[Ffmpeg-devel-irc] ffmpeg-devel.log.20120424

[burek]

[02:43] <CIA-17> ffmpeg: 03Martin Storsjö 07master * r269cb6751b 10ffmpeg/libavformat/mpegts.c: (log message trimmed)
[02:43] <CIA-17> ffmpeg: mpegts: Try seeking back even for nonseekable protocols
[02:43] <CIA-17> ffmpeg: The mpegts demuxer reads 5 KB at startup just for discovering
[02:43] <CIA-17> ffmpeg: the packet size. Since the default avio buffer size is 32 KB,
[02:43] <CIA-17> ffmpeg: the seek back to the start will in most cases be within the
[02:43] <CIA-17> ffmpeg: avio buffer, and will in most cases succeed even if the actual
[02:43] <CIA-17> ffmpeg: protocol isn't seekable.
[02:43] <CIA-17> ffmpeg: 03Diego Biurrun 07master * r9b75ae05cf 10ffmpeg/libswscale/input.c: swscale: K&R formatting cosmetics (part III)
[02:43] <CIA-17> ffmpeg: 03Dale Curtis 07master * r8336eb6f85 10ffmpeg/ (libavformat/matroskadec.c tests/ref/seek/lavf_mkv): (log message trimmed)
[02:43] <CIA-17> ffmpeg: matroska: Add incremental parsing of clusters.
[02:43] <CIA-17> ffmpeg: Reduces the amount of upfront data required for cluster parsing
[02:43] <CIA-17> ffmpeg: thus decreasing latency on seek and startup.
[02:43] <CIA-17> ffmpeg: The change in the seek-lavf_mkv FATE test is due to incremental
[02:43] <CIA-17> ffmpeg: parsing no longer reading as much data as the old parser and
[02:43] <CIA-17> ffmpeg: thus not having that additional data to generate index entries
[02:43] <CIA-17> ffmpeg: 03Diego Biurrun 07master * ra229d6c285 10ffmpeg/ (7 files in 3 dirs): 
[02:43] <CIA-17> ffmpeg: Remove libdirac decoder.
[02:43] <CIA-17> ffmpeg: libschroedinger is the preferred way to decode Dirac video.
[02:43] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r0ca4414d0f 10ffmpeg/libavformat/au.c: 
[02:43] <CIA-17> ffmpeg: padding for SIMD-optimized functions.
[02:44] <CIA-17> ffmpeg: 03Luca Barbato 07master * r1381081cdb 10ffmpeg/avconv.c: 
[02:44] <CIA-17> ffmpeg: avconv: fix off by one check in complex_filter
[02:44] <CIA-17> ffmpeg: nb_input_files is already an invalid index.
[02:44] <CIA-17> ffmpeg: 03Dale Curtis 07master * r7521c4bab2 10ffmpeg/libavformat/matroskadec.c: (log message trimmed)
[02:44] <CIA-17> ffmpeg: matroska: Clear prev_pkt between seeks.
[02:44] <CIA-17> ffmpeg: The new incremental parser doesn't always clear prev_pkt,
[02:44] <CIA-17> ffmpeg: however the packet queue is cleared when seeking. Which leads
[02:44] <CIA-17> ffmpeg: to a use-after-free.
[02:44] <CIA-17> (2 lines omitted)
[11:13] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r9c99bc433e 10ffmpeg/libavcodec/mpc8.c: 
[11:13] <CIA-17> ffmpeg: mpc8: print error messages on maxband(s) errors
[11:13] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[11:14] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r88ee2aa5ad 10ffmpeg/libavcodec/mpc8.c: 
[11:14] <CIA-17> ffmpeg: mpc8: make maxband check less picky.
[11:14] <CIA-17> ffmpeg: Fixes Ticket1245
[11:14] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:20] <mkoch> hi! I just came form #ffmpeg. I'm building a plugin which uses ffmpeg as a lib to process video streams. can I make ffmpeg to use vda on Mac? 
[12:21] <mkoch> http://ffmpeg.org/doxygen/trunk/vda_8h-source.html <-- I found this
[13:04] <mkoch> is that possible that this VDA works in 32bit mode only?
[13:06] <Compn> sorry mkoch , i dunno about vda 
[13:10] <michaelni> mkoch, dilaroga is our vda expert, he isnt here though ATM ...
[13:13] <mkoch> thanks, I try to wait for him
[14:00] <funman> mkoch: why do you think it is 32 bits only ?
[14:01] <funman> http://dilaroga.free.fr/vlc_vda/vlc-2.1.0-git.dmg is built with VDA
[15:59] <CIA-17> ffmpeg: 03Nicolas George 07master * r62b745a6d3 10ffmpeg/ffmpeg.c: 
[15:59] <CIA-17> ffmpeg: ffmpeg: do not report EOF as an error.
[15:59] <CIA-17> ffmpeg: Not all processing goes through filters,
[15:59] <CIA-17> ffmpeg: so processing can continue even when
[15:59] <CIA-17> ffmpeg: all filtergraphs have reached EOF on all sinks.
[16:45] <mkoch> funman: it was my fault, I missed to link a lib
[16:45] <mkoch> now it seems to work, but I still have no proof that it's using VDA actually
[16:45] <mkoch> how can I find it out in runtime?
[16:47] <CIA-17> ffmpeg: 03Nicolas George 07master * rfd2cd64fc9 10ffmpeg/libavfilter/avfiltergraph.c: (log message trimmed)
[16:47] <CIA-17> ffmpeg: avfiltergraph: fix format selection.
[16:47] <CIA-17> ffmpeg: The old code had two bugs:
[16:47] <CIA-17> ffmpeg: For audio filters, the format was not set.
[16:47] <CIA-17> ffmpeg: For video filters, if several links reference the same format list,
[16:47] <CIA-17> ffmpeg: the same format must be selected in the end. This is done by
[16:47] <CIA-17> ffmpeg: setting formats->format_count to 1: the other links sharing
[17:06] <michaelni> mkoch, put a av_log or abort() or something in the vda code, if you see the effects of that its being executed beyond doubt
[17:24] <CIA-17> ffmpeg: 03Carl Eugen Hoyos 07master * rae0591f913 10ffmpeg/libavcodec/shorten.c: Add missing newline in shorten decoder.
[19:07] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r44bd2f90a9 10ffmpeg/ffmpeg.c: 
[19:07] <CIA-17> ffmpeg: ffmpeg: Fix initial dts for streamcopy, if the input packets lack dts.
[19:07] <CIA-17> ffmpeg: Fixes Ticket1138
[19:07] <CIA-17> ffmpeg: Based on a patch by: Eelco Lempsink <eml at tupil.com>
[19:07] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:58] <funman> http://mailman.videolan.org/pipermail/vlc-devel/2012-April/088604.html
[20:04] <j-b> this is quite out of topic, if I may say.
[20:07] <Daemon404> wtf
[20:08] <funman> j-b: well VLC is still the biggest distributor of FFmpeg
[20:08] <funman> or perhaps it's Chrome?
[20:08] Action: Daemon404 senses some dirty laundry is being aired in public, but noone has any idea whats going on
[20:09] <ubitux> nice drama, up to ffmpeg/libav/mplayer* community :)
[20:13] <j-b> funman: I believe it is Chrome.
[20:13] <j-b> funman: even if they use libav, and strip it down quite a bit.
[20:13] <Daemon404> they still use ffmpeg iirc.
[20:14] <ubitux> so maybe even BBB at google is using ffmpeg without admitting it
[20:14] <ubitux> :°
[20:15] <ubitux> i wonder if steam is not one of the biggest distributor too ;)
[20:15] <Daemon404> since when does steam distribute it?
[20:15] <ubitux> ./Program Files/Steam/bin/avformat-52.dll
[20:15] <ubitux> ./Program Files/Steam/bin/avutil-50.dll
[20:15] <ubitux> ./Program Files/Steam/bin/avcodec-52.dll
[20:16] <ubitux> quite a while :p
[20:16] <Daemon404> interesting.
[20:17] <JEEB> hah, never noticed that either
[20:17] <ubitux> iive mentionned it a few months ago
[20:17] <Daemon404> and id imagine more people have chrome than steam.
[20:47] <Compn> j-b : offtopic flames are always on topic here ;D
[20:48] <Compn> and wow thats a strange looking mail :D
[21:15] <Compn> funman : so are you and remi going pistols at dawn ?
[21:15] Action: Compn is always in the mood for a duel
[21:15] <funman> would be nice
[21:15] <Compn> 10 paces ?
[21:15] <funman> one last night to write our final maths theorems
[21:15] <Compn> blindfolded.
[21:16] <funman> I chose water pistols
[21:16] <Compn> bah 
[21:17] <funman> but I'll take a high pressure kärsher
[21:18] <Compn> naked, on the cold tundra ?
[21:19] Action: Compn should write a movie or something
[21:24] <Compn> funman : btw your link in that mail requires a login
[21:24] <Compn> :P
[21:24] <Compn> the irc logs i think
[21:25] <funman> he just removed it
[21:25] <funman> (the login)
[21:25] <Compn> Resolving dionoea.chewa.net... 91.121.6.101
[21:25] <Compn> Connecting to dionoea.chewa.net|91.121.6.101|:80... connected.
[21:26] <Compn> HTTP request sent, awaiting response... 401 Authorization Required
[21:26] <Compn> Authorization failed.
[21:26] <funman> ah it was here for a few seconds
[21:27] <funman> http://dionoea.chewa.net/irc/ yep it's all deleted , must have been my browser's cache acting
[21:28] <Compn> that remi's server ?
[21:28] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r2b336df3cb 10ffmpeg/libavutil/parseutils.c: 
[21:28] <CIA-17> ffmpeg: parseutils: reliably detect out-of-range alpha.
[21:28] <CIA-17> ffmpeg: This should fix the FATE test on ARM (not tested),
[21:28] <CIA-17> ffmpeg: but it should also detect alpha values like 2^128
[21:28] <CIA-17> ffmpeg: reliably as invalid which would be another out-of-range
[21:28] <CIA-17> ffmpeg: case with implementation-dependant behaviour.
[21:28] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger at gmx.de>
[21:34] <funman> yeah it's shared between a bunch of ex ECP students
[21:40] <Compn> whoa
[21:40] <Compn> http://mailman.videolan.org/pipermail/vlc-devel/2012-April/ is all 404
[21:41] <Compn> mailman.videolan.org that is
[21:41] <j-b> yep
[21:41] <Compn> j-b : you guys having fun over there ?
[21:41] <j-b> Compn: not really.
[21:42] <iive> looks like this is contagious. Mplayer infected FFmpeg, and now it spreads 
[21:42] <Compn> wait how can you blame mplayer for this ? :P
[21:42] <Compn> oh wait, nm i remember 
[21:42] <funman> no it's quite far from being fun
[21:43] <iive> it is never fun when it is in your house.
[21:43] <j-b> Compn: I dislike people insulting each other.
[21:45] <Compn> j-b : well i wish it would all work out for everyone.
[21:45] <j-b> Compn: it won't.
[21:45] <j-b> Compn: cause at the end of this, VLC dies.
[21:45] <Compn> iive : i think we should blame google instead :P
[21:46] <Compn> or fsf :P
[21:46] <iive> why?
[21:46] <j-b> Compn: cause I will quit the VideoLAN project, and believe me the project will die slowly.
[21:46] <ubitux> huh?
[21:46] <Compn> j-b : yeah, i know you work pretty hard at it :(
[21:46] <ubitux> why would you do that?
[21:47] <j-b> ubitux: just to annoy you and your cat :)
[21:47] <iive> ubitux, you have a cat?
[21:47] <ubitux> :D
[21:47] <ubitux> no i don't iive :)
[21:47] <Compn> videolan has really owned the video player market for far too long
[21:47] <j-b> Compn: that is totally untrue.
[21:47] <Compn> its too easy to use :)
[21:47] <ubitux> i have a logitech mouse though
[21:48] <Compn> ubitux : how big is your /usr/bin/cat ?
[21:48] <Compn> 1/11/1999  04:00a              22,016 cat.exe
[21:48] <ubitux> HUGEZ
[21:49] <Compn> cat (GNU textutils) 2.0
[21:51] <Compn> Daemon404 : you think more people have chrome vs steam ?
[21:51] <Compn> he concurrent users peak was 5 million on January 2, 2012. (steam)
[21:54] <Compn> Google Chrome Browser Has 200 Million Users
[21:54] <Compn> whoa
[21:54] <Compn> guess so ;)
[21:54] <j-b> We are at around 100M
[21:55] Action: gnafu uses VLC, doesn't use Chrome or Steam.
[22:33] <VooDooNOFX_> To followup my earlier conversation, r3d appears to now have a version 2 of their file format with new cameras.
[22:46] <Compn> VooDooNOFX_ : well samples and specs and patches welcome :)
[23:49] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r3158160334 10ffmpeg/libavcodec/ffv1.c: 
[23:49] <CIA-17> ffmpeg: ffv1: move the decoder side ac->vlc switch to decode_slice()
[23:49] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:49] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * re37419a806 10ffmpeg/libavcodec/ffv1.c: 
[23:49] <CIA-17> ffmpeg: ffv1: move ac->vlc switch to encode_slice()
[23:49] <CIA-17> ffmpeg: This will allow us to use vlc coding in 1.3 too
[23:49] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:49] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r44f789c092 10ffmpeg/libavcodec/ffv1.c: 
[23:49] <CIA-17> ffmpeg: ffv1: allow selecting 1.3 by using -level
[23:49] <CIA-17> ffmpeg: Note, 1.3 is not finalized and the bitstream will still change
[23:49] <CIA-17> ffmpeg: do not use it yet. This option is just to make playing with it
[23:49] <CIA-17> ffmpeg: easier, otherwise one would have to edit the source
[23:49] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:49] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rfe5bc46f60 10ffmpeg/libavcodec/ffv1.c: 
[23:49] <CIA-17> ffmpeg: ffv1: support vlc mode in 1.3
[23:49] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:00] --- Wed Apr 25 2012