1 \input texinfo @c -*- texinfo -*-
3 @settitle Developer Documentation
5 @center @titlefont{Developer Documentation}
12 @chapter Developers Guide
16 @item libavcodec is the library containing the codecs (both encoding and
17 decoding). Look at @file{doc/examples/decoding_encoding.c} to see how to use
20 @item libavformat is the library containing the file format handling (mux and
21 demux code for several formats). Look at @file{ffplay.c} to use it in a
22 player. See @file{doc/examples/muxing.c} to use it to generate audio or video
27 @section Integrating libavcodec or libavformat in your program
29 You can integrate all the source code of the libraries to link them
30 statically to avoid any version problem. All you need is to provide a
31 'config.mak' and a 'config.h' in the parent directory. See the defines
32 generated by ./configure to understand what is needed.
34 You can use libavcodec or libavformat in your commercial program, but
35 @emph{any patch you make must be published}. The best way to proceed is
36 to send your patches to the FFmpeg mailing list.
40 There are 3 ways by which code gets into ffmpeg.
42 @item Submitting Patches to the main developer mailing list
43 see @ref{Submitting patches} for details.
44 @item Directly committing changes to the main tree.
45 @item Committing changes to a git clone, for example on github.com or
46 gitorious.org. And asking us to merge these changes.
49 Whichever way, changes should be reviewed by the maintainer of the code
50 before they are committed. And they should follow the @ref{Coding Rules}.
51 The developer making the commit and the author are responsible for their changes
52 and should try to fix issues their commit causes.
57 @subsection Code formatting conventions
59 There are the following guidelines regarding the indentation in files:
64 The TAB character is forbidden outside of Makefiles as is any
65 form of trailing whitespace. Commits containing either will be
66 rejected by the git repository.
68 You should try to limit your code lines to 80 characters; however, do so if
69 and only if this improves readability.
71 The presentation is one inspired by 'indent -i4 -kr -nut'.
73 The main priority in FFmpeg is simplicity and small code size in order to
74 minimize the bug count.
77 Use the JavaDoc/Doxygen format (see examples below) so that code documentation
78 can be generated automatically. All nontrivial functions should have a comment
79 above them explaining what the function does, even if it is just one sentence.
80 All structures and their member variables should be documented, too.
82 Avoid Qt-style and similar Doxygen syntax with @code{!} in it, i.e. replace
83 @code{//!} with @code{///} and similar. Also @@ syntax should be employed
84 for markup commands, i.e. use @code{@@param} and not @code{\param}.
98 typedef struct Foobar@{
99 int var1; /**< var1 description */
100 int var2; ///< var2 description
101 /** var3 description */
109 * @@param my_parameter description of my_parameter
110 * @@return return value description
112 int myfunc(int my_parameter)
116 @subsection C language features
118 FFmpeg is programmed in the ISO C90 language with a few additional
119 features from ISO C99, namely:
122 the @samp{inline} keyword;
126 designated struct initializers (@samp{struct s x = @{ .i = 17 @};})
128 compound literals (@samp{x = (struct s) @{ 17, 23 @};})
131 These features are supported by all compilers we care about, so we will not
132 accept patches to remove their use unless they absolutely do not impair
133 clarity and performance.
135 All code must compile with recent versions of GCC and a number of other
136 currently supported compilers. To ensure compatibility, please do not use
137 additional C99 features or GCC extensions. Especially watch out for:
140 mixing statements and declarations;
142 @samp{long long} (use @samp{int64_t} instead);
144 @samp{__attribute__} not protected by @samp{#ifdef __GNUC__} or similar;
146 GCC statement expressions (@samp{(x = (@{ int y = 4; y; @})}).
149 @subsection Naming conventions
150 All names should be composed with underscores (_), not CamelCase. For example,
151 @samp{avfilter_get_video_buffer} is an acceptable function name and
152 @samp{AVFilterGetVideo} is not. The exception from this are type names, like
153 for example structs and enums; they should always be in the CamelCase
155 There are the following conventions for naming variables and functions:
158 For local variables no prefix is required.
160 For variables and functions declared as @code{static} no prefix is required.
162 For variables and functions used internally by a library an @code{ff_}
163 prefix should be used, e.g. @samp{ff_w64_demuxer}.
165 For variables and functions used internally across multiple libraries, use
166 @code{avpriv_}. For example, @samp{avpriv_aac_parse_header}.
168 Each library has its own prefix for public symbols, in addition to the
169 commonly used @code{av_} (@code{avformat_} for libavformat,
170 @code{avcodec_} for libavcodec, @code{swr_} for libswresample, etc).
171 Check the existing code and choose names accordingly.
172 Note that some symbols without these prefixes are also exported for
173 retro-compatibility reasons. These exceptions are declared in the
174 @code{lib<name>/lib<name>.v} files.
177 Furthermore, name space reserved for the system should not be invaded.
178 Identifiers ending in @code{_t} are reserved by
179 @url{http://pubs.opengroup.org/onlinepubs/007904975/functions/xsh_chap02_02.html#tag_02_02_02, POSIX}.
180 Also avoid names starting with @code{__} or @code{_} followed by an uppercase
181 letter as they are reserved by the C standard. Names starting with @code{_}
182 are reserved at the file level and may not be used for externally visible
183 symbols. If in doubt, just avoid names starting with @code{_} altogether.
185 @subsection Miscellaneous conventions
188 fprintf and printf are forbidden in libavformat and libavcodec,
189 please use av_log() instead.
191 Casts should be used only when necessary. Unneeded parentheses
192 should also be avoided if they don't make the code easier to understand.
195 @subsection Editor configuration
196 In order to configure Vim to follow FFmpeg formatting conventions, paste
197 the following snippet into your @file{.vimrc}:
199 " indentation rules for FFmpeg: 4 spaces, no tabs
205 " Allow tabs in Makefiles.
206 autocmd FileType make set noexpandtab shiftwidth=8 softtabstop=8
207 " Trailing whitespace and tabs are forbidden, so highlight them.
208 highlight ForbiddenWhitespace ctermbg=red guibg=red
209 match ForbiddenWhitespace /\s\+$\|\t/
210 " Do not highlight spaces at the end of line while typing on that line.
211 autocmd InsertEnter * match ForbiddenWhitespace /\t\|\s\+\%#\@@<!$/
214 For Emacs, add these roughly equivalent lines to your @file{.emacs.d/init.el}:
216 (c-add-style "ffmpeg"
219 (indent-tabs-mode . nil)
220 (show-trailing-whitespace . t)
222 (statement-cont . (c-lineup-assignments +)))
225 (setq c-default-style "ffmpeg")
228 @section Development Policy
232 Contributions should be licensed under the
233 @uref{http://www.gnu.org/licenses/lgpl-2.1.html, LGPL 2.1},
234 including an "or any later version" clause, or, if you prefer
235 a gift-style license, the
236 @uref{http://www.isc.org/software/license/, ISC} or
237 @uref{http://mit-license.org/, MIT} license.
238 @uref{http://www.gnu.org/licenses/gpl-2.0.html, GPL 2} including
239 an "or any later version" clause is also acceptable, but LGPL is
242 You must not commit code which breaks FFmpeg! (Meaning unfinished but
243 enabled code which breaks compilation or compiles but does not work or
244 breaks the regression tests)
245 You can commit unfinished stuff (for testing etc), but it must be disabled
246 (#ifdef etc) by default so it does not interfere with other developers'
249 The commit message should have a short first line in the form of
250 a @samp{topic: short description} as a header, separated by a newline
251 from the body consisting of an explanation of why the change is necessary.
252 If the commit fixes a known bug on the bug tracker, the commit message
253 should include its bug ID. Referring to the issue on the bug tracker does
254 not exempt you from writing an excerpt of the bug in the commit message.
256 You do not have to over-test things. If it works for you, and you think it
257 should work for others, then commit. If your code has problems
258 (portability, triggers compiler bugs, unusual environment etc) they will be
259 reported and eventually fixed.
261 Do not commit unrelated changes together, split them into self-contained
262 pieces. Also do not forget that if part B depends on part A, but A does not
263 depend on B, then A can and should be committed first and separate from B.
264 Keeping changes well split into self-contained parts makes reviewing and
265 understanding them on the commit log mailing list easier. This also helps
266 in case of debugging later on.
267 Also if you have doubts about splitting or not splitting, do not hesitate to
268 ask/discuss it on the developer mailing list.
270 Do not change behavior of the programs (renaming options etc) or public
271 API or ABI without first discussing it on the ffmpeg-devel mailing list.
272 Do not remove functionality from the code. Just improve!
274 Note: Redundant code can be removed.
276 Do not commit changes to the build system (Makefiles, configure script)
277 which change behavior, defaults etc, without asking first. The same
278 applies to compiler warning fixes, trivial looking fixes and to code
279 maintained by other developers. We usually have a reason for doing things
280 the way we do. Send your changes as patches to the ffmpeg-devel mailing
281 list, and if the code maintainers say OK, you may commit. This does not
282 apply to files you wrote and/or maintain.
284 We refuse source indentation and other cosmetic changes if they are mixed
285 with functional changes, such commits will be rejected and removed. Every
286 developer has his own indentation style, you should not change it. Of course
287 if you (re)write something, you can use your own style, even though we would
288 prefer if the indentation throughout FFmpeg was consistent (Many projects
289 force a given indentation style - we do not.). If you really need to make
290 indentation changes (try to avoid this), separate them strictly from real
293 NOTE: If you had to put if()@{ .. @} over a large (> 5 lines) chunk of code,
294 then either do NOT change the indentation of the inner part within (do not
295 move it to the right)! or do so in a separate commit
297 Always fill out the commit log message. Describe in a few lines what you
298 changed and why. You can refer to mailing list postings if you fix a
299 particular bug. Comments such as "fixed!" or "Changed it." are unacceptable.
301 area changed: Short 1 line description
303 details describing what and why and giving references.
305 Make sure the author of the commit is set correctly. (see git commit --author)
306 If you apply a patch, send an
307 answer to ffmpeg-devel (or wherever you got the patch from) saying that
308 you applied the patch.
310 When applying patches that have been discussed (at length) on the mailing
311 list, reference the thread in the log message.
313 Do NOT commit to code actively maintained by others without permission.
314 Send a patch to ffmpeg-devel instead. If no one answers within a reasonable
315 timeframe (12h for build failures and security fixes, 3 days small changes,
316 1 week for big patches) then commit your patch if you think it is OK.
317 Also note, the maintainer can simply ask for more time to review!
319 Subscribe to the ffmpeg-cvslog mailing list. The diffs of all commits
320 are sent there and reviewed by all the other developers. Bugs and possible
321 improvements or general questions regarding commits are discussed there. We
322 expect you to react if problems with your code are uncovered.
324 Update the documentation if you change behavior or add features. If you are
325 unsure how best to do this, send a patch to ffmpeg-devel, the documentation
326 maintainer(s) will review and commit your stuff.
328 Try to keep important discussions and requests (also) on the public
329 developer mailing list, so that all developers can benefit from them.
331 Never write to unallocated memory, never write over the end of arrays,
332 always check values read from some untrusted source before using them
333 as array index or other risky things.
335 Remember to check if you need to bump versions for the specific libav*
336 parts (libavutil, libavcodec, libavformat) you are changing. You need
337 to change the version integer.
338 Incrementing the first component means no backward compatibility to
339 previous versions (e.g. removal of a function from the public API).
340 Incrementing the second component means backward compatible change
341 (e.g. addition of a function to the public API or extension of an
342 existing data structure).
343 Incrementing the third component means a noteworthy binary compatible
344 change (e.g. encoder bug fix that matters for the decoder). The third
345 component always starts at 100 to distinguish FFmpeg from Libav.
347 Compiler warnings indicate potential bugs or code with bad style. If a type of
348 warning always points to correct and clean code, that warning should
349 be disabled, not the code changed.
350 Thus the remaining warnings can either be bugs or correct code.
351 If it is a bug, the bug has to be fixed. If it is not, the code should
352 be changed to not generate a warning unless that causes a slowdown
353 or obfuscates the code.
355 If you add a new file, give it a proper license header. Do not copy and
356 paste it from a random place, use an existing file as template.
359 We think our rules are not too hard. If you have comments, contact us.
361 @anchor{Submitting patches}
362 @section Submitting patches
364 First, read the @ref{Coding Rules} above if you did not yet, in particular
365 the rules regarding patch submission.
367 When you submit your patch, please use @code{git format-patch} or
368 @code{git send-email}. We cannot read other diffs :-)
370 Also please do not submit a patch which contains several unrelated changes.
371 Split it into separate, self-contained pieces. This does not mean splitting
372 file by file. Instead, make the patch as small as possible while still
373 keeping it as a logical unit that contains an individual change, even
374 if it spans multiple files. This makes reviewing your patches much easier
375 for us and greatly increases your chances of getting your patch applied.
377 Use the patcheck tool of FFmpeg to check your patch.
378 The tool is located in the tools directory.
380 Run the @ref{Regression tests} before submitting a patch in order to verify
381 it does not cause unexpected problems.
383 It also helps quite a bit if you tell us what the patch does (for example
384 'replaces lrint by lrintf'), and why (for example '*BSD isn't C99 compliant
387 Also please if you send several patches, send each patch as a separate mail,
388 do not attach several unrelated patches to the same mail.
390 Patches should be posted to the
391 @uref{http://lists.ffmpeg.org/mailman/listinfo/ffmpeg-devel, ffmpeg-devel}
392 mailing list. Use @code{git send-email} when possible since it will properly
393 send patches without requiring extra care. If you cannot, then send patches
394 as base64-encoded attachments, so your patch is not trashed during
397 Your patch will be reviewed on the mailing list. You will likely be asked
398 to make some changes and are expected to send in an improved version that
399 incorporates the requests from the review. This process may go through
400 several iterations. Once your patch is deemed good enough, some developer
401 will pick it up and commit it to the official FFmpeg tree.
403 Give us a few days to react. But if some time passes without reaction,
404 send a reminder by email. Your patch should eventually be dealt with.
407 @section New codecs or formats checklist
411 Did you use av_cold for codec initialization and close functions?
413 Did you add a long_name under NULL_IF_CONFIG_SMALL to the AVCodec or
414 AVInputFormat/AVOutputFormat struct?
416 Did you bump the minor version number (and reset the micro version
417 number) in @file{libavcodec/version.h} or @file{libavformat/version.h}?
419 Did you register it in @file{allcodecs.c} or @file{allformats.c}?
421 Did you add the AVCodecID to @file{avcodec.h}?
422 When adding new codec IDs, also add an entry to the codec descriptor
423 list in @file{libavcodec/codec_desc.c}.
425 If it has a FourCC, did you add it to @file{libavformat/riff.c},
426 even if it is only a decoder?
428 Did you add a rule to compile the appropriate files in the Makefile?
429 Remember to do this even if you're just adding a format to a file that is
430 already being compiled by some other rule, like a raw demuxer.
432 Did you add an entry to the table of supported formats or codecs in
433 @file{doc/general.texi}?
435 Did you add an entry in the Changelog?
437 If it depends on a parser or a library, did you add that dependency in
440 Did you @code{git add} the appropriate files before committing?
442 Did you make sure it compiles standalone, i.e. with
443 @code{configure --disable-everything --enable-decoder=foo}
444 (or @code{--enable-demuxer} or whatever your component is)?
448 @section patch submission checklist
452 Does @code{make fate} pass with the patch applied?
454 Was the patch generated with git format-patch or send-email?
456 Did you sign off your patch? (git commit -s)
457 See @url{http://git.kernel.org/?p=linux/kernel/git/torvalds/linux.git;a=blob_plain;f=Documentation/SubmittingPatches} for the meaning
460 Did you provide a clear git commit log message?
462 Is the patch against latest FFmpeg git master branch?
464 Are you subscribed to ffmpeg-devel?
465 (the list is subscribers only due to spam)
467 Have you checked that the changes are minimal, so that the same cannot be
468 achieved with a smaller patch and/or simpler final code?
470 If the change is to speed critical code, did you benchmark it?
472 If you did any benchmarks, did you provide them in the mail?
474 Have you checked that the patch does not introduce buffer overflows or
475 other security issues?
477 Did you test your decoder or demuxer against damaged data? If no, see
478 tools/trasher, the noise bitstream filter, and
479 @uref{http://caca.zoy.org/wiki/zzuf, zzuf}. Your decoder or demuxer
480 should not crash, end in a (near) infinite loop, or allocate ridiculous
481 amounts of memory when fed damaged data.
483 Does the patch not mix functional and cosmetic changes?
485 Did you add tabs or trailing whitespace to the code? Both are forbidden.
487 Is the patch attached to the email you send?
489 Is the mime type of the patch correct? It should be text/x-diff or
490 text/x-patch or at least text/plain and not application/octet-stream.
492 If the patch fixes a bug, did you provide a verbose analysis of the bug?
494 If the patch fixes a bug, did you provide enough information, including
495 a sample, so the bug can be reproduced and the fix can be verified?
496 Note please do not attach samples >100k to mails but rather provide a
497 URL, you can upload to ftp://upload.ffmpeg.org
499 Did you provide a verbose summary about what the patch does change?
501 Did you provide a verbose explanation why it changes things like it does?
503 Did you provide a verbose summary of the user visible advantages and
504 disadvantages if the patch is applied?
506 Did you provide an example so we can verify the new feature added by the
509 If you added a new file, did you insert a license header? It should be
510 taken from FFmpeg, not randomly copied and pasted from somewhere else.
512 You should maintain alphabetical order in alphabetically ordered lists as
513 long as doing so does not break API/ABI compatibility.
515 Lines with similar content should be aligned vertically when doing so
516 improves readability.
518 Consider to add a regression test for your code.
520 If you added YASM code please check that things still work with --disable-yasm
522 Make sure you check the return values of function and return appropriate
523 error codes. Especially memory allocation functions like @code{av_malloc()}
524 are notoriously left unchecked, which is a serious problem.
526 Test your code with valgrind and or Address Sanitizer to ensure it's free
527 of leaks, out of array accesses, etc.
530 @section Patch review process
532 All patches posted to ffmpeg-devel will be reviewed, unless they contain a
533 clear note that the patch is not for the git master branch.
534 Reviews and comments will be posted as replies to the patch on the
535 mailing list. The patch submitter then has to take care of every comment,
536 that can be by resubmitting a changed patch or by discussion. Resubmitted
537 patches will themselves be reviewed like any other patch. If at some point
538 a patch passes review with no comments then it is approved, that can for
539 simple and small patches happen immediately while large patches will generally
540 have to be changed and reviewed many times before they are approved.
541 After a patch is approved it will be committed to the repository.
543 We will review all submitted patches, but sometimes we are quite busy so
544 especially for large patches this can take several weeks.
546 If you feel that the review process is too slow and you are willing to try to
547 take over maintainership of the area of code you change then just clone
548 git master and maintain the area of code there. We will merge each area from
549 where its best maintained.
551 When resubmitting patches, please do not make any significant changes
552 not related to the comments received during review. Such patches will
553 be rejected. Instead, submit significant changes or new features as
556 @anchor{Regression tests}
557 @section Regression tests
559 Before submitting a patch (or committing to the repository), you should at least
560 test that you did not break anything.
562 Running 'make fate' accomplishes this, please see @url{fate.html} for details.
564 [Of course, some patches may change the results of the regression tests. In
565 this case, the reference results of the regression tests shall be modified
568 @subsection Adding files to the fate-suite dataset
570 When there is no muxer or encoder available to generate test media for a
571 specific test then the media has to be inlcuded in the fate-suite.
572 First please make sure that the sample file is as small as possible to test the
573 respective decoder or demuxer sufficiently. Large files increase network
574 bandwidth and disk space requirements.
575 Once you have a working fate test and fate sample, provide in the commit
576 message or introductionary message for the patch series that you post to
577 the ffmpeg-devel mailing list, a direct link to download the sample media.