1 <!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN" "http://www.w3.org/TR/html401/loose.dtd">
3 <!-- Created on May 18, 2011 by texi2html 1.82
4 texi2html was written by:
5 Lionel Cons <Lionel.Cons@cern.ch> (original author)
6 Karl Berry <karl@freefriends.org>
7 Olaf Bachmann <obachman@mathematik.uni-kl.de>
9 Maintained by: Many creative people.
10 Send bugs and suggestions to <texi2html-bug@nongnu.org>
13 <title>Developer Documentation</title>
15 <meta name="description" content="Developer Documentation">
16 <meta name="keywords" content="Developer Documentation">
17 <meta name="resource-type" content="document">
18 <meta name="distribution" content="global">
19 <meta name="Generator" content="texi2html 1.82">
20 <meta http-equiv="Content-Type" content="text/html; charset=utf-8">
21 <style type="text/css">
23 a.summary-letter {text-decoration: none}
24 blockquote.smallquotation {font-size: smaller}
25 pre.display {font-family: serif}
26 pre.format {font-family: serif}
27 pre.menu-comment {font-family: serif}
28 pre.menu-preformatted {font-family: serif}
29 pre.smalldisplay {font-family: serif; font-size: smaller}
30 pre.smallexample {font-size: smaller}
31 pre.smallformat {font-family: serif; font-size: smaller}
32 pre.smalllisp {font-size: smaller}
33 span.roman {font-family:serif; font-weight:normal;}
34 span.sansserif {font-family:sans-serif; font-weight:normal;}
35 ul.toc {list-style: none}
42 <body lang="en" bgcolor="#FFFFFF" text="#000000" link="#0000FF" vlink="#800080" alink="#FF0000">
44 <a name="SEC_Top"></a>
45 <h1 class="settitle">Developer Documentation</h1>
47 <a name="SEC_Contents"></a>
48 <h1>Table of Contents</h1>
49 <div class="contents">
52 <li><a name="toc-Developers-Guide" href="#Developers-Guide">1. Developers Guide</a>
54 <li><a name="toc-API" href="#API">1.1 API</a></li>
55 <li><a name="toc-Integrating-libavcodec-or-libavformat-in-your-program" href="#Integrating-libavcodec-or-libavformat-in-your-program">1.2 Integrating libavcodec or libavformat in your program</a></li>
56 <li><a name="toc-Coding-Rules-1" href="#Coding-Rules-1">1.3 Coding Rules</a></li>
57 <li><a name="toc-Development-Policy" href="#Development-Policy">1.4 Development Policy</a></li>
58 <li><a name="toc-Submitting-patches" href="#Submitting-patches">1.5 Submitting patches</a></li>
59 <li><a name="toc-New-codecs-or-formats-checklist" href="#New-codecs-or-formats-checklist">1.6 New codecs or formats checklist</a></li>
60 <li><a name="toc-patch-submission-checklist" href="#patch-submission-checklist">1.7 patch submission checklist</a></li>
61 <li><a name="toc-Patch-review-process" href="#Patch-review-process">1.8 Patch review process</a></li>
62 <li><a name="toc-Regression-tests" href="#Regression-tests">1.9 Regression tests</a></li>
69 <a name="Developers-Guide"></a>
70 <h1 class="chapter"><a href="developer.html#toc-Developers-Guide">1. Developers Guide</a></h1>
73 <h2 class="section"><a href="developer.html#toc-API">1.1 API</a></h2>
75 <li> libavcodec is the library containing the codecs (both encoding and
76 decoding). Look at ‘<tt>libavcodec/apiexample.c</tt>’ to see how to use it.
78 </li><li> libavformat is the library containing the file format handling (mux and
79 demux code for several formats). Look at ‘<tt>ffplay.c</tt>’ to use it in a
80 player. See ‘<tt>libavformat/output-example.c</tt>’ to use it to generate
81 audio or video streams.
85 <a name="Integrating-libavcodec-or-libavformat-in-your-program"></a>
86 <h2 class="section"><a href="developer.html#toc-Integrating-libavcodec-or-libavformat-in-your-program">1.2 Integrating libavcodec or libavformat in your program</a></h2>
88 <p>You can integrate all the source code of the libraries to link them
89 statically to avoid any version problem. All you need is to provide a
90 ’config.mak’ and a ’config.h’ in the parent directory. See the defines
91 generated by ./configure to understand what is needed.
93 <p>You can use libavcodec or libavformat in your commercial program, but
94 <em>any patch you make must be published</em>. The best way to proceed is
95 to send your patches to the FFmpeg mailing list.
97 <p><a name="Coding-Rules"></a>
98 </p><a name="Coding-Rules-1"></a>
99 <h2 class="section"><a href="developer.html#toc-Coding-Rules-1">1.3 Coding Rules</a></h2>
101 <p>FFmpeg is programmed in the ISO C90 language with a few additional
102 features from ISO C99, namely:
105 the ‘<samp>inline</samp>’ keyword;
107 ‘<samp>//</samp>’ comments;
109 designated struct initializers (‘<samp>struct s x = { .i = 17 };</samp>’)
111 compound literals (‘<samp>x = (struct s) { 17, 23 };</samp>’)
114 <p>These features are supported by all compilers we care about, so we will not
115 accept patches to remove their use unless they absolutely do not impair
116 clarity and performance.
118 <p>All code must compile with GCC 2.95 and GCC 3.3. Currently, FFmpeg also
119 compiles with several other compilers, such as the Compaq ccc compiler
120 or Sun Studio 9, and we would like to keep it that way unless it would
121 be exceedingly involved. To ensure compatibility, please do not use any
122 additional C99 features or GCC extensions. Especially watch out for:
125 mixing statements and declarations;
127 ‘<samp>long long</samp>’ (use ‘<samp>int64_t</samp>’ instead);
129 ‘<samp>__attribute__</samp>’ not protected by ‘<samp>#ifdef __GNUC__</samp>’ or similar;
131 GCC statement expressions (‘<samp>(x = ({ int y = 4; y; })</samp>’).
135 The presentation is one inspired by ’indent -i4 -kr -nut’.
136 The TAB character is forbidden outside of Makefiles as is any
137 form of trailing whitespace. Commits containing either will be
138 rejected by the git repository.
140 <p>The main priority in FFmpeg is simplicity and small code size in order to
141 minimize the bug count.
143 <p>Comments: Use the JavaDoc/Doxygen
144 format (see examples below) so that code documentation
145 can be generated automatically. All nontrivial functions should have a comment
146 above them explaining what the function does, even if it is just one sentence.
147 All structures and their member variables should be documented, too.
148 </p><table><tr><td> </td><td><pre class="example">/**
159 typedef struct Foobar{
160 int var1; /**< var1 description */
161 int var2; ///< var2 description
162 /** var3 description */
170 * @param my_parameter description of my_parameter
171 * @return return value description
173 int myfunc(int my_parameter)
175 </pre></td></tr></table>
177 <p>fprintf and printf are forbidden in libavformat and libavcodec,
178 please use av_log() instead.
180 <p>Casts should be used only when necessary. Unneeded parentheses
181 should also be avoided if they don’t make the code easier to understand.
183 <a name="Development-Policy"></a>
184 <h2 class="section"><a href="developer.html#toc-Development-Policy">1.4 Development Policy</a></h2>
188 Contributions should be licensed under the LGPL 2.1, including an
189 "or any later version" clause, or the MIT license. GPL 2 including
190 an "or any later version" clause is also acceptable, but LGPL is
193 You must not commit code which breaks FFmpeg! (Meaning unfinished but
194 enabled code which breaks compilation or compiles but does not work or
195 breaks the regression tests)
196 You can commit unfinished stuff (for testing etc), but it must be disabled
197 (#ifdef etc) by default so it does not interfere with other developers’
200 You do not have to over-test things. If it works for you, and you think it
201 should work for others, then commit. If your code has problems
202 (portability, triggers compiler bugs, unusual environment etc) they will be
203 reported and eventually fixed.
205 Do not commit unrelated changes together, split them into self-contained
206 pieces. Also do not forget that if part B depends on part A, but A does not
207 depend on B, then A can and should be committed first and separate from B.
208 Keeping changes well split into self-contained parts makes reviewing and
209 understanding them on the commit log mailing list easier. This also helps
210 in case of debugging later on.
211 Also if you have doubts about splitting or not splitting, do not hesitate to
212 ask/discuss it on the developer mailing list.
214 Do not change behavior of the programs (renaming options etc) or public
215 API or ABI without first discussing it on the ffmpeg-devel mailing list.
216 Do not remove functionality from the code. Just improve!
218 <p> Note: Redundant code can be removed.
220 Do not commit changes to the build system (Makefiles, configure script)
221 which change behavior, defaults etc, without asking first. The same
222 applies to compiler warning fixes, trivial looking fixes and to code
223 maintained by other developers. We usually have a reason for doing things
224 the way we do. Send your changes as patches to the ffmpeg-devel mailing
225 list, and if the code maintainers say OK, you may commit. This does not
226 apply to files you wrote and/or maintain.
228 We refuse source indentation and other cosmetic changes if they are mixed
229 with functional changes, such commits will be rejected and removed. Every
230 developer has his own indentation style, you should not change it. Of course
231 if you (re)write something, you can use your own style, even though we would
232 prefer if the indentation throughout FFmpeg was consistent (Many projects
233 force a given indentation style - we do not.). If you really need to make
234 indentation changes (try to avoid this), separate them strictly from real
237 <p> NOTE: If you had to put if(){ .. } over a large (> 5 lines) chunk of code,
238 then either do NOT change the indentation of the inner part within (do not
239 move it to the right)! or do so in a separate commit
241 Always fill out the commit log message. Describe in a few lines what you
242 changed and why. You can refer to mailing list postings if you fix a
243 particular bug. Comments such as "fixed!" or "Changed it." are unacceptable.
245 area changed: Short 1 line description
247 <p> details describing what and why and giving references.
249 Make sure the author of the commit is set correctly. (see git commit –author)
250 If you apply a patch, send an
251 answer to ffmpeg-devel (or wherever you got the patch from) saying that
252 you applied the patch.
254 When applying patches that have been discussed (at length) on the mailing
255 list, reference the thread in the log message.
257 Do NOT commit to code actively maintained by others without permission.
258 Send a patch to ffmpeg-devel instead. If no one answers within a reasonable
259 timeframe (12h for build failures and security fixes, 3 days small changes,
260 1 week for big patches) then commit your patch if you think it is OK.
261 Also note, the maintainer can simply ask for more time to review!
263 Subscribe to the ffmpeg-cvslog mailing list. The diffs of all commits
264 are sent there and reviewed by all the other developers. Bugs and possible
265 improvements or general questions regarding commits are discussed there. We
266 expect you to react if problems with your code are uncovered.
268 Update the documentation if you change behavior or add features. If you are
269 unsure how best to do this, send a patch to ffmpeg-devel, the documentation
270 maintainer(s) will review and commit your stuff.
272 Try to keep important discussions and requests (also) on the public
273 developer mailing list, so that all developers can benefit from them.
275 Never write to unallocated memory, never write over the end of arrays,
276 always check values read from some untrusted source before using them
277 as array index or other risky things.
279 Remember to check if you need to bump versions for the specific libav
280 parts (libavutil, libavcodec, libavformat) you are changing. You need
281 to change the version integer.
282 Incrementing the first component means no backward compatibility to
283 previous versions (e.g. removal of a function from the public API).
284 Incrementing the second component means backward compatible change
285 (e.g. addition of a function to the public API or extension of an
286 existing data structure).
287 Incrementing the third component means a noteworthy binary compatible
288 change (e.g. encoder bug fix that matters for the decoder).
290 Compiler warnings indicate potential bugs or code with bad style. If a type of
291 warning always points to correct and clean code, that warning should
292 be disabled, not the code changed.
293 Thus the remaining warnings can either be bugs or correct code.
294 If it is a bug, the bug has to be fixed. If it is not, the code should
295 be changed to not generate a warning unless that causes a slowdown
296 or obfuscates the code.
298 If you add a new file, give it a proper license header. Do not copy and
299 paste it from a random place, use an existing file as template.
302 <p>We think our rules are not too hard. If you have comments, contact us.
304 <p>Note, these rules are mostly borrowed from the MPlayer project.
306 <a name="Submitting-patches"></a>
307 <h2 class="section"><a href="developer.html#toc-Submitting-patches">1.5 Submitting patches</a></h2>
309 <p>First, read the (see <a href="#Coding-Rules">Coding Rules</a>) above if you did not yet.
311 <p>When you submit your patch, please use <code>git format-patch</code> or
312 <code>git send-email</code>. We cannot read other diffs :-)
314 <p>Also please do not submit a patch which contains several unrelated changes.
315 Split it into separate, self-contained pieces. This does not mean splitting
316 file by file. Instead, make the patch as small as possible while still
317 keeping it as a logical unit that contains an individual change, even
318 if it spans multiple files. This makes reviewing your patches much easier
319 for us and greatly increases your chances of getting your patch applied.
321 <p>Use the patcheck tool of FFmpeg to check your patch.
322 The tool is located in the tools directory.
324 <p>Run the regression tests before submitting a patch so that you can
325 verify that there are no big problems.
327 <p>Patches should be posted as base64 encoded attachments (or any other
328 encoding which ensures that the patch will not be trashed during
329 transmission) to the ffmpeg-devel mailing list, see
330 <a href="http://lists.ffmpeg.org/mailman/listinfo/ffmpeg-devel">http://lists.ffmpeg.org/mailman/listinfo/ffmpeg-devel</a>
332 <p>It also helps quite a bit if you tell us what the patch does (for example
333 ’replaces lrint by lrintf’), and why (for example ’*BSD isn’t C99 compliant
334 and has no lrint()’)
336 <p>Also please if you send several patches, send each patch as a separate mail,
337 do not attach several unrelated patches to the same mail.
339 <p>Your patch will be reviewed on the mailing list. You will likely be asked
340 to make some changes and are expected to send in an improved version that
341 incorporates the requests from the review. This process may go through
342 several iterations. Once your patch is deemed good enough, some developer
343 will pick it up and commit it to the official FFmpeg tree.
345 <p>Give us a few days to react. But if some time passes without reaction,
346 send a reminder by email. Your patch should eventually be dealt with.
349 <a name="New-codecs-or-formats-checklist"></a>
350 <h2 class="section"><a href="developer.html#toc-New-codecs-or-formats-checklist">1.6 New codecs or formats checklist</a></h2>
354 Did you use av_cold for codec initialization and close functions?
356 Did you add a long_name under NULL_IF_CONFIG_SMALL to the AVCodec or
357 AVInputFormat/AVOutputFormat struct?
359 Did you bump the minor version number (and reset the micro version
360 number) in ‘<tt>avcodec.h</tt>’ or ‘<tt>avformat.h</tt>’?
362 Did you register it in ‘<tt>allcodecs.c</tt>’ or ‘<tt>allformats.c</tt>’?
364 Did you add the CodecID to ‘<tt>avcodec.h</tt>’?
366 If it has a fourcc, did you add it to ‘<tt>libavformat/riff.c</tt>’,
367 even if it is only a decoder?
369 Did you add a rule to compile the appropriate files in the Makefile?
370 Remember to do this even if you’re just adding a format to a file that is
371 already being compiled by some other rule, like a raw demuxer.
373 Did you add an entry to the table of supported formats or codecs in
374 ‘<tt>doc/general.texi</tt>’?
376 Did you add an entry in the Changelog?
378 If it depends on a parser or a library, did you add that dependency in
381 Did you <code>git add</code> the appropriate files before committing?
383 Did you make sure it compiles standalone, i.e. with
384 <code>configure --disable-everything --enable-decoder=foo</code>
385 (or <code>--enable-demuxer</code> or whatever your component is)?
389 <a name="patch-submission-checklist"></a>
390 <h2 class="section"><a href="developer.html#toc-patch-submission-checklist">1.7 patch submission checklist</a></h2>
394 Does ’make fate’ pass with the patch applied?
396 Was the patch generated with git format-patch or send-email?
398 Did you sign off your patch? (git commit -s)
399 See <a href="http://kerneltrap.org/files/Jeremy/DCO.txt">http://kerneltrap.org/files/Jeremy/DCO.txt</a> for the meaning
402 Did you provide a clear git commit log message?
404 Is the patch against latest FFmpeg git master branch?
406 Are you subscribed to ffmpeg-dev?
407 (the list is subscribers only due to spam)
409 Have you checked that the changes are minimal, so that the same cannot be
410 achieved with a smaller patch and/or simpler final code?
412 If the change is to speed critical code, did you benchmark it?
414 If you did any benchmarks, did you provide them in the mail?
416 Have you checked that the patch does not introduce buffer overflows or
417 other security issues?
419 Did you test your decoder or demuxer against damaged data? If no, see
420 tools/trasher and the noise bitstream filter. Your decoder or demuxer
421 should not crash or end in a (near) infinite loop when fed damaged data.
423 Does the patch not mix functional and cosmetic changes?
425 Did you add tabs or trailing whitespace to the code? Both are forbidden.
427 Is the patch attached to the email you send?
429 Is the mime type of the patch correct? It should be text/x-diff or
430 text/x-patch or at least text/plain and not application/octet-stream.
432 If the patch fixes a bug, did you provide a verbose analysis of the bug?
434 If the patch fixes a bug, did you provide enough information, including
435 a sample, so the bug can be reproduced and the fix can be verified?
436 Note please do not attach samples >100k to mails but rather provide a
437 URL, you can upload to ftp://upload.ffmpeg.org
439 Did you provide a verbose summary about what the patch does change?
441 Did you provide a verbose explanation why it changes things like it does?
443 Did you provide a verbose summary of the user visible advantages and
444 disadvantages if the patch is applied?
446 Did you provide an example so we can verify the new feature added by the
449 If you added a new file, did you insert a license header? It should be
450 taken from FFmpeg, not randomly copied and pasted from somewhere else.
452 You should maintain alphabetical order in alphabetically ordered lists as
453 long as doing so does not break API/ABI compatibility.
455 Lines with similar content should be aligned vertically when doing so
456 improves readability.
458 Consider to add a regression test for your code.
461 <a name="Patch-review-process"></a>
462 <h2 class="section"><a href="developer.html#toc-Patch-review-process">1.8 Patch review process</a></h2>
464 <p>All patches posted to ffmpeg-devel will be reviewed, unless they contain a
465 clear note that the patch is not for the git master branch.
466 Reviews and comments will be posted as replies to the patch on the
467 mailing list. The patch submitter then has to take care of every comment,
468 that can be by resubmitting a changed patch or by discussion. Resubmitted
469 patches will themselves be reviewed like any other patch. If at some point
470 a patch passes review with no comments then it is approved, that can for
471 simple and small patches happen immediately while large patches will generally
472 have to be changed and reviewed many times before they are approved.
473 After a patch is approved it will be committed to the repository.
475 <p>We will review all submitted patches, but sometimes we are quite busy so
476 especially for large patches this can take several weeks.
478 <p>When resubmitting patches, please do not make any significant changes
479 not related to the comments received during review. Such patches will
480 be rejected. Instead, submit significant changes or new features as
483 <a name="Regression-tests"></a>
484 <h2 class="section"><a href="developer.html#toc-Regression-tests">1.9 Regression tests</a></h2>
486 <p>Before submitting a patch (or committing to the repository), you should at least
487 test that you did not break anything.
489 <p>The regression tests build a synthetic video stream and a synthetic
490 audio stream. These are then encoded and decoded with all codecs or
491 formats. The CRC (or MD5) of each generated file is recorded in a
492 result file. A ’diff’ is launched to compare the reference results and
493 the result file. The output is checked immediately after each test
496 <p>The regression tests then go on to test the FFserver code with a
497 limited set of streams. It is important that this step runs correctly
500 <p>Run ’make test’ to test all the codecs and formats. Commands like
501 ’make regtest-mpeg2’ can be used to run a single test. By default,
502 make will abort if any test fails. To run all tests regardless,
503 use make -k. To get a more verbose output, use ’make V=1 test’ or
504 ’make V=2 test’.
506 <p>Run ’make fulltest’ to test all the codecs, formats and FFserver.
508 <p>[Of course, some patches may change the results of the regression tests. In
509 this case, the reference results of the regression tests shall be modified
515 This document was generated by <em>Kyle Schwarz</em> on <em>May 18, 2011</em> using <a href="http://www.nongnu.org/texi2html/"><em>texi2html 1.82</em></a>.