[Ffmpeg-devel-irc] ffmpeg-devel.log.20120107

burek burek021 at gmail.com
Sun Jan 8 02:05:03 CET 2012

[00:08] <laurion> how do i find someone that provides ffmpeg free acces to their specifications?
[00:08] <iive> what specification? I want them too.
[00:09] <kierank> i have a lot of specs
[00:09] <kierank> what do you want
[00:11] <iive> everything :)
[00:12] <laurion> ISO, ITU or SMPTE
[00:13] <kierank> why
[00:13] <kierank> just for the sake of it or what
[00:13] <laurion> kierank, are you involved with one of these orgs?
[00:14] <kierank> no
[00:14] <kierank> but i have specs
[00:14] <JEEB> kierank, I think he's trying that GCI task to find people who'd be officially willing to give specs out for free >_>
[00:14] <kierank> oh
[00:14] <kierank> nobody will be
[00:14] <kierank> that's not even a task
[00:14] <laurion> JEEB, right
[00:14] <JEEB> indeed
[00:15] <JEEB> it's there for the laughs or for people to just try I guess
[00:15] <laurion> oh really?
[00:15] <laurion> trolling?
[00:15] <laurion> or what
[00:15] <JEEB> well, not really
[00:15] <JEEB> but it'll be really hard to get it seriously done :P
[00:15] <kierank> the organisations who give out specs for free already do it
[00:15] <JEEB> given the fact that those organizations are getting money off those specs officially
[00:15] <JEEB> yeah
[00:15] <JEEB> and that, too
[00:16] <laurion> can't we fix a bug and take that task instead?
[00:17] <laurion> if it's not going to be done anyway
[00:17] <kierank> pfft it was a bad task anyway
[00:18] <kierank> like asking the bank for free money
[00:19] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r4925b6e784 10ffmpeg/libavformat/tta.c: 
[00:19] <CIA-101> ffmpeg: tta: fix off be 1 error in the end detection.
[00:19] <CIA-101> ffmpeg: Fixes use of uninitialized values.
[00:19] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:19] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r877f6eb5fb 10ffmpeg/libavformat/tta.c: 
[00:19] <CIA-101> ffmpeg: tta: Fix returned error code at EOF
[00:19] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:19] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r5152beeba8 10ffmpeg/libavcodec/mjpeg.h: 
[00:20] <CIA-101> ffmpeg: ljpeg: Add predictor=0 support.
[00:20] <CIA-101> ffmpeg: Fixed first half of ticket889
[00:20] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:20] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * r12d8340f16 10ffmpeg/libavcodec/mjpegdec.c: 
[00:20] <CIA-101> ffmpeg: ljpeg: Fix bits != 8 or 16 support
[00:20] <CIA-101> ffmpeg: Fixes 2nd half of Ticket889
[00:20] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:24] <michaelni> laurion, our bugs arent easy to fix either
[00:25] <ubitux> don't we have typo & spelling tasks?
[00:25] <laurion> michaelni, maybe some crashes,  invalid reads...
[00:25] <ubitux> for code comments, documentation and website
[00:26] <michaelni> ubitux, no
[00:27] <michaelni> that is IIRC
[00:31] <michaelni> whats the oppinion of the other people in here about using the "get all specs" tasks for "close a bug" ?
[00:32] <laurion> michaelni, what is iirc
[00:35] <michaelni> http://www.urbandictionary.com/define.php?term=iirc
[02:10] <OanaStratulat> michaelni: was my 4xm FPE patch pushed ?
[02:21] <michaelni> OanaStratulat, appears not, ill apply it soon
[02:29] <OanaStratulat> michaelni: ok
[02:32] <OanaStratulat> michaelni: how can i see what patches have been applied to the master branch of ffmpeg ? 
[02:32] <OanaStratulat> all the commits, with patches for the decoders/encoders
[02:33] <michaelni> git log -p master
[02:44] <CIA-101> ffmpeg: 03Anssi Hannula 07master * r580bb77936 10ffmpeg/libavformat/spdifenc.c: 
[02:44] <CIA-101> ffmpeg: spdifenc: use special alignment for DTS-HD length_code
[02:44] <CIA-101> ffmpeg: Align IEC 61937 length_code for DTS-HD so that
[02:44] <CIA-101> ffmpeg: (length_code & 0xf) == 0x8. This is reportedly needed with some
[02:44] <CIA-101> ffmpeg: receivers.
[02:44] <CIA-101> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje at gmail.com>
[02:44] <CIA-101> ffmpeg: 03Janne Grunau 07master * rf907615f08 10ffmpeg/libavcodec/ (h264_parser.c mpeg4video_parser.c vc1_parser.c): (log message trimmed)
[02:44] <CIA-101> ffmpeg: parsers: initialize MpegEncContext.slice_context_count to 1
[02:44] <CIA-101> ffmpeg: The mpeg4 video, H264 and VC-1 parser hold (directly or indirectly)
[02:44] <CIA-101> ffmpeg: a MpegEncContext in their private context. Since they do not call the
[02:44] <CIA-101> ffmpeg: common mpegvideo init function slice_context_count has explicitly set
[02:44] <CIA-101> ffmpeg: to 1.
[02:44] <CIA-101> ffmpeg: Prevents a null pointer dereference in the h264 parser and fixes
[02:44] <CIA-101> ffmpeg: 03Justin Ruggles 07master * r6e8bf6db48 10ffmpeg/libavcodec/bytestream.h: add bytestream2_tell() and bytestream2_seek() functions
[02:44] <CIA-101> ffmpeg: 03Diego Biurrun 07master * r301fb92131 10ffmpeg/libavcodec/h264.c: 
[02:44] <CIA-101> ffmpeg: h264: Only use symbols from the SVQ3 decoder under proper conditionals.
[02:45] <CIA-101> ffmpeg: Fixes bug #126.
[02:45] <CIA-101> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles at gmail.com>
[02:45] <CIA-101> ffmpeg: 03Diego Biurrun 07master * r3aa3fc45fe 10ffmpeg/libavcodec/cabac.h: cabac: remove unused function renorm_cabac_decoder
[02:58] <CIA-101> ffmpeg: 03Oana Stratulat 07master * r44cc1936be 10ffmpeg/ffmpeg.c: 
[02:58] <CIA-101> ffmpeg: ffmpeg: Fix segfault with zzufed 4xm file.
[02:58] <CIA-101> ffmpeg: Fixes ticket 885
[02:58] <CIA-101> ffmpeg: Signed-off-by: Oana Stratulat <oanaandreeastratulat at gmail.com>
[02:58] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:09] <CIA-101> ffmpeg: 03Stefano Sabatini 07master * rceef1ee767 10ffmpeg/ (avconv.c ffmpeg.c ffplay.c ffprobe.c): 
[11:09] <CIA-101> ffmpeg: cmdutils: make show_usage() use av_log()
[11:09] <CIA-101> ffmpeg: Avoid printing on stdout when show_usage is used in an error message.
[11:09] <CIA-101> ffmpeg: 03Stefano Sabatini 07master * r9ecccd6e5a 10ffmpeg/ffprobe.c: 
[11:09] <CIA-101> ffmpeg: ffprobe: move header and trailer print from probe_file() to main()
[11:09] <CIA-101> ffmpeg: Simplify printing from the main() context, in case probe_file() is not
[11:09] <CIA-101> ffmpeg: called, as required by the pending -show_version option.
[11:09] <CIA-101> ffmpeg: 03Stefano Sabatini 07master * re869d08cbc 10ffmpeg/ (cmdutils.c cmdutils.h): 
[11:09] <CIA-101> ffmpeg: cmdutils: make this_year extern, so it can be referenced from other .o files
[11:09] <CIA-101> ffmpeg: Required by a pending change in ffprobe.
[11:09] <CIA-101> ffmpeg: 03Stefano Sabatini 07master * rfa7d1c39fd 10ffmpeg/ffprobe.c: 
[11:09] <CIA-101> ffmpeg: ffprobe: move writer context registration and initialization in main()
[11:09] <CIA-101> ffmpeg: Simplify pending changes, as the writer context will be used in the
[11:09] <CIA-101> ffmpeg: main() routine.
[13:14] <CIA-101> ffmpeg: 03Carl Eugen Hoyos 07master * ree884c3023 10ffmpeg/libavformat/idcin.c: Fix OOM error condition in idcin demuxer.
[14:31] Action: cbsrobot likes how quicktime 7 crashes when deleting 2 timecode tracks
[14:32] <merbanan> you are pushing it to hard
[14:32] Action: cbsrobot likes how quicktime 7 crashes even deleting one timecode track
[14:34] Action: cbsrobot is sure quicktime 7 even crashes just staring at a timecode track long enough 
[14:38] <cbsrobot> ha - 3 minutes staring at it is enough
[16:08] <ubitux> is there anyone on windows to make a test for me please?
[16:09] <ubitux> i'd like someone to apply this:
[16:09] <ubitux> -static const uint8_t color[] = { 12, 12, 12, 14, 7, 7, 7 };
[16:09] <ubitux> +static const uint8_t color[] = { 12, 12, 12, 14, 7, 7, 10 };
[16:09] <ubitux> in libavutil/log.c
[16:10] <ubitux> and make me a screenshot of a misc decode with -debug 1
[16:10] <ubitux> just to confirm it's indeed green, and not unreadable
[16:26] <OanaStratulat> ubitux: michaelni ping
[16:26] <ubitux> yup
[16:26] <OanaStratulat> ubitux: ptp-roman.org/ffmpeg_crashes/dude
[16:26] <OanaStratulat> does it crash for you?
[16:27] <ubitux> yep
[16:27] <ubitux> bug in the fraps decoder
[16:28] <OanaStratulat> ubitux: in fraps2_decode_plane no?
[16:29] <ubitux> seems to be in decode_frame according to valgrind but i didn't look closely
[16:30] <OanaStratulat> my valgrind is http://pastebin.com/9tsunaW3 
[16:30] <OanaStratulat> but it says in fraps2_decode_plane
[16:31] <ubitux> it's strange how clueless is valgrind in your output
[16:31] <ubitux> http://pastie.org/3143669
[16:31] <ubitux> this is my output
[16:32] -OanaStratulat:#ffmpeg-devel- Cannot pull with rebase: You have unstaged changes.
[16:32] -OanaStratulat:#ffmpeg-devel- Please commit or stash them.
[16:32] <OanaStratulat> ubitux: when i want to do git pull --rebase i get ^
[16:33] <ubitux> you need to commit your changes in the current branch (or another), or trash them
[16:33] <nevcairiel> or stash them. :)
[16:33] <OanaStratulat> how to trash them?
[16:33] <ubitux> git checkout themodifiedfile.c
[16:33] <ubitux> or git reset --hard
[16:34] <ubitux> (you won't get your changes back)
[16:35] <OanaStratulat> thx
[16:38] <OanaStratulat> ubitux: now i have remade with the latest changes and this is my valgrind http://pastebin.com/usGdQtK3
[16:38] <OanaStratulat> what valgrind version do you have
[16:39] <ubitux> 3.7.0
[16:39] <OanaStratulat> i ahve the same
[16:39] <OanaStratulat> and the latest ffmeg src files
[16:39] <ubitux> you are using ffmpeg in that latest paste
[16:40] <OanaStratulat> yes
[16:40] <OanaStratulat> maybe this is
[16:40] <ubitux> instead of the non-stripped _g one
[16:40] <ubitux> but your first paste seems to be with ffmpeg_g
[16:41] <ubitux> maybe the osx valgrind is just crappy, no idea
[16:41] <OanaStratulat> yes...i think this is
[16:41] <OanaStratulat> you are in gmt+1 ?
[16:43] <ubitux> sth like that
[16:43] <ubitux> why?
[16:44] <OanaStratulat> will you have time 2night to discuss about this bug?
[16:44] <OanaStratulat> or are you busy/
[16:45] <ubitux> feel free to ask here any related question at anytime
[16:45] <ubitux> i generally read the irc backlog anyway
[16:45] <OanaStratulat> i`m leaving now and that`s why
[16:45] <OanaStratulat> anyway speak 2night if you are here
[16:45] <OanaStratulat> bye everyone
[18:43] <durandal_1707> there is spammer on trac
[18:44] <durandal_1707> http://ffmpeg.org/trac/ffmpeg/wiki/WikiStart?action=diff&version=17
[18:44] <durandal_1707> j
[19:10] <michaelni> durandal_1707, spammer deleted, please fix the wiki if you didnt already, thx
[19:23] <CIA-101> ffmpeg: 03Carl Eugen Hoyos 07master * r19b945dc2a 10ffmpeg/libavcodec/pnm.c: 
[19:23] <CIA-101> ffmpeg: Cosmetics: Fix indentation.
[19:23] <CIA-101> ffmpeg: Found by: Alexander Strasser
[19:40] <cbsrobot> ubitux ping
[19:45] <ubitux> cbsrobot: pong
[19:47] <cbsrobot> ubitux you wrote the timecode decoding for mov lately
[19:48] <cbsrobot> I guess you only pass the drop frame flag to the avcodeccontext - right ?
[19:50] <cbsrobot> would it be useful to also pas the "24 hour", "negative times"  and "counter" to the context ?
[19:50] <ubitux> you mean in the codec flags2?
[19:50] <cbsrobot> yes
[19:51] <cbsrobot> I ask because I write the encoding part atm
[19:52] <ubitux> i guess it could be added, but why do you need it for the encoding part?
[19:53] <cbsrobot> well to support encoding timecode tracks ...
[19:54] <ubitux> the only reason i'd see it necessary would be for transcoding
[19:55] <cbsrobot> you know the "-dcodec copy" issue on track
[19:55] <ubitux> but if you want to add the writing support, adding -timecode opt doesn't require it
[19:55] <ubitux> ok, so in order to keep the data in a transcode
[19:55] <cbsrobot> yeah
[19:56] <ubitux> timecode transcode is not yet supported indeed
[19:56] <cbsrobot>  I guess the context would get really big, if we support every possible aspect of the mov container ...
[19:56] <cbsrobot> I'm wondering how much is really necessary
[20:55] <CIA-101> ffmpeg: 03Michael Niedermayer 07master * rc8e309c2a6 10ffmpeg/libavformat/utils.c: 
[20:55] <CIA-101> ffmpeg: lavf: Do not attempt to use frame multi threading when probing.
[20:55] <CIA-101> ffmpeg: This fixes various problems with getting stream info. For example playback of the
[20:55] <CIA-101> ffmpeg: file of Ticket88. Multithreaded find_stream_info should be reenabled
[20:55] <CIA-101> ffmpeg: once it works correctly
[20:55] <CIA-101> ffmpeg: This partly reverts 212fd3a1f1121a9db20c478f1fccf15e8bb53c94
[20:55] <CIA-101> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:04] <CIA-101> ffmpeg: 03Stefano Sabatini 07master * r9997d41672 10ffmpeg/ (Changelog doc/ffprobe.texi doc/ffprobe.xsd ffprobe.c): 
[23:04] <CIA-101> ffmpeg: ffprobe: add support to video frame information printing
[23:04] <CIA-101> ffmpeg: Add -show_frames option to ffprobe.
[23:04] <CIA-101> ffmpeg: Partially based on the work of Thomas Kuehnel <kuehnelth at googlemail.com>
[23:04] <CIA-101> ffmpeg: for SOCIS 2011.
[23:04] <CIA-101> ffmpeg: The wicked idea of creating a special "packets_and_frames" container for
[23:04] <CIA-101> ffmpeg: structured formats (JSON and XML) comes from Clément.
[23:04] <CIA-101> ffmpeg: 03Stefano Sabatini 07master * r78c47e0a1e 10ffmpeg/doc/filters.texi: doc/filters.texi: fix typo in volume description
[23:09] <ubitux> :)
[00:00] --- Sun Jan  8 2012

More information about the Ffmpeg-devel-irc mailing list