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 May 5, 2008
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 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 1.2
35 or any later version published by the Free Software Foundation;
36 with no Invariant Sections, with no
37 Front-Cover Texts, and with no Back-Cover Texts.
38 A copy of the license is included in the section entitled ``GNU
39 Free Documentation License''.
43 @title GNU Coding Standards
44 @author Richard Stallman, et al.
45 @author last updated @value{lastupdate}
47 @vskip 0pt plus 1filll
54 @node Top, Preface, (dir), (dir)
61 * Preface:: About the GNU Coding Standards.
62 * Legal Issues:: Keeping free software free.
63 * Design Advice:: General program design.
64 * Program Behavior:: Program behavior for all programs
65 * Writing C:: Making the best use of C.
66 * Documentation:: Documenting programs.
67 * Managing Releases:: The release process.
68 * References:: Mentioning non-free software or documentation.
69 * GNU Free Documentation License:: Copying and sharing this manual.
75 @chapter About the GNU Coding Standards
77 The GNU Coding Standards were written by Richard Stallman and other GNU
78 Project volunteers. Their purpose is to make the GNU system clean,
79 consistent, and easy to install. This document can also be read as a
80 guide to writing portable, robust and reliable programs. It focuses on
81 programs written in C, but many of the rules and principles are useful
82 even if you write in another programming language. The rules often
83 state reasons for writing in a certain way.
85 This release of the GNU Coding Standards was last updated
88 @cindex where to obtain @code{standards.texi}
89 @cindex downloading this manual
90 If you did not obtain this file directly from the GNU project and
91 recently, please check for a newer version. You can get the GNU
92 Coding Standards from the GNU web server in many
93 different formats, including the Texinfo source, PDF, HTML, DVI, plain
94 text, and more, at: @uref{http://www.gnu.org/prep/standards/}.
96 Corrections or suggestions for this document should be sent to
97 @email{bug-standards@@gnu.org}. If you make a suggestion, please include a
98 suggested new wording for it; our time is limited. We prefer a context
99 diff to the @file{standards.texi} or @file{make-stds.texi} files, but if
100 you don't have those files, please mail your suggestion anyway.
102 These standards cover the minimum of what is important when writing a
103 GNU package. Likely, the need for additional standards will come up.
104 Sometimes, you might suggest that such standards be added to this
105 document. If you think your standards would be generally useful, please
108 You should also set standards for your package on many questions not
109 addressed or not firmly specified here. The most important point is to
110 be self-consistent---try to stick to the conventions you pick, and try
111 to document them as much as possible. That way, your program will be
112 more maintainable by others.
114 The GNU Hello program serves as an example of how to follow the GNU
115 coding standards for a trivial program.
116 @uref{http://www.gnu.org/software/hello/hello.html}.
119 @chapter Keeping Free Software Free
120 @cindex legal aspects
122 This chapter discusses how you can make sure that GNU software
123 avoids legal difficulties, and other related issues.
126 * Reading Non-Free Code:: Referring to proprietary programs.
127 * Contributions:: Accepting contributions.
128 * Trademarks:: How we deal with trademark issues.
131 @node Reading Non-Free Code
132 @section Referring to Proprietary Programs
133 @cindex proprietary programs
134 @cindex avoiding proprietary code
136 Don't in any circumstances refer to Unix source code for or during
137 your work on GNU! (Or to any other proprietary programs.)
139 If you have a vague recollection of the internals of a Unix program,
140 this does not absolutely mean you can't write an imitation of it, but
141 do try to organize the imitation internally along different lines,
142 because this is likely to make the details of the Unix version
143 irrelevant and dissimilar to your results.
145 For example, Unix utilities were generally optimized to minimize
146 memory use; if you go for speed instead, your program will be very
147 different. You could keep the entire input file in memory and scan it
148 there instead of using stdio. Use a smarter algorithm discovered more
149 recently than the Unix program. Eliminate use of temporary files. Do
150 it in one pass instead of two (we did this in the assembler).
152 Or, on the contrary, emphasize simplicity instead of speed. For some
153 applications, the speed of today's computers makes simpler algorithms
156 Or go for generality. For example, Unix programs often have static
157 tables or fixed-size strings, which make for arbitrary limits; use
158 dynamic allocation instead. Make sure your program handles NULs and
159 other funny characters in the input files. Add a programming language
160 for extensibility and write part of the program in that language.
162 Or turn some parts of the program into independently usable libraries.
163 Or use a simple garbage collector instead of tracking precisely when
164 to free memory, or use a new GNU facility such as obstacks.
167 @section Accepting Contributions
169 @cindex accepting contributions
171 If the program you are working on is copyrighted by the Free Software
172 Foundation, then when someone else sends you a piece of code to add to
173 the program, we need legal papers to use it---just as we asked you to
174 sign papers initially. @emph{Each} person who makes a nontrivial
175 contribution to a program must sign some sort of legal papers in order
176 for us to have clear title to the program; the main author alone is not
179 So, before adding in any contributions from other people, please tell
180 us, so we can arrange to get the papers. Then wait until we tell you
181 that we have received the signed papers, before you actually use the
184 This applies both before you release the program and afterward. If
185 you receive diffs to fix a bug, and they make significant changes, we
186 need legal papers for that change.
188 This also applies to comments and documentation files. For copyright
189 law, comments and code are just text. Copyright applies to all kinds of
190 text, so we need legal papers for all kinds.
192 We know it is frustrating to ask for legal papers; it's frustrating for
193 us as well. But if you don't wait, you are going out on a limb---for
194 example, what if the contributor's employer won't sign a disclaimer?
195 You might have to take that code out again!
197 You don't need papers for changes of a few lines here or there, since
198 they are not significant for copyright purposes. Also, you don't need
199 papers if all you get from the suggestion is some ideas, not actual code
200 which you use. For example, if someone sent you one implementation, but
201 you write a different implementation of the same idea, you don't need to
204 The very worst thing is if you forget to tell us about the other
205 contributor. We could be very embarrassed in court some day as a
208 We have more detailed advice for maintainers of programs; if you have
209 reached the stage of actually maintaining a program for GNU (whether
210 released or not), please ask us for a copy. It is also available
211 online for your perusal: @uref{http://www.gnu.org/prep/maintain/}.
217 Please do not include any trademark acknowledgements in GNU software
218 packages or documentation.
220 Trademark acknowledgements are the statements that such-and-such is a
221 trademark of so-and-so. The GNU Project has no objection to the basic
222 idea of trademarks, but these acknowledgements feel like kowtowing,
223 and there is no legal requirement for them, so we don't use them.
225 What is legally required, as regards other people's trademarks, is to
226 avoid using them in ways which a reader might reasonably understand as
227 naming or labeling our own programs or activities. For example, since
228 ``Objective C'' is (or at least was) a trademark, we made sure to say
229 that we provide a ``compiler for the Objective C language'' rather
230 than an ``Objective C compiler''. The latter would have been meant as
231 a shorter way of saying the former, but it does not explicitly state
232 the relationship, so it could be misinterpreted as using ``Objective
233 C'' as a label for the compiler rather than for the language.
235 Please don't use ``win'' as an abbreviation for Microsoft Windows in
236 GNU software or documentation. In hacker terminology, calling
237 something a ``win'' is a form of praise. If you wish to praise
238 Microsoft Windows when speaking on your own, by all means do so, but
239 not in GNU software. Usually we write the name ``Windows'' in full,
240 but when brevity is very important (as in file names and sometimes
241 symbol names), we abbreviate it to ``w''. For instance, the files and
242 functions in Emacs that deal with Windows start with @samp{w32}.
245 @chapter General Program Design
246 @cindex program design
248 This chapter discusses some of the issues you should take into
249 account when designing your program.
251 @c Standard or ANSI C
253 @c In 1989 the American National Standards Institute (ANSI) standardized
254 @c C as standard X3.159-1989. In December of that year the
255 @c International Standards Organization ISO adopted the ANSI C standard
256 @c making minor changes. In 1990 ANSI then re-adopted ISO standard
257 @c C. This version of C is known as either ANSI C or Standard C.
259 @c A major revision of the C Standard appeared in 1999.
262 * Source Language:: Which languages to use.
263 * Compatibility:: Compatibility with other implementations.
264 * Using Extensions:: Using non-standard features.
265 * Standard C:: Using standard C features.
266 * Conditional Compilation:: Compiling code only if a conditional is true.
269 @node Source Language
270 @section Which Languages to Use
271 @cindex programming languages
273 When you want to use a language that gets compiled and runs at high
274 speed, the best language to use is C. Using another language is like
275 using a non-standard feature: it will cause trouble for users. Even if
276 GCC supports the other language, users may find it inconvenient to have
277 to install the compiler for that other language in order to build your
278 program. For example, if you write your program in C++, people will
279 have to install the GNU C++ compiler in order to compile your program.
281 C has one other advantage over C++ and other compiled languages: more
282 people know C, so more people will find it easy to read and modify the
283 program if it is written in C.
285 So in general it is much better to use C, rather than the
286 comparable alternatives.
288 But there are two exceptions to that conclusion:
292 It is no problem to use another language to write a tool specifically
293 intended for use with that language. That is because the only people
294 who want to build the tool will be those who have installed the other
298 If an application is of interest only to a narrow part of the community,
299 then the question of which language it is written in has less effect on
300 other people, so you may as well please yourself.
303 Many programs are designed to be extensible: they include an interpreter
304 for a language that is higher level than C. Often much of the program
305 is written in that language, too. The Emacs editor pioneered this
309 @cindex GNOME and Guile
310 The standard extensibility interpreter for GNU software is Guile
311 (@uref{http://www.gnu.org/@/software/@/guile/}), which implements the
312 language Scheme (an especially clean and simple dialect of Lisp).
313 Guile also includes bindings for GTK+/GNOME, making it practical to
314 write modern GUI functionality within Guile. We don't reject programs
315 written in other ``scripting languages'' such as Perl and Python, but
316 using Guile is very important for the overall consistency of the GNU
321 @section Compatibility with Other Implementations
322 @cindex compatibility with C and @sc{posix} standards
323 @cindex @sc{posix} compatibility
325 With occasional exceptions, utility programs and libraries for GNU
326 should be upward compatible with those in Berkeley Unix, and upward
327 compatible with Standard C if Standard C specifies their
328 behavior, and upward compatible with @sc{posix} if @sc{posix} specifies
331 When these standards conflict, it is useful to offer compatibility
332 modes for each of them.
334 @cindex options for compatibility
335 Standard C and @sc{posix} prohibit many kinds of extensions. Feel
336 free to make the extensions anyway, and include a @samp{--ansi},
337 @samp{--posix}, or @samp{--compatible} option to turn them off.
338 However, if the extension has a significant chance of breaking any real
339 programs or scripts, then it is not really upward compatible. So you
340 should try to redesign its interface to make it upward compatible.
342 @cindex @code{POSIXLY_CORRECT}, environment variable
343 Many GNU programs suppress extensions that conflict with @sc{posix} if the
344 environment variable @code{POSIXLY_CORRECT} is defined (even if it is
345 defined with a null value). Please make your program recognize this
346 variable if appropriate.
348 When a feature is used only by users (not by programs or command
349 files), and it is done poorly in Unix, feel free to replace it
350 completely with something totally different and better. (For example,
351 @code{vi} is replaced with Emacs.) But it is nice to offer a compatible
352 feature as well. (There is a free @code{vi} clone, so we offer it.)
354 Additional useful features are welcome regardless of whether
355 there is any precedent for them.
357 @node Using Extensions
358 @section Using Non-standard Features
359 @cindex non-standard extensions
361 Many GNU facilities that already exist support a number of convenient
362 extensions over the comparable Unix facilities. Whether to use these
363 extensions in implementing your program is a difficult question.
365 On the one hand, using the extensions can make a cleaner program.
366 On the other hand, people will not be able to build the program
367 unless the other GNU tools are available. This might cause the
368 program to work on fewer kinds of machines.
370 With some extensions, it might be easy to provide both alternatives.
371 For example, you can define functions with a ``keyword'' @code{INLINE}
372 and define that as a macro to expand into either @code{inline} or
373 nothing, depending on the compiler.
375 In general, perhaps it is best not to use the extensions if you can
376 straightforwardly do without them, but to use the extensions if they
377 are a big improvement.
379 An exception to this rule are the large, established programs (such as
380 Emacs) which run on a great variety of systems. Using GNU extensions in
381 such programs would make many users unhappy, so we don't do that.
383 Another exception is for programs that are used as part of compilation:
384 anything that must be compiled with other compilers in order to
385 bootstrap the GNU compilation facilities. If these require the GNU
386 compiler, then no one can compile them without having them installed
387 already. That would be extremely troublesome in certain cases.
390 @section Standard C and Pre-Standard C
391 @cindex @sc{ansi} C standard
393 1989 Standard C is widespread enough now that it is ok to use its
394 features in new programs. There is one exception: do not ever use the
395 ``trigraph'' feature of Standard C.
397 1999 Standard C is not widespread yet, so please do not require its
398 features in programs. It is ok to use its features if they are present.
400 However, it is easy to support pre-standard compilers in most programs,
401 so if you know how to do that, feel free. If a program you are
402 maintaining has such support, you should try to keep it working.
404 @cindex function prototypes
405 To support pre-standard C, instead of writing function definitions in
406 standard prototype form,
415 write the definition in pre-standard style like this,
425 and use a separate declaration to specify the argument prototype:
431 You need such a declaration anyway, in a header file, to get the benefit
432 of prototypes in all the files where the function is called. And once
433 you have the declaration, you normally lose nothing by writing the
434 function definition in the pre-standard style.
436 This technique does not work for integer types narrower than @code{int}.
437 If you think of an argument as being of a type narrower than @code{int},
438 declare it as @code{int} instead.
440 There are a few special cases where this technique is hard to use. For
441 example, if a function argument needs to hold the system type
442 @code{dev_t}, you run into trouble, because @code{dev_t} is shorter than
443 @code{int} on some machines; but you cannot use @code{int} instead,
444 because @code{dev_t} is wider than @code{int} on some machines. There
445 is no type you can safely use on all machines in a non-standard
446 definition. The only way to support non-standard C and pass such an
447 argument is to check the width of @code{dev_t} using Autoconf and choose
448 the argument type accordingly. This may not be worth the trouble.
450 In order to support pre-standard compilers that do not recognize
451 prototypes, you may want to use a preprocessor macro like this:
454 /* Declare the prototype for a general external function. */
455 #if defined (__STDC__) || defined (WINDOWSNT)
456 #define P_(proto) proto
462 @node Conditional Compilation
463 @section Conditional Compilation
465 When supporting configuration options already known when building your
466 program we prefer using @code{if (... )} over conditional compilation,
467 as in the former case the compiler is able to perform more extensive
468 checking of all possible code paths.
470 For example, please write
490 A modern compiler such as GCC will generate exactly the same code in
491 both cases, and we have been using similar techniques with good success
492 in several projects. Of course, the former method assumes that
493 @code{HAS_FOO} is defined as either 0 or 1.
495 While this is not a silver bullet solving all portability problems,
496 and is not always appropriate, following this policy would have saved
497 GCC developers many hours, or even days, per year.
499 In the case of function-like macros like @code{REVERSIBLE_CC_MODE} in
500 GCC which cannot be simply used in @code{if( ...)} statements, there is
501 an easy workaround. Simply introduce another macro
502 @code{HAS_REVERSIBLE_CC_MODE} as in the following example:
505 #ifdef REVERSIBLE_CC_MODE
506 #define HAS_REVERSIBLE_CC_MODE 1
508 #define HAS_REVERSIBLE_CC_MODE 0
512 @node Program Behavior
513 @chapter Program Behavior for All Programs
515 This chapter describes conventions for writing robust
516 software. It also describes general standards for error messages, the
517 command line interface, and how libraries should behave.
520 * Non-GNU Standards:: We consider standards such as POSIX;
521 we don't "obey" them.
522 * Semantics:: Writing robust programs.
523 * Libraries:: Library behavior.
524 * Errors:: Formatting error messages.
525 * User Interfaces:: Standards about interfaces generally.
526 * Graphical Interfaces:: Standards for graphical interfaces.
527 * Command-Line Interfaces:: Standards for command line interfaces.
528 * Option Table:: Table of long options.
529 * Memory Usage:: When and how to care about memory needs.
530 * File Usage:: Which files to use, and where.
533 @node Non-GNU Standards
534 @section Non-GNU Standards
536 The GNU Project regards standards published by other organizations as
537 suggestions, not orders. We consider those standards, but we do not
538 ``obey'' them. In developing a GNU program, you should implement
539 an outside standard's specifications when that makes the GNU system
540 better overall in an objective sense. When it doesn't, you shouldn't.
542 In most cases, following published standards is convenient for
543 users---it means that their programs or scripts will work more
544 portably. For instance, GCC implements nearly all the features of
545 Standard C as specified by that standard. C program developers would
546 be unhappy if it did not. And GNU utilities mostly follow
547 specifications of POSIX.2; shell script writers and users would be
548 unhappy if our programs were incompatible.
550 But we do not follow either of these specifications rigidly, and there
551 are specific points on which we decided not to follow them, so as to
552 make the GNU system better for users.
554 For instance, Standard C says that nearly all extensions to C are
555 prohibited. How silly! GCC implements many extensions, some of which
556 were later adopted as part of the standard. If you want these
557 constructs to give an error message as ``required'' by the standard,
558 you must specify @samp{--pedantic}, which was implemented only so that
559 we can say ``GCC is a 100% implementation of the standard,'' not
560 because there is any reason to actually use it.
562 POSIX.2 specifies that @samp{df} and @samp{du} must output sizes by
563 default in units of 512 bytes. What users want is units of 1k, so
564 that is what we do by default. If you want the ridiculous behavior
565 ``required'' by POSIX, you must set the environment variable
566 @samp{POSIXLY_CORRECT} (which was originally going to be named
567 @samp{POSIX_ME_HARDER}).
569 GNU utilities also depart from the letter of the POSIX.2 specification
570 when they support long-named command-line options, and intermixing
571 options with ordinary arguments. This minor incompatibility with
572 POSIX is never a problem in practice, and it is very useful.
574 In particular, don't reject a new feature, or remove an old one,
575 merely because a standard says it is ``forbidden'' or ``deprecated.''
578 @section Writing Robust Programs
580 @cindex arbitrary limits on data
581 Avoid arbitrary limits on the length or number of @emph{any} data
582 structure, including file names, lines, files, and symbols, by allocating
583 all data structures dynamically. In most Unix utilities, ``long lines
584 are silently truncated''. This is not acceptable in a GNU utility.
586 @cindex @code{NUL} characters
587 Utilities reading files should not drop NUL characters, or any other
588 nonprinting characters @emph{including those with codes above 0177}.
589 The only sensible exceptions would be utilities specifically intended
590 for interface to certain types of terminals or printers
591 that can't handle those characters.
592 Whenever possible, try to make programs work properly with
593 sequences of bytes that represent multibyte characters, using encodings
594 such as UTF-8 and others.
596 @cindex error messages
597 Check every system call for an error return, unless you know you wish to
598 ignore errors. Include the system error text (from @code{perror} or
599 equivalent) in @emph{every} error message resulting from a failing
600 system call, as well as the name of the file if any and the name of the
601 utility. Just ``cannot open foo.c'' or ``stat failed'' is not
604 @cindex @code{malloc} return value
605 @cindex memory allocation failure
606 Check every call to @code{malloc} or @code{realloc} to see if it
607 returned zero. Check @code{realloc} even if you are making the block
608 smaller; in a system that rounds block sizes to a power of 2,
609 @code{realloc} may get a different block if you ask for less space.
611 In Unix, @code{realloc} can destroy the storage block if it returns
612 zero. GNU @code{realloc} does not have this bug: if it fails, the
613 original block is unchanged. Feel free to assume the bug is fixed. If
614 you wish to run your program on Unix, and wish to avoid lossage in this
615 case, you can use the GNU @code{malloc}.
617 You must expect @code{free} to alter the contents of the block that was
618 freed. Anything you want to fetch from the block, you must fetch before
621 If @code{malloc} fails in a noninteractive program, make that a fatal
622 error. In an interactive program (one that reads commands from the
623 user), it is better to abort the command and return to the command
624 reader loop. This allows the user to kill other processes to free up
625 virtual memory, and then try the command again.
627 @cindex command-line arguments, decoding
628 Use @code{getopt_long} to decode arguments, unless the argument syntax
629 makes this unreasonable.
631 When static storage is to be written in during program execution, use
632 explicit C code to initialize it. Reserve C initialized declarations
633 for data that will not be changed.
636 Try to avoid low-level interfaces to obscure Unix data structures (such
637 as file directories, utmp, or the layout of kernel memory), since these
638 are less likely to work compatibly. If you need to find all the files
639 in a directory, use @code{readdir} or some other high-level interface.
640 These are supported compatibly by GNU.
642 @cindex signal handling
643 The preferred signal handling facilities are the BSD variant of
644 @code{signal}, and the @sc{posix} @code{sigaction} function; the
645 alternative USG @code{signal} interface is an inferior design.
647 Nowadays, using the @sc{posix} signal functions may be the easiest way
648 to make a program portable. If you use @code{signal}, then on GNU/Linux
649 systems running GNU libc version 1, you should include
650 @file{bsd/signal.h} instead of @file{signal.h}, so as to get BSD
651 behavior. It is up to you whether to support systems where
652 @code{signal} has only the USG behavior, or give up on them.
654 @cindex impossible conditions
655 In error checks that detect ``impossible'' conditions, just abort.
656 There is usually no point in printing any message. These checks
657 indicate the existence of bugs. Whoever wants to fix the bugs will have
658 to read the source code and run a debugger. So explain the problem with
659 comments in the source. The relevant data will be in variables, which
660 are easy to examine with the debugger, so there is no point moving them
663 Do not use a count of errors as the exit status for a program.
664 @emph{That does not work}, because exit status values are limited to 8
665 bits (0 through 255). A single run of the program might have 256
666 errors; if you try to return 256 as the exit status, the parent process
667 will see 0 as the status, and it will appear that the program succeeded.
669 @cindex temporary files
670 @cindex @code{TMPDIR} environment variable
671 If you make temporary files, check the @code{TMPDIR} environment
672 variable; if that variable is defined, use the specified directory
673 instead of @file{/tmp}.
675 In addition, be aware that there is a possible security problem when
676 creating temporary files in world-writable directories. In C, you can
677 avoid this problem by creating temporary files in this manner:
680 fd = open(filename, O_WRONLY | O_CREAT | O_EXCL, 0600);
684 or by using the @code{mkstemps} function from libiberty.
686 In bash, use @code{set -C} to avoid this problem.
689 @section Library Behavior
692 Try to make library functions reentrant. If they need to do dynamic
693 storage allocation, at least try to avoid any nonreentrancy aside from
694 that of @code{malloc} itself.
696 Here are certain name conventions for libraries, to avoid name
699 Choose a name prefix for the library, more than two characters long.
700 All external function and variable names should start with this
701 prefix. In addition, there should only be one of these in any given
702 library member. This usually means putting each one in a separate
705 An exception can be made when two external symbols are always used
706 together, so that no reasonable program could use one without the
707 other; then they can both go in the same file.
709 External symbols that are not documented entry points for the user
710 should have names beginning with @samp{_}. The @samp{_} should be
711 followed by the chosen name prefix for the library, to prevent
712 collisions with other libraries. These can go in the same files with
713 user entry points if you like.
715 Static functions and variables can be used as you like and need not
716 fit any naming convention.
719 @section Formatting Error Messages
720 @cindex formatting error messages
721 @cindex error messages, formatting
723 Error messages from compilers should look like this:
726 @var{source-file-name}:@var{lineno}: @var{message}
730 If you want to mention the column number, use one of these formats:
733 @var{source-file-name}:@var{lineno}:@var{column}: @var{message}
734 @var{source-file-name}:@var{lineno}.@var{column}: @var{message}
739 Line numbers should start from 1 at the beginning of the file, and
740 column numbers should start from 1 at the beginning of the line. (Both
741 of these conventions are chosen for compatibility.) Calculate column
742 numbers assuming that space and all ASCII printing characters have
743 equal width, and assuming tab stops every 8 columns.
745 The error message can also give both the starting and ending positions
746 of the erroneous text. There are several formats so that you can
747 avoid redundant information such as a duplicate line number.
748 Here are the possible formats:
751 @var{source-file-name}:@var{lineno-1}.@var{column-1}-@var{lineno-2}.@var{column-2}: @var{message}
752 @var{source-file-name}:@var{lineno-1}.@var{column-1}-@var{column-2}: @var{message}
753 @var{source-file-name}:@var{lineno-1}-@var{lineno-2}: @var{message}
757 When an error is spread over several files, you can use this format:
760 @var{file-1}:@var{lineno-1}.@var{column-1}-@var{file-2}:@var{lineno-2}.@var{column-2}: @var{message}
763 Error messages from other noninteractive programs should look like this:
766 @var{program}:@var{source-file-name}:@var{lineno}: @var{message}
770 when there is an appropriate source file, or like this:
773 @var{program}: @var{message}
777 when there is no relevant source file.
779 If you want to mention the column number, use this format:
782 @var{program}:@var{source-file-name}:@var{lineno}:@var{column}: @var{message}
785 In an interactive program (one that is reading commands from a
786 terminal), it is better not to include the program name in an error
787 message. The place to indicate which program is running is in the
788 prompt or with the screen layout. (When the same program runs with
789 input from a source other than a terminal, it is not interactive and
790 would do best to print error messages using the noninteractive style.)
792 The string @var{message} should not begin with a capital letter when
793 it follows a program name and/or file name, because that isn't the
794 beginning of a sentence. (The sentence conceptually starts at the
795 beginning of the line.) Also, it should not end with a period.
797 Error messages from interactive programs, and other messages such as
798 usage messages, should start with a capital letter. But they should not
801 @node User Interfaces
802 @section Standards for Interfaces Generally
804 @cindex program name and its behavior
805 @cindex behavior, dependent on program's name
806 Please don't make the behavior of a utility depend on the name used
807 to invoke it. It is useful sometimes to make a link to a utility
808 with a different name, and that should not change what it does.
810 Instead, use a run time option or a compilation switch or both
811 to select among the alternate behaviors.
813 @cindex output device and program's behavior
814 Likewise, please don't make the behavior of the program depend on the
815 type of output device it is used with. Device independence is an
816 important principle of the system's design; do not compromise it merely
817 to save someone from typing an option now and then. (Variation in error
818 message syntax when using a terminal is ok, because that is a side issue
819 that people do not depend on.)
821 If you think one behavior is most useful when the output is to a
822 terminal, and another is most useful when the output is a file or a
823 pipe, then it is usually best to make the default behavior the one that
824 is useful with output to a terminal, and have an option for the other
827 Compatibility requires certain programs to depend on the type of output
828 device. It would be disastrous if @code{ls} or @code{sh} did not do so
829 in the way all users expect. In some of these cases, we supplement the
830 program with a preferred alternate version that does not depend on the
831 output device type. For example, we provide a @code{dir} program much
832 like @code{ls} except that its default output format is always
836 @node Graphical Interfaces
837 @section Standards for Graphical Interfaces
838 @cindex graphical user interface
841 When you write a program that provides a graphical user interface,
842 please make it work with X Windows and the GTK+ toolkit unless the
843 functionality specifically requires some alternative (for example,
844 ``displaying jpeg images while in console mode'').
846 In addition, please provide a command-line interface to control the
847 functionality. (In many cases, the graphical user interface can be a
848 separate program which invokes the command-line program.) This is
849 so that the same jobs can be done from scripts.
853 Please also consider providing a CORBA interface (for use from GNOME), a
854 library interface (for use from C), and perhaps a keyboard-driven
855 console interface (for use by users from console mode). Once you are
856 doing the work to provide the functionality and the graphical interface,
857 these won't be much extra work.
860 @node Command-Line Interfaces
861 @section Standards for Command Line Interfaces
862 @cindex command-line interface
865 It is a good idea to follow the @sc{posix} guidelines for the
866 command-line options of a program. The easiest way to do this is to use
867 @code{getopt} to parse them. Note that the GNU version of @code{getopt}
868 will normally permit options anywhere among the arguments unless the
869 special argument @samp{--} is used. This is not what @sc{posix}
870 specifies; it is a GNU extension.
872 @cindex long-named options
873 Please define long-named options that are equivalent to the
874 single-letter Unix-style options. We hope to make GNU more user
875 friendly this way. This is easy to do with the GNU function
878 One of the advantages of long-named options is that they can be
879 consistent from program to program. For example, users should be able
880 to expect the ``verbose'' option of any GNU program which has one, to be
881 spelled precisely @samp{--verbose}. To achieve this uniformity, look at
882 the table of common long-option names when you choose the option names
883 for your program (@pxref{Option Table}).
885 It is usually a good idea for file names given as ordinary arguments to
886 be input files only; any output files would be specified using options
887 (preferably @samp{-o} or @samp{--output}). Even if you allow an output
888 file name as an ordinary argument for compatibility, try to provide an
889 option as another way to specify it. This will lead to more consistency
890 among GNU utilities, and fewer idiosyncrasies for users to remember.
892 @cindex standard command-line options
893 @cindex options, standard command-line
894 @cindex CGI programs, standard options for
895 @cindex PATH_INFO, specifying standard options as
896 All programs should support two standard options: @samp{--version}
897 and @samp{--help}. CGI programs should accept these as command-line
898 options, and also if given as the @env{PATH_INFO}; for instance,
899 visiting @url{http://example.org/p.cgi/--help} in a browser should
900 output the same information as invoking @samp{p.cgi --help} from the
904 * --version:: The standard output for --version.
905 * --help:: The standard output for --help.
909 @subsection @option{--version}
911 @cindex @samp{--version} output
913 The standard @code{--version} option should direct the program to
914 print information about its name, version, origin and legal status,
915 all on standard output, and then exit successfully. Other options and
916 arguments should be ignored once this is seen, and the program should
917 not perform its normal function.
919 @cindex canonical name of a program
920 @cindex program's canonical name
921 The first line is meant to be easy for a program to parse; the version
922 number proper starts after the last space. In addition, it contains
923 the canonical name for this program, in this format:
930 The program's name should be a constant string; @emph{don't} compute it
931 from @code{argv[0]}. The idea is to state the standard or canonical
932 name for the program, not its file name. There are other ways to find
933 out the precise file name where a command is found in @code{PATH}.
935 If the program is a subsidiary part of a larger package, mention the
936 package name in parentheses, like this:
939 emacsserver (GNU Emacs) 19.30
943 If the package has a version number which is different from this
944 program's version number, you can mention the package version number
945 just before the close-parenthesis.
947 If you @emph{need} to mention the version numbers of libraries which
948 are distributed separately from the package which contains this program,
949 you can do so by printing an additional line of version info for each
950 library you want to mention. Use the same format for these lines as for
953 Please do not mention all of the libraries that the program uses ``just
954 for completeness''---that would produce a lot of unhelpful clutter.
955 Please mention library version numbers only if you find in practice that
956 they are very important to you in debugging.
958 The following line, after the version number line or lines, should be a
959 copyright notice. If more than one copyright notice is called for, put
960 each on a separate line.
962 Next should follow a line stating the license, preferably using one of
963 abbrevations below, and a brief statement that the program is free
964 software, and that users are free to copy and change it. Also mention
965 that there is no warranty, to the extent permitted by law. See
966 recommended wording below.
968 It is ok to finish the output with a list of the major authors of the
969 program, as a way of giving credit.
971 Here's an example of output that follows these rules:
975 Copyright (C) 2007 Free Software Foundation, Inc.
976 License GPLv3+: GNU GPL version 3 or later <http://gnu.org/licenses/gpl.html>
977 This is free software: you are free to change and redistribute it.
978 There is NO WARRANTY, to the extent permitted by law.
981 You should adapt this to your program, of course, filling in the proper
982 year, copyright holder, name of program, and the references to
983 distribution terms, and changing the rest of the wording as necessary.
985 This copyright notice only needs to mention the most recent year in
986 which changes were made---there's no need to list the years for previous
987 versions' changes. You don't have to mention the name of the program in
988 these notices, if that is inconvenient, since it appeared in the first
989 line. (The rules are different for copyright notices in source files;
990 @pxref{Copyright Notices,,,maintain,Information for GNU Maintainers}.)
992 Translations of the above lines must preserve the validity of the
993 copyright notices (@pxref{Internationalization}). If the translation's
994 character set supports it, the @samp{(C)} should be replaced with the
995 copyright symbol, as follows:
998 (the official copyright symbol, which is the letter C in a circle);
1004 Write the word ``Copyright'' exactly like that, in English. Do not
1005 translate it into another language. International treaties recognize
1006 the English word ``Copyright''; translations into other languages do not
1007 have legal significance.
1009 Finally, here is the table of our suggested license abbreviations.
1010 Any abbreviation can be followed by @samp{v@var{version}[+]}, meaning
1011 that particular version, or later versions with the @samp{+}, as shown
1014 In the case of exceptions for extra permissions with the GPL, we use
1015 @samp{/} for a separator; the version number can follow the license
1016 abbreviation as usual, as in the examples below.
1020 GNU General Public License, @url{http://www.gnu.org/@/licenses/@/gpl.html}.
1023 GNU Lesser General Public License, @url{http://www.gnu.org/@/licenses/@/lgpl.html}.
1026 GNU GPL with the exception for Guile; for example, GPLv3+/Guile means
1027 the GNU GPL version 3 or later, with the extra exception for Guile.
1030 GNU GPL with the exception for Ada.
1033 The Apache Software Foundation license,
1034 @url{http://www.apache.org/@/licenses}.
1037 The Artistic license used for Perl, @url{http://www.perlfoundation.org/@/legal}.
1040 The Expat license, @url{http://www.jclark.com/@/xml/@/copying.txt}.
1043 The Mozilla Public License, @url{http://www.mozilla.org/@/MPL/}.
1046 The original (4-clause) BSD license, incompatible with the GNU GPL
1047 @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#6}.
1050 The license used for PHP, @url{http://www.php.net/@/license/}.
1053 The non-license that is being in the public domain,
1054 @url{http://www.gnu.org/@/licenses/@/license-list.html#PublicDomain}.
1057 The license for Python, @url{http://www.python.org/@/2.0.1/@/license.html}.
1060 The revised (3-clause) BSD, compatible with the GNU GPL,
1061 @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#5}.
1064 The simple non-copyleft license used for most versions of the X Window
1065 system, @url{http://www.xfree86.org/@/3.3.6/@/COPYRIGHT2.html#3}.
1068 The license for Zlib, @url{http://www.gzip.org/@/zlib/@/zlib_license.html}.
1072 More information about these licenses and many more are on the GNU
1073 licensing web pages,
1074 @url{http://www.gnu.org/@/licenses/@/license-list.html}.
1078 @subsection @option{--help}
1080 @cindex @samp{--help} output
1082 The standard @code{--help} option should output brief documentation
1083 for how to invoke the program, on standard output, then exit
1084 successfully. Other options and arguments should be ignored once this
1085 is seen, and the program should not perform its normal function.
1087 @cindex address for bug reports
1089 Near the end of the @samp{--help} option's output there should be a line
1090 that says where to mail bug reports. It should have this format:
1093 Report bugs to @var{mailing-address}.
1098 @section Table of Long Options
1099 @cindex long option names
1100 @cindex table of long options
1102 Here is a table of long options used by GNU programs. It is surely
1103 incomplete, but we aim to list all the options that a new program might
1104 want to be compatible with. If you use names not already in the table,
1105 please send @email{bug-standards@@gnu.org} a list of them, with their
1106 meanings, so we can update the table.
1108 @c Please leave newlines between items in this table; it's much easier
1109 @c to update when it isn't completely squashed together and unreadable.
1110 @c When there is more than one short option for a long option name, put
1111 @c a semicolon between the lists of the programs that use them, not a
1112 @c period. --friedman
1116 @samp{-N} in @code{tar}.
1119 @samp{-a} in @code{du}, @code{ls}, @code{nm}, @code{stty}, @code{uname},
1120 and @code{unexpand}.
1123 @samp{-a} in @code{diff}.
1126 @samp{-A} in @code{ls}.
1129 @samp{-a} in @code{etags}, @code{tee}, @code{time};
1130 @samp{-r} in @code{tar}.
1133 @samp{-a} in @code{cp}.
1136 @samp{-n} in @code{shar}.
1139 @samp{-l} in @code{m4}.
1142 @samp{-a} in @code{diff}.
1145 @samp{-v} in @code{gawk}.
1154 @samp{-a} in @code{recode}.
1157 @samp{-a} in @code{wdiff}.
1159 @item auto-reference
1160 @samp{-A} in @code{ptx}.
1163 @samp{-n} in @code{wdiff}.
1166 For server programs, run in the background.
1168 @item backward-search
1169 @samp{-B} in @code{ctags}.
1172 @samp{-f} in @code{shar}.
1181 @samp{-b} in @code{tac}.
1184 @samp{-b} in @code{cpio} and @code{diff}.
1187 @samp{-b} in @code{shar}.
1190 Used in @code{cpio} and @code{tar}.
1193 @samp{-b} in @code{head} and @code{tail}.
1196 @samp{-b} in @code{ptx}.
1199 Used in various programs to make output shorter.
1202 @samp{-c} in @code{head}, @code{split}, and @code{tail}.
1205 @samp{-C} in @code{etags}.
1208 @samp{-A} in @code{tar}.
1211 Used in various programs to specify the directory to use.
1214 @samp{-c} in @code{chgrp} and @code{chown}.
1217 @samp{-F} in @code{ls}.
1220 @samp{-c} in @code{recode}.
1223 @samp{-c} in @code{su};
1227 @samp{-d} in @code{tar}.
1230 Used in @code{gawk}.
1233 @samp{-Z} in @code{tar} and @code{shar}.
1236 @samp{-A} in @code{tar}.
1239 @samp{-w} in @code{tar}.
1242 Used in @code{diff}.
1245 @samp{-W copyleft} in @code{gawk}.
1248 @samp{-C} in @code{ptx}, @code{recode}, and @code{wdiff};
1249 @samp{-W copyright} in @code{gawk}.
1255 @samp{-q} in @code{who}.
1258 @samp{-l} in @code{du}.
1261 Used in @code{tar} and @code{cpio}.
1264 @samp{-c} in @code{shar}.
1267 @samp{-x} in @code{ctags}.
1270 @samp{-d} in @code{touch}.
1273 @samp{-d} in Make and @code{m4};
1277 @samp{-D} in @code{m4}.
1280 @samp{-d} in Bison and @code{ctags}.
1283 @samp{-D} in @code{tar}.
1286 @samp{-L} in @code{chgrp}, @code{chown}, @code{cpio}, @code{du},
1287 @code{ls}, and @code{tar}.
1289 @item dereference-args
1290 @samp{-D} in @code{du}.
1293 Specify an I/O device (special file name).
1296 @samp{-d} in @code{recode}.
1298 @item dictionary-order
1299 @samp{-d} in @code{look}.
1302 @samp{-d} in @code{tar}.
1305 @samp{-n} in @code{csplit}.
1308 Specify the directory to use, in various programs. In @code{ls}, it
1309 means to show directories themselves rather than their contents. In
1310 @code{rm} and @code{ln}, it means to not treat links to directories
1314 @samp{-x} in @code{strip}.
1316 @item discard-locals
1317 @samp{-X} in @code{strip}.
1323 @samp{-e} in @code{diff}.
1325 @item elide-empty-files
1326 @samp{-z} in @code{csplit}.
1329 @samp{-x} in @code{wdiff}.
1332 @samp{-z} in @code{wdiff}.
1334 @item entire-new-file
1335 @samp{-N} in @code{diff}.
1337 @item environment-overrides
1341 @samp{-e} in @code{xargs}.
1347 Used in @code{makeinfo}.
1350 @samp{-o} in @code{m4}.
1353 @samp{-b} in @code{ls}.
1356 @samp{-X} in @code{tar}.
1362 @samp{-x} in @code{xargs}.
1365 @samp{-e} in @code{unshar}.
1368 @samp{-t} in @code{diff}.
1371 @samp{-e} in @code{sed}.
1374 @samp{-g} in @code{nm}.
1377 @samp{-i} in @code{cpio};
1378 @samp{-x} in @code{tar}.
1381 @samp{-f} in @code{finger}.
1384 @samp{-f} in @code{su}.
1386 @item fatal-warnings
1387 @samp{-E} in @code{m4}.
1390 @samp{-f} in @code{info}, @code{gawk}, Make, @code{mt}, and @code{tar};
1391 @samp{-n} in @code{sed};
1392 @samp{-r} in @code{touch}.
1394 @item field-separator
1395 @samp{-F} in @code{gawk}.
1401 @samp{-F} in @code{ls}.
1404 @samp{-T} in @code{tar}.
1407 Used in @code{makeinfo}.
1409 @item flag-truncation
1410 @samp{-F} in @code{ptx}.
1412 @item fixed-output-files
1416 @samp{-f} in @code{tail}.
1418 @item footnote-style
1419 Used in @code{makeinfo}.
1422 @samp{-f} in @code{cp}, @code{ln}, @code{mv}, and @code{rm}.
1425 @samp{-F} in @code{shar}.
1428 For server programs, run in the foreground;
1429 in other words, don't do anything special to run the server
1433 Used in @code{ls}, @code{time}, and @code{ptx}.
1436 @samp{-F} in @code{m4}.
1442 @samp{-g} in @code{ptx}.
1445 @samp{-x} in @code{tar}.
1448 @samp{-i} in @code{ul}.
1451 @samp{-g} in @code{recode}.
1454 @samp{-g} in @code{install}.
1457 @samp{-z} in @code{tar} and @code{shar}.
1460 @samp{-H} in @code{m4}.
1463 @samp{-h} in @code{objdump} and @code{recode}
1466 @samp{-H} in @code{who}.
1469 Used to ask for brief usage information.
1471 @item here-delimiter
1472 @samp{-d} in @code{shar}.
1474 @item hide-control-chars
1475 @samp{-q} in @code{ls}.
1478 In @code{makeinfo}, output HTML.
1481 @samp{-u} in @code{who}.
1484 @samp{-D} in @code{diff}.
1487 @samp{-I} in @code{ls};
1488 @samp{-x} in @code{recode}.
1490 @item ignore-all-space
1491 @samp{-w} in @code{diff}.
1493 @item ignore-backups
1494 @samp{-B} in @code{ls}.
1496 @item ignore-blank-lines
1497 @samp{-B} in @code{diff}.
1500 @samp{-f} in @code{look} and @code{ptx};
1501 @samp{-i} in @code{diff} and @code{wdiff}.
1507 @samp{-i} in @code{ptx}.
1509 @item ignore-indentation
1510 @samp{-I} in @code{etags}.
1512 @item ignore-init-file
1515 @item ignore-interrupts
1516 @samp{-i} in @code{tee}.
1518 @item ignore-matching-lines
1519 @samp{-I} in @code{diff}.
1521 @item ignore-space-change
1522 @samp{-b} in @code{diff}.
1525 @samp{-i} in @code{tar}.
1528 @samp{-i} in @code{etags};
1529 @samp{-I} in @code{m4}.
1535 @samp{-G} in @code{tar}.
1538 @samp{-i}, @samp{-l}, and @samp{-m} in Finger.
1541 In some programs, specify the name of the file to read as the user's
1545 @samp{-i} in @code{expand}.
1548 @samp{-T} in @code{diff}.
1551 @samp{-i} in @code{ls}.
1554 @samp{-i} in @code{cp}, @code{ln}, @code{mv}, @code{rm};
1555 @samp{-e} in @code{m4};
1556 @samp{-p} in @code{xargs};
1557 @samp{-w} in @code{tar}.
1560 @samp{-p} in @code{shar}.
1575 @samp{-k} in @code{csplit}.
1578 @samp{-k} in @code{du} and @code{ls}.
1581 @samp{-l} in @code{etags}.
1584 @samp{-l} in @code{wdiff}.
1586 @item level-for-gzip
1587 @samp{-g} in @code{shar}.
1590 @samp{-C} in @code{split}.
1593 Used in @code{split}, @code{head}, and @code{tail}.
1596 @samp{-l} in @code{cpio}.
1600 Used in @code{gawk}.
1603 @samp{-t} in @code{cpio};
1604 @samp{-l} in @code{recode}.
1607 @samp{-t} in @code{tar}.
1610 @samp{-N} in @code{ls}.
1619 Used in @code{uname}.
1622 @samp{-M} in @code{ptx}.
1625 @samp{-m} in @code{hello} and @code{uname}.
1627 @item make-directories
1628 @samp{-d} in @code{cpio}.
1637 @samp{-n} in @code{xargs}.
1640 @samp{-n} in @code{xargs}.
1643 @samp{-l} in @code{xargs}.
1649 @samp{-P} in @code{xargs}.
1652 @samp{-T} in @code{who}.
1655 @samp{-T} in @code{who}.
1658 @samp{-d} in @code{diff}.
1660 @item mixed-uuencode
1661 @samp{-M} in @code{shar}.
1664 @samp{-m} in @code{install}, @code{mkdir}, and @code{mkfifo}.
1666 @item modification-time
1667 @samp{-m} in @code{tar}.
1670 @samp{-M} in @code{tar}.
1676 @samp{-L} in @code{m4}.
1679 @samp{-a} in @code{shar}.
1684 @item no-builtin-rules
1687 @item no-character-count
1688 @samp{-w} in @code{shar}.
1690 @item no-check-existing
1691 @samp{-x} in @code{shar}.
1694 @samp{-3} in @code{wdiff}.
1697 @samp{-c} in @code{touch}.
1700 @samp{-D} in @code{etags}.
1703 @samp{-1} in @code{wdiff}.
1705 @item no-dereference
1706 @samp{-d} in @code{cp}.
1709 @samp{-2} in @code{wdiff}.
1718 @samp{-P} in @code{shar}.
1721 @samp{-e} in @code{gprof}.
1724 @samp{-R} in @code{etags}.
1727 @samp{-p} in @code{nm}.
1730 Don't print a startup splash screen.
1733 Used in @code{makeinfo}.
1736 @samp{-a} in @code{gprof}.
1739 @samp{-E} in @code{gprof}.
1742 @samp{-m} in @code{shar}.
1745 Used in @code{makeinfo}.
1748 Used in @code{emacsclient}.
1751 Used in various programs to inhibit warnings.
1754 @samp{-n} in @code{info}.
1757 @samp{-n} in @code{uname}.
1760 @samp{-f} in @code{cpio}.
1763 @samp{-n} in @code{objdump}.
1766 @samp{-0} in @code{xargs}.
1769 @samp{-n} in @code{cat}.
1771 @item number-nonblank
1772 @samp{-b} in @code{cat}.
1775 @samp{-n} in @code{nm}.
1777 @item numeric-uid-gid
1778 @samp{-n} in @code{cpio} and @code{ls}.
1784 @samp{-o} in @code{tar}.
1789 @item one-file-system
1790 @samp{-l} in @code{tar}, @code{cp}, and @code{du}.
1793 @samp{-o} in @code{ptx}.
1796 @samp{-f} in @code{gprof}.
1799 @samp{-F} in @code{gprof}.
1802 @samp{-o} in @code{getopt}, @code{fdlist}, @code{fdmount},
1803 @code{fdmountd}, and @code{fdumount}.
1806 In various programs, specify the output file name.
1809 @samp{-o} in @code{shar}.
1812 @samp{-o} in @code{rm}.
1815 @samp{-c} in @code{unshar}.
1818 @samp{-o} in @code{install}.
1821 @samp{-l} in @code{diff}.
1823 @item paragraph-indent
1824 Used in @code{makeinfo}.
1827 @samp{-p} in @code{mkdir} and @code{rmdir}.
1830 @samp{-p} in @code{ul}.
1833 @samp{-p} in @code{cpio}.
1836 @samp{-P} in @code{finger}.
1839 @samp{-c} in @code{cpio} and @code{tar}.
1842 Used in @code{gawk}.
1844 @item prefix-builtins
1845 @samp{-P} in @code{m4}.
1848 @samp{-f} in @code{csplit}.
1851 Used in @code{tar} and @code{cp}.
1853 @item preserve-environment
1854 @samp{-p} in @code{su}.
1856 @item preserve-modification-time
1857 @samp{-m} in @code{cpio}.
1859 @item preserve-order
1860 @samp{-s} in @code{tar}.
1862 @item preserve-permissions
1863 @samp{-p} in @code{tar}.
1866 @samp{-l} in @code{diff}.
1869 @samp{-L} in @code{cmp}.
1871 @item print-data-base
1874 @item print-directory
1877 @item print-file-name
1878 @samp{-o} in @code{nm}.
1881 @samp{-s} in @code{nm}.
1884 @samp{-p} in @code{wdiff}.
1887 @samp{-p} in @code{ed}.
1890 Specify an HTTP proxy.
1893 @samp{-X} in @code{shar}.
1899 Used in many programs to inhibit the usual output. Every
1900 program accepting @samp{--quiet} should accept @samp{--silent} as a
1904 @samp{-Q} in @code{shar}
1907 @samp{-Q} in @code{ls}.
1910 @samp{-n} in @code{diff}.
1913 Used in @code{gawk}.
1915 @item read-full-blocks
1916 @samp{-B} in @code{tar}.
1925 @samp{-R} in @code{tar}.
1928 Used in @code{chgrp}, @code{chown}, @code{cp}, @code{ls}, @code{diff},
1931 @item reference-limit
1932 Used in @code{makeinfo}.
1935 @samp{-r} in @code{ptx}.
1938 @samp{-r} in @code{tac} and @code{etags}.
1941 @samp{-r} in @code{uname}.
1944 @samp{-R} in @code{m4}.
1947 @samp{-r} in @code{objdump}.
1950 @samp{-r} in @code{cpio}.
1953 @samp{-i} in @code{xargs}.
1955 @item report-identical-files
1956 @samp{-s} in @code{diff}.
1958 @item reset-access-time
1959 @samp{-a} in @code{cpio}.
1962 @samp{-r} in @code{ls} and @code{nm}.
1965 @samp{-f} in @code{diff}.
1967 @item right-side-defs
1968 @samp{-R} in @code{ptx}.
1971 @samp{-s} in @code{tar}.
1973 @item same-permissions
1974 @samp{-p} in @code{tar}.
1977 @samp{-g} in @code{stty}.
1982 @item sentence-regexp
1983 @samp{-S} in @code{ptx}.
1986 @samp{-S} in @code{du}.
1989 @samp{-s} in @code{tac}.
1992 Used by @code{recode} to chose files or pipes for sequencing passes.
1995 @samp{-s} in @code{su}.
1998 @samp{-A} in @code{cat}.
2000 @item show-c-function
2001 @samp{-p} in @code{diff}.
2004 @samp{-E} in @code{cat}.
2006 @item show-function-line
2007 @samp{-F} in @code{diff}.
2010 @samp{-T} in @code{cat}.
2013 Used in many programs to inhibit the usual output.
2014 Every program accepting
2015 @samp{--silent} should accept @samp{--quiet} as a synonym.
2018 @samp{-s} in @code{ls}.
2021 Specify a file descriptor for a network server to use for its socket,
2022 instead of opening and binding a new socket. This provides a way to
2023 run, in a non-privileged process, a server that normally needs a
2024 reserved port number.
2030 @samp{-W source} in @code{gawk}.
2033 @samp{-S} in @code{tar}.
2035 @item speed-large-files
2036 @samp{-H} in @code{diff}.
2039 @samp{-E} in @code{unshar}.
2041 @item split-size-limit
2042 @samp{-L} in @code{shar}.
2045 @samp{-s} in @code{cat}.
2048 @samp{-w} in @code{wdiff}.
2051 @samp{-y} in @code{wdiff}.
2054 Used in @code{tar} and @code{diff} to specify which file within
2055 a directory to start processing with.
2058 @samp{-s} in @code{wdiff}.
2060 @item stdin-file-list
2061 @samp{-S} in @code{shar}.
2067 @samp{-s} in @code{recode}.
2070 @samp{-s} in @code{install}.
2073 @samp{-s} in @code{strip}.
2076 @samp{-S} in @code{strip}.
2079 @samp{-s} in @code{shar}.
2082 @samp{-S} in @code{cp}, @code{ln}, @code{mv}.
2085 @samp{-b} in @code{csplit}.
2088 @samp{-s} in @code{gprof}.
2091 @samp{-s} in @code{du}.
2094 @samp{-s} in @code{ln}.
2097 Used in GDB and @code{objdump}.
2100 @samp{-s} in @code{m4}.
2103 @samp{-s} in @code{uname}.
2106 @samp{-t} in @code{expand} and @code{unexpand}.
2109 @samp{-T} in @code{ls}.
2112 @samp{-T} in @code{tput} and @code{ul}.
2113 @samp{-t} in @code{wdiff}.
2116 @samp{-a} in @code{diff}.
2119 @samp{-T} in @code{shar}.
2122 Used in @code{ls} and @code{touch}.
2125 Specify how long to wait before giving up on some operation.
2128 @samp{-O} in @code{tar}.
2131 @samp{-c} in @code{du}.
2134 @samp{-t} in Make, @code{ranlib}, and @code{recode}.
2137 @samp{-t} in @code{m4}.
2140 @samp{-t} in @code{hello};
2141 @samp{-W traditional} in @code{gawk};
2142 @samp{-G} in @code{ed}, @code{m4}, and @code{ptx}.
2148 @samp{-t} in @code{ctags}.
2150 @item typedefs-and-c++
2151 @samp{-T} in @code{ctags}.
2154 @samp{-t} in @code{ptx}.
2157 @samp{-z} in @code{tar}.
2160 @samp{-u} in @code{cpio}.
2163 @samp{-U} in @code{m4}.
2165 @item undefined-only
2166 @samp{-u} in @code{nm}.
2169 @samp{-u} in @code{cp}, @code{ctags}, @code{mv}, @code{tar}.
2172 Used in @code{gawk}; same as @samp{--help}.
2175 @samp{-B} in @code{shar}.
2177 @item vanilla-operation
2178 @samp{-V} in @code{shar}.
2181 Print more information about progress. Many programs support this.
2184 @samp{-W} in @code{tar}.
2187 Print the version number.
2189 @item version-control
2190 @samp{-V} in @code{cp}, @code{ln}, @code{mv}.
2193 @samp{-v} in @code{ctags}.
2196 @samp{-V} in @code{tar}.
2201 @item whole-size-limit
2202 @samp{-l} in @code{shar}.
2205 @samp{-w} in @code{ls} and @code{ptx}.
2208 @samp{-W} in @code{ptx}.
2211 @samp{-T} in @code{who}.
2214 @samp{-z} in @code{gprof}.
2218 @section Memory Usage
2219 @cindex memory usage
2221 If a program typically uses just a few meg of memory, don't bother making any
2222 effort to reduce memory usage. For example, if it is impractical for
2223 other reasons to operate on files more than a few meg long, it is
2224 reasonable to read entire input files into memory to operate on them.
2226 However, for programs such as @code{cat} or @code{tail}, that can
2227 usefully operate on very large files, it is important to avoid using a
2228 technique that would artificially limit the size of files it can handle.
2229 If a program works by lines and could be applied to arbitrary
2230 user-supplied input files, it should keep only a line in memory, because
2231 this is not very hard and users will want to be able to operate on input
2232 files that are bigger than will fit in memory all at once.
2234 If your program creates complicated data structures, just make them in
2235 memory and give a fatal error if @code{malloc} returns zero.
2241 Programs should be prepared to operate when @file{/usr} and @file{/etc}
2242 are read-only file systems. Thus, if the program manages log files,
2243 lock files, backup files, score files, or any other files which are
2244 modified for internal purposes, these files should not be stored in
2245 @file{/usr} or @file{/etc}.
2247 There are two exceptions. @file{/etc} is used to store system
2248 configuration information; it is reasonable for a program to modify
2249 files in @file{/etc} when its job is to update the system configuration.
2250 Also, if the user explicitly asks to modify one file in a directory, it
2251 is reasonable for the program to store other files in the same
2255 @chapter Making The Best Use of C
2257 This chapter provides advice on how best to use the C language
2258 when writing GNU software.
2261 * Formatting:: Formatting your source code.
2262 * Comments:: Commenting your work.
2263 * Syntactic Conventions:: Clean use of C constructs.
2264 * Names:: Naming variables, functions, and files.
2265 * System Portability:: Portability among different operating systems.
2266 * CPU Portability:: Supporting the range of CPU types.
2267 * System Functions:: Portability and ``standard'' library functions.
2268 * Internationalization:: Techniques for internationalization.
2269 * Character Set:: Use ASCII by default.
2270 * Quote Characters:: Use `...' in the C locale.
2271 * Mmap:: How you can safely use @code{mmap}.
2275 @section Formatting Your Source Code
2276 @cindex formatting source code
2279 @cindex braces, in C source
2280 It is important to put the open-brace that starts the body of a C
2281 function in column one, so that they will start a defun. Several
2282 tools look for open-braces in column one to find the beginnings of C
2283 functions. These tools will not work on code not formatted that way.
2285 Avoid putting open-brace, open-parenthesis or open-bracket in column
2286 one when they are inside a function, so that they won't start a defun.
2287 The open-brace that starts a @code{struct} body can go in column one
2288 if you find it useful to treat that definition as a defun.
2290 It is also important for function definitions to start the name of the
2291 function in column one. This helps people to search for function
2292 definitions, and may also help certain tools recognize them. Thus,
2293 using Standard C syntax, the format is this:
2297 concat (char *s1, char *s2)
2304 or, if you want to use traditional C syntax, format the definition like
2309 concat (s1, s2) /* Name starts in column one here */
2311 @{ /* Open brace in column one here */
2316 In Standard C, if the arguments don't fit nicely on one line,
2321 lots_of_args (int an_integer, long a_long, short a_short,
2322 double a_double, float a_float)
2326 The rest of this section gives our recommendations for other aspects of
2327 C formatting style, which is also the default style of the @code{indent}
2328 program in version 1.2 and newer. It corresponds to the options
2331 -nbad -bap -nbc -bbo -bl -bli2 -bls -ncdb -nce -cp1 -cs -di2
2332 -ndj -nfc1 -nfca -hnl -i2 -ip5 -lp -pcs -psl -nsc -nsob
2335 We don't think of these recommendations as requirements, because it
2336 causes no problems for users if two different programs have different
2339 But whatever style you use, please use it consistently, since a mixture
2340 of styles within one program tends to look ugly. If you are
2341 contributing changes to an existing program, please follow the style of
2344 For the body of the function, our recommended style looks like this:
2356 return ++x + bar ();
2360 @cindex spaces before open-paren
2361 We find it easier to read a program when it has spaces before the
2362 open-parentheses and after the commas. Especially after the commas.
2364 When you split an expression into multiple lines, split it
2365 before an operator, not after one. Here is the right way:
2367 @cindex expressions, splitting
2369 if (foo_this_is_long && bar > win (x, y, z)
2370 && remaining_condition)
2373 Try to avoid having two operators of different precedence at the same
2374 level of indentation. For example, don't write this:
2377 mode = (inmode[j] == VOIDmode
2378 || GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])
2379 ? outmode[j] : inmode[j]);
2382 Instead, use extra parentheses so that the indentation shows the nesting:
2385 mode = ((inmode[j] == VOIDmode
2386 || (GET_MODE_SIZE (outmode[j]) > GET_MODE_SIZE (inmode[j])))
2387 ? outmode[j] : inmode[j]);
2390 Insert extra parentheses so that Emacs will indent the code properly.
2391 For example, the following indentation looks nice if you do it by hand,
2394 v = rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2395 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000;
2399 but Emacs would alter it. Adding a set of parentheses produces
2400 something that looks equally nice, and which Emacs will preserve:
2403 v = (rup->ru_utime.tv_sec*1000 + rup->ru_utime.tv_usec/1000
2404 + rup->ru_stime.tv_sec*1000 + rup->ru_stime.tv_usec/1000);
2407 Format do-while statements like this:
2419 Please use formfeed characters (control-L) to divide the program into
2420 pages at logical places (but not within a function). It does not matter
2421 just how long the pages are, since they do not have to fit on a printed
2422 page. The formfeeds should appear alone on lines by themselves.
2425 @section Commenting Your Work
2428 Every program should start with a comment saying briefly what it is for.
2429 Example: @samp{fmt - filter for simple filling of text}. This comment
2430 should be at the top of the source file containing the @samp{main}
2431 function of the program.
2433 Also, please write a brief comment at the start of each source file,
2434 with the file name and a line or two about the overall purpose of the
2437 Please write the comments in a GNU program in English, because English
2438 is the one language that nearly all programmers in all countries can
2439 read. If you do not write English well, please write comments in
2440 English as well as you can, then ask other people to help rewrite them.
2441 If you can't write comments in English, please find someone to work with
2442 you and translate your comments into English.
2444 Please put a comment on each function saying what the function does,
2445 what sorts of arguments it gets, and what the possible values of
2446 arguments mean and are used for. It is not necessary to duplicate in
2447 words the meaning of the C argument declarations, if a C type is being
2448 used in its customary fashion. If there is anything nonstandard about
2449 its use (such as an argument of type @code{char *} which is really the
2450 address of the second character of a string, not the first), or any
2451 possible values that would not work the way one would expect (such as,
2452 that strings containing newlines are not guaranteed to work), be sure
2455 Also explain the significance of the return value, if there is one.
2457 Please put two spaces after the end of a sentence in your comments, so
2458 that the Emacs sentence commands will work. Also, please write
2459 complete sentences and capitalize the first word. If a lower-case
2460 identifier comes at the beginning of a sentence, don't capitalize it!
2461 Changing the spelling makes it a different identifier. If you don't
2462 like starting a sentence with a lower case letter, write the sentence
2463 differently (e.g., ``The identifier lower-case is @dots{}'').
2465 The comment on a function is much clearer if you use the argument
2466 names to speak about the argument values. The variable name itself
2467 should be lower case, but write it in upper case when you are speaking
2468 about the value rather than the variable itself. Thus, ``the inode
2469 number NODE_NUM'' rather than ``an inode''.
2471 There is usually no purpose in restating the name of the function in
2472 the comment before it, because the reader can see that for himself.
2473 There might be an exception when the comment is so long that the function
2474 itself would be off the bottom of the screen.
2476 There should be a comment on each static variable as well, like this:
2479 /* Nonzero means truncate lines in the display;
2480 zero means continue them. */
2484 @cindex conditionals, comments for
2485 @cindex @code{#endif}, commenting
2486 Every @samp{#endif} should have a comment, except in the case of short
2487 conditionals (just a few lines) that are not nested. The comment should
2488 state the condition of the conditional that is ending, @emph{including
2489 its sense}. @samp{#else} should have a comment describing the condition
2490 @emph{and sense} of the code that follows. For example:
2498 #endif /* not foo */
2508 but, by contrast, write the comments this way for a @samp{#ifndef}:
2521 #endif /* not foo */
2525 @node Syntactic Conventions
2526 @section Clean Use of C Constructs
2527 @cindex syntactic conventions
2529 @cindex implicit @code{int}
2530 @cindex function argument, declaring
2531 Please explicitly declare the types of all objects. For example, you
2532 should explicitly declare all arguments to functions, and you should
2533 declare functions to return @code{int} rather than omitting the
2536 @cindex compiler warnings
2537 @cindex @samp{-Wall} compiler option
2538 Some programmers like to use the GCC @samp{-Wall} option, and change the
2539 code whenever it issues a warning. If you want to do this, then do.
2540 Other programmers prefer not to use @samp{-Wall}, because it gives
2541 warnings for valid and legitimate code which they do not want to change.
2542 If you want to do this, then do. The compiler should be your servant,
2545 Declarations of external functions and functions to appear later in the
2546 source file should all go in one place near the beginning of the file
2547 (somewhere before the first function definition in the file), or else
2548 should go in a header file. Don't put @code{extern} declarations inside
2551 @cindex temporary variables
2552 It used to be common practice to use the same local variables (with
2553 names like @code{tem}) over and over for different values within one
2554 function. Instead of doing this, it is better to declare a separate local
2555 variable for each distinct purpose, and give it a name which is
2556 meaningful. This not only makes programs easier to understand, it also
2557 facilitates optimization by good compilers. You can also move the
2558 declaration of each local variable into the smallest scope that includes
2559 all its uses. This makes the program even cleaner.
2561 Don't use local variables or parameters that shadow global identifiers.
2563 @cindex multiple variables in a line
2564 Don't declare multiple variables in one declaration that spans lines.
2565 Start a new declaration on each line, instead. For example, instead
2591 (If they are global variables, each should have a comment preceding it
2594 When you have an @code{if}-@code{else} statement nested in another
2595 @code{if} statement, always put braces around the @code{if}-@code{else}.
2596 Thus, never write like this:
2619 If you have an @code{if} statement nested inside of an @code{else}
2620 statement, either write @code{else if} on one line, like this,
2630 with its @code{then}-part indented like the preceding @code{then}-part,
2631 or write the nested @code{if} within braces like this:
2643 Don't declare both a structure tag and variables or typedefs in the
2644 same declaration. Instead, declare the structure tag separately
2645 and then use it to declare the variables or typedefs.
2647 Try to avoid assignments inside @code{if}-conditions (assignments
2648 inside @code{while}-conditions are ok). For example, don't write
2652 if ((foo = (char *) malloc (sizeof *foo)) == 0)
2653 fatal ("virtual memory exhausted");
2657 instead, write this:
2660 foo = (char *) malloc (sizeof *foo);
2662 fatal ("virtual memory exhausted");
2666 Don't make the program ugly to placate @code{lint}. Please don't insert any
2667 casts to @code{void}. Zero without a cast is perfectly fine as a null
2668 pointer constant, except when calling a varargs function.
2671 @section Naming Variables, Functions, and Files
2673 @cindex names of variables, functions, and files
2674 The names of global variables and functions in a program serve as
2675 comments of a sort. So don't choose terse names---instead, look for
2676 names that give useful information about the meaning of the variable or
2677 function. In a GNU program, names should be English, like other
2680 Local variable names can be shorter, because they are used only within
2681 one context, where (presumably) comments explain their purpose.
2683 Try to limit your use of abbreviations in symbol names. It is ok to
2684 make a few abbreviations, explain what they mean, and then use them
2685 frequently, but don't use lots of obscure abbreviations.
2687 Please use underscores to separate words in a name, so that the Emacs
2688 word commands can be useful within them. Stick to lower case; reserve
2689 upper case for macros and @code{enum} constants, and for name-prefixes
2690 that follow a uniform convention.
2692 For example, you should use names like @code{ignore_space_change_flag};
2693 don't use names like @code{iCantReadThis}.
2695 Variables that indicate whether command-line options have been
2696 specified should be named after the meaning of the option, not after
2697 the option-letter. A comment should state both the exact meaning of
2698 the option and its letter. For example,
2702 /* Ignore changes in horizontal whitespace (-b). */
2703 int ignore_space_change_flag;
2707 When you want to define names with constant integer values, use
2708 @code{enum} rather than @samp{#define}. GDB knows about enumeration
2711 @cindex file-name limitations
2713 You might want to make sure that none of the file names would conflict
2714 if the files were loaded onto an MS-DOS file system which shortens the
2715 names. You can use the program @code{doschk} to test for this.
2717 Some GNU programs were designed to limit themselves to file names of 14
2718 characters or less, to avoid file name conflicts if they are read into
2719 older System V systems. Please preserve this feature in the existing
2720 GNU programs that have it, but there is no need to do this in new GNU
2721 programs. @code{doschk} also reports file names longer than 14
2724 @node System Portability
2725 @section Portability between System Types
2726 @cindex portability, between system types
2728 In the Unix world, ``portability'' refers to porting to different Unix
2729 versions. For a GNU program, this kind of portability is desirable, but
2732 The primary purpose of GNU software is to run on top of the GNU kernel,
2733 compiled with the GNU C compiler, on various types of @sc{cpu}. So the
2734 kinds of portability that are absolutely necessary are quite limited.
2735 But it is important to support Linux-based GNU systems, since they
2736 are the form of GNU that is popular.
2738 Beyond that, it is good to support the other free operating systems
2739 (*BSD), and it is nice to support other Unix-like systems if you want
2740 to. Supporting a variety of Unix-like systems is desirable, although
2741 not paramount. It is usually not too hard, so you may as well do it.
2742 But you don't have to consider it an obligation, if it does turn out to
2746 The easiest way to achieve portability to most Unix-like systems is to
2747 use Autoconf. It's unlikely that your program needs to know more
2748 information about the host platform than Autoconf can provide, simply
2749 because most of the programs that need such knowledge have already been
2752 Avoid using the format of semi-internal data bases (e.g., directories)
2753 when there is a higher-level alternative (@code{readdir}).
2755 @cindex non-@sc{posix} systems, and portability
2756 As for systems that are not like Unix, such as MSDOS, Windows, VMS, MVS,
2757 and older Macintosh systems, supporting them is often a lot of work.
2758 When that is the case, it is better to spend your time adding features
2759 that will be useful on GNU and GNU/Linux, rather than on supporting
2760 other incompatible systems.
2762 If you do support Windows, please do not abbreviate it as ``win''. In
2763 hacker terminology, calling something a ``win'' is a form of praise.
2764 You're free to praise Microsoft Windows on your own if you want, but
2765 please don't do this in GNU packages. Instead of abbreviating
2766 ``Windows'' to ``win'', you can write it in full or abbreviate it to
2767 ``woe'' or ``w''. In GNU Emacs, for instance, we use @samp{w32} in
2768 file names of Windows-specific files, but the macro for Windows
2769 conditionals is called @code{WINDOWSNT}.
2771 It is a good idea to define the ``feature test macro''
2772 @code{_GNU_SOURCE} when compiling your C files. When you compile on GNU
2773 or GNU/Linux, this will enable the declarations of GNU library extension
2774 functions, and that will usually give you a compiler error message if
2775 you define the same function names in some other way in your program.
2776 (You don't have to actually @emph{use} these functions, if you prefer
2777 to make the program more portable to other systems.)
2779 But whether or not you use these GNU extensions, you should avoid
2780 using their names for any other meanings. Doing so would make it hard
2781 to move your code into other GNU programs.
2783 @node CPU Portability
2784 @section Portability between @sc{cpu}s
2786 @cindex data types, and portability
2787 @cindex portability, and data types
2788 Even GNU systems will differ because of differences among @sc{cpu}
2789 types---for example, difference in byte ordering and alignment
2790 requirements. It is absolutely essential to handle these differences.
2791 However, don't make any effort to cater to the possibility that an
2792 @code{int} will be less than 32 bits. We don't support 16-bit machines
2795 Similarly, don't make any effort to cater to the possibility that
2796 @code{long} will be smaller than predefined types like @code{size_t}.
2797 For example, the following code is ok:
2800 printf ("size = %lu\n", (unsigned long) sizeof array);
2801 printf ("diff = %ld\n", (long) (pointer2 - pointer1));
2804 1989 Standard C requires this to work, and we know of only one
2805 counterexample: 64-bit programs on Microsoft Windows. We will
2806 leave it to those who want to port GNU programs to that environment
2807 to figure out how to do it.
2809 Predefined file-size types like @code{off_t} are an exception: they are
2810 longer than @code{long} on many platforms, so code like the above won't
2811 work with them. One way to print an @code{off_t} value portably is to
2812 print its digits yourself, one by one.
2814 Don't assume that the address of an @code{int} object is also the
2815 address of its least-significant byte. This is false on big-endian
2816 machines. Thus, don't make the following mistake:
2821 while ((c = getchar ()) != EOF)
2822 write (file_descriptor, &c, 1);
2825 @noindent Instead, use @code{unsigned char} as follows. (The @code{unsigned}
2826 is for portability to unusual systems where @code{char} is signed and
2827 where there is integer overflow checking.)
2831 while ((c = getchar ()) != EOF)
2833 unsigned char u = c;
2834 write (file_descriptor, &u, 1);
2838 It used to be ok to not worry about the difference between pointers
2839 and integers when passing arguments to functions. However, on most
2840 modern 64-bit machines pointers are wider than @code{int}.
2841 Conversely, integer types like @code{long long int} and @code{off_t}
2842 are wider than pointers on most modern 32-bit machines. Hence it's
2843 often better nowadays to use prototypes to define functions whose
2844 argument types are not trivial.
2846 In particular, if functions accept varying argument counts or types
2847 they should be declared using prototypes containing @samp{...} and
2848 defined using @file{stdarg.h}. For an example of this, please see the
2849 @uref{http://www.gnu.org/software/gnulib/, Gnulib} error module, which
2850 declares and defines the following function:
2853 /* Print a message with `fprintf (stderr, FORMAT, ...)';
2854 if ERRNUM is nonzero, follow it with ": " and strerror (ERRNUM).
2855 If STATUS is nonzero, terminate the program with `exit (STATUS)'. */
2857 void error (int status, int errnum, const char *format, ...);
2860 A simple way to use the Gnulib error module is to obtain the two
2861 source files @file{error.c} and @file{error.h} from the Gnulib library
2862 source code repository at
2863 @uref{http://git.savannah.gnu.org/@/gitweb/@/?p=gnulib.git}.
2864 Here's a sample use:
2871 char *program_name = "myprogram";
2874 xfopen (char const *name)
2876 FILE *fp = fopen (name, "r");
2878 error (1, errno, "cannot read %s", name);
2883 @cindex casting pointers to integers
2884 Avoid casting pointers to integers if you can. Such casts greatly
2885 reduce portability, and in most programs they are easy to avoid. In the
2886 cases where casting pointers to integers is essential---such as, a Lisp
2887 interpreter which stores type information as well as an address in one
2888 word---you'll have to make explicit provisions to handle different word
2889 sizes. You will also need to make provision for systems in which the
2890 normal range of addresses you can get from @code{malloc} starts far away
2893 @node System Functions
2894 @section Calling System Functions
2895 @cindex library functions, and portability
2896 @cindex portability, and library functions
2898 C implementations differ substantially. Standard C reduces but does
2899 not eliminate the incompatibilities; meanwhile, many GNU packages still
2900 support pre-standard compilers because this is not hard to do. This
2901 chapter gives recommendations for how to use the more-or-less standard C
2902 library functions to avoid unnecessary loss of portability.
2906 Don't use the return value of @code{sprintf}. It returns the number of
2907 characters written on some systems, but not on all systems.
2910 Be aware that @code{vfprintf} is not always available.
2913 @code{main} should be declared to return type @code{int}. It should
2914 terminate either by calling @code{exit} or by returning the integer
2915 status code; make sure it cannot ever return an undefined value.
2917 @cindex declaration for system functions
2919 Don't declare system functions explicitly.
2921 Almost any declaration for a system function is wrong on some system.
2922 To minimize conflicts, leave it to the system header files to declare
2923 system functions. If the headers don't declare a function, let it
2926 While it may seem unclean to use a function without declaring it, in
2927 practice this works fine for most system library functions on the
2928 systems where this really happens; thus, the disadvantage is only
2929 theoretical. By contrast, actual declarations have frequently caused
2933 If you must declare a system function, don't specify the argument types.
2934 Use an old-style declaration, not a Standard C prototype. The more you
2935 specify about the function, the more likely a conflict.
2938 In particular, don't unconditionally declare @code{malloc} or
2941 Most GNU programs use those functions just once, in functions
2942 conventionally named @code{xmalloc} and @code{xrealloc}. These
2943 functions call @code{malloc} and @code{realloc}, respectively, and
2946 Because @code{xmalloc} and @code{xrealloc} are defined in your program,
2947 you can declare them in other files without any risk of type conflict.
2949 On most systems, @code{int} is the same length as a pointer; thus, the
2950 calls to @code{malloc} and @code{realloc} work fine. For the few
2951 exceptional systems (mostly 64-bit machines), you can use
2952 @strong{conditionalized} declarations of @code{malloc} and
2953 @code{realloc}---or put these declarations in configuration files
2954 specific to those systems.
2956 @cindex string library functions
2958 The string functions require special treatment. Some Unix systems have
2959 a header file @file{string.h}; others have @file{strings.h}. Neither
2960 file name is portable. There are two things you can do: use Autoconf to
2961 figure out which file to include, or don't include either file.
2964 If you don't include either strings file, you can't get declarations for
2965 the string functions from the header file in the usual way.
2967 That causes less of a problem than you might think. The newer standard
2968 string functions should be avoided anyway because many systems still
2969 don't support them. The string functions you can use are these:
2972 strcpy strncpy strcat strncat
2973 strlen strcmp strncmp
2977 The copy and concatenate functions work fine without a declaration as
2978 long as you don't use their values. Using their values without a
2979 declaration fails on systems where the width of a pointer differs from
2980 the width of @code{int}, and perhaps in other cases. It is trivial to
2981 avoid using their values, so do that.
2983 The compare functions and @code{strlen} work fine without a declaration
2984 on most systems, possibly all the ones that GNU software runs on.
2985 You may find it necessary to declare them @strong{conditionally} on a
2988 The search functions must be declared to return @code{char *}. Luckily,
2989 there is no variation in the data type they return. But there is
2990 variation in their names. Some systems give these functions the names
2991 @code{index} and @code{rindex}; other systems use the names
2992 @code{strchr} and @code{strrchr}. Some systems support both pairs of
2993 names, but neither pair works on all systems.
2995 You should pick a single pair of names and use it throughout your
2996 program. (Nowadays, it is better to choose @code{strchr} and
2997 @code{strrchr} for new programs, since those are the standard
2998 names.) Declare both of those names as functions returning @code{char
2999 *}. On systems which don't support those names, define them as macros
3000 in terms of the other pair. For example, here is what to put at the
3001 beginning of your file (or in a header) if you want to use the names
3002 @code{strchr} and @code{strrchr} throughout:
3006 #define strchr index
3008 #ifndef HAVE_STRRCHR
3009 #define strrchr rindex
3017 Here we assume that @code{HAVE_STRCHR} and @code{HAVE_STRRCHR} are
3018 macros defined in systems where the corresponding functions exist.
3019 One way to get them properly defined is to use Autoconf.
3021 @node Internationalization
3022 @section Internationalization
3023 @cindex internationalization
3026 GNU has a library called GNU gettext that makes it easy to translate the
3027 messages in a program into various languages. You should use this
3028 library in every program. Use English for the messages as they appear
3029 in the program, and let gettext provide the way to translate them into
3032 Using GNU gettext involves putting a call to the @code{gettext} macro
3033 around each string that might need translation---like this:
3036 printf (gettext ("Processing file `%s'..."));
3040 This permits GNU gettext to replace the string @code{"Processing file
3041 `%s'..."} with a translated version.
3043 Once a program uses gettext, please make a point of writing calls to
3044 @code{gettext} when you add new strings that call for translation.
3046 Using GNU gettext in a package involves specifying a @dfn{text domain
3047 name} for the package. The text domain name is used to separate the
3048 translations for this package from the translations for other packages.
3049 Normally, the text domain name should be the same as the name of the
3050 package---for example, @samp{coreutils} for the GNU core utilities.
3052 @cindex message text, and internationalization
3053 To enable gettext to work well, avoid writing code that makes
3054 assumptions about the structure of words or sentences. When you want
3055 the precise text of a sentence to vary depending on the data, use two or
3056 more alternative string constants each containing a complete sentences,
3057 rather than inserting conditionalized words or phrases into a single
3060 Here is an example of what not to do:
3063 printf ("%s is full", capacity > 5000000 ? "disk" : "floppy disk");
3066 If you apply gettext to all strings, like this,
3069 printf (gettext ("%s is full"),
3070 capacity > 5000000 ? gettext ("disk") : gettext ("floppy disk"));
3074 the translator will hardly know that "disk" and "floppy disk" are meant to
3075 be substituted in the other string. Worse, in some languages (like French)
3076 the construction will not work: the translation of the word "full" depends
3077 on the gender of the first part of the sentence; it happens to be not the
3078 same for "disk" as for "floppy disk".
3080 Complete sentences can be translated without problems:
3083 printf (capacity > 5000000 ? gettext ("disk is full")
3084 : gettext ("floppy disk is full"));
3087 A similar problem appears at the level of sentence structure with this
3091 printf ("# Implicit rule search has%s been done.\n",
3092 f->tried_implicit ? "" : " not");
3096 Adding @code{gettext} calls to this code cannot give correct results for
3097 all languages, because negation in some languages requires adding words
3098 at more than one place in the sentence. By contrast, adding
3099 @code{gettext} calls does the job straightforwardly if the code starts
3103 printf (f->tried_implicit
3104 ? "# Implicit rule search has been done.\n",
3105 : "# Implicit rule search has not been done.\n");
3108 Another example is this one:
3111 printf ("%d file%s processed", nfiles,
3112 nfiles != 1 ? "s" : "");
3116 The problem with this example is that it assumes that plurals are made
3117 by adding `s'. If you apply gettext to the format string, like this,
3120 printf (gettext ("%d file%s processed"), nfiles,
3121 nfiles != 1 ? "s" : "");
3125 the message can use different words, but it will still be forced to use
3126 `s' for the plural. Here is a better way, with gettext being applied to
3127 the two strings independently:
3130 printf ((nfiles != 1 ? gettext ("%d files processed")
3131 : gettext ("%d file processed")),
3136 But this still doesn't work for languages like Polish, which has three
3137 plural forms: one for nfiles == 1, one for nfiles == 2, 3, 4, 22, 23, 24, ...
3138 and one for the rest. The GNU @code{ngettext} function solves this problem:
3141 printf (ngettext ("%d files processed", "%d file processed", nfiles),
3147 @section Character Set
3148 @cindex character set
3150 @cindex ASCII characters
3151 @cindex non-ASCII characters
3153 Sticking to the ASCII character set (plain text, 7-bit characters) is
3154 preferred in GNU source code comments, text documents, and other
3155 contexts, unless there is good reason to do something else because of
3156 the application domain. For example, if source code deals with the
3157 French Revolutionary calendar, it is OK if its literal strings contain
3158 accented characters in month names like ``Flor@'eal''. Also, it is OK
3159 to use non-ASCII characters to represent proper names of contributors in
3160 change logs (@pxref{Change Logs}).
3162 If you need to use non-ASCII characters, you should normally stick with
3163 one encoding, as one cannot in general mix encodings reliably.
3166 @node Quote Characters
3167 @section Quote Characters
3168 @cindex quote characters
3169 @cindex locale-specific quote characters
3171 @cindex grave accent
3173 In the C locale, GNU programs should stick to plain ASCII for quotation
3174 characters in messages to users: preferably 0x60 (@samp{`}) for left
3175 quotes and 0x27 (@samp{'}) for right quotes. It is ok, but not
3176 required, to use locale-specific quotes in other locales.
3178 The @uref{http://www.gnu.org/software/gnulib/, Gnulib} @code{quote} and
3179 @code{quotearg} modules provide a reasonably straightforward way to
3180 support locale-specific quote characters, as well as taking care of
3181 other issues, such as quoting a filename that itself contains a quote
3182 character. See the Gnulib documentation for usage details.
3184 In any case, the documentation for your program should clearly specify
3185 how it does quoting, if different than the preferred method of @samp{`}
3186 and @samp{'}. This is especially important if the output of your
3187 program is ever likely to be parsed by another program.
3189 Quotation characters are a difficult area in the computing world at
3190 this time: there are no true left or right quote characters in Latin1;
3191 the @samp{`} character we use was standardized there as a grave
3192 accent. Moreover, Latin1 is still not universally usable.
3194 Unicode contains the unambiguous quote characters required, and its
3195 common encoding UTF-8 is upward compatible with Latin1. However,
3196 Unicode and UTF-8 are not universally well-supported, either.
3198 This may change over the next few years, and then we will revisit
3206 Don't assume that @code{mmap} either works on all files or fails
3207 for all files. It may work on some files and fail on others.
3209 The proper way to use @code{mmap} is to try it on the specific file for
3210 which you want to use it---and if @code{mmap} doesn't work, fall back on
3211 doing the job in another way using @code{read} and @code{write}.
3213 The reason this precaution is needed is that the GNU kernel (the HURD)
3214 provides a user-extensible file system, in which there can be many
3215 different kinds of ``ordinary files.'' Many of them support
3216 @code{mmap}, but some do not. It is important to make programs handle
3217 all these kinds of files.
3220 @chapter Documenting Programs
3221 @cindex documentation
3223 A GNU program should ideally come with full free documentation, adequate
3224 for both reference and tutorial purposes. If the package can be
3225 programmed or extended, the documentation should cover programming or
3226 extending it, as well as just using it.
3229 * GNU Manuals:: Writing proper manuals.
3230 * Doc Strings and Manuals:: Compiling doc strings doesn't make a manual.
3231 * Manual Structure Details:: Specific structure conventions.
3232 * License for Manuals:: Writing the distribution terms for a manual.
3233 * Manual Credits:: Giving credit to documentation contributors.
3234 * Printed Manuals:: Mentioning the printed manual.
3235 * NEWS File:: NEWS files supplement manuals.
3236 * Change Logs:: Recording changes.
3237 * Man Pages:: Man pages are secondary.
3238 * Reading other Manuals:: How far you can go in learning
3243 @section GNU Manuals
3245 The preferred document format for the GNU system is the Texinfo
3246 formatting language. Every GNU package should (ideally) have
3247 documentation in Texinfo both for reference and for learners. Texinfo
3248 makes it possible to produce a good quality formatted book, using
3249 @TeX{}, and to generate an Info file. It is also possible to generate
3250 HTML output from Texinfo source. See the Texinfo manual, either the
3251 hardcopy, or the on-line version available through @code{info} or the
3252 Emacs Info subsystem (@kbd{C-h i}).
3254 Nowadays some other formats such as Docbook and Sgmltexi can be
3255 converted automatically into Texinfo. It is ok to produce the Texinfo
3256 documentation by conversion this way, as long as it gives good results.
3258 Make sure your manual is clear to a reader who knows nothing about the
3259 topic and reads it straight through. This means covering basic topics
3260 at the beginning, and advanced topics only later. This also means
3261 defining every specialized term when it is first used.
3263 Programmers tend to carry over the structure of the program as the
3264 structure for its documentation. But this structure is not
3265 necessarily good for explaining how to use the program; it may be
3266 irrelevant and confusing for a user.
3268 Instead, the right way to structure documentation is according to the
3269 concepts and questions that a user will have in mind when reading it.
3270 This principle applies at every level, from the lowest (ordering
3271 sentences in a paragraph) to the highest (ordering of chapter topics
3272 within the manual). Sometimes this structure of ideas matches the
3273 structure of the implementation of the software being documented---but
3274 often they are different. An important part of learning to write good
3275 documentation is to learn to notice when you have unthinkingly
3276 structured the documentation like the implementation, stop yourself,
3277 and look for better alternatives.
3279 For example, each program in the GNU system probably ought to be
3280 documented in one manual; but this does not mean each program should
3281 have its own manual. That would be following the structure of the
3282 implementation, rather than the structure that helps the user
3285 Instead, each manual should cover a coherent @emph{topic}. For example,
3286 instead of a manual for @code{diff} and a manual for @code{diff3}, we
3287 have one manual for ``comparison of files'' which covers both of those
3288 programs, as well as @code{cmp}. By documenting these programs
3289 together, we can make the whole subject clearer.
3291 The manual which discusses a program should certainly document all of
3292 the program's command-line options and all of its commands. It should
3293 give examples of their use. But don't organize the manual as a list
3294 of features. Instead, organize it logically, by subtopics. Address
3295 the questions that a user will ask when thinking about the job that
3296 the program does. Don't just tell the reader what each feature can
3297 do---say what jobs it is good for, and show how to use it for those
3298 jobs. Explain what is recommended usage, and what kinds of usage
3301 In general, a GNU manual should serve both as tutorial and reference.
3302 It should be set up for convenient access to each topic through Info,
3303 and for reading straight through (appendixes aside). A GNU manual
3304 should give a good introduction to a beginner reading through from the
3305 start, and should also provide all the details that hackers want.
3306 The Bison manual is a good example of this---please take a look at it
3307 to see what we mean.
3309 That is not as hard as it first sounds. Arrange each chapter as a
3310 logical breakdown of its topic, but order the sections, and write their
3311 text, so that reading the chapter straight through makes sense. Do
3312 likewise when structuring the book into chapters, and when structuring a
3313 section into paragraphs. The watchword is, @emph{at each point, address
3314 the most fundamental and important issue raised by the preceding text.}
3316 If necessary, add extra chapters at the beginning of the manual which
3317 are purely tutorial and cover the basics of the subject. These provide
3318 the framework for a beginner to understand the rest of the manual. The
3319 Bison manual provides a good example of how to do this.
3321 To serve as a reference, a manual should have an Index that list all the
3322 functions, variables, options, and important concepts that are part of
3323 the program. One combined Index should do for a short manual, but
3324 sometimes for a complex package it is better to use multiple indices.
3325 The Texinfo manual includes advice on preparing good index entries, see
3326 @ref{Index Entries, , Making Index Entries, texinfo, GNU Texinfo}, and
3327 see @ref{Indexing Commands, , Defining the Entries of an
3328 Index, texinfo, GNU Texinfo}.
3330 Don't use Unix man pages as a model for how to write GNU documentation;
3331 most of them are terse, badly structured, and give inadequate
3332 explanation of the underlying concepts. (There are, of course, some
3333 exceptions.) Also, Unix man pages use a particular format which is
3334 different from what we use in GNU manuals.
3336 Please include an email address in the manual for where to report
3337 bugs @emph{in the text of the manual}.
3339 Please do not use the term ``pathname'' that is used in Unix
3340 documentation; use ``file name'' (two words) instead. We use the term
3341 ``path'' only for search paths, which are lists of directory names.
3343 Please do not use the term ``illegal'' to refer to erroneous input to
3344 a computer program. Please use ``invalid'' for this, and reserve the
3345 term ``illegal'' for activities prohibited by law.
3347 Please do not write @samp{()} after a function name just to indicate
3348 it is a function. @code{foo ()} is not a function, it is a function
3349 call with no arguments.
3351 @node Doc Strings and Manuals
3352 @section Doc Strings and Manuals
3354 Some programming systems, such as Emacs, provide a documentation string
3355 for each function, command or variable. You may be tempted to write a
3356 reference manual by compiling the documentation strings and writing a
3357 little additional text to go around them---but you must not do it. That
3358 approach is a fundamental mistake. The text of well-written
3359 documentation strings will be entirely wrong for a manual.
3361 A documentation string needs to stand alone---when it appears on the
3362 screen, there will be no other text to introduce or explain it.
3363 Meanwhile, it can be rather informal in style.
3365 The text describing a function or variable in a manual must not stand
3366 alone; it appears in the context of a section or subsection. Other text
3367 at the beginning of the section should explain some of the concepts, and
3368 should often make some general points that apply to several functions or
3369 variables. The previous descriptions of functions and variables in the
3370 section will also have given information about the topic. A description
3371 written to stand alone would repeat some of that information; this
3372 redundancy looks bad. Meanwhile, the informality that is acceptable in
3373 a documentation string is totally unacceptable in a manual.
3375 The only good way to use documentation strings in writing a good manual
3376 is to use them as a source of information for writing good text.
3378 @node Manual Structure Details
3379 @section Manual Structure Details
3380 @cindex manual structure
3382 The title page of the manual should state the version of the programs or
3383 packages documented in the manual. The Top node of the manual should
3384 also contain this information. If the manual is changing more
3385 frequently than or independent of the program, also state a version
3386 number for the manual in both of these places.
3388 Each program documented in the manual should have a node named
3389 @samp{@var{program} Invocation} or @samp{Invoking @var{program}}. This
3390 node (together with its subnodes, if any) should describe the program's
3391 command line arguments and how to run it (the sort of information people
3392 would look for in a man page). Start with an @samp{@@example}
3393 containing a template for all the options and arguments that the program
3396 Alternatively, put a menu item in some menu whose item name fits one of
3397 the above patterns. This identifies the node which that item points to
3398 as the node for this purpose, regardless of the node's actual name.
3400 The @samp{--usage} feature of the Info reader looks for such a node
3401 or menu item in order to find the relevant text, so it is essential
3402 for every Texinfo file to have one.
3404 If one manual describes several programs, it should have such a node for
3405 each program described in the manual.
3407 @node License for Manuals
3408 @section License for Manuals
3409 @cindex license for manuals
3411 Please use the GNU Free Documentation License for all GNU manuals that
3412 are more than a few pages long. Likewise for a collection of short
3413 documents---you only need one copy of the GNU FDL for the whole
3414 collection. For a single short document, you can use a very permissive
3415 non-copyleft license, to avoid taking up space with a long license.
3417 See @uref{http://www.gnu.org/copyleft/fdl-howto.html} for more explanation
3418 of how to employ the GFDL.
3420 Note that it is not obligatory to include a copy of the GNU GPL or GNU
3421 LGPL in a manual whose license is neither the GPL nor the LGPL. It can
3422 be a good idea to include the program's license in a large manual; in a
3423 short manual, whose size would be increased considerably by including
3424 the program's license, it is probably better not to include it.
3426 @node Manual Credits
3427 @section Manual Credits
3428 @cindex credits for manuals
3430 Please credit the principal human writers of the manual as the authors,
3431 on the title page of the manual. If a company sponsored the work, thank
3432 the company in a suitable place in the manual, but do not cite the
3433 company as an author.
3435 @node Printed Manuals
3436 @section Printed Manuals
3438 The FSF publishes some GNU manuals in printed form. To encourage sales
3439 of these manuals, the on-line versions of the manual should mention at
3440 the very start that the printed manual is available and should point at
3441 information for getting it---for instance, with a link to the page
3442 @url{http://www.gnu.org/order/order.html}. This should not be included
3443 in the printed manual, though, because there it is redundant.
3445 It is also useful to explain in the on-line forms of the manual how the
3446 user can print out the manual from the sources.
3449 @section The NEWS File
3450 @cindex @file{NEWS} file
3452 In addition to its manual, the package should have a file named
3453 @file{NEWS} which contains a list of user-visible changes worth
3454 mentioning. In each new release, add items to the front of the file and
3455 identify the version they pertain to. Don't discard old items; leave
3456 them in the file after the newer items. This way, a user upgrading from
3457 any previous version can see what is new.
3459 If the @file{NEWS} file gets very long, move some of the older items
3460 into a file named @file{ONEWS} and put a note at the end referring the
3464 @section Change Logs
3467 Keep a change log to describe all the changes made to program source
3468 files. The purpose of this is so that people investigating bugs in the
3469 future will know about the changes that might have introduced the bug.
3470 Often a new bug can be found by looking at what was recently changed.
3471 More importantly, change logs can help you eliminate conceptual
3472 inconsistencies between different parts of a program, by giving you a
3473 history of how the conflicting concepts arose and who they came from.
3476 * Change Log Concepts::
3477 * Style of Change Logs::
3479 * Conditional Changes::
3480 * Indicating the Part Changed::
3483 @node Change Log Concepts
3484 @subsection Change Log Concepts
3486 You can think of the change log as a conceptual ``undo list'' which
3487 explains how earlier versions were different from the current version.
3488 People can see the current version; they don't need the change log
3489 to tell them what is in it. What they want from a change log is a
3490 clear explanation of how the earlier version differed.
3492 The change log file is normally called @file{ChangeLog} and covers an
3493 entire directory. Each directory can have its own change log, or a
3494 directory can use the change log of its parent directory--it's up to
3497 Another alternative is to record change log information with a version
3498 control system such as RCS or CVS. This can be converted automatically
3499 to a @file{ChangeLog} file using @code{rcs2log}; in Emacs, the command
3500 @kbd{C-x v a} (@code{vc-update-change-log}) does the job.
3502 There's no need to describe the full purpose of the changes or how they
3503 work together. If you think that a change calls for explanation, you're
3504 probably right. Please do explain it---but please put the explanation
3505 in comments in the code, where people will see it whenever they see the
3506 code. For example, ``New function'' is enough for the change log when
3507 you add a function, because there should be a comment before the
3508 function definition to explain what it does.
3510 In the past, we recommended not mentioning changes in non-software
3511 files (manuals, help files, etc.) in change logs. However, we've been
3512 advised that it is a good idea to include them, for the sake of
3515 However, sometimes it is useful to write one line to describe the
3516 overall purpose of a batch of changes.
3518 The easiest way to add an entry to @file{ChangeLog} is with the Emacs
3519 command @kbd{M-x add-change-log-entry}. An entry should have an
3520 asterisk, the name of the changed file, and then in parentheses the name
3521 of the changed functions, variables or whatever, followed by a colon.
3522 Then describe the changes you made to that function or variable.
3524 @node Style of Change Logs
3525 @subsection Style of Change Logs
3526 @cindex change logs, style
3528 Here are some simple examples of change log entries, starting with the
3529 header line that says who made the change and when it was installed,
3530 followed by descriptions of specific changes. (These examples are
3531 drawn from Emacs and GCC.)
3534 1998-08-17 Richard Stallman <rms@@gnu.org>
3536 * register.el (insert-register): Return nil.
3537 (jump-to-register): Likewise.
3539 * sort.el (sort-subr): Return nil.
3541 * tex-mode.el (tex-bibtex-file, tex-file, tex-region):
3542 Restart the tex shell if process is gone or stopped.
3543 (tex-shell-running): New function.
3545 * expr.c (store_one_arg): Round size up for move_block_to_reg.
3546 (expand_call): Round up when emitting USE insns.
3547 * stmt.c (assign_parms): Round size up for move_block_from_reg.
3550 It's important to name the changed function or variable in full. Don't
3551 abbreviate function or variable names, and don't combine them.
3552 Subsequent maintainers will often search for a function name to find all
3553 the change log entries that pertain to it; if you abbreviate the name,
3554 they won't find it when they search.
3556 For example, some people are tempted to abbreviate groups of function
3557 names by writing @samp{* register.el (@{insert,jump-to@}-register)};
3558 this is not a good idea, since searching for @code{jump-to-register} or
3559 @code{insert-register} would not find that entry.
3561 Separate unrelated change log entries with blank lines. When two
3562 entries represent parts of the same change, so that they work together,
3563 then don't put blank lines between them. Then you can omit the file
3564 name and the asterisk when successive entries are in the same file.
3566 Break long lists of function names by closing continued lines with
3567 @samp{)}, rather than @samp{,}, and opening the continuation with
3568 @samp{(} as in this example:
3571 * keyboard.c (menu_bar_items, tool_bar_items)
3572 (Fexecute_extended_command): Deal with `keymap' property.
3575 When you install someone else's changes, put the contributor's name in
3576 the change log entry rather than in the text of the entry. In other
3580 2002-07-14 John Doe <jdoe@@gnu.org>
3582 * sewing.c: Make it sew.
3589 2002-07-14 Usual Maintainer <usual@@gnu.org>
3591 * sewing.c: Make it sew. Patch by jdoe@@gnu.org.
3594 As for the date, that should be the date you applied the change.
3596 @node Simple Changes
3597 @subsection Simple Changes
3599 Certain simple kinds of changes don't need much detail in the change
3602 When you change the calling sequence of a function in a simple fashion,
3603 and you change all the callers of the function to use the new calling
3604 sequence, there is no need to make individual entries for all the
3605 callers that you changed. Just write in the entry for the function
3606 being called, ``All callers changed''---like this:
3609 * keyboard.c (Fcommand_execute): New arg SPECIAL.
3610 All callers changed.
3613 When you change just comments or doc strings, it is enough to write an
3614 entry for the file, without mentioning the functions. Just ``Doc
3615 fixes'' is enough for the change log.
3617 There's no technical need to make change log entries for documentation
3618 files. This is because documentation is not susceptible to bugs that
3619 are hard to fix. Documentation does not consist of parts that must
3620 interact in a precisely engineered fashion. To correct an error, you
3621 need not know the history of the erroneous passage; it is enough to
3622 compare what the documentation says with the way the program actually
3625 However, you should keep change logs for documentation files when the
3626 project gets copyright assignments from its contributors, so as to
3627 make the records of authorship more accurate.
3629 @node Conditional Changes
3630 @subsection Conditional Changes
3631 @cindex conditional changes, and change logs
3632 @cindex change logs, conditional changes
3634 C programs often contain compile-time @code{#if} conditionals. Many
3635 changes are conditional; sometimes you add a new definition which is
3636 entirely contained in a conditional. It is very useful to indicate in
3637 the change log the conditions for which the change applies.
3639 Our convention for indicating conditional changes is to use square
3640 brackets around the name of the condition.
3642 Here is a simple example, describing a change which is conditional but
3643 does not have a function or entity name associated with it:
3646 * xterm.c [SOLARIS2]: Include string.h.
3649 Here is an entry describing a new definition which is entirely
3650 conditional. This new definition for the macro @code{FRAME_WINDOW_P} is
3651 used only when @code{HAVE_X_WINDOWS} is defined:
3654 * frame.h [HAVE_X_WINDOWS] (FRAME_WINDOW_P): Macro defined.
3657 Here is an entry for a change within the function @code{init_display},
3658 whose definition as a whole is unconditional, but the changes themselves
3659 are contained in a @samp{#ifdef HAVE_LIBNCURSES} conditional:
3662 * dispnew.c (init_display) [HAVE_LIBNCURSES]: If X, call tgetent.
3665 Here is an entry for a change that takes affect only when
3666 a certain macro is @emph{not} defined:
3669 (gethostname) [!HAVE_SOCKETS]: Replace with winsock version.
3672 @node Indicating the Part Changed
3673 @subsection Indicating the Part Changed
3675 Indicate the part of a function which changed by using angle brackets
3676 enclosing an indication of what the changed part does. Here is an entry
3677 for a change in the part of the function @code{sh-while-getopts} that
3678 deals with @code{sh} commands:
3681 * progmodes/sh-script.el (sh-while-getopts) <sh>: Handle case that
3682 user-specified option string is empty.
3690 In the GNU project, man pages are secondary. It is not necessary or
3691 expected for every GNU program to have a man page, but some of them do.
3692 It's your choice whether to include a man page in your program.
3694 When you make this decision, consider that supporting a man page
3695 requires continual effort each time the program is changed. The time
3696 you spend on the man page is time taken away from more useful work.
3698 For a simple program which changes little, updating the man page may be
3699 a small job. Then there is little reason not to include a man page, if
3702 For a large program that changes a great deal, updating a man page may
3703 be a substantial burden. If a user offers to donate a man page, you may
3704 find this gift costly to accept. It may be better to refuse the man
3705 page unless the same person agrees to take full responsibility for
3706 maintaining it---so that you can wash your hands of it entirely. If
3707 this volunteer later ceases to do the job, then don't feel obliged to
3708 pick it up yourself; it may be better to withdraw the man page from the
3709 distribution until someone else agrees to update it.
3711 When a program changes only a little, you may feel that the
3712 discrepancies are small enough that the man page remains useful without
3713 updating. If so, put a prominent note near the beginning of the man
3714 page explaining that you don't maintain it and that the Texinfo manual
3715 is more authoritative. The note should say how to access the Texinfo
3718 Be sure that man pages include a copyright statement and free
3719 license. The simple all-permissive license is appropriate for simple
3723 Copying and distribution of this file, with or without modification,
3724 are permitted in any medium without royalty provided the copyright
3725 notice and this notice are preserved.
3728 For long man pages, with enough explanation and documentation that
3729 they can be considered true manuals, use the GFDL (@pxref{License for
3732 Finally, the GNU help2man program
3733 (@uref{http://www.gnu.org/software/help2man/}) is one way to automate
3734 generation of a man page, in this case from @option{--help} output.
3735 This is sufficient in many cases.
3737 @node Reading other Manuals
3738 @section Reading other Manuals
3740 There may be non-free books or documentation files that describe the
3741 program you are documenting.
3743 It is ok to use these documents for reference, just as the author of a
3744 new algebra textbook can read other books on algebra. A large portion
3745 of any non-fiction book consists of facts, in this case facts about how
3746 a certain program works, and these facts are necessarily the same for
3747 everyone who writes about the subject. But be careful not to copy your
3748 outline structure, wording, tables or examples from preexisting non-free
3749 documentation. Copying from free documentation may be ok; please check
3750 with the FSF about the individual case.
3752 @node Managing Releases
3753 @chapter The Release Process
3756 Making a release is more than just bundling up your source files in a
3757 tar file and putting it up for FTP. You should set up your software so
3758 that it can be configured to run on a variety of systems. Your Makefile
3759 should conform to the GNU standards described below, and your directory
3760 layout should also conform to the standards discussed below. Doing so
3761 makes it easy to include your package into the larger framework of
3765 * Configuration:: How configuration of GNU packages should work.
3766 * Makefile Conventions:: Makefile conventions.
3767 * Releases:: Making releases
3771 @section How Configuration Should Work
3772 @cindex program configuration
3775 Each GNU distribution should come with a shell script named
3776 @code{configure}. This script is given arguments which describe the
3777 kind of machine and system you want to compile the program for.
3779 The @code{configure} script must record the configuration options so
3780 that they affect compilation.
3782 One way to do this is to make a link from a standard name such as
3783 @file{config.h} to the proper configuration file for the chosen system.
3784 If you use this technique, the distribution should @emph{not} contain a
3785 file named @file{config.h}. This is so that people won't be able to
3786 build the program without configuring it first.
3788 Another thing that @code{configure} can do is to edit the Makefile. If
3789 you do this, the distribution should @emph{not} contain a file named
3790 @file{Makefile}. Instead, it should include a file @file{Makefile.in} which
3791 contains the input used for editing. Once again, this is so that people
3792 won't be able to build the program without configuring it first.
3794 If @code{configure} does write the @file{Makefile}, then @file{Makefile}
3795 should have a target named @file{Makefile} which causes @code{configure}
3796 to be rerun, setting up the same configuration that was set up last
3797 time. The files that @code{configure} reads should be listed as
3798 dependencies of @file{Makefile}.
3800 All the files which are output from the @code{configure} script should
3801 have comments at the beginning explaining that they were generated
3802 automatically using @code{configure}. This is so that users won't think
3803 of trying to edit them by hand.
3805 The @code{configure} script should write a file named @file{config.status}
3806 which describes which configuration options were specified when the
3807 program was last configured. This file should be a shell script which,
3808 if run, will recreate the same configuration.
3810 The @code{configure} script should accept an option of the form
3811 @samp{--srcdir=@var{dirname}} to specify the directory where sources are found
3812 (if it is not the current directory). This makes it possible to build
3813 the program in a separate directory, so that the actual source directory
3816 If the user does not specify @samp{--srcdir}, then @code{configure} should
3817 check both @file{.} and @file{..} to see if it can find the sources. If
3818 it finds the sources in one of these places, it should use them from
3819 there. Otherwise, it should report that it cannot find the sources, and
3820 should exit with nonzero status.
3822 Usually the easy way to support @samp{--srcdir} is by editing a
3823 definition of @code{VPATH} into the Makefile. Some rules may need to
3824 refer explicitly to the specified source directory. To make this
3825 possible, @code{configure} can add to the Makefile a variable named
3826 @code{srcdir} whose value is precisely the specified directory.
3828 In addition, the @samp{configure} script should take options
3829 corresponding to most of the standard directory variables
3830 (@pxref{Directory Variables}). Here is the list:
3833 --prefix --exec-prefix --bindir --sbindir --libexecdir --sysconfdir
3834 --sharedstatedir --localstatedir --libdir --includedir --oldincludedir
3835 --datarootdir --datadir --infodir --localedir --mandir --docdir
3836 --htmldir --dvidir --pdfdir --psdir
3839 The @code{configure} script should also take an argument which specifies the
3840 type of system to build the program for. This argument should look like
3844 @var{cpu}-@var{company}-@var{system}
3847 For example, an Athlon-based GNU/Linux system might be
3848 @samp{i686-pc-linux-gnu}.
3850 The @code{configure} script needs to be able to decode all plausible
3851 alternatives for how to describe a machine. Thus,
3852 @samp{athlon-pc-gnu/linux} would be a valid alias. There is a shell
3854 @uref{http://git.savannah.gnu.org/@/gitweb/@/?p=config.git;a=blob_plain;f=config.sub;hb=HEAD,
3855 @file{config.sub}} that you can use as a subroutine to validate system
3856 types and canonicalize aliases.
3858 The @code{configure} script should also take the option
3859 @option{--build=@var{buildtype}}, which should be equivalent to a
3860 plain @var{buildtype} argument. For example, @samp{configure
3861 --build=i686-pc-linux-gnu} is equivalent to @samp{configure
3862 i686-pc-linux-gnu}. When the build type is not specified by an option
3863 or argument, the @code{configure} script should normally guess it using
3865 @uref{http://git.savannah.gnu.org/@/gitweb/@/?p=config.git;a=blob_plain;f=config.guess;hb=HEAD,
3866 @file{config.guess}}.
3868 @cindex optional features, configure-time
3869 Other options are permitted to specify in more detail the software
3870 or hardware present on the machine, to include or exclude optional parts
3871 of the package, or to adjust the name of some tools or arguments to them:
3874 @item --enable-@var{feature}@r{[}=@var{parameter}@r{]}
3875 Configure the package to build and install an optional user-level
3876 facility called @var{feature}. This allows users to choose which
3877 optional features to include. Giving an optional @var{parameter} of
3878 @samp{no} should omit @var{feature}, if it is built by default.
3880 No @samp{--enable} option should @strong{ever} cause one feature to
3881 replace another. No @samp{--enable} option should ever substitute one
3882 useful behavior for another useful behavior. The only proper use for
3883 @samp{--enable} is for questions of whether to build part of the program
3886 @item --with-@var{package}
3887 @c @r{[}=@var{parameter}@r{]}
3888 The package @var{package} will be installed, so configure this package
3889 to work with @var{package}.
3891 @c Giving an optional @var{parameter} of
3892 @c @samp{no} should omit @var{package}, if it is used by default.
3894 Possible values of @var{package} include
3895 @samp{gnu-as} (or @samp{gas}), @samp{gnu-ld}, @samp{gnu-libc},
3901 Do not use a @samp{--with} option to specify the file name to use to
3902 find certain files. That is outside the scope of what @samp{--with}
3905 @item @var{variable}=@var{value}
3906 Set the value of the variable @var{variable} to @var{value}. This is
3907 used to override the default values of commands or arguments in the
3908 build process. For example, the user could issue @samp{configure
3909 CFLAGS=-g CXXFLAGS=-g} to build with debugging information and without
3910 the default optimization.
3912 Specifying variables as arguments to @code{configure}, like this:
3916 is preferable to setting them in environment variables:
3920 as it helps to recreate the same configuration later with
3921 @file{config.status}.
3924 All @code{configure} scripts should accept all of the ``detail''
3925 options and the variable settings, whether or not they make any
3926 difference to the particular package at hand. In particular, they
3927 should accept any option that starts with @samp{--with-} or
3928 @samp{--enable-}. This is so users will be able to configure an
3929 entire GNU source tree at once with a single set of options.
3931 You will note that the categories @samp{--with-} and @samp{--enable-}
3932 are narrow: they @strong{do not} provide a place for any sort of option
3933 you might think of. That is deliberate. We want to limit the possible
3934 configuration options in GNU software. We do not want GNU programs to
3935 have idiosyncratic configuration options.
3937 Packages that perform part of the compilation process may support
3938 cross-compilation. In such a case, the host and target machines for the
3939 program may be different.
3941 The @code{configure} script should normally treat the specified type of
3942 system as both the host and the target, thus producing a program which
3943 works for the same type of machine that it runs on.
3945 To compile a program to run on a host type that differs from the build
3946 type, use the configure option @option{--host=@var{hosttype}}, where
3947 @var{hosttype} uses the same syntax as @var{buildtype}. The host type
3948 normally defaults to the build type.
3950 To configure a cross-compiler, cross-assembler, or what have you, you
3951 should specify a target different from the host, using the configure
3952 option @samp{--target=@var{targettype}}. The syntax for
3953 @var{targettype} is the same as for the host type. So the command would
3957 ./configure --host=@var{hosttype} --target=@var{targettype}
3960 The target type normally defaults to the host type.
3961 Programs for which cross-operation is not meaningful need not accept the
3962 @samp{--target} option, because configuring an entire operating system for
3963 cross-operation is not a meaningful operation.
3965 Some programs have ways of configuring themselves automatically. If
3966 your program is set up to do this, your @code{configure} script can simply
3967 ignore most of its arguments.
3969 @comment The makefile standards are in a separate file that is also
3970 @comment included by make.texinfo. Done by roland@gnu.ai.mit.edu on 1/6/93.
3971 @comment For this document, turn chapters into sections, etc.
3973 @include make-stds.texi
3977 @section Making Releases
3980 You should identify each release with a pair of version numbers, a
3981 major version and a minor. We have no objection to using more than
3982 two numbers, but it is very unlikely that you really need them.
3984 Package the distribution of @code{Foo version 69.96} up in a gzipped tar
3985 file with the name @file{foo-69.96.tar.gz}. It should unpack into a
3986 subdirectory named @file{foo-69.96}.
3988 Building and installing the program should never modify any of the files
3989 contained in the distribution. This means that all the files that form
3990 part of the program in any way must be classified into @dfn{source
3991 files} and @dfn{non-source files}. Source files are written by humans
3992 and never changed automatically; non-source files are produced from
3993 source files by programs under the control of the Makefile.
3995 @cindex @file{README} file
3996 The distribution should contain a file named @file{README} which gives
3997 the name of the package, and a general description of what it does. It
3998 is also good to explain the purpose of each of the first-level
3999 subdirectories in the package, if there are any. The @file{README} file
4000 should either state the version number of the package, or refer to where
4001 in the package it can be found.
4003 The @file{README} file should refer to the file @file{INSTALL}, which
4004 should contain an explanation of the installation procedure.
4006 The @file{README} file should also refer to the file which contains the
4007 copying conditions. The GNU GPL, if used, should be in a file called
4008 @file{COPYING}. If the GNU LGPL is used, it should be in a file called
4011 Naturally, all the source files must be in the distribution. It is okay
4012 to include non-source files in the distribution, provided they are
4013 up-to-date and machine-independent, so that building the distribution
4014 normally will never modify them. We commonly include non-source files
4015 produced by Bison, @code{lex}, @TeX{}, and @code{makeinfo}; this helps avoid
4016 unnecessary dependencies between our distributions, so that users can
4017 install whichever packages they want to install.
4019 Non-source files that might actually be modified by building and
4020 installing the program should @strong{never} be included in the
4021 distribution. So if you do distribute non-source files, always make
4022 sure they are up to date when you make a new distribution.
4024 Make sure that the directory into which the distribution unpacks (as
4025 well as any subdirectories) are all world-writable (octal mode 777).
4026 This is so that old versions of @code{tar} which preserve the
4027 ownership and permissions of the files from the tar archive will be
4028 able to extract all the files even if the user is unprivileged.
4030 Make sure that all the files in the distribution are world-readable.
4032 Don't include any symbolic links in the distribution itself. If the tar
4033 file contains symbolic links, then people cannot even unpack it on
4034 systems that don't support symbolic links. Also, don't use multiple
4035 names for one file in different directories, because certain file
4036 systems cannot handle this and that prevents unpacking the
4039 Try to make sure that all the file names will be unique on MS-DOS. A
4040 name on MS-DOS consists of up to 8 characters, optionally followed by a
4041 period and up to three characters. MS-DOS will truncate extra
4042 characters both before and after the period. Thus,
4043 @file{foobarhacker.c} and @file{foobarhacker.o} are not ambiguous; they
4044 are truncated to @file{foobarha.c} and @file{foobarha.o}, which are
4047 @cindex @file{texinfo.tex}, in a distribution
4048 Include in your distribution a copy of the @file{texinfo.tex} you used
4049 to test print any @file{*.texinfo} or @file{*.texi} files.
4051 Likewise, if your program uses small GNU software packages like regex,
4052 getopt, obstack, or termcap, include them in the distribution file.
4053 Leaving them out would make the distribution file a little smaller at
4054 the expense of possible inconvenience to a user who doesn't know what
4058 @chapter References to Non-Free Software and Documentation
4059 @cindex references to non-free material
4061 A GNU program should not recommend, promote, or grant legitimacy to
4062 the use of any non-free program. Proprietary software is a social and
4063 ethical problem, and our aim is to put an end to that problem. We
4064 can't stop some people from writing proprietary programs, or stop
4065 other people from using them, but we can and should refuse to
4066 advertise them to new potential customers, or to give the public the
4067 idea that their existence is ethical.
4069 The GNU definition of free software is found on the GNU web site at
4070 @url{http://www.gnu.org/@/philosophy/@/free-sw.html}, and the definition
4071 of free documentation is found at
4072 @url{http://www.gnu.org/@/philosophy/@/free-doc.html}. The terms ``free''
4073 and ``non-free'', used in this document, refer to those definitions.
4075 A list of important licenses and whether they qualify as free is in
4076 @url{http://www.gnu.org/@/licenses/@/license-list.html}. If it is not
4077 clear whether a license qualifies as free, please ask the GNU Project
4078 by writing to @email{licensing@@gnu.org}. We will answer, and if the
4079 license is an important one, we will add it to the list.
4081 When a non-free program or system is well known, you can mention it in
4082 passing---that is harmless, since users who might want to use it
4083 probably already know about it. For instance, it is fine to explain
4084 how to build your package on top of some widely used non-free
4085 operating system, or how to use it together with some widely used
4088 However, you should give only the necessary information to help those
4089 who already use the non-free program to use your program with
4090 it---don't give, or refer to, any further information about the
4091 proprietary program, and don't imply that the proprietary program
4092 enhances your program, or that its existence is in any way a good
4093 thing. The goal should be that people already using the proprietary
4094 program will get the advice they need about how to use your free
4095 program with it, while people who don't already use the proprietary
4096 program will not see anything likely to lead them to take an interest
4099 If a non-free program or system is obscure in your program's domain,
4100 your program should not mention or support it at all, since doing so
4101 would tend to popularize the non-free program more than it popularizes
4102 your program. (You cannot hope to find many additional users for your
4103 program among the users of Foobar, if the existence of Foobar is not
4104 generally known among people who might want to use your program.)
4106 Sometimes a program is free software in itself but depends on a
4107 non-free platform in order to run. For instance, many Java programs
4108 depend on some non-free Java libraries. To recommend or promote such
4109 a program is to promote the other programs it needs. This is why we
4110 are careful about listing Java programs in the Free Software
4111 Directory: we don't want to promote the non-free Java libraries.
4113 We hope this particular problem with Java will be gone by and by, as
4114 we replace the remaining non-free standard Java libraries with free
4115 software, but the general principle will remain the same: don't
4116 recommend, promote or legitimize programs that depend on non-free
4119 Some free programs strongly encourage the use of non-free software. A
4120 typical example is @command{mplayer}. It is free software in itself,
4121 and the free code can handle some kinds of files. However,
4122 @command{mplayer} recommends use of non-free codecs for other kinds of
4123 files, and users that install @command{mplayer} are very likely to
4124 install those codecs along with it. To recommend @command{mplayer}
4125 is, in effect, to promote use of the non-free codecs.
4127 Thus, you should not recommend programs that strongly encourage the
4128 use of non-free software. This is why we do not list
4129 @command{mplayer} in the Free Software Directory.
4131 A GNU package should not refer the user to any non-free documentation
4132 for free software. Free documentation that can be included in free
4133 operating systems is essential for completing the GNU system, or any
4134 free operating system, so encouraging it is a priority; to recommend
4135 use of documentation that we are not allowed to include undermines the
4136 impetus for the community to produce documentation that we can
4137 include. So GNU packages should never recommend non-free
4140 By contrast, it is ok to refer to journal articles and textbooks in
4141 the comments of a program for explanation of how it functions, even
4142 though they are non-free. This is because we don't include such
4143 things in the GNU system even they are free---they are outside the
4144 scope of what a software distribution needs to include.
4146 Referring to a web site that describes or recommends a non-free
4147 program is promoting that program, so please do not make links (or
4148 mention by name) web sites that contain such material. This policy is
4149 relevant particularly for the web pages for a GNU package.
4151 Following links from nearly any web site can lead eventually to
4152 non-free software; this is inherent in the nature of the web. So it
4153 makes no sense to criticize a site for having such links. As long as
4154 the site does not itself recommend a non-free program, there is no
4155 need to consider the question of the sites that it links to for other
4158 Thus, for example, you should not refer to AT&T's web site if that
4159 recommends AT&T's non-free software packages; you should not refer to
4160 a site that links to AT&T's site presenting it as a place to get some
4161 non-free program, because that link recommends and legitimizes the
4162 non-free program. However, that a site contains a link to AT&T's web
4163 site for some other purpose (such as long-distance telephone service)
4164 is not an objection against it.
4166 @node GNU Free Documentation License
4167 @appendix GNU Free Documentation License
4169 @cindex FDL, GNU Free Documentation License
4179 eval: (add-hook 'write-file-hooks 'time-stamp)
4180 time-stamp-start: "@set lastupdate "
4182 time-stamp-format: "%:b %:d, %:y"
4183 compile-command: "make just-standards"