[Ffmpeg-devel-irc] ffmpeg-devel.log.20121230

[burek]

[00:22] <ubitux> btw, reviews welcome :p
[00:49] <michaelni> ubitux, i have no oppinion on its removial because i dont know if someone uses/used it
[00:49] <ubitux> and about the question?
[02:24] <michaelni> ubitux, if the code that sets LIBPOSTPROC_VERSION_INT is droped then si can the if be droped i thinlk
[02:33] <cone-628> ffmpeg.git 03Clément BSsch 07master:518239cd6126: build: postproc cleanup.
[05:29] <cone-628> ffmpeg.git 03Michael Niedermayer 07master:6d38661920ec: aacdec: fix dual mono
[05:29] <cone-628> ffmpeg.git 03Michael Niedermayer 07master:644f021ccf4b: aacdec: simplify dmono
[05:29] <cone-628> ffmpeg.git 03Michael Niedermayer 07master:b113d4a83c82: aacdec: make dual mono mode selectable through AVOptions too.
[06:00] <ubitux> does anyone what the ass packet merge & split is for?
[06:00] <ubitux> +know
[06:01] <ubitux> it seems the ass lines are merged when timing are the same (with the matroska demuxer)
[06:01] <ubitux> then the ass decoder split them
[06:02] <ubitux> (the merge is done when pts are the same)
[06:24] <ubitux> wow, this ASS stuff is an incredible mess :)
[06:58] <cone-628> ffmpeg.git 03Clément BSsch 07master:c83002a4f804: lavc/ass_split: check for NULL pointer in ff_ass_split_override_codes().
[07:02] <ubitux> #2079 haha what :D
[07:03] <ubitux> Diego, breaking stuff with cosmetics since 2004
[07:18] <cone-628> ffmpeg.git 03Clément BSsch 07master:613001d75fee: Merge back some registering macros in all*.c.
[11:00] <nevcairiel> ubitux: that reminds me (re mov chapters), i fixed that bug but the patch got lost in my backlog so i never submitted it (and i never tested if tmcd was still parsed properly)
[11:00] <nevcairiel> feel free to steal: http://git.1f0.de/gitweb?p=ffmpeg.git;a=commitdiff;h=29959d27e73a1ec8a076e372f60a179704a82d2d
[11:11] <ubitux> mmh
[11:11] <ubitux> didn't we talk about such patch a long time ago on ffmpeg-devel?
[11:15] <nevcairiel> maybe
[11:16] <ubitux> do you have other secret patches like this? :)
[11:17] <nevcairiel> plenty, i need to take some time to sort, clean and submit again
[11:17] <ubitux> !
[11:30] <durandal_1707> fate no red !!! what changed?
[11:31] <ubitux> warning fixes?
[11:32] <durandal_1707> the one with strict aliasing rules?
[11:36] <ubitux> no idea
[11:36] <ubitux> what was red?
[11:38] <durandal_1707> for years there were machines that would never compile
[11:43] <saste> did they remove those machines?
[13:57] <wm4> is there a way to know whether a frame contains alpha?
[13:57] <wm4> the pixel formats don't always uniquely identify this
[13:58] <durandal_1707> pixel formats that have alpha component
[13:59] <wm4> as I said, that's not always definitive
[13:59] <wm4> most packed RGB formats have an implicit alpha component, that may or may not be used
[14:00] <durandal_1707> in such cases there are rgb0 like formats
[14:00] <wm4> these are new and rarely used
[14:01] <wm4> and none of such for formats like RGB444
[14:01] <durandal_1707> rgb444 does not have alpha
[14:01] <cone-681> ffmpeg.git 03Luca Barbato 07master:d8fd06c37de9: avstring: add av_basename and av_dirname
[14:01] <cone-681> ffmpeg.git 03Luca Barbato 07master:66f7b4862fe9: hlsenc: use the basename to generate the list entries
[14:01] <cone-681> ffmpeg.git 03Luca Barbato 07master:6dd93ee6f1b0: hlsenc: check append_entry return value
[14:01] <cone-681> ffmpeg.git 03Michael Niedermayer 07master:2ce43b37fc46: Merge commit '6dd93ee6f1b050ad7c4b247899e83efa293ee405'
[14:01] <durandal_1707> if it would, then it would be called rgba4444
[14:02] <wm4> and Libav doesn't even have the rgb0 formats
[14:03] <durandal_1707> your point?
[14:05] <durandal_1707> last time they laught on rgb0 and even made jokes of it in g2m "patch"
[14:08] <cone-681> ffmpeg.git 03Luca Barbato 07master:ae85d6c9c0dc: hlsenc: use the segment filename in the playlist entry
[14:08] <cone-681> ffmpeg.git 03Luca Barbato 07master:0448f26c97c5: hlsenc: keep the playlist to the correct number of items
[14:08] <cone-681> ffmpeg.git 03Michael Niedermayer 07master:19e1c8e4f203: Merge commit '0448f26c97c5ab4858d31e456a4f1738ae783242'
[14:21] <cone-681> ffmpeg.git 03Kanglin 07master:0d8cc7a3b223: hlsenc: use the correct AV_TIME_BASE macro
[14:21] <cone-681> ffmpeg.git 03Luca Barbato 07master:9b1370aced38: hlsenc: do not add timestamps in different timebases
[14:21] <cone-681> ffmpeg.git 03Michael Niedermayer 07master:56c71cd59986: Merge commit '9b1370aced385698bc783747917544ab69ecb373'
[14:27] <cone-681> ffmpeg.git 03Kanglin 07master:27a15e0af6fb: hlsenc: make EXT-X-MEDIA-SEQUENCE always increase
[14:28] <cone-681> ffmpeg.git 03Luca Barbato 07master:30a76487304e: hlsenc: make segment number unsigned
[14:28] <cone-681> ffmpeg.git 03Michael Niedermayer 07master:384dd9804d63: Merge commit '30a76487304e7250294c9c0e9fa179bf07fd822a'
[14:41] <saste> ubitux, how many subtitle formats in your todo list yet?
[14:51] <cone-681> ffmpeg.git 03Luca Barbato 07master:d615da60c1f9: snow: use VideoDSPContext
[14:51] <cone-681> ffmpeg.git 03Michael Niedermayer 07master:cbed2426a029: Merge remote-tracking branch 'qatar/master'
[15:21] <cone-681> ffmpeg.git 03Stefano Sabatini 07master:9e3b6285cbe8: doc/filters: drop duplicated description of resampler options in aresample docs
[15:21] <cone-681> ffmpeg.git 03Mark Himsley 07master:8997a0fa79cc: lavfi/tinterlace: add low-pass-filter for top/bottom interleave modes
[17:03] <cone-681> ffmpeg.git 03Michael Niedermayer 07master:c0c0b19644eb: lavfi: remove some draw_slice related code that has become unneeded
[17:03] <cone-681> ffmpeg.git 03Michael Niedermayer 07master:b84871b8eba5: vf_tinterlace: dont mix declarations and expressions
[17:13] <durandal_1707> ubitux: when you gonna push subtitles?
[18:09] <burek> one song dedicated to ubitux: Queen - Under Pressure :)
[18:29] <cone-681> ffmpeg.git 03Don Moir 07master:522cb6abf22a: gifdec: use transparent instead of background color index
[18:41] <cone-681> ffmpeg.git 03Nicolas George 07master:a978c04e228e: lavfi/video: ensure that filter_frame is called.
[18:41] <cone-681> ffmpeg.git 03Nicolas George 07master:c36302a7a44b: lavfi/sink_buffer: switch to filter_frame.
[18:47] <durandal_1707> there are two twinvq variants?
[18:51] <nevcairiel> i found a ogg vorbis file which plays with libav but fails with ffmpeg, the actual vorbisdec.c is exactly identical short of some minor cosmetics, what would be the next thing to check? :d
[18:52] <durandal_1707> demuxer obviously
[18:59] <beastd> nevcairiel: that would be lavf/oggdec and i think there are differences between the fork and ffmpeg. so investigation is welcome. if you do not intend to work on it, would be good to file a bug report so it is not forgotten.
[19:00] <durandal_1707> is there any mpeg twinvq sample? twvq fourcc in CAF
[19:01] <nevcairiel> there are quite big differences between the two oggdecs
[19:02] <durandal_1707> there was some chained thing added...
[19:19] <durandal_1707> nevcairiel: you have more such samples in your queue?
[19:20] <nevcairiel> its a recording from a specific streaming audio service, so i'm sure i could create more =p
[19:21] <nevcairiel> i'll keep an eye on it and post a bug report if i dont find anything myself
[19:21] <nevcairiel> i cant really say i know anything about ogg...
[19:22] <nevcairiel> should be a bisect'able error
[19:33] <durandal_1707> nevcairiel: i'meant any sample, not just related to this bug
[20:39] <ubitux> saste: i'm done
[20:39] <ubitux> durandal_1707: when i get reviews :p
[20:39] <ubitux> i may not wait too much though
[20:43] <ubitux> wow i got reviews
[20:43] <ubitux> (just woke up)
[20:43] <durandal_1707> your are blocking my patches
[20:44] <ubitux> i'm going to push some stuff in the next 2 hours
[20:45] <saste> durandal_1707, how it is blocking your patches?
[20:45] <saste> conflicts most happen in Changelogs/version files
[20:45] <saste> that's why i just update them before I push
[20:46] <beastd> ubitux: Should we add sth like "when in doubt use the concat filter" to the concat demuxer faq? i mean for cases of say codec/codec-settings mismatch and similar.
[20:47] <ubitux> beastd: concat filter is the safer yes
[20:47] <ubitux> not sure how to word it
[20:50] <ubitux> avfilter_get_audio_buffer_ref_from_arrays_nolayout() haha
[20:50] <wm4> wat, there's a concat filter and demuxer?
[20:50] <ubitux> wm4: yes
[20:50] <ubitux> wm4: https://ffmpeg.org/faq.html#How-can-I-concatenate-video-files_003f
[20:51] <beastd> maybe "If you are not absolutely sure about the streams contained in the files and their comaptibility use the concat filter."
[20:52] <beastd> I find it hard to word either. The concat muxer is a great tool but it is also kind of dangerous and should not be used without caution IMHO.
[20:55] <durandal_1707> show big red warning in caps?
[20:56] <saste> why is the demuxer not able to check it for itself?
[20:56] <beastd> saste: read reply from nicolas
[20:58] <beastd> saste: this one http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/156674/focus=156695
[21:20] <llogan> durandal_1707: will the g+ page have ffmpeg news or something? how/if is it related to news?
[21:23] <durandal_1707> if you want manage page/community you can add whatever you like
[21:24] <durandal_1707> is there some better place for badge, i'm listening
[21:24] <durandal_1707> i just put there where rss icon is....
[21:24] <llogan> ah, ok. i'm just not familiar with g+
[21:25] <llogan> if we get more of these badge things we can put them on the right side of "Project Description"
[21:26] <durandal_1707> aha
[21:29] <llogan> durandal_1707: are you the FFmpeg g+ "maintainer"
[21:30] <durandal_1707> yes, with others under way
[21:32] <durandal_1707> there are already 3 moderators for FFmpeg community
[21:36] <durandal_1707> i michaelni can push it after there is some more text on page, next news entry on main web page
[21:37] <llogan> should be soon from beastd
[21:40] <cone-681> ffmpeg.git 03Clément BSsch 07master:d927d8395d71: lavc/srtdec: make some sscanf to work at the end of a line.
[21:40] <cone-681> ffmpeg.git 03Clément BSsch 07master:3af3a3006f39: lavf/srtdec: remove line break hack.
[21:40] <cone-681> ffmpeg.git 03Clément BSsch 07master:1b673cd9ff32: lavc/srtenc: do not add trailing line breaks with SubRip.
[21:40] <cone-681> ffmpeg.git 03Clément BSsch 07master:e911f4ae720a: lavc/srtenc: just like the decoder, mark SRT encoder as deprecated.
[21:46] <ubitux> durandal_1707: feel free to push, i have to deal with the all*.c cosmetic mess anyway
[21:47] <durandal_1707> i can wait
[22:12] <beastd> No new comments on ffmpeg-devel. Will push the news soon with bold concat demuxer warning included. I mentioned it would also be possible to drop that part completely (e.g. post news about concat demuxer when it is a bit safer) but got no more comments since.
[22:13] <Compnn> i'd test it if the ffmpeg autobulds were working :P
[22:16] <beastd> Compnn: it?
[22:21] <Compnn> concat
[22:24] <cone-681> ffmpeg.git 03Clément BSsch 07master:36e61e24e7ac: lavc: add ff_bprint_to_extradata() helper and use it.
[22:24] <cone-681> ffmpeg.git 03Clément BSsch 07master:9db5f82032b1: ffserver: add missing padding to extradata.
[22:24] <cone-681> ffmpeg.git 03Clément BSsch 07master:2d3305e4f41f: lavc/avcodec: fix FF_INPUT_/BUFFER_/PADDING_SIZE typo.
[22:25] <beastd> Compnn: You want to try out creating broken files :D
[22:26] <ubitux> michaelni: the aliasing issues should be fixed now; if not, please tell me :p
[22:33] <beastd> ubitux: What is it with the trailing null in extradata? Who needs it? Did you answer Nicolas mail on that patch?
[22:35] <ubitux> yes i answered, and commented in the code
[22:35] <ubitux> the trailing NUL allows to parse it like a string
[22:35] <ubitux> without the end pointer check
[22:35] <ubitux> the NUL is not accounted anymore, but it is present
[22:36] <beastd> ah i see some mails now
[22:36] <ubitux> (either because of bprint, added manually, or because of FF_INPUT_BUFFER_PADDING_SIZE)
[22:36] <beastd> ubitux: yes, that is fine.
[22:37] <beastd> in general I do no think we have the padding requirement on extradata. but it is strictly needed if you want to use our bitstreamreader on it
[22:40] <cone-681> ffmpeg.git 03Clément BSsch 07master:5f02844c8d4f: Add MPL2 subtitles demuxer and decoder.
[22:51] <ubitux> oups i broke fate
[22:52] <ubitux> will fix in a moment
[22:52] <cone-681> ffmpeg.git 03Clément BSsch 07master:5c68aae90883: Add VPlayer subtitles demuxer and decoder.
[22:56] <cone-681> ffmpeg.git 03Clément BSsch 07master:67286fa98b1e: 10l: export ff_bprint_to_extradata between libs using avpriv_ prefix.
[23:00] <beastd> news is out. going off now. see you. bye...
[23:00] <ubitux> beastd: good job :)
[23:01] <beastd> thanks, subtitles filter is in and concat demuxer is left out for safety
[23:01] <ubitux> ok
[23:12] <cone-681> ffmpeg.git 03Clément BSsch 07master:d9ac8d296725: lavf: move srtdec:read_chunk() to subtitles utils.
[23:12] <cone-681> ffmpeg.git 03Clément BSsch 07master:725d6c615c78: Add MPlayer subtitles demuxer.
[23:14] <cone-681> ffmpeg.git 03Clément BSsch 07master:949506191a81: lavf/subtitles: fix CLRF/CRLF typo.
[23:32] <ubitux> michaelni: did i forgot to send you the mpsub samples?
[23:33] <ubitux> btw, sub/MicroDVD_capability_tester.srt can be removed
[23:33] <ubitux> michaelni: sub/MPSub_capability_tester_frames.sub is present but sub/MPSub_capability_tester.sub is missing
[23:41] <cone-681> ffmpeg.git 03Clément BSsch 07master:a1e4e352a624: Add AQTitle subtitles demuxer.
[23:43] <cone-681> ffmpeg.git 03Clément BSsch 07master:580ee973d16c: FATE: disable fate-sub-mpsub temporarly.
[23:44] <saste> ^ the day of the subtitles
[23:44] <ubitux> :)
[23:44] <ubitux> two left
[23:46] <ubitux> thanks everyone for the reviews btw
[23:59] <cone-681> ffmpeg.git 03Clément BSsch 07master:7b43402724b2: Add PJS subtitles demuxer and decoder.
[00:00] --- Mon Dec 31 2012