1 \input texinfo @c -*-texinfo-*-
3 @setfilename standards.info
4 @settitle GNU Coding Standards
5 @c This date is automagically updated when you save this file:
6 @set lastupdate March 23, 2010
9 @dircategory GNU organization
11 * Standards: (standards). GNU coding standards.
14 @c @setchapternewpage odd
15 @setchapternewpage off
17 @c Put everything in one index (arbitrarily chosen to be the concept index).
23 @c This is used by a cross ref in make-stds.texi
27 The GNU coding standards, last updated @value{lastupdate}.
29 Copyright @copyright{} 1992, 1993, 1994, 1995, 1996, 1997, 1998, 1999,
30 2000, 2001, 2002, 2003, 2004, 2005, 2006, 2007, 2008, 2009, 2010 Free Software
33 Permission is granted to copy, distribute and/or modify this document
34 under the terms of the GNU Free Documentation License, Version 1.3 or
35 any later version published by the Free Software Foundation; with no
36 Invariant Sections, with no Front-Cover Texts, and with no Back-Cover
37 Texts. A copy of the license is included in the section entitled
38 ``GNU Free Documentation License''.
42 @title GNU Coding Standards
43 @author Richard Stallman, et al.
44 @author last updated @value{lastupdate}
46 @vskip 0pt plus 1filll
53 @node Top, Preface, (dir), (dir)
60 * Preface:: About the GNU Coding Standards.
61 * Legal Issues:: Keeping free software free.
62 * Design Advice:: General program design.
63 * Program Behavior:: Program behavior for all programs
64 * Writing C:: Making the best use of C.
65 * Documentation:: Documenting programs.
66 * Managing Releases:: The release process.
67 * References:: Mentioning non-free software or documentation.
68 * GNU Free Documentation License:: Copying and sharing this manual.
74 @chapter About the GNU Coding Standards
76 The GNU Coding Standards were written by Richard Stallman and other GNU
77 Project volunteers. Their purpose is to make the GNU system clean,
78 consistent, and easy to install. This document can also be read as a
79 guide to writing portable, robust and reliable programs. It focuses on
80 programs written in C, but many of the rules and principles are useful
81 even if you write in another programming language. The rules often
82 state reasons for writing in a certain way.
84 @cindex where to obtain @code{standards.texi}
85 @cindex downloading this manual
86 If you did not obtain this file directly from the GNU project and
87 recently, please check for a newer version. You can get the GNU
88 Coding Standards from the GNU web server in many
89 different formats, including the Texinfo source, PDF, HTML, DVI, plain
90 text, and more, at: @uref{http://www.gnu.org/prep/standards/}.
92 If you are maintaining an official GNU package, in addition to this
93 document, please read and follow the GNU maintainer information
94 (@pxref{Top, , Contents, maintain, Information for Maintainers of GNU
97 @cindex @code{gnustandards-commit@@gnu.org} mailing list
98 If you want to receive diffs for every change to these GNU documents,
99 join the mailing list @code{gnustandards-commit@@gnu.org}, via the web
101 @url{http://lists.gnu.org/mailman/listinfo/gnustandards-commit}.
102 Archives are also available there.
104 @cindex @code{maintainers@@gnu.org} email address
105 Please send corrections or suggestions for this document to
106 @email{bug-standards@@gnu.org}. If you make a suggestion, please
107 include a suggested new wording for it, to help us consider the
108 suggestion efficiently. We prefer a context diff to the Texinfo
109 source, but if that's difficult for you, you can make a context diff
110 for some other version of this document, or propose it in any way that
111 makes it clear. The source repository for this document can be found
112 at @url{http://savannah.gnu.org/projects/gnustandards}.
114 These standards cover the minimum of what is important when writing a
115 GNU package. Likely, the need for additional standards will come up.
116 Sometimes, you might suggest that such standards be added to this
117 document. If you think your standards would be generally useful, please
120 You should also set standards for your package on many questions not
121 addressed or not firmly specified here. The most important point is to
122 be self-consistent---try to stick to the conventions you pick, and try
123 to document them as much as possible. That way, your program will be
124 more maintainable by others.
126 The GNU Hello program serves as an example of how to follow the GNU
127 coding standards for a trivial program.
128 @uref{http://www.gnu.org/software/hello/hello.html}.
130 This release of the GNU Coding Standards was last updated
135 @chapter Keeping Free Software Free
136 @cindex legal aspects
138 This chapter discusses how you can make sure that GNU software
139 avoids legal difficulties, and other related issues.
142 * Reading Non-Free Code:: Referring to proprietary programs.
143 * Contributions:: Accepting contributions.
144 * Trademarks:: How we deal with trademark issues.
147 @node Reading Non-Free Code
148 @section Referring to Proprietary Programs
149 @cindex proprietary programs
150 @cindex avoiding proprietary code
152 Don't in any circumstances refer to Unix source code for or during
153 your work on GNU! (Or to any other proprietary programs.)
155 If you have a vague recollection of the internals of a Unix program,
156 this does not absolutely mean you can't write an imitation of it, but
157 do try to organize the imitation internally along different lines,
158 because this is likely to make the details of the Unix version
159 irrelevant and dissimilar to your results.
161 For example, Unix utilities were generally optimized to minimize
162 memory use; if you go for speed instead, your program will be very
163 different. You could keep the entire input file in memory and scan it
164 there instead of using stdio. Use a smarter algorithm discovered more
165 recently than the Unix program. Eliminate use of temporary files. Do
166 it in one pass instead of two (we did this in the assembler).
168 Or, on the contrary, emphasize simplicity instead of speed. For some
169 applications, the speed of today's computers makes simpler algorithms
172 Or go for generality. For example, Unix programs often have static
173 tables or fixed-size strings, which make for arbitrary limits; use
174 dynamic allocation instead. Make sure your program handles NULs and
175 other funny characters in the input files. Add a programming language
176 for extensibility and write part of the program in that language.
178 Or turn some parts of the program into independently usable libraries.
179 Or use a simple garbage collector instead of tracking precisely when
180 to free memory, or use a new GNU facility such as obstacks.
183 @section Accepting Contributions
185 @cindex accepting contributions
187 If the program you are working on is copyrighted by the Free Software
188 Foundation, then when someone else sends you a piece of code to add to
189 the program, we need legal papers to use it---just as we asked you to
190 sign papers initially. @emph{Each} person who makes a nontrivial
191 contribution to a program must sign some sort of legal papers in order
192 for us to have clear title to the program; the main author alone is not
195 So, before adding in any contributions from other people, please tell
196 us, so we can arrange to get the papers. Then wait until we tell you
197 that we have received the signed papers, before you actually use the
200 This applies both before you release the program and afterward. If
201 you receive diffs to fix a bug, and they make significant changes, we
202 need legal papers for that change.
204 This also applies to comments and documentation files. For copyright
205 law, comments and code are just text. Copyright applies to all kinds of
206 text, so we need legal papers for all kinds.
208 We know it is frustrating to ask for legal papers; it's frustrating for
209 us as well. But if you don't wait, you are going out on a limb---for
210 example, what if the contributor's employer won't sign a disclaimer?
211 You might have to take that code out again!
213 You don't need papers for changes of a few lines here or there, since
214 they are not significant for copyright purposes. Also, you don't need
215 papers if all you get from the suggestion is some ideas, not actual code
216 which you use. For example, if someone sent you one implementation, but
217 you write a different implementation of the same idea, you don't need to
220 The very worst thing is if you forget to tell us about the other
221 contributor. We could be very embarrassed in court some day as a
224 We have more detailed advice for maintainers of programs; if you have
225 reached the stage of actually maintaining a program for GNU (whether
226 released or not), please ask us for a copy. It is also available
227 online for your perusal: @uref{http://www.gnu.org/prep/maintain/}.
233 Please do not include any trademark acknowledgements in GNU software
234 packages or documentation.
236 Trademark acknowledgements are the statements that such-and-such is a
237 trademark of so-and-so. The GNU Project has no objection to the basic
238 idea of trademarks, but these acknowledgements feel like kowtowing,
239 and there is no legal requirement for them, so we don't use them.
241 What is legally required, as regards other people's trademarks, is to
242 avoid using them in ways which a reader might reasonably understand as
243 naming or labeling our own programs or activities. For example, since
244 ``Objective C'' is (or at least was) a trademark, we made sure to say
245 that we provide a ``compiler for the Objective C language'' rather
246 than an ``Objective C compiler''. The latter would have been meant as
247 a shorter way of saying the former, but it does not explicitly state
248 the relationship, so it could be misinterpreted as using ``Objective
249 C'' as a label for the compiler rather than for the language.
251 Please don't use ``win'' as an abbreviation for Microsoft Windows in
252 GNU software or documentation. In hacker terminology, calling
253 something a ``win'' is a form of praise. If you wish to praise
254 Microsoft Windows when speaking on your own, by all means do so, but
255 not in GNU software. Usually we write the name ``Windows'' in full,
256 but when brevity is very important (as in file names and sometimes
257 symbol names), we abbreviate it to ``w''. For instance, the files and
258 functions in Emacs that deal with Windows start with @samp{w32}.
261 @chapter General Program Design
262 @cindex program design
264 This chapter discusses some of the issues you should take into
265 account when designing your program.
267 @c Standard or ANSI C
269 @c In 1989 the American National Standards Institute (ANSI) standardized
270 @c C as standard X3.159-1989. In December of that year the
271 @c International Standards Organization ISO adopted the ANSI C standard
272 @c making minor changes. In 1990 ANSI then re-adopted ISO standard
273 @c C. This version of C is known as either ANSI C or Standard C.
275 @c A major revision of the C Standard appeared in 1999.
278 * Source Language:: Which languages to use.
279 * Compatibility:: Compatibility with other implementations.
280 * Using Extensions:: Using non-standard features.
281 * Standard C:: Using standard C features.
282 * Conditional Compilation:: Compiling code only if a conditional is true.
285 @node Source Language
286 @section Which Languages to Use
287 @cindex programming languages
289 When you want to use a language that gets compiled and runs at high
290 speed, the best language to use is C. Using another language is like
291 using a non-standard feature: it will cause trouble for users. Even if
292 GCC supports the other language, users may find it inconvenient to have
293 to install the compiler for that other language in order to build your
294 program. For example, if you write your program in C++, people will
295 have to install the GNU C++ compiler in order to compile your program.
297 C has one other advantage over C++ and other compiled languages: more
298 people know C, so more people will find it easy to read and modify the
299 program if it is written in C.
301 So in general it is much better to use C, rather than the
302 comparable alternatives.
304 But there are two exceptions to that conclusion:
308 It is no problem to use another language to write a tool specifically
309 intended for use with that language. That is because the only people
310 who want to build the tool will be those who have installed the other
314 If an application is of interest only to a narrow part of the community,
315 then the question of which language it is written in has less effect on
316 other people, so you may as well please yourself.
319 Many programs are designed to be extensible: they include an interpreter
320 for a language that is higher level than C. Often much of the program
321 is written in that language, too. The Emacs editor pioneered this
325 @cindex GNOME and Guile
326 The standard extensibility interpreter for GNU software is Guile
327 (@uref{http://www.gnu.org/@/software/@/guile/}), which implements the
328 language Scheme (an especially clean and simple dialect of Lisp).
329 Guile also includes bindings for GTK+/GNOME, making it practical to
330 write modern GUI functionality within Guile. We don't reject programs
331 written in other ``scripting languages'' such as Perl and Python, but
332 using Guile is very important for the overall consistency of the GNU
337 @section Compatibility with Other Implementations
338 @cindex compatibility with C and @sc{posix} standards
339 @cindex @sc{posix} compatibility
341 With occasional exceptions, utility programs and libraries for GNU
342 should be upward compatible with those in Berkeley Unix, and upward
343 compatible with Standard C if Standard C specifies their
344 behavior, and upward compatible with @sc{posix} if @sc{posix} specifies
347 When these standards conflict, it is useful to offer compatibility
348 modes for each of them.
350 @cindex options for compatibility
351 Standard C and @sc{posix} prohibit many kinds of extensions. Feel
352 free to make the extensions anyway, and include a @samp{--ansi},
353 @samp{--posix}, or @samp{--compatible} option to turn them off.
354 However, if the extension has a significant chance of breaking any real
355 programs or scripts, then it is not really upward compatible. So you
356 should try to redesign its interface to make it upward compatible.
358 @cindex @code{POSIXLY_CORRECT}, environment variable
359 Many GNU programs suppress extensions that conflict with @sc{posix} if the
360 environment variable @code{POSIXLY_CORRECT} is defined (even if it is
361 defined with a null value). Please make your program recognize this
362 variable if appropriate.
364 When a feature is used only by users (not by programs or command
365 files), and it is done poorly in Unix, feel free to replace it
366 completely with something totally different and better. (For example,
367 @code{vi} is replaced with Emacs.) But it is nice to offer a compatible
368 feature as well. (There is a free @code{vi} clone, so we offer it.)
370 Additional useful features are welcome regardless of whether
371 there is any precedent for them.
373 @node Using Extensions
374 @section Using Non-standard Features
375 @cindex non-standard extensions
377 Many GNU facilities that already exist support a number of convenient
378 extensions over the comparable Unix facilities. Whether to use these
379 extensions in implementing your program is a difficult question.
381 On the one hand, using the extensions can make a cleaner program.
382 On the other hand, people will not be able to build the program
383 unless the other GNU tools are available. This might cause the
384 program to work on fewer kinds of machines.
386 With some extensions, it might be easy to provide both alternatives.
387 For example, you can define functions with a ``keyword'' @code{INLINE}
388 and define that as a macro to expand into either @code{inline} or
389 nothing, depending on the compiler.
391 In general, perhaps it is best not to use the extensions if you can
392 straightforwardly do without them, but to use the extensions if they
393 are a big improvement.
395 An exception to this rule are the large, established programs (such as
396 Emacs) which run on a great variety of systems. Using GNU extensions in
397 such programs would make many users unhappy, so we don't do that.
399 Another exception is for programs that are used as part of compilation:
400 anything that must be compiled with other compilers in order to
401 bootstrap the GNU compilation facilities. If these require the GNU
402 compiler, then no one can compile them without having them installed
403 already. That would be extremely troublesome in certain cases.
406 @section Standard C and Pre-Standard C
407 @cindex @sc{ansi} C standard
409 1989 Standard C is widespread enough now that it is ok to use its
410 features in new programs. There is one exception: do not ever use the
411 ``trigraph'' feature of Standard C.
413 1999 Standard C is not widespread yet, so please do not require its
414 features in programs. It is ok to use its features if they are present.
416 However, it is easy to support pre-standard compilers in most programs,
417 so if you know how to do that, feel free. If a program you are
418 maintaining has such support, you should try to keep it working.
420 @cindex function prototypes
421 To support pre-standard C, instead of writing function definitions in
422 standard prototype form,
431 write the definition in pre-standard style like this,
441 and use a separate declaration to specify the argument prototype:
447 You need such a declaration anyway, in a header file, to get the benefit
448 of prototypes in all the files where the function is called. And once
449 you have the declaration, you normally lose nothing by writing the
450 function definition in the pre-standard style.
452 This technique does not work for integer types narrower than @code{int}.
453 If you think of an argument as being of a type narrower than @code{int},
454 declare it as @code{int} instead.
456 There are a few special cases where this technique is hard to use. For
457 example, if a function argument needs to hold the system type
458 @code{dev_t}, you run into trouble, because @code{dev_t} is shorter than
459 @code{int} on some machines; but you cannot use @code{int} instead,
460 because @code{dev_t} is wider than @code{int} on some machines. There
461 is no type you can safely use on all machines in a non-standard
462 definition. The only way to support non-standard C and pass such an
463 argument is to check the width of @code{dev_t} using Autoconf and choose
464 the argument type accordingly. This may not be worth the trouble.
466 In order to support pre-standard compilers that do not recognize
467 prototypes, you may want to use a preprocessor macro like this:
470 /* Declare the prototype for a general external function. */
471 #if defined (__STDC__) || defined (WINDOWSNT)
472 #define P_(proto) proto
478 @node Conditional Compilation
479 @section Conditional Compilation
481 When supporting configuration options already known when building your
482 program we prefer using @code{if (... )} over conditional compilation,
483 as in the former case the compiler is able to perform more extensive
484 checking of all possible code paths.
486 For example, please write
506 A modern compiler such as GCC will generate exactly the same code in
507 both cases, and we have been using similar techniques with good success
508 in several projects. Of course, the former method assumes that
509 @code{HAS_FOO} is defined as either 0 or 1.
511 While this is not a silver bullet solving all portability problems,
512 and is not always appropriate, following this policy would have saved
513 GCC developers many hours, or even days, per year.
515 In the case of function-like macros like @code{REVERSIBLE_CC_MODE} in
516 GCC which cannot be simply used in @code{if (...)} statements, there is
517 an easy workaround. Simply introduce another macro
518 @code{HAS_REVERSIBLE_CC_MODE} as in the following example:
521 #ifdef REVERSIBLE_CC_MODE
522 #define HAS_REVERSIBLE_CC_MODE 1
524 #define HAS_REVERSIBLE_CC_MODE 0
528 @node Program Behavior
529 @chapter Program Behavior for All Programs
531 This chapter describes conventions for writing robust
532 software. It also describes general standards for error messages, the
533 command line interface, and how libraries should behave.
536 * Non-GNU Standards:: We consider standards such as POSIX;
537 we don't "obey" them.
538 * Semantics:: Writing robust programs.
539 * Libraries:: Library behavior.
540 * Errors:: Formatting error messages.
541 * User Interfaces:: Standards about interfaces generally.
542 * Graphical Interfaces:: Standards for graphical interfaces.
543 * Command-Line Interfaces:: Standards for command line interfaces.
544 * Option Table:: Table of long options.
545 * OID Allocations:: Table of OID slots for GNU.
546 * Memory Usage:: When and how to care about memory needs.
547 * File Usage:: Which files to use, and where.
550 @node Non-GNU Standards
551 @section Non-GNU Standards
553 The GNU Project regards standards published by other organizations as
554 suggestions, not orders. We consider those standards, but we do not
555 ``obey'' them. In developing a GNU program, you should implement
556 an outside standard's specifications when that makes the GNU system
557 better overall in an objective sense. When it doesn't, you shouldn't.
559 In most cases, following published standards is convenient for
560 users---it means that their programs or scripts will work more
561 portably. For instance, GCC implements nearly all the features of
562 Standard C as specified by that standard. C program developers would
563 be unhappy if it did not. And GNU utilities mostly follow
564 specifications of POSIX.2; shell script writers and users would be
565 unhappy if our programs were incompatible.
567 But we do not follow either of these specifications rigidly, and there
568 are specific points on which we decided not to follow them, so as to
569 make the GNU system better for users.
571 For instance, Standard C says that nearly all extensions to C are
572 prohibited. How silly! GCC implements many extensions, some of which
573 were later adopted as part of the standard. If you want these
574 constructs to give an error message as ``required'' by the standard,
575 you must specify @samp{--pedantic}, which was implemented only so that
576 we can say ``GCC is a 100% implementation of the standard,'' not
577 because there is any reason to actually use it.
579 POSIX.2 specifies that @samp{df} and @samp{du} must output sizes by
580 default in units of 512 bytes. What users want is units of 1k, so
581 that is what we do by default. If you want the ridiculous behavior
582 ``required'' by POSIX, you must set the environment variable
583 @samp{POSIXLY_CORRECT} (which was originally going to be named
584 @samp{POSIX_ME_HARDER}).
586 GNU utilities also depart from the letter of the POSIX.2 specification
587 when they support long-named command-line options, and intermixing
588 options with ordinary arguments. This minor incompatibility with
589 POSIX is never a problem in practice, and it is very useful.
591 In particular, don't reject a new feature, or remove an old one,
592 merely because a standard says it is ``forbidden'' or ``deprecated.''
595 @section Writing Robust Programs
597 @cindex arbitrary limits on data
598 Avoid arbitrary limits on the length or number of @emph{any} data
599 structure, including file names, lines, files, and symbols, by allocating
600 all data structures dynamically. In most Unix utilities, ``long lines
601 are silently truncated''. This is not acceptable in a GNU utility.
603 @cindex @code{NUL} characters
604 Utilities reading files should not drop NUL characters, or any other
605 nonprinting characters @emph{including those with codes above 0177}.
606 The only sensible exceptions would be utilities specifically intended
607 for interface to certain types of terminals or printers
608 that can't handle those characters.
609 Whenever possible, try to make programs work properly with
610 sequences of bytes that represent multibyte characters, using encodings
611 such as UTF-8 and others.
613 @cindex error messages
614 Check every system call for an error return, unless you know you wish to
615 ignore errors. Include the system error text (from @code{perror} or
616 equivalent) in @emph{every} error message resulting from a failing
617 system call, as well as the name of the file if any and the name of the
618 utility. Just ``cannot open foo.c'' or ``stat failed'' is not
621 @cindex @code{malloc} return value
622 @cindex memory allocation failure
623 Check every call to @code{malloc} or @code{realloc} to see if it
624 returned zero. Check @code{realloc} even if you are making the block
625 smaller; in a system that rounds block sizes to a power of 2,
626 @code{realloc} may get a different block if you ask for less space.
628 In Unix, @code{realloc} can destroy the storage block if it returns
629 zero. GNU @code{realloc} does not have this bug: if it fails, the
630 original block is unchanged. Feel free to assume the bug is fixed. If
631 you wish to run your program on Unix, and wish to avoid lossage in this
632 case, you can use the GNU @code{malloc}.
634 You must expect @code{free} to alter the contents of the block that was
635 freed. Anything you want to fetch from the block, you must fetch before
638 If @code{malloc} fails in a noninteractive program, make that a fatal
639 error. In an interactive program (one that reads commands from the
640 user), it is better to abort the command and return to the command
641 reader loop. This allows the user to kill other processes to free up
642 virtual memory, and then try the command again.
644 @cindex command-line arguments, decoding
645 Use @code{getopt_long} to decode arguments, unless the argument syntax
646 makes this unreasonable.
648 When static storage is to be written in during program execution, use
649 explicit C code to initialize it. Reserve C initialized declarations
650 for data that will not be changed.
653 Try to avoid low-level interfaces to obscure Unix data structures (such
654 as file directories, utmp, or the layout of kernel memory), since these
655 are less likely to work compatibly. If you need to find all the files
656 in a directory, use @code{readdir} or some other high-level interface.
657 These are supported compatibly by GNU.
659 @cindex signal handling
660 The preferred signal handling facilities are the BSD variant of
661 @code{signal}, and the @sc{posix} @code{sigaction} function; the
662 alternative USG @code{signal} interface is an inferior design.
664 Nowadays, using the @sc{posix} signal functions may be the easiest way
665 to make a program portable. If you use @code{signal}, then on GNU/Linux
666 systems running GNU libc version 1, you should include
667 @file{bsd/signal.h} instead of @file{signal.h}, so as to get BSD
668 behavior. It is up to you whether to support systems where
669 @code{signal} has only the USG behavior, or give up on them.
671 @cindex impossible conditions
672 In error checks that detect ``impossible'' conditions, just abort.
673 There is usually no point in printing any message. These checks
674 indicate the existence of bugs. Whoever wants to fix the bugs will have
675 to read the source code and run a debugger. So explain the problem with
676 comments in the source. The relevant data will be in variables, which
677 are easy to examine with the debugger, so there is no point moving them
680 Do not use a count of errors as the exit status for a program.
681 @emph{That does not work}, because exit status values are limited to 8
682 bits (0 through 255). A single run of the program might have 256
683 errors; if you try to return 256 as the exit status, the parent process
684 will see 0 as the status, and it will appear that the program succeeded.
686 @cindex temporary files
687 @cindex @code{TMPDIR} environment variable
688 If you make temporary files, check the @code{TMPDIR} environment
689 variable; if that variable is defined, use the specified directory
690 instead of @file{/tmp}.
692 In addition, be aware that there is a possible security problem when
693 creating temporary files in world-writable directories. In C, you can
694 avoid this problem by creating temporary files in this manner:
697 fd = open (filename, O_WRONLY | O_CREAT | O_EXCL, 0600);
701 or by using the @code{mkstemps} function from libiberty.
703 In bash, use @code{set -C} to avoid this problem.
706 @section Library Behavior
709 Try to make library functions reentrant. If they need to do dynamic
710 storage allocation, at least try to avoid any nonreentrancy aside from
711 that of @code{malloc} itself.
713 Here are certain name conventions for libraries, to avoid name
716 Choose a name prefix for the library, more than two characters long.
717 All external function and variable names should start with this
718 prefix. In addition, there should only be one of these in any given
719 library member. This usually means putting each one in a separate
722 An exception can be made when two external symbols are always used
723 together, so that no reasonable program could use one without the
724 other; then they can both go in the same file.
726 External symbols that are not documented entry points for the user
727 should have names beginning with @samp{_}. The @samp{_} should be
728 followed by the chosen name prefix for the library, to prevent
729 collisions with other libraries. These can go in the same files with
730 user entry points if you like.
732 Static functions and variables can be used as you like and need not
733 fit any naming convention.
736 @section Formatting Error Messages
737 @cindex formatting error messages
738 @cindex error messages, formatting
740 Error messages from compilers should look like this:
743 @var{source-file-name}:@var{lineno}: @var{message}
747 If you want to mention the column number, use one of these formats:
750 @var{source-file-name}:@var{lineno}:@var{column}: @var{message}
751 @var{source-file-name}:@var{lineno}.@var{column}: @var{message}
756 Line numbers should start from 1 at the beginning of the file, and
757 column numbers should start from 1 at the beginning of the line. (Both
758 of these conventions are chosen for compatibility.) Calculate column
759 numbers assuming that space and all ASCII printing characters have
760 equal width, and assuming tab stops every 8 columns.
762 The error message can also give both the starting and ending positions
763 of the erroneous text. There are several formats so that you can
764 avoid redundant information such as a duplicate line number.
765 Here are the possible formats:
768 @var{source-file-name}:@var{lineno-1}.@var{column-1}-@var{lineno-2}.@var{column-2}: @var{message}
769 @var{source-file-name}:@var{lineno-1}.@var{column-1}-@var{column-2}: @var{message}
770 @var{source-file-name}:@var{lineno-1}-@var{lineno-2}: @var{message}
774 When an error is spread over several files, you can use this format:
777 @var{file-1}:@var{lineno-1}.@var{column-1}-@var{file-2}:@var{lineno-2}.@var{column-2}: @var{message}
780 Error messages from other noninteractive programs should look like this:
783 @var{program}:@var{source-file-name}:@var{lineno}: @var{message}
787 when there is an appropriate source file, or like this:
790 @var{program}: @var{message}
794 when there is no relevant source file.
796 If you want to mention the column number, use this format:
799 @var{program}:@var{source-file-name}:@var{lineno}:@var{column}: @var{message}
802 In an interactive program (one that is reading commands from a
803 terminal), it is better not to include the program name in an error
804 message. The place to indicate which program is running is in the
805 prompt or with the screen layout. (When the same program runs with
806 input from a source other than a terminal, it is not interactive and
807 would do best to print error messages using the noninteractive style.)
809 The string @var{message} should not begin with a capital letter when
810 it follows a program name and/or file name, because that isn't the
811 beginning of a sentence. (The sentence conceptually starts at the
812 beginning of the line.) Also, it should not end with a period.
814 Error messages from interactive programs, and other messages such as
815 usage messages, should start with a capital letter. But they should not
818 @node User Interfaces
819 @section Standards for Interfaces Generally
821 @cindex program name and its behavior
822 @cindex behavior, dependent on program's name
823 Please don't make the behavior of a utility depend on the name used
824 to invoke it. It is useful sometimes to make a link to a utility
825 with a different name, and that should not change what it does.
827 Instead, use a run time option or a compilation switch or both
828 to select among the alternate behaviors.
830 @cindex output device and program's behavior
831 Likewise, please don't make the behavior of the program depend on the
832 type of output device it is used with. Device independence is an
833 important principle of the system's design; do not compromise it merely
834 to save someone from typing an option now and then. (Variation in error
835 message syntax when using a terminal is ok, because that is a side issue
836 that people do not depend on.)
838 If you think one behavior is most useful when the output is to a
839 terminal, and another is most useful when the output is a file or a
840 pipe, then it is usually best to make the default behavior the one that
841 is useful with output to a terminal, and have an option for the other
844 Compatibility requires certain programs to depend on the type of output
845 device. It would be disastrous if @code{ls} or @code{sh} did not do so
846 in the way all users expect. In some of these cases, we supplement the
847 program with a preferred alternate version that does not depend on the
848 output device type. For example, we provide a @code{dir} program much
849 like @code{ls} except that its default output format is always
853 @node Graphical Interfaces
854 @section Standards for Graphical Interfaces
855 @cindex graphical user interface
858 When you write a program that provides a graphical user interface,
859 please make it work with the X Window System and the GTK+ toolkit
860 unless the functionality specifically requires some alternative (for
861 example, ``displaying jpeg images while in console mode'').
863 In addition, please provide a command-line interface to control the
864 functionality. (In many cases, the graphical user interface can be a
865 separate program which invokes the command-line program.) This is
866 so that the same jobs can be done from scripts.
870 Please also consider providing a CORBA interface (for use from GNOME), a
871 library interface (for use from C), and perhaps a keyboard-driven
872 console interface (for use by users from console mode). Once you are
873 doing the work to provide the functionality and the graphical interface,
874 these won't be much extra work.
877 @node Command-Line Interfaces
878 @section Standards for Command Line Interfaces
879 @cindex command-line interface
882 It is a good idea to follow the @sc{posix} guidelines for the
883 command-line options of a program. The easiest way to do this is to use
884 @code{getopt} to parse them. Note that the GNU version of @code{getopt}
885 will normally permit options anywhere among the arguments unless the
886 special argument @samp{--} is used. This is not what @sc{posix}
887 specifies; it is a GNU extension.
889 @cindex long-named options
890 Please define long-named options that are equivalent to the
891 single-letter Unix-style options. We hope to make GNU more user
892 friendly this way. This is easy to do with the GNU function
895 One of the advantages of long-named options is that they can be
896 consistent from program to program. For example, users should be able
897 to expect the ``verbose'' option of any GNU program which has one, to be
898 spelled precisely @samp{--verbose}. To achieve this uniformity, look at
899 the table of common long-option names when you choose the option names
900 for your program (@pxref{Option Table}).
902 It is usually a good idea for file names given as ordinary arguments to
903 be input files only; any output files would be specified using options
904 (preferably @samp{-o} or @samp{--output}). Even if you allow an output
905 file name as an ordinary argument for compatibility, try to provide an
906 option as another way to specify it. This will lead to more consistency
907 among GNU utilities, and fewer idiosyncrasies for users to remember.
909 @cindex standard command-line options
910 @cindex options, standard command-line
911 @cindex CGI programs, standard options for
912 @cindex PATH_INFO, specifying standard options as
913 All programs should support two standard options: @samp{--version}
914 and @samp{--help}. CGI programs should accept these as command-line
915 options, and also if given as the @env{PATH_INFO}; for instance,
916 visiting @url{http://example.org/p.cgi/--help} in a browser should
917 output the same information as invoking @samp{p.cgi --help} from the
921 * --version:: The standard output for --version.
922 * --help:: The standard output for --help.
926 @subsection @option{--version}
928 @cindex @samp{--version} output
930 The standard @code{--version} option should direct the program to
931 print information about its name, version, origin and legal status,
932 all on standard output, and then exit successfully. Other options and
933 arguments should be ignored once this is seen, and the program should
934 not perform its normal function.
936 @cindex canonical name of a program
937 @cindex program's canonical name
938 The first line is meant to be easy for a program to parse; the version
939 number proper starts after the last space. In addition, it contains
940 the canonical name for this program, in this format:
947 The program's name should be a constant string; @emph{don't} compute it
948 from @code{argv[0]}. The idea is to state the standard or canonical
949 name for the program, not its file name. There are other ways to find
950 out the precise file name where a command is found in @code{PATH}.
952 If the program is a subsidiary part of a larger package, mention the
953 package name in parentheses, like this:
956 emacsserver (GNU Emacs) 19.30
960 If the package has a version number which is different from this
961 program's version number, you can mention the package version number
962 just before the close-parenthesis.
964 If you @emph{need} to mention the version numbers of libraries which
965 are distributed separately from the package which contains this program,
966 you can do so by printing an additional line of version info for each
967 library you want to mention. Use the same format for these lines as for
970 Please do not mention all of the libraries that the program uses ``just
971 for completeness''---that would produce a lot of unhelpful clutter.
972 Please mention library version numbers only if you find in practice that
973 they are very important to you in debugging.
975 The following line, after the version number line or lines, should be a
976 copyright notice. If more than one copyright notice is called for, put
977 each on a separate line.
979 Next should follow a line stating the license, preferably using one of
980 abbrevations below, and a brief statement that the program is free
981 software, and that users are free to copy and change it. Also mention
982 that there is no warranty, to the extent permitted by law. See
983 recommended wording below.
985 It is ok to finish the output with a list of the major authors of the
986 program, as a way of giving credit.
988 Here's an example of output that follows these rules:
992 Copyright (C) 2007 Free Software Foundation, Inc.
993 License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
994 This is free software: you are free to change and redistribute it.
995 There is NO WARRANTY, to the extent permitted by law.
998 You should adapt this to your program, of course, filling in the proper
999 year, copyright holder, name of program, and the references to
1000 distribution terms, and changing the rest of the wording as necessary.
1002 This copyright notice only needs to mention the most recent year in
1003 which changes were made---there's no need to list the years for previous
1004 versions' changes. You don't have to mention the name of the program in
1005 these notices, if that is inconvenient, since it appeared in the first
1006 line. (The rules are different for copyright notices in source files;
1007 @pxref{Copyright Notices,,,maintain,Information for GNU Maintainers}.)
1009 Translations of the above lines must preserve the validity of the
1010 copyright notices (@pxref{Internationalization}). If the translation's
1011 character set supports it, the @samp{(C)} should be replaced with the
1012 copyright symbol, as follows:
1015 (the official copyright symbol, which is the letter C in a circle);
1021 Write the word ``Copyright'' exactly like that, in English. Do not
1022 translate it into another language. International treaties recognize
1023 the English word ``Copyright''; translations into other languages do not
1024 have legal significance.
1026 Finally, here is the table of our suggested license abbreviations.
1027 Any abbreviation can be followed by @samp{v@var{version}[+]}, meaning
1028 that particular version, or later versions with the @samp{+}, as shown
1031 In the case of exceptions for extra permissions with the GPL, we use
1032 @samp{/} for a separator; the version number can follow the license
1033 abbreviation as usual, as in the examples below.
1037 GNU General Public License, @url{http://www.gnu.org/@/licenses/@/gpl.html}.
1040 GNU Lesser General Public License, @url{http://www.gnu.org/@/licenses/@/lgpl.html}.
1043 GNU GPL with the exception for Guile; for example, GPLv3+/Guile means
1044 the GNU GPL version 3 or later, with the extra exception for Guile.
1047 GNU GPL with the exception for Ada.
1050 The Apache Software Foundation license,
1051 @url{http://www.apache.org/@/licenses}.
1054 The Artistic license used for Perl, @url{http://www.perlfoundation.org/@/legal}.
1057 The Expat license, @url{http://www.jclark.com/@/xml/@/copying.txt}.
1060 The Mozilla Public License, @url{http://www.mozilla.org/@/MPL/}.
1063 The original (4-clause) BSD license, incompatible with the GNU GPL
1064 @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#6}.
1067 The license used for PHP, @url{http://www.php.net/@/license/}.
1070 The non-license that is being in the public domain,
1071 @url{http://www.gnu.org/@/licenses/@/license-list.html#PublicDomain}.
1074 The license for Python, @url{http://www.python.org/@/2.0.1/@/license.html}.
1077 The revised (3-clause) BSD, compatible with the GNU GPL,
1078 @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#5}.
1081 The simple non-copyleft license used for most versions of the X Window
1082 System, @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#3}.
1085 The license for Zlib, @url{http://www.gzip.org/@/zlib/@/zlib_license.html}.
1089 More information about these licenses and many more are on the GNU
1090 licensing web pages,
1091 @url{http://www.gnu.org/@/licenses/@/license-list.html}.
1095 @subsection @option{--help}
1097 @cindex @samp{--help} output
1099 The standard @code{--help} option should output brief documentation
1100 for how to invoke the program, on standard output, then exit
1101 successfully. Other options and arguments should be ignored once this
1102 is seen, and the program should not perform its normal function.
1104 @cindex address for bug reports
1106 Near the end of the @samp{--help} option's output, please place lines
1107 giving the email address for bug reports, the package's home page
1108 (normally @indicateurl{http://www.gnu.org/software/@var{pkg}}, and the
1109 general page for help using GNU programs. The format should be like this:
1112 Report bugs to: @var{mailing-address}
1113 @var{pkg} home page: <http://www.gnu.org/software/@var{pkg}/>
1114 General help using GNU software: <http://www.gnu.org/gethelp/>
1117 It is ok to mention other appropriate mailing lists and web pages.
1121 @section Table of Long Options
1122 @cindex long option names
1123 @cindex table of long options
1125 Here is a table of long options used by GNU programs. It is surely
1126 incomplete, but we aim to list all the options that a new program might
1127 want to be compatible with. If you use names not already in the table,
1128 please send @email{bug-standards@@gnu.org} a list of them, with their
1129 meanings, so we can update the table.
1131 @c Please leave newlines between items in this table; it's much easier
1132 @c to update when it isn't completely squashed together and unreadable.
1133 @c When there is more than one short option for a long option name, put
1134 @c a semicolon between the lists of the programs that use them, not a
1135 @c period. --friedman
1139 @samp{-N} in @code{tar}.
1142 @samp{-a} in @code{du}, @code{ls}, @code{nm}, @code{stty}, @code{uname},
1143 and @code{unexpand}.
1146 @samp{-a} in @code{diff}.
1149 @samp{-A} in @code{ls}.
1152 @samp{-a} in @code{etags}, @code{tee}, @code{time};
1153 @samp{-r} in @code{tar}.
1156 @samp{-a} in @code{cp}.
1159 @samp{-n} in @code{shar}.
1162 @samp{-l} in @code{m4}.
1165 @samp{-a} in @code{diff}.
1168 @samp{-v} in @code{gawk}.
1171 @samp{-W} in @code{make}.
1174 @samp{-o} in @code{make}.
1177 @samp{-a} in @code{recode}.
1180 @samp{-a} in @code{wdiff}.
1182 @item auto-reference
1183 @samp{-A} in @code{ptx}.
1186 @samp{-n} in @code{wdiff}.
1189 For server programs, run in the background.
1191 @item backward-search
1192 @samp{-B} in @code{ctags}.
1195 @samp{-f} in @code{shar}.
1204 @samp{-b} in @code{tac}.
1207 @samp{-b} in @code{cpio} and @code{diff}.
1210 @samp{-b} in @code{shar}.
1213 Used in @code{cpio} and @code{tar}.
1216 @samp{-b} in @code{head} and @code{tail}.
1219 @samp{-b} in @code{ptx}.
1222 Used in various programs to make output shorter.
1225 @samp{-c} in @code{head}, @code{split}, and @code{tail}.
1228 @samp{-C} in @code{etags}.
1231 @samp{-A} in @code{tar}.
1234 Used in various programs to specify the directory to use.
1237 @samp{-c} in @code{chgrp} and @code{chown}.
1240 @samp{-F} in @code{ls}.
1243 @samp{-c} in @code{recode}.
1246 @samp{-c} in @code{su};
1250 @samp{-d} in @code{tar}.
1253 Used in @code{gawk}.
1256 @samp{-Z} in @code{tar} and @code{shar}.
1259 @samp{-A} in @code{tar}.
1262 @samp{-w} in @code{tar}.
1265 Used in @code{diff}.
1268 @samp{-W copyleft} in @code{gawk}.
1271 @samp{-C} in @code{ptx}, @code{recode}, and @code{wdiff};
1272 @samp{-W copyright} in @code{gawk}.
1278 @samp{-q} in @code{who}.
1281 @samp{-l} in @code{du}.
1284 Used in @code{tar} and @code{cpio}.
1287 @samp{-c} in @code{shar}.
1290 @samp{-x} in @code{ctags}.
1293 @samp{-d} in @code{touch}.
1296 @samp{-d} in @code{make} and @code{m4};
1300 @samp{-D} in @code{m4}.
1303 @samp{-d} in Bison and @code{ctags}.
1306 @samp{-D} in @code{tar}.
1309 @samp{-L} in @code{chgrp}, @code{chown}, @code{cpio}, @code{du},
1310 @code{ls}, and @code{tar}.
1312 @item dereference-args
1313 @samp{-D} in @code{du}.
1316 Specify an I/O device (special file name).
1319 @samp{-d} in @code{recode}.
1321 @item dictionary-order
1322 @samp{-d} in @code{look}.
1325 @samp{-d} in @code{tar}.
1328 @samp{-n} in @code{csplit}.
1331 Specify the directory to use, in various programs. In @code{ls}, it
1332 means to show directories themselves rather than their contents. In
1333 @code{rm} and @code{ln}, it means to not treat links to directories
1337 @samp{-x} in @code{strip}.
1339 @item discard-locals
1340 @samp{-X} in @code{strip}.
1343 @samp{-n} in @code{make}.
1346 @samp{-e} in @code{diff}.
1348 @item elide-empty-files
1349 @samp{-z} in @code{csplit}.
1352 @samp{-x} in @code{wdiff}.
1355 @samp{-z} in @code{wdiff}.
1357 @item entire-new-file
1358 @samp{-N} in @code{diff}.
1360 @item environment-overrides
1361 @samp{-e} in @code{make}.
1364 @samp{-e} in @code{xargs}.
1370 Used in @code{makeinfo}.
1373 @samp{-o} in @code{m4}.
1376 @samp{-b} in @code{ls}.
1379 @samp{-X} in @code{tar}.
1385 @samp{-x} in @code{xargs}.
1388 @samp{-e} in @code{unshar}.
1391 @samp{-t} in @code{diff}.
1394 @samp{-e} in @code{sed}.
1397 @samp{-g} in @code{nm}.
1400 @samp{-i} in @code{cpio};
1401 @samp{-x} in @code{tar}.
1404 @samp{-f} in @code{finger}.
1407 @samp{-f} in @code{su}.
1409 @item fatal-warnings
1410 @samp{-E} in @code{m4}.
1413 @samp{-f} in @code{gawk}, @code{info}, @code{make}, @code{mt},
1414 @code{sed}, and @code{tar}.
1416 @item field-separator
1417 @samp{-F} in @code{gawk}.
1423 @samp{-F} in @code{ls}.
1426 @samp{-T} in @code{tar}.
1429 Used in @code{makeinfo}.
1431 @item flag-truncation
1432 @samp{-F} in @code{ptx}.
1434 @item fixed-output-files
1438 @samp{-f} in @code{tail}.
1440 @item footnote-style
1441 Used in @code{makeinfo}.
1444 @samp{-f} in @code{cp}, @code{ln}, @code{mv}, and @code{rm}.
1447 @samp{-F} in @code{shar}.
1450 For server programs, run in the foreground;
1451 in other words, don't do anything special to run the server
1455 Used in @code{ls}, @code{time}, and @code{ptx}.
1458 @samp{-F} in @code{m4}.
1464 @samp{-g} in @code{ptx}.
1467 @samp{-x} in @code{tar}.
1470 @samp{-i} in @code{ul}.
1473 @samp{-g} in @code{recode}.
1476 @samp{-g} in @code{install}.
1479 @samp{-z} in @code{tar} and @code{shar}.
1482 @samp{-H} in @code{m4}.
1485 @samp{-h} in @code{objdump} and @code{recode}
1488 @samp{-H} in @code{who}.
1491 Used to ask for brief usage information.
1493 @item here-delimiter
1494 @samp{-d} in @code{shar}.
1496 @item hide-control-chars
1497 @samp{-q} in @code{ls}.
1500 In @code{makeinfo}, output HTML.
1503 @samp{-u} in @code{who}.
1506 @samp{-D} in @code{diff}.
1509 @samp{-I} in @code{ls};
1510 @samp{-x} in @code{recode}.
1512 @item ignore-all-space
1513 @samp{-w} in @code{diff}.
1515 @item ignore-backups
1516 @samp{-B} in @code{ls}.
1518 @item ignore-blank-lines
1519 @samp{-B} in @code{diff}.
1522 @samp{-f} in @code{look} and @code{ptx};
1523 @samp{-i} in @code{diff} and @code{wdiff}.
1526 @samp{-i} in @code{make}.
1529 @samp{-i} in @code{ptx}.
1531 @item ignore-indentation
1532 @samp{-I} in @code{etags}.
1534 @item ignore-init-file
1537 @item ignore-interrupts
1538 @samp{-i} in @code{tee}.
1540 @item ignore-matching-lines
1541 @samp{-I} in @code{diff}.
1543 @item ignore-space-change
1544 @samp{-b} in @code{diff}.
1547 @samp{-i} in @code{tar}.
1550 @samp{-i} in @code{etags};
1551 @samp{-I} in @code{m4}.
1554 @samp{-I} in @code{make}.
1557 @samp{-G} in @code{tar}.
1560 @samp{-i}, @samp{-l}, and @samp{-m} in Finger.
1563 In some programs, specify the name of the file to read as the user's
1567 @samp{-i} in @code{expand}.
1570 @samp{-T} in @code{diff}.
1573 @samp{-i} in @code{ls}.
1576 @samp{-i} in @code{cp}, @code{ln}, @code{mv}, @code{rm};
1577 @samp{-e} in @code{m4};
1578 @samp{-p} in @code{xargs};
1579 @samp{-w} in @code{tar}.
1582 @samp{-p} in @code{shar}.
1588 @samp{-j} in @code{make}.
1591 @samp{-n} in @code{make}.
1594 @samp{-k} in @code{make}.
1597 @samp{-k} in @code{csplit}.
1600 @samp{-k} in @code{du} and @code{ls}.
1603 @samp{-l} in @code{etags}.
1606 @samp{-l} in @code{wdiff}.
1608 @item level-for-gzip
1609 @samp{-g} in @code{shar}.
1612 @samp{-C} in @code{split}.
1615 Used in @code{split}, @code{head}, and @code{tail}.
1618 @samp{-l} in @code{cpio}.
1622 Used in @code{gawk}.
1625 @samp{-t} in @code{cpio};
1626 @samp{-l} in @code{recode}.
1629 @samp{-t} in @code{tar}.
1632 @samp{-N} in @code{ls}.
1635 @samp{-l} in @code{make}.
1641 Used in @code{uname}.
1644 @samp{-M} in @code{ptx}.
1647 @samp{-m} in @code{hello} and @code{uname}.
1649 @item make-directories
1650 @samp{-d} in @code{cpio}.
1653 @samp{-f} in @code{make}.
1659 @samp{-n} in @code{xargs}.
1662 @samp{-n} in @code{xargs}.
1665 @samp{-l} in @code{xargs}.
1668 @samp{-l} in @code{make}.
1671 @samp{-P} in @code{xargs}.
1674 @samp{-T} in @code{who}.
1677 @samp{-T} in @code{who}.
1680 @samp{-d} in @code{diff}.
1682 @item mixed-uuencode
1683 @samp{-M} in @code{shar}.
1686 @samp{-m} in @code{install}, @code{mkdir}, and @code{mkfifo}.
1688 @item modification-time
1689 @samp{-m} in @code{tar}.
1692 @samp{-M} in @code{tar}.
1698 @samp{-L} in @code{m4}.
1701 @samp{-a} in @code{shar}.
1704 @samp{-W} in @code{make}.
1706 @item no-builtin-rules
1707 @samp{-r} in @code{make}.
1709 @item no-character-count
1710 @samp{-w} in @code{shar}.
1712 @item no-check-existing
1713 @samp{-x} in @code{shar}.
1716 @samp{-3} in @code{wdiff}.
1719 @samp{-c} in @code{touch}.
1722 @samp{-D} in @code{etags}.
1725 @samp{-1} in @code{wdiff}.
1727 @item no-dereference
1728 @samp{-d} in @code{cp}.
1731 @samp{-2} in @code{wdiff}.
1734 @samp{-S} in @code{make}.
1740 @samp{-P} in @code{shar}.
1743 @samp{-e} in @code{gprof}.
1746 @samp{-R} in @code{etags}.
1749 @samp{-p} in @code{nm}.
1752 Don't print a startup splash screen.
1755 Used in @code{makeinfo}.
1758 @samp{-a} in @code{gprof}.
1761 @samp{-E} in @code{gprof}.
1764 @samp{-m} in @code{shar}.
1767 Used in @code{makeinfo}.
1770 Used in @code{emacsclient}.
1773 Used in various programs to inhibit warnings.
1776 @samp{-n} in @code{info}.
1779 @samp{-n} in @code{uname}.
1782 @samp{-f} in @code{cpio}.
1785 @samp{-n} in @code{objdump}.
1788 @samp{-0} in @code{xargs}.
1791 @samp{-n} in @code{cat}.
1793 @item number-nonblank
1794 @samp{-b} in @code{cat}.
1797 @samp{-n} in @code{nm}.
1799 @item numeric-uid-gid
1800 @samp{-n} in @code{cpio} and @code{ls}.
1806 @samp{-o} in @code{tar}.
1809 @samp{-o} in @code{make}.
1811 @item one-file-system
1812 @samp{-l} in @code{tar}, @code{cp}, and @code{du}.
1815 @samp{-o} in @code{ptx}.
1818 @samp{-f} in @code{gprof}.
1821 @samp{-F} in @code{gprof}.
1824 @samp{-o} in @code{getopt}, @code{fdlist}, @code{fdmount},
1825 @code{fdmountd}, and @code{fdumount}.
1828 In various programs, specify the output file name.
1831 @samp{-o} in @code{shar}.
1834 @samp{-o} in @code{rm}.
1837 @samp{-c} in @code{unshar}.
1840 @samp{-o} in @code{install}.
1843 @samp{-l} in @code{diff}.
1845 @item paragraph-indent
1846 Used in @code{makeinfo}.
1849 @samp{-p} in @code{mkdir} and @code{rmdir}.
1852 @samp{-p} in @code{ul}.
1855 @samp{-p} in @code{cpio}.
1858 @samp{-P} in @code{finger}.
1861 @samp{-c} in @code{cpio} and @code{tar}.
1864 Used in @code{gawk}.
1866 @item prefix-builtins
1867 @samp{-P} in @code{m4}.
1870 @samp{-f} in @code{csplit}.
1873 Used in @code{tar} and @code{cp}.
1875 @item preserve-environment
1876 @samp{-p} in @code{su}.
1878 @item preserve-modification-time
1879 @samp{-m} in @code{cpio}.
1881 @item preserve-order
1882 @samp{-s} in @code{tar}.
1884 @item preserve-permissions
1885 @samp{-p} in @code{tar}.
1888 @samp{-l} in @code{diff}.
1891 @samp{-L} in @code{cmp}.
1893 @item print-data-base
1894 @samp{-p} in @code{make}.
1896 @item print-directory
1897 @samp{-w} in @code{make}.
1899 @item print-file-name
1900 @samp{-o} in @code{nm}.
1903 @samp{-s} in @code{nm}.
1906 @samp{-p} in @code{wdiff}.
1909 @samp{-p} in @code{ed}.
1912 Specify an HTTP proxy.
1915 @samp{-X} in @code{shar}.
1918 @samp{-q} in @code{make}.
1921 Used in many programs to inhibit the usual output. Every
1922 program accepting @samp{--quiet} should accept @samp{--silent} as a
1926 @samp{-Q} in @code{shar}
1929 @samp{-Q} in @code{ls}.
1932 @samp{-n} in @code{diff}.
1935 Used in @code{gawk}.
1937 @item read-full-blocks
1938 @samp{-B} in @code{tar}.
1944 @samp{-n} in @code{make}.
1947 @samp{-R} in @code{tar}.
1950 Used in @code{chgrp}, @code{chown}, @code{cp}, @code{ls}, @code{diff},
1954 @samp{-r} in @code{touch}.
1957 @samp{-r} in @code{ptx}.
1960 @samp{-r} in @code{tac} and @code{etags}.
1963 @samp{-r} in @code{uname}.
1966 @samp{-R} in @code{m4}.
1969 @samp{-r} in @code{objdump}.
1972 @samp{-r} in @code{cpio}.
1975 @samp{-i} in @code{xargs}.
1977 @item report-identical-files
1978 @samp{-s} in @code{diff}.
1980 @item reset-access-time
1981 @samp{-a} in @code{cpio}.
1984 @samp{-r} in @code{ls} and @code{nm}.
1987 @samp{-f} in @code{diff}.
1989 @item right-side-defs
1990 @samp{-R} in @code{ptx}.
1993 @samp{-s} in @code{tar}.
1995 @item same-permissions
1996 @samp{-p} in @code{tar}.
1999 @samp{-g} in @code{stty}.
2004 @item sentence-regexp
2005 @samp{-S} in @code{ptx}.
2008 @samp{-S} in @code{du}.
2011 @samp{-s} in @code{tac}.
2014 Used by @code{recode} to chose files or pipes for sequencing passes.
2017 @samp{-s} in @code{su}.
2020 @samp{-A} in @code{cat}.
2022 @item show-c-function
2023 @samp{-p} in @code{diff}.
2026 @samp{-E} in @code{cat}.
2028 @item show-function-line
2029 @samp{-F} in @code{diff}.
2032 @samp{-T} in @code{cat}.
2035 Used in many programs to inhibit the usual output.
2036 Every program accepting
2037 @samp{--silent} should accept @samp{--quiet} as a synonym.
2040 @samp{-s} in @code{ls}.
2043 Specify a file descriptor for a network server to use for its socket,
2044 instead of opening and binding a new socket. This provides a way to
2045 run, in a non-privileged process, a server that normally needs a
2046 reserved port number.
2052 @samp{-W source} in @code{gawk}.
2055 @samp{-S} in @code{tar}.
2057 @item speed-large-files
2058 @samp{-H} in @code{diff}.
2061 @samp{-E} in @code{unshar}.
2063 @item split-size-limit
2064 @samp{-L} in @code{shar}.
2067 @samp{-s} in @code{cat}.
2070 @samp{-w} in @code{wdiff}.
2073 @samp{-y} in @code{wdiff}.
2076 Used in @code{tar} and @code{diff} to specify which file within
2077 a directory to start processing with.
2080 @samp{-s} in @code{wdiff}.
2082 @item stdin-file-list
2083 @samp{-S} in @code{shar}.
2086 @samp{-S} in @code{make}.
2089 @samp{-s} in @code{recode}.
2092 @samp{-s} in @code{install}.
2095 @samp{-s} in @code{strip}.
2098 @samp{-S} in @code{strip}.
2101 @samp{-s} in @code{shar}.
2104 @samp{-S} in @code{cp}, @code{ln}, @code{mv}.
2107 @samp{-b} in @code{csplit}.
2110 @samp{-s} in @code{gprof}.
2113 @samp{-s} in @code{du}.
2116 @samp{-s} in @code{ln}.
2119 Used in GDB and @code{objdump}.
2122 @samp{-s} in @code{m4}.
2125 @samp{-s} in @code{uname}.
2128 @samp{-t} in @code{expand} and @code{unexpand}.
2131 @samp{-T} in @code{ls}.
2134 @samp{-T} in @code{tput} and @code{ul}.
2135 @samp{-t} in @code{wdiff}.
2138 @samp{-a} in @code{diff}.
2141 @samp{-T} in @code{shar}.
2144 Used in @code{ls} and @code{touch}.
2147 Specify how long to wait before giving up on some operation.
2150 @samp{-O} in @code{tar}.
2153 @samp{-c} in @code{du}.
2156 @samp{-t} in @code{make}, @code{ranlib}, and @code{recode}.
2159 @samp{-t} in @code{m4}.
2162 @samp{-t} in @code{hello};
2163 @samp{-W traditional} in @code{gawk};
2164 @samp{-G} in @code{ed}, @code{m4}, and @code{ptx}.
2170 @samp{-t} in @code{ctags}.
2172 @item typedefs-and-c++
2173 @samp{-T} in @code{ctags}.
2176 @samp{-t} in @code{ptx}.
2179 @samp{-z} in @code{tar}.
2182 @samp{-u} in @code{cpio}.
2185 @samp{-U} in @code{m4}.
2187 @item undefined-only
2188 @samp{-u} in @code{nm}.
2191 @samp{-u} in @code{cp}, @code{ctags}, @code{mv}, @code{tar}.
2194 Used in @code{gawk}; same as @samp{--help}.
2197 @samp{-B} in @code{shar}.
2199 @item vanilla-operation
2200 @samp{-V} in @code{shar}.
2203 Print more information about progress. Many programs support this.
2206 @samp{-W} in @code{tar}.
2209 Print the version number.
2211 @item version-control
2212 @samp{-V} in @code{cp}, @code{ln}, @code{mv}.
2215 @samp{-v} in @code{ctags}.
2218 @samp{-V} in @code{tar}.
2221 @samp{-W} in @code{make}.
2223 @item whole-size-limit
2224 @samp{-l} in @code{shar}.
2227 @samp{-w} in @code{ls} and @code{ptx}.
2230 @samp{-W} in @code{ptx}.
2233 @samp{-T} in @code{who}.
2236 @samp{-z} in @code{gprof}.
2239 @node OID Allocations
2240 @section OID Allocations
2241 @cindex OID allocations for GNU
2246 The OID (object identifier) 1.3.6.1.4.1.11591 has been assigned to the
2247 GNU Project (thanks to Werner Koch). These are used for SNMP, LDAP,
2248 X.509 certificates, and so on. The web site
2249 @url{http://www.alvestrand.no/objectid} has a (voluntary) listing of
2250 many OID assignments.
2252 If you need a new slot for your GNU package, write
2253 @email{maintainers@@gnu.org}. Here is a list of arcs currently
2257 @include gnu-oids.texi
2262 @section Memory Usage
2263 @cindex memory usage
2265 If a program typically uses just a few meg of memory, don't bother making any
2266 effort to reduce memory usage. For example, if it is impractical for
2267 other reasons to operate on files more than a few meg long, it is
2268 reasonable to read entire input files into memory to operate on them.
2270 However, for programs such as @code{cat} or @code{tail}, that can
2271 usefully operate on very large files, it is important to avoid using a
2272 technique that would artificially limit the size of files it can handle.
2273 If a program works by lines and could be applied to arbitrary
2274 user-supplied input files, it should keep only a line in memory, because
2275 this is not very hard and users will want to be able to operate on input
2276 files that are bigger than will fit in memory all at once.
2278 If your program creates complicated data structures, just make them in
2279 memory and give a fatal error if @code{malloc} returns zero.
2285 Programs should be prepared to operate when @file{/usr} and @file{/etc}
2286 are read-only file systems. Thus, if the program manages log files,
2287 lock files, backup files, score files, or any other files which are
2288 modified for internal purposes, these files should not be stored in
2289 @file{/usr} or @file{/etc}.
2291 There are two exceptions. @file{/etc} is used to store system
2292 configuration information; it is reasonable for a program to modify
2293 files in @file{/etc} when its job is to update the system configuration.
2294 Also, if the user explicitly asks to modify one file in a directory, it
2295 is reasonable for the program to store other files in the same
2299 @chapter Making The Best Use of C
2301 This chapter provides advice on how best to use the C language
2302 when writing GNU software.
2305 * Formatting:: Formatting your source code.
2306 * Comments:: Commenting your work.
2307 * Syntactic Conventions:: Clean use of C constructs.
2308 * Names:: Naming variables, functions, and files.
2309 * System Portability:: Portability among different operating systems.
2310 * CPU Portability:: Supporting the range of CPU types.
2311 * System Functions:: Portability and ``standard'' library functions.
2312 * Internationalization:: Techniques for internationalization.
2313 * Character Set:: Use ASCII by default.
2314 * Quote Characters:: Use `...' in the C locale.
2315 * Mmap:: How you can safely use @code{mmap}.
2319 @section Formatting Your Source Code
2320 @cindex formatting source code
2323 @cindex braces, in C source
2324 It is important to put the open-brace that starts the body of a C
2325 function in column one, so that they will start a defun. Several
2326 tools look for open-braces in column one to find the beginnings of C
2327 functions. These tools will not work on code not formatted that way.
2329 Avoid putting open-brace, open-parenthesis or open-bracket in column
2330 one when they are inside a function, so that they won't start a defun.
2331 The open-brace that starts a @code{struct} body can go in column one
2332 if you find it useful to treat that definition as a defun.
2334 It is also important for function definitions to start the name of the
2335 function in column one. This helps people to search for function
2336 definitions, and may also help certain tools recognize them. Thus,
2337 using Standard C syntax, the format is this:
2341 concat (char *s1, char *s2)
2348 or, if you want to use traditional C syntax, format the definition like
2353 concat (s1, s2) /* Name starts in column one here */
2355 @{ /* Open brace in column one here */
2360 In Standard C, if the arguments don't fit nicely on one line,
2365 lots_of_args (int an_integer, long a_long, short a_short,
2366 double a_double, float a_float)
2370 The rest of this section gives our recommendations for other aspects of
2371 C formatting style, which is also the default style of the @code{indent}
2372 program in version 1.2 and newer. It corresponds to the options
2375 -nbad -bap -nbc -bbo -bl -bli2 -bls -ncdb -nce -cp1 -cs -di2
2376 -ndj -nfc1 -nfca -hnl -i2 -ip5 -lp -pcs -psl -nsc -nsob
2379 We don't think of these recommendations as requirements, because it
2380 causes no problems for users if two different programs have different
2383 But whatever style you use, please use it consistently, since a mixture
2384 of styles within one program tends to look ugly. If you are
2385 contributing changes to an existing program, please follow the style of
2388 For the body of the function, our recommended style looks like this:
2400 return ++x + bar ();
2404 @cindex spaces before open-paren
2405 We find it easier to read a program when it has spaces before the
2406 open-parentheses and after the commas. Especially after the commas.
2408 When you split an expression into multiple lines, split it
2409 before an operator, not after one. Here is the right way:
2411 @cindex expressions, splitting
2413 if (foo_this_is_long && bar > win (x, y, z)
2414 && remaining_condition)
2417 Try to avoid having two operators of different precedence at the same
2418 level of indentation. For example, don't write this:
2421 mode = (inmode[j] == VOIDmode
2422 || GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])
2423 ? outmode[j] : inmode[j]);
2426 Instead, use extra parentheses so that the indentation shows the nesting:
2429 mode = ((inmode[j] == VOIDmode
2430 || (GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])))
2431 ? outmode[j] : inmode[j]);
2434 Insert extra parentheses so that Emacs will indent the code properly.
2435 For example, the following indentation looks nice if you do it by hand,
2438 v = rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2439 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000;
2443 but Emacs would alter it. Adding a set of parentheses produces
2444 something that looks equally nice, and which Emacs will preserve:
2447 v = (rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2448 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000);
2451 Format do-while statements like this:
2463 Please use formfeed characters (control-L) to divide the program into
2464 pages at logical places (but not within a function). It does not matter
2465 just how long the pages are, since they do not have to fit on a printed
2466 page. The formfeeds should appear alone on lines by themselves.
2469 @section Commenting Your Work
2472 Every program should start with a comment saying briefly what it is for.
2473 Example: @samp{fmt - filter for simple filling of text}. This comment
2474 should be at the top of the source file containing the @samp{main}
2475 function of the program.
2477 Also, please write a brief comment at the start of each source file,
2478 with the file name and a line or two about the overall purpose of the
2481 Please write the comments in a GNU program in English, because English
2482 is the one language that nearly all programmers in all countries can
2483 read. If you do not write English well, please write comments in
2484 English as well as you can, then ask other people to help rewrite them.
2485 If you can't write comments in English, please find someone to work with
2486 you and translate your comments into English.
2488 Please put a comment on each function saying what the function does,
2489 what sorts of arguments it gets, and what the possible values of
2490 arguments mean and are used for. It is not necessary to duplicate in
2491 words the meaning of the C argument declarations, if a C type is being
2492 used in its customary fashion. If there is anything nonstandard about
2493 its use (such as an argument of type @code{char *} which is really the
2494 address of the second character of a string, not the first), or any
2495 possible values that would not work the way one would expect (such as,
2496 that strings containing newlines are not guaranteed to work), be sure
2499 Also explain the significance of the return value, if there is one.
2501 Please put two spaces after the end of a sentence in your comments, so
2502 that the Emacs sentence commands will work. Also, please write
2503 complete sentences and capitalize the first word. If a lower-case
2504 identifier comes at the beginning of a sentence, don't capitalize it!
2505 Changing the spelling makes it a different identifier. If you don't
2506 like starting a sentence with a lower case letter, write the sentence
2507 differently (e.g., ``The identifier lower-case is @dots{}'').
2509 The comment on a function is much clearer if you use the argument
2510 names to speak about the argument values. The variable name itself
2511 should be lower case, but write it in upper case when you are speaking
2512 about the value rather than the variable itself. Thus, ``the inode
2513 number NODE_NUM'' rather than ``an inode''.
2515 There is usually no purpose in restating the name of the function in
2516 the comment before it, because the reader can see that for himself.
2517 There might be an exception when the comment is so long that the function
2518 itself would be off the bottom of the screen.
2520 There should be a comment on each static variable as well, like this:
2523 /* Nonzero means truncate lines in the display;
2524 zero means continue them. */
2528 @cindex conditionals, comments for
2529 @cindex @code{#endif}, commenting
2530 Every @samp{#endif} should have a comment, except in the case of short
2531 conditionals (just a few lines) that are not nested. The comment should
2532 state the condition of the conditional that is ending, @emph{including
2533 its sense}. @samp{#else} should have a comment describing the condition
2534 @emph{and sense} of the code that follows. For example:
2542 #endif /* not foo */
2552 but, by contrast, write the comments this way for a @samp{#ifndef}:
2565 #endif /* not foo */
2569 @node Syntactic Conventions
2570 @section Clean Use of C Constructs
2571 @cindex syntactic conventions
2573 @cindex implicit @code{int}
2574 @cindex function argument, declaring
2575 Please explicitly declare the types of all objects. For example, you
2576 should explicitly declare all arguments to functions, and you should
2577 declare functions to return @code{int} rather than omitting the
2580 @cindex compiler warnings
2581 @cindex @samp{-Wall} compiler option
2582 Some programmers like to use the GCC @samp{-Wall} option, and change the
2583 code whenever it issues a warning. If you want to do this, then do.
2584 Other programmers prefer not to use @samp{-Wall}, because it gives
2585 warnings for valid and legitimate code which they do not want to change.
2586 If you want to do this, then do. The compiler should be your servant,
2589 Declarations of external functions and functions to appear later in the
2590 source file should all go in one place near the beginning of the file
2591 (somewhere before the first function definition in the file), or else
2592 should go in a header file. Don't put @code{extern} declarations inside
2595 @cindex temporary variables
2596 It used to be common practice to use the same local variables (with
2597 names like @code{tem}) over and over for different values within one
2598 function. Instead of doing this, it is better to declare a separate local
2599 variable for each distinct purpose, and give it a name which is
2600 meaningful. This not only makes programs easier to understand, it also
2601 facilitates optimization by good compilers. You can also move the
2602 declaration of each local variable into the smallest scope that includes
2603 all its uses. This makes the program even cleaner.
2605 Don't use local variables or parameters that shadow global identifiers.
2607 @cindex multiple variables in a line
2608 Don't declare multiple variables in one declaration that spans lines.
2609 Start a new declaration on each line, instead. For example, instead
2635 (If they are global variables, each should have a comment preceding it
2638 When you have an @code{if}-@code{else} statement nested in another
2639 @code{if} statement, always put braces around the @code{if}-@code{else}.
2640 Thus, never write like this:
2663 If you have an @code{if} statement nested inside of an @code{else}
2664 statement, either write @code{else if} on one line, like this,
2674 with its @code{then}-part indented like the preceding @code{then}-part,
2675 or write the nested @code{if} within braces like this:
2687 Don't declare both a structure tag and variables or typedefs in the
2688 same declaration. Instead, declare the structure tag separately
2689 and then use it to declare the variables or typedefs.
2691 Try to avoid assignments inside @code{if}-conditions (assignments
2692 inside @code{while}-conditions are ok). For example, don't write
2696 if ((foo = (char *) malloc (sizeof *foo)) == 0)
2697 fatal ("virtual memory exhausted");
2701 instead, write this:
2704 foo = (char *) malloc (sizeof *foo);
2706 fatal ("virtual memory exhausted");
2710 Don't make the program ugly to placate @code{lint}. Please don't insert any
2711 casts to @code{void}. Zero without a cast is perfectly fine as a null
2712 pointer constant, except when calling a varargs function.
2715 @section Naming Variables, Functions, and Files
2717 @cindex names of variables, functions, and files
2718 The names of global variables and functions in a program serve as
2719 comments of a sort. So don't choose terse names---instead, look for
2720 names that give useful information about the meaning of the variable or
2721 function. In a GNU program, names should be English, like other
2724 Local variable names can be shorter, because they are used only within
2725 one context, where (presumably) comments explain their purpose.
2727 Try to limit your use of abbreviations in symbol names. It is ok to
2728 make a few abbreviations, explain what they mean, and then use them
2729 frequently, but don't use lots of obscure abbreviations.
2731 Please use underscores to separate words in a name, so that the Emacs
2732 word commands can be useful within them. Stick to lower case; reserve
2733 upper case for macros and @code{enum} constants, and for name-prefixes
2734 that follow a uniform convention.
2736 For example, you should use names like @code{ignore_space_change_flag};
2737 don't use names like @code{iCantReadThis}.
2739 Variables that indicate whether command-line options have been
2740 specified should be named after the meaning of the option, not after
2741 the option-letter. A comment should state both the exact meaning of
2742 the option and its letter. For example,
2746 /* Ignore changes in horizontal whitespace (-b). */
2747 int ignore_space_change_flag;
2751 When you want to define names with constant integer values, use
2752 @code{enum} rather than @samp{#define}. GDB knows about enumeration
2755 @cindex file-name limitations
2757 You might want to make sure that none of the file names would conflict
2758 if the files were loaded onto an MS-DOS file system which shortens the
2759 names. You can use the program @code{doschk} to test for this.
2761 Some GNU programs were designed to limit themselves to file names of 14
2762 characters or less, to avoid file name conflicts if they are read into
2763 older System V systems. Please preserve this feature in the existing
2764 GNU programs that have it, but there is no need to do this in new GNU
2765 programs. @code{doschk} also reports file names longer than 14
2768 @node System Portability
2769 @section Portability between System Types
2770 @cindex portability, between system types
2772 In the Unix world, ``portability'' refers to porting to different Unix
2773 versions. For a GNU program, this kind of portability is desirable, but
2776 The primary purpose of GNU software is to run on top of the GNU kernel,
2777 compiled with the GNU C compiler, on various types of @sc{cpu}. So the
2778 kinds of portability that are absolutely necessary are quite limited.
2779 But it is important to support Linux-based GNU systems, since they
2780 are the form of GNU that is popular.
2782 Beyond that, it is good to support the other free operating systems
2783 (*BSD), and it is nice to support other Unix-like systems if you want
2784 to. Supporting a variety of Unix-like systems is desirable, although
2785 not paramount. It is usually not too hard, so you may as well do it.
2786 But you don't have to consider it an obligation, if it does turn out to
2790 The easiest way to achieve portability to most Unix-like systems is to
2791 use Autoconf. It's unlikely that your program needs to know more
2792 information about the host platform than Autoconf can provide, simply
2793 because most of the programs that need such knowledge have already been
2796 Avoid using the format of semi-internal data bases (e.g., directories)
2797 when there is a higher-level alternative (@code{readdir}).
2799 @cindex non-@sc{posix} systems, and portability
2800 As for systems that are not like Unix, such as MSDOS, Windows, VMS, MVS,
2801 and older Macintosh systems, supporting them is often a lot of work.
2802 When that is the case, it is better to spend your time adding features
2803 that will be useful on GNU and GNU/Linux, rather than on supporting
2804 other incompatible systems.
2806 If you do support Windows, please do not abbreviate it as ``win''. In
2807 hacker terminology, calling something a ``win'' is a form of praise.
2808 You're free to praise Microsoft Windows on your own if you want, but
2809 please don't do this in GNU packages. Instead of abbreviating
2810 ``Windows'' to ``win'', you can write it in full or abbreviate it to
2811 ``woe'' or ``w''. In GNU Emacs, for instance, we use @samp{w32} in
2812 file names of Windows-specific files, but the macro for Windows
2813 conditionals is called @code{WINDOWSNT}.
2815 It is a good idea to define the ``feature test macro''
2816 @code{_GNU_SOURCE} when compiling your C files. When you compile on GNU
2817 or GNU/Linux, this will enable the declarations of GNU library extension
2818 functions, and that will usually give you a compiler error message if
2819 you define the same function names in some other way in your program.
2820 (You don't have to actually @emph{use} these functions, if you prefer
2821 to make the program more portable to other systems.)
2823 But whether or not you use these GNU extensions, you should avoid
2824 using their names for any other meanings. Doing so would make it hard
2825 to move your code into other GNU programs.
2827 @node CPU Portability
2828 @section Portability between @sc{cpu}s
2830 @cindex data types, and portability
2831 @cindex portability, and data types
2832 Even GNU systems will differ because of differences among @sc{cpu}
2833 types---for example, difference in byte ordering and alignment
2834 requirements. It is absolutely essential to handle these differences.
2835 However, don't make any effort to cater to the possibility that an
2836 @code{int} will be less than 32 bits. We don't support 16-bit machines
2839 Similarly, don't make any effort to cater to the possibility that
2840 @code{long} will be smaller than predefined types like @code{size_t}.
2841 For example, the following code is ok:
2844 printf ("size = %lu\n", (unsigned long) sizeof array);
2845 printf ("diff = %ld\n", (long) (pointer2 - pointer1));
2848 1989 Standard C requires this to work, and we know of only one
2849 counterexample: 64-bit programs on Microsoft Windows. We will
2850 leave it to those who want to port GNU programs to that environment
2851 to figure out how to do it.
2853 Predefined file-size types like @code{off_t} are an exception: they are
2854 longer than @code{long} on many platforms, so code like the above won't
2855 work with them. One way to print an @code{off_t} value portably is to
2856 print its digits yourself, one by one.
2858 Don't assume that the address of an @code{int} object is also the
2859 address of its least-significant byte. This is false on big-endian
2860 machines. Thus, don't make the following mistake:
2865 while ((c = getchar ()) != EOF)
2866 write (file_descriptor, &c, 1);
2869 @noindent Instead, use @code{unsigned char} as follows. (The @code{unsigned}
2870 is for portability to unusual systems where @code{char} is signed and
2871 where there is integer overflow checking.)
2875 while ((c = getchar ()) != EOF)
2877 unsigned char u = c;
2878 write (file_descriptor, &u, 1);
2882 It used to be ok to not worry about the difference between pointers
2883 and integers when passing arguments to functions. However, on most
2884 modern 64-bit machines pointers are wider than @code{int}.
2885 Conversely, integer types like @code{long long int} and @code{off_t}
2886 are wider than pointers on most modern 32-bit machines. Hence it's
2887 often better nowadays to use prototypes to define functions whose
2888 argument types are not trivial.
2890 In particular, if functions accept varying argument counts or types
2891 they should be declared using prototypes containing @samp{...} and
2892 defined using @file{stdarg.h}. For an example of this, please see the
2893 @uref{http://www.gnu.org/software/gnulib/, Gnulib} error module, which
2894 declares and defines the following function:
2897 /* Print a message with `fprintf (stderr, FORMAT, ...)';
2898 if ERRNUM is nonzero, follow it with ": " and strerror (ERRNUM).
2899 If STATUS is nonzero, terminate the program with `exit (STATUS)'. */
2901 void error (int status, int errnum, const char *format, ...);
2904 A simple way to use the Gnulib error module is to obtain the two
2905 source files @file{error.c} and @file{error.h} from the Gnulib library
2906 source code repository at
2907 @uref{http://git.savannah.gnu.org/@/gitweb/@/?p=gnulib.git}.
2908 Here's a sample use:
2915 char *program_name = "myprogram";
2918 xfopen (char const *name)
2920 FILE *fp = fopen (name, "r");
2922 error (1, errno, "cannot read %s", name);
2927 @cindex casting pointers to integers
2928 Avoid casting pointers to integers if you can. Such casts greatly
2929 reduce portability, and in most programs they are easy to avoid. In the
2930 cases where casting pointers to integers is essential---such as, a Lisp
2931 interpreter which stores type information as well as an address in one
2932 word---you'll have to make explicit provisions to handle different word
2933 sizes. You will also need to make provision for systems in which the
2934 normal range of addresses you can get from @code{malloc} starts far away
2937 @node System Functions
2938 @section Calling System Functions
2939 @cindex library functions, and portability
2940 @cindex portability, and library functions
2942 C implementations differ substantially. Standard C reduces but does
2943 not eliminate the incompatibilities; meanwhile, many GNU packages still
2944 support pre-standard compilers because this is not hard to do. This
2945 chapter gives recommendations for how to use the more-or-less standard C
2946 library functions to avoid unnecessary loss of portability.
2950 Don't use the return value of @code{sprintf}. It returns the number of
2951 characters written on some systems, but not on all systems.
2954 Be aware that @code{vfprintf} is not always available.
2957 @code{main} should be declared to return type @code{int}. It should
2958 terminate either by calling @code{exit} or by returning the integer
2959 status code; make sure it cannot ever return an undefined value.
2961 @cindex declaration for system functions
2963 Don't declare system functions explicitly.
2965 Almost any declaration for a system function is wrong on some system.
2966 To minimize conflicts, leave it to the system header files to declare
2967 system functions. If the headers don't declare a function, let it
2970 While it may seem unclean to use a function without declaring it, in
2971 practice this works fine for most system library functions on the
2972 systems where this really happens; thus, the disadvantage is only
2973 theoretical. By contrast, actual declarations have frequently caused
2977 If you must declare a system function, don't specify the argument types.
2978 Use an old-style declaration, not a Standard C prototype. The more you
2979 specify about the function, the more likely a conflict.
2982 In particular, don't unconditionally declare @code{malloc} or
2985 Most GNU programs use those functions just once, in functions
2986 conventionally named @code{xmalloc} and @code{xrealloc}. These
2987 functions call @code{malloc} and @code{realloc}, respectively, and
2990 Because @code{xmalloc} and @code{xrealloc} are defined in your program,
2991 you can declare them in other files without any risk of type conflict.
2993 On most systems, @code{int} is the same length as a pointer; thus, the
2994 calls to @code{malloc} and @code{realloc} work fine. For the few
2995 exceptional systems (mostly 64-bit machines), you can use
2996 @strong{conditionalized} declarations of @code{malloc} and
2997 @code{realloc}---or put these declarations in configuration files
2998 specific to those systems.
3000 @cindex string library functions
3002 The string functions require special treatment. Some Unix systems have
3003 a header file @file{string.h}; others have @file{strings.h}. Neither
3004 file name is portable. There are two things you can do: use Autoconf to
3005 figure out which file to include, or don't include either file.
3008 If you don't include either strings file, you can't get declarations for
3009 the string functions from the header file in the usual way.
3011 That causes less of a problem than you might think. The newer standard
3012 string functions should be avoided anyway because many systems still
3013 don't support them. The string functions you can use are these:
3016 strcpy strncpy strcat strncat
3017 strlen strcmp strncmp
3021 The copy and concatenate functions work fine without a declaration as
3022 long as you don't use their values. Using their values without a
3023 declaration fails on systems where the width of a pointer differs from
3024 the width of @code{int}, and perhaps in other cases. It is trivial to
3025 avoid using their values, so do that.
3027 The compare functions and @code{strlen} work fine without a declaration
3028 on most systems, possibly all the ones that GNU software runs on.
3029 You may find it necessary to declare them @strong{conditionally} on a
3032 The search functions must be declared to return @code{char *}. Luckily,
3033 there is no variation in the data type they return. But there is
3034 variation in their names. Some systems give these functions the names
3035 @code{index} and @code{rindex}; other systems use the names
3036 @code{strchr} and @code{strrchr}. Some systems support both pairs of
3037 names, but neither pair works on all systems.
3039 You should pick a single pair of names and use it throughout your
3040 program. (Nowadays, it is better to choose @code{strchr} and
3041 @code{strrchr} for new programs, since those are the standard
3042 names.) Declare both of those names as functions returning @code{char
3043 *}. On systems which don't support those names, define them as macros
3044 in terms of the other pair. For example, here is what to put at the
3045 beginning of your file (or in a header) if you want to use the names
3046 @code{strchr} and @code{strrchr} throughout:
3050 #define strchr index
3052 #ifndef HAVE_STRRCHR
3053 #define strrchr rindex
3061 Here we assume that @code{HAVE_STRCHR} and @code{HAVE_STRRCHR} are
3062 macros defined in systems where the corresponding functions exist.
3063 One way to get them properly defined is to use Autoconf.
3065 @node Internationalization
3066 @section Internationalization
3067 @cindex internationalization
3070 GNU has a library called GNU gettext that makes it easy to translate the
3071 messages in a program into various languages. You should use this
3072 library in every program. Use English for the messages as they appear
3073 in the program, and let gettext provide the way to translate them into
3076 Using GNU gettext involves putting a call to the @code{gettext} macro
3077 around each string that might need translation---like this:
3080 printf (gettext ("Processing file `%s'..."));
3084 This permits GNU gettext to replace the string @code{"Processing file
3085 `%s'..."} with a translated version.
3087 Once a program uses gettext, please make a point of writing calls to
3088 @code{gettext} when you add new strings that call for translation.
3090 Using GNU gettext in a package involves specifying a @dfn{text domain
3091 name} for the package. The text domain name is used to separate the
3092 translations for this package from the translations for other packages.
3093 Normally, the text domain name should be the same as the name of the
3094 package---for example, @samp{coreutils} for the GNU core utilities.
3096 @cindex message text, and internationalization
3097 To enable gettext to work well, avoid writing code that makes
3098 assumptions about the structure of words or sentences. When you want
3099 the precise text of a sentence to vary depending on the data, use two or
3100 more alternative string constants each containing a complete sentences,
3101 rather than inserting conditionalized words or phrases into a single
3104 Here is an example of what not to do:
3107 printf ("%s is full", capacity > 5000000 ? "disk" : "floppy disk");
3110 If you apply gettext to all strings, like this,
3113 printf (gettext ("%s is full"),
3114 capacity > 5000000 ? gettext ("disk") : gettext ("floppy disk"));
3118 the translator will hardly know that "disk" and "floppy disk" are meant to
3119 be substituted in the other string. Worse, in some languages (like French)
3120 the construction will not work: the translation of the word "full" depends
3121 on the gender of the first part of the sentence; it happens to be not the
3122 same for "disk" as for "floppy disk".
3124 Complete sentences can be translated without problems:
3127 printf (capacity > 5000000 ? gettext ("disk is full")
3128 : gettext ("floppy disk is full"));
3131 A similar problem appears at the level of sentence structure with this
3135 printf ("# Implicit rule search has%s been done.\n",
3136 f->tried_implicit ? "" : " not");
3140 Adding @code{gettext} calls to this code cannot give correct results for
3141 all languages, because negation in some languages requires adding words
3142 at more than one place in the sentence. By contrast, adding
3143 @code{gettext} calls does the job straightforwardly if the code starts
3147 printf (f->tried_implicit
3148 ? "# Implicit rule search has been done.\n",
3149 : "# Implicit rule search has not been done.\n");
3152 Another example is this one:
3155 printf ("%d file%s processed", nfiles,
3156 nfiles != 1 ? "s" : "");
3160 The problem with this example is that it assumes that plurals are made
3161 by adding `s'. If you apply gettext to the format string, like this,
3164 printf (gettext ("%d file%s processed"), nfiles,
3165 nfiles != 1 ? "s" : "");
3169 the message can use different words, but it will still be forced to use
3170 `s' for the plural. Here is a better way, with gettext being applied to
3171 the two strings independently:
3174 printf ((nfiles != 1 ? gettext ("%d files processed")
3175 : gettext ("%d file processed")),
3180 But this still doesn't work for languages like Polish, which has three
3181 plural forms: one for nfiles == 1, one for nfiles == 2, 3, 4, 22, 23, 24, ...
3182 and one for the rest. The GNU @code{ngettext} function solves this problem:
3185 printf (ngettext ("%d files processed", "%d file processed", nfiles),
3191 @section Character Set
3192 @cindex character set
3194 @cindex ASCII characters
3195 @cindex non-ASCII characters
3197 Sticking to the ASCII character set (plain text, 7-bit characters) is
3198 preferred in GNU source code comments, text documents, and other
3199 contexts, unless there is good reason to do something else because of
3200 the application domain. For example, if source code deals with the
3201 French Revolutionary calendar, it is OK if its literal strings contain
3202 accented characters in month names like ``Flor@'eal''. Also, it is OK
3203 to use non-ASCII characters to represent proper names of contributors in
3204 change logs (@pxref{Change Logs}).
3206 If you need to use non-ASCII characters, you should normally stick with
3207 one encoding, as one cannot in general mix encodings reliably.
3210 @node Quote Characters
3211 @section Quote Characters
3212 @cindex quote characters
3213 @cindex locale-specific quote characters
3215 @cindex grave accent
3217 In the C locale, GNU programs should stick to plain ASCII for quotation
3218 characters in messages to users: preferably 0x60 (@samp{`}) for left
3219 quotes and 0x27 (@samp{'}) for right quotes. It is ok, but not
3220 required, to use locale-specific quotes in other locales.
3222 The @uref{http://www.gnu.org/software/gnulib/, Gnulib} @code{quote} and
3223 @code{quotearg} modules provide a reasonably straightforward way to
3224 support locale-specific quote characters, as well as taking care of
3225 other issues, such as quoting a filename that itself contains a quote
3226 character. See the Gnulib documentation for usage details.
3228 In any case, the documentation for your program should clearly specify
3229 how it does quoting, if different than the preferred method of @samp{`}
3230 and @samp{'}. This is especially important if the output of your
3231 program is ever likely to be parsed by another program.
3233 Quotation characters are a difficult area in the computing world at
3234 this time: there are no true left or right quote characters in Latin1;
3235 the @samp{`} character we use was standardized there as a grave
3236 accent. Moreover, Latin1 is still not universally usable.
3238 Unicode contains the unambiguous quote characters required, and its
3239 common encoding UTF-8 is upward compatible with Latin1. However,
3240 Unicode and UTF-8 are not universally well-supported, either.
3242 This may change over the next few years, and then we will revisit
3250 Don't assume that @code{mmap} either works on all files or fails
3251 for all files. It may work on some files and fail on others.
3253 The proper way to use @code{mmap} is to try it on the specific file for
3254 which you want to use it---and if @code{mmap} doesn't work, fall back on
3255 doing the job in another way using @code{read} and @code{write}.
3257 The reason this precaution is needed is that the GNU kernel (the HURD)
3258 provides a user-extensible file system, in which there can be many
3259 different kinds of ``ordinary files.'' Many of them support
3260 @code{mmap}, but some do not. It is important to make programs handle
3261 all these kinds of files.
3264 @chapter Documenting Programs
3265 @cindex documentation
3267 A GNU program should ideally come with full free documentation, adequate
3268 for both reference and tutorial purposes. If the package can be
3269 programmed or extended, the documentation should cover programming or
3270 extending it, as well as just using it.
3273 * GNU Manuals:: Writing proper manuals.
3274 * Doc Strings and Manuals:: Compiling doc strings doesn't make a manual.
3275 * Manual Structure Details:: Specific structure conventions.
3276 * License for Manuals:: Writing the distribution terms for a manual.
3277 * Manual Credits:: Giving credit to documentation contributors.
3278 * Printed Manuals:: Mentioning the printed manual.
3279 * NEWS File:: NEWS files supplement manuals.
3280 * Change Logs:: Recording changes.
3281 * Man Pages:: Man pages are secondary.
3282 * Reading other Manuals:: How far you can go in learning
3287 @section GNU Manuals
3289 The preferred document format for the GNU system is the Texinfo
3290 formatting language. Every GNU package should (ideally) have
3291 documentation in Texinfo both for reference and for learners. Texinfo
3292 makes it possible to produce a good quality formatted book, using
3293 @TeX{}, and to generate an Info file. It is also possible to generate
3294 HTML output from Texinfo source. See the Texinfo manual, either the
3295 hardcopy, or the on-line version available through @code{info} or the
3296 Emacs Info subsystem (@kbd{C-h i}).
3298 Nowadays some other formats such as Docbook and Sgmltexi can be
3299 converted automatically into Texinfo. It is ok to produce the Texinfo
3300 documentation by conversion this way, as long as it gives good results.
3302 Make sure your manual is clear to a reader who knows nothing about the
3303 topic and reads it straight through. This means covering basic topics
3304 at the beginning, and advanced topics only later. This also means
3305 defining every specialized term when it is first used.
3307 Programmers tend to carry over the structure of the program as the
3308 structure for its documentation. But this structure is not
3309 necessarily good for explaining how to use the program; it may be
3310 irrelevant and confusing for a user.
3312 Instead, the right way to structure documentation is according to the
3313 concepts and questions that a user will have in mind when reading it.
3314 This principle applies at every level, from the lowest (ordering
3315 sentences in a paragraph) to the highest (ordering of chapter topics
3316 within the manual). Sometimes this structure of ideas matches the
3317 structure of the implementation of the software being documented---but
3318 often they are different. An important part of learning to write good
3319 documentation is to learn to notice when you have unthinkingly
3320 structured the documentation like the implementation, stop yourself,
3321 and look for better alternatives.
3323 For example, each program in the GNU system probably ought to be
3324 documented in one manual; but this does not mean each program should
3325 have its own manual. That would be following the structure of the
3326 implementation, rather than the structure that helps the user
3329 Instead, each manual should cover a coherent @emph{topic}. For example,
3330 instead of a manual for @code{diff} and a manual for @code{diff3}, we
3331 have one manual for ``comparison of files'' which covers both of those
3332 programs, as well as @code{cmp}. By documenting these programs
3333 together, we can make the whole subject clearer.
3335 The manual which discusses a program should certainly document all of
3336 the program's command-line options and all of its commands. It should
3337 give examples of their use. But don't organize the manual as a list
3338 of features. Instead, organize it logically, by subtopics. Address
3339 the questions that a user will ask when thinking about the job that
3340 the program does. Don't just tell the reader what each feature can
3341 do---say what jobs it is good for, and show how to use it for those
3342 jobs. Explain what is recommended usage, and what kinds of usage
3345 In general, a GNU manual should serve both as tutorial and reference.
3346 It should be set up for convenient access to each topic through Info,
3347 and for reading straight through (appendixes aside). A GNU manual
3348 should give a good introduction to a beginner reading through from the
3349 start, and should also provide all the details that hackers want.
3350 The Bison manual is a good example of this---please take a look at it
3351 to see what we mean.
3353 That is not as hard as it first sounds. Arrange each chapter as a
3354 logical breakdown of its topic, but order the sections, and write their
3355 text, so that reading the chapter straight through makes sense. Do
3356 likewise when structuring the book into chapters, and when structuring a
3357 section into paragraphs. The watchword is, @emph{at each point, address
3358 the most fundamental and important issue raised by the preceding text.}
3360 If necessary, add extra chapters at the beginning of the manual which
3361 are purely tutorial and cover the basics of the subject. These provide
3362 the framework for a beginner to understand the rest of the manual. The
3363 Bison manual provides a good example of how to do this.
3365 To serve as a reference, a manual should have an Index that list all the
3366 functions, variables, options, and important concepts that are part of
3367 the program. One combined Index should do for a short manual, but
3368 sometimes for a complex package it is better to use multiple indices.
3369 The Texinfo manual includes advice on preparing good index entries, see
3370 @ref{Index Entries, , Making Index Entries, texinfo, GNU Texinfo}, and
3371 see @ref{Indexing Commands, , Defining the Entries of an
3372 Index, texinfo, GNU Texinfo}.
3374 Don't use Unix man pages as a model for how to write GNU documentation;
3375 most of them are terse, badly structured, and give inadequate
3376 explanation of the underlying concepts. (There are, of course, some
3377 exceptions.) Also, Unix man pages use a particular format which is
3378 different from what we use in GNU manuals.
3380 Please include an email address in the manual for where to report
3381 bugs @emph{in the text of the manual}.
3383 Please do not use the term ``pathname'' that is used in Unix
3384 documentation; use ``file name'' (two words) instead. We use the term
3385 ``path'' only for search paths, which are lists of directory names.
3387 Please do not use the term ``illegal'' to refer to erroneous input to
3388 a computer program. Please use ``invalid'' for this, and reserve the
3389 term ``illegal'' for activities prohibited by law.
3391 Please do not write @samp{()} after a function name just to indicate
3392 it is a function. @code{foo ()} is not a function, it is a function
3393 call with no arguments.
3395 @node Doc Strings and Manuals
3396 @section Doc Strings and Manuals
3398 Some programming systems, such as Emacs, provide a documentation string
3399 for each function, command or variable. You may be tempted to write a
3400 reference manual by compiling the documentation strings and writing a
3401 little additional text to go around them---but you must not do it. That
3402 approach is a fundamental mistake. The text of well-written
3403 documentation strings will be entirely wrong for a manual.
3405 A documentation string needs to stand alone---when it appears on the
3406 screen, there will be no other text to introduce or explain it.
3407 Meanwhile, it can be rather informal in style.
3409 The text describing a function or variable in a manual must not stand
3410 alone; it appears in the context of a section or subsection. Other text
3411 at the beginning of the section should explain some of the concepts, and
3412 should often make some general points that apply to several functions or
3413 variables. The previous descriptions of functions and variables in the
3414 section will also have given information about the topic. A description
3415 written to stand alone would repeat some of that information; this
3416 redundancy looks bad. Meanwhile, the informality that is acceptable in
3417 a documentation string is totally unacceptable in a manual.
3419 The only good way to use documentation strings in writing a good manual
3420 is to use them as a source of information for writing good text.
3422 @node Manual Structure Details
3423 @section Manual Structure Details
3424 @cindex manual structure
3426 The title page of the manual should state the version of the programs or
3427 packages documented in the manual. The Top node of the manual should
3428 also contain this information. If the manual is changing more
3429 frequently than or independent of the program, also state a version
3430 number for the manual in both of these places.
3432 Each program documented in the manual should have a node named
3433 @samp{@var{program} Invocation} or @samp{Invoking @var{program}}. This
3434 node (together with its subnodes, if any) should describe the program's
3435 command line arguments and how to run it (the sort of information people
3436 would look for in a man page). Start with an @samp{@@example}
3437 containing a template for all the options and arguments that the program
3440 Alternatively, put a menu item in some menu whose item name fits one of
3441 the above patterns. This identifies the node which that item points to
3442 as the node for this purpose, regardless of the node's actual name.
3444 The @samp{--usage} feature of the Info reader looks for such a node
3445 or menu item in order to find the relevant text, so it is essential
3446 for every Texinfo file to have one.
3448 If one manual describes several programs, it should have such a node for
3449 each program described in the manual.
3451 @node License for Manuals
3452 @section License for Manuals
3453 @cindex license for manuals
3455 Please use the GNU Free Documentation License for all GNU manuals that
3456 are more than a few pages long. Likewise for a collection of short
3457 documents---you only need one copy of the GNU FDL for the whole
3458 collection. For a single short document, you can use a very permissive
3459 non-copyleft license, to avoid taking up space with a long license.
3461 See @uref{http://www.gnu.org/copyleft/fdl-howto.html} for more explanation
3462 of how to employ the GFDL.
3464 Note that it is not obligatory to include a copy of the GNU GPL or GNU
3465 LGPL in a manual whose license is neither the GPL nor the LGPL. It can
3466 be a good idea to include the program's license in a large manual; in a
3467 short manual, whose size would be increased considerably by including
3468 the program's license, it is probably better not to include it.
3470 @node Manual Credits
3471 @section Manual Credits
3472 @cindex credits for manuals
3474 Please credit the principal human writers of the manual as the authors,
3475 on the title page of the manual. If a company sponsored the work, thank
3476 the company in a suitable place in the manual, but do not cite the
3477 company as an author.
3479 @node Printed Manuals
3480 @section Printed Manuals
3482 The FSF publishes some GNU manuals in printed form. To encourage sales
3483 of these manuals, the on-line versions of the manual should mention at
3484 the very start that the printed manual is available and should point at
3485 information for getting it---for instance, with a link to the page
3486 @url{http://www.gnu.org/order/order.html}. This should not be included
3487 in the printed manual, though, because there it is redundant.
3489 It is also useful to explain in the on-line forms of the manual how the
3490 user can print out the manual from the sources.
3493 @section The NEWS File
3494 @cindex @file{NEWS} file
3496 In addition to its manual, the package should have a file named
3497 @file{NEWS} which contains a list of user-visible changes worth
3498 mentioning. In each new release, add items to the front of the file and
3499 identify the version they pertain to. Don't discard old items; leave
3500 them in the file after the newer items. This way, a user upgrading from
3501 any previous version can see what is new.
3503 If the @file{NEWS} file gets very long, move some of the older items
3504 into a file named @file{ONEWS} and put a note at the end referring the
3508 @section Change Logs
3511 Keep a change log to describe all the changes made to program source
3512 files. The purpose of this is so that people investigating bugs in the
3513 future will know about the changes that might have introduced the bug.
3514 Often a new bug can be found by looking at what was recently changed.
3515 More importantly, change logs can help you eliminate conceptual
3516 inconsistencies between different parts of a program, by giving you a
3517 history of how the conflicting concepts arose and who they came from.
3520 * Change Log Concepts::
3521 * Style of Change Logs::
3523 * Conditional Changes::
3524 * Indicating the Part Changed::
3527 @node Change Log Concepts
3528 @subsection Change Log Concepts
3530 You can think of the change log as a conceptual ``undo list'' which
3531 explains how earlier versions were different from the current version.
3532 People can see the current version; they don't need the change log
3533 to tell them what is in it. What they want from a change log is a
3534 clear explanation of how the earlier version differed.
3536 The change log file is normally called @file{ChangeLog} and covers an
3537 entire directory. Each directory can have its own change log, or a
3538 directory can use the change log of its parent directory---it's up to
3541 Another alternative is to record change log information with a version
3542 control system such as RCS or CVS. This can be converted automatically
3543 to a @file{ChangeLog} file using @code{rcs2log}; in Emacs, the command
3544 @kbd{C-x v a} (@code{vc-update-change-log}) does the job.
3546 There's no need to describe the full purpose of the changes or how
3547 they work together. However, sometimes it is useful to write one line
3548 to describe the overall purpose of a change or a batch of changes. If
3549 you think that a change calls for explanation, you're probably right.
3550 Please do explain it---but please put the full explanation in comments
3551 in the code, where people will see it whenever they see the code. For
3552 example, ``New function'' is enough for the change log when you add a
3553 function, because there should be a comment before the function
3554 definition to explain what it does.
3556 In the past, we recommended not mentioning changes in non-software
3557 files (manuals, help files, etc.) in change logs. However, we've been
3558 advised that it is a good idea to include them, for the sake of
3561 The easiest way to add an entry to @file{ChangeLog} is with the Emacs
3562 command @kbd{M-x add-change-log-entry}. An entry should have an
3563 asterisk, the name of the changed file, and then in parentheses the name
3564 of the changed functions, variables or whatever, followed by a colon.
3565 Then describe the changes you made to that function or variable.
3567 @node Style of Change Logs
3568 @subsection Style of Change Logs
3569 @cindex change logs, style
3571 Here are some simple examples of change log entries, starting with the
3572 header line that says who made the change and when it was installed,
3573 followed by descriptions of specific changes. (These examples are
3574 drawn from Emacs and GCC.)
3577 1998-08-17 Richard Stallman <rms@@gnu.org>
3579 * register.el (insert-register): Return nil.
3580 (jump-to-register): Likewise.
3582 * sort.el (sort-subr): Return nil.
3584 * tex-mode.el (tex-bibtex-file, tex-file, tex-region):
3585 Restart the tex shell if process is gone or stopped.
3586 (tex-shell-running): New function.
3588 * expr.c (store_one_arg): Round size up for move_block_to_reg.
3589 (expand_call): Round up when emitting USE insns.
3590 * stmt.c (assign_parms): Round size up for move_block_from_reg.
3593 It's important to name the changed function or variable in full. Don't
3594 abbreviate function or variable names, and don't combine them.
3595 Subsequent maintainers will often search for a function name to find all
3596 the change log entries that pertain to it; if you abbreviate the name,
3597 they won't find it when they search.
3599 For example, some people are tempted to abbreviate groups of function
3600 names by writing @samp{* register.el (@{insert,jump-to@}-register)};
3601 this is not a good idea, since searching for @code{jump-to-register} or
3602 @code{insert-register} would not find that entry.
3604 Separate unrelated change log entries with blank lines. When two
3605 entries represent parts of the same change, so that they work together,
3606 then don't put blank lines between them. Then you can omit the file
3607 name and the asterisk when successive entries are in the same file.
3609 Break long lists of function names by closing continued lines with
3610 @samp{)}, rather than @samp{,}, and opening the continuation with
3611 @samp{(} as in this example:
3614 * keyboard.c (menu_bar_items, tool_bar_items)
3615 (Fexecute_extended_command): Deal with `keymap' property.
3618 When you install someone else's changes, put the contributor's name in
3619 the change log entry rather than in the text of the entry. In other
3623 2002-07-14 John Doe <jdoe@@gnu.org>
3625 * sewing.c: Make it sew.
3632 2002-07-14 Usual Maintainer <usual@@gnu.org>
3634 * sewing.c: Make it sew. Patch by jdoe@@gnu.org.
3637 As for the date, that should be the date you applied the change.
3639 @node Simple Changes
3640 @subsection Simple Changes
3642 Certain simple kinds of changes don't need much detail in the change
3645 When you change the calling sequence of a function in a simple fashion,
3646 and you change all the callers of the function to use the new calling
3647 sequence, there is no need to make individual entries for all the
3648 callers that you changed. Just write in the entry for the function
3649 being called, ``All callers changed''---like this:
3652 * keyboard.c (Fcommand_execute): New arg SPECIAL.
3653 All callers changed.
3656 When you change just comments or doc strings, it is enough to write an
3657 entry for the file, without mentioning the functions. Just ``Doc
3658 fixes'' is enough for the change log.
3660 There's no technical need to make change log entries for documentation
3661 files. This is because documentation is not susceptible to bugs that
3662 are hard to fix. Documentation does not consist of parts that must
3663 interact in a precisely engineered fashion. To correct an error, you
3664 need not know the history of the erroneous passage; it is enough to
3665 compare what the documentation says with the way the program actually
3668 However, you should keep change logs for documentation files when the
3669 project gets copyright assignments from its contributors, so as to
3670 make the records of authorship more accurate.
3672 @node Conditional Changes
3673 @subsection Conditional Changes
3674 @cindex conditional changes, and change logs
3675 @cindex change logs, conditional changes
3677 C programs often contain compile-time @code{#if} conditionals. Many
3678 changes are conditional; sometimes you add a new definition which is
3679 entirely contained in a conditional. It is very useful to indicate in
3680 the change log the conditions for which the change applies.
3682 Our convention for indicating conditional changes is to use square
3683 brackets around the name of the condition.
3685 Here is a simple example, describing a change which is conditional but
3686 does not have a function or entity name associated with it:
3689 * xterm.c [SOLARIS2]: Include string.h.
3692 Here is an entry describing a new definition which is entirely
3693 conditional. This new definition for the macro @code{FRAME_WINDOW_P} is
3694 used only when @code{HAVE_X_WINDOWS} is defined:
3697 * frame.h [HAVE_X_WINDOWS] (FRAME_WINDOW_P): Macro defined.
3700 Here is an entry for a change within the function @code{init_display},
3701 whose definition as a whole is unconditional, but the changes themselves
3702 are contained in a @samp{#ifdef HAVE_LIBNCURSES} conditional:
3705 * dispnew.c (init_display) [HAVE_LIBNCURSES]: If X, call tgetent.
3708 Here is an entry for a change that takes affect only when
3709 a certain macro is @emph{not} defined:
3712 (gethostname) [!HAVE_SOCKETS]: Replace with winsock version.
3715 @node Indicating the Part Changed
3716 @subsection Indicating the Part Changed
3718 Indicate the part of a function which changed by using angle brackets
3719 enclosing an indication of what the changed part does. Here is an entry
3720 for a change in the part of the function @code{sh-while-getopts} that
3721 deals with @code{sh} commands:
3724 * progmodes/sh-script.el (sh-while-getopts) <sh>: Handle case that
3725 user-specified option string is empty.
3733 In the GNU project, man pages are secondary. It is not necessary or
3734 expected for every GNU program to have a man page, but some of them do.
3735 It's your choice whether to include a man page in your program.
3737 When you make this decision, consider that supporting a man page
3738 requires continual effort each time the program is changed. The time
3739 you spend on the man page is time taken away from more useful work.
3741 For a simple program which changes little, updating the man page may be
3742 a small job. Then there is little reason not to include a man page, if
3745 For a large program that changes a great deal, updating a man page may
3746 be a substantial burden. If a user offers to donate a man page, you may
3747 find this gift costly to accept. It may be better to refuse the man
3748 page unless the same person agrees to take full responsibility for
3749 maintaining it---so that you can wash your hands of it entirely. If
3750 this volunteer later ceases to do the job, then don't feel obliged to
3751 pick it up yourself; it may be better to withdraw the man page from the
3752 distribution until someone else agrees to update it.
3754 When a program changes only a little, you may feel that the
3755 discrepancies are small enough that the man page remains useful without
3756 updating. If so, put a prominent note near the beginning of the man
3757 page explaining that you don't maintain it and that the Texinfo manual
3758 is more authoritative. The note should say how to access the Texinfo
3761 Be sure that man pages include a copyright statement and free license.
3762 The simple all-permissive license is appropriate for simple man pages
3763 (@pxref{License Notices for Other Files,,,maintain,Information for GNU
3766 For long man pages, with enough explanation and documentation that
3767 they can be considered true manuals, use the GFDL (@pxref{License for
3770 Finally, the GNU help2man program
3771 (@uref{http://www.gnu.org/software/help2man/}) is one way to automate
3772 generation of a man page, in this case from @option{--help} output.
3773 This is sufficient in many cases.
3775 @node Reading other Manuals
3776 @section Reading other Manuals
3778 There may be non-free books or documentation files that describe the
3779 program you are documenting.
3781 It is ok to use these documents for reference, just as the author of a
3782 new algebra textbook can read other books on algebra. A large portion
3783 of any non-fiction book consists of facts, in this case facts about how
3784 a certain program works, and these facts are necessarily the same for
3785 everyone who writes about the subject. But be careful not to copy your
3786 outline structure, wording, tables or examples from preexisting non-free
3787 documentation. Copying from free documentation may be ok; please check
3788 with the FSF about the individual case.
3790 @node Managing Releases
3791 @chapter The Release Process
3794 Making a release is more than just bundling up your source files in a
3795 tar file and putting it up for FTP. You should set up your software so
3796 that it can be configured to run on a variety of systems. Your Makefile
3797 should conform to the GNU standards described below, and your directory
3798 layout should also conform to the standards discussed below. Doing so
3799 makes it easy to include your package into the larger framework of
3803 * Configuration:: How configuration of GNU packages should work.
3804 * Makefile Conventions:: Makefile conventions.
3805 * Releases:: Making releases
3809 @section How Configuration Should Work
3810 @cindex program configuration
3813 Each GNU distribution should come with a shell script named
3814 @code{configure}. This script is given arguments which describe the
3815 kind of machine and system you want to compile the program for.
3816 The @code{configure} script must record the configuration options so
3817 that they affect compilation.
3819 The description here is the specification of the interface for the
3820 @code{configure} script in GNU packages. Many packages implement it
3821 using GNU Autoconf (@pxref{Top,, Introduction, autoconf, Autoconf})
3822 and/or GNU Automake (@pxref{Top,, Introduction, automake, Automake}),
3823 but you do not have to use these tools. You can implement it any way
3824 you like; for instance, by making @code{configure} be a wrapper around
3825 a completely different configuration system.
3827 Another way for the @code{configure} script to operate is to make a
3828 link from a standard name such as @file{config.h} to the proper
3829 configuration file for the chosen system. If you use this technique,
3830 the distribution should @emph{not} contain a file named
3831 @file{config.h}. This is so that people won't be able to build the
3832 program without configuring it first.
3834 Another thing that @code{configure} can do is to edit the Makefile. If
3835 you do this, the distribution should @emph{not} contain a file named
3836 @file{Makefile}. Instead, it should include a file @file{Makefile.in} which
3837 contains the input used for editing. Once again, this is so that people
3838 won't be able to build the program without configuring it first.
3840 If @code{configure} does write the @file{Makefile}, then @file{Makefile}
3841 should have a target named @file{Makefile} which causes @code{configure}
3842 to be rerun, setting up the same configuration that was set up last
3843 time. The files that @code{configure} reads should be listed as
3844 dependencies of @file{Makefile}.
3846 All the files which are output from the @code{configure} script should
3847 have comments at the beginning explaining that they were generated
3848 automatically using @code{configure}. This is so that users won't think
3849 of trying to edit them by hand.
3851 The @code{configure} script should write a file named @file{config.status}
3852 which describes which configuration options were specified when the
3853 program was last configured. This file should be a shell script which,
3854 if run, will recreate the same configuration.
3856 The @code{configure} script should accept an option of the form
3857 @samp{--srcdir=@var{dirname}} to specify the directory where sources are found
3858 (if it is not the current directory). This makes it possible to build
3859 the program in a separate directory, so that the actual source directory
3862 If the user does not specify @samp{--srcdir}, then @code{configure} should
3863 check both @file{.} and @file{..} to see if it can find the sources. If
3864 it finds the sources in one of these places, it should use them from
3865 there. Otherwise, it should report that it cannot find the sources, and
3866 should exit with nonzero status.
3868 Usually the easy way to support @samp{--srcdir} is by editing a
3869 definition of @code{VPATH} into the Makefile. Some rules may need to
3870 refer explicitly to the specified source directory. To make this
3871 possible, @code{configure} can add to the Makefile a variable named
3872 @code{srcdir} whose value is precisely the specified directory.
3874 In addition, the @samp{configure} script should take options
3875 corresponding to most of the standard directory variables
3876 (@pxref{Directory Variables}). Here is the list:
3879 --prefix --exec-prefix --bindir --sbindir --libexecdir --sysconfdir
3880 --sharedstatedir --localstatedir --libdir --includedir --oldincludedir
3881 --datarootdir --datadir --infodir --localedir --mandir --docdir
3882 --htmldir --dvidir --pdfdir --psdir
3885 The @code{configure} script should also take an argument which specifies the
3886 type of system to build the program for. This argument should look like
3890 @var{cpu}-@var{company}-@var{system}
3893 For example, an Athlon-based GNU/Linux system might be
3894 @samp{i686-pc-linux-gnu}.
3896 The @code{configure} script needs to be able to decode all plausible
3897 alternatives for how to describe a machine. Thus,
3898 @samp{athlon-pc-gnu/linux} would be a valid alias. There is a shell
3900 @uref{http://git.savannah.gnu.org/@/gitweb/@/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD,
3901 @file{config.sub}} that you can use as a subroutine to validate system
3902 types and canonicalize aliases.
3904 The @code{configure} script should also take the option
3905 @option{--build=@var{buildtype}}, which should be equivalent to a
3906 plain @var{buildtype} argument. For example, @samp{configure
3907 --build=i686-pc-linux-gnu} is equivalent to @samp{configure
3908 i686-pc-linux-gnu}. When the build type is not specified by an option
3909 or argument, the @code{configure} script should normally guess it using
3911 @uref{http://git.savannah.gnu.org/@/gitweb/@/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD,
3912 @file{config.guess}}.
3914 @cindex optional features, configure-time
3915 Other options are permitted to specify in more detail the software
3916 or hardware present on the machine, to include or exclude optional parts
3917 of the package, or to adjust the name of some tools or arguments to them:
3920 @item --enable-@var{feature}@r{[}=@var{parameter}@r{]}
3921 Configure the package to build and install an optional user-level
3922 facility called @var{feature}. This allows users to choose which
3923 optional features to include. Giving an optional @var{parameter} of
3924 @samp{no} should omit @var{feature}, if it is built by default.
3926 No @samp{--enable} option should @strong{ever} cause one feature to
3927 replace another. No @samp{--enable} option should ever substitute one
3928 useful behavior for another useful behavior. The only proper use for
3929 @samp{--enable} is for questions of whether to build part of the program
3932 @item --with-@var{package}
3933 @c @r{[}=@var{parameter}@r{]}
3934 The package @var{package} will be installed, so configure this package
3935 to work with @var{package}.
3937 @c Giving an optional @var{parameter} of
3938 @c @samp{no} should omit @var{package}, if it is used by default.
3940 Possible values of @var{package} include
3941 @samp{gnu-as} (or @samp{gas}), @samp{gnu-ld}, @samp{gnu-libc},
3947 Do not use a @samp{--with} option to specify the file name to use to
3948 find certain files. That is outside the scope of what @samp{--with}
3951 @item @var{variable}=@var{value}
3952 Set the value of the variable @var{variable} to @var{value}. This is
3953 used to override the default values of commands or arguments in the
3954 build process. For example, the user could issue @samp{configure
3955 CFLAGS=-g CXXFLAGS=-g} to build with debugging information and without
3956 the default optimization.
3958 Specifying variables as arguments to @code{configure}, like this:
3962 is preferable to setting them in environment variables:
3966 as it helps to recreate the same configuration later with
3967 @file{config.status}. However, both methods should be supported.
3970 All @code{configure} scripts should accept all of the ``detail''
3971 options and the variable settings, whether or not they make any
3972 difference to the particular package at hand. In particular, they
3973 should accept any option that starts with @samp{--with-} or
3974 @samp{--enable-}. This is so users will be able to configure an
3975 entire GNU source tree at once with a single set of options.
3977 You will note that the categories @samp{--with-} and @samp{--enable-}
3978 are narrow: they @strong{do not} provide a place for any sort of option
3979 you might think of. That is deliberate. We want to limit the possible
3980 configuration options in GNU software. We do not want GNU programs to
3981 have idiosyncratic configuration options.
3983 Packages that perform part of the compilation process may support
3984 cross-compilation. In such a case, the host and target machines for the
3985 program may be different.
3987 The @code{configure} script should normally treat the specified type of
3988 system as both the host and the target, thus producing a program which
3989 works for the same type of machine that it runs on.
3991 To compile a program to run on a host type that differs from the build
3992 type, use the configure option @option{--host=@var{hosttype}}, where
3993 @var{hosttype} uses the same syntax as @var{buildtype}. The host type
3994 normally defaults to the build type.
3996 To configure a cross-compiler, cross-assembler, or what have you, you
3997 should specify a target different from the host, using the configure
3998 option @samp{--target=@var{targettype}}. The syntax for
3999 @var{targettype} is the same as for the host type. So the command would
4003 ./configure --host=@var{hosttype} --target=@var{targettype}
4006 The target type normally defaults to the host type.
4007 Programs for which cross-operation is not meaningful need not accept the
4008 @samp{--target} option, because configuring an entire operating system for
4009 cross-operation is not a meaningful operation.
4011 Some programs have ways of configuring themselves automatically. If
4012 your program is set up to do this, your @code{configure} script can simply
4013 ignore most of its arguments.
4015 @comment The makefile standards are in a separate file that is also
4016 @comment included by make.texinfo. Done by roland@gnu.ai.mit.edu on 1/6/93.
4017 @comment For this document, turn chapters into sections, etc.
4019 @include make-stds.texi
4023 @section Making Releases
4026 You should identify each release with a pair of version numbers, a
4027 major version and a minor. We have no objection to using more than
4028 two numbers, but it is very unlikely that you really need them.
4030 Package the distribution of @code{Foo version 69.96} up in a gzipped tar
4031 file with the name @file{foo-69.96.tar.gz}. It should unpack into a
4032 subdirectory named @file{foo-69.96}.
4034 Building and installing the program should never modify any of the files
4035 contained in the distribution. This means that all the files that form
4036 part of the program in any way must be classified into @dfn{source
4037 files} and @dfn{non-source files}. Source files are written by humans
4038 and never changed automatically; non-source files are produced from
4039 source files by programs under the control of the Makefile.
4041 @cindex @file{README} file
4042 The distribution should contain a file named @file{README} which gives
4043 the name of the package, and a general description of what it does. It
4044 is also good to explain the purpose of each of the first-level
4045 subdirectories in the package, if there are any. The @file{README} file
4046 should either state the version number of the package, or refer to where
4047 in the package it can be found.
4049 The @file{README} file should refer to the file @file{INSTALL}, which
4050 should contain an explanation of the installation procedure.
4052 The @file{README} file should also refer to the file which contains the
4053 copying conditions. The GNU GPL, if used, should be in a file called
4054 @file{COPYING}. If the GNU LGPL is used, it should be in a file called
4055 @file{COPYING.LESSER}.
4057 Naturally, all the source files must be in the distribution. It is okay
4058 to include non-source files in the distribution, provided they are
4059 up-to-date and machine-independent, so that building the distribution
4060 normally will never modify them. We commonly include non-source files
4061 produced by Bison, @code{lex}, @TeX{}, and @code{makeinfo}; this helps avoid
4062 unnecessary dependencies between our distributions, so that users can
4063 install whichever packages they want to install.
4065 Non-source files that might actually be modified by building and
4066 installing the program should @strong{never} be included in the
4067 distribution. So if you do distribute non-source files, always make
4068 sure they are up to date when you make a new distribution.
4070 Make sure that all the files in the distribution are world-readable, and
4071 that directories are world-readable and world-searchable (octal mode 755).
4072 We used to recommend that all directories in the distribution also be
4073 world-writable (octal mode 777), because ancient versions of @code{tar}
4074 would otherwise not cope when extracting the archive as an unprivileged
4075 user. That can easily lead to security issues when creating the archive,
4076 however, so now we recommend against that.
4078 Don't include any symbolic links in the distribution itself. If the tar
4079 file contains symbolic links, then people cannot even unpack it on
4080 systems that don't support symbolic links. Also, don't use multiple
4081 names for one file in different directories, because certain file
4082 systems cannot handle this and that prevents unpacking the
4085 Try to make sure that all the file names will be unique on MS-DOS. A
4086 name on MS-DOS consists of up to 8 characters, optionally followed by a
4087 period and up to three characters. MS-DOS will truncate extra
4088 characters both before and after the period. Thus,
4089 @file{foobarhacker.c} and @file{foobarhacker.o} are not ambiguous; they
4090 are truncated to @file{foobarha.c} and @file{foobarha.o}, which are
4093 @cindex @file{texinfo.tex}, in a distribution
4094 Include in your distribution a copy of the @file{texinfo.tex} you used
4095 to test print any @file{*.texinfo} or @file{*.texi} files.
4097 Likewise, if your program uses small GNU software packages like regex,
4098 getopt, obstack, or termcap, include them in the distribution file.
4099 Leaving them out would make the distribution file a little smaller at
4100 the expense of possible inconvenience to a user who doesn't know what
4104 @chapter References to Non-Free Software and Documentation
4105 @cindex references to non-free material
4107 A GNU program should not recommend, promote, or grant legitimacy to
4108 the use of any non-free program. Proprietary software is a social and
4109 ethical problem, and our aim is to put an end to that problem. We
4110 can't stop some people from writing proprietary programs, or stop
4111 other people from using them, but we can and should refuse to
4112 advertise them to new potential customers, or to give the public the
4113 idea that their existence is ethical.
4115 The GNU definition of free software is found on the GNU web site at
4116 @url{http://www.gnu.org/@/philosophy/@/free-sw.html}, and the definition
4117 of free documentation is found at
4118 @url{http://www.gnu.org/@/philosophy/@/free-doc.html}. The terms ``free''
4119 and ``non-free'', used in this document, refer to those definitions.
4121 A list of important licenses and whether they qualify as free is in
4122 @url{http://www.gnu.org/@/licenses/@/license-list.html}. If it is not
4123 clear whether a license qualifies as free, please ask the GNU Project
4124 by writing to @email{licensing@@gnu.org}. We will answer, and if the
4125 license is an important one, we will add it to the list.
4127 When a non-free program or system is well known, you can mention it in
4128 passing---that is harmless, since users who might want to use it
4129 probably already know about it. For instance, it is fine to explain
4130 how to build your package on top of some widely used non-free
4131 operating system, or how to use it together with some widely used
4134 However, you should give only the necessary information to help those
4135 who already use the non-free program to use your program with
4136 it---don't give, or refer to, any further information about the
4137 proprietary program, and don't imply that the proprietary program
4138 enhances your program, or that its existence is in any way a good
4139 thing. The goal should be that people already using the proprietary
4140 program will get the advice they need about how to use your free
4141 program with it, while people who don't already use the proprietary
4142 program will not see anything likely to lead them to take an interest
4145 If a non-free program or system is obscure in your program's domain,
4146 your program should not mention or support it at all, since doing so
4147 would tend to popularize the non-free program more than it popularizes
4148 your program. (You cannot hope to find many additional users for your
4149 program among the users of Foobar, if the existence of Foobar is not
4150 generally known among people who might want to use your program.)
4152 Sometimes a program is free software in itself but depends on a
4153 non-free platform in order to run. For instance, many Java programs
4154 depend on some non-free Java libraries. To recommend or promote such
4155 a program is to promote the other programs it needs. This is why we
4156 are careful about listing Java programs in the Free Software
4157 Directory: we don't want to promote the non-free Java libraries.
4159 We hope this particular problem with Java will be gone by and by, as
4160 we replace the remaining non-free standard Java libraries with free
4161 software, but the general principle will remain the same: don't
4162 recommend, promote or legitimize programs that depend on non-free
4165 Some free programs strongly encourage the use of non-free software. A
4166 typical example is @command{mplayer}. It is free software in itself,
4167 and the free code can handle some kinds of files. However,
4168 @command{mplayer} recommends use of non-free codecs for other kinds of
4169 files, and users that install @command{mplayer} are very likely to
4170 install those codecs along with it. To recommend @command{mplayer}
4171 is, in effect, to promote use of the non-free codecs.
4173 Thus, you should not recommend programs that strongly encourage the
4174 use of non-free software. This is why we do not list
4175 @command{mplayer} in the Free Software Directory.
4177 A GNU package should not refer the user to any non-free documentation
4178 for free software. Free documentation that can be included in free
4179 operating systems is essential for completing the GNU system, or any
4180 free operating system, so encouraging it is a priority; to recommend
4181 use of documentation that we are not allowed to include undermines the
4182 impetus for the community to produce documentation that we can
4183 include. So GNU packages should never recommend non-free
4186 By contrast, it is ok to refer to journal articles and textbooks in
4187 the comments of a program for explanation of how it functions, even
4188 though they are non-free. This is because we don't include such
4189 things in the GNU system even they are free---they are outside the
4190 scope of what a software distribution needs to include.
4192 Referring to a web site that describes or recommends a non-free
4193 program is promoting that program, so please do not make links (or
4194 mention by name) web sites that contain such material. This policy is
4195 relevant particularly for the web pages for a GNU package.
4197 Following links from nearly any web site can lead eventually to
4198 non-free software; this is inherent in the nature of the web. So it
4199 makes no sense to criticize a site for having such links. As long as
4200 the site does not itself recommend a non-free program, there is no
4201 need to consider the question of the sites that it links to for other
4204 Thus, for example, you should not refer to AT&T's web site if that
4205 recommends AT&T's non-free software packages; you should not refer to
4206 a site that links to AT&T's site presenting it as a place to get some
4207 non-free program, because that link recommends and legitimizes the
4208 non-free program. However, that a site contains a link to AT&T's web
4209 site for some other purpose (such as long-distance telephone service)
4210 is not an objection against it.
4212 @node GNU Free Documentation License
4213 @appendix GNU Free Documentation License
4215 @cindex FDL, GNU Free Documentation License
4225 eval: (add-hook 'write-file-hooks 'time-stamp)
4226 time-stamp-start: "@set lastupdate "
4228 time-stamp-format: "%:b %:d, %:y"
4229 compile-command: "cd work.s && make"