[Ffmpeg-devel-irc] ffmpeg-devel.log.20160413

burek burek021 at gmail.com
Thu Apr 14 02:05:03 CEST 2016


[02:51:20 CEST] <jamrial_> BBB: checkasm started failing on some fate clients
[03:09:01 CEST] <BBB> jamrial_: I believe michaelni just fixed that
[03:09:16 CEST] <BBB> http://git.videolan.org/?p=ffmpeg.git;a=commit;h=4d59d075a96c7bc9fc7a118f96015fbf5156708a
[03:09:50 CEST] <BBB> it may take a while for that to relay to all clients
[03:29:58 CEST] <jamrial_> oh, ok
[03:30:06 CEST] <jamrial_> sorry for the ping then
[06:42:22 CEST] <cone-720> ffmpeg 03James Almer 07master:3674a53d1710: avformat/uncodedframecrc: fix incompatible pointer type warning
[08:40:55 CEST] <cone-720> ffmpeg 03Tobias Rapp 07master:ee104580c5b8: avfilter/vf_drawtext: add optional default value to metadata function
[09:18:24 CEST] <durandal_1707> nevcairiel: what you used for decoding 09 wmal track?
[09:41:45 CEST] <nevcairiel> durandal_1707:  like i said on trac, the Microsoft decoder
[09:51:49 CEST] <durandal_1707> nevcairiel: that is utility/application/lib?
[09:53:41 CEST] <nevcairiel> "WMAudio Decoder DMO"
[09:53:53 CEST] <nevcairiel> (wmadmod.dll)
[10:28:48 CEST] <cone-720> ffmpeg 03Rodger Combs 07master:b20d3bf4a45d: lavc/audiotoolboxdec: avoid relying on consumer-provided params when possible
[10:28:49 CEST] <cone-720> ffmpeg 03Rodger Combs 07master:acd5910e39be: lavc/audiotoolboxdec: reindent
[10:28:50 CEST] <cone-720> ffmpeg 03Rodger Combs 07master:c11157c09a11: lavf/audiotoolboxdec: only send extradata for formats that use it
[10:28:51 CEST] <cone-720> ffmpeg 03Rodger Combs 07master:c890bcbacf32: lavf/audiotoolboxdec: only provide block alignment for ILBC
[11:29:36 CEST] <rcombs> Daemon404: sent new patches rebased against latest master (i.e. with codecpar); no warnings; no longer breaks DASH by default; works fine here
[11:56:12 CEST] <wm4> so how do I build ffmpeg in a way you can debug it? I tried with  --enable-debug --disable-optimizations
[11:58:01 CEST] <wm4> ah need to add --disable-stripping
[11:58:04 CEST] <wm4> why is this so absurd
[11:58:39 CEST] <funman> isn't ffmpeg_g always generated regardless of --disable-stripping?
[11:58:44 CEST] <nevcairiel> you could just debug the ffmpeg_g variant which is not strpped
[11:59:21 CEST] <wm4> what if I'm not using ffmpeg CLI
[11:59:44 CEST] <wm4> will ffmpeg_g have the lib debug symbols even if they're built as dynamic libs?
[12:00:09 CEST] <wm4> maybe there should be something like --developer-mode that enables all the obscure stuff you need to actually debug libavcodec etc.
[12:00:41 CEST] <nevcairiel> shared libraries are only stripped in the install step, if you reference them from the source folder then they are intact
[12:01:29 CEST] <wm4> you can't build external code like this, unless you do convoluted things
[12:01:44 CEST] <nevcairiel> well then just build with disable stripping =p
[12:02:12 CEST] <funman> you can use LD_LIBRARY_PATH when running gdb so the unstripped libs are loaded
[12:02:16 CEST] <wm4> I'm using this with success: --enable-debug --disable-optimizations --disable-stripping
[12:02:35 CEST] <wm4> but now imagine someone wants to debug a program using libavcodec, and the problem appears to be in libavcodec
[12:02:57 CEST] <nevcairiel> enable debug isnt needed btw, its enabled by default
[12:03:25 CEST] <wm4> so what's the point of having that enabled by default, and stirpping it off by default
[12:03:41 CEST] <nevcairiel> because it produces ffmpeg_g as well
[12:04:04 CEST] <nevcairiel> if you work on any of the libs its often enough
[12:12:20 CEST] <pa> sorry for the OT question: do you guys know whether there exists an AV container that can be seeked before the file is closed? that is while it is being written
[12:15:29 CEST] <wm4> a what?
[12:15:30 CEST] <pa> for example if i want to transcode the stream from a DVB card, but i want to seek inside while ffmpeg is still capturing/transcoding
[12:15:49 CEST] <nevcairiel> use a streaming container like mpegts then
[12:15:53 CEST] <wm4> so you want to play a file that is being written by ffmpeg with a different program?
[12:15:59 CEST] <pa> yes
[12:16:09 CEST] <pa> like mplayer and then pgup /pgdown
[12:16:13 CEST] <pa> or <- / ->
[12:16:18 CEST] <ubitux> flv, mkv, webm should do
[12:16:31 CEST] <pa> ubitux, mkv? doesn't that write the index at the end?
[12:16:32 CEST] <nevcairiel> ubitux: mkv without a seeking index doesnt work well
[12:16:45 CEST] <wm4> mkv without seeking index works perfectly well, but can be slow
[12:16:58 CEST] <nevcairiel> being very slow isnt working "well" =p
[12:17:10 CEST] <ubitux> even avi should work, right?
[12:17:12 CEST] <wm4> depends how fast your IO is
[12:17:14 CEST] <nevcairiel> most demuxers will try to build their own index, or just blindly seek around in it
[12:17:32 CEST] <pa> well i'm asking for a container that somehow builds and store an incremental index
[12:17:43 CEST] <nevcairiel> for the second case, you might as well use mpegts then, which seems appropriate when recording a mpegts DVB stream anyway
[12:17:53 CEST] <pa> right..
[12:18:05 CEST] <pa> so with mpegts i can mplayer -ss quickly, right?
[12:18:22 CEST] <DHE> mpegts has no index. you just seek based on bitrate estimates, resync and look at the timestamps
[12:18:28 CEST] <nevcairiel> but i'm not aware of any container that writes an incremental index, the problem with that is that it would have to reserve space in the file somewhere, but it has no clue how much since it doesnt know how long the file is going to be
[12:18:31 CEST] <pa> ah okay
[12:18:46 CEST] <nevcairiel> hence seeking indexes are often at the end
[12:18:47 CEST] <pa> well it could build indexes in chunks
[12:18:55 CEST] <pa> and link to the next chunk at file offset
[12:18:57 CEST] <DHE> a fragmented index?
[12:19:01 CEST] <pa> yes
[12:19:16 CEST] <pa> each chunk could hold one hour or so
[12:19:19 CEST] <pa> configuragble perhaps
[12:19:27 CEST] <nevcairiel> honestly its not much of a practical use-case to design a container around =p
[12:19:38 CEST] <nevcairiel> seeking in mpegts is usually acceptably fast
[12:19:44 CEST] <pa> well if mpegts does it well, it's not needed indeed
[12:20:28 CEST] <nevcairiel> having an index would be faster of course, but its not like you really wait for an mpegts seek unless your IO is high latency
[12:20:49 CEST] <pa> right
[12:20:54 CEST] <pa> can try it out, thanks :)
[12:20:56 CEST] <av500> I think devices like PVRs build their own index files on the side
[12:21:06 CEST] <nevcairiel> thats always possible
[12:21:07 CEST] <av500> they record TS but also keep an index
[12:21:09 CEST] <pa> ah that would be an alternative
[12:21:14 CEST] <av500> fur stuff like timeshifting
[12:21:14 CEST] <pa> and embed it at the end
[12:21:28 CEST] <av500> TS cannot embed an index
[12:21:30 CEST] <pa> but does ffmpeg have such option?
[12:21:32 CEST] <av500> no
[12:21:52 CEST] <nevcairiel> somehow i thought there actually was some odd approach to load an index from somewhere else
[12:21:57 CEST] <nevcairiel> but maybe i'm thinking of something else
[12:22:15 CEST] <pa> that could be nice.. like when creating an mkv for example
[12:22:18 CEST] <pa> index file on the side
[12:22:21 CEST] <pa> embed it at the end
[12:22:30 CEST] <pa> or use it as separate index before the end
[12:23:23 CEST] <av500> but that would mean that the user app would need to know the filepos from the current frame in the MKV file
[12:23:28 CEST] <av500> from the muxer
[12:23:30 CEST] <pa> yes
[12:23:36 CEST] <av500> to write to the index
[12:23:39 CEST] <pa> it could for example be written in the container
[12:23:43 CEST] <pa> before closing it
[12:24:01 CEST] <pa> so at the end in place of the URL of the index, the actual index could be written
[12:24:22 CEST] <av500> well, the MKV muxer keeps an index in memory
[12:24:29 CEST] <av500> that it will write when you call close()
[12:24:35 CEST] <av500> so there is already an index
[12:24:42 CEST] <pa> right
[12:24:48 CEST] <pa> but can't be used by a player
[12:25:03 CEST] <pa> maybe i could try to hack ffmpeg
[12:25:06 CEST] <pa> to do that :)
[12:25:10 CEST] <av500> sure :)
[12:25:21 CEST] <av500> add an api to query the muxer index
[12:25:44 CEST] <pa> iirc mkv has options to add some empty space between the header and the first block
[12:26:02 CEST] <pa> where i could write the index url
[12:26:09 CEST] <av500> url?
[12:26:13 CEST] <pa> like file path
[12:26:33 CEST] <av500> "index in the cloud"
[12:26:36 CEST] <nevcairiel> referencing external files from media files is always a huge PITA
[12:26:38 CEST] <av500> make it a service :)
[12:26:55 CEST] <pa> nevcairiel, wouldn't abs path work?
[14:54:08 CEST] <Daemon404> rcombs, i will look
[14:54:15 CEST] <rcombs> thanks
[15:14:27 CEST] <Daemon404> rcombs, you should add maybe a verbose log or some documentation for the autobsf disabling behavior
[15:32:38 CEST] <cone-774> ffmpeg 03Diego Biurrun 07master:1a094af63828: fft: Split MDCT bits off from FFT
[15:32:38 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:546bf316cad2: Merge commit '1a094af638281295bf087945923d258b5acd1ab1'
[15:36:50 CEST] <cone-774> ffmpeg 03Luca Barbato 07master:d4066a702407: indeo2data: K&R formatting cosmetics
[15:36:51 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:6b1a0f205868: Merge commit 'd4066a702407352a0648af882c34ea81a404fa2b'
[15:38:50 CEST] <cone-774> ffmpeg 03Luca Barbato 07master:f8c34f4b8d62: indeo2: Fix banding artefacts
[15:38:51 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:78af369189eb: Merge commit 'f8c34f4b8d62afad3f63cf3d9617d73735bef8c1'
[15:39:40 CEST] <Daemon404> nevcairiel / michaelni - ping
[15:39:46 CEST] <Daemon404> next commit needs a new fate sample upload
[15:41:37 CEST] <Daemon404> who else can upload?
[15:42:06 CEST] <nevcairiel> you know how it goes, sample and place
[15:42:35 CEST] <Daemon404> http://fate-suite.libav.org/rt21/ISKATE.AVI
[15:42:42 CEST] <Daemon404> same place in our fate server
[15:45:11 CEST] <nevcairiel> Daemon404: done, please check
[15:45:33 CEST] <Daemon404> nevcairiel, rsync worked thanks
[15:50:57 CEST] <cone-774> ffmpeg 03Vittorio Giovara 07master:b39ab8549a53: fate: Add test for indeo2 with delta frames
[15:50:58 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:622f18a2bf46: Merge commit 'b39ab8549a53e2fc7978ab9db50e5c2ba6a6602d'
[16:06:44 CEST] <durandal_1707> nevcairiel: wmadmod.dll you used is for 64 bit?
[16:07:04 CEST] <nevcairiel> 32-bit i think, but i probably have both
[16:09:19 CEST] <durandal_1707> nevcairiel: md5 hash of it?
[16:10:16 CEST] <durandal_1707> mine is 03dea3b0b07a91fa145578a9973c24af
[16:10:18 CEST] <nevcairiel> durandal_1707: 0b7c5790893f3650162bed4bea35d9a6 *WMADMOD.DLL
[16:10:39 CEST] <nevcairiel> version 10.0.10586.63
[16:10:43 CEST] <nevcairiel> ie. latest windows 10 version
[16:21:18 CEST] <cone-774> ffmpeg 03Diego Biurrun 07master:11843ededacd: fate: Add separate target for all indeo3 tests
[16:21:19 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:884dd175f061: Merge commit '11843ededacd0157aea642771837557549b5b417'
[16:22:01 CEST] <Daemon404> nevcairiel, https://git.libav.org/?p=libav.git;a=commitdiff;h=1ceb07eb313c2d51383408025e57a2fe50ccd164
[16:22:08 CEST] <Daemon404> merge y/n?
[16:22:47 CEST] <nevcairiel> seems sensible, does it impact any fate results?
[16:22:54 CEST] <Daemon404> im going to run it
[16:36:24 CEST] <Daemon404> nevcairiel, no fate failures
[16:38:29 CEST] <cone-774> ffmpeg 03Anton Khirnov 07master:1ceb07eb313c: avformat_find_stream_info: move duration guessing after updating codec parameters
[16:38:30 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:3c461eecd48b: Merge commit '1ceb07eb313c2d51383408025e57a2fe50ccd164'
[16:47:53 CEST] <Daemon404> wow our asfenc code is REALLY different
[16:51:00 CEST] <RiCON> isn't it the old one?
[16:51:27 CEST] <RiCON> the libav asf is asf_o
[16:51:38 CEST] <Daemon404> asfenc
[16:51:48 CEST] <RiCON> disregard
[16:52:24 CEST] <Daemon404> 1ceff0859df1c4f6bfacd6c1cd9dbdcceb039423 asfenc: rename some variables
[16:52:31 CEST] <Daemon404> except it also changes where some stuff is stored
[16:52:35 CEST] <Daemon404> without explanation
[16:52:42 CEST] <Daemon404> so much for hunting for reasons why stuff is the way it is
[17:09:27 CEST] <cone-774> ffmpeg 03Anton Khirnov 07master:ff3db937ef3a: asfenc: fix some possible integer overflows
[17:09:28 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:bfd0f42277d1: Merge commit 'ff3db937ef3aa30046a3936146f86ad48ee2ff90'
[17:10:28 CEST] <cone-774> ffmpeg 03Anton Khirnov 07master:84b5dcf27589: asfenc: remove an unused variable
[17:10:29 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:3bf368a2a643: Merge commit '84b5dcf27589b32713a4ba0723a129156b4d2408'
[17:11:56 CEST] <cone-774> ffmpeg 03wm4 07master:7a6cf2771414: lavu: improve documentation of some AVFrame functions
[17:11:57 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:be6b58a98171: Merge commit '7a6cf2771414c7ab8bca0811d589f6091a6e2b71'
[17:15:05 CEST] <cone-774> ffmpeg 03wm4 07master:2e2f8534ebde: lavc: factor apply_param_change() AV_EF_EXPLODE handling
[17:15:06 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:f5b93c46e93a: Merge commit '2e2f8534ebde47d3a3909fe64c2e66204bc56874'
[17:15:22 CEST] <Daemon404> wm4, ping
[17:15:32 CEST] <Daemon404> whats up with "avconv: remove sub-frame warning"
[17:15:36 CEST] <Daemon404> it's in libav but not ours?
[17:15:41 CEST] <Daemon404> mailing list bikeshedding?
[17:15:59 CEST] <wm4> it's in ffmpeg.c too
[17:16:06 CEST] <wm4> you can probably skip it for now
[17:16:13 CEST] <wm4> michaelni was against removing it this way
[17:17:04 CEST] <Daemon404> ok
[17:17:59 CEST] <cone-774> ffmpeg 03wm4 07master:0b6e5d6b32b9: avconv: remove sub-frame warning
[17:18:00 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:e7ac968f60dd: Merge commit '0b6e5d6b32b91c6da79cd919a3c2ede9d682f838'
[17:22:33 CEST] <cone-774> ffmpeg 03Vittorio Giovara 07master:d40cb726d271: mov: Trim dref absolute path
[17:22:34 CEST] <cone-774> ffmpeg 03Vittorio Giovara 07master:e10b7ef2fe56: vdpau: Add missing deprecation guards
[17:22:35 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:feb1f7abc587: Merge commit 'd40cb726d271b0284642a1ba159eb26a5c579f77'
[17:22:36 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:09dc6845662b: Merge commit 'e10b7ef2fe56603fb1baac6b20fd6bd0a3fdd0d0'
[17:26:11 CEST] <cone-774> ffmpeg 03Katerina Barone-Adesi 07master:1389b4c18d10: idct8x8: Fix undefined negative shifts
[17:26:12 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:eff2b4311750: Merge commit '1389b4c18d1042c196603ba66c25113bcee1738b'
[17:29:52 CEST] <cone-774> ffmpeg 03Luca Barbato 07master:7d4a1ff344cb: mpegvideo: Fix undefined negative shifts in ff_init_block_index
[17:29:54 CEST] <cone-774> ffmpeg 03Luca Barbato 07master:024235139064: mpegvideo: Fix undefined negative shifts in mpeg_motion_internal
[17:29:54 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:c2f9f8165009: Merge commit '7d4a1ff344cbf969ac648642a0fd8484fd5b8637'
[17:29:55 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:d3e1c6f97502: Merge commit '0242351390643d176b10600c2eb854414f9559e6'
[17:31:56 CEST] <cone-774> ffmpeg 03Luca Barbato 07master:39a2d3288e82: mpegvideo: Refactor emulated_edge_mc calls
[17:31:56 CEST] <michaelni> Daemon404, 3c461eecd48ba2cf7616d98e6f99954de3ad4b06 breaks chapter durations
[17:31:57 CEST] <cone-774> ffmpeg 03Derek Buitenhuis 07master:f8b09e90e92c: Merge commit '39a2d3288e82e4e576c03efb32179ef5a19fff50'
[17:32:18 CEST] <Daemon404> michaelni, breaks how
[17:32:22 CEST] <Daemon404> also we should add a fate test
[17:32:37 CEST] <michaelni> ./ffprobe -i ~/tickets/1833/vorbis_chapter_extension_demo.ogg
[17:32:43 CEST] <michaelni> Chapter #0:3: start 15.000000, end 15.000000
[17:32:56 CEST] <michaelni> was: Chapter #0:3: start 15.000000, end 19.849000
[17:33:01 CEST] <nevcairiel> probably should move the compute_chapters_end to later in the function as well
[17:33:24 CEST] <nevcairiel> since it uses the duration
[17:33:28 CEST] <Daemon404> that would probablt fix it yes
[17:34:18 CEST] <Daemon404> michaelni, should be able to add a fate test with that file, using -show_chapters
[17:35:28 CEST] <michaelni> yes, please add one if you can, ill try to test more
[17:35:45 CEST] <Daemon404> building right now
[17:38:40 CEST] <michaelni> '39a2d3288e82e4e576c03efb32179ef5a19fff50' also changes stream-program assignment in tickets//2441/Pirunpelto_cut.ts
[17:38:57 CEST] <michaelni> Stream #0:0[0x44](fin): Audio: mp2 is no longer in program 1 but "No Program" after it
[17:39:14 CEST] <nevcairiel> that seems odd, why would a avcodec change do that
[17:40:23 CEST] <Daemon404> michaelni, yum did you quote teh wrong hash
[17:40:32 CEST] <Daemon404> 39a2d3288e82e4e576c03efb32179ef5a19fff50 merge is a no-op
[17:41:46 CEST] <michaelni> 3c461eecd48ba2cf7616d98e6f99954de3ad4b06 dunno how i got that other hash
[17:42:21 CEST] <Daemon404> michaelni, we can just revert if you wish
[17:43:18 CEST] <nevcairiel> moving that causes program assignment to change? still seems odd
[17:45:38 CEST] <Daemon404> nevcairiel, ofc
[17:47:14 CEST] <michaelni> i also see changes in subtitles from it
[17:48:26 CEST] <Daemon404> how come none of this was caught by fate
[17:48:50 CEST] <michaelni> ./ffmpeg  -i ~/tickets/1332/Starship_Troopers.vob -vn -an -scodec copy -f framecrc -
[17:48:53 CEST] <Daemon404> seems pretty weird to have all these changes from moving that down
[17:49:02 CEST] <michaelni> i dont get it either
[17:49:09 CEST] <nevcairiel> seems unlikely infact
[17:50:58 CEST] <Daemon404> we have a ton of code in between that libav doesnt
[17:51:03 CEST] <Daemon404> i may have moved it too far down
[17:51:29 CEST] <michaelni> it probably interacts with av_opt_set(ic, "skip_clear", "0", AV_OPT_SEARCH_CHILDREN);
[17:52:15 CEST] <michaelni> i slightly tend toward a revert atm but ive not looked at later commits much yet
[17:52:34 CEST] <Daemon404> most of the later commits are no-ops.
[17:55:54 CEST] <michaelni> ahh, if i commet out the timings entirely subtitles in Starship_Troopers.vob disappear entirely
[17:56:37 CEST] <michaelni> so they get detected during estimate_timings() probably and the loop it was moved over needs the subtitles to be detected
[17:57:07 CEST] <Daemon404> thats seems odd to be... detecting in estimate_timings...
[18:00:00 CEST] <michaelni> also by moving estimate_timings outside skip_clear it will remove streams from programs
[18:00:36 CEST] <michaelni> that is when during timing search it encounters a program table thingy without the stream
[18:01:22 CEST] <michaelni> ATM i think this commit should be reverted, did what it fixed even fail in ffmpeg ?
[18:02:17 CEST] <Daemon404>     avformat_find_stream_info: move duration guessing after updating codec parameters
[18:02:20 CEST] <Daemon404> from the commit message
[18:02:22 CEST] <Daemon404>     This bitrate might not be known otherwise.
[18:02:25 CEST] <Daemon404>     Bug-Id: 926
[18:02:54 CEST] <Daemon404> https://bugzilla.libav.org/show_bug.cgi?id=926
[18:04:17 CEST] <michaelni> where can i find the file mentioned in the bug ? or am i silly and teres a lnk?
[18:04:39 CEST] <Daemon404> "on any mp3"
[18:06:53 CEST] <michaelni> ahh, the bug also says "Note duration is N/A. Now compare the same with ffprobe:" <-- i guess that means it worked fine in ffmpeg
[18:07:08 CEST] <Daemon404> this is pre-codecpar ffmpeg though
[18:07:15 CEST] <Daemon404> i would check just in case
[18:08:36 CEST] <Daemon404> trying..
[18:08:43 CEST] <michaelni> pre codecpar Duration: 00:04:53.30, start: 0.000000, bitrate: 256 kb/s, post codecpar: Duration: 00:04:53.30, start: 0.000000, bitrate: 256 kb/s
[18:08:49 CEST] <michaelni> with random mp3
[18:08:55 CEST] <Daemon404> .  Duration: 00:00:01.04, start: 0.025057, bitrate: 131 kb/s
[18:09:01 CEST] <Daemon404> i reverted the merge
[18:09:04 CEST] <Daemon404> and i still get a duration
[18:09:08 CEST] <Daemon404> (mp3 is from fate)
[18:09:25 CEST] <Daemon404> so it is probably safe to revert
[18:09:31 CEST] <michaelni> yes
[18:09:50 CEST] <Daemon404> ill push it
[18:11:30 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:2691a8399f57: Revert "Merge commit '1ceb07eb313c2d51383408025e57a2fe50ccd164'"
[18:17:01 CEST] <michaelni> thx
[18:26:31 CEST] <cone-248> ffmpeg 03Anton Khirnov 07master:7480d001312d: pixfmt: fix the AV_PIX_FMT_VAAPI_VLD doxy
[18:26:32 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:27558679a1e0: Merge commit '7480d001312d9ba706333ec970264ed9df3f82cb'
[18:27:13 CEST] <cone-248> ffmpeg 03Anton Khirnov 07master:328e9a15c568: buffer: drop a reference to a non-existing function from the docs
[18:27:14 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:c849ed8b2fa2: Merge commit '328e9a15c568843580ff3ff490748d545f16def8'
[18:29:28 CEST] <cone-248> ffmpeg 03Luca Barbato 07master:c11a85862645: configure: Support msan as toolchain
[18:29:29 CEST] <cone-248> ffmpeg 03Luca Barbato 07master:59b9d2f684f1: configure: Add support for clang llvm-cov
[18:29:30 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:6e6b1fe7f714: Merge commit 'c11a8586264520e6afcddc52156f4a1fd2fb07b2'
[18:29:31 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:37f4cdb93780: Merge commit '59b9d2f684f1ff66627ca2b7d2dd05771ade62f0'
[18:31:05 CEST] <cone-248> ffmpeg 03Luca Barbato 07master:7e01d48cfd16: mov: Check the entries value when parsing dref boxes
[18:31:06 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:bdd6275691dc: Merge commit '7e01d48cfd168c3dfc663f03a3b6a98e0ecba328'
[18:31:48 CEST] <cone-248> ffmpeg 03Luca Barbato 07master:92c1a83ee939: qsv: Fix loading multiple plugins
[18:31:49 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:c1828450686c: Merge commit '92c1a83ee9394b39d68f6affd9104752a03714f8'
[18:34:35 CEST] <cone-248> ffmpeg 03Luca Barbato 07master:8b4b1c1eea9d: matroska: Support V_QUICKTIME as written in the specification
[18:34:36 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:e5bb7d98f8c0: Merge commit '8b4b1c1eea9daa4e2003aa0935e73f56aab8102d'
[18:35:55 CEST] <cone-248> ffmpeg 03Sean McGovern 07master:2f4a1bb9bfb2: cmdutils: update copyright year to 2016
[18:35:56 CEST] <cone-248> ffmpeg 03Derek Buitenhuis 07master:89ec4d46eedf: Merge commit '2f4a1bb9bfb29112711ba904e1dc0dd58e24f361'
[19:32:20 CEST] <jamrial> michaelni: could you look at the updated ref files for force_key_frames and gapless-mp3 tests in the framecrc patch?
[19:32:44 CEST] <jamrial> either the one you sent some weeks ago or mine from last night
[19:33:09 CEST] <jamrial> is that new output intended?
[19:35:18 CEST] <michaelni> jamrial, i think  these md5s are taken from framecrc or something data, the change is probably ok but they probably shouldt be md5s that way to begin with
[19:36:02 CEST] <michaelni> that is md5s of "text" from framecrc
[19:36:22 CEST] <michaelni> the size and "PSNR" also suggest that
[19:36:30 CEST] <jamrial> alright
[19:37:19 CEST] <jamrial> just re-ran fate to update the couple tests Daemon404 merged earlier today as well. going to push now
[19:40:28 CEST] <cone-248> ffmpeg 03James Almer 07master:33aa8a62219d: avformat/framecrc: enable new output
[19:43:40 CEST] <cone-248> ffmpeg 03James Almer 07master:5557e881c9e5: avformat/framehash: add extradata checksum
[20:45:06 CEST] <cone-248> ffmpeg 03Lou Logan 07master:fa0f59d55d93: doc/demuxers: fix "Quicktme" typo
[21:41:49 CEST] <durandal_1707> ok to push wmal patch, nobody cares for speedup drop?
[21:44:21 CEST] <nevcairiel> if you add it to floatdsp then someone could make it fast later
[21:44:25 CEST] <nevcairiel> eh losslessdsp
[21:44:26 CEST] <nevcairiel> sorry
[21:44:43 CEST] <nevcairiel> otherwise accurate decoding is more important than speed
[21:58:17 CEST] <durandal_1707> its not perfect, but much better than before
[22:40:57 CEST] <cone-248> ffmpeg 03Michael Niedermayer 07master:3c0511f29e9e: tests/checkasm/vf_colorspace: Make bpp_mask const
[23:02:33 CEST] <cone-248> ffmpeg 03Paul B Mahol 07master:5ac71e9db84c: avcodec/wmalosslessdec: improve >2 channel support
[23:02:34 CEST] <cone-248> ffmpeg 03Paul B Mahol 07master:56759f69a601: avcodec/wmalosslessdec: improve 24bit support
[00:00:00 CEST] --- Thu Apr 14 2016


More information about the Ffmpeg-devel-irc mailing list