttercup
9749 In the diagram, each box represents a word of computer memory that
9750 holds a Lisp object, usually in the form of a memory address. The boxes,
9751 i.e.@: the addresses, are in pairs. Each arrow points to what the address
9752 is the address of, either an atom or another pair of addresses. The
9753 first box is the electronic address of @samp{rose} and the arrow points
9754 to @samp{rose}; the second box is the address of the next pair of boxes,
9755 the first part of which is the address of @samp{violet} and the second
9756 part of which is the address of the next pair. The very last box
9757 points to the symbol @code{nil}, which marks the end of the list.
9760 When a variable is set to a list with a function such as @code{setq},
9761 it stores the address of the first box in the variable. Thus,
9762 evaluation of the expression
9765 (setq bouquet '(rose violet buttercup))
9770 creates a situation like this:
9772 @c cons-cell-diagram #2
9778 | ___ ___ ___ ___ ___ ___
9779 --> |___|___|--> |___|___|--> |___|___|--> nil
9782 --> rose --> violet --> buttercup
9786 @ifset print-postscript-figures
9789 @center @image{cons-2}
9790 %%%% old method of including an image
9791 % \input /usr/local/lib/tex/inputs/psfig.tex
9792 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2.eps}}
9797 @ifclear print-postscript-figures
9803 | ___ ___ ___ ___ ___ ___
9804 --> |___|___|--> |___|___|--> |___|___|--> nil
9807 --> rose --> violet --> buttercup
9814 In this example, the symbol @code{bouquet} holds the address of the first
9818 This same list can be illustrated in a different sort of box notation
9821 @c cons-cell-diagram #2a
9827 | -------------- --------------- ----------------
9828 | | car | cdr | | car | cdr | | car | cdr |
9829 -->| rose | o------->| violet | o------->| butter- | nil |
9830 | | | | | | | cup | |
9831 -------------- --------------- ----------------
9835 @ifset print-postscript-figures
9838 @center @image{cons-2a}
9839 %%%% old method of including an image
9840 % \input /usr/local/lib/tex/inputs/psfig.tex
9841 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-2a.eps}}
9846 @ifclear print-postscript-figures
9852 | -------------- --------------- ----------------
9853 | | car | cdr | | car | cdr | | car | cdr |
9854 -->| rose | o------->| violet | o------->| butter- | nil |
9855 | | | | | | | cup | |
9856 -------------- --------------- ----------------
9862 (Symbols consist of more than pairs of addresses, but the structure of
9863 a symbol is made up of addresses. Indeed, the symbol @code{bouquet}
9864 consists of a group of address-boxes, one of which is the address of
9865 the printed word @samp{bouquet}, a second of which is the address of a
9866 function definition attached to the symbol, if any, a third of which
9867 is the address of the first pair of address-boxes for the list
9868 @code{(rose violet buttercup)}, and so on. Here we are showing that
9869 the symbol's third address-box points to the first pair of
9870 address-boxes for the list.)
9872 If a symbol is set to the @sc{cdr} of a list, the list itself is not
9873 changed; the symbol simply has an address further down the list. (In
9874 the jargon, @sc{car} and @sc{cdr} are `non-destructive'.) Thus,
9875 evaluation of the following expression
9878 (setq flowers (cdr bouquet))
9885 @c cons-cell-diagram #3
9892 | ___ ___ | ___ ___ ___ ___
9893 --> | | | --> | | | | | |
9894 |___|___|----> |___|___|--> |___|___|--> nil
9897 --> rose --> violet --> buttercup
9902 @ifset print-postscript-figures
9905 @center @image{cons-3}
9906 %%%% old method of including an image
9907 % \input /usr/local/lib/tex/inputs/psfig.tex
9908 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-3.eps}}
9913 @ifclear print-postscript-figures
9920 | ___ ___ | ___ ___ ___ ___
9921 --> | | | --> | | | | | |
9922 |___|___|----> |___|___|--> |___|___|--> nil
9925 --> rose --> violet --> buttercup
9933 The value of @code{flowers} is @code{(violet buttercup)}, which is
9934 to say, the symbol @code{flowers} holds the address of the pair of
9935 address-boxes, the first of which holds the address of @code{violet},
9936 and the second of which holds the address of @code{buttercup}.
9938 A pair of address-boxes is called a @dfn{cons cell} or @dfn{dotted
9939 pair}. @xref{Cons Cell Type, , Cons Cell and List Types, elisp, The GNU Emacs Lisp
9940 Reference Manual}, and @ref{Dotted Pair Notation, , Dotted Pair
9941 Notation, elisp, The GNU Emacs Lisp Reference Manual}, for more
9942 information about cons cells and dotted pairs.
9945 The function @code{cons} adds a new pair of addresses to the front of
9946 a series of addresses like that shown above. For example, evaluating
9950 (setq bouquet (cons 'lily bouquet))
9957 @c cons-cell-diagram #4
9964 | ___ ___ ___ ___ | ___ ___ ___ ___
9965 --> | | | | | | --> | | | | | |
9966 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9969 --> lily --> rose --> violet --> buttercup
9974 @ifset print-postscript-figures
9977 @center @image{cons-4}
9978 %%%% old method of including an image
9979 % \input /usr/local/lib/tex/inputs/psfig.tex
9980 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-4.eps}}
9985 @ifclear print-postscript-figures
9992 | ___ ___ ___ ___ | ___ ___ ___ ___
9993 --> | | | | | | --> | | | | | |
9994 |___|___|----> |___|___|----> |___|___|---->|___|___|--> nil
9997 --> lily --> rose --> violet --> buttercup
10006 However, this does not change the value of the symbol
10007 @code{flowers}, as you can see by evaluating the following,
10010 (eq (cdr (cdr bouquet)) flowers)
10014 which returns @code{t} for true.
10016 Until it is reset, @code{flowers} still has the value
10017 @code{(violet buttercup)}; that is, it has the address of the cons
10018 cell whose first address is of @code{violet}. Also, this does not
10019 alter any of the pre-existing cons cells; they are all still there.
10021 Thus, in Lisp, to get the @sc{cdr} of a list, you just get the address
10022 of the next cons cell in the series; to get the @sc{car} of a list,
10023 you get the address of the first element of the list; to @code{cons} a
10024 new element on a list, you add a new cons cell to the front of the list.
10025 That is all there is to it! The underlying structure of Lisp is
10026 brilliantly simple!
10028 And what does the last address in a series of cons cells refer to? It
10029 is the address of the empty list, of @code{nil}.
10031 In summary, when a Lisp variable is set to a value, it is provided with
10032 the address of the list to which the variable refers.
10034 @node Symbols as Chest, List Exercise, Lists diagrammed, List Implementation
10035 @section Symbols as a Chest of Drawers
10036 @cindex Symbols as a Chest of Drawers
10037 @cindex Chest of Drawers, metaphor for a symbol
10038 @cindex Drawers, Chest of, metaphor for a symbol
10040 In an earlier section, I suggested that you might imagine a symbol as
10041 being a chest of drawers. The function definition is put in one
10042 drawer, the value in another, and so on. What is put in the drawer
10043 holding the value can be changed without affecting the contents of the
10044 drawer holding the function definition, and vice-verse.
10046 Actually, what is put in each drawer is the address of the value or
10047 function definition. It is as if you found an old chest in the attic,
10048 and in one of its drawers you found a map giving you directions to
10049 where the buried treasure lies.
10051 (In addition to its name, symbol definition, and variable value, a
10052 symbol has a `drawer' for a @dfn{property list} which can be used to
10053 record other information. Property lists are not discussed here; see
10054 @ref{Property Lists, , Property Lists, elisp, The GNU Emacs Lisp
10055 Reference Manual}.)
10058 Here is a fanciful representation:
10060 @c chest-of-drawers diagram
10065 Chest of Drawers Contents of Drawers
10069 ---------------------
10070 | directions to | [map to]
10071 | symbol name | bouquet
10073 +---------------------+
10075 | symbol definition | [none]
10077 +---------------------+
10078 | directions to | [map to]
10079 | variable value | (rose violet buttercup)
10081 +---------------------+
10083 | property list | [not described here]
10085 +---------------------+
10091 @ifset print-postscript-figures
10094 @center @image{drawers}
10095 %%%% old method of including an image
10096 % \input /usr/local/lib/tex/inputs/psfig.tex
10097 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/drawers.eps}}
10102 @ifclear print-postscript-figures
10107 Chest of Drawers Contents of Drawers
10111 ---------------------
10112 | directions to | [map to]
10113 | symbol name | bouquet
10115 +---------------------+
10117 | symbol definition | [none]
10119 +---------------------+
10120 | directions to | [map to]
10121 | variable value | (rose violet buttercup)
10123 +---------------------+
10125 | property list | [not described here]
10127 +---------------------+
10135 @node List Exercise, , Symbols as Chest, List Implementation
10138 Set @code{flowers} to @code{violet} and @code{buttercup}. Cons two
10139 more flowers on to this list and set this new list to
10140 @code{more-flowers}. Set the @sc{car} of @code{flowers} to a fish.
10141 What does the @code{more-flowers} list now contain?
10143 @node Yanking, Loops & Recursion, List Implementation, Top
10144 @comment node-name, next, previous, up
10145 @chapter Yanking Text Back
10147 @cindex Text retrieval
10148 @cindex Retrieving text
10149 @cindex Pasting text
10151 Whenever you cut text out of a buffer with a `kill' command in GNU Emacs,
10152 you can bring it back with a `yank' command. The text that is cut out of
10153 the buffer is put in the kill ring and the yank commands insert the
10154 appropriate contents of the kill ring back into a buffer (not necessarily
10155 the original buffer).
10157 A simple @kbd{C-y} (@code{yank}) command inserts the first item from
10158 the kill ring into the current buffer. If the @kbd{C-y} command is
10159 followed immediately by @kbd{M-y}, the first element is replaced by
10160 the second element. Successive @kbd{M-y} commands replace the second
10161 element with the third, fourth, or fifth element, and so on. When the
10162 last element in the kill ring is reached, it is replaced by the first
10163 element and the cycle is repeated. (Thus the kill ring is called a
10164 `ring' rather than just a `list'. However, the actual data structure
10165 that holds the text is a list.
10166 @xref{Kill Ring, , Handling the Kill Ring}, for the details of how the
10167 list is handled as a ring.)
10170 * Kill Ring Overview::
10171 * kill-ring-yank-pointer:: The kill ring is a list.
10172 * yank nthcdr Exercises:: The @code{kill-ring-yank-pointer} variable.
10175 @node Kill Ring Overview, kill-ring-yank-pointer, Yanking, Yanking
10176 @comment node-name, next, previous, up
10177 @section Kill Ring Overview
10178 @cindex Kill ring overview
10180 The kill ring is a list of textual strings. This is what it looks like:
10183 ("some text" "a different piece of text" "yet more text")
10186 If this were the contents of my kill ring and I pressed @kbd{C-y}, the
10187 string of characters saying @samp{some text} would be inserted in this
10188 buffer where my cursor is located.
10190 The @code{yank} command is also used for duplicating text by copying it.
10191 The copied text is not cut from the buffer, but a copy of it is put on the
10192 kill ring and is inserted by yanking it back.
10194 Three functions are used for bringing text back from the kill ring:
10195 @code{yank}, which is usually bound to @kbd{C-y}; @code{yank-pop},
10196 which is usually bound to @kbd{M-y}; and @code{rotate-yank-pointer},
10197 which is used by the two other functions.
10199 These functions refer to the kill ring through a variable called the
10200 @code{kill-ring-yank-pointer}. Indeed, the insertion code for both the
10201 @code{yank} and @code{yank-pop} functions is:
10204 (insert (car kill-ring-yank-pointer))
10208 (Well, no more. In GNU Emacs 22, the function has been replaced by
10209 @code{insert-for-yank} which calls @code{insert-for-yank-1}
10210 repetitively for each @code{yank-handler} segment. In turn,
10211 @code{insert-for-yank-1} strips text properties from the inserted text
10212 according to @code{yank-excluded-properties}. Otherwise, it is just
10213 like @code{insert}. We will stick with plain @code{insert} since it
10214 is easier to understand.)
10216 To begin to understand how @code{yank} and @code{yank-pop} work, it is
10217 first necessary to look at the @code{kill-ring-yank-pointer} variable.
10219 @node kill-ring-yank-pointer, yank nthcdr Exercises, Kill Ring Overview, Yanking
10220 @comment node-name, next, previous, up
10221 @section The @code{kill-ring-yank-pointer} Variable
10223 @code{kill-ring-yank-pointer} is a variable, just as @code{kill-ring} is
10224 a variable. It points to something by being bound to the value of what
10225 it points to, like any other Lisp variable.
10228 Thus, if the value of the kill ring is:
10231 ("some text" "a different piece of text" "yet more text")
10236 and the @code{kill-ring-yank-pointer} points to the second clause, the
10237 value of @code{kill-ring-yank-pointer} is:
10240 ("a different piece of text" "yet more text")
10243 As explained in the previous chapter (@pxref{List Implementation}), the
10244 computer does not keep two different copies of the text being pointed to
10245 by both the @code{kill-ring} and the @code{kill-ring-yank-pointer}. The
10246 words ``a different piece of text'' and ``yet more text'' are not
10247 duplicated. Instead, the two Lisp variables point to the same pieces of
10248 text. Here is a diagram:
10250 @c cons-cell-diagram #5
10254 kill-ring kill-ring-yank-pointer
10256 | ___ ___ | ___ ___ ___ ___
10257 ---> | | | --> | | | | | |
10258 |___|___|----> |___|___|--> |___|___|--> nil
10261 | | --> "yet more text"
10263 | --> "a different piece of text"
10270 @ifset print-postscript-figures
10273 @center @image{cons-5}
10274 %%%% old method of including an image
10275 % \input /usr/local/lib/tex/inputs/psfig.tex
10276 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/cons-5.eps}}
10281 @ifclear print-postscript-figures
10285 kill-ring kill-ring-yank-pointer
10287 | ___ ___ | ___ ___ ___ ___
10288 ---> | | | --> | | | | | |
10289 |___|___|----> |___|___|--> |___|___|--> nil
10292 | | --> "yet more text"
10294 | --> "a different piece of text
10303 Both the variable @code{kill-ring} and the variable
10304 @code{kill-ring-yank-pointer} are pointers. But the kill ring itself is
10305 usually described as if it were actually what it is composed of. The
10306 @code{kill-ring} is spoken of as if it were the list rather than that it
10307 points to the list. Conversely, the @code{kill-ring-yank-pointer} is
10308 spoken of as pointing to a list.
10310 These two ways of talking about the same thing sound confusing at first but
10311 make sense on reflection. The kill ring is generally thought of as the
10312 complete structure of data that holds the information of what has recently
10313 been cut out of the Emacs buffers. The @code{kill-ring-yank-pointer}
10314 on the other hand, serves to indicate---that is, to `point to'---that part
10315 of the kill ring of which the first element (the @sc{car}) will be
10319 In GNU Emacs 22, the @code{kill-new} function calls
10321 @code{(setq kill-ring-yank-pointer kill-ring)}
10323 (defun rotate-yank-pointer (arg)
10324 "Rotate the yanking point in the kill ring.
10325 With argument, rotate that many kills forward (or backward, if negative)."
10327 (current-kill arg))
10329 (defun current-kill (n &optional do-not-move)
10330 "Rotate the yanking point by N places, and then return that kill.
10331 If N is zero, `interprogram-paste-function' is set, and calling it
10332 returns a string, then that string is added to the front of the
10333 kill ring and returned as the latest kill.
10334 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
10335 yanking point; just return the Nth kill forward."
10336 (let ((interprogram-paste (and (= n 0)
10337 interprogram-paste-function
10338 (funcall interprogram-paste-function))))
10339 (if interprogram-paste
10341 ;; Disable the interprogram cut function when we add the new
10342 ;; text to the kill ring, so Emacs doesn't try to own the
10343 ;; selection, with identical text.
10344 (let ((interprogram-cut-function nil))
10345 (kill-new interprogram-paste))
10346 interprogram-paste)
10347 (or kill-ring (error "Kill ring is empty"))
10348 (let ((ARGth-kill-element
10349 (nthcdr (mod (- n (length kill-ring-yank-pointer))
10350 (length kill-ring))
10353 (setq kill-ring-yank-pointer ARGth-kill-element))
10354 (car ARGth-kill-element)))))
10359 @node yank nthcdr Exercises, , kill-ring-yank-pointer, Yanking
10360 @section Exercises with @code{yank} and @code{nthcdr}
10364 Using @kbd{C-h v} (@code{describe-variable}), look at the value of
10365 your kill ring. Add several items to your kill ring; look at its
10366 value again. Using @kbd{M-y} (@code{yank-pop)}, move all the way
10367 around the kill ring. How many items were in your kill ring? Find
10368 the value of @code{kill-ring-max}. Was your kill ring full, or could
10369 you have kept more blocks of text within it?
10372 Using @code{nthcdr} and @code{car}, construct a series of expressions
10373 to return the first, second, third, and fourth elements of a list.
10376 @node Loops & Recursion, Regexp Search, Yanking, Top
10377 @comment node-name, next, previous, up
10378 @chapter Loops and Recursion
10379 @cindex Loops and recursion
10380 @cindex Recursion and loops
10381 @cindex Repetition (loops)
10383 Emacs Lisp has two primary ways to cause an expression, or a series of
10384 expressions, to be evaluated repeatedly: one uses a @code{while}
10385 loop, and the other uses @dfn{recursion}.
10387 Repetition can be very valuable. For example, to move forward four
10388 sentences, you need only write a program that will move forward one
10389 sentence and then repeat the process four times. Since a computer does
10390 not get bored or tired, such repetitive action does not have the
10391 deleterious effects that excessive or the wrong kinds of repetition can
10394 People mostly write Emacs Lisp functions using @code{while} loops and
10395 their kin; but you can use recursion, which provides a very powerful
10396 way to think about and then to solve problems@footnote{You can write
10397 recursive functions to be frugal or wasteful of mental or computer
10398 resources; as it happens, methods that people find easy---that are
10399 frugal of `mental resources'---sometimes use considerable computer
10400 resources. Emacs was designed to run on machines that we now consider
10401 limited and its default settings are conservative. You may want to
10402 increase the values of @code{max-specpdl-size} and
10403 @code{max-lisp-eval-depth}. In my @file{.emacs} file, I set them to
10404 15 and 30 times their default value.}.
10407 * while:: Causing a stretch of code to repeat.
10409 * Recursion:: Causing a function to call itself.
10410 * Looping exercise::
10413 @node while, dolist dotimes, Loops & Recursion, Loops & Recursion
10414 @comment node-name, next, previous, up
10415 @section @code{while}
10419 The @code{while} special form tests whether the value returned by
10420 evaluating its first argument is true or false. This is similar to what
10421 the Lisp interpreter does with an @code{if}; what the interpreter does
10422 next, however, is different.
10424 In a @code{while} expression, if the value returned by evaluating the
10425 first argument is false, the Lisp interpreter skips the rest of the
10426 expression (the @dfn{body} of the expression) and does not evaluate it.
10427 However, if the value is true, the Lisp interpreter evaluates the body
10428 of the expression and then again tests whether the first argument to
10429 @code{while} is true or false. If the value returned by evaluating the
10430 first argument is again true, the Lisp interpreter again evaluates the
10431 body of the expression.
10434 The template for a @code{while} expression looks like this:
10438 (while @var{true-or-false-test}
10444 * Looping with while:: Repeat so long as test returns true.
10445 * Loop Example:: A @code{while} loop that uses a list.
10446 * print-elements-of-list:: Uses @code{while}, @code{car}, @code{cdr}.
10447 * Incrementing Loop:: A loop with an incrementing counter.
10448 * Incrementing Loop Details::
10449 * Decrementing Loop:: A loop with a decrementing counter.
10452 @node Looping with while, Loop Example, while, while
10454 @unnumberedsubsec Looping with @code{while}
10457 So long as the true-or-false-test of the @code{while} expression
10458 returns a true value when it is evaluated, the body is repeatedly
10459 evaluated. This process is called a loop since the Lisp interpreter
10460 repeats the same thing again and again, like an airplane doing a loop.
10461 When the result of evaluating the true-or-false-test is false, the
10462 Lisp interpreter does not evaluate the rest of the @code{while}
10463 expression and `exits the loop'.
10465 Clearly, if the value returned by evaluating the first argument to
10466 @code{while} is always true, the body following will be evaluated
10467 again and again @dots{} and again @dots{} forever. Conversely, if the
10468 value returned is never true, the expressions in the body will never
10469 be evaluated. The craft of writing a @code{while} loop consists of
10470 choosing a mechanism such that the true-or-false-test returns true
10471 just the number of times that you want the subsequent expressions to
10472 be evaluated, and then have the test return false.
10474 The value returned by evaluating a @code{while} is the value of the
10475 true-or-false-test. An interesting consequence of this is that a
10476 @code{while} loop that evaluates without error will return @code{nil}
10477 or false regardless of whether it has looped 1 or 100 times or none at
10478 all. A @code{while} expression that evaluates successfully never
10479 returns a true value! What this means is that @code{while} is always
10480 evaluated for its side effects, which is to say, the consequences of
10481 evaluating the expressions within the body of the @code{while} loop.
10482 This makes sense. It is not the mere act of looping that is desired,
10483 but the consequences of what happens when the expressions in the loop
10484 are repeatedly evaluated.
10486 @node Loop Example, print-elements-of-list, Looping with while, while
10487 @comment node-name, next, previous, up
10488 @subsection A @code{while} Loop and a List
10490 A common way to control a @code{while} loop is to test whether a list
10491 has any elements. If it does, the loop is repeated; but if it does not,
10492 the repetition is ended. Since this is an important technique, we will
10493 create a short example to illustrate it.
10495 A simple way to test whether a list has elements is to evaluate the
10496 list: if it has no elements, it is an empty list and will return the
10497 empty list, @code{()}, which is a synonym for @code{nil} or false. On
10498 the other hand, a list with elements will return those elements when it
10499 is evaluated. Since Emacs Lisp considers as true any value that is not
10500 @code{nil}, a list that returns elements will test true in a
10504 For example, you can set the variable @code{empty-list} to @code{nil} by
10505 evaluating the following @code{setq} expression:
10508 (setq empty-list ())
10512 After evaluating the @code{setq} expression, you can evaluate the
10513 variable @code{empty-list} in the usual way, by placing the cursor after
10514 the symbol and typing @kbd{C-x C-e}; @code{nil} will appear in your
10521 On the other hand, if you set a variable to be a list with elements, the
10522 list will appear when you evaluate the variable, as you can see by
10523 evaluating the following two expressions:
10527 (setq animals '(gazelle giraffe lion tiger))
10533 Thus, to create a @code{while} loop that tests whether there are any
10534 items in the list @code{animals}, the first part of the loop will be
10545 When the @code{while} tests its first argument, the variable
10546 @code{animals} is evaluated. It returns a list. So long as the list
10547 has elements, the @code{while} considers the results of the test to be
10548 true; but when the list is empty, it considers the results of the test
10551 To prevent the @code{while} loop from running forever, some mechanism
10552 needs to be provided to empty the list eventually. An oft-used
10553 technique is to have one of the subsequent forms in the @code{while}
10554 expression set the value of the list to be the @sc{cdr} of the list.
10555 Each time the @code{cdr} function is evaluated, the list will be made
10556 shorter, until eventually only the empty list will be left. At this
10557 point, the test of the @code{while} loop will return false, and the
10558 arguments to the @code{while} will no longer be evaluated.
10560 For example, the list of animals bound to the variable @code{animals}
10561 can be set to be the @sc{cdr} of the original list with the
10562 following expression:
10565 (setq animals (cdr animals))
10569 If you have evaluated the previous expressions and then evaluate this
10570 expression, you will see @code{(giraffe lion tiger)} appear in the echo
10571 area. If you evaluate the expression again, @code{(lion tiger)} will
10572 appear in the echo area. If you evaluate it again and yet again,
10573 @code{(tiger)} appears and then the empty list, shown by @code{nil}.
10575 A template for a @code{while} loop that uses the @code{cdr} function
10576 repeatedly to cause the true-or-false-test eventually to test false
10581 (while @var{test-whether-list-is-empty}
10583 @var{set-list-to-cdr-of-list})
10587 This test and use of @code{cdr} can be put together in a function that
10588 goes through a list and prints each element of the list on a line of its
10591 @node print-elements-of-list, Incrementing Loop, Loop Example, while
10592 @subsection An Example: @code{print-elements-of-list}
10593 @findex print-elements-of-list
10595 The @code{print-elements-of-list} function illustrates a @code{while}
10598 @cindex @file{*scratch*} buffer
10599 The function requires several lines for its output. If you are
10600 reading this in a recent instance of GNU Emacs,
10601 @c GNU Emacs 21, GNU Emacs 22, or a later version,
10602 you can evaluate the following expression inside of Info, as usual.
10604 If you are using an earlier version of Emacs, you need to copy the
10605 necessary expressions to your @file{*scratch*} buffer and evaluate
10606 them there. This is because the echo area had only one line in the
10609 You can copy the expressions by marking the beginning of the region
10610 with @kbd{C-@key{SPC}} (@code{set-mark-command}), moving the cursor to
10611 the end of the region and then copying the region using @kbd{M-w}
10612 (@code{kill-ring-save}, which calls @code{copy-region-as-kill} and
10613 then provides visual feedback). In the @file{*scratch*}
10614 buffer, you can yank the expressions back by typing @kbd{C-y}
10617 After you have copied the expressions to the @file{*scratch*} buffer,
10618 evaluate each expression in turn. Be sure to evaluate the last
10619 expression, @code{(print-elements-of-list animals)}, by typing
10620 @kbd{C-u C-x C-e}, that is, by giving an argument to
10621 @code{eval-last-sexp}. This will cause the result of the evaluation
10622 to be printed in the @file{*scratch*} buffer instead of being printed
10623 in the echo area. (Otherwise you will see something like this in your
10624 echo area: @code{^Jgazelle^J^Jgiraffe^J^Jlion^J^Jtiger^Jnil}, in which
10625 each @samp{^J} stands for a `newline'.)
10628 In a recent instance of GNU Emacs, you can evaluate these expressions
10629 directly in the Info buffer, and the echo area will grow to show the
10634 (setq animals '(gazelle giraffe lion tiger))
10636 (defun print-elements-of-list (list)
10637 "Print each element of LIST on a line of its own."
10640 (setq list (cdr list))))
10642 (print-elements-of-list animals)
10648 When you evaluate the three expressions in sequence, you will see
10664 Each element of the list is printed on a line of its own (that is what
10665 the function @code{print} does) and then the value returned by the
10666 function is printed. Since the last expression in the function is the
10667 @code{while} loop, and since @code{while} loops always return
10668 @code{nil}, a @code{nil} is printed after the last element of the list.
10670 @node Incrementing Loop, Incrementing Loop Details, print-elements-of-list, while
10671 @comment node-name, next, previous, up
10672 @subsection A Loop with an Incrementing Counter
10674 A loop is not useful unless it stops when it ought. Besides
10675 controlling a loop with a list, a common way of stopping a loop is to
10676 write the first argument as a test that returns false when the correct
10677 number of repetitions are complete. This means that the loop must
10678 have a counter---an expression that counts how many times the loop
10681 @node Incrementing Loop Details, Decrementing Loop, Incrementing Loop, while
10683 @unnumberedsubsec Details of an Incrementing Loop
10686 The test for a loop with an incrementing counter can be an expression
10687 such as @code{(< count desired-number)} which returns @code{t} for
10688 true if the value of @code{count} is less than the
10689 @code{desired-number} of repetitions and @code{nil} for false if the
10690 value of @code{count} is equal to or is greater than the
10691 @code{desired-number}. The expression that increments the count can
10692 be a simple @code{setq} such as @code{(setq count (1+ count))}, where
10693 @code{1+} is a built-in function in Emacs Lisp that adds 1 to its
10694 argument. (The expression @w{@code{(1+ count)}} has the same result
10695 as @w{@code{(+ count 1)}}, but is easier for a human to read.)
10698 The template for a @code{while} loop controlled by an incrementing
10699 counter looks like this:
10703 @var{set-count-to-initial-value}
10704 (while (< count desired-number) ; @r{true-or-false-test}
10706 (setq count (1+ count))) ; @r{incrementer}
10711 Note that you need to set the initial value of @code{count}; usually it
10715 * Incrementing Example:: Counting pebbles in a triangle.
10716 * Inc Example parts:: The parts of the function definition.
10717 * Inc Example altogether:: Putting the function definition together.
10720 @node Incrementing Example, Inc Example parts, Incrementing Loop Details, Incrementing Loop Details
10721 @unnumberedsubsubsec Example with incrementing counter
10723 Suppose you are playing on the beach and decide to make a triangle of
10724 pebbles, putting one pebble in the first row, two in the second row,
10725 three in the third row and so on, like this:
10743 @bullet{} @bullet{}
10744 @bullet{} @bullet{} @bullet{}
10745 @bullet{} @bullet{} @bullet{} @bullet{}
10752 (About 2500 years ago, Pythagoras and others developed the beginnings of
10753 number theory by considering questions such as this.)
10755 Suppose you want to know how many pebbles you will need to make a
10756 triangle with 7 rows?
10758 Clearly, what you need to do is add up the numbers from 1 to 7. There
10759 are two ways to do this; start with the smallest number, one, and add up
10760 the list in sequence, 1, 2, 3, 4 and so on; or start with the largest
10761 number and add the list going down: 7, 6, 5, 4 and so on. Because both
10762 mechanisms illustrate common ways of writing @code{while} loops, we will
10763 create two examples, one counting up and the other counting down. In
10764 this first example, we will start with 1 and add 2, 3, 4 and so on.
10766 If you are just adding up a short list of numbers, the easiest way to do
10767 it is to add up all the numbers at once. However, if you do not know
10768 ahead of time how many numbers your list will have, or if you want to be
10769 prepared for a very long list, then you need to design your addition so
10770 that what you do is repeat a simple process many times instead of doing
10771 a more complex process once.
10773 For example, instead of adding up all the pebbles all at once, what you
10774 can do is add the number of pebbles in the first row, 1, to the number
10775 in the second row, 2, and then add the total of those two rows to the
10776 third row, 3. Then you can add the number in the fourth row, 4, to the
10777 total of the first three rows; and so on.
10779 The critical characteristic of the process is that each repetitive
10780 action is simple. In this case, at each step we add only two numbers,
10781 the number of pebbles in the row and the total already found. This
10782 process of adding two numbers is repeated again and again until the last
10783 row has been added to the total of all the preceding rows. In a more
10784 complex loop the repetitive action might not be so simple, but it will
10785 be simpler than doing everything all at once.
10787 @node Inc Example parts, Inc Example altogether, Incrementing Example, Incrementing Loop Details
10788 @unnumberedsubsubsec The parts of the function definition
10790 The preceding analysis gives us the bones of our function definition:
10791 first, we will need a variable that we can call @code{total} that will
10792 be the total number of pebbles. This will be the value returned by
10795 Second, we know that the function will require an argument: this
10796 argument will be the total number of rows in the triangle. It can be
10797 called @code{number-of-rows}.
10799 Finally, we need a variable to use as a counter. We could call this
10800 variable @code{counter}, but a better name is @code{row-number}. That
10801 is because what the counter does in this function is count rows, and a
10802 program should be written to be as understandable as possible.
10804 When the Lisp interpreter first starts evaluating the expressions in the
10805 function, the value of @code{total} should be set to zero, since we have
10806 not added anything to it. Then the function should add the number of
10807 pebbles in the first row to the total, and then add the number of
10808 pebbles in the second to the total, and then add the number of
10809 pebbles in the third row to the total, and so on, until there are no
10810 more rows left to add.
10812 Both @code{total} and @code{row-number} are used only inside the
10813 function, so they can be declared as local variables with @code{let}
10814 and given initial values. Clearly, the initial value for @code{total}
10815 should be 0. The initial value of @code{row-number} should be 1,
10816 since we start with the first row. This means that the @code{let}
10817 statement will look like this:
10827 After the internal variables are declared and bound to their initial
10828 values, we can begin the @code{while} loop. The expression that serves
10829 as the test should return a value of @code{t} for true so long as the
10830 @code{row-number} is less than or equal to the @code{number-of-rows}.
10831 (If the expression tests true only so long as the row number is less
10832 than the number of rows in the triangle, the last row will never be
10833 added to the total; hence the row number has to be either less than or
10834 equal to the number of rows.)
10837 @findex <= @r{(less than or equal)}
10838 Lisp provides the @code{<=} function that returns true if the value of
10839 its first argument is less than or equal to the value of its second
10840 argument and false otherwise. So the expression that the @code{while}
10841 will evaluate as its test should look like this:
10844 (<= row-number number-of-rows)
10847 The total number of pebbles can be found by repeatedly adding the number
10848 of pebbles in a row to the total already found. Since the number of
10849 pebbles in the row is equal to the row number, the total can be found by
10850 adding the row number to the total. (Clearly, in a more complex
10851 situation, the number of pebbles in the row might be related to the row
10852 number in a more complicated way; if this were the case, the row number
10853 would be replaced by the appropriate expression.)
10856 (setq total (+ total row-number))
10860 What this does is set the new value of @code{total} to be equal to the
10861 sum of adding the number of pebbles in the row to the previous total.
10863 After setting the value of @code{total}, the conditions need to be
10864 established for the next repetition of the loop, if there is one. This
10865 is done by incrementing the value of the @code{row-number} variable,
10866 which serves as a counter. After the @code{row-number} variable has
10867 been incremented, the true-or-false-test at the beginning of the
10868 @code{while} loop tests whether its value is still less than or equal to
10869 the value of the @code{number-of-rows} and if it is, adds the new value
10870 of the @code{row-number} variable to the @code{total} of the previous
10871 repetition of the loop.
10874 The built-in Emacs Lisp function @code{1+} adds 1 to a number, so the
10875 @code{row-number} variable can be incremented with this expression:
10878 (setq row-number (1+ row-number))
10881 @node Inc Example altogether, , Inc Example parts, Incrementing Loop Details
10882 @unnumberedsubsubsec Putting the function definition together
10884 We have created the parts for the function definition; now we need to
10888 First, the contents of the @code{while} expression:
10892 (while (<= row-number number-of-rows) ; @r{true-or-false-test}
10893 (setq total (+ total row-number))
10894 (setq row-number (1+ row-number))) ; @r{incrementer}
10898 Along with the @code{let} expression varlist, this very nearly
10899 completes the body of the function definition. However, it requires
10900 one final element, the need for which is somewhat subtle.
10902 The final touch is to place the variable @code{total} on a line by
10903 itself after the @code{while} expression. Otherwise, the value returned
10904 by the whole function is the value of the last expression that is
10905 evaluated in the body of the @code{let}, and this is the value
10906 returned by the @code{while}, which is always @code{nil}.
10908 This may not be evident at first sight. It almost looks as if the
10909 incrementing expression is the last expression of the whole function.
10910 But that expression is part of the body of the @code{while}; it is the
10911 last element of the list that starts with the symbol @code{while}.
10912 Moreover, the whole of the @code{while} loop is a list within the body
10916 In outline, the function will look like this:
10920 (defun @var{name-of-function} (@var{argument-list})
10921 "@var{documentation}@dots{}"
10922 (let (@var{varlist})
10923 (while (@var{true-or-false-test})
10924 @var{body-of-while}@dots{} )
10925 @dots{} )) ; @r{Need final expression here.}
10929 The result of evaluating the @code{let} is what is going to be returned
10930 by the @code{defun} since the @code{let} is not embedded within any
10931 containing list, except for the @code{defun} as a whole. However, if
10932 the @code{while} is the last element of the @code{let} expression, the
10933 function will always return @code{nil}. This is not what we want!
10934 Instead, what we want is the value of the variable @code{total}. This
10935 is returned by simply placing the symbol as the last element of the list
10936 starting with @code{let}. It gets evaluated after the preceding
10937 elements of the list are evaluated, which means it gets evaluated after
10938 it has been assigned the correct value for the total.
10940 It may be easier to see this by printing the list starting with
10941 @code{let} all on one line. This format makes it evident that the
10942 @var{varlist} and @code{while} expressions are the second and third
10943 elements of the list starting with @code{let}, and the @code{total} is
10948 (let (@var{varlist}) (while (@var{true-or-false-test}) @var{body-of-while}@dots{} ) total)
10953 Putting everything together, the @code{triangle} function definition
10958 (defun triangle (number-of-rows) ; @r{Version with}
10959 ; @r{ incrementing counter.}
10960 "Add up the number of pebbles in a triangle.
10961 The first row has one pebble, the second row two pebbles,
10962 the third row three pebbles, and so on.
10963 The argument is NUMBER-OF-ROWS."
10968 (while (<= row-number number-of-rows)
10969 (setq total (+ total row-number))
10970 (setq row-number (1+ row-number)))
10976 After you have installed @code{triangle} by evaluating the function, you
10977 can try it out. Here are two examples:
10988 The sum of the first four numbers is 10 and the sum of the first seven
10991 @node Decrementing Loop, , Incrementing Loop Details, while
10992 @comment node-name, next, previous, up
10993 @subsection Loop with a Decrementing Counter
10995 Another common way to write a @code{while} loop is to write the test
10996 so that it determines whether a counter is greater than zero. So long
10997 as the counter is greater than zero, the loop is repeated. But when
10998 the counter is equal to or less than zero, the loop is stopped. For
10999 this to work, the counter has to start out greater than zero and then
11000 be made smaller and smaller by a form that is evaluated
11003 The test will be an expression such as @code{(> counter 0)} which
11004 returns @code{t} for true if the value of @code{counter} is greater
11005 than zero, and @code{nil} for false if the value of @code{counter} is
11006 equal to or less than zero. The expression that makes the number
11007 smaller and smaller can be a simple @code{setq} such as @code{(setq
11008 counter (1- counter))}, where @code{1-} is a built-in function in
11009 Emacs Lisp that subtracts 1 from its argument.
11012 The template for a decrementing @code{while} loop looks like this:
11016 (while (> counter 0) ; @r{true-or-false-test}
11018 (setq counter (1- counter))) ; @r{decrementer}
11023 * Decrementing Example:: More pebbles on the beach.
11024 * Dec Example parts:: The parts of the function definition.
11025 * Dec Example altogether:: Putting the function definition together.
11028 @node Decrementing Example, Dec Example parts, Decrementing Loop, Decrementing Loop
11029 @unnumberedsubsubsec Example with decrementing counter
11031 To illustrate a loop with a decrementing counter, we will rewrite the
11032 @code{triangle} function so the counter decreases to zero.
11034 This is the reverse of the earlier version of the function. In this
11035 case, to find out how many pebbles are needed to make a triangle with
11036 3 rows, add the number of pebbles in the third row, 3, to the number
11037 in the preceding row, 2, and then add the total of those two rows to
11038 the row that precedes them, which is 1.
11040 Likewise, to find the number of pebbles in a triangle with 7 rows, add
11041 the number of pebbles in the seventh row, 7, to the number in the
11042 preceding row, which is 6, and then add the total of those two rows to
11043 the row that precedes them, which is 5, and so on. As in the previous
11044 example, each addition only involves adding two numbers, the total of
11045 the rows already added up and the number of pebbles in the row that is
11046 being added to the total. This process of adding two numbers is
11047 repeated again and again until there are no more pebbles to add.
11049 We know how many pebbles to start with: the number of pebbles in the
11050 last row is equal to the number of rows. If the triangle has seven
11051 rows, the number of pebbles in the last row is 7. Likewise, we know how
11052 many pebbles are in the preceding row: it is one less than the number in
11055 @node Dec Example parts, Dec Example altogether, Decrementing Example, Decrementing Loop
11056 @unnumberedsubsubsec The parts of the function definition
11058 We start with three variables: the total number of rows in the
11059 triangle; the number of pebbles in a row; and the total number of
11060 pebbles, which is what we want to calculate. These variables can be
11061 named @code{number-of-rows}, @code{number-of-pebbles-in-row}, and
11062 @code{total}, respectively.
11064 Both @code{total} and @code{number-of-pebbles-in-row} are used only
11065 inside the function and are declared with @code{let}. The initial
11066 value of @code{total} should, of course, be zero. However, the
11067 initial value of @code{number-of-pebbles-in-row} should be equal to
11068 the number of rows in the triangle, since the addition will start with
11072 This means that the beginning of the @code{let} expression will look
11078 (number-of-pebbles-in-row number-of-rows))
11083 The total number of pebbles can be found by repeatedly adding the number
11084 of pebbles in a row to the total already found, that is, by repeatedly
11085 evaluating the following expression:
11088 (setq total (+ total number-of-pebbles-in-row))
11092 After the @code{number-of-pebbles-in-row} is added to the @code{total},
11093 the @code{number-of-pebbles-in-row} should be decremented by one, since
11094 the next time the loop repeats, the preceding row will be
11095 added to the total.
11097 The number of pebbles in a preceding row is one less than the number of
11098 pebbles in a row, so the built-in Emacs Lisp function @code{1-} can be
11099 used to compute the number of pebbles in the preceding row. This can be
11100 done with the following expression:
11104 (setq number-of-pebbles-in-row
11105 (1- number-of-pebbles-in-row))
11109 Finally, we know that the @code{while} loop should stop making repeated
11110 additions when there are no pebbles in a row. So the test for
11111 the @code{while} loop is simply:
11114 (while (> number-of-pebbles-in-row 0)
11117 @node Dec Example altogether, , Dec Example parts, Decrementing Loop
11118 @unnumberedsubsubsec Putting the function definition together
11120 We can put these expressions together to create a function definition
11121 that works. However, on examination, we find that one of the local
11122 variables is unneeded!
11125 The function definition looks like this:
11129 ;;; @r{First subtractive version.}
11130 (defun triangle (number-of-rows)
11131 "Add up the number of pebbles in a triangle."
11133 (number-of-pebbles-in-row number-of-rows))
11134 (while (> number-of-pebbles-in-row 0)
11135 (setq total (+ total number-of-pebbles-in-row))
11136 (setq number-of-pebbles-in-row
11137 (1- number-of-pebbles-in-row)))
11142 As written, this function works.
11144 However, we do not need @code{number-of-pebbles-in-row}.
11146 @cindex Argument as local variable
11147 When the @code{triangle} function is evaluated, the symbol
11148 @code{number-of-rows} will be bound to a number, giving it an initial
11149 value. That number can be changed in the body of the function as if
11150 it were a local variable, without any fear that such a change will
11151 effect the value of the variable outside of the function. This is a
11152 very useful characteristic of Lisp; it means that the variable
11153 @code{number-of-rows} can be used anywhere in the function where
11154 @code{number-of-pebbles-in-row} is used.
11157 Here is a second version of the function written a bit more cleanly:
11161 (defun triangle (number) ; @r{Second version.}
11162 "Return sum of numbers 1 through NUMBER inclusive."
11164 (while (> number 0)
11165 (setq total (+ total number))
11166 (setq number (1- number)))
11171 In brief, a properly written @code{while} loop will consist of three parts:
11175 A test that will return false after the loop has repeated itself the
11176 correct number of times.
11179 An expression the evaluation of which will return the value desired
11180 after being repeatedly evaluated.
11183 An expression to change the value passed to the true-or-false-test so
11184 that the test returns false after the loop has repeated itself the right
11188 @node dolist dotimes, Recursion, while, Loops & Recursion
11189 @comment node-name, next, previous, up
11190 @section Save your time: @code{dolist} and @code{dotimes}
11192 In addition to @code{while}, both @code{dolist} and @code{dotimes}
11193 provide for looping. Sometimes these are quicker to write than the
11194 equivalent @code{while} loop. Both are Lisp macros. (@xref{Macros, ,
11195 Macros, elisp, The GNU Emacs Lisp Reference Manual}. )
11197 @code{dolist} works like a @code{while} loop that `@sc{cdr}s down a
11198 list': @code{dolist} automatically shortens the list each time it
11199 loops---takes the @sc{cdr} of the list---and binds the @sc{car} of
11200 each shorter version of the list to the first of its arguments.
11202 @code{dotimes} loops a specific number of times: you specify the number.
11209 @node dolist, dotimes, dolist dotimes, dolist dotimes
11210 @unnumberedsubsubsec The @code{dolist} Macro
11213 Suppose, for example, you want to reverse a list, so that
11214 ``first'' ``second'' ``third'' becomes ``third'' ``second'' ``first''.
11217 In practice, you would use the @code{reverse} function, like this:
11221 (setq animals '(gazelle giraffe lion tiger))
11229 Here is how you could reverse the list using a @code{while} loop:
11233 (setq animals '(gazelle giraffe lion tiger))
11235 (defun reverse-list-with-while (list)
11236 "Using while, reverse the order of LIST."
11237 (let (value) ; make sure list starts empty
11239 (setq value (cons (car list) value))
11240 (setq list (cdr list)))
11243 (reverse-list-with-while animals)
11249 And here is how you could use the @code{dolist} macro:
11253 (setq animals '(gazelle giraffe lion tiger))
11255 (defun reverse-list-with-dolist (list)
11256 "Using dolist, reverse the order of LIST."
11257 (let (value) ; make sure list starts empty
11258 (dolist (element list value)
11259 (setq value (cons element value)))))
11261 (reverse-list-with-dolist animals)
11267 In Info, you can place your cursor after the closing parenthesis of
11268 each expression and type @kbd{C-x C-e}; in each case, you should see
11271 (tiger lion giraffe gazelle)
11277 For this example, the existing @code{reverse} function is obviously best.
11278 The @code{while} loop is just like our first example (@pxref{Loop
11279 Example, , A @code{while} Loop and a List}). The @code{while} first
11280 checks whether the list has elements; if so, it constructs a new list
11281 by adding the first element of the list to the existing list (which in
11282 the first iteration of the loop is @code{nil}). Since the second
11283 element is prepended in front of the first element, and the third
11284 element is prepended in front of the second element, the list is reversed.
11286 In the expression using a @code{while} loop,
11287 the @w{@code{(setq list (cdr list))}}
11288 expression shortens the list, so the @code{while} loop eventually
11289 stops. In addition, it provides the @code{cons} expression with a new
11290 first element by creating a new and shorter list at each repetition of
11293 The @code{dolist} expression does very much the same as the
11294 @code{while} expression, except that the @code{dolist} macro does some
11295 of the work you have to do when writing a @code{while} expression.
11297 Like a @code{while} loop, a @code{dolist} loops. What is different is
11298 that it automatically shortens the list each time it loops --- it
11299 `@sc{cdr}s down the list' on its own --- and it automatically binds
11300 the @sc{car} of each shorter version of the list to the first of its
11303 In the example, the @sc{car} of each shorter version of the list is
11304 referred to using the symbol @samp{element}, the list itself is called
11305 @samp{list}, and the value returned is called @samp{value}. The
11306 remainder of the @code{dolist} expression is the body.
11308 The @code{dolist} expression binds the @sc{car} of each shorter
11309 version of the list to @code{element} and then evaluates the body of
11310 the expression; and repeats the loop. The result is returned in
11313 @node dotimes, , dolist, dolist dotimes
11314 @unnumberedsubsubsec The @code{dotimes} Macro
11317 The @code{dotimes} macro is similar to @code{dolist}, except that it
11318 loops a specific number of times.
11320 The first argument to @code{dotimes} is assigned the numbers 0, 1, 2
11321 and so forth each time around the loop, and the value of the third
11322 argument is returned. You need to provide the value of the second
11323 argument, which is how many times the macro loops.
11326 For example, the following binds the numbers from 0 up to, but not
11327 including, the number 3 to the first argument, @var{number}, and then
11328 constructs a list of the three numbers. (The first number is 0, the
11329 second number is 1, and the third number is 2; this makes a total of
11330 three numbers in all, starting with zero as the first number.)
11334 (let (value) ; otherwise a value is a void variable
11335 (dotimes (number 3 value)
11336 (setq value (cons number value))))
11343 @code{dotimes} returns @code{value}, so the way to use
11344 @code{dotimes} is to operate on some expression @var{number} number of
11345 times and then return the result, either as a list or an atom.
11348 Here is an example of a @code{defun} that uses @code{dotimes} to add
11349 up the number of pebbles in a triangle.
11353 (defun triangle-using-dotimes (number-of-rows)
11354 "Using dotimes, add up the number of pebbles in a triangle."
11355 (let ((total 0)) ; otherwise a total is a void variable
11356 (dotimes (number number-of-rows total)
11357 (setq total (+ total (1+ number))))))
11359 (triangle-using-dotimes 4)
11363 @node Recursion, Looping exercise, dolist dotimes, Loops & Recursion
11364 @comment node-name, next, previous, up
11368 A recursive function contains code that tells the Lisp interpreter to
11369 call a program that runs exactly like itself, but with slightly
11370 different arguments. The code runs exactly the same because it has
11371 the same name. However, even though the program has the same name, it
11372 is not the same entity. It is different. In the jargon, it is a
11373 different `instance'.
11375 Eventually, if the program is written correctly, the `slightly
11376 different arguments' will become sufficiently different from the first
11377 arguments that the final instance will stop.
11380 * Building Robots:: Same model, different serial number ...
11381 * Recursive Definition Parts:: Walk until you stop ...
11382 * Recursion with list:: Using a list as the test whether to recurse.
11383 * Recursive triangle function::
11384 * Recursion with cond::
11385 * Recursive Patterns:: Often used templates.
11386 * No Deferment:: Don't store up work ...
11387 * No deferment solution::
11390 @node Building Robots, Recursive Definition Parts, Recursion, Recursion
11391 @comment node-name, next, previous, up
11392 @subsection Building Robots: Extending the Metaphor
11393 @cindex Building robots
11394 @cindex Robots, building
11396 It is sometimes helpful to think of a running program as a robot that
11397 does a job. In doing its job, a recursive function calls on a second
11398 robot to help it. The second robot is identical to the first in every
11399 way, except that the second robot helps the first and has been
11400 passed different arguments than the first.
11402 In a recursive function, the second robot may call a third; and the
11403 third may call a fourth, and so on. Each of these is a different
11404 entity; but all are clones.
11406 Since each robot has slightly different instructions---the arguments
11407 will differ from one robot to the next---the last robot should know
11410 Let's expand on the metaphor in which a computer program is a robot.
11412 A function definition provides the blueprints for a robot. When you
11413 install a function definition, that is, when you evaluate a
11414 @code{defun} special form, you install the necessary equipment to
11415 build robots. It is as if you were in a factory, setting up an
11416 assembly line. Robots with the same name are built according to the
11417 same blueprints. So they have, as it were, the same `model number',
11418 but a different `serial number'.
11420 We often say that a recursive function `calls itself'. What we mean
11421 is that the instructions in a recursive function cause the Lisp
11422 interpreter to run a different function that has the same name and
11423 does the same job as the first, but with different arguments.
11425 It is important that the arguments differ from one instance to the
11426 next; otherwise, the process will never stop.
11428 @node Recursive Definition Parts, Recursion with list, Building Robots, Recursion
11429 @comment node-name, next, previous, up
11430 @subsection The Parts of a Recursive Definition
11431 @cindex Parts of a Recursive Definition
11432 @cindex Recursive Definition Parts
11434 A recursive function typically contains a conditional expression which
11439 A true-or-false-test that determines whether the function is called
11440 again, here called the @dfn{do-again-test}.
11443 The name of the function. When this name is called, a new instance of
11444 the function---a new robot, as it were---is created and told what to do.
11447 An expression that returns a different value each time the function is
11448 called, here called the @dfn{next-step-expression}. Consequently, the
11449 argument (or arguments) passed to the new instance of the function
11450 will be different from that passed to the previous instance. This
11451 causes the conditional expression, the @dfn{do-again-test}, to test
11452 false after the correct number of repetitions.
11455 Recursive functions can be much simpler than any other kind of
11456 function. Indeed, when people first start to use them, they often look
11457 so mysteriously simple as to be incomprehensible. Like riding a
11458 bicycle, reading a recursive function definition takes a certain knack
11459 which is hard at first but then seems simple.
11462 There are several different common recursive patterns. A very simple
11463 pattern looks like this:
11467 (defun @var{name-of-recursive-function} (@var{argument-list})
11468 "@var{documentation}@dots{}"
11469 (if @var{do-again-test}
11471 (@var{name-of-recursive-function}
11472 @var{next-step-expression})))
11476 Each time a recursive function is evaluated, a new instance of it is
11477 created and told what to do. The arguments tell the instance what to do.
11479 An argument is bound to the value of the next-step-expression. Each
11480 instance runs with a different value of the next-step-expression.
11482 The value in the next-step-expression is used in the do-again-test.
11484 The value returned by the next-step-expression is passed to the new
11485 instance of the function, which evaluates it (or some
11486 transmogrification of it) to determine whether to continue or stop.
11487 The next-step-expression is designed so that the do-again-test returns
11488 false when the function should no longer be repeated.
11490 The do-again-test is sometimes called the @dfn{stop condition},
11491 since it stops the repetitions when it tests false.
11493 @node Recursion with list, Recursive triangle function, Recursive Definition Parts, Recursion
11494 @comment node-name, next, previous, up
11495 @subsection Recursion with a List
11497 The example of a @code{while} loop that printed the elements of a list
11498 of numbers can be written recursively. Here is the code, including
11499 an expression to set the value of the variable @code{animals} to a list.
11501 If you are using GNU Emacs 20 or before, this example must be copied
11502 to the @file{*scratch*} buffer and each expression must be evaluated
11503 there. Use @kbd{C-u C-x C-e} to evaluate the
11504 @code{(print-elements-recursively animals)} expression so that the
11505 results are printed in the buffer; otherwise the Lisp interpreter will
11506 try to squeeze the results into the one line of the echo area.
11508 Also, place your cursor immediately after the last closing parenthesis
11509 of the @code{print-elements-recursively} function, before the comment.
11510 Otherwise, the Lisp interpreter will try to evaluate the comment.
11512 If you are using a more recent version of Emacs, you can evaluate this
11513 expression directly in Info.
11515 @findex print-elements-recursively
11518 (setq animals '(gazelle giraffe lion tiger))
11520 (defun print-elements-recursively (list)
11521 "Print each element of LIST on a line of its own.
11523 (when list ; @r{do-again-test}
11524 (print (car list)) ; @r{body}
11525 (print-elements-recursively ; @r{recursive call}
11526 (cdr list)))) ; @r{next-step-expression}
11528 (print-elements-recursively animals)
11532 The @code{print-elements-recursively} function first tests whether
11533 there is any content in the list; if there is, the function prints the
11534 first element of the list, the @sc{car} of the list. Then the
11535 function `invokes itself', but gives itself as its argument, not the
11536 whole list, but the second and subsequent elements of the list, the
11537 @sc{cdr} of the list.
11539 Put another way, if the list is not empty, the function invokes
11540 another instance of code that is similar to the initial code, but is a
11541 different thread of execution, with different arguments than the first
11544 Put in yet another way, if the list is not empty, the first robot
11545 assembles a second robot and tells it what to do; the second robot is
11546 a different individual from the first, but is the same model.
11548 When the second evaluation occurs, the @code{when} expression is
11549 evaluated and if true, prints the first element of the list it
11550 receives as its argument (which is the second element of the original
11551 list). Then the function `calls itself' with the @sc{cdr} of the list
11552 it is invoked with, which (the second time around) is the @sc{cdr} of
11553 the @sc{cdr} of the original list.
11555 Note that although we say that the function `calls itself', what we
11556 mean is that the Lisp interpreter assembles and instructs a new
11557 instance of the program. The new instance is a clone of the first,
11558 but is a separate individual.
11560 Each time the function `invokes itself', it invokes itself on a
11561 shorter version of the original list. It creates a new instance that
11562 works on a shorter list.
11564 Eventually, the function invokes itself on an empty list. It creates
11565 a new instance whose argument is @code{nil}. The conditional expression
11566 tests the value of @code{list}. Since the value of @code{list} is
11567 @code{nil}, the @code{when} expression tests false so the then-part is
11568 not evaluated. The function as a whole then returns @code{nil}.
11571 When you evaluate the expression @code{(print-elements-recursively
11572 animals)} in the @file{*scratch*} buffer, you see this result:
11588 @node Recursive triangle function, Recursion with cond, Recursion with list, Recursion
11589 @comment node-name, next, previous, up
11590 @subsection Recursion in Place of a Counter
11591 @findex triangle-recursively
11594 The @code{triangle} function described in a previous section can also
11595 be written recursively. It looks like this:
11599 (defun triangle-recursively (number)
11600 "Return the sum of the numbers 1 through NUMBER inclusive.
11602 (if (= number 1) ; @r{do-again-test}
11604 (+ number ; @r{else-part}
11605 (triangle-recursively ; @r{recursive call}
11606 (1- number))))) ; @r{next-step-expression}
11608 (triangle-recursively 7)
11613 You can install this function by evaluating it and then try it by
11614 evaluating @code{(triangle-recursively 7)}. (Remember to put your
11615 cursor immediately after the last parenthesis of the function
11616 definition, before the comment.) The function evaluates to 28.
11618 To understand how this function works, let's consider what happens in the
11619 various cases when the function is passed 1, 2, 3, or 4 as the value of
11623 * Recursive Example arg of 1 or 2::
11624 * Recursive Example arg of 3 or 4::
11627 @node Recursive Example arg of 1 or 2, Recursive Example arg of 3 or 4, Recursive triangle function, Recursive triangle function
11629 @unnumberedsubsubsec An argument of 1 or 2
11632 First, what happens if the value of the argument is 1?
11634 The function has an @code{if} expression after the documentation
11635 string. It tests whether the value of @code{number} is equal to 1; if
11636 so, Emacs evaluates the then-part of the @code{if} expression, which
11637 returns the number 1 as the value of the function. (A triangle with
11638 one row has one pebble in it.)
11640 Suppose, however, that the value of the argument is 2. In this case,
11641 Emacs evaluates the else-part of the @code{if} expression.
11644 The else-part consists of an addition, the recursive call to
11645 @code{triangle-recursively} and a decrementing action; and it looks like
11649 (+ number (triangle-recursively (1- number)))
11652 When Emacs evaluates this expression, the innermost expression is
11653 evaluated first; then the other parts in sequence. Here are the steps
11657 @item Step 1 @w{ } Evaluate the innermost expression.
11659 The innermost expression is @code{(1- number)} so Emacs decrements the
11660 value of @code{number} from 2 to 1.
11662 @item Step 2 @w{ } Evaluate the @code{triangle-recursively} function.
11664 The Lisp interpreter creates an individual instance of
11665 @code{triangle-recursively}. It does not matter that this function is
11666 contained within itself. Emacs passes the result Step 1 as the
11667 argument used by this instance of the @code{triangle-recursively}
11670 In this case, Emacs evaluates @code{triangle-recursively} with an
11671 argument of 1. This means that this evaluation of
11672 @code{triangle-recursively} returns 1.
11674 @item Step 3 @w{ } Evaluate the value of @code{number}.
11676 The variable @code{number} is the second element of the list that
11677 starts with @code{+}; its value is 2.
11679 @item Step 4 @w{ } Evaluate the @code{+} expression.
11681 The @code{+} expression receives two arguments, the first
11682 from the evaluation of @code{number} (Step 3) and the second from the
11683 evaluation of @code{triangle-recursively} (Step 2).
11685 The result of the addition is the sum of 2 plus 1, and the number 3 is
11686 returned, which is correct. A triangle with two rows has three
11690 @node Recursive Example arg of 3 or 4, , Recursive Example arg of 1 or 2, Recursive triangle function
11691 @unnumberedsubsubsec An argument of 3 or 4
11693 Suppose that @code{triangle-recursively} is called with an argument of
11697 @item Step 1 @w{ } Evaluate the do-again-test.
11699 The @code{if} expression is evaluated first. This is the do-again
11700 test and returns false, so the else-part of the @code{if} expression
11701 is evaluated. (Note that in this example, the do-again-test causes
11702 the function to call itself when it tests false, not when it tests
11705 @item Step 2 @w{ } Evaluate the innermost expression of the else-part.
11707 The innermost expression of the else-part is evaluated, which decrements
11708 3 to 2. This is the next-step-expression.
11710 @item Step 3 @w{ } Evaluate the @code{triangle-recursively} function.
11712 The number 2 is passed to the @code{triangle-recursively} function.
11714 We already know what happens when Emacs evaluates @code{triangle-recursively} with
11715 an argument of 2. After going through the sequence of actions described
11716 earlier, it returns a value of 3. So that is what will happen here.
11718 @item Step 4 @w{ } Evaluate the addition.
11720 3 will be passed as an argument to the addition and will be added to the
11721 number with which the function was called, which is 3.
11725 The value returned by the function as a whole will be 6.
11727 Now that we know what will happen when @code{triangle-recursively} is
11728 called with an argument of 3, it is evident what will happen if it is
11729 called with an argument of 4:
11733 In the recursive call, the evaluation of
11736 (triangle-recursively (1- 4))
11741 will return the value of evaluating
11744 (triangle-recursively 3)
11748 which is 6 and this value will be added to 4 by the addition in the
11753 The value returned by the function as a whole will be 10.
11755 Each time @code{triangle-recursively} is evaluated, it evaluates a
11756 version of itself---a different instance of itself---with a smaller
11757 argument, until the argument is small enough so that it does not
11760 Note that this particular design for a recursive function
11761 requires that operations be deferred.
11763 Before @code{(triangle-recursively 7)} can calculate its answer, it
11764 must call @code{(triangle-recursively 6)}; and before
11765 @code{(triangle-recursively 6)} can calculate its answer, it must call
11766 @code{(triangle-recursively 5)}; and so on. That is to say, the
11767 calculation that @code{(triangle-recursively 7)} makes must be
11768 deferred until @code{(triangle-recursively 6)} makes its calculation;
11769 and @code{(triangle-recursively 6)} must defer until
11770 @code{(triangle-recursively 5)} completes; and so on.
11772 If each of these instances of @code{triangle-recursively} are thought
11773 of as different robots, the first robot must wait for the second to
11774 complete its job, which must wait until the third completes, and so
11777 There is a way around this kind of waiting, which we will discuss in
11778 @ref{No Deferment, , Recursion without Deferments}.
11780 @node Recursion with cond, Recursive Patterns, Recursive triangle function, Recursion
11781 @comment node-name, next, previous, up
11782 @subsection Recursion Example Using @code{cond}
11785 The version of @code{triangle-recursively} described earlier is written
11786 with the @code{if} special form. It can also be written using another
11787 special form called @code{cond}. The name of the special form
11788 @code{cond} is an abbreviation of the word @samp{conditional}.
11790 Although the @code{cond} special form is not used as often in the
11791 Emacs Lisp sources as @code{if}, it is used often enough to justify
11795 The template for a @code{cond} expression looks like this:
11805 where the @var{body} is a series of lists.
11808 Written out more fully, the template looks like this:
11813 (@var{first-true-or-false-test} @var{first-consequent})
11814 (@var{second-true-or-false-test} @var{second-consequent})
11815 (@var{third-true-or-false-test} @var{third-consequent})
11820 When the Lisp interpreter evaluates the @code{cond} expression, it
11821 evaluates the first element (the @sc{car} or true-or-false-test) of
11822 the first expression in a series of expressions within the body of the
11825 If the true-or-false-test returns @code{nil} the rest of that
11826 expression, the consequent, is skipped and the true-or-false-test of the
11827 next expression is evaluated. When an expression is found whose
11828 true-or-false-test returns a value that is not @code{nil}, the
11829 consequent of that expression is evaluated. The consequent can be one
11830 or more expressions. If the consequent consists of more than one
11831 expression, the expressions are evaluated in sequence and the value of
11832 the last one is returned. If the expression does not have a consequent,
11833 the value of the true-or-false-test is returned.
11835 If none of the true-or-false-tests test true, the @code{cond} expression
11836 returns @code{nil}.
11839 Written using @code{cond}, the @code{triangle} function looks like this:
11843 (defun triangle-using-cond (number)
11844 (cond ((<= number 0) 0)
11847 (+ number (triangle-using-cond (1- number))))))
11852 In this example, the @code{cond} returns 0 if the number is less than or
11853 equal to 0, it returns 1 if the number is 1 and it evaluates @code{(+
11854 number (triangle-using-cond (1- number)))} if the number is greater than
11857 @node Recursive Patterns, No Deferment, Recursion with cond, Recursion
11858 @comment node-name, next, previous, up
11859 @subsection Recursive Patterns
11860 @cindex Recursive Patterns
11862 Here are three common recursive patterns. Each involves a list.
11863 Recursion does not need to involve lists, but Lisp is designed for lists
11864 and this provides a sense of its primal capabilities.
11872 @node Every, Accumulate, Recursive Patterns, Recursive Patterns
11873 @comment node-name, next, previous, up
11874 @unnumberedsubsubsec Recursive Pattern: @emph{every}
11875 @cindex Every, type of recursive pattern
11876 @cindex Recursive pattern: every
11878 In the @code{every} recursive pattern, an action is performed on every
11882 The basic pattern is:
11886 If a list be empty, return @code{nil}.
11888 Else, act on the beginning of the list (the @sc{car} of the list)
11891 through a recursive call by the function on the rest (the
11892 @sc{cdr}) of the list,
11894 and, optionally, combine the acted-on element, using @code{cons},
11895 with the results of acting on the rest.
11904 (defun square-each (numbers-list)
11905 "Square each of a NUMBERS LIST, recursively."
11906 (if (not numbers-list) ; do-again-test
11909 (* (car numbers-list) (car numbers-list))
11910 (square-each (cdr numbers-list))))) ; next-step-expression
11914 (square-each '(1 2 3))
11921 If @code{numbers-list} is empty, do nothing. But if it has content,
11922 construct a list combining the square of the first number in the list
11923 with the result of the recursive call.
11925 (The example follows the pattern exactly: @code{nil} is returned if
11926 the numbers' list is empty. In practice, you would write the
11927 conditional so it carries out the action when the numbers' list is not
11930 The @code{print-elements-recursively} function (@pxref{Recursion with
11931 list, , Recursion with a List}) is another example of an @code{every}
11932 pattern, except in this case, rather than bring the results together
11933 using @code{cons}, we print each element of output.
11936 The @code{print-elements-recursively} function looks like this:
11940 (setq animals '(gazelle giraffe lion tiger))
11944 (defun print-elements-recursively (list)
11945 "Print each element of LIST on a line of its own.
11947 (when list ; @r{do-again-test}
11948 (print (car list)) ; @r{body}
11949 (print-elements-recursively ; @r{recursive call}
11950 (cdr list)))) ; @r{next-step-expression}
11952 (print-elements-recursively animals)
11957 The pattern for @code{print-elements-recursively} is:
11961 When the list is empty, do nothing.
11963 But when the list has at least one element,
11966 act on the beginning of the list (the @sc{car} of the list),
11968 and make a recursive call on the rest (the @sc{cdr}) of the list.
11972 @node Accumulate, Keep, Every, Recursive Patterns
11973 @comment node-name, next, previous, up
11974 @unnumberedsubsubsec Recursive Pattern: @emph{accumulate}
11975 @cindex Accumulate, type of recursive pattern
11976 @cindex Recursive pattern: accumulate
11978 Another recursive pattern is called the @code{accumulate} pattern. In
11979 the @code{accumulate} recursive pattern, an action is performed on
11980 every element of a list and the result of that action is accumulated
11981 with the results of performing the action on the other elements.
11983 This is very like the `every' pattern using @code{cons}, except that
11984 @code{cons} is not used, but some other combiner.
11991 If a list be empty, return zero or some other constant.
11993 Else, act on the beginning of the list (the @sc{car} of the list),
11996 and combine that acted-on element, using @code{+} or
11997 some other combining function, with
11999 a recursive call by the function on the rest (the @sc{cdr}) of the list.
12004 Here is an example:
12008 (defun add-elements (numbers-list)
12009 "Add the elements of NUMBERS-LIST together."
12010 (if (not numbers-list)
12012 (+ (car numbers-list) (add-elements (cdr numbers-list)))))
12016 (add-elements '(1 2 3 4))
12021 @xref{Files List, , Making a List of Files}, for an example of the
12022 accumulate pattern.
12024 @node Keep, , Accumulate, Recursive Patterns
12025 @comment node-name, next, previous, up
12026 @unnumberedsubsubsec Recursive Pattern: @emph{keep}
12027 @cindex Keep, type of recursive pattern
12028 @cindex Recursive pattern: keep
12030 A third recursive pattern is called the @code{keep} pattern.
12031 In the @code{keep} recursive pattern, each element of a list is tested;
12032 the element is acted on and the results are kept only if the element
12035 Again, this is very like the `every' pattern, except the element is
12036 skipped unless it meets a criterion.
12039 The pattern has three parts:
12043 If a list be empty, return @code{nil}.
12045 Else, if the beginning of the list (the @sc{car} of the list) passes
12049 act on that element and combine it, using @code{cons} with
12051 a recursive call by the function on the rest (the @sc{cdr}) of the list.
12054 Otherwise, if the beginning of the list (the @sc{car} of the list) fails
12058 skip on that element,
12060 and, recursively call the function on the rest (the @sc{cdr}) of the list.
12065 Here is an example that uses @code{cond}:
12069 (defun keep-three-letter-words (word-list)
12070 "Keep three letter words in WORD-LIST."
12072 ;; First do-again-test: stop-condition
12073 ((not word-list) nil)
12075 ;; Second do-again-test: when to act
12076 ((eq 3 (length (symbol-name (car word-list))))
12077 ;; combine acted-on element with recursive call on shorter list
12078 (cons (car word-list) (keep-three-letter-words (cdr word-list))))
12080 ;; Third do-again-test: when to skip element;
12081 ;; recursively call shorter list with next-step expression
12082 (t (keep-three-letter-words (cdr word-list)))))
12086 (keep-three-letter-words '(one two three four five six))
12087 @result{} (one two six)
12091 It goes without saying that you need not use @code{nil} as the test for
12092 when to stop; and you can, of course, combine these patterns.
12094 @node No Deferment, No deferment solution, Recursive Patterns, Recursion
12095 @subsection Recursion without Deferments
12096 @cindex Deferment in recursion
12097 @cindex Recursion without Deferments
12099 Let's consider again what happens with the @code{triangle-recursively}
12100 function. We will find that the intermediate calculations are
12101 deferred until all can be done.
12104 Here is the function definition:
12108 (defun triangle-recursively (number)
12109 "Return the sum of the numbers 1 through NUMBER inclusive.
12111 (if (= number 1) ; @r{do-again-test}
12113 (+ number ; @r{else-part}
12114 (triangle-recursively ; @r{recursive call}
12115 (1- number))))) ; @r{next-step-expression}
12119 What happens when we call this function with a argument of 7?
12121 The first instance of the @code{triangle-recursively} function adds
12122 the number 7 to the value returned by a second instance of
12123 @code{triangle-recursively}, an instance that has been passed an
12124 argument of 6. That is to say, the first calculation is:
12127 (+ 7 (triangle-recursively 6))
12131 The first instance of @code{triangle-recursively}---you may want to
12132 think of it as a little robot---cannot complete its job. It must hand
12133 off the calculation for @code{(triangle-recursively 6)} to a second
12134 instance of the program, to a second robot. This second individual is
12135 completely different from the first one; it is, in the jargon, a
12136 `different instantiation'. Or, put another way, it is a different
12137 robot. It is the same model as the first; it calculates triangle
12138 numbers recursively; but it has a different serial number.
12140 And what does @code{(triangle-recursively 6)} return? It returns the
12141 number 6 added to the value returned by evaluating
12142 @code{triangle-recursively} with an argument of 5. Using the robot
12143 metaphor, it asks yet another robot to help it.
12149 (+ 7 6 (triangle-recursively 5))
12153 And what happens next?
12156 (+ 7 6 5 (triangle-recursively 4))
12159 Each time @code{triangle-recursively} is called, except for the last
12160 time, it creates another instance of the program---another robot---and
12161 asks it to make a calculation.
12164 Eventually, the full addition is set up and performed:
12170 This design for the function defers the calculation of the first step
12171 until the second can be done, and defers that until the third can be
12172 done, and so on. Each deferment means the computer must remember what
12173 is being waited on. This is not a problem when there are only a few
12174 steps, as in this example. But it can be a problem when there are
12177 @node No deferment solution, , No Deferment, Recursion
12178 @subsection No Deferment Solution
12179 @cindex No deferment solution
12180 @cindex Defermentless solution
12181 @cindex Solution without deferment
12183 The solution to the problem of deferred operations is to write in a
12184 manner that does not defer operations@footnote{The phrase @dfn{tail
12185 recursive} is used to describe such a process, one that uses
12186 `constant space'.}. This requires
12187 writing to a different pattern, often one that involves writing two
12188 function definitions, an `initialization' function and a `helper'
12191 The `initialization' function sets up the job; the `helper' function
12195 Here are the two function definitions for adding up numbers. They are
12196 so simple, I find them hard to understand.
12200 (defun triangle-initialization (number)
12201 "Return the sum of the numbers 1 through NUMBER inclusive.
12202 This is the `initialization' component of a two function
12203 duo that uses recursion."
12204 (triangle-recursive-helper 0 0 number))
12210 (defun triangle-recursive-helper (sum counter number)
12211 "Return SUM, using COUNTER, through NUMBER inclusive.
12212 This is the `helper' component of a two function duo
12213 that uses recursion."
12214 (if (> counter number)
12216 (triangle-recursive-helper (+ sum counter) ; @r{sum}
12217 (1+ counter) ; @r{counter}
12218 number))) ; @r{number}
12223 Install both function definitions by evaluating them, then call
12224 @code{triangle-initialization} with 2 rows:
12228 (triangle-initialization 2)
12233 The `initialization' function calls the first instance of the `helper'
12234 function with three arguments: zero, zero, and a number which is the
12235 number of rows in the triangle.
12237 The first two arguments passed to the `helper' function are
12238 initialization values. These values are changed when
12239 @code{triangle-recursive-helper} invokes new instances.@footnote{The
12240 jargon is mildly confusing: @code{triangle-recursive-helper} uses a
12241 process that is iterative in a procedure that is recursive. The
12242 process is called iterative because the computer need only record the
12243 three values, @code{sum}, @code{counter}, and @code{number}; the
12244 procedure is recursive because the function `calls itself'. On the
12245 other hand, both the process and the procedure used by
12246 @code{triangle-recursively} are called recursive. The word
12247 `recursive' has different meanings in the two contexts.}
12249 Let's see what happens when we have a triangle that has one row. (This
12250 triangle will have one pebble in it!)
12253 @code{triangle-initialization} will call its helper with
12254 the arguments @w{@code{0 0 1}}. That function will run the conditional
12255 test whether @code{(> counter number)}:
12263 and find that the result is false, so it will invoke
12264 the else-part of the @code{if} clause:
12268 (triangle-recursive-helper
12269 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12270 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12271 number) ; @r{number stays the same}
12277 which will first compute:
12281 (triangle-recursive-helper (+ 0 0) ; @r{sum}
12282 (1+ 0) ; @r{counter}
12286 (triangle-recursive-helper 0 1 1)
12290 Again, @code{(> counter number)} will be false, so again, the Lisp
12291 interpreter will evaluate @code{triangle-recursive-helper}, creating a
12292 new instance with new arguments.
12295 This new instance will be;
12299 (triangle-recursive-helper
12300 (+ sum counter) ; @r{sum plus counter} @result{} @r{sum}
12301 (1+ counter) ; @r{increment counter} @result{} @r{counter}
12302 number) ; @r{number stays the same}
12306 (triangle-recursive-helper 1 2 1)
12310 In this case, the @code{(> counter number)} test will be true! So the
12311 instance will return the value of the sum, which will be 1, as
12314 Now, let's pass @code{triangle-initialization} an argument
12315 of 2, to find out how many pebbles there are in a triangle with two rows.
12317 That function calls @code{(triangle-recursive-helper 0 0 2)}.
12320 In stages, the instances called will be:
12324 @r{sum counter number}
12325 (triangle-recursive-helper 0 1 2)
12327 (triangle-recursive-helper 1 2 2)
12329 (triangle-recursive-helper 3 3 2)
12333 When the last instance is called, the @code{(> counter number)} test
12334 will be true, so the instance will return the value of @code{sum},
12337 This kind of pattern helps when you are writing functions that can use
12338 many resources in a computer.
12341 @node Looping exercise, , Recursion, Loops & Recursion
12342 @section Looping Exercise
12346 Write a function similar to @code{triangle} in which each row has a
12347 value which is the square of the row number. Use a @code{while} loop.
12350 Write a function similar to @code{triangle} that multiplies instead of
12354 Rewrite these two functions recursively. Rewrite these functions
12357 @c comma in printed title causes problem in Info cross reference
12359 Write a function for Texinfo mode that creates an index entry at the
12360 beginning of a paragraph for every @samp{@@dfn} within the paragraph.
12361 (In a Texinfo file, @samp{@@dfn} marks a definition. This book is
12362 written in Texinfo.)
12364 Many of the functions you will need are described in two of the
12365 previous chapters, @ref{Cutting & Storing Text, , Cutting and Storing
12366 Text}, and @ref{Yanking, , Yanking Text Back}. If you use
12367 @code{forward-paragraph} to put the index entry at the beginning of
12368 the paragraph, you will have to use @w{@kbd{C-h f}}
12369 (@code{describe-function}) to find out how to make the command go
12372 For more information, see
12374 @ref{Indicating, , Indicating Definitions, texinfo}.
12377 @ref{Indicating, , Indicating, texinfo, Texinfo Manual}, which goes to
12378 a Texinfo manual in the current directory. Or, if you are on the
12380 @uref{http://www.gnu.org/software/texinfo/manual/texinfo/}
12383 ``Indicating Definitions, Commands, etc.'' in @cite{Texinfo, The GNU
12384 Documentation Format}.
12388 @node Regexp Search, Counting Words, Loops & Recursion, Top
12389 @comment node-name, next, previous, up
12390 @chapter Regular Expression Searches
12391 @cindex Searches, illustrating
12392 @cindex Regular expression searches
12393 @cindex Patterns, searching for
12394 @cindex Motion by sentence and paragraph
12395 @cindex Sentences, movement by
12396 @cindex Paragraphs, movement by
12398 Regular expression searches are used extensively in GNU Emacs. The
12399 two functions, @code{forward-sentence} and @code{forward-paragraph},
12400 illustrate these searches well. They use regular expressions to find
12401 where to move point. The phrase `regular expression' is often written
12404 Regular expression searches are described in @ref{Regexp Search, ,
12405 Regular Expression Search, emacs, The GNU Emacs Manual}, as well as in
12406 @ref{Regular Expressions, , , elisp, The GNU Emacs Lisp Reference
12407 Manual}. In writing this chapter, I am presuming that you have at
12408 least a mild acquaintance with them. The major point to remember is
12409 that regular expressions permit you to search for patterns as well as
12410 for literal strings of characters. For example, the code in
12411 @code{forward-sentence} searches for the pattern of possible
12412 characters that could mark the end of a sentence, and moves point to
12415 Before looking at the code for the @code{forward-sentence} function, it
12416 is worth considering what the pattern that marks the end of a sentence
12417 must be. The pattern is discussed in the next section; following that
12418 is a description of the regular expression search function,
12419 @code{re-search-forward}. The @code{forward-sentence} function
12420 is described in the section following. Finally, the
12421 @code{forward-paragraph} function is described in the last section of
12422 this chapter. @code{forward-paragraph} is a complex function that
12423 introduces several new features.
12426 * sentence-end:: The regular expression for @code{sentence-end}.
12427 * re-search-forward:: Very similar to @code{search-forward}.
12428 * forward-sentence:: A straightforward example of regexp search.
12429 * forward-paragraph:: A somewhat complex example.
12430 * etags:: How to create your own @file{TAGS} table.
12432 * re-search Exercises::
12435 @node sentence-end, re-search-forward, Regexp Search, Regexp Search
12436 @comment node-name, next, previous, up
12437 @section The Regular Expression for @code{sentence-end}
12438 @findex sentence-end
12440 The symbol @code{sentence-end} is bound to the pattern that marks the
12441 end of a sentence. What should this regular expression be?
12443 Clearly, a sentence may be ended by a period, a question mark, or an
12444 exclamation mark. Indeed, in English, only clauses that end with one
12445 of those three characters should be considered the end of a sentence.
12446 This means that the pattern should include the character set:
12452 However, we do not want @code{forward-sentence} merely to jump to a
12453 period, a question mark, or an exclamation mark, because such a character
12454 might be used in the middle of a sentence. A period, for example, is
12455 used after abbreviations. So other information is needed.
12457 According to convention, you type two spaces after every sentence, but
12458 only one space after a period, a question mark, or an exclamation mark in
12459 the body of a sentence. So a period, a question mark, or an exclamation
12460 mark followed by two spaces is a good indicator of an end of sentence.
12461 However, in a file, the two spaces may instead be a tab or the end of a
12462 line. This means that the regular expression should include these three
12463 items as alternatives.
12466 This group of alternatives will look like this:
12477 Here, @samp{$} indicates the end of the line, and I have pointed out
12478 where the tab and two spaces are inserted in the expression. Both are
12479 inserted by putting the actual characters into the expression.
12481 Two backslashes, @samp{\\}, are required before the parentheses and
12482 vertical bars: the first backslash quotes the following backslash in
12483 Emacs; and the second indicates that the following character, the
12484 parenthesis or the vertical bar, is special.
12487 Also, a sentence may be followed by one or more carriage returns, like
12498 Like tabs and spaces, a carriage return is inserted into a regular
12499 expression by inserting it literally. The asterisk indicates that the
12500 @key{RET} is repeated zero or more times.
12502 But a sentence end does not consist only of a period, a question mark or
12503 an exclamation mark followed by appropriate space: a closing quotation
12504 mark or a closing brace of some kind may precede the space. Indeed more
12505 than one such mark or brace may precede the space. These require a
12506 expression that looks like this:
12512 In this expression, the first @samp{]} is the first character in the
12513 expression; the second character is @samp{"}, which is preceded by a
12514 @samp{\} to tell Emacs the @samp{"} is @emph{not} special. The last
12515 three characters are @samp{'}, @samp{)}, and @samp{@}}.
12517 All this suggests what the regular expression pattern for matching the
12518 end of a sentence should be; and, indeed, if we evaluate
12519 @code{sentence-end} we find that it returns the following value:
12524 @result{} "[.?!][]\"')@}]*\\($\\| \\| \\)[
12530 (Well, not in GNU Emacs 22; that is because of an effort to make the
12531 process simpler and to handle more glyphs and languages. When the
12532 value of @code{sentence-end} is @code{nil}, then use the value defined
12533 by the function @code{sentence-end}. (Here is a use of the difference
12534 between a value and a function in Emacs Lisp.) The function returns a
12535 value constructed from the variables @code{sentence-end-base},
12536 @code{sentence-end-double-space}, @code{sentence-end-without-period},
12537 and @code{sentence-end-without-space}. The critical variable is
12538 @code{sentence-end-base}; its global value is similar to the one
12539 described above but it also contains two additional quotation marks.
12540 These have differing degrees of curliness. The
12541 @code{sentence-end-without-period} variable, when true, tells Emacs
12542 that a sentence may end without a period, such as text in Thai.)
12546 (Note that here the @key{TAB}, two spaces, and @key{RET} are shown
12547 literally in the pattern.)
12549 This regular expression can be deciphered as follows:
12553 The first part of the pattern is the three characters, a period, a question
12554 mark and an exclamation mark, within square brackets. The pattern must
12555 begin with one or other of these characters.
12558 The second part of the pattern is the group of closing braces and
12559 quotation marks, which can appear zero or more times. These may follow
12560 the period, question mark or exclamation mark. In a regular expression,
12561 the backslash, @samp{\}, followed by the double quotation mark,
12562 @samp{"}, indicates the class of string-quote characters. Usually, the
12563 double quotation mark is the only character in this class. The
12564 asterisk, @samp{*}, indicates that the items in the previous group (the
12565 group surrounded by square brackets, @samp{[]}) may be repeated zero or
12568 @item \\($\\| \\| \\)
12569 The third part of the pattern is one or other of: either the end of a
12570 line, or two blank spaces, or a tab. The double back-slashes are used
12571 to prevent Emacs from reading the parentheses and vertical bars as part
12572 of the search pattern; the parentheses are used to mark the group and
12573 the vertical bars are used to indicated that the patterns to either side
12574 of them are alternatives. The dollar sign is used to indicate the end
12575 of a line and both the two spaces and the tab are each inserted as is to
12576 indicate what they are.
12579 Finally, the last part of the pattern indicates that the end of the line
12580 or the whitespace following the period, question mark or exclamation
12581 mark may, but need not, be followed by one or more carriage returns. In
12582 the pattern, the carriage return is inserted as an actual carriage
12583 return between square brackets but here it is shown as @key{RET}.
12587 @node re-search-forward, forward-sentence, sentence-end, Regexp Search
12588 @comment node-name, next, previous, up
12589 @section The @code{re-search-forward} Function
12590 @findex re-search-forward
12592 The @code{re-search-forward} function is very like the
12593 @code{search-forward} function. (@xref{search-forward, , The
12594 @code{search-forward} Function}.)
12596 @code{re-search-forward} searches for a regular expression. If the
12597 search is successful, it leaves point immediately after the last
12598 character in the target. If the search is backwards, it leaves point
12599 just before the first character in the target. You may tell
12600 @code{re-search-forward} to return @code{t} for true. (Moving point
12601 is therefore a `side effect'.)
12603 Like @code{search-forward}, the @code{re-search-forward} function takes
12608 The first argument is the regular expression that the function searches
12609 for. The regular expression will be a string between quotations marks.
12612 The optional second argument limits how far the function will search; it is a
12613 bound, which is specified as a position in the buffer.
12616 The optional third argument specifies how the function responds to
12617 failure: @code{nil} as the third argument causes the function to
12618 signal an error (and print a message) when the search fails; any other
12619 value causes it to return @code{nil} if the search fails and @code{t}
12620 if the search succeeds.
12623 The optional fourth argument is the repeat count. A negative repeat
12624 count causes @code{re-search-forward} to search backwards.
12628 The template for @code{re-search-forward} looks like this:
12632 (re-search-forward "@var{regular-expression}"
12633 @var{limit-of-search}
12634 @var{what-to-do-if-search-fails}
12635 @var{repeat-count})
12639 The second, third, and fourth arguments are optional. However, if you
12640 want to pass a value to either or both of the last two arguments, you
12641 must also pass a value to all the preceding arguments. Otherwise, the
12642 Lisp interpreter will mistake which argument you are passing the value
12646 In the @code{forward-sentence} function, the regular expression will be
12647 the value of the variable @code{sentence-end}. In simple form, that is:
12651 "[.?!][]\"')@}]*\\($\\| \\| \\)[
12657 The limit of the search will be the end of the paragraph (since a
12658 sentence cannot go beyond a paragraph). If the search fails, the
12659 function will return @code{nil}; and the repeat count will be provided
12660 by the argument to the @code{forward-sentence} function.
12662 @node forward-sentence, forward-paragraph, re-search-forward, Regexp Search
12663 @comment node-name, next, previous, up
12664 @section @code{forward-sentence}
12665 @findex forward-sentence
12667 The command to move the cursor forward a sentence is a straightforward
12668 illustration of how to use regular expression searches in Emacs Lisp.
12669 Indeed, the function looks longer and more complicated than it is; this
12670 is because the function is designed to go backwards as well as forwards;
12671 and, optionally, over more than one sentence. The function is usually
12672 bound to the key command @kbd{M-e}.
12675 * Complete forward-sentence::
12676 * fwd-sentence while loops:: Two @code{while} loops.
12677 * fwd-sentence re-search:: A regular expression search.
12680 @node Complete forward-sentence, fwd-sentence while loops, forward-sentence, forward-sentence
12682 @unnumberedsubsec Complete @code{forward-sentence} function definition
12686 Here is the code for @code{forward-sentence}:
12691 (defun forward-sentence (&optional arg)
12692 "Move forward to next `sentence-end'. With argument, repeat.
12693 With negative argument, move backward repeatedly to `sentence-beginning'.
12695 The variable `sentence-end' is a regular expression that matches ends of
12696 sentences. Also, every paragraph boundary terminates sentences as well."
12700 (or arg (setq arg 1))
12701 (let ((opoint (point))
12702 (sentence-end (sentence-end)))
12704 (let ((pos (point))
12705 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12706 (if (and (re-search-backward sentence-end par-beg t)
12707 (or (< (match-end 0) pos)
12708 (re-search-backward sentence-end par-beg t)))
12709 (goto-char (match-end 0))
12710 (goto-char par-beg)))
12711 (setq arg (1+ arg)))
12715 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12716 (if (re-search-forward sentence-end par-end t)
12717 (skip-chars-backward " \t\n")
12718 (goto-char par-end)))
12719 (setq arg (1- arg)))
12720 (constrain-to-field nil opoint t)))
12728 (defun forward-sentence (&optional arg)
12729 "Move forward to next sentence-end. With argument, repeat.
12730 With negative argument, move backward repeatedly to sentence-beginning.
12731 Sentence ends are identified by the value of sentence-end
12732 treated as a regular expression. Also, every paragraph boundary
12733 terminates sentences as well."
12737 (or arg (setq arg 1))
12740 (save-excursion (start-of-paragraph-text) (point))))
12741 (if (re-search-backward
12742 (concat sentence-end "[^ \t\n]") par-beg t)
12743 (goto-char (1- (match-end 0)))
12744 (goto-char par-beg)))
12745 (setq arg (1+ arg)))
12748 (save-excursion (end-of-paragraph-text) (point))))
12749 (if (re-search-forward sentence-end par-end t)
12750 (skip-chars-backward " \t\n")
12751 (goto-char par-end)))
12752 (setq arg (1- arg))))
12757 The function looks long at first sight and it is best to look at its
12758 skeleton first, and then its muscle. The way to see the skeleton is to
12759 look at the expressions that start in the left-most columns:
12763 (defun forward-sentence (&optional arg)
12764 "@var{documentation}@dots{}"
12766 (or arg (setq arg 1))
12767 (let ((opoint (point)) (sentence-end (sentence-end)))
12769 (let ((pos (point))
12770 (par-beg (save-excursion (start-of-paragraph-text) (point))))
12771 @var{rest-of-body-of-while-loop-when-going-backwards}
12773 (let ((par-end (save-excursion (end-of-paragraph-text) (point))))
12774 @var{rest-of-body-of-while-loop-when-going-forwards}
12775 @var{handle-forms-and-equivalent}
12779 This looks much simpler! The function definition consists of
12780 documentation, an @code{interactive} expression, an @code{or}
12781 expression, a @code{let} expression, and @code{while} loops.
12783 Let's look at each of these parts in turn.
12785 We note that the documentation is thorough and understandable.
12787 The function has an @code{interactive "p"} declaration. This means
12788 that the processed prefix argument, if any, is passed to the
12789 function as its argument. (This will be a number.) If the function
12790 is not passed an argument (it is optional) then the argument
12791 @code{arg} will be bound to 1.
12793 When @code{forward-sentence} is called non-interactively without an
12794 argument, @code{arg} is bound to @code{nil}. The @code{or} expression
12795 handles this. What it does is either leave the value of @code{arg} as
12796 it is, but only if @code{arg} is bound to a value; or it sets the
12797 value of @code{arg} to 1, in the case when @code{arg} is bound to
12800 Next is a @code{let}. That specifies the values of two local
12801 variables, @code{point} and @code{sentence-end}. The local value of
12802 point, from before the search, is used in the
12803 @code{constrain-to-field} function which handles forms and
12804 equivalents. The @code{sentence-end} variable is set by the
12805 @code{sentence-end} function.
12807 @node fwd-sentence while loops, fwd-sentence re-search, Complete forward-sentence, forward-sentence
12808 @unnumberedsubsec The @code{while} loops
12810 Two @code{while} loops follow. The first @code{while} has a
12811 true-or-false-test that tests true if the prefix argument for
12812 @code{forward-sentence} is a negative number. This is for going
12813 backwards. The body of this loop is similar to the body of the second
12814 @code{while} clause, but it is not exactly the same. We will skip
12815 this @code{while} loop and concentrate on the second @code{while}
12819 The second @code{while} loop is for moving point forward. Its skeleton
12824 (while (> arg 0) ; @r{true-or-false-test}
12826 (if (@var{true-or-false-test})
12829 (setq arg (1- arg)))) ; @code{while} @r{loop decrementer}
12833 The @code{while} loop is of the decrementing kind.
12834 (@xref{Decrementing Loop, , A Loop with a Decrementing Counter}.) It
12835 has a true-or-false-test that tests true so long as the counter (in
12836 this case, the variable @code{arg}) is greater than zero; and it has a
12837 decrementer that subtracts 1 from the value of the counter every time
12840 If no prefix argument is given to @code{forward-sentence}, which is
12841 the most common way the command is used, this @code{while} loop will
12842 run once, since the value of @code{arg} will be 1.
12844 The body of the @code{while} loop consists of a @code{let} expression,
12845 which creates and binds a local variable, and has, as its body, an
12846 @code{if} expression.
12849 The body of the @code{while} loop looks like this:
12854 (save-excursion (end-of-paragraph-text) (point))))
12855 (if (re-search-forward sentence-end par-end t)
12856 (skip-chars-backward " \t\n")
12857 (goto-char par-end)))
12861 The @code{let} expression creates and binds the local variable
12862 @code{par-end}. As we shall see, this local variable is designed to
12863 provide a bound or limit to the regular expression search. If the
12864 search fails to find a proper sentence ending in the paragraph, it will
12865 stop on reaching the end of the paragraph.
12867 But first, let us examine how @code{par-end} is bound to the value of
12868 the end of the paragraph. What happens is that the @code{let} sets the
12869 value of @code{par-end} to the value returned when the Lisp interpreter
12870 evaluates the expression
12874 (save-excursion (end-of-paragraph-text) (point))
12879 In this expression, @code{(end-of-paragraph-text)} moves point to the
12880 end of the paragraph, @code{(point)} returns the value of point, and then
12881 @code{save-excursion} restores point to its original position. Thus,
12882 the @code{let} binds @code{par-end} to the value returned by the
12883 @code{save-excursion} expression, which is the position of the end of
12884 the paragraph. (The @code{end-of-paragraph-text} function uses
12885 @code{forward-paragraph}, which we will discuss shortly.)
12888 Emacs next evaluates the body of the @code{let}, which is an @code{if}
12889 expression that looks like this:
12893 (if (re-search-forward sentence-end par-end t) ; @r{if-part}
12894 (skip-chars-backward " \t\n") ; @r{then-part}
12895 (goto-char par-end))) ; @r{else-part}
12899 The @code{if} tests whether its first argument is true and if so,
12900 evaluates its then-part; otherwise, the Emacs Lisp interpreter
12901 evaluates the else-part. The true-or-false-test of the @code{if}
12902 expression is the regular expression search.
12904 It may seem odd to have what looks like the `real work' of
12905 the @code{forward-sentence} function buried here, but this is a common
12906 way this kind of operation is carried out in Lisp.
12908 @node fwd-sentence re-search, , fwd-sentence while loops, forward-sentence
12909 @unnumberedsubsec The regular expression search
12911 The @code{re-search-forward} function searches for the end of the
12912 sentence, that is, for the pattern defined by the @code{sentence-end}
12913 regular expression. If the pattern is found---if the end of the sentence is
12914 found---then the @code{re-search-forward} function does two things:
12918 The @code{re-search-forward} function carries out a side effect, which
12919 is to move point to the end of the occurrence found.
12922 The @code{re-search-forward} function returns a value of true. This is
12923 the value received by the @code{if}, and means that the search was
12928 The side effect, the movement of point, is completed before the
12929 @code{if} function is handed the value returned by the successful
12930 conclusion of the search.
12932 When the @code{if} function receives the value of true from a successful
12933 call to @code{re-search-forward}, the @code{if} evaluates the then-part,
12934 which is the expression @code{(skip-chars-backward " \t\n")}. This
12935 expression moves backwards over any blank spaces, tabs or carriage
12936 returns until a printed character is found and then leaves point after
12937 the character. Since point has already been moved to the end of the
12938 pattern that marks the end of the sentence, this action leaves point
12939 right after the closing printed character of the sentence, which is
12942 On the other hand, if the @code{re-search-forward} function fails to
12943 find a pattern marking the end of the sentence, the function returns
12944 false. The false then causes the @code{if} to evaluate its third
12945 argument, which is @code{(goto-char par-end)}: it moves point to the
12946 end of the paragraph.
12948 (And if the text is in a form or equivalent, and point may not move
12949 fully, then the @code{constrain-to-field} function comes into play.)
12951 Regular expression searches are exceptionally useful and the pattern
12952 illustrated by @code{re-search-forward}, in which the search is the
12953 test of an @code{if} expression, is handy. You will see or write code
12954 incorporating this pattern often.
12956 @node forward-paragraph, etags, forward-sentence, Regexp Search
12957 @comment node-name, next, previous, up
12958 @section @code{forward-paragraph}: a Goldmine of Functions
12959 @findex forward-paragraph
12963 (defun forward-paragraph (&optional arg)
12964 "Move forward to end of paragraph.
12965 With argument ARG, do it ARG times;
12966 a negative argument ARG = -N means move backward N paragraphs.
12968 A line which `paragraph-start' matches either separates paragraphs
12969 \(if `paragraph-separate' matches it also) or is the first line of a paragraph.
12970 A paragraph end is the beginning of a line which is not part of the paragraph
12971 to which the end of the previous line belongs, or the end of the buffer.
12972 Returns the count of paragraphs left to move."
12974 (or arg (setq arg 1))
12975 (let* ((opoint (point))
12976 (fill-prefix-regexp
12977 (and fill-prefix (not (equal fill-prefix ""))
12978 (not paragraph-ignore-fill-prefix)
12979 (regexp-quote fill-prefix)))
12980 ;; Remove ^ from paragraph-start and paragraph-sep if they are there.
12981 ;; These regexps shouldn't be anchored, because we look for them
12982 ;; starting at the left-margin. This allows paragraph commands to
12983 ;; work normally with indented text.
12984 ;; This hack will not find problem cases like "whatever\\|^something".
12985 (parstart (if (and (not (equal "" paragraph-start))
12986 (equal ?^ (aref paragraph-start 0)))
12987 (substring paragraph-start 1)
12989 (parsep (if (and (not (equal "" paragraph-separate))
12990 (equal ?^ (aref paragraph-separate 0)))
12991 (substring paragraph-separate 1)
12992 paragraph-separate))
12994 (if fill-prefix-regexp
12995 (concat parsep "\\|"
12996 fill-prefix-regexp "[ \t]*$")
12998 ;; This is used for searching.
12999 (sp-parstart (concat "^[ \t]*\\(?:" parstart "\\|" parsep "\\)"))
13001 (while (and (< arg 0) (not (bobp)))
13002 (if (and (not (looking-at parsep))
13003 (re-search-backward "^\n" (max (1- (point)) (point-min)) t)
13004 (looking-at parsep))
13005 (setq arg (1+ arg))
13006 (setq start (point))
13007 ;; Move back over paragraph-separating lines.
13008 (forward-char -1) (beginning-of-line)
13009 (while (and (not (bobp))
13010 (progn (move-to-left-margin)
13011 (looking-at parsep)))
13015 (setq arg (1+ arg))
13016 ;; Go to end of the previous (non-separating) line.
13018 ;; Search back for line that starts or separates paragraphs.
13019 (if (if fill-prefix-regexp
13020 ;; There is a fill prefix; it overrides parstart.
13021 (let (multiple-lines)
13022 (while (and (progn (beginning-of-line) (not (bobp)))
13023 (progn (move-to-left-margin)
13024 (not (looking-at parsep)))
13025 (looking-at fill-prefix-regexp))
13026 (unless (= (point) start)
13027 (setq multiple-lines t))
13029 (move-to-left-margin)
13030 ;; This deleted code caused a long hanging-indent line
13031 ;; not to be filled together with the following lines.
13032 ;; ;; Don't move back over a line before the paragraph
13033 ;; ;; which doesn't start with fill-prefix
13034 ;; ;; unless that is the only line we've moved over.
13035 ;; (and (not (looking-at fill-prefix-regexp))
13037 ;; (forward-line 1))
13039 (while (and (re-search-backward sp-parstart nil 1)
13040 (setq found-start t)
13041 ;; Found a candidate, but need to check if it is a
13043 (progn (setq start (point))
13044 (move-to-left-margin)
13045 (not (looking-at parsep)))
13046 (not (and (looking-at parstart)
13047 (or (not use-hard-newlines)
13050 (1- start) 'hard)))))
13051 (setq found-start nil)
13056 ;; Move forward over paragraph separators.
13057 ;; We know this cannot reach the place we started
13058 ;; because we know we moved back over a non-separator.
13059 (while (and (not (eobp))
13060 (progn (move-to-left-margin)
13061 (looking-at parsep)))
13063 ;; If line before paragraph is just margin, back up to there.
13065 (if (> (current-column) (current-left-margin))
13067 (skip-chars-backward " \t")
13069 (forward-line 1))))
13070 ;; No starter or separator line => use buffer beg.
13071 (goto-char (point-min))))))
13073 (while (and (> arg 0) (not (eobp)))
13074 ;; Move forward over separator lines...
13075 (while (and (not (eobp))
13076 (progn (move-to-left-margin) (not (eobp)))
13077 (looking-at parsep))
13079 (unless (eobp) (setq arg (1- arg)))
13080 ;; ... and one more line.
13082 (if fill-prefix-regexp
13083 ;; There is a fill prefix; it overrides parstart.
13084 (while (and (not (eobp))
13085 (progn (move-to-left-margin) (not (eobp)))
13086 (not (looking-at parsep))
13087 (looking-at fill-prefix-regexp))
13089 (while (and (re-search-forward sp-parstart nil 1)
13090 (progn (setq start (match-beginning 0))
13093 (progn (move-to-left-margin)
13094 (not (looking-at parsep)))
13095 (or (not (looking-at parstart))
13096 (and use-hard-newlines
13097 (not (get-text-property (1- start) 'hard)))))
13099 (if (< (point) (point-max))
13100 (goto-char start))))
13101 (constrain-to-field nil opoint t)
13102 ;; Return the number of steps that could not be done.
13106 The @code{forward-paragraph} function moves point forward to the end
13107 of the paragraph. It is usually bound to @kbd{M-@}} and makes use of a
13108 number of functions that are important in themselves, including
13109 @code{let*}, @code{match-beginning}, and @code{looking-at}.
13111 The function definition for @code{forward-paragraph} is considerably
13112 longer than the function definition for @code{forward-sentence}
13113 because it works with a paragraph, each line of which may begin with a
13116 A fill prefix consists of a string of characters that are repeated at
13117 the beginning of each line. For example, in Lisp code, it is a
13118 convention to start each line of a paragraph-long comment with
13119 @samp{;;; }. In Text mode, four blank spaces make up another common
13120 fill prefix, creating an indented paragraph. (@xref{Fill Prefix, , ,
13121 emacs, The GNU Emacs Manual}, for more information about fill
13124 The existence of a fill prefix means that in addition to being able to
13125 find the end of a paragraph whose lines begin on the left-most
13126 column, the @code{forward-paragraph} function must be able to find the
13127 end of a paragraph when all or many of the lines in the buffer begin
13128 with the fill prefix.
13130 Moreover, it is sometimes practical to ignore a fill prefix that
13131 exists, especially when blank lines separate paragraphs.
13132 This is an added complication.
13135 * forward-paragraph in brief:: Key parts of the function definition.
13136 * fwd-para let:: The @code{let*} expression.
13137 * fwd-para while:: The forward motion @code{while} loop.
13140 @node forward-paragraph in brief, fwd-para let, forward-paragraph, forward-paragraph
13142 @unnumberedsubsec Shortened @code{forward-paragraph} function definition
13145 Rather than print all of the @code{forward-paragraph} function, we
13146 will only print parts of it. Read without preparation, the function
13150 In outline, the function looks like this:
13154 (defun forward-paragraph (&optional arg)
13155 "@var{documentation}@dots{}"
13157 (or arg (setq arg 1))
13160 (while (and (< arg 0) (not (bobp))) ; @r{backward-moving-code}
13162 (while (and (> arg 0) (not (eobp))) ; @r{forward-moving-code}
13167 The first parts of the function are routine: the function's argument
13168 list consists of one optional argument. Documentation follows.
13170 The lower case @samp{p} in the @code{interactive} declaration means
13171 that the processed prefix argument, if any, is passed to the function.
13172 This will be a number, and is the repeat count of how many paragraphs
13173 point will move. The @code{or} expression in the next line handles
13174 the common case when no argument is passed to the function, which occurs
13175 if the function is called from other code rather than interactively.
13176 This case was described earlier. (@xref{forward-sentence, The
13177 @code{forward-sentence} function}.) Now we reach the end of the
13178 familiar part of this function.
13180 @node fwd-para let, fwd-para while, forward-paragraph in brief, forward-paragraph
13181 @unnumberedsubsec The @code{let*} expression
13183 The next line of the @code{forward-paragraph} function begins a
13184 @code{let*} expression. This is a different than @code{let}. The
13185 symbol is @code{let*} not @code{let}.
13187 The @code{let*} special form is like @code{let} except that Emacs sets
13188 each variable in sequence, one after another, and variables in the
13189 latter part of the varlist can make use of the values to which Emacs
13190 set variables in the earlier part of the varlist.
13193 ( refappend save-excursion, , code save-excursion in code append-to-buffer .)
13196 (@ref{append save-excursion, , @code{save-excursion} in @code{append-to-buffer}}.)
13198 In the @code{let*} expression in this function, Emacs binds a total of
13199 seven variables: @code{opoint}, @code{fill-prefix-regexp},
13200 @code{parstart}, @code{parsep}, @code{sp-parstart}, @code{start}, and
13201 @code{found-start}.
13203 The variable @code{parsep} appears twice, first, to remove instances
13204 of @samp{^}, and second, to handle fill prefixes.
13206 The variable @code{opoint} is just the value of @code{point}. As you
13207 can guess, it is used in a @code{constrain-to-field} expression, just
13208 as in @code{forward-sentence}.
13210 The variable @code{fill-prefix-regexp} is set to the value returned by
13211 evaluating the following list:
13216 (not (equal fill-prefix ""))
13217 (not paragraph-ignore-fill-prefix)
13218 (regexp-quote fill-prefix))
13223 This is an expression whose first element is the @code{and} special form.
13225 As we learned earlier (@pxref{kill-new function, , The @code{kill-new}
13226 function}), the @code{and} special form evaluates each of its
13227 arguments until one of the arguments returns a value of @code{nil}, in
13228 which case the @code{and} expression returns @code{nil}; however, if
13229 none of the arguments returns a value of @code{nil}, the value
13230 resulting from evaluating the last argument is returned. (Since such
13231 a value is not @code{nil}, it is considered true in Lisp.) In other
13232 words, an @code{and} expression returns a true value only if all its
13233 arguments are true.
13236 In this case, the variable @code{fill-prefix-regexp} is bound to a
13237 non-@code{nil} value only if the following four expressions produce a
13238 true (i.e., a non-@code{nil}) value when they are evaluated; otherwise,
13239 @code{fill-prefix-regexp} is bound to @code{nil}.
13243 When this variable is evaluated, the value of the fill prefix, if any,
13244 is returned. If there is no fill prefix, this variable returns
13247 @item (not (equal fill-prefix "")
13248 This expression checks whether an existing fill prefix is an empty
13249 string, that is, a string with no characters in it. An empty string is
13250 not a useful fill prefix.
13252 @item (not paragraph-ignore-fill-prefix)
13253 This expression returns @code{nil} if the variable
13254 @code{paragraph-ignore-fill-prefix} has been turned on by being set to a
13255 true value such as @code{t}.
13257 @item (regexp-quote fill-prefix)
13258 This is the last argument to the @code{and} special form. If all the
13259 arguments to the @code{and} are true, the value resulting from
13260 evaluating this expression will be returned by the @code{and} expression
13261 and bound to the variable @code{fill-prefix-regexp},
13264 @findex regexp-quote
13266 The result of evaluating this @code{and} expression successfully is that
13267 @code{fill-prefix-regexp} will be bound to the value of
13268 @code{fill-prefix} as modified by the @code{regexp-quote} function.
13269 What @code{regexp-quote} does is read a string and return a regular
13270 expression that will exactly match the string and match nothing else.
13271 This means that @code{fill-prefix-regexp} will be set to a value that
13272 will exactly match the fill prefix if the fill prefix exists.
13273 Otherwise, the variable will be set to @code{nil}.
13275 The next two local variables in the @code{let*} expression are
13276 designed to remove instances of @samp{^} from @code{parstart} and
13277 @code{parsep}, the local variables which indicate the paragraph start
13278 and the paragraph separator. The next expression sets @code{parsep}
13279 again. That is to handle fill prefixes.
13281 This is the setting that requires the definition call @code{let*}
13282 rather than @code{let}. The true-or-false-test for the @code{if}
13283 depends on whether the variable @code{fill-prefix-regexp} evaluates to
13284 @code{nil} or some other value.
13286 If @code{fill-prefix-regexp} does not have a value, Emacs evaluates
13287 the else-part of the @code{if} expression and binds @code{parsep} to
13288 its local value. (@code{parsep} is a regular expression that matches
13289 what separates paragraphs.)
13291 But if @code{fill-prefix-regexp} does have a value, Emacs evaluates
13292 the then-part of the @code{if} expression and binds @code{parsep} to a
13293 regular expression that includes the @code{fill-prefix-regexp} as part
13296 Specifically, @code{parsep} is set to the original value of the
13297 paragraph separate regular expression concatenated with an alternative
13298 expression that consists of the @code{fill-prefix-regexp} followed by
13299 optional whitespace to the end of the line. The whitespace is defined
13300 by @w{@code{"[ \t]*$"}}.) The @samp{\\|} defines this portion of the
13301 regexp as an alternative to @code{parsep}.
13303 According to a comment in the code, the next local variable,
13304 @code{sp-parstart}, is used for searching, and then the final two,
13305 @code{start} and @code{found-start}, are set to @code{nil}.
13307 Now we get into the body of the @code{let*}. The first part of the body
13308 of the @code{let*} deals with the case when the function is given a
13309 negative argument and is therefore moving backwards. We will skip this
13312 @node fwd-para while, , fwd-para let, forward-paragraph
13313 @unnumberedsubsec The forward motion @code{while} loop
13315 The second part of the body of the @code{let*} deals with forward
13316 motion. It is a @code{while} loop that repeats itself so long as the
13317 value of @code{arg} is greater than zero. In the most common use of
13318 the function, the value of the argument is 1, so the body of the
13319 @code{while} loop is evaluated exactly once, and the cursor moves
13320 forward one paragraph.
13323 (while (and (> arg 0) (not (eobp)))
13325 ;; Move forward over separator lines...
13326 (while (and (not (eobp))
13327 (progn (move-to-left-margin) (not (eobp)))
13328 (looking-at parsep))
13330 (unless (eobp) (setq arg (1- arg)))
13331 ;; ... and one more line.
13334 (if fill-prefix-regexp
13335 ;; There is a fill prefix; it overrides parstart.
13336 (while (and (not (eobp))
13337 (progn (move-to-left-margin) (not (eobp)))
13338 (not (looking-at parsep))
13339 (looking-at fill-prefix-regexp))
13342 (while (and (re-search-forward sp-parstart nil 1)
13343 (progn (setq start (match-beginning 0))
13346 (progn (move-to-left-margin)
13347 (not (looking-at parsep)))
13348 (or (not (looking-at parstart))
13349 (and use-hard-newlines
13350 (not (get-text-property (1- start) 'hard)))))
13353 (if (< (point) (point-max))
13354 (goto-char start))))
13357 This part handles three situations: when point is between paragraphs,
13358 when there is a fill prefix and when there is no fill prefix.
13361 The @code{while} loop looks like this:
13365 ;; @r{going forwards and not at the end of the buffer}
13366 (while (and (> arg 0) (not (eobp)))
13368 ;; @r{between paragraphs}
13369 ;; Move forward over separator lines...
13370 (while (and (not (eobp))
13371 (progn (move-to-left-margin) (not (eobp)))
13372 (looking-at parsep))
13374 ;; @r{This decrements the loop}
13375 (unless (eobp) (setq arg (1- arg)))
13376 ;; ... and one more line.
13381 (if fill-prefix-regexp
13382 ;; There is a fill prefix; it overrides parstart;
13383 ;; we go forward line by line
13384 (while (and (not (eobp))
13385 (progn (move-to-left-margin) (not (eobp)))
13386 (not (looking-at parsep))
13387 (looking-at fill-prefix-regexp))
13392 ;; There is no fill prefix;
13393 ;; we go forward character by character
13394 (while (and (re-search-forward sp-parstart nil 1)
13395 (progn (setq start (match-beginning 0))
13398 (progn (move-to-left-margin)
13399 (not (looking-at parsep)))
13400 (or (not (looking-at parstart))
13401 (and use-hard-newlines
13402 (not (get-text-property (1- start) 'hard)))))
13407 ;; and if there is no fill prefix and if we are not at the end,
13408 ;; go to whatever was found in the regular expression search
13410 (if (< (point) (point-max))
13411 (goto-char start))))
13416 We can see that this is a decrementing counter @code{while} loop,
13417 using the expression @code{(setq arg (1- arg))} as the decrementer.
13418 That expression is not far from the @code{while}, but is hidden in
13419 another Lisp macro, an @code{unless} macro. Unless we are at the end
13420 of the buffer --- that is what the @code{eobp} function determines; it
13421 is an abbreviation of @samp{End Of Buffer P} --- we decrease the value
13422 of @code{arg} by one.
13424 (If we are at the end of the buffer, we cannot go forward any more and
13425 the next loop of the @code{while} expression will test false since the
13426 test is an @code{and} with @code{(not (eobp))}. The @code{not}
13427 function means exactly as you expect; it is another name for
13428 @code{null}, a function that returns true when its argument is false.)
13430 Interestingly, the loop count is not decremented until we leave the
13431 space between paragraphs, unless we come to the end of buffer or stop
13432 seeing the local value of the paragraph separator.
13434 That second @code{while} also has a @code{(move-to-left-margin)}
13435 expression. The function is self-explanatory. It is inside a
13436 @code{progn} expression and not the last element of its body, so it is
13437 only invoked for its side effect, which is to move point to the left
13438 margin of the current line.
13441 The @code{looking-at} function is also self-explanatory; it returns
13442 true if the text after point matches the regular expression given as
13445 The rest of the body of the loop looks difficult at first, but makes
13446 sense as you come to understand it.
13449 First consider what happens if there is a fill prefix:
13453 (if fill-prefix-regexp
13454 ;; There is a fill prefix; it overrides parstart;
13455 ;; we go forward line by line
13456 (while (and (not (eobp))
13457 (progn (move-to-left-margin) (not (eobp)))
13458 (not (looking-at parsep))
13459 (looking-at fill-prefix-regexp))
13465 This expression moves point forward line by line so long
13466 as four conditions are true:
13470 Point is not at the end of the buffer.
13473 We can move to the left margin of the text and are
13474 not at the end of the buffer.
13477 The text following point does not separate paragraphs.
13480 The pattern following point is the fill prefix regular expression.
13483 The last condition may be puzzling, until you remember that point was
13484 moved to the beginning of the line early in the @code{forward-paragraph}
13485 function. This means that if the text has a fill prefix, the
13486 @code{looking-at} function will see it.
13489 Consider what happens when there is no fill prefix.
13493 (while (and (re-search-forward sp-parstart nil 1)
13494 (progn (setq start (match-beginning 0))
13497 (progn (move-to-left-margin)
13498 (not (looking-at parsep)))
13499 (or (not (looking-at parstart))
13500 (and use-hard-newlines
13501 (not (get-text-property (1- start) 'hard)))))
13507 This @code{while} loop has us searching forward for
13508 @code{sp-parstart}, which is the combination of possible whitespace
13509 with a the local value of the start of a paragraph or of a paragraph
13510 separator. (The latter two are within an expression starting
13511 @code{\(?:} so that they are not referenced by the
13512 @code{match-beginning} function.)
13515 The two expressions,
13519 (setq start (match-beginning 0))
13525 mean go to the start of the text matched by the regular expression
13528 The @code{(match-beginning 0)} expression is new. It returns a number
13529 specifying the location of the start of the text that was matched by
13532 The @code{match-beginning} function is used here because of a
13533 characteristic of a forward search: a successful forward search,
13534 regardless of whether it is a plain search or a regular expression
13535 search, moves point to the end of the text that is found. In this
13536 case, a successful search moves point to the end of the pattern for
13537 @code{sp-parstart}.
13539 However, we want to put point at the end of the current paragraph, not
13540 somewhere else. Indeed, since the search possibly includes the
13541 paragraph separator, point may end up at the beginning of the next one
13542 unless we use an expression that includes @code{match-beginning}.
13544 @findex match-beginning
13545 When given an argument of 0, @code{match-beginning} returns the
13546 position that is the start of the text matched by the most recent
13547 search. In this case, the most recent search looks for
13548 @code{sp-parstart}. The @code{(match-beginning 0)} expression returns
13549 the beginning position of that pattern, rather than the end position
13552 (Incidentally, when passed a positive number as an argument, the
13553 @code{match-beginning} function returns the location of point at that
13554 parenthesized expression in the last search unless that parenthesized
13555 expression begins with @code{\(?:}. I don't know why @code{\(?:}
13556 appears here since the argument is 0.)
13559 The last expression when there is no fill prefix is
13563 (if (< (point) (point-max))
13564 (goto-char start))))
13569 This says that if there is no fill prefix and if we are not at the
13570 end, point should move to the beginning of whatever was found by the
13571 regular expression search for @code{sp-parstart}.
13573 The full definition for the @code{forward-paragraph} function not only
13574 includes code for going forwards, but also code for going backwards.
13576 If you are reading this inside of GNU Emacs and you want to see the
13577 whole function, you can type @kbd{C-h f} (@code{describe-function})
13578 and the name of the function. This gives you the function
13579 documentation and the name of the library containing the function's
13580 source. Place point over the name of the library and press the RET
13581 key; you will be taken directly to the source. (Be sure to install
13582 your sources! Without them, you are like a person who tries to drive
13583 a car with his eyes shut!)
13585 @node etags, Regexp Review, forward-paragraph, Regexp Search
13586 @section Create Your Own @file{TAGS} File
13588 @cindex @file{TAGS} file, create own
13590 Besides @kbd{C-h f} (@code{describe-function}), another way to see the
13591 source of a function is to type @kbd{M-.} (@code{find-tag}) and the
13592 name of the function when prompted for it. This is a good habit to
13593 get into. The @kbd{M-.} (@code{find-tag}) command takes you directly
13594 to the source for a function, variable, or node. The function depends
13595 on tags tables to tell it where to go.
13597 If the @code{find-tag} function first asks you for the name of a
13598 @file{TAGS} table, give it the name of a @file{TAGS} file such as
13599 @file{/usr/local/src/emacs/src/TAGS}. (The exact path to your
13600 @file{TAGS} file depends on how your copy of Emacs was installed. I
13601 just told you the location that provides both my C and my Emacs Lisp
13604 You can also create your own @file{TAGS} file for directories that
13607 You often need to build and install tags tables yourself. They are
13608 not built automatically. A tags table is called a @file{TAGS} file;
13609 the name is in upper case letters.
13611 You can create a @file{TAGS} file by calling the @code{etags} program
13612 that comes as a part of the Emacs distribution. Usually, @code{etags}
13613 is compiled and installed when Emacs is built. (@code{etags} is not
13614 an Emacs Lisp function or a part of Emacs; it is a C program.)
13617 To create a @file{TAGS} file, first switch to the directory in which
13618 you want to create the file. In Emacs you can do this with the
13619 @kbd{M-x cd} command, or by visiting a file in the directory, or by
13620 listing the directory with @kbd{C-x d} (@code{dired}). Then run the
13621 compile command, with @w{@code{etags *.el}} as the command to execute
13624 M-x compile RET etags *.el RET
13628 to create a @file{TAGS} file for Emacs Lisp.
13630 For example, if you have a large number of files in your
13631 @file{~/emacs} directory, as I do---I have 137 @file{.el} files in it,
13632 of which I load 12---you can create a @file{TAGS} file for the Emacs
13633 Lisp files in that directory.
13636 The @code{etags} program takes all the usual shell `wildcards'. For
13637 example, if you have two directories for which you want a single
13638 @file{TAGS} file, type @w{@code{etags *.el ../elisp/*.el}}, where
13639 @file{../elisp/} is the second directory:
13642 M-x compile RET etags *.el ../elisp/*.el RET
13649 M-x compile RET etags --help RET
13653 to see a list of the options accepted by @code{etags} as well as a
13654 list of supported languages.
13656 The @code{etags} program handles more than 20 languages, including
13657 Emacs Lisp, Common Lisp, Scheme, C, C++, Ada, Fortran, HTML, Java,
13658 LaTeX, Pascal, Perl, Postscript, Python, TeX, Texinfo, makefiles, and
13659 most assemblers. The program has no switches for specifying the
13660 language; it recognizes the language in an input file according to its
13661 file name and contents.
13663 @file{etags} is very helpful when you are writing code yourself and
13664 want to refer back to functions you have already written. Just run
13665 @code{etags} again at intervals as you write new functions, so they
13666 become part of the @file{TAGS} file.
13668 If you think an appropriate @file{TAGS} file already exists for what
13669 you want, but do not know where it is, you can use the @code{locate}
13670 program to attempt to find it.
13672 Type @w{@kbd{M-x locate @key{RET} TAGS @key{RET}}} and Emacs will list
13673 for you the full path names of all your @file{TAGS} files. On my
13674 system, this command lists 34 @file{TAGS} files. On the other hand, a
13675 `plain vanilla' system I recently installed did not contain any
13678 If the tags table you want has been created, you can use the @code{M-x
13679 visit-tags-table} command to specify it. Otherwise, you will need to
13680 create the tag table yourself and then use @code{M-x
13683 @subsubheading Building Tags in the Emacs sources
13684 @cindex Building Tags in the Emacs sources
13685 @cindex Tags in the Emacs sources
13688 The GNU Emacs sources come with a @file{Makefile} that contains a
13689 sophisticated @code{etags} command that creates, collects, and merges
13690 tags tables from all over the Emacs sources and puts the information
13691 into one @file{TAGS} file in the @file{src/} directory. (The
13692 @file{src/} directory is below the top level of your Emacs directory.)
13695 To build this @file{TAGS} file, go to the top level of your Emacs
13696 source directory and run the compile command @code{make tags}:
13699 M-x compile RET make tags RET
13703 (The @code{make tags} command works well with the GNU Emacs sources,
13704 as well as with some other source packages.)
13706 For more information, see @ref{Tags, , Tag Tables, emacs, The GNU Emacs
13709 @node Regexp Review, re-search Exercises, etags, Regexp Search
13710 @comment node-name, next, previous, up
13713 Here is a brief summary of some recently introduced functions.
13717 Repeatedly evaluate the body of the expression so long as the first
13718 element of the body tests true. Then return @code{nil}. (The
13719 expression is evaluated only for its side effects.)
13728 (insert (format "foo is %d.\n" foo))
13729 (setq foo (1- foo))))
13731 @result{} foo is 2.
13738 (The @code{insert} function inserts its arguments at point; the
13739 @code{format} function returns a string formatted from its arguments
13740 the way @code{message} formats its arguments; @code{\n} produces a new
13743 @item re-search-forward
13744 Search for a pattern, and if the pattern is found, move point to rest
13748 Takes four arguments, like @code{search-forward}:
13752 A regular expression that specifies the pattern to search for.
13753 (Remember to put quotation marks around this argument!)
13756 Optionally, the limit of the search.
13759 Optionally, what to do if the search fails, return @code{nil} or an
13763 Optionally, how many times to repeat the search; if negative, the
13764 search goes backwards.
13768 Bind some variables locally to particular values,
13769 and then evaluate the remaining arguments, returning the value of the
13770 last one. While binding the local variables, use the local values of
13771 variables bound earlier, if any.
13780 (message "`bar' is %d." bar))
13781 @result{} `bar' is 21.
13785 @item match-beginning
13786 Return the position of the start of the text found by the last regular
13790 Return @code{t} for true if the text after point matches the argument,
13791 which should be a regular expression.
13794 Return @code{t} for true if point is at the end of the accessible part
13795 of a buffer. The end of the accessible part is the end of the buffer
13796 if the buffer is not narrowed; it is the end of the narrowed part if
13797 the buffer is narrowed.
13801 @node re-search Exercises, , Regexp Review, Regexp Search
13802 @section Exercises with @code{re-search-forward}
13806 Write a function to search for a regular expression that matches two
13807 or more blank lines in sequence.
13810 Write a function to search for duplicated words, such as `the the'.
13811 @xref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
13812 Manual}, for information on how to write a regexp (a regular
13813 expression) to match a string that is composed of two identical
13814 halves. You can devise several regexps; some are better than others.
13815 The function I use is described in an appendix, along with several
13816 regexps. @xref{the-the, , @code{the-the} Duplicated Words Function}.
13819 @node Counting Words, Words in a defun, Regexp Search, Top
13820 @chapter Counting: Repetition and Regexps
13821 @cindex Repetition for word counting
13822 @cindex Regular expressions for word counting
13824 Repetition and regular expression searches are powerful tools that you
13825 often use when you write code in Emacs Lisp. This chapter illustrates
13826 the use of regular expression searches through the construction of
13827 word count commands using @code{while} loops and recursion.
13830 * Why Count Words::
13831 * count-words-region:: Use a regexp, but find a problem.
13832 * recursive-count-words:: Start with case of no words in region.
13833 * Counting Exercise::
13836 @node Why Count Words, count-words-region, Counting Words, Counting Words
13838 @unnumberedsec Counting words
13841 The standard Emacs distribution contains a function for counting the
13842 number of lines within a region. However, there is no corresponding
13843 function for counting words.
13845 Certain types of writing ask you to count words. Thus, if you write
13846 an essay, you may be limited to 800 words; if you write a novel, you
13847 may discipline yourself to write 1000 words a day. It seems odd to me
13848 that Emacs lacks a word count command. Perhaps people use Emacs
13849 mostly for code or types of documentation that do not require word
13850 counts; or perhaps they restrict themselves to the operating system
13851 word count command, @code{wc}. Alternatively, people may follow
13852 the publishers' convention and compute a word count by dividing the
13853 number of characters in a document by five. In any event, here are
13854 commands to count words.
13856 @node count-words-region, recursive-count-words, Why Count Words, Counting Words
13857 @comment node-name, next, previous, up
13858 @section The @code{count-words-region} Function
13859 @findex count-words-region
13861 A word count command could count words in a line, paragraph, region,
13862 or buffer. What should the command cover? You could design the
13863 command to count the number of words in a complete buffer. However,
13864 the Emacs tradition encourages flexibility---you may want to count
13865 words in just a section, rather than all of a buffer. So it makes
13866 more sense to design the command to count the number of words in a
13867 region. Once you have a @code{count-words-region} command, you can,
13868 if you wish, count words in a whole buffer by marking it with
13869 @w{@kbd{C-x h}} (@code{mark-whole-buffer}).
13871 Clearly, counting words is a repetitive act: starting from the
13872 beginning of the region, you count the first word, then the second
13873 word, then the third word, and so on, until you reach the end of the
13874 region. This means that word counting is ideally suited to recursion
13875 or to a @code{while} loop.
13878 * Design count-words-region:: The definition using a @code{while} loop.
13879 * Whitespace Bug:: The Whitespace Bug in @code{count-words-region}.
13882 @node Design count-words-region, Whitespace Bug, count-words-region, count-words-region
13884 @unnumberedsubsec Designing @code{count-words-region}
13887 First, we will implement the word count command with a @code{while}
13888 loop, then with recursion. The command will, of course, be
13892 The template for an interactive function definition is, as always:
13896 (defun @var{name-of-function} (@var{argument-list})
13897 "@var{documentation}@dots{}"
13898 (@var{interactive-expression}@dots{})
13903 What we need to do is fill in the slots.
13905 The name of the function should be self-explanatory and similar to the
13906 existing @code{count-lines-region} name. This makes the name easier
13907 to remember. @code{count-words-region} is a good choice.
13909 The function counts words within a region. This means that the
13910 argument list must contain symbols that are bound to the two
13911 positions, the beginning and end of the region. These two positions
13912 can be called @samp{beginning} and @samp{end} respectively. The first
13913 line of the documentation should be a single sentence, since that is
13914 all that is printed as documentation by a command such as
13915 @code{apropos}. The interactive expression will be of the form
13916 @samp{(interactive "r")}, since that will cause Emacs to pass the
13917 beginning and end of the region to the function's argument list. All
13920 The body of the function needs to be written to do three tasks:
13921 first, to set up conditions under which the @code{while} loop can
13922 count words, second, to run the @code{while} loop, and third, to send
13923 a message to the user.
13925 When a user calls @code{count-words-region}, point may be at the
13926 beginning or the end of the region. However, the counting process
13927 must start at the beginning of the region. This means we will want
13928 to put point there if it is not already there. Executing
13929 @code{(goto-char beginning)} ensures this. Of course, we will want to
13930 return point to its expected position when the function finishes its
13931 work. For this reason, the body must be enclosed in a
13932 @code{save-excursion} expression.
13934 The central part of the body of the function consists of a
13935 @code{while} loop in which one expression jumps point forward word by
13936 word, and another expression counts those jumps. The true-or-false-test
13937 of the @code{while} loop should test true so long as point should jump
13938 forward, and false when point is at the end of the region.
13940 We could use @code{(forward-word 1)} as the expression for moving point
13941 forward word by word, but it is easier to see what Emacs identifies as a
13942 `word' if we use a regular expression search.
13944 A regular expression search that finds the pattern for which it is
13945 searching leaves point after the last character matched. This means
13946 that a succession of successful word searches will move point forward
13949 As a practical matter, we want the regular expression search to jump
13950 over whitespace and punctuation between words as well as over the
13951 words themselves. A regexp that refuses to jump over interword
13952 whitespace would never jump more than one word! This means that
13953 the regexp should include the whitespace and punctuation that follows
13954 a word, if any, as well as the word itself. (A word may end a buffer
13955 and not have any following whitespace or punctuation, so that part of
13956 the regexp must be optional.)
13958 Thus, what we want for the regexp is a pattern defining one or more
13959 word constituent characters followed, optionally, by one or more
13960 characters that are not word constituents. The regular expression for
13968 The buffer's syntax table determines which characters are and are not
13969 word constituents. (@xref{Syntax, , What Constitutes a Word or
13970 Symbol?}, for more about syntax. Also, see @ref{Syntax, Syntax, The
13971 Syntax Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, ,
13972 Syntax Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
13975 The search expression looks like this:
13978 (re-search-forward "\\w+\\W*")
13982 (Note that paired backslashes precede the @samp{w} and @samp{W}. A
13983 single backslash has special meaning to the Emacs Lisp interpreter.
13984 It indicates that the following character is interpreted differently
13985 than usual. For example, the two characters, @samp{\n}, stand for
13986 @samp{newline}, rather than for a backslash followed by @samp{n}. Two
13987 backslashes in a row stand for an ordinary, `unspecial' backslash, so
13988 Emacs Lisp interpreter ends of seeing a single backslash followed by a
13989 letter. So it discovers the letter is special.)
13991 We need a counter to count how many words there are; this variable
13992 must first be set to 0 and then incremented each time Emacs goes
13993 around the @code{while} loop. The incrementing expression is simply:
13996 (setq count (1+ count))
13999 Finally, we want to tell the user how many words there are in the
14000 region. The @code{message} function is intended for presenting this
14001 kind of information to the user. The message has to be phrased so
14002 that it reads properly regardless of how many words there are in the
14003 region: we don't want to say that ``there are 1 words in the region''.
14004 The conflict between singular and plural is ungrammatical. We can
14005 solve this problem by using a conditional expression that evaluates
14006 different messages depending on the number of words in the region.
14007 There are three possibilities: no words in the region, one word in the
14008 region, and more than one word. This means that the @code{cond}
14009 special form is appropriate.
14012 All this leads to the following function definition:
14016 ;;; @r{First version; has bugs!}
14017 (defun count-words-region (beginning end)
14018 "Print number of words in the region.
14019 Words are defined as at least one word-constituent
14020 character followed by at least one character that
14021 is not a word-constituent. The buffer's syntax
14022 table determines which characters these are."
14024 (message "Counting words in region ... ")
14028 ;;; @r{1. Set up appropriate conditions.}
14030 (goto-char beginning)
14035 ;;; @r{2. Run the} while @r{loop.}
14036 (while (< (point) end)
14037 (re-search-forward "\\w+\\W*")
14038 (setq count (1+ count)))
14042 ;;; @r{3. Send a message to the user.}
14043 (cond ((zerop count)
14045 "The region does NOT have any words."))
14048 "The region has 1 word."))
14051 "The region has %d words." count))))))
14056 As written, the function works, but not in all circumstances.
14058 @node Whitespace Bug, , Design count-words-region, count-words-region
14059 @comment node-name, next, previous, up
14060 @subsection The Whitespace Bug in @code{count-words-region}
14062 The @code{count-words-region} command described in the preceding
14063 section has two bugs, or rather, one bug with two manifestations.
14064 First, if you mark a region containing only whitespace in the middle
14065 of some text, the @code{count-words-region} command tells you that the
14066 region contains one word! Second, if you mark a region containing
14067 only whitespace at the end of the buffer or the accessible portion of
14068 a narrowed buffer, the command displays an error message that looks
14072 Search failed: "\\w+\\W*"
14075 If you are reading this in Info in GNU Emacs, you can test for these
14078 First, evaluate the function in the usual manner to install it.
14080 Here is a copy of the definition. Place your cursor after the closing
14081 parenthesis and type @kbd{C-x C-e} to install it.
14085 ;; @r{First version; has bugs!}
14086 (defun count-words-region (beginning end)
14087 "Print number of words in the region.
14088 Words are defined as at least one word-constituent character followed
14089 by at least one character that is not a word-constituent. The buffer's
14090 syntax table determines which characters these are."
14094 (message "Counting words in region ... ")
14098 ;;; @r{1. Set up appropriate conditions.}
14100 (goto-char beginning)
14105 ;;; @r{2. Run the} while @r{loop.}
14106 (while (< (point) end)
14107 (re-search-forward "\\w+\\W*")
14108 (setq count (1+ count)))
14112 ;;; @r{3. Send a message to the user.}
14113 (cond ((zerop count)
14114 (message "The region does NOT have any words."))
14115 ((= 1 count) (message "The region has 1 word."))
14116 (t (message "The region has %d words." count))))))
14122 If you wish, you can also install this keybinding by evaluating it:
14125 (global-set-key "\C-c=" 'count-words-region)
14128 To conduct the first test, set mark and point to the beginning and end
14129 of the following line and then type @kbd{C-c =} (or @kbd{M-x
14130 count-words-region} if you have not bound @kbd{C-c =}):
14137 Emacs will tell you, correctly, that the region has three words.
14139 Repeat the test, but place mark at the beginning of the line and place
14140 point just @emph{before} the word @samp{one}. Again type the command
14141 @kbd{C-c =} (or @kbd{M-x count-words-region}). Emacs should tell you
14142 that the region has no words, since it is composed only of the
14143 whitespace at the beginning of the line. But instead Emacs tells you
14144 that the region has one word!
14146 For the third test, copy the sample line to the end of the
14147 @file{*scratch*} buffer and then type several spaces at the end of the
14148 line. Place mark right after the word @samp{three} and point at the
14149 end of line. (The end of the line will be the end of the buffer.)
14150 Type @kbd{C-c =} (or @kbd{M-x count-words-region}) as you did before.
14151 Again, Emacs should tell you that the region has no words, since it is
14152 composed only of the whitespace at the end of the line. Instead,
14153 Emacs displays an error message saying @samp{Search failed}.
14155 The two bugs stem from the same problem.
14157 Consider the first manifestation of the bug, in which the command
14158 tells you that the whitespace at the beginning of the line contains
14159 one word. What happens is this: The @code{M-x count-words-region}
14160 command moves point to the beginning of the region. The @code{while}
14161 tests whether the value of point is smaller than the value of
14162 @code{end}, which it is. Consequently, the regular expression search
14163 looks for and finds the first word. It leaves point after the word.
14164 @code{count} is set to one. The @code{while} loop repeats; but this
14165 time the value of point is larger than the value of @code{end}, the
14166 loop is exited; and the function displays a message saying the number
14167 of words in the region is one. In brief, the regular expression
14168 search looks for and finds the word even though it is outside
14171 In the second manifestation of the bug, the region is whitespace at
14172 the end of the buffer. Emacs says @samp{Search failed}. What happens
14173 is that the true-or-false-test in the @code{while} loop tests true, so
14174 the search expression is executed. But since there are no more words
14175 in the buffer, the search fails.
14177 In both manifestations of the bug, the search extends or attempts to
14178 extend outside of the region.
14180 The solution is to limit the search to the region---this is a fairly
14181 simple action, but as you may have come to expect, it is not quite as
14182 simple as you might think.
14184 As we have seen, the @code{re-search-forward} function takes a search
14185 pattern as its first argument. But in addition to this first,
14186 mandatory argument, it accepts three optional arguments. The optional
14187 second argument bounds the search. The optional third argument, if
14188 @code{t}, causes the function to return @code{nil} rather than signal
14189 an error if the search fails. The optional fourth argument is a
14190 repeat count. (In Emacs, you can see a function's documentation by
14191 typing @kbd{C-h f}, the name of the function, and then @key{RET}.)
14193 In the @code{count-words-region} definition, the value of the end of
14194 the region is held by the variable @code{end} which is passed as an
14195 argument to the function. Thus, we can add @code{end} as an argument
14196 to the regular expression search expression:
14199 (re-search-forward "\\w+\\W*" end)
14202 However, if you make only this change to the @code{count-words-region}
14203 definition and then test the new version of the definition on a
14204 stretch of whitespace, you will receive an error message saying
14205 @samp{Search failed}.
14207 What happens is this: the search is limited to the region, and fails
14208 as you expect because there are no word-constituent characters in the
14209 region. Since it fails, we receive an error message. But we do not
14210 want to receive an error message in this case; we want to receive the
14211 message that "The region does NOT have any words."
14213 The solution to this problem is to provide @code{re-search-forward}
14214 with a third argument of @code{t}, which causes the function to return
14215 @code{nil} rather than signal an error if the search fails.
14217 However, if you make this change and try it, you will see the message
14218 ``Counting words in region ... '' and @dots{} you will keep on seeing
14219 that message @dots{}, until you type @kbd{C-g} (@code{keyboard-quit}).
14221 Here is what happens: the search is limited to the region, as before,
14222 and it fails because there are no word-constituent characters in the
14223 region, as expected. Consequently, the @code{re-search-forward}
14224 expression returns @code{nil}. It does nothing else. In particular,
14225 it does not move point, which it does as a side effect if it finds the
14226 search target. After the @code{re-search-forward} expression returns
14227 @code{nil}, the next expression in the @code{while} loop is evaluated.
14228 This expression increments the count. Then the loop repeats. The
14229 true-or-false-test tests true because the value of point is still less
14230 than the value of end, since the @code{re-search-forward} expression
14231 did not move point. @dots{} and the cycle repeats @dots{}
14233 The @code{count-words-region} definition requires yet another
14234 modification, to cause the true-or-false-test of the @code{while} loop
14235 to test false if the search fails. Put another way, there are two
14236 conditions that must be satisfied in the true-or-false-test before the
14237 word count variable is incremented: point must still be within the
14238 region and the search expression must have found a word to count.
14240 Since both the first condition and the second condition must be true
14241 together, the two expressions, the region test and the search
14242 expression, can be joined with an @code{and} special form and embedded in
14243 the @code{while} loop as the true-or-false-test, like this:
14246 (and (< (point) end) (re-search-forward "\\w+\\W*" end t))
14249 @c colon in printed section title causes problem in Info cross reference
14250 @c also trouble with an overfull hbox
14253 (For information about @code{and}, see
14254 @ref{kill-new function, , The @code{kill-new} function}.)
14258 (@xref{kill-new function, , The @code{kill-new} function}, for
14259 information about @code{and}.)
14262 The @code{re-search-forward} expression returns @code{t} if the search
14263 succeeds and as a side effect moves point. Consequently, as words are
14264 found, point is moved through the region. When the search expression
14265 fails to find another word, or when point reaches the end of the
14266 region, the true-or-false-test tests false, the @code{while} loop
14267 exits, and the @code{count-words-region} function displays one or
14268 other of its messages.
14270 After incorporating these final changes, the @code{count-words-region}
14271 works without bugs (or at least, without bugs that I have found!).
14272 Here is what it looks like:
14276 ;;; @r{Final version:} @code{while}
14277 (defun count-words-region (beginning end)
14278 "Print number of words in the region."
14280 (message "Counting words in region ... ")
14284 ;;; @r{1. Set up appropriate conditions.}
14287 (goto-char beginning)
14291 ;;; @r{2. Run the} while @r{loop.}
14292 (while (and (< (point) end)
14293 (re-search-forward "\\w+\\W*" end t))
14294 (setq count (1+ count)))
14298 ;;; @r{3. Send a message to the user.}
14299 (cond ((zerop count)
14301 "The region does NOT have any words."))
14304 "The region has 1 word."))
14307 "The region has %d words." count))))))
14311 @node recursive-count-words, Counting Exercise, count-words-region, Counting Words
14312 @comment node-name, next, previous, up
14313 @section Count Words Recursively
14314 @cindex Count words recursively
14315 @cindex Recursively counting words
14316 @cindex Words, counted recursively
14318 You can write the function for counting words recursively as well as
14319 with a @code{while} loop. Let's see how this is done.
14321 First, we need to recognize that the @code{count-words-region}
14322 function has three jobs: it sets up the appropriate conditions for
14323 counting to occur; it counts the words in the region; and it sends a
14324 message to the user telling how many words there are.
14326 If we write a single recursive function to do everything, we will
14327 receive a message for every recursive call. If the region contains 13
14328 words, we will receive thirteen messages, one right after the other.
14329 We don't want this! Instead, we must write two functions to do the
14330 job, one of which (the recursive function) will be used inside of the
14331 other. One function will set up the conditions and display the
14332 message; the other will return the word count.
14334 Let us start with the function that causes the message to be displayed.
14335 We can continue to call this @code{count-words-region}.
14337 This is the function that the user will call. It will be interactive.
14338 Indeed, it will be similar to our previous versions of this
14339 function, except that it will call @code{recursive-count-words} to
14340 determine how many words are in the region.
14343 We can readily construct a template for this function, based on our
14348 ;; @r{Recursive version; uses regular expression search}
14349 (defun count-words-region (beginning end)
14350 "@var{documentation}@dots{}"
14351 (@var{interactive-expression}@dots{})
14355 ;;; @r{1. Set up appropriate conditions.}
14356 (@var{explanatory message})
14357 (@var{set-up functions}@dots{}
14361 ;;; @r{2. Count the words.}
14362 @var{recursive call}
14366 ;;; @r{3. Send a message to the user.}
14367 @var{message providing word count}))
14371 The definition looks straightforward, except that somehow the count
14372 returned by the recursive call must be passed to the message
14373 displaying the word count. A little thought suggests that this can be
14374 done by making use of a @code{let} expression: we can bind a variable
14375 in the varlist of a @code{let} expression to the number of words in
14376 the region, as returned by the recursive call; and then the
14377 @code{cond} expression, using binding, can display the value to the
14380 Often, one thinks of the binding within a @code{let} expression as
14381 somehow secondary to the `primary' work of a function. But in this
14382 case, what you might consider the `primary' job of the function,
14383 counting words, is done within the @code{let} expression.
14386 Using @code{let}, the function definition looks like this:
14390 (defun count-words-region (beginning end)
14391 "Print number of words in the region."
14396 ;;; @r{1. Set up appropriate conditions.}
14397 (message "Counting words in region ... ")
14399 (goto-char beginning)
14403 ;;; @r{2. Count the words.}
14404 (let ((count (recursive-count-words end)))
14408 ;;; @r{3. Send a message to the user.}
14409 (cond ((zerop count)
14411 "The region does NOT have any words."))
14414 "The region has 1 word."))
14417 "The region has %d words." count))))))
14421 Next, we need to write the recursive counting function.
14423 A recursive function has at least three parts: the `do-again-test', the
14424 `next-step-expression', and the recursive call.
14426 The do-again-test determines whether the function will or will not be
14427 called again. Since we are counting words in a region and can use a
14428 function that moves point forward for every word, the do-again-test
14429 can check whether point is still within the region. The do-again-test
14430 should find the value of point and determine whether point is before,
14431 at, or after the value of the end of the region. We can use the
14432 @code{point} function to locate point. Clearly, we must pass the
14433 value of the end of the region to the recursive counting function as an
14436 In addition, the do-again-test should also test whether the search finds a
14437 word. If it does not, the function should not call itself again.
14439 The next-step-expression changes a value so that when the recursive
14440 function is supposed to stop calling itself, it stops. More
14441 precisely, the next-step-expression changes a value so that at the
14442 right time, the do-again-test stops the recursive function from
14443 calling itself again. In this case, the next-step-expression can be
14444 the expression that moves point forward, word by word.
14446 The third part of a recursive function is the recursive call.
14448 Somewhere, also, we also need a part that does the `work' of the
14449 function, a part that does the counting. A vital part!
14452 But already, we have an outline of the recursive counting function:
14456 (defun recursive-count-words (region-end)
14457 "@var{documentation}@dots{}"
14458 @var{do-again-test}
14459 @var{next-step-expression}
14460 @var{recursive call})
14464 Now we need to fill in the slots. Let's start with the simplest cases
14465 first: if point is at or beyond the end of the region, there cannot
14466 be any words in the region, so the function should return zero.
14467 Likewise, if the search fails, there are no words to count, so the
14468 function should return zero.
14470 On the other hand, if point is within the region and the search
14471 succeeds, the function should call itself again.
14474 Thus, the do-again-test should look like this:
14478 (and (< (point) region-end)
14479 (re-search-forward "\\w+\\W*" region-end t))
14483 Note that the search expression is part of the do-again-test---the
14484 function returns @code{t} if its search succeeds and @code{nil} if it
14485 fails. (@xref{Whitespace Bug, , The Whitespace Bug in
14486 @code{count-words-region}}, for an explanation of how
14487 @code{re-search-forward} works.)
14489 The do-again-test is the true-or-false test of an @code{if} clause.
14490 Clearly, if the do-again-test succeeds, the then-part of the @code{if}
14491 clause should call the function again; but if it fails, the else-part
14492 should return zero since either point is outside the region or the
14493 search failed because there were no words to find.
14495 But before considering the recursive call, we need to consider the
14496 next-step-expression. What is it? Interestingly, it is the search
14497 part of the do-again-test.
14499 In addition to returning @code{t} or @code{nil} for the
14500 do-again-test, @code{re-search-forward} moves point forward as a side
14501 effect of a successful search. This is the action that changes the
14502 value of point so that the recursive function stops calling itself
14503 when point completes its movement through the region. Consequently,
14504 the @code{re-search-forward} expression is the next-step-expression.
14507 In outline, then, the body of the @code{recursive-count-words}
14508 function looks like this:
14512 (if @var{do-again-test-and-next-step-combined}
14514 @var{recursive-call-returning-count}
14520 How to incorporate the mechanism that counts?
14522 If you are not used to writing recursive functions, a question like
14523 this can be troublesome. But it can and should be approached
14526 We know that the counting mechanism should be associated in some way
14527 with the recursive call. Indeed, since the next-step-expression moves
14528 point forward by one word, and since a recursive call is made for
14529 each word, the counting mechanism must be an expression that adds one
14530 to the value returned by a call to @code{recursive-count-words}.
14533 Consider several cases:
14537 If there are two words in the region, the function should return
14538 a value resulting from adding one to the value returned when it counts
14539 the first word, plus the number returned when it counts the remaining
14540 words in the region, which in this case is one.
14543 If there is one word in the region, the function should return
14544 a value resulting from adding one to the value returned when it counts
14545 that word, plus the number returned when it counts the remaining
14546 words in the region, which in this case is zero.
14549 If there are no words in the region, the function should return zero.
14552 From the sketch we can see that the else-part of the @code{if} returns
14553 zero for the case of no words. This means that the then-part of the
14554 @code{if} must return a value resulting from adding one to the value
14555 returned from a count of the remaining words.
14558 The expression will look like this, where @code{1+} is a function that
14559 adds one to its argument.
14562 (1+ (recursive-count-words region-end))
14566 The whole @code{recursive-count-words} function will then look like
14571 (defun recursive-count-words (region-end)
14572 "@var{documentation}@dots{}"
14574 ;;; @r{1. do-again-test}
14575 (if (and (< (point) region-end)
14576 (re-search-forward "\\w+\\W*" region-end t))
14580 ;;; @r{2. then-part: the recursive call}
14581 (1+ (recursive-count-words region-end))
14583 ;;; @r{3. else-part}
14589 Let's examine how this works:
14591 If there are no words in the region, the else part of the @code{if}
14592 expression is evaluated and consequently the function returns zero.
14594 If there is one word in the region, the value of point is less than
14595 the value of @code{region-end} and the search succeeds. In this case,
14596 the true-or-false-test of the @code{if} expression tests true, and the
14597 then-part of the @code{if} expression is evaluated. The counting
14598 expression is evaluated. This expression returns a value (which will
14599 be the value returned by the whole function) that is the sum of one
14600 added to the value returned by a recursive call.
14602 Meanwhile, the next-step-expression has caused point to jump over the
14603 first (and in this case only) word in the region. This means that
14604 when @code{(recursive-count-words region-end)} is evaluated a second
14605 time, as a result of the recursive call, the value of point will be
14606 equal to or greater than the value of region end. So this time,
14607 @code{recursive-count-words} will return zero. The zero will be added
14608 to one, and the original evaluation of @code{recursive-count-words}
14609 will return one plus zero, which is one, which is the correct amount.
14611 Clearly, if there are two words in the region, the first call to
14612 @code{recursive-count-words} returns one added to the value returned
14613 by calling @code{recursive-count-words} on a region containing the
14614 remaining word---that is, it adds one to one, producing two, which is
14615 the correct amount.
14617 Similarly, if there are three words in the region, the first call to
14618 @code{recursive-count-words} returns one added to the value returned
14619 by calling @code{recursive-count-words} on a region containing the
14620 remaining two words---and so on and so on.
14624 With full documentation the two functions look like this:
14628 The recursive function:
14630 @findex recursive-count-words
14633 (defun recursive-count-words (region-end)
14634 "Number of words between point and REGION-END."
14638 ;;; @r{1. do-again-test}
14639 (if (and (< (point) region-end)
14640 (re-search-forward "\\w+\\W*" region-end t))
14644 ;;; @r{2. then-part: the recursive call}
14645 (1+ (recursive-count-words region-end))
14647 ;;; @r{3. else-part}
14658 ;;; @r{Recursive version}
14659 (defun count-words-region (beginning end)
14660 "Print number of words in the region.
14664 Words are defined as at least one word-constituent
14665 character followed by at least one character that is
14666 not a word-constituent. The buffer's syntax table
14667 determines which characters these are."
14671 (message "Counting words in region ... ")
14673 (goto-char beginning)
14674 (let ((count (recursive-count-words end)))
14677 (cond ((zerop count)
14679 "The region does NOT have any words."))
14683 (message "The region has 1 word."))
14686 "The region has %d words." count))))))
14690 @node Counting Exercise, , recursive-count-words, Counting Words
14691 @section Exercise: Counting Punctuation
14693 Using a @code{while} loop, write a function to count the number of
14694 punctuation marks in a region---period, comma, semicolon, colon,
14695 exclamation mark, and question mark. Do the same using recursion.
14697 @node Words in a defun, Readying a Graph, Counting Words, Top
14698 @chapter Counting Words in a @code{defun}
14699 @cindex Counting words in a @code{defun}
14700 @cindex Word counting in a @code{defun}
14702 Our next project is to count the number of words in a function
14703 definition. Clearly, this can be done using some variant of
14704 @code{count-word-region}. @xref{Counting Words, , Counting Words:
14705 Repetition and Regexps}. If we are just going to count the words in
14706 one definition, it is easy enough to mark the definition with the
14707 @kbd{C-M-h} (@code{mark-defun}) command, and then call
14708 @code{count-word-region}.
14710 However, I am more ambitious: I want to count the words and symbols in
14711 every definition in the Emacs sources and then print a graph that
14712 shows how many functions there are of each length: how many contain 40
14713 to 49 words or symbols, how many contain 50 to 59 words or symbols,
14714 and so on. I have often been curious how long a typical function is,
14715 and this will tell.
14718 * Divide and Conquer::
14719 * Words and Symbols:: What to count?
14720 * Syntax:: What constitutes a word or symbol?
14721 * count-words-in-defun:: Very like @code{count-words}.
14722 * Several defuns:: Counting several defuns in a file.
14723 * Find a File:: Do you want to look at a file?
14724 * lengths-list-file:: A list of the lengths of many definitions.
14725 * Several files:: Counting in definitions in different files.
14726 * Several files recursively:: Recursively counting in different files.
14727 * Prepare the data:: Prepare the data for display in a graph.
14730 @node Divide and Conquer, Words and Symbols, Words in a defun, Words in a defun
14732 @unnumberedsec Divide and Conquer
14735 Described in one phrase, the histogram project is daunting; but
14736 divided into numerous small steps, each of which we can take one at a
14737 time, the project becomes less fearsome. Let us consider what the
14742 First, write a function to count the words in one definition. This
14743 includes the problem of handling symbols as well as words.
14746 Second, write a function to list the numbers of words in each function
14747 in a file. This function can use the @code{count-words-in-defun}
14751 Third, write a function to list the numbers of words in each function
14752 in each of several files. This entails automatically finding the
14753 various files, switching to them, and counting the words in the
14754 definitions within them.
14757 Fourth, write a function to convert the list of numbers that we
14758 created in step three to a form that will be suitable for printing as
14762 Fifth, write a function to print the results as a graph.
14765 This is quite a project! But if we take each step slowly, it will not
14768 @node Words and Symbols, Syntax, Divide and Conquer, Words in a defun
14769 @section What to Count?
14770 @cindex Words and symbols in defun
14772 When we first start thinking about how to count the words in a
14773 function definition, the first question is (or ought to be) what are
14774 we going to count? When we speak of `words' with respect to a Lisp
14775 function definition, we are actually speaking, in large part, of
14776 `symbols'. For example, the following @code{multiply-by-seven}
14777 function contains the five symbols @code{defun},
14778 @code{multiply-by-seven}, @code{number}, @code{*}, and @code{7}. In
14779 addition, in the documentation string, it contains the four words
14780 @samp{Multiply}, @samp{NUMBER}, @samp{by}, and @samp{seven}. The
14781 symbol @samp{number} is repeated, so the definition contains a total
14782 of ten words and symbols.
14786 (defun multiply-by-seven (number)
14787 "Multiply NUMBER by seven."
14793 However, if we mark the @code{multiply-by-seven} definition with
14794 @kbd{C-M-h} (@code{mark-defun}), and then call
14795 @code{count-words-region} on it, we will find that
14796 @code{count-words-region} claims the definition has eleven words, not
14797 ten! Something is wrong!
14799 The problem is twofold: @code{count-words-region} does not count the
14800 @samp{*} as a word, and it counts the single symbol,
14801 @code{multiply-by-seven}, as containing three words. The hyphens are
14802 treated as if they were interword spaces rather than intraword
14803 connectors: @samp{multiply-by-seven} is counted as if it were written
14804 @samp{multiply by seven}.
14806 The cause of this confusion is the regular expression search within
14807 the @code{count-words-region} definition that moves point forward word
14808 by word. In the canonical version of @code{count-words-region}, the
14816 This regular expression is a pattern defining one or more word
14817 constituent characters possibly followed by one or more characters
14818 that are not word constituents. What is meant by `word constituent
14819 characters' brings us to the issue of syntax, which is worth a section
14822 @node Syntax, count-words-in-defun, Words and Symbols, Words in a defun
14823 @section What Constitutes a Word or Symbol?
14824 @cindex Syntax categories and tables
14826 Emacs treats different characters as belonging to different
14827 @dfn{syntax categories}. For example, the regular expression,
14828 @samp{\\w+}, is a pattern specifying one or more @emph{word
14829 constituent} characters. Word constituent characters are members of
14830 one syntax category. Other syntax categories include the class of
14831 punctuation characters, such as the period and the comma, and the
14832 class of whitespace characters, such as the blank space and the tab
14833 character. (For more information, see @ref{Syntax, Syntax, The Syntax
14834 Table, emacs, The GNU Emacs Manual}, and @ref{Syntax Tables, , Syntax
14835 Tables, elisp, The GNU Emacs Lisp Reference Manual}.)
14837 Syntax tables specify which characters belong to which categories.
14838 Usually, a hyphen is not specified as a `word constituent character'.
14839 Instead, it is specified as being in the `class of characters that are
14840 part of symbol names but not words.' This means that the
14841 @code{count-words-region} function treats it in the same way it treats
14842 an interword white space, which is why @code{count-words-region}
14843 counts @samp{multiply-by-seven} as three words.
14845 There are two ways to cause Emacs to count @samp{multiply-by-seven} as
14846 one symbol: modify the syntax table or modify the regular expression.
14848 We could redefine a hyphen as a word constituent character by
14849 modifying the syntax table that Emacs keeps for each mode. This
14850 action would serve our purpose, except that a hyphen is merely the
14851 most common character within symbols that is not typically a word
14852 constituent character; there are others, too.
14854 Alternatively, we can redefine the regular expression used in the
14855 @code{count-words} definition so as to include symbols. This
14856 procedure has the merit of clarity, but the task is a little tricky.
14859 The first part is simple enough: the pattern must match ``at least one
14860 character that is a word or symbol constituent''. Thus:
14863 "\\(\\w\\|\\s_\\)+"
14867 The @samp{\\(} is the first part of the grouping construct that
14868 includes the @samp{\\w} and the @samp{\\s_} as alternatives, separated
14869 by the @samp{\\|}. The @samp{\\w} matches any word-constituent
14870 character and the @samp{\\s_} matches any character that is part of a
14871 symbol name but not a word-constituent character. The @samp{+}
14872 following the group indicates that the word or symbol constituent
14873 characters must be matched at least once.
14875 However, the second part of the regexp is more difficult to design.
14876 What we want is to follow the first part with ``optionally one or more
14877 characters that are not constituents of a word or symbol''. At first,
14878 I thought I could define this with the following:
14881 "\\(\\W\\|\\S_\\)*"
14885 The upper case @samp{W} and @samp{S} match characters that are
14886 @emph{not} word or symbol constituents. Unfortunately, this
14887 expression matches any character that is either not a word constituent
14888 or not a symbol constituent. This matches any character!
14890 I then noticed that every word or symbol in my test region was
14891 followed by white space (blank space, tab, or newline). So I tried
14892 placing a pattern to match one or more blank spaces after the pattern
14893 for one or more word or symbol constituents. This failed, too. Words
14894 and symbols are often separated by whitespace, but in actual code
14895 parentheses may follow symbols and punctuation may follow words. So
14896 finally, I designed a pattern in which the word or symbol constituents
14897 are followed optionally by characters that are not white space and
14898 then followed optionally by white space.
14901 Here is the full regular expression:
14904 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
14907 @node count-words-in-defun, Several defuns, Syntax, Words in a defun
14908 @section The @code{count-words-in-defun} Function
14909 @cindex Counting words in a @code{defun}
14911 We have seen that there are several ways to write a
14912 @code{count-word-region} function. To write a
14913 @code{count-words-in-defun}, we need merely adapt one of these
14916 The version that uses a @code{while} loop is easy to understand, so I
14917 am going to adapt that. Because @code{count-words-in-defun} will be
14918 part of a more complex program, it need not be interactive and it need
14919 not display a message but just return the count. These considerations
14920 simplify the definition a little.
14922 On the other hand, @code{count-words-in-defun} will be used within a
14923 buffer that contains function definitions. Consequently, it is
14924 reasonable to ask that the function determine whether it is called
14925 when point is within a function definition, and if it is, to return
14926 the count for that definition. This adds complexity to the
14927 definition, but saves us from needing to pass arguments to the
14931 These considerations lead us to prepare the following template:
14935 (defun count-words-in-defun ()
14936 "@var{documentation}@dots{}"
14937 (@var{set up}@dots{}
14938 (@var{while loop}@dots{})
14939 @var{return count})
14944 As usual, our job is to fill in the slots.
14948 We are presuming that this function will be called within a buffer
14949 containing function definitions. Point will either be within a
14950 function definition or not. For @code{count-words-in-defun} to work,
14951 point must move to the beginning of the definition, a counter must
14952 start at zero, and the counting loop must stop when point reaches the
14953 end of the definition.
14955 The @code{beginning-of-defun} function searches backwards for an
14956 opening delimiter such as a @samp{(} at the beginning of a line, and
14957 moves point to that position, or else to the limit of the search. In
14958 practice, this means that @code{beginning-of-defun} moves point to the
14959 beginning of an enclosing or preceding function definition, or else to
14960 the beginning of the buffer. We can use @code{beginning-of-defun} to
14961 place point where we wish to start.
14963 The @code{while} loop requires a counter to keep track of the words or
14964 symbols being counted. A @code{let} expression can be used to create
14965 a local variable for this purpose, and bind it to an initial value of zero.
14967 The @code{end-of-defun} function works like @code{beginning-of-defun}
14968 except that it moves point to the end of the definition.
14969 @code{end-of-defun} can be used as part of an expression that
14970 determines the position of the end of the definition.
14972 The set up for @code{count-words-in-defun} takes shape rapidly: first
14973 we move point to the beginning of the definition, then we create a
14974 local variable to hold the count, and finally, we record the position
14975 of the end of the definition so the @code{while} loop will know when to stop
14979 The code looks like this:
14983 (beginning-of-defun)
14985 (end (save-excursion (end-of-defun) (point))))
14990 The code is simple. The only slight complication is likely to concern
14991 @code{end}: it is bound to the position of the end of the definition
14992 by a @code{save-excursion} expression that returns the value of point
14993 after @code{end-of-defun} temporarily moves it to the end of the
14996 The second part of the @code{count-words-in-defun}, after the set up,
14997 is the @code{while} loop.
14999 The loop must contain an expression that jumps point forward word by
15000 word and symbol by symbol, and another expression that counts the
15001 jumps. The true-or-false-test for the @code{while} loop should test
15002 true so long as point should jump forward, and false when point is at
15003 the end of the definition. We have already redefined the regular
15004 expression for this (@pxref{Syntax}), so the loop is straightforward:
15008 (while (and (< (point) end)
15010 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*" end t)
15011 (setq count (1+ count)))
15015 The third part of the function definition returns the count of words
15016 and symbols. This part is the last expression within the body of the
15017 @code{let} expression, and can be, very simply, the local variable
15018 @code{count}, which when evaluated returns the count.
15021 Put together, the @code{count-words-in-defun} definition looks like this:
15023 @findex count-words-in-defun
15026 (defun count-words-in-defun ()
15027 "Return the number of words and symbols in a defun."
15028 (beginning-of-defun)
15030 (end (save-excursion (end-of-defun) (point))))
15034 (and (< (point) end)
15036 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
15038 (setq count (1+ count)))
15043 How to test this? The function is not interactive, but it is easy to
15044 put a wrapper around the function to make it interactive; we can use
15045 almost the same code as for the recursive version of
15046 @code{count-words-region}:
15050 ;;; @r{Interactive version.}
15051 (defun count-words-defun ()
15052 "Number of words and symbols in a function definition."
15055 "Counting words and symbols in function definition ... ")
15058 (let ((count (count-words-in-defun)))
15062 "The definition does NOT have any words or symbols."))
15067 "The definition has 1 word or symbol."))
15070 "The definition has %d words or symbols." count)))))
15076 Let's re-use @kbd{C-c =} as a convenient keybinding:
15079 (global-set-key "\C-c=" 'count-words-defun)
15082 Now we can try out @code{count-words-defun}: install both
15083 @code{count-words-in-defun} and @code{count-words-defun}, and set the
15084 keybinding, and then place the cursor within the following definition:
15088 (defun multiply-by-seven (number)
15089 "Multiply NUMBER by seven."
15096 Success! The definition has 10 words and symbols.
15098 The next problem is to count the numbers of words and symbols in
15099 several definitions within a single file.
15101 @node Several defuns, Find a File, count-words-in-defun, Words in a defun
15102 @section Count Several @code{defuns} Within a File
15104 A file such as @file{simple.el} may have a hundred or more function
15105 definitions within it. Our long term goal is to collect statistics on
15106 many files, but as a first step, our immediate goal is to collect
15107 statistics on one file.
15109 The information will be a series of numbers, each number being the
15110 length of a function definition. We can store the numbers in a list.
15112 We know that we will want to incorporate the information regarding one
15113 file with information about many other files; this means that the
15114 function for counting definition lengths within one file need only
15115 return the list of lengths. It need not and should not display any
15118 The word count commands contain one expression to jump point forward
15119 word by word and another expression to count the jumps. The function
15120 to return the lengths of definitions can be designed to work the same
15121 way, with one expression to jump point forward definition by
15122 definition and another expression to construct the lengths' list.
15124 This statement of the problem makes it elementary to write the
15125 function definition. Clearly, we will start the count at the
15126 beginning of the file, so the first command will be @code{(goto-char
15127 (point-min))}. Next, we start the @code{while} loop; and the
15128 true-or-false test of the loop can be a regular expression search for
15129 the next function definition---so long as the search succeeds, point
15130 is moved forward and then the body of the loop is evaluated. The body
15131 needs an expression that constructs the lengths' list. @code{cons},
15132 the list construction command, can be used to create the list. That
15133 is almost all there is to it.
15136 Here is what this fragment of code looks like:
15140 (goto-char (point-min))
15141 (while (re-search-forward "^(defun" nil t)
15143 (cons (count-words-in-defun) lengths-list)))
15147 What we have left out is the mechanism for finding the file that
15148 contains the function definitions.
15150 In previous examples, we either used this, the Info file, or we
15151 switched back and forth to some other buffer, such as the
15152 @file{*scratch*} buffer.
15154 Finding a file is a new process that we have not yet discussed.
15156 @node Find a File, lengths-list-file, Several defuns, Words in a defun
15157 @comment node-name, next, previous, up
15158 @section Find a File
15159 @cindex Find a File
15161 To find a file in Emacs, you use the @kbd{C-x C-f} (@code{find-file})
15162 command. This command is almost, but not quite right for the lengths
15166 Let's look at the source for @code{find-file}:
15170 (defun find-file (filename)
15171 "Edit file FILENAME.
15172 Switch to a buffer visiting file FILENAME,
15173 creating one if none already exists."
15174 (interactive "FFind file: ")
15175 (switch-to-buffer (find-file-noselect filename)))
15180 (The most recent version of the @code{find-file} function definition
15181 permits you to specify optional wildcards to visit multiple files; that
15182 makes the definition more complex and we will not discuss it here,
15183 since it is not relevant. You can see its source using either
15184 @kbd{M-.} (@code{find-tag}) or @kbd{C-h f} (@code{describe-function}).)
15188 (defun find-file (filename &optional wildcards)
15189 "Edit file FILENAME.
15190 Switch to a buffer visiting file FILENAME,
15191 creating one if none already exists.
15192 Interactively, the default if you just type RET is the current directory,
15193 but the visited file name is available through the minibuffer history:
15194 type M-n to pull it into the minibuffer.
15196 Interactively, or if WILDCARDS is non-nil in a call from Lisp,
15197 expand wildcards (if any) and visit multiple files. You can
15198 suppress wildcard expansion by setting `find-file-wildcards' to nil.
15200 To visit a file without any kind of conversion and without
15201 automatically choosing a major mode, use \\[find-file-literally]."
15202 (interactive (find-file-read-args "Find file: " nil))
15203 (let ((value (find-file-noselect filename nil nil wildcards)))
15205 (mapcar 'switch-to-buffer (nreverse value))
15206 (switch-to-buffer value))))
15209 The definition I am showing possesses short but complete documentation
15210 and an interactive specification that prompts you for a file name when
15211 you use the command interactively. The body of the definition
15212 contains two functions, @code{find-file-noselect} and
15213 @code{switch-to-buffer}.
15215 According to its documentation as shown by @kbd{C-h f} (the
15216 @code{describe-function} command), the @code{find-file-noselect}
15217 function reads the named file into a buffer and returns the buffer.
15218 (Its most recent version includes an optional wildcards argument,
15219 too, as well as another to read a file literally and an other you
15220 suppress warning messages. These optional arguments are irrelevant.)
15222 However, the @code{find-file-noselect} function does not select the
15223 buffer in which it puts the file. Emacs does not switch its attention
15224 (or yours if you are using @code{find-file-noselect}) to the selected
15225 buffer. That is what @code{switch-to-buffer} does: it switches the
15226 buffer to which Emacs attention is directed; and it switches the
15227 buffer displayed in the window to the new buffer. We have discussed
15228 buffer switching elsewhere. (@xref{Switching Buffers}.)
15230 In this histogram project, we do not need to display each file on the
15231 screen as the program determines the length of each definition within
15232 it. Instead of employing @code{switch-to-buffer}, we can work with
15233 @code{set-buffer}, which redirects the attention of the computer
15234 program to a different buffer but does not redisplay it on the screen.
15235 So instead of calling on @code{find-file} to do the job, we must write
15236 our own expression.
15238 The task is easy: use @code{find-file-noselect} and @code{set-buffer}.
15240 @node lengths-list-file, Several files, Find a File, Words in a defun
15241 @section @code{lengths-list-file} in Detail
15243 The core of the @code{lengths-list-file} function is a @code{while}
15244 loop containing a function to move point forward `defun by defun' and
15245 a function to count the number of words and symbols in each defun.
15246 This core must be surrounded by functions that do various other tasks,
15247 including finding the file, and ensuring that point starts out at the
15248 beginning of the file. The function definition looks like this:
15249 @findex lengths-list-file
15253 (defun lengths-list-file (filename)
15254 "Return list of definitions' lengths within FILE.
15255 The returned list is a list of numbers.
15256 Each number is the number of words or
15257 symbols in one function definition."
15260 (message "Working on `%s' ... " filename)
15262 (let ((buffer (find-file-noselect filename))
15264 (set-buffer buffer)
15265 (setq buffer-read-only t)
15267 (goto-char (point-min))
15268 (while (re-search-forward "^(defun" nil t)
15270 (cons (count-words-in-defun) lengths-list)))
15271 (kill-buffer buffer)
15277 The function is passed one argument, the name of the file on which it
15278 will work. It has four lines of documentation, but no interactive
15279 specification. Since people worry that a computer is broken if they
15280 don't see anything going on, the first line of the body is a
15283 The next line contains a @code{save-excursion} that returns Emacs'
15284 attention to the current buffer when the function completes. This is
15285 useful in case you embed this function in another function that
15286 presumes point is restored to the original buffer.
15288 In the varlist of the @code{let} expression, Emacs finds the file and
15289 binds the local variable @code{buffer} to the buffer containing the
15290 file. At the same time, Emacs creates @code{lengths-list} as a local
15293 Next, Emacs switches its attention to the buffer.
15295 In the following line, Emacs makes the buffer read-only. Ideally,
15296 this line is not necessary. None of the functions for counting words
15297 and symbols in a function definition should change the buffer.
15298 Besides, the buffer is not going to be saved, even if it were changed.
15299 This line is entirely the consequence of great, perhaps excessive,
15300 caution. The reason for the caution is that this function and those
15301 it calls work on the sources for Emacs and it is inconvenient if they
15302 are inadvertently modified. It goes without saying that I did not
15303 realize a need for this line until an experiment went awry and started
15304 to modify my Emacs source files @dots{}
15306 Next comes a call to widen the buffer if it is narrowed. This
15307 function is usually not needed---Emacs creates a fresh buffer if none
15308 already exists; but if a buffer visiting the file already exists Emacs
15309 returns that one. In this case, the buffer may be narrowed and must
15310 be widened. If we wanted to be fully `user-friendly', we would
15311 arrange to save the restriction and the location of point, but we
15314 The @code{(goto-char (point-min))} expression moves point to the
15315 beginning of the buffer.
15317 Then comes a @code{while} loop in which the `work' of the function is
15318 carried out. In the loop, Emacs determines the length of each
15319 definition and constructs a lengths' list containing the information.
15321 Emacs kills the buffer after working through it. This is to save
15322 space inside of Emacs. My version of GNU Emacs 19 contained over 300
15323 source files of interest; GNU Emacs 22 contains over a thousand source
15324 files. Another function will apply @code{lengths-list-file} to each
15327 Finally, the last expression within the @code{let} expression is the
15328 @code{lengths-list} variable; its value is returned as the value of
15329 the whole function.
15331 You can try this function by installing it in the usual fashion. Then
15332 place your cursor after the following expression and type @kbd{C-x
15333 C-e} (@code{eval-last-sexp}).
15335 @c !!! 22.1.1 lisp sources location here
15338 "/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el")
15342 (You may need to change the pathname of the file; the one here is for
15343 GNU Emacs version 22.1.1. To change the expression, copy it to
15344 the @file{*scratch*} buffer and edit it.
15348 (Also, to see the full length of the list, rather than a truncated
15349 version, you may have to evaluate the following:
15352 (custom-set-variables '(eval-expression-print-length nil))
15356 (@xref{defcustom, , Specifying Variables using @code{defcustom}}.
15357 Then evaluate the @code{lengths-list-file} expression.)
15360 The lengths' list for @file{debug.el} takes less than a second to
15361 produce and looks like this in GNU Emacs 22:
15364 (83 113 105 144 289 22 30 97 48 89 25 52 52 88 28 29 77 49 43 290 232 587)
15368 (Using my old machine, the version 19 lengths' list for @file{debug.el}
15369 took seven seconds to produce and looked like this:
15372 (75 41 80 62 20 45 44 68 45 12 34 235)
15375 (The newer version of @file{debug.el} contains more defuns than the
15376 earlier one; and my new machine is much faster than the old one.)
15378 Note that the length of the last definition in the file is first in
15381 @node Several files, Several files recursively, lengths-list-file, Words in a defun
15382 @section Count Words in @code{defuns} in Different Files
15384 In the previous section, we created a function that returns a list of
15385 the lengths of each definition in a file. Now, we want to define a
15386 function to return a master list of the lengths of the definitions in
15389 Working on each of a list of files is a repetitious act, so we can use
15390 either a @code{while} loop or recursion.
15393 * lengths-list-many-files:: Return a list of the lengths of defuns.
15394 * append:: Attach one list to another.
15397 @node lengths-list-many-files, append, Several files, Several files
15399 @unnumberedsubsec Determine the lengths of @code{defuns}
15402 The design using a @code{while} loop is routine. The argument passed
15403 the function is a list of files. As we saw earlier (@pxref{Loop
15404 Example}), you can write a @code{while} loop so that the body of the
15405 loop is evaluated if such a list contains elements, but to exit the
15406 loop if the list is empty. For this design to work, the body of the
15407 loop must contain an expression that shortens the list each time the
15408 body is evaluated, so that eventually the list is empty. The usual
15409 technique is to set the value of the list to the value of the @sc{cdr}
15410 of the list each time the body is evaluated.
15413 The template looks like this:
15417 (while @var{test-whether-list-is-empty}
15419 @var{set-list-to-cdr-of-list})
15423 Also, we remember that a @code{while} loop returns @code{nil} (the
15424 result of evaluating the true-or-false-test), not the result of any
15425 evaluation within its body. (The evaluations within the body of the
15426 loop are done for their side effects.) However, the expression that
15427 sets the lengths' list is part of the body---and that is the value
15428 that we want returned by the function as a whole. To do this, we
15429 enclose the @code{while} loop within a @code{let} expression, and
15430 arrange that the last element of the @code{let} expression contains
15431 the value of the lengths' list. (@xref{Incrementing Example, , Loop
15432 Example with an Incrementing Counter}.)
15434 @findex lengths-list-many-files
15436 These considerations lead us directly to the function itself:
15440 ;;; @r{Use @code{while} loop.}
15441 (defun lengths-list-many-files (list-of-files)
15442 "Return list of lengths of defuns in LIST-OF-FILES."
15445 (let (lengths-list)
15447 ;;; @r{true-or-false-test}
15448 (while list-of-files
15453 ;;; @r{Generate a lengths' list.}
15455 (expand-file-name (car list-of-files)))))
15459 ;;; @r{Make files' list shorter.}
15460 (setq list-of-files (cdr list-of-files)))
15462 ;;; @r{Return final value of lengths' list.}
15467 @code{expand-file-name} is a built-in function that converts a file
15468 name to the absolute, long, path name form. The function employs the
15469 name of the directory in which the function is called.
15471 @c !!! 22.1.1 lisp sources location here
15473 Thus, if @code{expand-file-name} is called on @code{debug.el} when
15474 Emacs is visiting the
15475 @file{/usr/local/share/emacs/22.1.1/lisp/emacs-lisp/} directory,
15485 @c !!! 22.1.1 lisp sources location here
15487 /usr/local/share/emacs/22.1.1/lisp/emacs-lisp/debug.el
15490 The only other new element of this function definition is the as yet
15491 unstudied function @code{append}, which merits a short section for
15494 @node append, , lengths-list-many-files, Several files
15495 @subsection The @code{append} Function
15498 The @code{append} function attaches one list to another. Thus,
15501 (append '(1 2 3 4) '(5 6 7 8))
15512 This is exactly how we want to attach two lengths' lists produced by
15513 @code{lengths-list-file} to each other. The results contrast with
15517 (cons '(1 2 3 4) '(5 6 7 8))
15522 which constructs a new list in which the first argument to @code{cons}
15523 becomes the first element of the new list:
15526 ((1 2 3 4) 5 6 7 8)
15529 @node Several files recursively, Prepare the data, Several files, Words in a defun
15530 @section Recursively Count Words in Different Files
15532 Besides a @code{while} loop, you can work on each of a list of files
15533 with recursion. A recursive version of @code{lengths-list-many-files}
15534 is short and simple.
15536 The recursive function has the usual parts: the `do-again-test', the
15537 `next-step-expression', and the recursive call. The `do-again-test'
15538 determines whether the function should call itself again, which it
15539 will do if the @code{list-of-files} contains any remaining elements;
15540 the `next-step-expression' resets the @code{list-of-files} to the
15541 @sc{cdr} of itself, so eventually the list will be empty; and the
15542 recursive call calls itself on the shorter list. The complete
15543 function is shorter than this description!
15544 @findex recursive-lengths-list-many-files
15548 (defun recursive-lengths-list-many-files (list-of-files)
15549 "Return list of lengths of each defun in LIST-OF-FILES."
15550 (if list-of-files ; @r{do-again-test}
15553 (expand-file-name (car list-of-files)))
15554 (recursive-lengths-list-many-files
15555 (cdr list-of-files)))))
15560 In a sentence, the function returns the lengths' list for the first of
15561 the @code{list-of-files} appended to the result of calling itself on
15562 the rest of the @code{list-of-files}.
15564 Here is a test of @code{recursive-lengths-list-many-files}, along with
15565 the results of running @code{lengths-list-file} on each of the files
15568 Install @code{recursive-lengths-list-many-files} and
15569 @code{lengths-list-file}, if necessary, and then evaluate the
15570 following expressions. You may need to change the files' pathnames;
15571 those here work when this Info file and the Emacs sources are located
15572 in their customary places. To change the expressions, copy them to
15573 the @file{*scratch*} buffer, edit them, and then evaluate them.
15575 The results are shown after the @samp{@result{}}. (These results are
15576 for files from Emacs version 22.1.1; files from other versions of
15577 Emacs may produce different results.)
15579 @c !!! 22.1.1 lisp sources location here
15582 (cd "/usr/local/share/emacs/22.1.1/")
15584 (lengths-list-file "./lisp/macros.el")
15585 @result{} (283 263 480 90)
15589 (lengths-list-file "./lisp/mail/mailalias.el")
15590 @result{} (38 32 29 95 178 180 321 218 324)
15594 (lengths-list-file "./lisp/makesum.el")
15599 (recursive-lengths-list-many-files
15600 '("./lisp/macros.el"
15601 "./lisp/mail/mailalias.el"
15602 "./lisp/makesum.el"))
15603 @result{} (283 263 480 90 38 32 29 95 178 180 321 218 324 85 181)
15607 The @code{recursive-lengths-list-many-files} function produces the
15610 The next step is to prepare the data in the list for display in a graph.
15612 @node Prepare the data, , Several files recursively, Words in a defun
15613 @section Prepare the Data for Display in a Graph
15615 The @code{recursive-lengths-list-many-files} function returns a list
15616 of numbers. Each number records the length of a function definition.
15617 What we need to do now is transform this data into a list of numbers
15618 suitable for generating a graph. The new list will tell how many
15619 functions definitions contain less than 10 words and
15620 symbols, how many contain between 10 and 19 words and symbols, how
15621 many contain between 20 and 29 words and symbols, and so on.
15623 In brief, we need to go through the lengths' list produced by the
15624 @code{recursive-lengths-list-many-files} function and count the number
15625 of defuns within each range of lengths, and produce a list of those
15629 * Data for Display in Detail::
15630 * Sorting:: Sorting lists.
15631 * Files List:: Making a list of files.
15632 * Counting function definitions::
15635 @node Data for Display in Detail, Sorting, Prepare the data, Prepare the data
15637 @unnumberedsubsec The Data for Display in Detail
15640 Based on what we have done before, we can readily foresee that it
15641 should not be too hard to write a function that `@sc{cdr}s' down the
15642 lengths' list, looks at each element, determines which length range it
15643 is in, and increments a counter for that range.
15645 However, before beginning to write such a function, we should consider
15646 the advantages of sorting the lengths' list first, so the numbers are
15647 ordered from smallest to largest. First, sorting will make it easier
15648 to count the numbers in each range, since two adjacent numbers will
15649 either be in the same length range or in adjacent ranges. Second, by
15650 inspecting a sorted list, we can discover the highest and lowest
15651 number, and thereby determine the largest and smallest length range
15654 @node Sorting, Files List, Data for Display in Detail, Prepare the data
15655 @subsection Sorting Lists
15658 Emacs contains a function to sort lists, called (as you might guess)
15659 @code{sort}. The @code{sort} function takes two arguments, the list
15660 to be sorted, and a predicate that determines whether the first of
15661 two list elements is ``less'' than the second.
15663 As we saw earlier (@pxref{Wrong Type of Argument, , Using the Wrong
15664 Type Object as an Argument}), a predicate is a function that
15665 determines whether some property is true or false. The @code{sort}
15666 function will reorder a list according to whatever property the
15667 predicate uses; this means that @code{sort} can be used to sort
15668 non-numeric lists by non-numeric criteria---it can, for example,
15669 alphabetize a list.
15672 The @code{<} function is used when sorting a numeric list. For example,
15675 (sort '(4 8 21 17 33 7 21 7) '<)
15683 (4 7 7 8 17 21 21 33)
15687 (Note that in this example, both the arguments are quoted so that the
15688 symbols are not evaluated before being passed to @code{sort} as
15691 Sorting the list returned by the
15692 @code{recursive-lengths-list-many-files} function is straightforward;
15693 it uses the @code{<} function:
15697 In GNU Emacs 22, eval
15699 (cd "/usr/local/share/emacs/22.0.50/")
15701 (recursive-lengths-list-many-files
15702 '("./lisp/macros.el"
15703 "./lisp/mail/mailalias.el"
15704 "./lisp/makesum.el"))
15712 (recursive-lengths-list-many-files
15713 '("./lisp/macros.el"
15714 "./lisp/mailalias.el"
15715 "./lisp/makesum.el"))
15725 (29 32 38 85 90 95 178 180 181 218 263 283 321 324 480)
15729 (Note that in this example, the first argument to @code{sort} is not
15730 quoted, since the expression must be evaluated so as to produce the
15731 list that is passed to @code{sort}.)
15733 @node Files List, Counting function definitions, Sorting, Prepare the data
15734 @subsection Making a List of Files
15736 The @code{recursive-lengths-list-many-files} function requires a list
15737 of files as its argument. For our test examples, we constructed such
15738 a list by hand; but the Emacs Lisp source directory is too large for
15739 us to do for that. Instead, we will write a function to do the job
15740 for us. In this function, we will use both a @code{while} loop and a
15743 @findex directory-files
15744 We did not have to write a function like this for older versions of
15745 GNU Emacs, since they placed all the @samp{.el} files in one
15746 directory. Instead, we were able to use the @code{directory-files}
15747 function, which lists the names of files that match a specified
15748 pattern within a single directory.
15750 However, recent versions of Emacs place Emacs Lisp files in
15751 sub-directories of the top level @file{lisp} directory. This
15752 re-arrangement eases navigation. For example, all the mail related
15753 files are in a @file{lisp} sub-directory called @file{mail}. But at
15754 the same time, this arrangement forces us to create a file listing
15755 function that descends into the sub-directories.
15757 @findex files-in-below-directory
15758 We can create this function, called @code{files-in-below-directory},
15759 using familiar functions such as @code{car}, @code{nthcdr}, and
15760 @code{substring} in conjunction with an existing function called
15761 @code{directory-files-and-attributes}. This latter function not only
15762 lists all the filenames in a directory, including the names
15763 of sub-directories, but also their attributes.
15765 To restate our goal: to create a function that will enable us
15766 to feed filenames to @code{recursive-lengths-list-many-files}
15767 as a list that looks like this (but with more elements):
15771 ("./lisp/macros.el"
15772 "./lisp/mail/rmail.el"
15773 "./lisp/makesum.el")
15777 The @code{directory-files-and-attributes} function returns a list of
15778 lists. Each of the lists within the main list consists of 13
15779 elements. The first element is a string that contains the name of the
15780 file -- which, in GNU/Linux, may be a `directory file', that is to
15781 say, a file with the special attributes of a directory. The second
15782 element of the list is @code{t} for a directory, a string
15783 for symbolic link (the string is the name linked to), or @code{nil}.
15785 For example, the first @samp{.el} file in the @file{lisp/} directory
15786 is @file{abbrev.el}. Its name is
15787 @file{/usr/local/share/emacs/22.1.1/lisp/abbrev.el} and it is not a
15788 directory or a symbolic link.
15791 This is how @code{directory-files-and-attributes} lists that file and
15817 On the other hand, @file{mail/} is a directory within the @file{lisp/}
15818 directory. The beginning of its listing looks like this:
15829 (To learn about the different attributes, look at the documentation of
15830 @code{file-attributes}. Bear in mind that the @code{file-attributes}
15831 function does not list the filename, so its first element is
15832 @code{directory-files-and-attributes}'s second element.)
15834 We will want our new function, @code{files-in-below-directory}, to
15835 list the @samp{.el} files in the directory it is told to check, and in
15836 any directories below that directory.
15838 This gives us a hint on how to construct
15839 @code{files-in-below-directory}: within a directory, the function
15840 should add @samp{.el} filenames to a list; and if, within a directory,
15841 the function comes upon a sub-directory, it should go into that
15842 sub-directory and repeat its actions.
15844 However, we should note that every directory contains a name that
15845 refers to itself, called @file{.}, (``dot'') and a name that refers to
15846 its parent directory, called @file{..} (``double dot''). (In
15847 @file{/}, the root directory, @file{..} refers to itself, since
15848 @file{/} has no parent.) Clearly, we do not want our
15849 @code{files-in-below-directory} function to enter those directories,
15850 since they always lead us, directly or indirectly, to the current
15853 Consequently, our @code{files-in-below-directory} function must do
15858 Check to see whether it is looking at a filename that ends in
15859 @samp{.el}; and if so, add its name to a list.
15862 Check to see whether it is looking at a filename that is the name of a
15863 directory; and if so,
15867 Check to see whether it is looking at @file{.} or @file{..}; and if
15871 Or else, go into that directory and repeat the process.
15875 Let's write a function definition to do these tasks. We will use a
15876 @code{while} loop to move from one filename to another within a
15877 directory, checking what needs to be done; and we will use a recursive
15878 call to repeat the actions on each sub-directory. The recursive
15879 pattern is `accumulate'
15880 (@pxref{Accumulate, , Recursive Pattern: @emph{accumulate}}),
15881 using @code{append} as the combiner.
15884 (directory-files "/usr/local/src/emacs/lisp/" t "\\.el$")
15885 (shell-command "find /usr/local/src/emacs/lisp/ -name '*.el'")
15887 (directory-files "/usr/local/share/emacs/22.1.1/lisp/" t "\\.el$")
15888 (shell-command "find /usr/local/share/emacs/22.1.1/lisp/ -name '*.el'")
15891 @c /usr/local/share/emacs/22.1.1/lisp/
15894 Here is the function:
15898 (defun files-in-below-directory (directory)
15899 "List the .el files in DIRECTORY and in its sub-directories."
15900 ;; Although the function will be used non-interactively,
15901 ;; it will be easier to test if we make it interactive.
15902 ;; The directory will have a name such as
15903 ;; "/usr/local/share/emacs/22.1.1/lisp/"
15904 (interactive "DDirectory name: ")
15907 (let (el-files-list
15908 (current-directory-list
15909 (directory-files-and-attributes directory t)))
15910 ;; while we are in the current directory
15911 (while current-directory-list
15915 ;; check to see whether filename ends in `.el'
15916 ;; and if so, append its name to a list.
15917 ((equal ".el" (substring (car (car current-directory-list)) -3))
15918 (setq el-files-list
15919 (cons (car (car current-directory-list)) el-files-list)))
15922 ;; check whether filename is that of a directory
15923 ((eq t (car (cdr (car current-directory-list))))
15924 ;; decide whether to skip or recurse
15927 (substring (car (car current-directory-list)) -1))
15928 ;; then do nothing since filename is that of
15929 ;; current directory or parent, "." or ".."
15933 ;; else descend into the directory and repeat the process
15934 (setq el-files-list
15936 (files-in-below-directory
15937 (car (car current-directory-list)))
15939 ;; move to the next filename in the list; this also
15940 ;; shortens the list so the while loop eventually comes to an end
15941 (setq current-directory-list (cdr current-directory-list)))
15942 ;; return the filenames
15947 @c (files-in-below-directory "/usr/local/src/emacs/lisp/")
15948 @c (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15950 The @code{files-in-below-directory} @code{directory-files} function
15951 takes one argument, the name of a directory.
15954 Thus, on my system,
15956 @c (length (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15958 @c !!! 22.1.1 lisp sources location here
15962 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/"))
15967 tells me that in and below my Lisp sources directory are 1031
15970 @code{files-in-below-directory} returns a list in reverse alphabetical
15971 order. An expression to sort the list in alphabetical order looks
15977 (files-in-below-directory "/usr/local/share/emacs/22.1.1/lisp/")
15984 "Test how long it takes to find lengths of all sorted elisp defuns."
15985 (insert "\n" (current-time-string) "\n")
15988 (recursive-lengths-list-many-files
15989 (files-in-below-directory "/usr/local/src/emacs/lisp/"))
15991 (insert (format "%s" (current-time-string))))
15994 @node Counting function definitions, , Files List, Prepare the data
15995 @subsection Counting function definitions
15997 Our immediate goal is to generate a list that tells us how many
15998 function definitions contain fewer than 10 words and symbols, how many
15999 contain between 10 and 19 words and symbols, how many contain between
16000 20 and 29 words and symbols, and so on.
16002 With a sorted list of numbers, this is easy: count how many elements
16003 of the list are smaller than 10, then, after moving past the numbers
16004 just counted, count how many are smaller than 20, then, after moving
16005 past the numbers just counted, count how many are smaller than 30, and
16006 so on. Each of the numbers, 10, 20, 30, 40, and the like, is one
16007 larger than the top of that range. We can call the list of such
16008 numbers the @code{top-of-ranges} list.
16011 If we wished, we could generate this list automatically, but it is
16012 simpler to write a list manually. Here it is:
16013 @vindex top-of-ranges
16017 (defvar top-of-ranges
16020 110 120 130 140 150
16021 160 170 180 190 200
16022 210 220 230 240 250
16023 260 270 280 290 300)
16024 "List specifying ranges for `defuns-per-range'.")
16028 To change the ranges, we edit this list.
16030 Next, we need to write the function that creates the list of the
16031 number of definitions within each range. Clearly, this function must
16032 take the @code{sorted-lengths} and the @code{top-of-ranges} lists
16035 The @code{defuns-per-range} function must do two things again and
16036 again: it must count the number of definitions within a range
16037 specified by the current top-of-range value; and it must shift to the
16038 next higher value in the @code{top-of-ranges} list after counting the
16039 number of definitions in the current range. Since each of these
16040 actions is repetitive, we can use @code{while} loops for the job.
16041 One loop counts the number of definitions in the range defined by the
16042 current top-of-range value, and the other loop selects each of the
16043 top-of-range values in turn.
16045 Several entries of the @code{sorted-lengths} list are counted for each
16046 range; this means that the loop for the @code{sorted-lengths} list
16047 will be inside the loop for the @code{top-of-ranges} list, like a
16048 small gear inside a big gear.
16050 The inner loop counts the number of definitions within the range. It
16051 is a simple counting loop of the type we have seen before.
16052 (@xref{Incrementing Loop, , A loop with an incrementing counter}.)
16053 The true-or-false test of the loop tests whether the value from the
16054 @code{sorted-lengths} list is smaller than the current value of the
16055 top of the range. If it is, the function increments the counter and
16056 tests the next value from the @code{sorted-lengths} list.
16059 The inner loop looks like this:
16063 (while @var{length-element-smaller-than-top-of-range}
16064 (setq number-within-range (1+ number-within-range))
16065 (setq sorted-lengths (cdr sorted-lengths)))
16069 The outer loop must start with the lowest value of the
16070 @code{top-of-ranges} list, and then be set to each of the succeeding
16071 higher values in turn. This can be done with a loop like this:
16075 (while top-of-ranges
16076 @var{body-of-loop}@dots{}
16077 (setq top-of-ranges (cdr top-of-ranges)))
16082 Put together, the two loops look like this:
16086 (while top-of-ranges
16088 ;; @r{Count the number of elements within the current range.}
16089 (while @var{length-element-smaller-than-top-of-range}
16090 (setq number-within-range (1+ number-within-range))
16091 (setq sorted-lengths (cdr sorted-lengths)))
16093 ;; @r{Move to next range.}
16094 (setq top-of-ranges (cdr top-of-ranges)))
16098 In addition, in each circuit of the outer loop, Emacs should record
16099 the number of definitions within that range (the value of
16100 @code{number-within-range}) in a list. We can use @code{cons} for
16101 this purpose. (@xref{cons, , @code{cons}}.)
16103 The @code{cons} function works fine, except that the list it
16104 constructs will contain the number of definitions for the highest
16105 range at its beginning and the number of definitions for the lowest
16106 range at its end. This is because @code{cons} attaches new elements
16107 of the list to the beginning of the list, and since the two loops are
16108 working their way through the lengths' list from the lower end first,
16109 the @code{defuns-per-range-list} will end up largest number first.
16110 But we will want to print our graph with smallest values first and the
16111 larger later. The solution is to reverse the order of the
16112 @code{defuns-per-range-list}. We can do this using the
16113 @code{nreverse} function, which reverses the order of a list.
16120 (nreverse '(1 2 3 4))
16131 Note that the @code{nreverse} function is ``destructive''---that is,
16132 it changes the list to which it is applied; this contrasts with the
16133 @code{car} and @code{cdr} functions, which are non-destructive. In
16134 this case, we do not want the original @code{defuns-per-range-list},
16135 so it does not matter that it is destroyed. (The @code{reverse}
16136 function provides a reversed copy of a list, leaving the original list
16141 Put all together, the @code{defuns-per-range} looks like this:
16145 (defun defuns-per-range (sorted-lengths top-of-ranges)
16146 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
16147 (let ((top-of-range (car top-of-ranges))
16148 (number-within-range 0)
16149 defuns-per-range-list)
16154 (while top-of-ranges
16160 ;; @r{Need number for numeric test.}
16161 (car sorted-lengths)
16162 (< (car sorted-lengths) top-of-range))
16166 ;; @r{Count number of definitions within current range.}
16167 (setq number-within-range (1+ number-within-range))
16168 (setq sorted-lengths (cdr sorted-lengths)))
16170 ;; @r{Exit inner loop but remain within outer loop.}
16174 (setq defuns-per-range-list
16175 (cons number-within-range defuns-per-range-list))
16176 (setq number-within-range 0) ; @r{Reset count to zero.}
16180 ;; @r{Move to next range.}
16181 (setq top-of-ranges (cdr top-of-ranges))
16182 ;; @r{Specify next top of range value.}
16183 (setq top-of-range (car top-of-ranges)))
16187 ;; @r{Exit outer loop and count the number of defuns larger than}
16188 ;; @r{ the largest top-of-range value.}
16189 (setq defuns-per-range-list
16191 (length sorted-lengths)
16192 defuns-per-range-list))
16196 ;; @r{Return a list of the number of definitions within each range,}
16197 ;; @r{ smallest to largest.}
16198 (nreverse defuns-per-range-list)))
16204 The function is straightforward except for one subtle feature. The
16205 true-or-false test of the inner loop looks like this:
16209 (and (car sorted-lengths)
16210 (< (car sorted-lengths) top-of-range))
16216 instead of like this:
16219 (< (car sorted-lengths) top-of-range)
16222 The purpose of the test is to determine whether the first item in the
16223 @code{sorted-lengths} list is less than the value of the top of the
16226 The simple version of the test works fine unless the
16227 @code{sorted-lengths} list has a @code{nil} value. In that case, the
16228 @code{(car sorted-lengths)} expression function returns
16229 @code{nil}. The @code{<} function cannot compare a number to
16230 @code{nil}, which is an empty list, so Emacs signals an error and
16231 stops the function from attempting to continue to execute.
16233 The @code{sorted-lengths} list always becomes @code{nil} when the
16234 counter reaches the end of the list. This means that any attempt to
16235 use the @code{defuns-per-range} function with the simple version of
16236 the test will fail.
16238 We solve the problem by using the @code{(car sorted-lengths)}
16239 expression in conjunction with the @code{and} expression. The
16240 @code{(car sorted-lengths)} expression returns a non-@code{nil}
16241 value so long as the list has at least one number within it, but
16242 returns @code{nil} if the list is empty. The @code{and} expression
16243 first evaluates the @code{(car sorted-lengths)} expression, and
16244 if it is @code{nil}, returns false @emph{without} evaluating the
16245 @code{<} expression. But if the @code{(car sorted-lengths)}
16246 expression returns a non-@code{nil} value, the @code{and} expression
16247 evaluates the @code{<} expression, and returns that value as the value
16248 of the @code{and} expression.
16250 @c colon in printed section title causes problem in Info cross reference
16251 This way, we avoid an error.
16254 (For information about @code{and}, see
16255 @ref{kill-new function, , The @code{kill-new} function}.)
16259 (@xref{kill-new function, , The @code{kill-new} function}, for
16260 information about @code{and}.)
16263 Here is a short test of the @code{defuns-per-range} function. First,
16264 evaluate the expression that binds (a shortened)
16265 @code{top-of-ranges} list to the list of values, then evaluate the
16266 expression for binding the @code{sorted-lengths} list, and then
16267 evaluate the @code{defuns-per-range} function.
16271 ;; @r{(Shorter list than we will use later.)}
16272 (setq top-of-ranges
16273 '(110 120 130 140 150
16274 160 170 180 190 200))
16276 (setq sorted-lengths
16277 '(85 86 110 116 122 129 154 176 179 200 265 300 300))
16279 (defuns-per-range sorted-lengths top-of-ranges)
16285 The list returned looks like this:
16288 (2 2 2 0 0 1 0 2 0 0 4)
16292 Indeed, there are two elements of the @code{sorted-lengths} list
16293 smaller than 110, two elements between 110 and 119, two elements
16294 between 120 and 129, and so on. There are four elements with a value
16297 @c The next step is to turn this numbers' list into a graph.
16298 @node Readying a Graph, Emacs Initialization, Words in a defun, Top
16299 @chapter Readying a Graph
16300 @cindex Readying a graph
16301 @cindex Graph prototype
16302 @cindex Prototype graph
16303 @cindex Body of graph
16305 Our goal is to construct a graph showing the numbers of function
16306 definitions of various lengths in the Emacs lisp sources.
16308 As a practical matter, if you were creating a graph, you would
16309 probably use a program such as @code{gnuplot} to do the job.
16310 (@code{gnuplot} is nicely integrated into GNU Emacs.) In this case,
16311 however, we create one from scratch, and in the process we will
16312 re-acquaint ourselves with some of what we learned before and learn
16315 In this chapter, we will first write a simple graph printing function.
16316 This first definition will be a @dfn{prototype}, a rapidly written
16317 function that enables us to reconnoiter this unknown graph-making
16318 territory. We will discover dragons, or find that they are myth.
16319 After scouting the terrain, we will feel more confident and enhance
16320 the function to label the axes automatically.
16323 * Columns of a graph::
16324 * graph-body-print:: How to print the body of a graph.
16325 * recursive-graph-body-print::
16327 * Line Graph Exercise::
16330 @node Columns of a graph, graph-body-print, Readying a Graph, Readying a Graph
16332 @unnumberedsec Printing the Columns of a Graph
16335 Since Emacs is designed to be flexible and work with all kinds of
16336 terminals, including character-only terminals, the graph will need to
16337 be made from one of the `typewriter' symbols. An asterisk will do; as
16338 we enhance the graph-printing function, we can make the choice of
16339 symbol a user option.
16341 We can call this function @code{graph-body-print}; it will take a
16342 @code{numbers-list} as its only argument. At this stage, we will not
16343 label the graph, but only print its body.
16345 The @code{graph-body-print} function inserts a vertical column of
16346 asterisks for each element in the @code{numbers-list}. The height of
16347 each line is determined by the value of that element of the
16348 @code{numbers-list}.
16350 Inserting columns is a repetitive act; that means that this function can
16351 be written either with a @code{while} loop or recursively.
16353 Our first challenge is to discover how to print a column of asterisks.
16354 Usually, in Emacs, we print characters onto a screen horizontally,
16355 line by line, by typing. We have two routes we can follow: write our
16356 own column-insertion function or discover whether one exists in Emacs.
16358 To see whether there is one in Emacs, we can use the @kbd{M-x apropos}
16359 command. This command is like the @kbd{C-h a} (@code{command-apropos})
16360 command, except that the latter finds only those functions that are
16361 commands. The @kbd{M-x apropos} command lists all symbols that match
16362 a regular expression, including functions that are not interactive.
16365 What we want to look for is some command that prints or inserts
16366 columns. Very likely, the name of the function will contain either
16367 the word `print' or the word `insert' or the word `column'.
16368 Therefore, we can simply type @kbd{M-x apropos RET
16369 print\|insert\|column RET} and look at the result. On my system, this
16370 command once too takes quite some time, and then produced a list of 79
16371 functions and variables. Now it does not take much time at all and
16372 produces a list of 211 functions and variables. Scanning down the
16373 list, the only function that looks as if it might do the job is
16374 @code{insert-rectangle}.
16377 Indeed, this is the function we want; its documentation says:
16382 Insert text of RECTANGLE with upper left corner at point.
16383 RECTANGLE's first line is inserted at point,
16384 its second line is inserted at a point vertically under point, etc.
16385 RECTANGLE should be a list of strings.
16386 After this command, the mark is at the upper left corner
16387 and point is at the lower right corner.
16391 We can run a quick test, to make sure it does what we expect of it.
16393 Here is the result of placing the cursor after the
16394 @code{insert-rectangle} expression and typing @kbd{C-u C-x C-e}
16395 (@code{eval-last-sexp}). The function inserts the strings
16396 @samp{"first"}, @samp{"second"}, and @samp{"third"} at and below
16397 point. Also the function returns @code{nil}.
16401 (insert-rectangle '("first" "second" "third"))first
16408 Of course, we won't be inserting the text of the
16409 @code{insert-rectangle} expression itself into the buffer in which we
16410 are making the graph, but will call the function from our program. We
16411 shall, however, have to make sure that point is in the buffer at the
16412 place where the @code{insert-rectangle} function will insert its
16415 If you are reading this in Info, you can see how this works by
16416 switching to another buffer, such as the @file{*scratch*} buffer,
16417 placing point somewhere in the buffer, typing @kbd{M-:}, typing the
16418 @code{insert-rectangle} expression into the minibuffer at the prompt,
16419 and then typing @key{RET}. This causes Emacs to evaluate the
16420 expression in the minibuffer, but to use as the value of point the
16421 position of point in the @file{*scratch*} buffer. (@kbd{M-:} is the
16422 keybinding for @code{eval-expression}. Also, @code{nil} does not
16423 appear in the @file{*scratch*} buffer since the expression is
16424 evaluated in the minibuffer.)
16426 We find when we do this that point ends up at the end of the last
16427 inserted line---that is to say, this function moves point as a
16428 side-effect. If we were to repeat the command, with point at this
16429 position, the next insertion would be below and to the right of the
16430 previous insertion. We don't want this! If we are going to make a
16431 bar graph, the columns need to be beside each other.
16433 So we discover that each cycle of the column-inserting @code{while}
16434 loop must reposition point to the place we want it, and that place
16435 will be at the top, not the bottom, of the column. Moreover, we
16436 remember that when we print a graph, we do not expect all the columns
16437 to be the same height. This means that the top of each column may be
16438 at a different height from the previous one. We cannot simply
16439 reposition point to the same line each time, but moved over to the
16440 right---or perhaps we can@dots{}
16442 We are planning to make the columns of the bar graph out of asterisks.
16443 The number of asterisks in the column is the number specified by the
16444 current element of the @code{numbers-list}. We need to construct a
16445 list of asterisks of the right length for each call to
16446 @code{insert-rectangle}. If this list consists solely of the requisite
16447 number of asterisks, then we will have position point the right number
16448 of lines above the base for the graph to print correctly. This could
16451 Alternatively, if we can figure out some way to pass
16452 @code{insert-rectangle} a list of the same length each time, then we
16453 can place point on the same line each time, but move it over one
16454 column to the right for each new column. If we do this, however, some
16455 of the entries in the list passed to @code{insert-rectangle} must be
16456 blanks rather than asterisks. For example, if the maximum height of
16457 the graph is 5, but the height of the column is 3, then
16458 @code{insert-rectangle} requires an argument that looks like this:
16461 (" " " " "*" "*" "*")
16464 This last proposal is not so difficult, so long as we can determine
16465 the column height. There are two ways for us to specify the column
16466 height: we can arbitrarily state what it will be, which would work
16467 fine for graphs of that height; or we can search through the list of
16468 numbers and use the maximum height of the list as the maximum height
16469 of the graph. If the latter operation were difficult, then the former
16470 procedure would be easiest, but there is a function built into Emacs
16471 that determines the maximum of its arguments. We can use that
16472 function. The function is called @code{max} and it returns the
16473 largest of all its arguments, which must be numbers. Thus, for
16481 returns 7. (A corresponding function called @code{min} returns the
16482 smallest of all its arguments.)
16486 However, we cannot simply call @code{max} on the @code{numbers-list};
16487 the @code{max} function expects numbers as its argument, not a list of
16488 numbers. Thus, the following expression,
16491 (max '(3 4 6 5 7 3))
16496 produces the following error message;
16499 Wrong type of argument: number-or-marker-p, (3 4 6 5 7 3)
16503 We need a function that passes a list of arguments to a function.
16504 This function is @code{apply}. This function `applies' its first
16505 argument (a function) to its remaining arguments, the last of which
16512 (apply 'max 3 4 7 3 '(4 8 5))
16518 (Incidentally, I don't know how you would learn of this function
16519 without a book such as this. It is possible to discover other
16520 functions, like @code{search-forward} or @code{insert-rectangle}, by
16521 guessing at a part of their names and then using @code{apropos}. Even
16522 though its base in metaphor is clear---`apply' its first argument to
16523 the rest---I doubt a novice would come up with that particular word
16524 when using @code{apropos} or other aid. Of course, I could be wrong;
16525 after all, the function was first named by someone who had to invent
16528 The second and subsequent arguments to @code{apply} are optional, so
16529 we can use @code{apply} to call a function and pass the elements of a
16530 list to it, like this, which also returns 8:
16533 (apply 'max '(4 8 5))
16536 This latter way is how we will use @code{apply}. The
16537 @code{recursive-lengths-list-many-files} function returns a numbers'
16538 list to which we can apply @code{max} (we could also apply @code{max} to
16539 the sorted numbers' list; it does not matter whether the list is
16543 Hence, the operation for finding the maximum height of the graph is this:
16546 (setq max-graph-height (apply 'max numbers-list))
16549 Now we can return to the question of how to create a list of strings
16550 for a column of the graph. Told the maximum height of the graph
16551 and the number of asterisks that should appear in the column, the
16552 function should return a list of strings for the
16553 @code{insert-rectangle} command to insert.
16555 Each column is made up of asterisks or blanks. Since the function is
16556 passed the value of the height of the column and the number of
16557 asterisks in the column, the number of blanks can be found by
16558 subtracting the number of asterisks from the height of the column.
16559 Given the number of blanks and the number of asterisks, two
16560 @code{while} loops can be used to construct the list:
16564 ;;; @r{First version.}
16565 (defun column-of-graph (max-graph-height actual-height)
16566 "Return list of strings that is one column of a graph."
16567 (let ((insert-list nil)
16568 (number-of-top-blanks
16569 (- max-graph-height actual-height)))
16573 ;; @r{Fill in asterisks.}
16574 (while (> actual-height 0)
16575 (setq insert-list (cons "*" insert-list))
16576 (setq actual-height (1- actual-height)))
16580 ;; @r{Fill in blanks.}
16581 (while (> number-of-top-blanks 0)
16582 (setq insert-list (cons " " insert-list))
16583 (setq number-of-top-blanks
16584 (1- number-of-top-blanks)))
16588 ;; @r{Return whole list.}
16593 If you install this function and then evaluate the following
16594 expression you will see that it returns the list as desired:
16597 (column-of-graph 5 3)
16605 (" " " " "*" "*" "*")
16608 As written, @code{column-of-graph} contains a major flaw: the symbols
16609 used for the blank and for the marked entries in the column are
16610 `hard-coded' as a space and asterisk. This is fine for a prototype,
16611 but you, or another user, may wish to use other symbols. For example,
16612 in testing the graph function, you many want to use a period in place
16613 of the space, to make sure the point is being repositioned properly
16614 each time the @code{insert-rectangle} function is called; or you might
16615 want to substitute a @samp{+} sign or other symbol for the asterisk.
16616 You might even want to make a graph-column that is more than one
16617 display column wide. The program should be more flexible. The way to
16618 do that is to replace the blank and the asterisk with two variables
16619 that we can call @code{graph-blank} and @code{graph-symbol} and define
16620 those variables separately.
16622 Also, the documentation is not well written. These considerations
16623 lead us to the second version of the function:
16627 (defvar graph-symbol "*"
16628 "String used as symbol in graph, usually an asterisk.")
16632 (defvar graph-blank " "
16633 "String used as blank in graph, usually a blank space.
16634 graph-blank must be the same number of columns wide
16640 (For an explanation of @code{defvar}, see
16641 @ref{defvar, , Initializing a Variable with @code{defvar}}.)
16645 ;;; @r{Second version.}
16646 (defun column-of-graph (max-graph-height actual-height)
16647 "Return MAX-GRAPH-HEIGHT strings; ACTUAL-HEIGHT are graph-symbols.
16651 The graph-symbols are contiguous entries at the end
16653 The list will be inserted as one column of a graph.
16654 The strings are either graph-blank or graph-symbol."
16658 (let ((insert-list nil)
16659 (number-of-top-blanks
16660 (- max-graph-height actual-height)))
16664 ;; @r{Fill in @code{graph-symbols}.}
16665 (while (> actual-height 0)
16666 (setq insert-list (cons graph-symbol insert-list))
16667 (setq actual-height (1- actual-height)))
16671 ;; @r{Fill in @code{graph-blanks}.}
16672 (while (> number-of-top-blanks 0)
16673 (setq insert-list (cons graph-blank insert-list))
16674 (setq number-of-top-blanks
16675 (1- number-of-top-blanks)))
16677 ;; @r{Return whole list.}
16682 If we wished, we could rewrite @code{column-of-graph} a third time to
16683 provide optionally for a line graph as well as for a bar graph. This
16684 would not be hard to do. One way to think of a line graph is that it
16685 is no more than a bar graph in which the part of each bar that is
16686 below the top is blank. To construct a column for a line graph, the
16687 function first constructs a list of blanks that is one shorter than
16688 the value, then it uses @code{cons} to attach a graph symbol to the
16689 list; then it uses @code{cons} again to attach the `top blanks' to
16692 It is easy to see how to write such a function, but since we don't
16693 need it, we will not do it. But the job could be done, and if it were
16694 done, it would be done with @code{column-of-graph}. Even more
16695 important, it is worth noting that few changes would have to be made
16696 anywhere else. The enhancement, if we ever wish to make it, is
16699 Now, finally, we come to our first actual graph printing function.
16700 This prints the body of a graph, not the labels for the vertical and
16701 horizontal axes, so we can call this @code{graph-body-print}.
16703 @node graph-body-print, recursive-graph-body-print, Columns of a graph, Readying a Graph
16704 @section The @code{graph-body-print} Function
16705 @findex graph-body-print
16707 After our preparation in the preceding section, the
16708 @code{graph-body-print} function is straightforward. The function
16709 will print column after column of asterisks and blanks, using the
16710 elements of a numbers' list to specify the number of asterisks in each
16711 column. This is a repetitive act, which means we can use a
16712 decrementing @code{while} loop or recursive function for the job. In
16713 this section, we will write the definition using a @code{while} loop.
16715 The @code{column-of-graph} function requires the height of the graph
16716 as an argument, so we should determine and record that as a local variable.
16718 This leads us to the following template for the @code{while} loop
16719 version of this function:
16723 (defun graph-body-print (numbers-list)
16724 "@var{documentation}@dots{}"
16725 (let ((height @dots{}
16730 (while numbers-list
16731 @var{insert-columns-and-reposition-point}
16732 (setq numbers-list (cdr numbers-list)))))
16737 We need to fill in the slots of the template.
16739 Clearly, we can use the @code{(apply 'max numbers-list)} expression to
16740 determine the height of the graph.
16742 The @code{while} loop will cycle through the @code{numbers-list} one
16743 element at a time. As it is shortened by the @code{(setq numbers-list
16744 (cdr numbers-list))} expression, the @sc{car} of each instance of the
16745 list is the value of the argument for @code{column-of-graph}.
16747 At each cycle of the @code{while} loop, the @code{insert-rectangle}
16748 function inserts the list returned by @code{column-of-graph}. Since
16749 the @code{insert-rectangle} function moves point to the lower right of
16750 the inserted rectangle, we need to save the location of point at the
16751 time the rectangle is inserted, move back to that position after the
16752 rectangle is inserted, and then move horizontally to the next place
16753 from which @code{insert-rectangle} is called.
16755 If the inserted columns are one character wide, as they will be if
16756 single blanks and asterisks are used, the repositioning command is
16757 simply @code{(forward-char 1)}; however, the width of a column may be
16758 greater than one. This means that the repositioning command should be
16759 written @code{(forward-char symbol-width)}. The @code{symbol-width}
16760 itself is the length of a @code{graph-blank} and can be found using
16761 the expression @code{(length graph-blank)}. The best place to bind
16762 the @code{symbol-width} variable to the value of the width of graph
16763 column is in the varlist of the @code{let} expression.
16766 These considerations lead to the following function definition:
16770 (defun graph-body-print (numbers-list)
16771 "Print a bar graph of the NUMBERS-LIST.
16772 The numbers-list consists of the Y-axis values."
16774 (let ((height (apply 'max numbers-list))
16775 (symbol-width (length graph-blank))
16780 (while numbers-list
16781 (setq from-position (point))
16783 (column-of-graph height (car numbers-list)))
16784 (goto-char from-position)
16785 (forward-char symbol-width)
16788 ;; @r{Draw graph column by column.}
16790 (setq numbers-list (cdr numbers-list)))
16793 ;; @r{Place point for X axis labels.}
16794 (forward-line height)
16801 The one unexpected expression in this function is the
16802 @w{@code{(sit-for 0)}} expression in the @code{while} loop. This
16803 expression makes the graph printing operation more interesting to
16804 watch than it would be otherwise. The expression causes Emacs to
16805 `sit' or do nothing for a zero length of time and then redraw the
16806 screen. Placed here, it causes Emacs to redraw the screen column by
16807 column. Without it, Emacs would not redraw the screen until the
16810 We can test @code{graph-body-print} with a short list of numbers.
16814 Install @code{graph-symbol}, @code{graph-blank},
16815 @code{column-of-graph}, which are in
16817 @ref{Readying a Graph, , Readying a Graph},
16820 @ref{Columns of a graph},
16822 and @code{graph-body-print}.
16826 Copy the following expression:
16829 (graph-body-print '(1 2 3 4 6 4 3 5 7 6 5 2 3))
16833 Switch to the @file{*scratch*} buffer and place the cursor where you
16834 want the graph to start.
16837 Type @kbd{M-:} (@code{eval-expression}).
16840 Yank the @code{graph-body-print} expression into the minibuffer
16841 with @kbd{C-y} (@code{yank)}.
16844 Press @key{RET} to evaluate the @code{graph-body-print} expression.
16848 Emacs will print a graph like this:
16862 @node recursive-graph-body-print, Printed Axes, graph-body-print, Readying a Graph
16863 @section The @code{recursive-graph-body-print} Function
16864 @findex recursive-graph-body-print
16866 The @code{graph-body-print} function may also be written recursively.
16867 The recursive solution is divided into two parts: an outside `wrapper'
16868 that uses a @code{let} expression to determine the values of several
16869 variables that need only be found once, such as the maximum height of
16870 the graph, and an inside function that is called recursively to print
16874 The `wrapper' is uncomplicated:
16878 (defun recursive-graph-body-print (numbers-list)
16879 "Print a bar graph of the NUMBERS-LIST.
16880 The numbers-list consists of the Y-axis values."
16881 (let ((height (apply 'max numbers-list))
16882 (symbol-width (length graph-blank))
16884 (recursive-graph-body-print-internal
16891 The recursive function is a little more difficult. It has four parts:
16892 the `do-again-test', the printing code, the recursive call, and the
16893 `next-step-expression'. The `do-again-test' is a @code{when}
16894 expression that determines whether the @code{numbers-list} contains
16895 any remaining elements; if it does, the function prints one column of
16896 the graph using the printing code and calls itself again. The
16897 function calls itself again according to the value produced by the
16898 `next-step-expression' which causes the call to act on a shorter
16899 version of the @code{numbers-list}.
16903 (defun recursive-graph-body-print-internal
16904 (numbers-list height symbol-width)
16905 "Print a bar graph.
16906 Used within recursive-graph-body-print function."
16911 (setq from-position (point))
16913 (column-of-graph height (car numbers-list)))
16916 (goto-char from-position)
16917 (forward-char symbol-width)
16918 (sit-for 0) ; @r{Draw graph column by column.}
16919 (recursive-graph-body-print-internal
16920 (cdr numbers-list) height symbol-width)))
16925 After installation, this expression can be tested; here is a sample:
16928 (recursive-graph-body-print '(3 2 5 6 7 5 3 4 6 4 3 2 1))
16932 Here is what @code{recursive-graph-body-print} produces:
16946 Either of these two functions, @code{graph-body-print} or
16947 @code{recursive-graph-body-print}, create the body of a graph.
16949 @node Printed Axes, Line Graph Exercise, recursive-graph-body-print, Readying a Graph
16950 @section Need for Printed Axes
16952 A graph needs printed axes, so you can orient yourself. For a do-once
16953 project, it may be reasonable to draw the axes by hand using Emacs'
16954 Picture mode; but a graph drawing function may be used more than once.
16956 For this reason, I have written enhancements to the basic
16957 @code{print-graph-body} function that automatically print labels for
16958 the horizontal and vertical axes. Since the label printing functions
16959 do not contain much new material, I have placed their description in
16960 an appendix. @xref{Full Graph, , A Graph with Labelled Axes}.
16962 @node Line Graph Exercise, , Printed Axes, Readying a Graph
16965 Write a line graph version of the graph printing functions.
16967 @node Emacs Initialization, Debugging, Readying a Graph, Top
16968 @chapter Your @file{.emacs} File
16969 @cindex @file{.emacs} file
16970 @cindex Customizing your @file{.emacs} file
16971 @cindex Initialization file
16973 ``You don't have to like Emacs to like it'' -- this seemingly
16974 paradoxical statement is the secret of GNU Emacs. The plain, `out of
16975 the box' Emacs is a generic tool. Most people who use it, customize
16976 it to suit themselves.
16978 GNU Emacs is mostly written in Emacs Lisp; this means that by writing
16979 expressions in Emacs Lisp you can change or extend Emacs.
16982 * Default Configuration::
16983 * Site-wide Init:: You can write site-wide init files.
16984 * defcustom:: Emacs will write code for you.
16985 * Beginning a .emacs File:: How to write a @code{.emacs file}.
16986 * Text and Auto-fill:: Automatically wrap lines.
16987 * Mail Aliases:: Use abbreviations for email addresses.
16988 * Indent Tabs Mode:: Don't use tabs with @TeX{}
16989 * Keybindings:: Create some personal keybindings.
16990 * Keymaps:: More about key binding.
16991 * Loading Files:: Load (i.e., evaluate) files automatically.
16992 * Autoload:: Make functions available.
16993 * Simple Extension:: Define a function; bind it to a key.
16994 * X11 Colors:: Colors in X.
16996 * Mode Line:: How to customize your mode line.
16999 @node Default Configuration, Site-wide Init, Emacs Initialization, Emacs Initialization
17001 @unnumberedsec Emacs' Default Configuration
17004 There are those who appreciate Emacs' default configuration. After
17005 all, Emacs starts you in C mode when you edit a C file, starts you in
17006 Fortran mode when you edit a Fortran file, and starts you in
17007 Fundamental mode when you edit an unadorned file. This all makes
17008 sense, if you do not know who is going to use Emacs. Who knows what a
17009 person hopes to do with an unadorned file? Fundamental mode is the
17010 right default for such a file, just as C mode is the right default for
17011 editing C code. (Enough programming languages have syntaxes
17012 that enable them to share or nearly share features, so C mode is
17013 now provided by CC mode, the `C Collection'.)
17015 But when you do know who is going to use Emacs---you,
17016 yourself---then it makes sense to customize Emacs.
17018 For example, I seldom want Fundamental mode when I edit an
17019 otherwise undistinguished file; I want Text mode. This is why I
17020 customize Emacs: so it suits me.
17022 You can customize and extend Emacs by writing or adapting a
17023 @file{~/.emacs} file. This is your personal initialization file; its
17024 contents, written in Emacs Lisp, tell Emacs what to do.@footnote{You
17025 may also add @file{.el} to @file{~/.emacs} and call it a
17026 @file{~/.emacs.el} file. In the past, you were forbidden to type the
17027 extra keystrokes that the name @file{~/.emacs.el} requires, but now
17028 you may. The new format is consistent with the Emacs Lisp file
17029 naming conventions; the old format saves typing.}
17031 A @file{~/.emacs} file contains Emacs Lisp code. You can write this
17032 code yourself; or you can use Emacs' @code{customize} feature to write
17033 the code for you. You can combine your own expressions and
17034 auto-written Customize expressions in your @file{.emacs} file.
17036 (I myself prefer to write my own expressions, except for those,
17037 particularly fonts, that I find easier to manipulate using the
17038 @code{customize} command. I combine the two methods.)
17040 Most of this chapter is about writing expressions yourself. It
17041 describes a simple @file{.emacs} file; for more information, see
17042 @ref{Init File, , The Init File, emacs, The GNU Emacs Manual}, and
17043 @ref{Init File, , The Init File, elisp, The GNU Emacs Lisp Reference
17046 @node Site-wide Init, defcustom, Default Configuration, Emacs Initialization
17047 @section Site-wide Initialization Files
17049 @cindex @file{default.el} init file
17050 @cindex @file{site-init.el} init file
17051 @cindex @file{site-load.el} init file
17052 In addition to your personal initialization file, Emacs automatically
17053 loads various site-wide initialization files, if they exist. These
17054 have the same form as your @file{.emacs} file, but are loaded by
17057 Two site-wide initialization files, @file{site-load.el} and
17058 @file{site-init.el}, are loaded into Emacs and then `dumped' if a
17059 `dumped' version of Emacs is created, as is most common. (Dumped
17060 copies of Emacs load more quickly. However, once a file is loaded and
17061 dumped, a change to it does not lead to a change in Emacs unless you
17062 load it yourself or re-dump Emacs. @xref{Building Emacs, , Building
17063 Emacs, elisp, The GNU Emacs Lisp Reference Manual}, and the
17064 @file{INSTALL} file.)
17066 Three other site-wide initialization files are loaded automatically
17067 each time you start Emacs, if they exist. These are
17068 @file{site-start.el}, which is loaded @emph{before} your @file{.emacs}
17069 file, and @file{default.el}, and the terminal type file, which are both
17070 loaded @emph{after} your @file{.emacs} file.
17072 Settings and definitions in your @file{.emacs} file will overwrite
17073 conflicting settings and definitions in a @file{site-start.el} file,
17074 if it exists; but the settings and definitions in a @file{default.el}
17075 or terminal type file will overwrite those in your @file{.emacs} file.
17076 (You can prevent interference from a terminal type file by setting
17077 @code{term-file-prefix} to @code{nil}. @xref{Simple Extension, , A
17078 Simple Extension}.)
17080 @c Rewritten to avoid overfull hbox.
17081 The @file{INSTALL} file that comes in the distribution contains
17082 descriptions of the @file{site-init.el} and @file{site-load.el} files.
17084 The @file{loadup.el}, @file{startup.el}, and @file{loaddefs.el} files
17085 control loading. These files are in the @file{lisp} directory of the
17086 Emacs distribution and are worth perusing.
17088 The @file{loaddefs.el} file contains a good many suggestions as to
17089 what to put into your own @file{.emacs} file, or into a site-wide
17090 initialization file.
17092 @node defcustom, Beginning a .emacs File, Site-wide Init, Emacs Initialization
17093 @section Specifying Variables using @code{defcustom}
17096 You can specify variables using @code{defcustom} so that you and
17097 others can then use Emacs' @code{customize} feature to set their
17098 values. (You cannot use @code{customize} to write function
17099 definitions; but you can write @code{defuns} in your @file{.emacs}
17100 file. Indeed, you can write any Lisp expression in your @file{.emacs}
17103 The @code{customize} feature depends on the @code{defcustom} special
17104 form. Although you can use @code{defvar} or @code{setq} for variables
17105 that users set, the @code{defcustom} special form is designed for the
17108 You can use your knowledge of @code{defvar} for writing the
17109 first three arguments for @code{defcustom}. The first argument to
17110 @code{defcustom} is the name of the variable. The second argument is
17111 the variable's initial value, if any; and this value is set only if
17112 the value has not already been set. The third argument is the
17115 The fourth and subsequent arguments to @code{defcustom} specify types
17116 and options; these are not featured in @code{defvar}. (These
17117 arguments are optional.)
17119 Each of these arguments consists of a keyword followed by a value.
17120 Each keyword starts with the colon character @samp{:}.
17123 For example, the customizable user option variable
17124 @code{text-mode-hook} looks like this:
17128 (defcustom text-mode-hook nil
17129 "Normal hook run when entering Text mode and many related modes."
17131 :options '(turn-on-auto-fill flyspell-mode)
17137 The name of the variable is @code{text-mode-hook}; it has no default
17138 value; and its documentation string tells you what it does.
17140 The @code{:type} keyword tells Emacs the kind of data to which
17141 @code{text-mode-hook} should be set and how to display the value in a
17142 Customization buffer.
17144 The @code{:options} keyword specifies a suggested list of values for
17145 the variable. Usually, @code{:options} applies to a hook.
17146 The list is only a suggestion; it is not exclusive; a person who sets
17147 the variable may set it to other values; the list shown following the
17148 @code{:options} keyword is intended to offer convenient choices to a
17151 Finally, the @code{:group} keyword tells the Emacs Customization
17152 command in which group the variable is located. This tells where to
17155 The @code{defcustom} function recognizes more than a dozen keywords.
17156 For more information, see @ref{Customization, , Writing Customization
17157 Definitions, elisp, The GNU Emacs Lisp Reference Manual}.
17159 Consider @code{text-mode-hook} as an example.
17161 There are two ways to customize this variable. You can use the
17162 customization command or write the appropriate expressions yourself.
17165 Using the customization command, you can type:
17172 and find that the group for editing files of data is called `data'.
17173 Enter that group. Text Mode Hook is the first member. You can click
17174 on its various options, such as @code{turn-on-auto-fill}, to set the
17175 values. After you click on the button to
17178 Save for Future Sessions
17182 Emacs will write an expression into your @file{.emacs} file.
17183 It will look like this:
17187 (custom-set-variables
17188 ;; custom-set-variables was added by Custom.
17189 ;; If you edit it by hand, you could mess it up, so be careful.
17190 ;; Your init file should contain only one such instance.
17191 ;; If there is more than one, they won't work right.
17192 '(text-mode-hook (quote (turn-on-auto-fill text-mode-hook-identify))))
17197 (The @code{text-mode-hook-identify} function tells
17198 @code{toggle-text-mode-auto-fill} which buffers are in Text mode.
17199 It comes on automatically.)
17201 The @code{custom-set-variables} function works somewhat differently
17202 than a @code{setq}. While I have never learned the differences, I
17203 modify the @code{custom-set-variables} expressions in my @file{.emacs}
17204 file by hand: I make the changes in what appears to me to be a
17205 reasonable manner and have not had any problems. Others prefer to use
17206 the Customization command and let Emacs do the work for them.
17208 Another @code{custom-set-@dots{}} function is @code{custom-set-faces}.
17209 This function sets the various font faces. Over time, I have set a
17210 considerable number of faces. Some of the time, I re-set them using
17211 @code{customize}; other times, I simply edit the
17212 @code{custom-set-faces} expression in my @file{.emacs} file itself.
17214 The second way to customize your @code{text-mode-hook} is to set it
17215 yourself in your @file{.emacs} file using code that has nothing to do
17216 with the @code{custom-set-@dots{}} functions.
17219 When you do this, and later use @code{customize}, you will see a
17223 CHANGED outside Customize; operating on it here may be unreliable.
17227 This message is only a warning. If you click on the button to
17230 Save for Future Sessions
17234 Emacs will write a @code{custom-set-@dots{}} expression near the end
17235 of your @file{.emacs} file that will be evaluated after your
17236 hand-written expression. It will, therefore, overrule your
17237 hand-written expression. No harm will be done. When you do this,
17238 however, be careful to remember which expression is active; if you
17239 forget, you may confuse yourself.
17241 So long as you remember where the values are set, you will have no
17242 trouble. In any event, the values are always set in your
17243 initialization file, which is usually called @file{.emacs}.
17245 I myself use @code{customize} for hardly anything. Mostly, I write
17246 expressions myself.
17250 Incidentally, to be more complete concerning defines: @code{defsubst}
17251 defines an inline function. The syntax is just like that of
17252 @code{defun}. @code{defconst} defines a symbol as a constant. The
17253 intent is that neither programs nor users should ever change a value
17254 set by @code{defconst}. (You can change it; the value set is a
17255 variable; but please do not.)
17257 @node Beginning a .emacs File, Text and Auto-fill, defcustom, Emacs Initialization
17258 @section Beginning a @file{.emacs} File
17259 @cindex @file{.emacs} file, beginning of
17261 When you start Emacs, it loads your @file{.emacs} file unless you tell
17262 it not to by specifying @samp{-q} on the command line. (The
17263 @code{emacs -q} command gives you a plain, out-of-the-box Emacs.)
17265 A @file{.emacs} file contains Lisp expressions. Often, these are no
17266 more than expressions to set values; sometimes they are function
17269 @xref{Init File, , The Init File @file{~/.emacs}, emacs, The GNU Emacs
17270 Manual}, for a short description of initialization files.
17272 This chapter goes over some of the same ground, but is a walk among
17273 extracts from a complete, long-used @file{.emacs} file---my own.
17275 The first part of the file consists of comments: reminders to myself.
17276 By now, of course, I remember these things, but when I started, I did
17282 ;;;; Bob's .emacs file
17283 ; Robert J. Chassell
17284 ; 26 September 1985
17289 Look at that date! I started this file a long time ago. I have been
17290 adding to it ever since.
17294 ; Each section in this file is introduced by a
17295 ; line beginning with four semicolons; and each
17296 ; entry is introduced by a line beginning with
17297 ; three semicolons.
17302 This describes the usual conventions for comments in Emacs Lisp.
17303 Everything on a line that follows a semicolon is a comment. Two,
17304 three, and four semicolons are used as subsection and section markers.
17305 (@xref{Comments, ,, elisp, The GNU Emacs Lisp Reference Manual}, for
17306 more about comments.)
17311 ; Control-h is the help key;
17312 ; after typing control-h, type a letter to
17313 ; indicate the subject about which you want help.
17314 ; For an explanation of the help facility,
17315 ; type control-h two times in a row.
17320 Just remember: type @kbd{C-h} two times for help.
17324 ; To find out about any mode, type control-h m
17325 ; while in that mode. For example, to find out
17326 ; about mail mode, enter mail mode and then type
17332 `Mode help', as I call this, is very helpful. Usually, it tells you
17333 all you need to know.
17335 Of course, you don't need to include comments like these in your
17336 @file{.emacs} file. I included them in mine because I kept forgetting
17337 about Mode help or the conventions for comments---but I was able to
17338 remember to look here to remind myself.
17340 @node Text and Auto-fill, Mail Aliases, Beginning a .emacs File, Emacs Initialization
17341 @section Text and Auto Fill Mode
17343 Now we come to the part that `turns on' Text mode and
17348 ;;; Text mode and Auto Fill mode
17349 ; The next two lines put Emacs into Text mode
17350 ; and Auto Fill mode, and are for writers who
17351 ; want to start writing prose rather than code.
17352 (setq default-major-mode 'text-mode)
17353 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17357 Here is the first part of this @file{.emacs} file that does something
17358 besides remind a forgetful human!
17360 The first of the two lines in parentheses tells Emacs to turn on Text
17361 mode when you find a file, @emph{unless} that file should go into some
17362 other mode, such as C mode.
17364 @cindex Per-buffer, local variables list
17365 @cindex Local variables list, per-buffer,
17366 @cindex Automatic mode selection
17367 @cindex Mode selection, automatic
17368 When Emacs reads a file, it looks at the extension to the file name,
17369 if any. (The extension is the part that comes after a @samp{.}.) If
17370 the file ends with a @samp{.c} or @samp{.h} extension then Emacs turns
17371 on C mode. Also, Emacs looks at first nonblank line of the file; if
17372 the line says @w{@samp{-*- C -*-}}, Emacs turns on C mode. Emacs
17373 possesses a list of extensions and specifications that it uses
17374 automatically. In addition, Emacs looks near the last page for a
17375 per-buffer, ``local variables list'', if any.
17378 @xref{Choosing Modes, , How Major Modes are Chosen, emacs, The GNU
17381 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17385 See sections ``How Major Modes are Chosen'' and ``Local Variables in
17386 Files'' in @cite{The GNU Emacs Manual}.
17389 Now, back to the @file{.emacs} file.
17392 Here is the line again; how does it work?
17394 @cindex Text Mode turned on
17396 (setq major-mode 'text-mode)
17400 This line is a short, but complete Emacs Lisp expression.
17402 We are already familiar with @code{setq}. It sets the following variable,
17403 @code{major-mode}, to the subsequent value, which is @code{text-mode}.
17404 The single quote mark before @code{text-mode} tells Emacs to deal directly
17405 with the @code{text-mode} symbol, not with whatever it might stand for.
17406 @xref{set & setq, , Setting the Value of a Variable},
17407 for a reminder of how @code{setq} works.
17408 The main point is that there is no difference between the procedure you
17409 use to set a value in your @file{.emacs} file and the procedure you use
17410 anywhere else in Emacs.
17413 Here is the next line:
17415 @cindex Auto Fill mode turned on
17418 (add-hook 'text-mode-hook 'turn-on-auto-fill)
17422 In this line, the @code{add-hook} command adds
17423 @code{turn-on-auto-fill} to the variable.
17425 @code{turn-on-auto-fill} is the name of a program, that, you guessed
17426 it!, turns on Auto Fill mode.
17428 Every time Emacs turns on Text mode, Emacs runs the commands `hooked'
17429 onto Text mode. So every time Emacs turns on Text mode, Emacs also
17430 turns on Auto Fill mode.
17432 In brief, the first line causes Emacs to enter Text mode when you edit a
17433 file, unless the file name extension, a first non-blank line, or local
17434 variables to tell Emacs otherwise.
17436 Text mode among other actions, sets the syntax table to work
17437 conveniently for writers. In Text mode, Emacs considers an apostrophe
17438 as part of a word like a letter; but Emacs does not consider a period
17439 or a space as part of a word. Thus, @kbd{M-f} moves you over
17440 @samp{it's}. On the other hand, in C mode, @kbd{M-f} stops just after
17441 the @samp{t} of @samp{it's}.
17443 The second line causes Emacs to turn on Auto Fill mode when it turns
17444 on Text mode. In Auto Fill mode, Emacs automatically breaks a line
17445 that is too wide and brings the excessively wide part of the line down
17446 to the next line. Emacs breaks lines between words, not within them.
17448 When Auto Fill mode is turned off, lines continue to the right as you
17449 type them. Depending on how you set the value of
17450 @code{truncate-lines}, the words you type either disappear off the
17451 right side of the screen, or else are shown, in a rather ugly and
17452 unreadable manner, as a continuation line on the screen.
17455 In addition, in this part of my @file{.emacs} file, I tell the Emacs
17456 fill commands to insert two spaces after a colon:
17459 (setq colon-double-space t)
17462 @node Mail Aliases, Indent Tabs Mode, Text and Auto-fill, Emacs Initialization
17463 @section Mail Aliases
17465 Here is a @code{setq} that `turns on' mail aliases, along with more
17471 ; To enter mail mode, type `C-x m'
17472 ; To enter RMAIL (for reading mail),
17474 (setq mail-aliases t)
17478 @cindex Mail aliases
17480 This @code{setq} command sets the value of the variable
17481 @code{mail-aliases} to @code{t}. Since @code{t} means true, the line
17482 says, in effect, ``Yes, use mail aliases.''
17484 Mail aliases are convenient short names for long email addresses or
17485 for lists of email addresses. The file where you keep your `aliases'
17486 is @file{~/.mailrc}. You write an alias like this:
17489 alias geo george@@foobar.wiz.edu
17493 When you write a message to George, address it to @samp{geo}; the
17494 mailer will automatically expand @samp{geo} to the full address.
17496 @node Indent Tabs Mode, Keybindings, Mail Aliases, Emacs Initialization
17497 @section Indent Tabs Mode
17498 @cindex Tabs, preventing
17499 @findex indent-tabs-mode
17501 By default, Emacs inserts tabs in place of multiple spaces when it
17502 formats a region. (For example, you might indent many lines of text
17503 all at once with the @code{indent-region} command.) Tabs look fine on
17504 a terminal or with ordinary printing, but they produce badly indented
17505 output when you use @TeX{} or Texinfo since @TeX{} ignores tabs.
17508 The following turns off Indent Tabs mode:
17512 ;;; Prevent Extraneous Tabs
17513 (setq-default indent-tabs-mode nil)
17517 Note that this line uses @code{setq-default} rather than the
17518 @code{setq} command that we have seen before. The @code{setq-default}
17519 command sets values only in buffers that do not have their own local
17520 values for the variable.
17523 @xref{Just Spaces, , Tabs vs. Spaces, emacs, The GNU Emacs Manual}.
17525 @xref{File Variables, , Local Variables in Files, emacs, The GNU Emacs
17529 See sections ``Tabs vs.@: Spaces'' and ``Local Variables in
17530 Files'' in @cite{The GNU Emacs Manual}.
17534 @node Keybindings, Keymaps, Indent Tabs Mode, Emacs Initialization
17535 @section Some Keybindings
17537 Now for some personal keybindings:
17541 ;;; Compare windows
17542 (global-set-key "\C-cw" 'compare-windows)
17546 @findex compare-windows
17547 @code{compare-windows} is a nifty command that compares the text in
17548 your current window with text in the next window. It makes the
17549 comparison by starting at point in each window, moving over text in
17550 each window as far as they match. I use this command all the time.
17552 This also shows how to set a key globally, for all modes.
17554 @cindex Setting a key globally
17555 @cindex Global set key
17556 @cindex Key setting globally
17557 @findex global-set-key
17558 The command is @code{global-set-key}. It is followed by the
17559 keybinding. In a @file{.emacs} file, the keybinding is written as
17560 shown: @code{\C-c} stands for `control-c', which means `press the
17561 control key and the @key{c} key at the same time'. The @code{w} means
17562 `press the @key{w} key'. The keybinding is surrounded by double
17563 quotation marks. In documentation, you would write this as
17564 @w{@kbd{C-c w}}. (If you were binding a @key{META} key, such as
17565 @kbd{M-c}, rather than a @key{CTRL} key, you would write
17566 @w{@code{\M-c}} in your @file{.emacs} file. @xref{Init Rebinding, ,
17567 Rebinding Keys in Your Init File, emacs, The GNU Emacs Manual}, for
17570 The command invoked by the keys is @code{compare-windows}. Note that
17571 @code{compare-windows} is preceded by a single quote; otherwise, Emacs
17572 would first try to evaluate the symbol to determine its value.
17574 These three things, the double quotation marks, the backslash before
17575 the @samp{C}, and the single quote mark are necessary parts of
17576 keybinding that I tend to forget. Fortunately, I have come to
17577 remember that I should look at my existing @file{.emacs} file, and
17578 adapt what is there.
17580 As for the keybinding itself: @kbd{C-c w}. This combines the prefix
17581 key, @kbd{C-c}, with a single character, in this case, @kbd{w}. This
17582 set of keys, @kbd{C-c} followed by a single character, is strictly
17583 reserved for individuals' own use. (I call these `own' keys, since
17584 these are for my own use.) You should always be able to create such a
17585 keybinding for your own use without stomping on someone else's
17586 keybinding. If you ever write an extension to Emacs, please avoid
17587 taking any of these keys for public use. Create a key like @kbd{C-c
17588 C-w} instead. Otherwise, we will run out of `own' keys.
17591 Here is another keybinding, with a comment:
17595 ;;; Keybinding for `occur'
17596 ; I use occur a lot, so let's bind it to a key:
17597 (global-set-key "\C-co" 'occur)
17602 The @code{occur} command shows all the lines in the current buffer
17603 that contain a match for a regular expression. Matching lines are
17604 shown in a buffer called @file{*Occur*}. That buffer serves as a menu
17605 to jump to occurrences.
17607 @findex global-unset-key
17608 @cindex Unbinding key
17609 @cindex Key unbinding
17611 Here is how to unbind a key, so it does not
17617 (global-unset-key "\C-xf")
17621 There is a reason for this unbinding: I found I inadvertently typed
17622 @w{@kbd{C-x f}} when I meant to type @kbd{C-x C-f}. Rather than find a
17623 file, as I intended, I accidentally set the width for filled text,
17624 almost always to a width I did not want. Since I hardly ever reset my
17625 default width, I simply unbound the key.
17627 @findex list-buffers, @r{rebound}
17628 @findex buffer-menu, @r{bound to key}
17630 The following rebinds an existing key:
17634 ;;; Rebind `C-x C-b' for `buffer-menu'
17635 (global-set-key "\C-x\C-b" 'buffer-menu)
17639 By default, @kbd{C-x C-b} runs the
17640 @code{list-buffers} command. This command lists
17641 your buffers in @emph{another} window. Since I
17642 almost always want to do something in that
17643 window, I prefer the @code{buffer-menu}
17644 command, which not only lists the buffers,
17645 but moves point into that window.
17647 @node Keymaps, Loading Files, Keybindings, Emacs Initialization
17650 @cindex Rebinding keys
17652 Emacs uses @dfn{keymaps} to record which keys call which commands.
17653 When you use @code{global-set-key} to set the keybinding for a single
17654 command in all parts of Emacs, you are specifying the keybinding in
17655 @code{current-global-map}.
17657 Specific modes, such as C mode or Text mode, have their own keymaps;
17658 the mode-specific keymaps override the global map that is shared by
17661 The @code{global-set-key} function binds, or rebinds, the global
17662 keymap. For example, the following binds the key @kbd{C-x C-b} to the
17663 function @code{buffer-menu}:
17666 (global-set-key "\C-x\C-b" 'buffer-menu)
17669 Mode-specific keymaps are bound using the @code{define-key} function,
17670 which takes a specific keymap as an argument, as well as the key and
17671 the command. For example, my @file{.emacs} file contains the
17672 following expression to bind the @code{texinfo-insert-@@group} command
17673 to @kbd{C-c C-c g}:
17677 (define-key texinfo-mode-map "\C-c\C-cg" 'texinfo-insert-@@group)
17682 The @code{texinfo-insert-@@group} function itself is a little extension
17683 to Texinfo mode that inserts @samp{@@group} into a Texinfo file. I
17684 use this command all the time and prefer to type the three strokes
17685 @kbd{C-c C-c g} rather than the six strokes @kbd{@@ g r o u p}.
17686 (@samp{@@group} and its matching @samp{@@end group} are commands that
17687 keep all enclosed text together on one page; many multi-line examples
17688 in this book are surrounded by @samp{@@group @dots{} @@end group}.)
17691 Here is the @code{texinfo-insert-@@group} function definition:
17695 (defun texinfo-insert-@@group ()
17696 "Insert the string @@group in a Texinfo buffer."
17698 (beginning-of-line)
17699 (insert "@@group\n"))
17703 (Of course, I could have used Abbrev mode to save typing, rather than
17704 write a function to insert a word; but I prefer key strokes consistent
17705 with other Texinfo mode key bindings.)
17707 You will see numerous @code{define-key} expressions in
17708 @file{loaddefs.el} as well as in the various mode libraries, such as
17709 @file{cc-mode.el} and @file{lisp-mode.el}.
17711 @xref{Key Bindings, , Customizing Key Bindings, emacs, The GNU Emacs
17712 Manual}, and @ref{Keymaps, , Keymaps, elisp, The GNU Emacs Lisp
17713 Reference Manual}, for more information about keymaps.
17715 @node Loading Files, Autoload, Keymaps, Emacs Initialization
17716 @section Loading Files
17717 @cindex Loading files
17720 Many people in the GNU Emacs community have written extensions to
17721 Emacs. As time goes by, these extensions are often included in new
17722 releases. For example, the Calendar and Diary packages are now part
17723 of the standard GNU Emacs, as is Calc.
17725 You can use a @code{load} command to evaluate a complete file and
17726 thereby install all the functions and variables in the file into Emacs.
17729 @c (auto-compression-mode t)
17732 (load "~/emacs/slowsplit")
17735 This evaluates, i.e.@: loads, the @file{slowsplit.el} file or if it
17736 exists, the faster, byte compiled @file{slowsplit.elc} file from the
17737 @file{emacs} sub-directory of your home directory. The file contains
17738 the function @code{split-window-quietly}, which John Robinson wrote in
17741 The @code{split-window-quietly} function splits a window with the
17742 minimum of redisplay. I installed it in 1989 because it worked well
17743 with the slow 1200 baud terminals I was then using. Nowadays, I only
17744 occasionally come across such a slow connection, but I continue to use
17745 the function because I like the way it leaves the bottom half of a
17746 buffer in the lower of the new windows and the top half in the upper
17750 To replace the key binding for the default
17751 @code{split-window-vertically}, you must also unset that key and bind
17752 the keys to @code{split-window-quietly}, like this:
17756 (global-unset-key "\C-x2")
17757 (global-set-key "\C-x2" 'split-window-quietly)
17762 If you load many extensions, as I do, then instead of specifying the
17763 exact location of the extension file, as shown above, you can specify
17764 that directory as part of Emacs' @code{load-path}. Then, when Emacs
17765 loads a file, it will search that directory as well as its default
17766 list of directories. (The default list is specified in @file{paths.h}
17767 when Emacs is built.)
17770 The following command adds your @file{~/emacs} directory to the
17771 existing load path:
17775 ;;; Emacs Load Path
17776 (setq load-path (cons "~/emacs" load-path))
17780 Incidentally, @code{load-library} is an interactive interface to the
17781 @code{load} function. The complete function looks like this:
17783 @findex load-library
17786 (defun load-library (library)
17787 "Load the library named LIBRARY.
17788 This is an interface to the function `load'."
17790 (list (completing-read "Load library: "
17791 (apply-partially 'locate-file-completion-table
17793 (get-load-suffixes)))))
17798 The name of the function, @code{load-library}, comes from the use of
17799 `library' as a conventional synonym for `file'. The source for the
17800 @code{load-library} command is in the @file{files.el} library.
17802 Another interactive command that does a slightly different job is
17803 @code{load-file}. @xref{Lisp Libraries, , Libraries of Lisp Code for
17804 Emacs, emacs, The GNU Emacs Manual}, for information on the
17805 distinction between @code{load-library} and this command.
17807 @node Autoload, Simple Extension, Loading Files, Emacs Initialization
17808 @section Autoloading
17811 Instead of installing a function by loading the file that contains it,
17812 or by evaluating the function definition, you can make the function
17813 available but not actually install it until it is first called. This
17814 is called @dfn{autoloading}.
17816 When you execute an autoloaded function, Emacs automatically evaluates
17817 the file that contains the definition, and then calls the function.
17819 Emacs starts quicker with autoloaded functions, since their libraries
17820 are not loaded right away; but you need to wait a moment when you
17821 first use such a function, while its containing file is evaluated.
17823 Rarely used functions are frequently autoloaded. The
17824 @file{loaddefs.el} library contains hundreds of autoloaded functions,
17825 from @code{bookmark-set} to @code{wordstar-mode}. Of course, you may
17826 come to use a `rare' function frequently. When you do, you should
17827 load that function's file with a @code{load} expression in your
17828 @file{.emacs} file.
17830 In my @file{.emacs} file, I load 14 libraries that contain functions
17831 that would otherwise be autoloaded. (Actually, it would have been
17832 better to include these files in my `dumped' Emacs, but I forgot.
17833 @xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
17834 Reference Manual}, and the @file{INSTALL} file for more about
17837 You may also want to include autoloaded expressions in your @file{.emacs}
17838 file. @code{autoload} is a built-in function that takes up to five
17839 arguments, the final three of which are optional. The first argument
17840 is the name of the function to be autoloaded; the second is the name
17841 of the file to be loaded. The third argument is documentation for the
17842 function, and the fourth tells whether the function can be called
17843 interactively. The fifth argument tells what type of
17844 object---@code{autoload} can handle a keymap or macro as well as a
17845 function (the default is a function).
17848 Here is a typical example:
17852 (autoload 'html-helper-mode
17853 "html-helper-mode" "Edit HTML documents" t)
17858 (@code{html-helper-mode} is an older alternative to @code{html-mode},
17859 which is a standard part of the distribution.)
17862 This expression autoloads the @code{html-helper-mode} function. It
17863 takes it from the @file{html-helper-mode.el} file (or from the byte
17864 compiled version @file{html-helper-mode.elc}, if that exists.) The
17865 file must be located in a directory specified by @code{load-path}.
17866 The documentation says that this is a mode to help you edit documents
17867 written in the HyperText Markup Language. You can call this mode
17868 interactively by typing @kbd{M-x html-helper-mode}. (You need to
17869 duplicate the function's regular documentation in the autoload
17870 expression because the regular function is not yet loaded, so its
17871 documentation is not available.)
17873 @xref{Autoload, , Autoload, elisp, The GNU Emacs Lisp Reference
17874 Manual}, for more information.
17876 @node Simple Extension, X11 Colors, Autoload, Emacs Initialization
17877 @section A Simple Extension: @code{line-to-top-of-window}
17878 @findex line-to-top-of-window
17879 @cindex Simple extension in @file{.emacs} file
17881 Here is a simple extension to Emacs that moves the line point is on to
17882 the top of the window. I use this all the time, to make text easier
17885 You can put the following code into a separate file and then load it
17886 from your @file{.emacs} file, or you can include it within your
17887 @file{.emacs} file.
17890 Here is the definition:
17894 ;;; Line to top of window;
17895 ;;; replace three keystroke sequence C-u 0 C-l
17896 (defun line-to-top-of-window ()
17897 "Move the line point is on to top of window."
17904 Now for the keybinding.
17906 Nowadays, function keys as well as mouse button events and
17907 non-@sc{ascii} characters are written within square brackets, without
17908 quotation marks. (In Emacs version 18 and before, you had to write
17909 different function key bindings for each different make of terminal.)
17911 I bind @code{line-to-top-of-window} to my @key{F6} function key like
17915 (global-set-key [f6] 'line-to-top-of-window)
17918 For more information, see @ref{Init Rebinding, , Rebinding Keys in
17919 Your Init File, emacs, The GNU Emacs Manual}.
17921 @cindex Conditional 'twixt two versions of Emacs
17922 @cindex Version of Emacs, choosing
17923 @cindex Emacs version, choosing
17924 If you run two versions of GNU Emacs, such as versions 22 and 23, and
17925 use one @file{.emacs} file, you can select which code to evaluate with
17926 the following conditional:
17931 ((= 22 emacs-major-version)
17932 ;; evaluate version 22 code
17934 ((= 23 emacs-major-version)
17935 ;; evaluate version 23 code
17940 For example, in contrast to version 20, more recent versions blink
17941 their cursors by default. I hate such blinking, as well as other
17942 features, so I placed the following in my @file{.emacs}
17943 file@footnote{When I start instances of Emacs that do not load my
17944 @file{.emacs} file or any site file, I also turn off blinking:
17947 emacs -q --no-site-file -eval '(blink-cursor-mode nil)'
17949 @exdent Or nowadays, using an even more sophisticated set of options,
17957 (when (>= emacs-major-version 21)
17958 (blink-cursor-mode 0)
17959 ;; Insert newline when you press `C-n' (next-line)
17960 ;; at the end of the buffer
17961 (setq next-line-add-newlines t)
17964 ;; Turn on image viewing
17965 (auto-image-file-mode t)
17968 ;; Turn on menu bar (this bar has text)
17969 ;; (Use numeric argument to turn on)
17973 ;; Turn off tool bar (this bar has icons)
17974 ;; (Use numeric argument to turn on)
17975 (tool-bar-mode nil)
17978 ;; Turn off tooltip mode for tool bar
17979 ;; (This mode causes icon explanations to pop up)
17980 ;; (Use numeric argument to turn on)
17982 ;; If tooltips turned on, make tips appear promptly
17983 (setq tooltip-delay 0.1) ; default is 0.7 second
17988 @node X11 Colors, Miscellaneous, Simple Extension, Emacs Initialization
17989 @section X11 Colors
17991 You can specify colors when you use Emacs with the MIT X Windowing
17994 I dislike the default colors and specify my own.
17997 Here are the expressions in my @file{.emacs}
17998 file that set values:
18002 ;; Set cursor color
18003 (set-cursor-color "white")
18006 (set-mouse-color "white")
18008 ;; Set foreground and background
18009 (set-foreground-color "white")
18010 (set-background-color "darkblue")
18014 ;;; Set highlighting colors for isearch and drag
18015 (set-face-foreground 'highlight "white")
18016 (set-face-background 'highlight "blue")
18020 (set-face-foreground 'region "cyan")
18021 (set-face-background 'region "blue")
18025 (set-face-foreground 'secondary-selection "skyblue")
18026 (set-face-background 'secondary-selection "darkblue")
18030 ;; Set calendar highlighting colors
18031 (setq calendar-load-hook
18033 (set-face-foreground 'diary-face "skyblue")
18034 (set-face-background 'holiday-face "slate blue")
18035 (set-face-foreground 'holiday-face "white")))
18039 The various shades of blue soothe my eye and prevent me from seeing
18040 the screen flicker.
18042 Alternatively, I could have set my specifications in various X
18043 initialization files. For example, I could set the foreground,
18044 background, cursor, and pointer (i.e., mouse) colors in my
18045 @file{~/.Xresources} file like this:
18049 Emacs*foreground: white
18050 Emacs*background: darkblue
18051 Emacs*cursorColor: white
18052 Emacs*pointerColor: white
18056 In any event, since it is not part of Emacs, I set the root color of
18057 my X window in my @file{~/.xinitrc} file, like this@footnote{I also
18058 run more modern window managers, such as Enlightenment, Gnome, or KDE;
18059 in those cases, I often specify an image rather than a plain color.}:
18062 xsetroot -solid Navy -fg white &
18066 @node Miscellaneous, Mode Line, X11 Colors, Emacs Initialization
18067 @section Miscellaneous Settings for a @file{.emacs} File
18070 Here are a few miscellaneous settings:
18075 Set the shape and color of the mouse cursor:
18079 ; Cursor shapes are defined in
18080 ; `/usr/include/X11/cursorfont.h';
18081 ; for example, the `target' cursor is number 128;
18082 ; the `top_left_arrow' cursor is number 132.
18086 (let ((mpointer (x-get-resource "*mpointer"
18087 "*emacs*mpointer")))
18088 ;; If you have not set your mouse pointer
18089 ;; then set it, otherwise leave as is:
18090 (if (eq mpointer nil)
18091 (setq mpointer "132")) ; top_left_arrow
18094 (setq x-pointer-shape (string-to-int mpointer))
18095 (set-mouse-color "white"))
18100 Or you can set the values of a variety of features in an alist, like
18106 default-frame-alist
18107 '((cursor-color . "white")
18108 (mouse-color . "white")
18109 (foreground-color . "white")
18110 (background-color . "DodgerBlue4")
18111 ;; (cursor-type . bar)
18112 (cursor-type . box)
18115 (tool-bar-lines . 0)
18116 (menu-bar-lines . 1)
18120 "-Misc-Fixed-Medium-R-Normal--20-200-75-75-C-100-ISO8859-1")
18126 Convert @kbd{@key{CTRL}-h} into @key{DEL} and @key{DEL}
18127 into @kbd{@key{CTRL}-h}.@*
18128 (Some older keyboards needed this, although I have not seen the
18133 ;; Translate `C-h' to <DEL>.
18134 ; (keyboard-translate ?\C-h ?\C-?)
18136 ;; Translate <DEL> to `C-h'.
18137 (keyboard-translate ?\C-? ?\C-h)
18141 @item Turn off a blinking cursor!
18145 (if (fboundp 'blink-cursor-mode)
18146 (blink-cursor-mode -1))
18151 or start GNU Emacs with the command @code{emacs -nbc}.
18154 @item When using `grep'@*
18155 @samp{-i}@w{ } Ignore case distinctions@*
18156 @samp{-n}@w{ } Prefix each line of output with line number@*
18157 @samp{-H}@w{ } Print the filename for each match.@*
18158 @samp{-e}@w{ } Protect patterns beginning with a hyphen character, @samp{-}
18161 (setq grep-command "grep -i -nH -e ")
18165 @c Evidently, no longer needed in GNU Emacs 22
18167 item Automatically uncompress compressed files when visiting them
18170 (load "uncompress")
18175 @item Find an existing buffer, even if it has a different name@*
18176 This avoids problems with symbolic links.
18179 (setq find-file-existing-other-name t)
18182 @item Set your language environment and default input method
18186 (set-language-environment "latin-1")
18187 ;; Remember you can enable or disable multilingual text input
18188 ;; with the @code{toggle-input-method'} (@kbd{C-\}) command
18189 (setq default-input-method "latin-1-prefix")
18193 If you want to write with Chinese `GB' characters, set this instead:
18197 (set-language-environment "Chinese-GB")
18198 (setq default-input-method "chinese-tonepy")
18203 @subsubheading Fixing Unpleasant Key Bindings
18204 @cindex Key bindings, fixing
18205 @cindex Bindings, key, fixing unpleasant
18207 Some systems bind keys unpleasantly. Sometimes, for example, the
18208 @key{CTRL} key appears in an awkward spot rather than at the far left
18211 Usually, when people fix these sorts of keybindings, they do not
18212 change their @file{~/.emacs} file. Instead, they bind the proper keys
18213 on their consoles with the @code{loadkeys} or @code{install-keymap}
18214 commands in their boot script and then include @code{xmodmap} commands
18215 in their @file{.xinitrc} or @file{.Xsession} file for X Windows.
18223 loadkeys /usr/share/keymaps/i386/qwerty/emacs2.kmap.gz
18225 install-keymap emacs2
18231 For a @file{.xinitrc} or @file{.Xsession} file when the @key{Caps
18232 Lock} key is at the far left of the home row:
18236 # Bind the key labeled `Caps Lock' to `Control'
18237 # (Such a broken user interface suggests that keyboard manufacturers
18238 # think that computers are typewriters from 1885.)
18240 xmodmap -e "clear Lock"
18241 xmodmap -e "add Control = Caps_Lock"
18247 In a @file{.xinitrc} or @file{.Xsession} file, to convert an @key{ALT}
18248 key to a @key{META} key:
18252 # Some ill designed keyboards have a key labeled ALT and no Meta
18253 xmodmap -e "keysym Alt_L = Meta_L Alt_L"
18258 @node Mode Line, , Miscellaneous, Emacs Initialization
18259 @section A Modified Mode Line
18260 @vindex default-mode-line-format
18261 @cindex Mode line format
18263 Finally, a feature I really like: a modified mode line.
18265 When I work over a network, I forget which machine I am using. Also,
18266 I tend to I lose track of where I am, and which line point is on.
18268 So I reset my mode line to look like this:
18271 -:-- foo.texi rattlesnake:/home/bob/ Line 1 (Texinfo Fill) Top
18274 I am visiting a file called @file{foo.texi}, on my machine
18275 @file{rattlesnake} in my @file{/home/bob} buffer. I am on line 1, in
18276 Texinfo mode, and am at the top of the buffer.
18279 My @file{.emacs} file has a section that looks like this:
18283 ;; Set a Mode Line that tells me which machine, which directory,
18284 ;; and which line I am on, plus the other customary information.
18285 (setq default-mode-line-format
18289 "mouse-1: select window, mouse-2: delete others ..."))
18290 mode-line-mule-info
18292 mode-line-frame-identification
18296 mode-line-buffer-identification
18299 (system-name) 0 (string-match "\\..+" (system-name))))
18304 "mouse-1: select window, mouse-2: delete others ..."))
18305 (line-number-mode " Line %l ")
18311 "mouse-1: select window, mouse-2: delete others ..."))
18312 (:eval (mode-line-mode-name))
18315 #("%n" 0 2 (help-echo "mouse-2: widen" local-map (keymap ...)))
18324 Here, I redefine the default mode line. Most of the parts are from
18325 the original; but I make a few changes. I set the @emph{default} mode
18326 line format so as to permit various modes, such as Info, to override
18329 Many elements in the list are self-explanatory:
18330 @code{mode-line-modified} is a variable that tells whether the buffer
18331 has been modified, @code{mode-name} tells the name of the mode, and so
18332 on. However, the format looks complicated because of two features we
18333 have not discussed.
18335 @cindex Properties, in mode line example
18336 The first string in the mode line is a dash or hyphen, @samp{-}. In
18337 the old days, it would have been specified simply as @code{"-"}. But
18338 nowadays, Emacs can add properties to a string, such as highlighting
18339 or, as in this case, a help feature. If you place your mouse cursor
18340 over the hyphen, some help information appears (By default, you must
18341 wait seven-tenths of a second before the information appears. You can
18342 change that timing by changing the value of @code{tooltip-delay}.)
18345 The new string format has a special syntax:
18348 #("-" 0 1 (help-echo "mouse-1: select window, ..."))
18352 The @code{#(} begins a list. The first element of the list is the
18353 string itself, just one @samp{-}. The second and third
18354 elements specify the range over which the fourth element applies. A
18355 range starts @emph{after} a character, so a zero means the range
18356 starts just before the first character; a 1 means that the range ends
18357 just after the first character. The third element is the property for
18358 the range. It consists of a property list, a
18359 property name, in this case, @samp{help-echo}, followed by a value, in this
18360 case, a string. The second, third, and fourth elements of this new
18361 string format can be repeated.
18363 @xref{Text Properties, , Text Properties, elisp, The GNU Emacs Lisp
18364 Reference Manual}, and see @ref{Mode Line Format, , Mode Line Format,
18365 elisp, The GNU Emacs Lisp Reference Manual}, for more information.
18367 @code{mode-line-buffer-identification}
18368 displays the current buffer name. It is a list
18369 beginning @code{(#("%12b" 0 4 @dots{}}.
18370 The @code{#(} begins the list.
18372 The @samp{"%12b"} displays the current buffer name, using the
18373 @code{buffer-name} function with which we are familiar; the `12'
18374 specifies the maximum number of characters that will be displayed.
18375 When a name has fewer characters, whitespace is added to fill out to
18376 this number. (Buffer names can and often should be longer than 12
18377 characters; this length works well in a typical 80 column wide
18380 @code{:eval} says to evaluate the following form and use the result as
18381 a string to display. In this case, the expression displays the first
18382 component of the full system name. The end of the first component is
18383 a @samp{.} (`period'), so I use the @code{string-match} function to
18384 tell me the length of the first component. The substring from the
18385 zeroth character to that length is the name of the machine.
18388 This is the expression:
18393 (system-name) 0 (string-match "\\..+" (system-name))))
18397 @samp{%[} and @samp{%]} cause a pair of square brackets
18398 to appear for each recursive editing level. @samp{%n} says `Narrow'
18399 when narrowing is in effect. @samp{%P} tells you the percentage of
18400 the buffer that is above the bottom of the window, or `Top', `Bottom',
18401 or `All'. (A lower case @samp{p} tell you the percentage above the
18402 @emph{top} of the window.) @samp{%-} inserts enough dashes to fill
18405 Remember, ``You don't have to like Emacs to like it'' --- your own
18406 Emacs can have different colors, different commands, and different
18407 keys than a default Emacs.
18409 On the other hand, if you want to bring up a plain `out of the box'
18410 Emacs, with no customization, type:
18417 This will start an Emacs that does @emph{not} load your
18418 @file{~/.emacs} initialization file. A plain, default Emacs. Nothing
18421 @node Debugging, Conclusion, Emacs Initialization, Top
18425 GNU Emacs has two debuggers, @code{debug} and @code{edebug}. The
18426 first is built into the internals of Emacs and is always with you;
18427 the second requires that you instrument a function before you can use it.
18429 Both debuggers are described extensively in @ref{Debugging, ,
18430 Debugging Lisp Programs, elisp, The GNU Emacs Lisp Reference Manual}.
18431 In this chapter, I will walk through a short example of each.
18434 * debug:: How to use the built-in debugger.
18435 * debug-on-entry:: Start debugging when you call a function.
18436 * debug-on-quit:: Start debugging when you quit with @kbd{C-g}.
18437 * edebug:: How to use Edebug, a source level debugger.
18438 * Debugging Exercises::
18441 @node debug, debug-on-entry, Debugging, Debugging
18442 @section @code{debug}
18445 Suppose you have written a function definition that is intended to
18446 return the sum of the numbers 1 through a given number. (This is the
18447 @code{triangle} function discussed earlier. @xref{Decrementing
18448 Example, , Example with Decrementing Counter}, for a discussion.)
18449 @c xref{Decrementing Loop,, Loop with a Decrementing Counter}, for a discussion.)
18451 However, your function definition has a bug. You have mistyped
18452 @samp{1=} for @samp{1-}. Here is the broken definition:
18454 @findex triangle-bugged
18457 (defun triangle-bugged (number)
18458 "Return sum of numbers 1 through NUMBER inclusive."
18460 (while (> number 0)
18461 (setq total (+ total number))
18462 (setq number (1= number))) ; @r{Error here.}
18467 If you are reading this in Info, you can evaluate this definition in
18468 the normal fashion. You will see @code{triangle-bugged} appear in the
18472 Now evaluate the @code{triangle-bugged} function with an
18476 (triangle-bugged 4)
18480 In a recent GNU Emacs, you will create and enter a @file{*Backtrace*}
18486 ---------- Buffer: *Backtrace* ----------
18487 Debugger entered--Lisp error: (void-function 1=)
18489 (setq number (1= number))
18490 (while (> number 0) (setq total (+ total number))
18491 (setq number (1= number)))
18492 (let ((total 0)) (while (> number 0) (setq total ...)
18493 (setq number ...)) total)
18497 eval((triangle-bugged 4))
18498 eval-last-sexp-1(nil)
18499 eval-last-sexp(nil)
18500 call-interactively(eval-last-sexp)
18501 ---------- Buffer: *Backtrace* ----------
18506 (I have reformatted this example slightly; the debugger does not fold
18507 long lines. As usual, you can quit the debugger by typing @kbd{q} in
18508 the @file{*Backtrace*} buffer.)
18510 In practice, for a bug as simple as this, the `Lisp error' line will
18511 tell you what you need to know to correct the definition. The
18512 function @code{1=} is `void'.
18516 In GNU Emacs 20 and before, you will see:
18519 Symbol's function definition is void:@: 1=
18523 which has the same meaning as the @file{*Backtrace*} buffer line in
18527 However, suppose you are not quite certain what is going on?
18528 You can read the complete backtrace.
18530 In this case, you need to run a recent GNU Emacs, which automatically
18531 starts the debugger that puts you in the @file{*Backtrace*} buffer; or
18532 else, you need to start the debugger manually as described below.
18534 Read the @file{*Backtrace*} buffer from the bottom up; it tells you
18535 what Emacs did that led to the error. Emacs made an interactive call
18536 to @kbd{C-x C-e} (@code{eval-last-sexp}), which led to the evaluation
18537 of the @code{triangle-bugged} expression. Each line above tells you
18538 what the Lisp interpreter evaluated next.
18541 The third line from the top of the buffer is
18544 (setq number (1= number))
18548 Emacs tried to evaluate this expression; in order to do so, it tried
18549 to evaluate the inner expression shown on the second line from the
18558 This is where the error occurred; as the top line says:
18561 Debugger entered--Lisp error: (void-function 1=)
18565 You can correct the mistake, re-evaluate the function definition, and
18566 then run your test again.
18568 @node debug-on-entry, debug-on-quit, debug, Debugging
18569 @section @code{debug-on-entry}
18570 @findex debug-on-entry
18572 A recent GNU Emacs starts the debugger automatically when your
18573 function has an error.
18576 GNU Emacs version 20 and before did not; it simply
18577 presented you with an error message. You had to start the debugger
18581 Incidentally, you can start the debugger manually for all versions of
18582 Emacs; the advantage is that the debugger runs even if you do not have
18583 a bug in your code. Sometimes your code will be free of bugs!
18585 You can enter the debugger when you call the function by calling
18586 @code{debug-on-entry}.
18593 M-x debug-on-entry RET triangle-bugged RET
18598 Now, evaluate the following:
18601 (triangle-bugged 5)
18605 All versions of Emacs will create a @file{*Backtrace*} buffer and tell
18606 you that it is beginning to evaluate the @code{triangle-bugged}
18611 ---------- Buffer: *Backtrace* ----------
18612 Debugger entered--entering a function:
18613 * triangle-bugged(5)
18614 eval((triangle-bugged 5))
18617 eval-last-sexp-1(nil)
18618 eval-last-sexp(nil)
18619 call-interactively(eval-last-sexp)
18620 ---------- Buffer: *Backtrace* ----------
18624 In the @file{*Backtrace*} buffer, type @kbd{d}. Emacs will evaluate
18625 the first expression in @code{triangle-bugged}; the buffer will look
18630 ---------- Buffer: *Backtrace* ----------
18631 Debugger entered--beginning evaluation of function call form:
18632 * (let ((total 0)) (while (> number 0) (setq total ...)
18633 (setq number ...)) total)
18634 * triangle-bugged(5)
18635 eval((triangle-bugged 5))
18638 eval-last-sexp-1(nil)
18639 eval-last-sexp(nil)
18640 call-interactively(eval-last-sexp)
18641 ---------- Buffer: *Backtrace* ----------
18646 Now, type @kbd{d} again, eight times, slowly. Each time you type
18647 @kbd{d}, Emacs will evaluate another expression in the function
18651 Eventually, the buffer will look like this:
18655 ---------- Buffer: *Backtrace* ----------
18656 Debugger entered--beginning evaluation of function call form:
18657 * (setq number (1= number))
18658 * (while (> number 0) (setq total (+ total number))
18659 (setq number (1= number)))
18662 * (let ((total 0)) (while (> number 0) (setq total ...)
18663 (setq number ...)) total)
18664 * triangle-bugged(5)
18665 eval((triangle-bugged 5))
18668 eval-last-sexp-1(nil)
18669 eval-last-sexp(nil)
18670 call-interactively(eval-last-sexp)
18671 ---------- Buffer: *Backtrace* ----------
18677 Finally, after you type @kbd{d} two more times, Emacs will reach the
18678 error, and the top two lines of the @file{*Backtrace*} buffer will look
18683 ---------- Buffer: *Backtrace* ----------
18684 Debugger entered--Lisp error: (void-function 1=)
18687 ---------- Buffer: *Backtrace* ----------
18691 By typing @kbd{d}, you were able to step through the function.
18693 You can quit a @file{*Backtrace*} buffer by typing @kbd{q} in it; this
18694 quits the trace, but does not cancel @code{debug-on-entry}.
18696 @findex cancel-debug-on-entry
18697 To cancel the effect of @code{debug-on-entry}, call
18698 @code{cancel-debug-on-entry} and the name of the function, like this:
18701 M-x cancel-debug-on-entry RET triangle-bugged RET
18705 (If you are reading this in Info, cancel @code{debug-on-entry} now.)
18707 @node debug-on-quit, edebug, debug-on-entry, Debugging
18708 @section @code{debug-on-quit} and @code{(debug)}
18710 In addition to setting @code{debug-on-error} or calling @code{debug-on-entry},
18711 there are two other ways to start @code{debug}.
18713 @findex debug-on-quit
18714 You can start @code{debug} whenever you type @kbd{C-g}
18715 (@code{keyboard-quit}) by setting the variable @code{debug-on-quit} to
18716 @code{t}. This is useful for debugging infinite loops.
18719 @cindex @code{(debug)} in code
18720 Or, you can insert a line that says @code{(debug)} into your code
18721 where you want the debugger to start, like this:
18725 (defun triangle-bugged (number)
18726 "Return sum of numbers 1 through NUMBER inclusive."
18728 (while (> number 0)
18729 (setq total (+ total number))
18730 (debug) ; @r{Start debugger.}
18731 (setq number (1= number))) ; @r{Error here.}
18736 The @code{debug} function is described in detail in @ref{Debugger, ,
18737 The Lisp Debugger, elisp, The GNU Emacs Lisp Reference Manual}.
18739 @node edebug, Debugging Exercises, debug-on-quit, Debugging
18740 @section The @code{edebug} Source Level Debugger
18741 @cindex Source level debugger
18744 Edebug is a source level debugger. Edebug normally displays the
18745 source of the code you are debugging, with an arrow at the left that
18746 shows which line you are currently executing.
18748 You can walk through the execution of a function, line by line, or run
18749 quickly until reaching a @dfn{breakpoint} where execution stops.
18751 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18752 Lisp Reference Manual}.
18755 Here is a bugged function definition for @code{triangle-recursively}.
18756 @xref{Recursive triangle function, , Recursion in place of a counter},
18757 for a review of it.
18761 (defun triangle-recursively-bugged (number)
18762 "Return sum of numbers 1 through NUMBER inclusive.
18767 (triangle-recursively-bugged
18768 (1= number))))) ; @r{Error here.}
18773 Normally, you would install this definition by positioning your cursor
18774 after the function's closing parenthesis and typing @kbd{C-x C-e}
18775 (@code{eval-last-sexp}) or else by positioning your cursor within the
18776 definition and typing @kbd{C-M-x} (@code{eval-defun}). (By default,
18777 the @code{eval-defun} command works only in Emacs Lisp mode or in Lisp
18781 However, to prepare this function definition for Edebug, you must
18782 first @dfn{instrument} the code using a different command. You can do
18783 this by positioning your cursor within or just after the definition
18787 M-x edebug-defun RET
18791 This will cause Emacs to load Edebug automatically if it is not
18792 already loaded, and properly instrument the function.
18794 After instrumenting the function, place your cursor after the
18795 following expression and type @kbd{C-x C-e} (@code{eval-last-sexp}):
18798 (triangle-recursively-bugged 3)
18802 You will be jumped back to the source for
18803 @code{triangle-recursively-bugged} and the cursor positioned at the
18804 beginning of the @code{if} line of the function. Also, you will see
18805 an arrowhead at the left hand side of that line. The arrowhead marks
18806 the line where the function is executing. (In the following examples,
18807 we show the arrowhead with @samp{=>}; in a windowing system, you may
18808 see the arrowhead as a solid triangle in the window `fringe'.)
18811 =>@point{}(if (= number 1)
18816 In the example, the location of point is displayed with a star,
18817 @samp{@point{}} (in Info, it is displayed as @samp{-!-}).
18820 In the example, the location of point is displayed as @samp{@point{}}
18821 (in a printed book, it is displayed with a five pointed star).
18824 If you now press @key{SPC}, point will move to the next expression to
18825 be executed; the line will look like this:
18828 =>(if @point{}(= number 1)
18832 As you continue to press @key{SPC}, point will move from expression to
18833 expression. At the same time, whenever an expression returns a value,
18834 that value will be displayed in the echo area. For example, after you
18835 move point past @code{number}, you will see the following:
18838 Result: 3 (#o3, #x3, ?\C-c)
18842 This means the value of @code{number} is 3, which is octal three,
18843 hexadecimal three, and @sc{ascii} `control-c' (the third letter of the
18844 alphabet, in case you need to know this information).
18846 You can continue moving through the code until you reach the line with
18847 the error. Before evaluation, that line looks like this:
18850 => @point{}(1= number))))) ; @r{Error here.}
18855 When you press @key{SPC} once again, you will produce an error message
18859 Symbol's function definition is void:@: 1=
18865 Press @kbd{q} to quit Edebug.
18867 To remove instrumentation from a function definition, simply
18868 re-evaluate it with a command that does not instrument it.
18869 For example, you could place your cursor after the definition's
18870 closing parenthesis and type @kbd{C-x C-e}.
18872 Edebug does a great deal more than walk with you through a function.
18873 You can set it so it races through on its own, stopping only at an
18874 error or at specified stopping points; you can cause it to display the
18875 changing values of various expressions; you can find out how many
18876 times a function is called, and more.
18878 Edebug is described in @ref{edebug, , Edebug, elisp, The GNU Emacs
18879 Lisp Reference Manual}.
18882 @node Debugging Exercises, , edebug, Debugging
18883 @section Debugging Exercises
18887 Install the @code{count-words-region} function and then cause it to
18888 enter the built-in debugger when you call it. Run the command on a
18889 region containing two words. You will need to press @kbd{d} a
18890 remarkable number of times. On your system, is a `hook' called after
18891 the command finishes? (For information on hooks, see @ref{Command
18892 Overview, , Command Loop Overview, elisp, The GNU Emacs Lisp Reference
18896 Copy @code{count-words-region} into the @file{*scratch*} buffer,
18897 instrument the function for Edebug, and walk through its execution.
18898 The function does not need to have a bug, although you can introduce
18899 one if you wish. If the function lacks a bug, the walk-through
18900 completes without problems.
18903 While running Edebug, type @kbd{?} to see a list of all the Edebug commands.
18904 (The @code{global-edebug-prefix} is usually @kbd{C-x X}, i.e.@:
18905 @kbd{@key{CTRL}-x} followed by an upper case @kbd{X}; use this prefix
18906 for commands made outside of the Edebug debugging buffer.)
18909 In the Edebug debugging buffer, use the @kbd{p}
18910 (@code{edebug-bounce-point}) command to see where in the region the
18911 @code{count-words-region} is working.
18914 Move point to some spot further down the function and then type the
18915 @kbd{h} (@code{edebug-goto-here}) command to jump to that location.
18918 Use the @kbd{t} (@code{edebug-trace-mode}) command to cause Edebug to
18919 walk through the function on its own; use an upper case @kbd{T} for
18920 @code{edebug-Trace-fast-mode}.
18923 Set a breakpoint, then run Edebug in Trace mode until it reaches the
18927 @node Conclusion, the-the, Debugging, Top
18928 @chapter Conclusion
18930 We have now reached the end of this Introduction. You have now
18931 learned enough about programming in Emacs Lisp to set values, to write
18932 simple @file{.emacs} files for yourself and your friends, and write
18933 simple customizations and extensions to Emacs.
18935 This is a place to stop. Or, if you wish, you can now go onward, and
18938 You have learned some of the basic nuts and bolts of programming. But
18939 only some. There are a great many more brackets and hinges that are
18940 easy to use that we have not touched.
18942 A path you can follow right now lies among the sources to GNU Emacs
18945 @cite{The GNU Emacs Lisp Reference Manual}.
18948 @ref{Top, , The GNU Emacs Lisp Reference Manual, elisp, The GNU
18949 Emacs Lisp Reference Manual}.
18952 The Emacs Lisp sources are an adventure. When you read the sources and
18953 come across a function or expression that is unfamiliar, you need to
18954 figure out or find out what it does.
18956 Go to the Reference Manual. It is a thorough, complete, and fairly
18957 easy-to-read description of Emacs Lisp. It is written not only for
18958 experts, but for people who know what you know. (The @cite{Reference
18959 Manual} comes with the standard GNU Emacs distribution. Like this
18960 introduction, it comes as a Texinfo source file, so you can read it
18961 on-line and as a typeset, printed book.)
18963 Go to the other on-line help that is part of GNU Emacs: the on-line
18964 documentation for all functions and variables, and @code{find-tags},
18965 the program that takes you to sources.
18967 Here is an example of how I explore the sources. Because of its name,
18968 @file{simple.el} is the file I looked at first, a long time ago. As
18969 it happens some of the functions in @file{simple.el} are complicated,
18970 or at least look complicated at first sight. The @code{open-line}
18971 function, for example, looks complicated.
18973 You may want to walk through this function slowly, as we did with the
18974 @code{forward-sentence} function. (@xref{forward-sentence, The
18975 @code{forward-sentence} function}.) Or you may want to skip that
18976 function and look at another, such as @code{split-line}. You don't
18977 need to read all the functions. According to
18978 @code{count-words-in-defun}, the @code{split-line} function contains
18979 102 words and symbols.
18981 Even though it is short, @code{split-line} contains expressions
18982 we have not studied: @code{skip-chars-forward}, @code{indent-to},
18983 @code{current-column} and @code{insert-and-inherit}.
18985 Consider the @code{skip-chars-forward} function. (It is part of the
18986 function definition for @code{back-to-indentation}, which is shown in
18987 @ref{Review, , Review}.)
18989 In GNU Emacs, you can find out more about @code{skip-chars-forward} by
18990 typing @kbd{C-h f} (@code{describe-function}) and the name of the
18991 function. This gives you the function documentation.
18993 You may be able to guess what is done by a well named function such as
18994 @code{indent-to}; or you can look it up, too. Incidentally, the
18995 @code{describe-function} function itself is in @file{help.el}; it is
18996 one of those long, but decipherable functions. You can look up
18997 @code{describe-function} using the @kbd{C-h f} command!
18999 In this instance, since the code is Lisp, the @file{*Help*} buffer
19000 contains the name of the library containing the function's source.
19001 You can put point over the name of the library and press the RET key,
19002 which in this situation is bound to @code{help-follow}, and be taken
19003 directly to the source, in the same way as @kbd{M-.}
19006 The definition for @code{describe-function} illustrates how to
19007 customize the @code{interactive} expression without using the standard
19008 character codes; and it shows how to create a temporary buffer.
19010 (The @code{indent-to} function is written in C rather than Emacs Lisp;
19011 it is a `built-in' function. @code{help-follow} takes you to its
19012 source as does @code{find-tag}, when properly set up.)
19014 You can look at a function's source using @code{find-tag}, which is
19015 bound to @kbd{M-.} Finally, you can find out what the Reference
19016 Manual has to say by visiting the manual in Info, and typing @kbd{i}
19017 (@code{Info-index}) and the name of the function, or by looking up the
19018 function in the index to a printed copy of the manual.
19020 Similarly, you can find out what is meant by
19021 @code{insert-and-inherit}.
19023 Other interesting source files include @file{paragraphs.el},
19024 @file{loaddefs.el}, and @file{loadup.el}. The @file{paragraphs.el}
19025 file includes short, easily understood functions as well as longer
19026 ones. The @file{loaddefs.el} file contains the many standard
19027 autoloads and many keymaps. I have never looked at it all; only at
19028 parts. @file{loadup.el} is the file that loads the standard parts of
19029 Emacs; it tells you a great deal about how Emacs is built.
19030 (@xref{Building Emacs, , Building Emacs, elisp, The GNU Emacs Lisp
19031 Reference Manual}, for more about building.)
19033 As I said, you have learned some nuts and bolts; however, and very
19034 importantly, we have hardly touched major aspects of programming; I
19035 have said nothing about how to sort information, except to use the
19036 predefined @code{sort} function; I have said nothing about how to store
19037 information, except to use variables and lists; I have said nothing
19038 about how to write programs that write programs. These are topics for
19039 another, and different kind of book, a different kind of learning.
19041 What you have done is learn enough for much practical work with GNU
19042 Emacs. What you have done is get started. This is the end of a
19045 @c ================ Appendix ================
19047 @node the-the, Kill Ring, Conclusion, Top
19048 @appendix The @code{the-the} Function
19050 @cindex Duplicated words function
19051 @cindex Words, duplicated
19053 Sometimes when you you write text, you duplicate words---as with ``you
19054 you'' near the beginning of this sentence. I find that most
19055 frequently, I duplicate ``the''; hence, I call the function for
19056 detecting duplicated words, @code{the-the}.
19059 As a first step, you could use the following regular expression to
19060 search for duplicates:
19063 \\(\\w+[ \t\n]+\\)\\1
19067 This regexp matches one or more word-constituent characters followed
19068 by one or more spaces, tabs, or newlines. However, it does not detect
19069 duplicated words on different lines, since the ending of the first
19070 word, the end of the line, is different from the ending of the second
19071 word, a space. (For more information about regular expressions, see
19072 @ref{Regexp Search, , Regular Expression Searches}, as well as
19073 @ref{Regexps, , Syntax of Regular Expressions, emacs, The GNU Emacs
19074 Manual}, and @ref{Regular Expressions, , Regular Expressions, elisp,
19075 The GNU Emacs Lisp Reference Manual}.)
19077 You might try searching just for duplicated word-constituent
19078 characters but that does not work since the pattern detects doubles
19079 such as the two occurrences of `th' in `with the'.
19081 Another possible regexp searches for word-constituent characters
19082 followed by non-word-constituent characters, reduplicated. Here,
19083 @w{@samp{\\w+}} matches one or more word-constituent characters and
19084 @w{@samp{\\W*}} matches zero or more non-word-constituent characters.
19087 \\(\\(\\w+\\)\\W*\\)\\1
19093 Here is the pattern that I use. It is not perfect, but good enough.
19094 @w{@samp{\\b}} matches the empty string, provided it is at the beginning
19095 or end of a word; @w{@samp{[^@@ \n\t]+}} matches one or more occurrences of
19096 any characters that are @emph{not} an @@-sign, space, newline, or tab.
19099 \\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b
19102 One can write more complicated expressions, but I found that this
19103 expression is good enough, so I use it.
19105 Here is the @code{the-the} function, as I include it in my
19106 @file{.emacs} file, along with a handy global key binding:
19111 "Search forward for for a duplicated word."
19113 (message "Searching for for duplicated words ...")
19117 ;; This regexp is not perfect
19118 ;; but is fairly good over all:
19119 (if (re-search-forward
19120 "\\b\\([^@@ \n\t]+\\)[ \n\t]+\\1\\b" nil 'move)
19121 (message "Found duplicated word.")
19122 (message "End of buffer")))
19126 ;; Bind `the-the' to C-c \
19127 (global-set-key "\C-c\\" 'the-the)
19136 one two two three four five
19141 You can substitute the other regular expressions shown above in the
19142 function definition and try each of them on this list.
19144 @node Kill Ring, Full Graph, the-the, Top
19145 @appendix Handling the Kill Ring
19146 @cindex Kill ring handling
19147 @cindex Handling the kill ring
19148 @cindex Ring, making a list like a
19150 The kill ring is a list that is transformed into a ring by the
19151 workings of the @code{current-kill} function. The @code{yank} and
19152 @code{yank-pop} commands use the @code{current-kill} function.
19154 This appendix describes the @code{current-kill} function as well as
19155 both the @code{yank} and the @code{yank-pop} commands, but first,
19156 consider the workings of the kill ring.
19159 * What the Kill Ring Does::
19161 * yank:: Paste a copy of a clipped element.
19162 * yank-pop:: Insert element pointed to.
19166 @node What the Kill Ring Does, current-kill, Kill Ring, Kill Ring
19168 @unnumberedsec What the Kill Ring Does
19172 The kill ring has a default maximum length of sixty items; this number
19173 is too large for an explanation. Instead, set it to four. Please
19174 evaluate the following:
19178 (setq old-kill-ring-max kill-ring-max)
19179 (setq kill-ring-max 4)
19184 Then, please copy each line of the following indented example into the
19185 kill ring. You may kill each line with @kbd{C-k} or mark it and copy
19189 (In a read-only buffer, such as the @file{*info*} buffer, the kill
19190 command, @kbd{C-k} (@code{kill-line}), will not remove the text,
19191 merely copy it to the kill ring. However, your machine may beep at
19192 you. Alternatively, for silence, you may copy the region of each line
19193 with the @kbd{M-w} (@code{kill-ring-save}) command. You must mark
19194 each line for this command to succeed, but it does not matter at which
19195 end you put point or mark.)
19199 Please invoke the calls in order, so that five elements attempt to
19200 fill the kill ring:
19205 second piece of text
19207 fourth line of text
19214 Then find the value of @code{kill-ring} by evaluating
19226 ("fifth bit of text" "fourth line of text"
19227 "third line" "second piece of text")
19232 The first element, @samp{first some text}, was dropped.
19235 To return to the old value for the length of the kill ring, evaluate:
19238 (setq kill-ring-max old-kill-ring-max)
19241 @node current-kill, yank, What the Kill Ring Does, Kill Ring
19242 @comment node-name, next, previous, up
19243 @appendixsec The @code{current-kill} Function
19244 @findex current-kill
19246 The @code{current-kill} function changes the element in the kill ring
19247 to which @code{kill-ring-yank-pointer} points. (Also, the
19248 @code{kill-new} function sets @code{kill-ring-yank-pointer} to point
19249 to the latest element of the kill ring. The @code{kill-new}
19250 function is used directly or indirectly by @code{kill-append},
19251 @code{copy-region-as-kill}, @code{kill-ring-save}, @code{kill-line},
19252 and @code{kill-region}.)
19255 * Code for current-kill::
19256 * Understanding current-kill::
19259 @node Code for current-kill, Understanding current-kill, current-kill, current-kill
19261 @unnumberedsubsec The code for @code{current-kill}
19266 The @code{current-kill} function is used by @code{yank} and by
19267 @code{yank-pop}. Here is the code for @code{current-kill}:
19271 (defun current-kill (n &optional do-not-move)
19272 "Rotate the yanking point by N places, and then return that kill.
19273 If N is zero, `interprogram-paste-function' is set, and calling it
19274 returns a string, then that string is added to the front of the
19275 kill ring and returned as the latest kill.
19278 If optional arg DO-NOT-MOVE is non-nil, then don't actually move the
19279 yanking point; just return the Nth kill forward."
19280 (let ((interprogram-paste (and (= n 0)
19281 interprogram-paste-function
19282 (funcall interprogram-paste-function))))
19285 (if interprogram-paste
19287 ;; Disable the interprogram cut function when we add the new
19288 ;; text to the kill ring, so Emacs doesn't try to own the
19289 ;; selection, with identical text.
19290 (let ((interprogram-cut-function nil))
19291 (kill-new interprogram-paste))
19292 interprogram-paste)
19295 (or kill-ring (error "Kill ring is empty"))
19296 (let ((ARGth-kill-element
19297 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19298 (length kill-ring))
19301 (setq kill-ring-yank-pointer ARGth-kill-element))
19302 (car ARGth-kill-element)))))
19306 Remember also that the @code{kill-new} function sets
19307 @code{kill-ring-yank-pointer} to the latest element of the kill
19308 ring, which means that all the functions that call it set the value
19309 indirectly: @code{kill-append}, @code{copy-region-as-kill},
19310 @code{kill-ring-save}, @code{kill-line}, and @code{kill-region}.
19313 Here is the line in @code{kill-new}, which is explained in
19314 @ref{kill-new function, , The @code{kill-new} function}.
19317 (setq kill-ring-yank-pointer kill-ring)
19320 @node Understanding current-kill, , Code for current-kill, current-kill
19322 @unnumberedsubsec @code{current-kill} in Outline
19325 The @code{current-kill} function looks complex, but as usual, it can
19326 be understood by taking it apart piece by piece. First look at it in
19331 (defun current-kill (n &optional do-not-move)
19332 "Rotate the yanking point by N places, and then return that kill."
19338 This function takes two arguments, one of which is optional. It has a
19339 documentation string. It is @emph{not} interactive.
19342 * Body of current-kill::
19343 * Digression concerning error:: How to mislead humans, but not computers.
19344 * Determining the Element::
19347 @node Body of current-kill, Digression concerning error, Understanding current-kill, Understanding current-kill
19349 @unnumberedsubsubsec The Body of @code{current-kill}
19352 The body of the function definition is a @code{let} expression, which
19353 itself has a body as well as a @var{varlist}.
19355 The @code{let} expression declares a variable that will be only usable
19356 within the bounds of this function. This variable is called
19357 @code{interprogram-paste} and is for copying to another program. It
19358 is not for copying within this instance of GNU Emacs. Most window
19359 systems provide a facility for interprogram pasting. Sadly, that
19360 facility usually provides only for the last element. Most windowing
19361 systems have not adopted a ring of many possibilities, even though
19362 Emacs has provided it for decades.
19364 The @code{if} expression has two parts, one if there exists
19365 @code{interprogram-paste} and one if not.
19368 Let us consider the `if not' or else-part of the @code{current-kill}
19369 function. (The then-part uses the @code{kill-new} function, which
19370 we have already described. @xref{kill-new function, , The
19371 @code{kill-new} function}.)
19375 (or kill-ring (error "Kill ring is empty"))
19376 (let ((ARGth-kill-element
19377 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19378 (length kill-ring))
19381 (setq kill-ring-yank-pointer ARGth-kill-element))
19382 (car ARGth-kill-element))
19387 The code first checks whether the kill ring has content; otherwise it
19391 Note that the @code{or} expression is very similar to testing length
19398 (if (zerop (length kill-ring)) ; @r{if-part}
19399 (error "Kill ring is empty")) ; @r{then-part}
19405 If there is not anything in the kill ring, its length must be zero and
19406 an error message sent to the user: @samp{Kill ring is empty}. The
19407 @code{current-kill} function uses an @code{or} expression which is
19408 simpler. But an @code{if} expression reminds us what goes on.
19410 This @code{if} expression uses the function @code{zerop} which returns
19411 true if the value it is testing is zero. When @code{zerop} tests
19412 true, the then-part of the @code{if} is evaluated. The then-part is a
19413 list starting with the function @code{error}, which is a function that
19414 is similar to the @code{message} function
19415 (@pxref{message, , The @code{message} Function}) in that
19416 it prints a one-line message in the echo area. However, in addition
19417 to printing a message, @code{error} also stops evaluation of the
19418 function within which it is embedded. This means that the rest of the
19419 function will not be evaluated if the length of the kill ring is zero.
19421 Then the @code{current-kill} function selects the element to return.
19422 The selection depends on the number of places that @code{current-kill}
19423 rotates and on where @code{kill-ring-yank-pointer} points.
19425 Next, either the optional @code{do-not-move} argument is true or the
19426 current value of @code{kill-ring-yank-pointer} is set to point to the
19427 list. Finally, another expression returns the first element of the
19428 list even if the @code{do-not-move} argument is true.
19430 @node Digression concerning error, Determining the Element, Body of current-kill, Understanding current-kill
19432 @unnumberedsubsubsec Digression about the word `error'
19435 In my opinion, it is slightly misleading, at least to humans, to use
19436 the term `error' as the name of the @code{error} function. A better
19437 term would be `cancel'. Strictly speaking, of course, you cannot
19438 point to, much less rotate a pointer to a list that has no length, so
19439 from the point of view of the computer, the word `error' is correct.
19440 But a human expects to attempt this sort of thing, if only to find out
19441 whether the kill ring is full or empty. This is an act of
19444 From the human point of view, the act of exploration and discovery is
19445 not necessarily an error, and therefore should not be labelled as one,
19446 even in the bowels of a computer. As it is, the code in Emacs implies
19447 that a human who is acting virtuously, by exploring his or her
19448 environment, is making an error. This is bad. Even though the computer
19449 takes the same steps as it does when there is an `error', a term such as
19450 `cancel' would have a clearer connotation.
19452 @node Determining the Element, , Digression concerning error, Understanding current-kill
19454 @unnumberedsubsubsec Determining the Element
19457 Among other actions, the else-part of the @code{if} expression sets
19458 the value of @code{kill-ring-yank-pointer} to
19459 @code{ARGth-kill-element} when the kill ring has something in it and
19460 the value of @code{do-not-move} is @code{nil}.
19463 The code looks like this:
19467 (nthcdr (mod (- n (length kill-ring-yank-pointer))
19468 (length kill-ring))
19473 This needs some examination. Unless it is not supposed to move the
19474 pointer, the @code{current-kill} function changes where
19475 @code{kill-ring-yank-pointer} points.
19477 @w{@code{(setq kill-ring-yank-pointer ARGth-kill-element))}}
19478 expression does. Also, clearly, @code{ARGth-kill-element} is being
19479 set to be equal to some @sc{cdr} of the kill ring, using the
19480 @code{nthcdr} function that is described in an earlier section.
19481 (@xref{copy-region-as-kill}.) How does it do this?
19483 As we have seen before (@pxref{nthcdr}), the @code{nthcdr} function
19484 works by repeatedly taking the @sc{cdr} of a list---it takes the
19485 @sc{cdr} of the @sc{cdr} of the @sc{cdr} @dots{}
19488 The two following expressions produce the same result:
19492 (setq kill-ring-yank-pointer (cdr kill-ring))
19494 (setq kill-ring-yank-pointer (nthcdr 1 kill-ring))
19498 However, the @code{nthcdr} expression is more complicated. It uses
19499 the @code{mod} function to determine which @sc{cdr} to select.
19501 (You will remember to look at inner functions first; indeed, we will
19502 have to go inside the @code{mod}.)
19504 The @code{mod} function returns the value of its first argument modulo
19505 the second; that is to say, it returns the remainder after dividing
19506 the first argument by the second. The value returned has the same
19507 sign as the second argument.
19515 @result{} 0 ;; @r{because there is no remainder}
19522 In this case, the first argument is often smaller than the second.
19534 We can guess what the @code{-} function does. It is like @code{+} but
19535 subtracts instead of adds; the @code{-} function subtracts its second
19536 argument from its first. Also, we already know what the @code{length}
19537 function does (@pxref{length}). It returns the length of a list.
19539 And @code{n} is the name of the required argument to the
19540 @code{current-kill} function.
19543 So when the first argument to @code{nthcdr} is zero, the @code{nthcdr}
19544 expression returns the whole list, as you can see by evaluating the
19549 ;; kill-ring-yank-pointer @r{and} kill-ring @r{have a length of four}
19550 ;; @r{and} (mod (- 0 4) 4) @result{} 0
19551 (nthcdr (mod (- 0 4) 4)
19552 '("fourth line of text"
19554 "second piece of text"
19555 "first some text"))
19560 When the first argument to the @code{current-kill} function is one,
19561 the @code{nthcdr} expression returns the list without its first
19566 (nthcdr (mod (- 1 4) 4)
19567 '("fourth line of text"
19569 "second piece of text"
19570 "first some text"))
19574 @cindex @samp{global variable} defined
19575 @cindex @samp{variable, global}, defined
19576 Incidentally, both @code{kill-ring} and @code{kill-ring-yank-pointer}
19577 are @dfn{global variables}. That means that any expression in Emacs
19578 Lisp can access them. They are not like the local variables set by
19579 @code{let} or like the symbols in an argument list.
19580 Local variables can only be accessed
19581 within the @code{let} that defines them or the function that specifies
19582 them in an argument list (and within expressions called by them).
19585 @c texi2dvi fails when the name of the section is within ifnottex ...
19586 (@xref{Prevent confusion, , @code{let} Prevents Confusion}, and
19587 @ref{defun, , The @code{defun} Special Form}.)
19590 @node yank, yank-pop, current-kill, Kill Ring
19591 @comment node-name, next, previous, up
19592 @appendixsec @code{yank}
19595 After learning about @code{current-kill}, the code for the
19596 @code{yank} function is almost easy.
19598 The @code{yank} function does not use the
19599 @code{kill-ring-yank-pointer} variable directly. It calls
19600 @code{insert-for-yank} which calls @code{current-kill} which sets the
19601 @code{kill-ring-yank-pointer} variable.
19604 The code looks like this:
19609 (defun yank (&optional arg)
19610 "Reinsert (\"paste\") the last stretch of killed text.
19611 More precisely, reinsert the stretch of killed text most recently
19612 killed OR yanked. Put point at end, and set mark at beginning.
19613 With just \\[universal-argument] as argument, same but put point at
19614 beginning (and mark at end). With argument N, reinsert the Nth most
19615 recently killed stretch of killed text.
19617 When this command inserts killed text into the buffer, it honors
19618 `yank-excluded-properties' and `yank-handler' as described in the
19619 doc string for `insert-for-yank-1', which see.
19621 See also the command \\[yank-pop]."
19625 (setq yank-window-start (window-start))
19626 ;; If we don't get all the way thru, make last-command indicate that
19627 ;; for the following command.
19628 (setq this-command t)
19629 (push-mark (point))
19632 (insert-for-yank (current-kill (cond
19637 ;; This is like exchange-point-and-mark,
19638 ;; but doesn't activate the mark.
19639 ;; It is cleaner to avoid activation, even though the command
19640 ;; loop would deactivate the mark because we inserted text.
19641 (goto-char (prog1 (mark t)
19642 (set-marker (mark-marker) (point) (current-buffer)))))
19645 ;; If we do get all the way thru, make this-command indicate that.
19646 (if (eq this-command t)
19647 (setq this-command 'yank))
19652 The key expression is @code{insert-for-yank}, which inserts the string
19653 returned by @code{current-kill}, but removes some text properties from
19656 However, before getting to that expression, the function sets the value
19657 of @code{yank-window-start} to the position returned by the
19658 @code{(window-start)} expression, the position at which the display
19659 currently starts. The @code{yank} function also sets
19660 @code{this-command} and pushes the mark.
19662 After it yanks the appropriate element, if the optional argument is a
19663 @sc{cons} rather than a number or nothing, it puts point at beginning
19664 of the yanked text and mark at its end.
19666 (The @code{prog1} function is like @code{progn} but returns the value
19667 of its first argument rather than the value of its last argument. Its
19668 first argument is forced to return the buffer's mark as an integer.
19669 You can see the documentation for these functions by placing point
19670 over them in this buffer and then typing @kbd{C-h f}
19671 (@code{describe-function}) followed by a @kbd{RET}; the default is the
19674 The last part of the function tells what to do when it succeeds.
19676 @node yank-pop, ring file, yank, Kill Ring
19677 @comment node-name, next, previous, up
19678 @appendixsec @code{yank-pop}
19681 After understanding @code{yank} and @code{current-kill}, you know how
19682 to approach the @code{yank-pop} function. Leaving out the
19683 documentation to save space, it looks like this:
19688 (defun yank-pop (&optional arg)
19691 (if (not (eq last-command 'yank))
19692 (error "Previous command was not a yank"))
19695 (setq this-command 'yank)
19696 (unless arg (setq arg 1))
19697 (let ((inhibit-read-only t)
19698 (before (< (point) (mark t))))
19702 (funcall (or yank-undo-function 'delete-region) (point) (mark t))
19703 (funcall (or yank-undo-function 'delete-region) (mark t) (point)))
19704 (setq yank-undo-function nil)
19707 (set-marker (mark-marker) (point) (current-buffer))
19708 (insert-for-yank (current-kill arg))
19709 ;; Set the window start back where it was in the yank command,
19711 (set-window-start (selected-window) yank-window-start t)
19715 ;; This is like exchange-point-and-mark,
19716 ;; but doesn't activate the mark.
19717 ;; It is cleaner to avoid activation, even though the command
19718 ;; loop would deactivate the mark because we inserted text.
19719 (goto-char (prog1 (mark t)
19720 (set-marker (mark-marker)
19722 (current-buffer))))))
19727 The function is interactive with a small @samp{p} so the prefix
19728 argument is processed and passed to the function. The command can
19729 only be used after a previous yank; otherwise an error message is
19730 sent. This check uses the variable @code{last-command} which is set
19731 by @code{yank} and is discussed elsewhere.
19732 (@xref{copy-region-as-kill}.)
19734 The @code{let} clause sets the variable @code{before} to true or false
19735 depending whether point is before or after mark and then the region
19736 between point and mark is deleted. This is the region that was just
19737 inserted by the previous yank and it is this text that will be
19740 @code{funcall} calls its first argument as a function, passing
19741 remaining arguments to it. The first argument is whatever the
19742 @code{or} expression returns. The two remaining arguments are the
19743 positions of point and mark set by the preceding @code{yank} command.
19745 There is more, but that is the hardest part.
19747 @node ring file, , yank-pop, Kill Ring
19748 @comment node-name, next, previous, up
19749 @appendixsec The @file{ring.el} File
19750 @cindex @file{ring.el} file
19752 Interestingly, GNU Emacs posses a file called @file{ring.el} that
19753 provides many of the features we just discussed. But functions such
19754 as @code{kill-ring-yank-pointer} do not use this library, possibly
19755 because they were written earlier.
19757 @node Full Graph, Free Software and Free Manuals, Kill Ring, Top
19758 @appendix A Graph with Labelled Axes
19760 Printed axes help you understand a graph. They convey scale. In an
19761 earlier chapter (@pxref{Readying a Graph, , Readying a Graph}), we
19762 wrote the code to print the body of a graph. Here we write the code
19763 for printing and labelling vertical and horizontal axes, along with the
19767 * Labelled Example::
19768 * print-graph Varlist:: @code{let} expression in @code{print-graph}.
19769 * print-Y-axis:: Print a label for the vertical axis.
19770 * print-X-axis:: Print a horizontal label.
19771 * Print Whole Graph:: The function to print a complete graph.
19774 @node Labelled Example, print-graph Varlist, Full Graph, Full Graph
19776 @unnumberedsec Labelled Example Graph
19779 Since insertions fill a buffer to the right and below point, the new
19780 graph printing function should first print the Y or vertical axis,
19781 then the body of the graph, and finally the X or horizontal axis.
19782 This sequence lays out for us the contents of the function:
19792 Print body of graph.
19799 Here is an example of how a finished graph should look:
19812 1 - ****************
19819 In this graph, both the vertical and the horizontal axes are labelled
19820 with numbers. However, in some graphs, the horizontal axis is time
19821 and would be better labelled with months, like this:
19835 Indeed, with a little thought, we can easily come up with a variety of
19836 vertical and horizontal labelling schemes. Our task could become
19837 complicated. But complications breed confusion. Rather than permit
19838 this, it is better choose a simple labelling scheme for our first
19839 effort, and to modify or replace it later.
19842 These considerations suggest the following outline for the
19843 @code{print-graph} function:
19847 (defun print-graph (numbers-list)
19848 "@var{documentation}@dots{}"
19849 (let ((height @dots{}
19853 (print-Y-axis height @dots{} )
19854 (graph-body-print numbers-list)
19855 (print-X-axis @dots{} )))
19859 We can work on each part of the @code{print-graph} function definition
19862 @node print-graph Varlist, print-Y-axis, Labelled Example, Full Graph
19863 @comment node-name, next, previous, up
19864 @appendixsec The @code{print-graph} Varlist
19865 @cindex @code{print-graph} varlist
19867 In writing the @code{print-graph} function, the first task is to write
19868 the varlist in the @code{let} expression. (We will leave aside for the
19869 moment any thoughts about making the function interactive or about the
19870 contents of its documentation string.)
19872 The varlist should set several values. Clearly, the top of the label
19873 for the vertical axis must be at least the height of the graph, which
19874 means that we must obtain this information here. Note that the
19875 @code{print-graph-body} function also requires this information. There
19876 is no reason to calculate the height of the graph in two different
19877 places, so we should change @code{print-graph-body} from the way we
19878 defined it earlier to take advantage of the calculation.
19880 Similarly, both the function for printing the X axis labels and the
19881 @code{print-graph-body} function need to learn the value of the width of
19882 each symbol. We can perform the calculation here and change the
19883 definition for @code{print-graph-body} from the way we defined it in the
19886 The length of the label for the horizontal axis must be at least as long
19887 as the graph. However, this information is used only in the function
19888 that prints the horizontal axis, so it does not need to be calculated here.
19890 These thoughts lead us directly to the following form for the varlist
19891 in the @code{let} for @code{print-graph}:
19895 (let ((height (apply 'max numbers-list)) ; @r{First version.}
19896 (symbol-width (length graph-blank)))
19901 As we shall see, this expression is not quite right.
19904 @node print-Y-axis, print-X-axis, print-graph Varlist, Full Graph
19905 @comment node-name, next, previous, up
19906 @appendixsec The @code{print-Y-axis} Function
19907 @cindex Axis, print vertical
19908 @cindex Y axis printing
19909 @cindex Vertical axis printing
19910 @cindex Print vertical axis
19912 The job of the @code{print-Y-axis} function is to print a label for
19913 the vertical axis that looks like this:
19931 The function should be passed the height of the graph, and then should
19932 construct and insert the appropriate numbers and marks.
19935 * print-Y-axis in Detail::
19936 * Height of label:: What height for the Y axis?
19937 * Compute a Remainder:: How to compute the remainder of a division.
19938 * Y Axis Element:: Construct a line for the Y axis.
19939 * Y-axis-column:: Generate a list of Y axis labels.
19940 * print-Y-axis Penultimate:: A not quite final version.
19943 @node print-Y-axis in Detail, Height of label, print-Y-axis, print-Y-axis
19945 @unnumberedsubsec The @code{print-Y-axis} Function in Detail
19948 It is easy enough to see in the figure what the Y axis label should
19949 look like; but to say in words, and then to write a function
19950 definition to do the job is another matter. It is not quite true to
19951 say that we want a number and a tic every five lines: there are only
19952 three lines between the @samp{1} and the @samp{5} (lines 2, 3, and 4),
19953 but four lines between the @samp{5} and the @samp{10} (lines 6, 7, 8,
19954 and 9). It is better to say that we want a number and a tic mark on
19955 the base line (number 1) and then that we want a number and a tic on
19956 the fifth line from the bottom and on every line that is a multiple of
19959 @node Height of label, Compute a Remainder, print-Y-axis in Detail, print-Y-axis
19961 @unnumberedsubsec What height should the label be?
19964 The next issue is what height the label should be? Suppose the maximum
19965 height of tallest column of the graph is seven. Should the highest
19966 label on the Y axis be @samp{5 -}, and should the graph stick up above
19967 the label? Or should the highest label be @samp{7 -}, and mark the peak
19968 of the graph? Or should the highest label be @code{10 -}, which is a
19969 multiple of five, and be higher than the topmost value of the graph?
19971 The latter form is preferred. Most graphs are drawn within rectangles
19972 whose sides are an integral number of steps long---5, 10, 15, and so
19973 on for a step distance of five. But as soon as we decide to use a
19974 step height for the vertical axis, we discover that the simple
19975 expression in the varlist for computing the height is wrong. The
19976 expression is @code{(apply 'max numbers-list)}. This returns the
19977 precise height, not the maximum height plus whatever is necessary to
19978 round up to the nearest multiple of five. A more complex expression
19981 As usual in cases like this, a complex problem becomes simpler if it is
19982 divided into several smaller problems.
19984 First, consider the case when the highest value of the graph is an
19985 integral multiple of five---when it is 5, 10, 15, or some higher
19986 multiple of five. We can use this value as the Y axis height.
19988 A fairly simply way to determine whether a number is a multiple of
19989 five is to divide it by five and see if the division results in a
19990 remainder. If there is no remainder, the number is a multiple of
19991 five. Thus, seven divided by five has a remainder of two, and seven
19992 is not an integral multiple of five. Put in slightly different
19993 language, more reminiscent of the classroom, five goes into seven
19994 once, with a remainder of two. However, five goes into ten twice,
19995 with no remainder: ten is an integral multiple of five.
19997 @node Compute a Remainder, Y Axis Element, Height of label, print-Y-axis
19998 @appendixsubsec Side Trip: Compute a Remainder
20000 @findex % @r{(remainder function)}
20001 @cindex Remainder function, @code{%}
20002 In Lisp, the function for computing a remainder is @code{%}. The
20003 function returns the remainder of its first argument divided by its
20004 second argument. As it happens, @code{%} is a function in Emacs Lisp
20005 that you cannot discover using @code{apropos}: you find nothing if you
20006 type @kbd{M-x apropos @key{RET} remainder @key{RET}}. The only way to
20007 learn of the existence of @code{%} is to read about it in a book such
20008 as this or in the Emacs Lisp sources.
20010 You can try the @code{%} function by evaluating the following two
20022 The first expression returns 2 and the second expression returns 0.
20024 To test whether the returned value is zero or some other number, we
20025 can use the @code{zerop} function. This function returns @code{t} if
20026 its argument, which must be a number, is zero.
20038 Thus, the following expression will return @code{t} if the height
20039 of the graph is evenly divisible by five:
20042 (zerop (% height 5))
20046 (The value of @code{height}, of course, can be found from @code{(apply
20047 'max numbers-list)}.)
20049 On the other hand, if the value of @code{height} is not a multiple of
20050 five, we want to reset the value to the next higher multiple of five.
20051 This is straightforward arithmetic using functions with which we are
20052 already familiar. First, we divide the value of @code{height} by five
20053 to determine how many times five goes into the number. Thus, five
20054 goes into twelve twice. If we add one to this quotient and multiply by
20055 five, we will obtain the value of the next multiple of five that is
20056 larger than the height. Five goes into twelve twice. Add one to two,
20057 and multiply by five; the result is fifteen, which is the next multiple
20058 of five that is higher than twelve. The Lisp expression for this is:
20061 (* (1+ (/ height 5)) 5)
20065 For example, if you evaluate the following, the result is 15:
20068 (* (1+ (/ 12 5)) 5)
20071 All through this discussion, we have been using `five' as the value
20072 for spacing labels on the Y axis; but we may want to use some other
20073 value. For generality, we should replace `five' with a variable to
20074 which we can assign a value. The best name I can think of for this
20075 variable is @code{Y-axis-label-spacing}.
20078 Using this term, and an @code{if} expression, we produce the
20083 (if (zerop (% height Y-axis-label-spacing))
20086 (* (1+ (/ height Y-axis-label-spacing))
20087 Y-axis-label-spacing))
20092 This expression returns the value of @code{height} itself if the height
20093 is an even multiple of the value of the @code{Y-axis-label-spacing} or
20094 else it computes and returns a value of @code{height} that is equal to
20095 the next higher multiple of the value of the @code{Y-axis-label-spacing}.
20097 We can now include this expression in the @code{let} expression of the
20098 @code{print-graph} function (after first setting the value of
20099 @code{Y-axis-label-spacing}):
20100 @vindex Y-axis-label-spacing
20104 (defvar Y-axis-label-spacing 5
20105 "Number of lines from one Y axis label to next.")
20110 (let* ((height (apply 'max numbers-list))
20111 (height-of-top-line
20112 (if (zerop (% height Y-axis-label-spacing))
20117 (* (1+ (/ height Y-axis-label-spacing))
20118 Y-axis-label-spacing)))
20119 (symbol-width (length graph-blank))))
20125 (Note use of the @code{let*} function: the initial value of height is
20126 computed once by the @code{(apply 'max numbers-list)} expression and
20127 then the resulting value of @code{height} is used to compute its
20128 final value. @xref{fwd-para let, , The @code{let*} expression}, for
20129 more about @code{let*}.)
20131 @node Y Axis Element, Y-axis-column, Compute a Remainder, print-Y-axis
20132 @appendixsubsec Construct a Y Axis Element
20134 When we print the vertical axis, we want to insert strings such as
20135 @w{@samp{5 -}} and @w{@samp{10 - }} every five lines.
20136 Moreover, we want the numbers and dashes to line up, so shorter
20137 numbers must be padded with leading spaces. If some of the strings
20138 use two digit numbers, the strings with single digit numbers must
20139 include a leading blank space before the number.
20141 @findex number-to-string
20142 To figure out the length of the number, the @code{length} function is
20143 used. But the @code{length} function works only with a string, not with
20144 a number. So the number has to be converted from being a number to
20145 being a string. This is done with the @code{number-to-string} function.
20150 (length (number-to-string 35))
20153 (length (number-to-string 100))
20159 (@code{number-to-string} is also called @code{int-to-string}; you will
20160 see this alternative name in various sources.)
20162 In addition, in each label, each number is followed by a string such
20163 as @w{@samp{ - }}, which we will call the @code{Y-axis-tic} marker.
20164 This variable is defined with @code{defvar}:
20169 (defvar Y-axis-tic " - "
20170 "String that follows number in a Y axis label.")
20174 The length of the Y label is the sum of the length of the Y axis tic
20175 mark and the length of the number of the top of the graph.
20178 (length (concat (number-to-string height) Y-axis-tic)))
20181 This value will be calculated by the @code{print-graph} function in
20182 its varlist as @code{full-Y-label-width} and passed on. (Note that we
20183 did not think to include this in the varlist when we first proposed it.)
20185 To make a complete vertical axis label, a tic mark is concatenated
20186 with a number; and the two together may be preceded by one or more
20187 spaces depending on how long the number is. The label consists of
20188 three parts: the (optional) leading spaces, the number, and the tic
20189 mark. The function is passed the value of the number for the specific
20190 row, and the value of the width of the top line, which is calculated
20191 (just once) by @code{print-graph}.
20195 (defun Y-axis-element (number full-Y-label-width)
20196 "Construct a NUMBERed label element.
20197 A numbered element looks like this ` 5 - ',
20198 and is padded as needed so all line up with
20199 the element for the largest number."
20202 (let* ((leading-spaces
20203 (- full-Y-label-width
20205 (concat (number-to-string number)
20210 (make-string leading-spaces ? )
20211 (number-to-string number)
20216 The @code{Y-axis-element} function concatenates together the leading
20217 spaces, if any; the number, as a string; and the tic mark.
20219 To figure out how many leading spaces the label will need, the
20220 function subtracts the actual length of the label---the length of the
20221 number plus the length of the tic mark---from the desired label width.
20223 @findex make-string
20224 Blank spaces are inserted using the @code{make-string} function. This
20225 function takes two arguments: the first tells it how long the string
20226 will be and the second is a symbol for the character to insert, in a
20227 special format. The format is a question mark followed by a blank
20228 space, like this, @samp{? }. @xref{Character Type, , Character Type,
20229 elisp, The GNU Emacs Lisp Reference Manual}, for a description of the
20230 syntax for characters. (Of course, you might want to replace the
20231 blank space by some other character @dots{} You know what to do.)
20233 The @code{number-to-string} function is used in the concatenation
20234 expression, to convert the number to a string that is concatenated
20235 with the leading spaces and the tic mark.
20237 @node Y-axis-column, print-Y-axis Penultimate, Y Axis Element, print-Y-axis
20238 @appendixsubsec Create a Y Axis Column
20240 The preceding functions provide all the tools needed to construct a
20241 function that generates a list of numbered and blank strings to insert
20242 as the label for the vertical axis:
20244 @findex Y-axis-column
20247 (defun Y-axis-column (height width-of-label)
20248 "Construct list of Y axis labels and blank strings.
20249 For HEIGHT of line above base and WIDTH-OF-LABEL."
20253 (while (> height 1)
20254 (if (zerop (% height Y-axis-label-spacing))
20255 ;; @r{Insert label.}
20258 (Y-axis-element height width-of-label)
20262 ;; @r{Else, insert blanks.}
20265 (make-string width-of-label ? )
20267 (setq height (1- height)))
20268 ;; @r{Insert base line.}
20270 (cons (Y-axis-element 1 width-of-label) Y-axis))
20271 (nreverse Y-axis)))
20275 In this function, we start with the value of @code{height} and
20276 repetitively subtract one from its value. After each subtraction, we
20277 test to see whether the value is an integral multiple of the
20278 @code{Y-axis-label-spacing}. If it is, we construct a numbered label
20279 using the @code{Y-axis-element} function; if not, we construct a
20280 blank label using the @code{make-string} function. The base line
20281 consists of the number one followed by a tic mark.
20284 @node print-Y-axis Penultimate, , Y-axis-column, print-Y-axis
20285 @appendixsubsec The Not Quite Final Version of @code{print-Y-axis}
20287 The list constructed by the @code{Y-axis-column} function is passed to
20288 the @code{print-Y-axis} function, which inserts the list as a column.
20290 @findex print-Y-axis
20293 (defun print-Y-axis (height full-Y-label-width)
20294 "Insert Y axis using HEIGHT and FULL-Y-LABEL-WIDTH.
20295 Height must be the maximum height of the graph.
20296 Full width is the width of the highest label element."
20297 ;; Value of height and full-Y-label-width
20298 ;; are passed by `print-graph'.
20301 (let ((start (point)))
20303 (Y-axis-column height full-Y-label-width))
20304 ;; @r{Place point ready for inserting graph.}
20306 ;; @r{Move point forward by value of} full-Y-label-width
20307 (forward-char full-Y-label-width)))
20311 The @code{print-Y-axis} uses the @code{insert-rectangle} function to
20312 insert the Y axis labels created by the @code{Y-axis-column} function.
20313 In addition, it places point at the correct position for printing the body of
20316 You can test @code{print-Y-axis}:
20324 Y-axis-label-spacing
20333 Copy the following expression:
20336 (print-Y-axis 12 5)
20340 Switch to the @file{*scratch*} buffer and place the cursor where you
20341 want the axis labels to start.
20344 Type @kbd{M-:} (@code{eval-expression}).
20347 Yank the @code{graph-body-print} expression into the minibuffer
20348 with @kbd{C-y} (@code{yank)}.
20351 Press @key{RET} to evaluate the expression.
20354 Emacs will print labels vertically, the top one being @w{@samp{10 -@w{
20355 }}}. (The @code{print-graph} function will pass the value of
20356 @code{height-of-top-line}, which in this case will end up as 15,
20357 thereby getting rid of what might appear as a bug.)
20360 @node print-X-axis, Print Whole Graph, print-Y-axis, Full Graph
20361 @appendixsec The @code{print-X-axis} Function
20362 @cindex Axis, print horizontal
20363 @cindex X axis printing
20364 @cindex Print horizontal axis
20365 @cindex Horizontal axis printing
20367 X axis labels are much like Y axis labels, except that the ticks are on a
20368 line above the numbers. Labels should look like this:
20377 The first tic is under the first column of the graph and is preceded by
20378 several blank spaces. These spaces provide room in rows above for the Y
20379 axis labels. The second, third, fourth, and subsequent ticks are all
20380 spaced equally, according to the value of @code{X-axis-label-spacing}.
20382 The second row of the X axis consists of numbers, preceded by several
20383 blank spaces and also separated according to the value of the variable
20384 @code{X-axis-label-spacing}.
20386 The value of the variable @code{X-axis-label-spacing} should itself be
20387 measured in units of @code{symbol-width}, since you may want to change
20388 the width of the symbols that you are using to print the body of the
20389 graph without changing the ways the graph is labelled.
20392 * Similarities differences:: Much like @code{print-Y-axis}, but not exactly.
20393 * X Axis Tic Marks:: Create tic marks for the horizontal axis.
20396 @node Similarities differences, X Axis Tic Marks, print-X-axis, print-X-axis
20398 @unnumberedsubsec Similarities and differences
20401 The @code{print-X-axis} function is constructed in more or less the
20402 same fashion as the @code{print-Y-axis} function except that it has
20403 two lines: the line of tic marks and the numbers. We will write a
20404 separate function to print each line and then combine them within the
20405 @code{print-X-axis} function.
20407 This is a three step process:
20411 Write a function to print the X axis tic marks, @code{print-X-axis-tic-line}.
20414 Write a function to print the X numbers, @code{print-X-axis-numbered-line}.
20417 Write a function to print both lines, the @code{print-X-axis} function,
20418 using @code{print-X-axis-tic-line} and
20419 @code{print-X-axis-numbered-line}.
20422 @node X Axis Tic Marks, , Similarities differences, print-X-axis
20423 @appendixsubsec X Axis Tic Marks
20425 The first function should print the X axis tic marks. We must specify
20426 the tic marks themselves and their spacing:
20430 (defvar X-axis-label-spacing
20431 (if (boundp 'graph-blank)
20432 (* 5 (length graph-blank)) 5)
20433 "Number of units from one X axis label to next.")
20438 (Note that the value of @code{graph-blank} is set by another
20439 @code{defvar}. The @code{boundp} predicate checks whether it has
20440 already been set; @code{boundp} returns @code{nil} if it has not. If
20441 @code{graph-blank} were unbound and we did not use this conditional
20442 construction, in a recent GNU Emacs, we would enter the debugger and
20443 see an error message saying @samp{@w{Debugger entered--Lisp error:}
20444 @w{(void-variable graph-blank)}}.)
20447 Here is the @code{defvar} for @code{X-axis-tic-symbol}:
20451 (defvar X-axis-tic-symbol "|"
20452 "String to insert to point to a column in X axis.")
20457 The goal is to make a line that looks like this:
20463 The first tic is indented so that it is under the first column, which is
20464 indented to provide space for the Y axis labels.
20466 A tic element consists of the blank spaces that stretch from one tic to
20467 the next plus a tic symbol. The number of blanks is determined by the
20468 width of the tic symbol and the @code{X-axis-label-spacing}.
20471 The code looks like this:
20475 ;;; X-axis-tic-element
20479 ;; @r{Make a string of blanks.}
20480 (- (* symbol-width X-axis-label-spacing)
20481 (length X-axis-tic-symbol))
20483 ;; @r{Concatenate blanks with tic symbol.}
20489 Next, we determine how many blanks are needed to indent the first tic
20490 mark to the first column of the graph. This uses the value of
20491 @code{full-Y-label-width} passed it by the @code{print-graph} function.
20494 The code to make @code{X-axis-leading-spaces}
20499 ;; X-axis-leading-spaces
20501 (make-string full-Y-label-width ? )
20506 We also need to determine the length of the horizontal axis, which is
20507 the length of the numbers list, and the number of ticks in the horizontal
20514 (length numbers-list)
20520 (* symbol-width X-axis-label-spacing)
20524 ;; number-of-X-ticks
20525 (if (zerop (% (X-length tic-width)))
20526 (/ (X-length tic-width))
20527 (1+ (/ (X-length tic-width))))
20532 All this leads us directly to the function for printing the X axis tic line:
20534 @findex print-X-axis-tic-line
20537 (defun print-X-axis-tic-line
20538 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
20539 "Print ticks for X axis."
20540 (insert X-axis-leading-spaces)
20541 (insert X-axis-tic-symbol) ; @r{Under first column.}
20544 ;; @r{Insert second tic in the right spot.}
20547 (- (* symbol-width X-axis-label-spacing)
20548 ;; @r{Insert white space up to second tic symbol.}
20549 (* 2 (length X-axis-tic-symbol)))
20551 X-axis-tic-symbol))
20554 ;; @r{Insert remaining ticks.}
20555 (while (> number-of-X-tics 1)
20556 (insert X-axis-tic-element)
20557 (setq number-of-X-tics (1- number-of-X-tics))))
20561 The line of numbers is equally straightforward:
20564 First, we create a numbered element with blank spaces before each number:
20566 @findex X-axis-element
20569 (defun X-axis-element (number)
20570 "Construct a numbered X axis element."
20571 (let ((leading-spaces
20572 (- (* symbol-width X-axis-label-spacing)
20573 (length (number-to-string number)))))
20574 (concat (make-string leading-spaces ? )
20575 (number-to-string number))))
20579 Next, we create the function to print the numbered line, starting with
20580 the number ``1'' under the first column:
20582 @findex print-X-axis-numbered-line
20585 (defun print-X-axis-numbered-line
20586 (number-of-X-tics X-axis-leading-spaces)
20587 "Print line of X-axis numbers"
20588 (let ((number X-axis-label-spacing))
20589 (insert X-axis-leading-spaces)
20595 ;; @r{Insert white space up to next number.}
20596 (- (* symbol-width X-axis-label-spacing) 2)
20598 (number-to-string number)))
20601 ;; @r{Insert remaining numbers.}
20602 (setq number (+ number X-axis-label-spacing))
20603 (while (> number-of-X-tics 1)
20604 (insert (X-axis-element number))
20605 (setq number (+ number X-axis-label-spacing))
20606 (setq number-of-X-tics (1- number-of-X-tics)))))
20610 Finally, we need to write the @code{print-X-axis} that uses
20611 @code{print-X-axis-tic-line} and
20612 @code{print-X-axis-numbered-line}.
20614 The function must determine the local values of the variables used by both
20615 @code{print-X-axis-tic-line} and @code{print-X-axis-numbered-line}, and
20616 then it must call them. Also, it must print the carriage return that
20617 separates the two lines.
20619 The function consists of a varlist that specifies five local variables,
20620 and calls to each of the two line printing functions:
20622 @findex print-X-axis
20625 (defun print-X-axis (numbers-list)
20626 "Print X axis labels to length of NUMBERS-LIST."
20627 (let* ((leading-spaces
20628 (make-string full-Y-label-width ? ))
20631 ;; symbol-width @r{is provided by} graph-body-print
20632 (tic-width (* symbol-width X-axis-label-spacing))
20633 (X-length (length numbers-list))
20641 ;; @r{Make a string of blanks.}
20642 (- (* symbol-width X-axis-label-spacing)
20643 (length X-axis-tic-symbol))
20647 ;; @r{Concatenate blanks with tic symbol.}
20648 X-axis-tic-symbol))
20652 (if (zerop (% X-length tic-width))
20653 (/ X-length tic-width)
20654 (1+ (/ X-length tic-width)))))
20657 (print-X-axis-tic-line tic-number leading-spaces X-tic)
20659 (print-X-axis-numbered-line tic-number leading-spaces)))
20664 You can test @code{print-X-axis}:
20668 Install @code{X-axis-tic-symbol}, @code{X-axis-label-spacing},
20669 @code{print-X-axis-tic-line}, as well as @code{X-axis-element},
20670 @code{print-X-axis-numbered-line}, and @code{print-X-axis}.
20673 Copy the following expression:
20678 (let ((full-Y-label-width 5)
20681 '(1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16))))
20686 Switch to the @file{*scratch*} buffer and place the cursor where you
20687 want the axis labels to start.
20690 Type @kbd{M-:} (@code{eval-expression}).
20693 Yank the test expression into the minibuffer
20694 with @kbd{C-y} (@code{yank)}.
20697 Press @key{RET} to evaluate the expression.
20701 Emacs will print the horizontal axis like this:
20711 @node Print Whole Graph, , print-X-axis, Full Graph
20712 @appendixsec Printing the Whole Graph
20713 @cindex Printing the whole graph
20714 @cindex Whole graph printing
20715 @cindex Graph, printing all
20717 Now we are nearly ready to print the whole graph.
20719 The function to print the graph with the proper labels follows the
20720 outline we created earlier (@pxref{Full Graph, , A Graph with Labelled
20721 Axes}), but with additions.
20724 Here is the outline:
20728 (defun print-graph (numbers-list)
20729 "@var{documentation}@dots{}"
20730 (let ((height @dots{}
20734 (print-Y-axis height @dots{} )
20735 (graph-body-print numbers-list)
20736 (print-X-axis @dots{} )))
20741 * The final version:: A few changes.
20742 * Test print-graph:: Run a short test.
20743 * Graphing words in defuns:: Executing the final code.
20744 * lambda:: How to write an anonymous function.
20745 * mapcar:: Apply a function to elements of a list.
20746 * Another Bug:: Yet another bug @dots{} most insidious.
20747 * Final printed graph:: The graph itself!
20750 @node The final version, Test print-graph, Print Whole Graph, Print Whole Graph
20752 @unnumberedsubsec Changes for the Final Version
20755 The final version is different from what we planned in two ways:
20756 first, it contains additional values calculated once in the varlist;
20757 second, it carries an option to specify the labels' increment per row.
20758 This latter feature turns out to be essential; otherwise, a graph may
20759 have more rows than fit on a display or on a sheet of paper.
20762 This new feature requires a change to the @code{Y-axis-column}
20763 function, to add @code{vertical-step} to it. The function looks like
20766 @findex Y-axis-column @r{Final version.}
20769 ;;; @r{Final version.}
20770 (defun Y-axis-column
20771 (height width-of-label &optional vertical-step)
20772 "Construct list of labels for Y axis.
20773 HEIGHT is maximum height of graph.
20774 WIDTH-OF-LABEL is maximum width of label.
20775 VERTICAL-STEP, an option, is a positive integer
20776 that specifies how much a Y axis label increments
20777 for each line. For example, a step of 5 means
20778 that each line is five units of the graph."
20782 (number-per-line (or vertical-step 1)))
20783 (while (> height 1)
20784 (if (zerop (% height Y-axis-label-spacing))
20787 ;; @r{Insert label.}
20791 (* height number-per-line)
20796 ;; @r{Else, insert blanks.}
20799 (make-string width-of-label ? )
20801 (setq height (1- height)))
20804 ;; @r{Insert base line.}
20805 (setq Y-axis (cons (Y-axis-element
20806 (or vertical-step 1)
20809 (nreverse Y-axis)))
20813 The values for the maximum height of graph and the width of a symbol
20814 are computed by @code{print-graph} in its @code{let} expression; so
20815 @code{graph-body-print} must be changed to accept them.
20817 @findex graph-body-print @r{Final version.}
20820 ;;; @r{Final version.}
20821 (defun graph-body-print (numbers-list height symbol-width)
20822 "Print a bar graph of the NUMBERS-LIST.
20823 The numbers-list consists of the Y-axis values.
20824 HEIGHT is maximum height of graph.
20825 SYMBOL-WIDTH is number of each column."
20828 (let (from-position)
20829 (while numbers-list
20830 (setq from-position (point))
20832 (column-of-graph height (car numbers-list)))
20833 (goto-char from-position)
20834 (forward-char symbol-width)
20837 ;; @r{Draw graph column by column.}
20839 (setq numbers-list (cdr numbers-list)))
20840 ;; @r{Place point for X axis labels.}
20841 (forward-line height)
20847 Finally, the code for the @code{print-graph} function:
20849 @findex print-graph @r{Final version.}
20852 ;;; @r{Final version.}
20854 (numbers-list &optional vertical-step)
20855 "Print labelled bar graph of the NUMBERS-LIST.
20856 The numbers-list consists of the Y-axis values.
20860 Optionally, VERTICAL-STEP, a positive integer,
20861 specifies how much a Y axis label increments for
20862 each line. For example, a step of 5 means that
20863 each row is five units."
20866 (let* ((symbol-width (length graph-blank))
20867 ;; @code{height} @r{is both the largest number}
20868 ;; @r{and the number with the most digits.}
20869 (height (apply 'max numbers-list))
20872 (height-of-top-line
20873 (if (zerop (% height Y-axis-label-spacing))
20876 (* (1+ (/ height Y-axis-label-spacing))
20877 Y-axis-label-spacing)))
20880 (vertical-step (or vertical-step 1))
20881 (full-Y-label-width
20887 (* height-of-top-line vertical-step))
20893 height-of-top-line full-Y-label-width vertical-step)
20897 numbers-list height-of-top-line symbol-width)
20898 (print-X-axis numbers-list)))
20902 @node Test print-graph, Graphing words in defuns, The final version, Print Whole Graph
20903 @appendixsubsec Testing @code{print-graph}
20906 We can test the @code{print-graph} function with a short list of numbers:
20910 Install the final versions of @code{Y-axis-column},
20911 @code{graph-body-print}, and @code{print-graph} (in addition to the
20915 Copy the following expression:
20918 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1))
20922 Switch to the @file{*scratch*} buffer and place the cursor where you
20923 want the axis labels to start.
20926 Type @kbd{M-:} (@code{eval-expression}).
20929 Yank the test expression into the minibuffer
20930 with @kbd{C-y} (@code{yank)}.
20933 Press @key{RET} to evaluate the expression.
20937 Emacs will print a graph that looks like this:
20958 On the other hand, if you pass @code{print-graph} a
20959 @code{vertical-step} value of 2, by evaluating this expression:
20962 (print-graph '(3 2 5 6 7 5 3 4 6 4 3 2 1) 2)
20967 The graph looks like this:
20988 (A question: is the `2' on the bottom of the vertical axis a bug or a
20989 feature? If you think it is a bug, and should be a `1' instead, (or
20990 even a `0'), you can modify the sources.)
20992 @node Graphing words in defuns, lambda, Test print-graph, Print Whole Graph
20993 @appendixsubsec Graphing Numbers of Words and Symbols
20995 Now for the graph for which all this code was written: a graph that
20996 shows how many function definitions contain fewer than 10 words and
20997 symbols, how many contain between 10 and 19 words and symbols, how
20998 many contain between 20 and 29 words and symbols, and so on.
21000 This is a multi-step process. First make sure you have loaded all the
21004 It is a good idea to reset the value of @code{top-of-ranges} in case
21005 you have set it to some different value. You can evaluate the
21010 (setq top-of-ranges
21013 110 120 130 140 150
21014 160 170 180 190 200
21015 210 220 230 240 250
21016 260 270 280 290 300)
21021 Next create a list of the number of words and symbols in each range.
21025 Evaluate the following:
21029 (setq list-for-graph
21032 (recursive-lengths-list-many-files
21033 (directory-files "/usr/local/emacs/lisp"
21041 On my old machine, this took about an hour. It looked though 303 Lisp
21042 files in my copy of Emacs version 19.23. After all that computing,
21043 the @code{list-for-graph} had this value:
21047 (537 1027 955 785 594 483 349 292 224 199 166 120 116 99
21048 90 80 67 48 52 45 41 33 28 26 25 20 12 28 11 13 220)
21053 This means that my copy of Emacs had 537 function definitions with
21054 fewer than 10 words or symbols in them, 1,027 function definitions
21055 with 10 to 19 words or symbols in them, 955 function definitions with
21056 20 to 29 words or symbols in them, and so on.
21058 Clearly, just by looking at this list we can see that most function
21059 definitions contain ten to thirty words and symbols.
21061 Now for printing. We do @emph{not} want to print a graph that is
21062 1,030 lines high @dots{} Instead, we should print a graph that is
21063 fewer than twenty-five lines high. A graph that height can be
21064 displayed on almost any monitor, and easily printed on a sheet of paper.
21066 This means that each value in @code{list-for-graph} must be reduced to
21067 one-fiftieth its present value.
21069 Here is a short function to do just that, using two functions we have
21070 not yet seen, @code{mapcar} and @code{lambda}.
21074 (defun one-fiftieth (full-range)
21075 "Return list, each number one-fiftieth of previous."
21076 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21080 @node lambda, mapcar, Graphing words in defuns, Print Whole Graph
21081 @appendixsubsec A @code{lambda} Expression: Useful Anonymity
21082 @cindex Anonymous function
21085 @code{lambda} is the symbol for an anonymous function, a function
21086 without a name. Every time you use an anonymous function, you need to
21087 include its whole body.
21094 (lambda (arg) (/ arg 50))
21098 is a function definition that says `return the value resulting from
21099 dividing whatever is passed to me as @code{arg} by 50'.
21102 Earlier, for example, we had a function @code{multiply-by-seven}; it
21103 multiplied its argument by 7. This function is similar, except it
21104 divides its argument by 50; and, it has no name. The anonymous
21105 equivalent of @code{multiply-by-seven} is:
21108 (lambda (number) (* 7 number))
21112 (@xref{defun, , The @code{defun} Special Form}.)
21116 If we want to multiply 3 by 7, we can write:
21118 @c !!! Clear print-postscript-figures if the computer formatting this
21119 @c document is too small and cannot handle all the diagrams and figures.
21120 @c clear print-postscript-figures
21121 @c set print-postscript-figures
21122 @c lambda example diagram #1
21126 (multiply-by-seven 3)
21127 \_______________/ ^
21133 @ifset print-postscript-figures
21136 @center @image{lambda-1}
21137 %%%% old method of including an image
21138 % \input /usr/local/lib/tex/inputs/psfig.tex
21139 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-1.eps}}
21144 @ifclear print-postscript-figures
21148 (multiply-by-seven 3)
21149 \_______________/ ^
21158 This expression returns 21.
21162 Similarly, we can write:
21164 @c lambda example diagram #2
21168 ((lambda (number) (* 7 number)) 3)
21169 \____________________________/ ^
21171 anonymous function argument
21175 @ifset print-postscript-figures
21178 @center @image{lambda-2}
21179 %%%% old method of including an image
21180 % \input /usr/local/lib/tex/inputs/psfig.tex
21181 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-2.eps}}
21186 @ifclear print-postscript-figures
21190 ((lambda (number) (* 7 number)) 3)
21191 \____________________________/ ^
21193 anonymous function argument
21201 If we want to divide 100 by 50, we can write:
21203 @c lambda example diagram #3
21207 ((lambda (arg) (/ arg 50)) 100)
21208 \______________________/ \_/
21210 anonymous function argument
21214 @ifset print-postscript-figures
21217 @center @image{lambda-3}
21218 %%%% old method of including an image
21219 % \input /usr/local/lib/tex/inputs/psfig.tex
21220 % \centerline{\psfig{figure=/usr/local/lib/emacs/man/lambda-3.eps}}
21225 @ifclear print-postscript-figures
21229 ((lambda (arg) (/ arg 50)) 100)
21230 \______________________/ \_/
21232 anonymous function argument
21239 This expression returns 2. The 100 is passed to the function, which
21240 divides that number by 50.
21242 @xref{Lambda Expressions, , Lambda Expressions, elisp, The GNU Emacs
21243 Lisp Reference Manual}, for more about @code{lambda}. Lisp and lambda
21244 expressions derive from the Lambda Calculus.
21246 @node mapcar, Another Bug, lambda, Print Whole Graph
21247 @appendixsubsec The @code{mapcar} Function
21250 @code{mapcar} is a function that calls its first argument with each
21251 element of its second argument, in turn. The second argument must be
21254 The @samp{map} part of the name comes from the mathematical phrase,
21255 `mapping over a domain', meaning to apply a function to each of the
21256 elements in a domain. The mathematical phrase is based on the
21257 metaphor of a surveyor walking, one step at a time, over an area he is
21258 mapping. And @samp{car}, of course, comes from the Lisp notion of the
21267 (mapcar '1+ '(2 4 6))
21273 The function @code{1+} which adds one to its argument, is executed on
21274 @emph{each} element of the list, and a new list is returned.
21276 Contrast this with @code{apply}, which applies its first argument to
21278 (@xref{Readying a Graph, , Readying a Graph}, for a explanation of
21282 In the definition of @code{one-fiftieth}, the first argument is the
21283 anonymous function:
21286 (lambda (arg) (/ arg 50))
21290 and the second argument is @code{full-range}, which will be bound to
21291 @code{list-for-graph}.
21294 The whole expression looks like this:
21297 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21300 @xref{Mapping Functions, , Mapping Functions, elisp, The GNU Emacs
21301 Lisp Reference Manual}, for more about @code{mapcar}.
21303 Using the @code{one-fiftieth} function, we can generate a list in
21304 which each element is one-fiftieth the size of the corresponding
21305 element in @code{list-for-graph}.
21309 (setq fiftieth-list-for-graph
21310 (one-fiftieth list-for-graph))
21315 The resulting list looks like this:
21319 (10 20 19 15 11 9 6 5 4 3 3 2 2
21320 1 1 1 1 0 1 0 0 0 0 0 0 0 0 0 0 0 4)
21325 This, we are almost ready to print! (We also notice the loss of
21326 information: many of the higher ranges are 0, meaning that fewer than
21327 50 defuns had that many words or symbols---but not necessarily meaning
21328 that none had that many words or symbols.)
21330 @node Another Bug, Final printed graph, mapcar, Print Whole Graph
21331 @appendixsubsec Another Bug @dots{} Most Insidious
21332 @cindex Bug, most insidious type
21333 @cindex Insidious type of bug
21335 I said `almost ready to print'! Of course, there is a bug in the
21336 @code{print-graph} function @dots{} It has a @code{vertical-step}
21337 option, but not a @code{horizontal-step} option. The
21338 @code{top-of-range} scale goes from 10 to 300 by tens. But the
21339 @code{print-graph} function will print only by ones.
21341 This is a classic example of what some consider the most insidious
21342 type of bug, the bug of omission. This is not the kind of bug you can
21343 find by studying the code, for it is not in the code; it is an omitted
21344 feature. Your best actions are to try your program early and often;
21345 and try to arrange, as much as you can, to write code that is easy to
21346 understand and easy to change. Try to be aware, whenever you can,
21347 that whatever you have written, @emph{will} be rewritten, if not soon,
21348 eventually. A hard maxim to follow.
21350 It is the @code{print-X-axis-numbered-line} function that needs the
21351 work; and then the @code{print-X-axis} and the @code{print-graph}
21352 functions need to be adapted. Not much needs to be done; there is one
21353 nicety: the numbers ought to line up under the tic marks. This takes
21357 Here is the corrected @code{print-X-axis-numbered-line}:
21361 (defun print-X-axis-numbered-line
21362 (number-of-X-tics X-axis-leading-spaces
21363 &optional horizontal-step)
21364 "Print line of X-axis numbers"
21365 (let ((number X-axis-label-spacing)
21366 (horizontal-step (or horizontal-step 1)))
21369 (insert X-axis-leading-spaces)
21370 ;; @r{Delete extra leading spaces.}
21373 (length (number-to-string horizontal-step)))))
21378 ;; @r{Insert white space.}
21380 X-axis-label-spacing)
21383 (number-to-string horizontal-step)))
21387 (* number horizontal-step))))
21390 ;; @r{Insert remaining numbers.}
21391 (setq number (+ number X-axis-label-spacing))
21392 (while (> number-of-X-tics 1)
21393 (insert (X-axis-element
21394 (* number horizontal-step)))
21395 (setq number (+ number X-axis-label-spacing))
21396 (setq number-of-X-tics (1- number-of-X-tics)))))
21401 If you are reading this in Info, you can see the new versions of
21402 @code{print-X-axis} @code{print-graph} and evaluate them. If you are
21403 reading this in a printed book, you can see the changed lines here
21404 (the full text is too much to print).
21409 (defun print-X-axis (numbers-list horizontal-step)
21411 (print-X-axis-numbered-line
21412 tic-number leading-spaces horizontal-step))
21420 &optional vertical-step horizontal-step)
21422 (print-X-axis numbers-list horizontal-step))
21430 (defun print-X-axis (numbers-list horizontal-step)
21431 "Print X axis labels to length of NUMBERS-LIST.
21432 Optionally, HORIZONTAL-STEP, a positive integer,
21433 specifies how much an X axis label increments for
21437 ;; Value of symbol-width and full-Y-label-width
21438 ;; are passed by `print-graph'.
21439 (let* ((leading-spaces
21440 (make-string full-Y-label-width ? ))
21441 ;; symbol-width @r{is provided by} graph-body-print
21442 (tic-width (* symbol-width X-axis-label-spacing))
21443 (X-length (length numbers-list))
21449 ;; @r{Make a string of blanks.}
21450 (- (* symbol-width X-axis-label-spacing)
21451 (length X-axis-tic-symbol))
21455 ;; @r{Concatenate blanks with tic symbol.}
21456 X-axis-tic-symbol))
21458 (if (zerop (% X-length tic-width))
21459 (/ X-length tic-width)
21460 (1+ (/ X-length tic-width)))))
21464 (print-X-axis-tic-line
21465 tic-number leading-spaces X-tic)
21467 (print-X-axis-numbered-line
21468 tic-number leading-spaces horizontal-step)))
21475 (numbers-list &optional vertical-step horizontal-step)
21476 "Print labelled bar graph of the NUMBERS-LIST.
21477 The numbers-list consists of the Y-axis values.
21481 Optionally, VERTICAL-STEP, a positive integer,
21482 specifies how much a Y axis label increments for
21483 each line. For example, a step of 5 means that
21484 each row is five units.
21488 Optionally, HORIZONTAL-STEP, a positive integer,
21489 specifies how much an X axis label increments for
21491 (let* ((symbol-width (length graph-blank))
21492 ;; @code{height} @r{is both the largest number}
21493 ;; @r{and the number with the most digits.}
21494 (height (apply 'max numbers-list))
21497 (height-of-top-line
21498 (if (zerop (% height Y-axis-label-spacing))
21501 (* (1+ (/ height Y-axis-label-spacing))
21502 Y-axis-label-spacing)))
21505 (vertical-step (or vertical-step 1))
21506 (full-Y-label-width
21510 (* height-of-top-line vertical-step))
21515 height-of-top-line full-Y-label-width vertical-step)
21517 numbers-list height-of-top-line symbol-width)
21518 (print-X-axis numbers-list horizontal-step)))
21525 Graphing Definitions Re-listed
21528 Here are all the graphing definitions in their final form:
21532 (defvar top-of-ranges
21535 110 120 130 140 150
21536 160 170 180 190 200
21537 210 220 230 240 250)
21538 "List specifying ranges for `defuns-per-range'.")
21542 (defvar graph-symbol "*"
21543 "String used as symbol in graph, usually an asterisk.")
21547 (defvar graph-blank " "
21548 "String used as blank in graph, usually a blank space.
21549 graph-blank must be the same number of columns wide
21554 (defvar Y-axis-tic " - "
21555 "String that follows number in a Y axis label.")
21559 (defvar Y-axis-label-spacing 5
21560 "Number of lines from one Y axis label to next.")
21564 (defvar X-axis-tic-symbol "|"
21565 "String to insert to point to a column in X axis.")
21569 (defvar X-axis-label-spacing
21570 (if (boundp 'graph-blank)
21571 (* 5 (length graph-blank)) 5)
21572 "Number of units from one X axis label to next.")
21578 (defun count-words-in-defun ()
21579 "Return the number of words and symbols in a defun."
21580 (beginning-of-defun)
21582 (end (save-excursion (end-of-defun) (point))))
21587 (and (< (point) end)
21589 "\\(\\w\\|\\s_\\)+[^ \t\n]*[ \t\n]*"
21591 (setq count (1+ count)))
21598 (defun lengths-list-file (filename)
21599 "Return list of definitions' lengths within FILE.
21600 The returned list is a list of numbers.
21601 Each number is the number of words or
21602 symbols in one function definition."
21606 (message "Working on `%s' ... " filename)
21608 (let ((buffer (find-file-noselect filename))
21610 (set-buffer buffer)
21611 (setq buffer-read-only t)
21613 (goto-char (point-min))
21617 (while (re-search-forward "^(defun" nil t)
21619 (cons (count-words-in-defun) lengths-list)))
21620 (kill-buffer buffer)
21627 (defun lengths-list-many-files (list-of-files)
21628 "Return list of lengths of defuns in LIST-OF-FILES."
21629 (let (lengths-list)
21630 ;;; @r{true-or-false-test}
21631 (while list-of-files
21637 ;;; @r{Generate a lengths' list.}
21639 (expand-file-name (car list-of-files)))))
21640 ;;; @r{Make files' list shorter.}
21641 (setq list-of-files (cdr list-of-files)))
21642 ;;; @r{Return final value of lengths' list.}
21649 (defun defuns-per-range (sorted-lengths top-of-ranges)
21650 "SORTED-LENGTHS defuns in each TOP-OF-RANGES range."
21651 (let ((top-of-range (car top-of-ranges))
21652 (number-within-range 0)
21653 defuns-per-range-list)
21658 (while top-of-ranges
21662 ;; @r{Need number for numeric test.}
21663 (car sorted-lengths)
21664 (< (car sorted-lengths) top-of-range))
21666 ;; @r{Count number of definitions within current range.}
21667 (setq number-within-range (1+ number-within-range))
21668 (setq sorted-lengths (cdr sorted-lengths)))
21672 ;; @r{Exit inner loop but remain within outer loop.}
21674 (setq defuns-per-range-list
21675 (cons number-within-range defuns-per-range-list))
21676 (setq number-within-range 0) ; @r{Reset count to zero.}
21678 ;; @r{Move to next range.}
21679 (setq top-of-ranges (cdr top-of-ranges))
21680 ;; @r{Specify next top of range value.}
21681 (setq top-of-range (car top-of-ranges)))
21685 ;; @r{Exit outer loop and count the number of defuns larger than}
21686 ;; @r{ the largest top-of-range value.}
21687 (setq defuns-per-range-list
21689 (length sorted-lengths)
21690 defuns-per-range-list))
21692 ;; @r{Return a list of the number of definitions within each range,}
21693 ;; @r{ smallest to largest.}
21694 (nreverse defuns-per-range-list)))
21700 (defun column-of-graph (max-graph-height actual-height)
21701 "Return list of MAX-GRAPH-HEIGHT strings;
21702 ACTUAL-HEIGHT are graph-symbols.
21703 The graph-symbols are contiguous entries at the end
21705 The list will be inserted as one column of a graph.
21706 The strings are either graph-blank or graph-symbol."
21710 (let ((insert-list nil)
21711 (number-of-top-blanks
21712 (- max-graph-height actual-height)))
21714 ;; @r{Fill in @code{graph-symbols}.}
21715 (while (> actual-height 0)
21716 (setq insert-list (cons graph-symbol insert-list))
21717 (setq actual-height (1- actual-height)))
21721 ;; @r{Fill in @code{graph-blanks}.}
21722 (while (> number-of-top-blanks 0)
21723 (setq insert-list (cons graph-blank insert-list))
21724 (setq number-of-top-blanks
21725 (1- number-of-top-blanks)))
21727 ;; @r{Return whole list.}
21734 (defun Y-axis-element (number full-Y-label-width)
21735 "Construct a NUMBERed label element.
21736 A numbered element looks like this ` 5 - ',
21737 and is padded as needed so all line up with
21738 the element for the largest number."
21741 (let* ((leading-spaces
21742 (- full-Y-label-width
21744 (concat (number-to-string number)
21749 (make-string leading-spaces ? )
21750 (number-to-string number)
21757 (defun print-Y-axis
21758 (height full-Y-label-width &optional vertical-step)
21759 "Insert Y axis by HEIGHT and FULL-Y-LABEL-WIDTH.
21760 Height must be the maximum height of the graph.
21761 Full width is the width of the highest label element.
21762 Optionally, print according to VERTICAL-STEP."
21765 ;; Value of height and full-Y-label-width
21766 ;; are passed by `print-graph'.
21767 (let ((start (point)))
21769 (Y-axis-column height full-Y-label-width vertical-step))
21772 ;; @r{Place point ready for inserting graph.}
21774 ;; @r{Move point forward by value of} full-Y-label-width
21775 (forward-char full-Y-label-width)))
21781 (defun print-X-axis-tic-line
21782 (number-of-X-tics X-axis-leading-spaces X-axis-tic-element)
21783 "Print ticks for X axis."
21784 (insert X-axis-leading-spaces)
21785 (insert X-axis-tic-symbol) ; @r{Under first column.}
21788 ;; @r{Insert second tic in the right spot.}
21791 (- (* symbol-width X-axis-label-spacing)
21792 ;; @r{Insert white space up to second tic symbol.}
21793 (* 2 (length X-axis-tic-symbol)))
21795 X-axis-tic-symbol))
21798 ;; @r{Insert remaining ticks.}
21799 (while (> number-of-X-tics 1)
21800 (insert X-axis-tic-element)
21801 (setq number-of-X-tics (1- number-of-X-tics))))
21807 (defun X-axis-element (number)
21808 "Construct a numbered X axis element."
21809 (let ((leading-spaces
21810 (- (* symbol-width X-axis-label-spacing)
21811 (length (number-to-string number)))))
21812 (concat (make-string leading-spaces ? )
21813 (number-to-string number))))
21819 (defun graph-body-print (numbers-list height symbol-width)
21820 "Print a bar graph of the NUMBERS-LIST.
21821 The numbers-list consists of the Y-axis values.
21822 HEIGHT is maximum height of graph.
21823 SYMBOL-WIDTH is number of each column."
21826 (let (from-position)
21827 (while numbers-list
21828 (setq from-position (point))
21830 (column-of-graph height (car numbers-list)))
21831 (goto-char from-position)
21832 (forward-char symbol-width)
21835 ;; @r{Draw graph column by column.}
21837 (setq numbers-list (cdr numbers-list)))
21838 ;; @r{Place point for X axis labels.}
21839 (forward-line height)
21846 (defun Y-axis-column
21847 (height width-of-label &optional vertical-step)
21848 "Construct list of labels for Y axis.
21849 HEIGHT is maximum height of graph.
21850 WIDTH-OF-LABEL is maximum width of label.
21853 VERTICAL-STEP, an option, is a positive integer
21854 that specifies how much a Y axis label increments
21855 for each line. For example, a step of 5 means
21856 that each line is five units of the graph."
21858 (number-per-line (or vertical-step 1)))
21861 (while (> height 1)
21862 (if (zerop (% height Y-axis-label-spacing))
21863 ;; @r{Insert label.}
21867 (* height number-per-line)
21872 ;; @r{Else, insert blanks.}
21875 (make-string width-of-label ? )
21877 (setq height (1- height)))
21880 ;; @r{Insert base line.}
21881 (setq Y-axis (cons (Y-axis-element
21882 (or vertical-step 1)
21885 (nreverse Y-axis)))
21891 (defun print-X-axis-numbered-line
21892 (number-of-X-tics X-axis-leading-spaces
21893 &optional horizontal-step)
21894 "Print line of X-axis numbers"
21895 (let ((number X-axis-label-spacing)
21896 (horizontal-step (or horizontal-step 1)))
21899 (insert X-axis-leading-spaces)
21901 (delete-char (- (1- (length (number-to-string horizontal-step)))))
21904 ;; @r{Insert white space up to next number.}
21905 (- (* symbol-width X-axis-label-spacing)
21906 (1- (length (number-to-string horizontal-step)))
21909 (number-to-string (* number horizontal-step))))
21912 ;; @r{Insert remaining numbers.}
21913 (setq number (+ number X-axis-label-spacing))
21914 (while (> number-of-X-tics 1)
21915 (insert (X-axis-element (* number horizontal-step)))
21916 (setq number (+ number X-axis-label-spacing))
21917 (setq number-of-X-tics (1- number-of-X-tics)))))
21923 (defun print-X-axis (numbers-list horizontal-step)
21924 "Print X axis labels to length of NUMBERS-LIST.
21925 Optionally, HORIZONTAL-STEP, a positive integer,
21926 specifies how much an X axis label increments for
21930 ;; Value of symbol-width and full-Y-label-width
21931 ;; are passed by `print-graph'.
21932 (let* ((leading-spaces
21933 (make-string full-Y-label-width ? ))
21934 ;; symbol-width @r{is provided by} graph-body-print
21935 (tic-width (* symbol-width X-axis-label-spacing))
21936 (X-length (length numbers-list))
21942 ;; @r{Make a string of blanks.}
21943 (- (* symbol-width X-axis-label-spacing)
21944 (length X-axis-tic-symbol))
21948 ;; @r{Concatenate blanks with tic symbol.}
21949 X-axis-tic-symbol))
21951 (if (zerop (% X-length tic-width))
21952 (/ X-length tic-width)
21953 (1+ (/ X-length tic-width)))))
21957 (print-X-axis-tic-line
21958 tic-number leading-spaces X-tic)
21960 (print-X-axis-numbered-line
21961 tic-number leading-spaces horizontal-step)))
21967 (defun one-fiftieth (full-range)
21968 "Return list, each number of which is 1/50th previous."
21969 (mapcar '(lambda (arg) (/ arg 50)) full-range))
21976 (numbers-list &optional vertical-step horizontal-step)
21977 "Print labelled bar graph of the NUMBERS-LIST.
21978 The numbers-list consists of the Y-axis values.
21982 Optionally, VERTICAL-STEP, a positive integer,
21983 specifies how much a Y axis label increments for
21984 each line. For example, a step of 5 means that
21985 each row is five units.
21989 Optionally, HORIZONTAL-STEP, a positive integer,
21990 specifies how much an X axis label increments for
21992 (let* ((symbol-width (length graph-blank))
21993 ;; @code{height} @r{is both the largest number}
21994 ;; @r{and the number with the most digits.}
21995 (height (apply 'max numbers-list))
21998 (height-of-top-line
21999 (if (zerop (% height Y-axis-label-spacing))
22002 (* (1+ (/ height Y-axis-label-spacing))
22003 Y-axis-label-spacing)))
22006 (vertical-step (or vertical-step 1))
22007 (full-Y-label-width
22011 (* height-of-top-line vertical-step))
22017 height-of-top-line full-Y-label-width vertical-step)
22019 numbers-list height-of-top-line symbol-width)
22020 (print-X-axis numbers-list horizontal-step)))
22027 @node Final printed graph, , Another Bug, Print Whole Graph
22028 @appendixsubsec The Printed Graph
22030 When made and installed, you can call the @code{print-graph} command
22036 (print-graph fiftieth-list-for-graph 50 10)
22066 50 - ***************** * *
22068 10 50 100 150 200 250 300 350
22075 The largest group of functions contain 10 -- 19 words and symbols each.
22077 @node Free Software and Free Manuals, GNU Free Documentation License, Full Graph, Top
22078 @appendix Free Software and Free Manuals
22080 @strong{by Richard M. Stallman}
22083 The biggest deficiency in free operating systems is not in the
22084 software---it is the lack of good free manuals that we can include in
22085 these systems. Many of our most important programs do not come with
22086 full manuals. Documentation is an essential part of any software
22087 package; when an important free software package does not come with a
22088 free manual, that is a major gap. We have many such gaps today.
22090 Once upon a time, many years ago, I thought I would learn Perl. I got
22091 a copy of a free manual, but I found it hard to read. When I asked
22092 Perl users about alternatives, they told me that there were better
22093 introductory manuals---but those were not free.
22095 Why was this? The authors of the good manuals had written them for
22096 O'Reilly Associates, which published them with restrictive terms---no
22097 copying, no modification, source files not available---which exclude
22098 them from the free software community.
22100 That wasn't the first time this sort of thing has happened, and (to
22101 our community's great loss) it was far from the last. Proprietary
22102 manual publishers have enticed a great many authors to restrict their
22103 manuals since then. Many times I have heard a GNU user eagerly tell me
22104 about a manual that he is writing, with which he expects to help the
22105 GNU project---and then had my hopes dashed, as he proceeded to explain
22106 that he had signed a contract with a publisher that would restrict it
22107 so that we cannot use it.
22109 Given that writing good English is a rare skill among programmers, we
22110 can ill afford to lose manuals this way.
22112 Free documentation, like free software, is a matter of freedom, not
22113 price. The problem with these manuals was not that O'Reilly Associates
22114 charged a price for printed copies---that in itself is fine. The Free
22115 Software Foundation @uref{http://shop.fsf.org, sells printed copies} of
22116 free @uref{http://www.gnu.org/doc/doc.html, GNU manuals}, too.
22117 But GNU manuals are available in source code form, while these manuals
22118 are available only on paper. GNU manuals come with permission to copy
22119 and modify; the Perl manuals do not. These restrictions are the
22122 The criterion for a free manual is pretty much the same as for free
22123 software: it is a matter of giving all users certain
22124 freedoms. Redistribution (including commercial redistribution) must be
22125 permitted, so that the manual can accompany every copy of the program,
22126 on-line or on paper. Permission for modification is crucial too.
22128 As a general rule, I don't believe that it is essential for people to
22129 have permission to modify all sorts of articles and books. The issues
22130 for writings are not necessarily the same as those for software. For
22131 example, I don't think you or I are obliged to give permission to
22132 modify articles like this one, which describe our actions and our
22135 But there is a particular reason why the freedom to modify is crucial
22136 for documentation for free software. When people exercise their right
22137 to modify the software, and add or change its features, if they are
22138 conscientious they will change the manual too---so they can provide
22139 accurate and usable documentation with the modified program. A manual
22140 which forbids programmers to be conscientious and finish the job, or
22141 more precisely requires them to write a new manual from scratch if
22142 they change the program, does not fill our community's needs.
22144 While a blanket prohibition on modification is unacceptable, some
22145 kinds of limits on the method of modification pose no problem. For
22146 example, requirements to preserve the original author's copyright
22147 notice, the distribution terms, or the list of authors, are ok. It is
22148 also no problem to require modified versions to include notice that
22149 they were modified, even to have entire sections that may not be
22150 deleted or changed, as long as these sections deal with nontechnical
22151 topics. (Some GNU manuals have them.)
22153 These kinds of restrictions are not a problem because, as a practical
22154 matter, they don't stop the conscientious programmer from adapting the
22155 manual to fit the modified program. In other words, they don't block
22156 the free software community from making full use of the manual.
22158 However, it must be possible to modify all the technical content of
22159 the manual, and then distribute the result in all the usual media,
22160 through all the usual channels; otherwise, the restrictions do block
22161 the community, the manual is not free, and so we need another manual.
22163 Unfortunately, it is often hard to find someone to write another
22164 manual when a proprietary manual exists. The obstacle is that many
22165 users think that a proprietary manual is good enough---so they don't
22166 see the need to write a free manual. They do not see that the free
22167 operating system has a gap that needs filling.
22169 Why do users think that proprietary manuals are good enough? Some have
22170 not considered the issue. I hope this article will do something to
22173 Other users consider proprietary manuals acceptable for the same
22174 reason so many people consider proprietary software acceptable: they
22175 judge in purely practical terms, not using freedom as a
22176 criterion. These people are entitled to their opinions, but since
22177 those opinions spring from values which do not include freedom, they
22178 are no guide for those of us who do value freedom.
22180 Please spread the word about this issue. We continue to lose manuals
22181 to proprietary publishing. If we spread the word that proprietary
22182 manuals are not sufficient, perhaps the next person who wants to help
22183 GNU by writing documentation will realize, before it is too late, that
22184 he must above all make it free.
22186 We can also encourage commercial publishers to sell free, copylefted
22187 manuals instead of proprietary ones. One way you can help this is to
22188 check the distribution terms of a manual before you buy it, and prefer
22189 copylefted manuals to non-copylefted ones.
22193 Note: The Free Software Foundation maintains a page on its Web site
22194 that lists free books available from other publishers:@*
22195 @uref{http://www.gnu.org/doc/other-free-books.html}
22197 @node GNU Free Documentation License, Index, Free Software and Free Manuals, Top
22198 @appendix GNU Free Documentation License
22200 @cindex FDL, GNU Free Documentation License
22201 @include doclicense.texi
22203 @node Index, About the Author, GNU Free Documentation License, Top
22204 @comment node-name, next, previous, up
22208 MENU ENTRY: NODE NAME.
22214 @c Place biographical information on right-hand (verso) page
22217 \par\vfill\supereject
22219 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22220 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22223 % \par\vfill\supereject
22224 \global\evenheadline={\hfil} \global\evenfootline={\hfil}
22225 \global\oddheadline={\hfil} \global\oddfootline={\hfil}
22226 %\page\hbox{}%\page
22227 %\page\hbox{}%\page
22234 @c ================ Biographical information ================
22238 @center About the Author
22243 @node About the Author, , Index, Top
22244 @unnumbered About the Author
22248 Robert J. Chassell has worked with GNU Emacs since 1985. He writes
22249 and edits, teaches Emacs and Emacs Lisp, and speaks throughout the
22250 world on software freedom. Chassell was a founding Director and
22251 Treasurer of the Free Software Foundation, Inc. He is co-author of
22252 the @cite{Texinfo} manual, and has edited more than a dozen other
22253 books. He graduated from Cambridge University, in England. He has an
22254 abiding interest in social and economic history and flies his own
22261 @c @c Prevent page number on blank verso, so eject it first.
22263 @c \par\vfill\supereject
22268 @c @evenheading @thispage @| @| @thistitle
22269 @c @oddheading @| @| @thispage
22275 arch-tag: da1a2154-531f-43a8-8e33-fc7faad10acf