[Ffmpeg-devel-irc] ffmpeg-devel.log.20140328

[burek]

[00:01] <wm4> it's being a bit impractical here, but I did it in another way already
[00:21] <wm4> ubitux: here's a patch for ASS utf16 support: http://sprunge.us/EHcY
[00:21] <cone-106> ffmpeg.git 03Michael Niedermayer 07master:64b79141bdfd: avcodec/libx264: move where x264opts is applied down so it isnt overridden by avctx & defaults
[00:21] <wm4> ubitux: other subtitle readers could be extended similarly, although they have more dependency on random other stuff
[00:29] <wm4> why do people do this https://github.com/chelyaev/ffmpeg-tutorial/blob/master/tutorial05.c#L529
[01:16] <kierank> wm4: you had to do that back in the days
[01:16] <kierank> with reordered opaque
[01:16] <wm4> even then you could do it in a cleaner way
[01:17] <wm4> at least in the last few years
[03:23] <cone-447> ffmpeg.git 03rogerdpack 07master:773eb74babe0: dshow: show device name when outputting buffer overflow log message
[03:32] <cone-447> ffmpeg.git 03Vittorio Giovara 07master:d37c96213a2a: lavc: restore copy_block{4,16} functions
[03:32] <cone-447> ffmpeg.git 03Michael Niedermayer 07master:03e4c2d8333d: Merge commit 'd37c96213a2a9e1fd8669122d5405f4ce6a99ed8'
[04:04] <Zeranoe> llogan: I think you have an admirer http://ffmpeg.zeranoe.com/forum/viewtopic.php?f=7&t=1358#p6489
[04:40] <cone-447> ffmpeg.git 03Paul B Mahol 07master:70daeacd6ef8: PAF demuxer and decoder
[04:40] <cone-447> ffmpeg.git 03Michael Niedermayer 07master:16ddc58bd7a2: Merge commit '70daeacd6ef8b354dd7d2d77ad393831a5bbf033'
[04:45] <cone-447> ffmpeg.git 03Paul B Mahol 07master:a7a5e3850ecd: fate: add PAF audio and video tests
[04:45] <cone-447> ffmpeg.git 03Michael Niedermayer 07master:8a8472dd49d7: Merge commit 'a7a5e3850ecd94e726ad2272295b9e6c91841cf8'
[04:46] <cone-447> ffmpeg.git 03Vittorio Giovara 07master:792e4c21f212: xbm: use av_frame_free on close
[04:46] <cone-447> ffmpeg.git 03Michael Niedermayer 07master:6bd05ed8bb30: Merge commit '792e4c21f212979f0e29bcdf107cb6b4f51645a4'
[05:56] <cone-447> ffmpeg.git 03Vittorio Giovara 07master:678082b409ac: X-Bitmap decoder
[05:56] <cone-447> ffmpeg.git 03Vittorio Giovara 07master:991362fab49b: fate: add XBM tests
[05:57] <cone-447> ffmpeg.git 03Michael Niedermayer 07master:296e63efa56f: Merge commit '678082b409aca711f9cf991df6b0200116489322'
[05:57] <cone-447> ffmpeg.git 03Michael Niedermayer 07master:4618084a5cdd: avcodec/xbmdec: merge ptr increase into dereference
[05:57] <cone-447> ffmpeg.git 03Michael Niedermayer 07master:1da186676738: avcodec/xbmdec: support X10 format
[05:57] <cone-447> ffmpeg.git 03Michael Niedermayer 07master:2b570c9569a2: Merge commit '991362fab49b60d34d89b7b7d6dd00fbec3ce022'
[09:23] <ubitux> wm4: interesting
[09:24] <ubitux> now it would probably be a good thing to make ass demuxer use the lavf/subtitles api, and make that one support utf16
[09:26] <ubitux> actually, make it use ff_get_line()
[09:27] <ubitux> mmh well, with AVBPrint you have a virtually unlimited buffer though
[09:30] <ubitux> wm4: anyway, we should probably have a ff_subtitles_read_{line,chunk}(), each of them taking that FFTextReader context
[09:31] <ubitux> most subtitles demuxers are already using ff_subtitles_read_chunk() and ff_get_line()
[09:33] <ubitux> that might require a bit more work thought, but would homogenize them
[09:34] <ubitux> ah, and you'll have ff_smil_extract_next_chunk as well (for sami & shit)
[09:41] <ubitux> that's pretty cool actually now that you have this "r8" read 
[11:17] <cone-453> ffmpeg.git 03Paul B Mahol 07master:fb5cf145b6bc: bmp: add a standalone parser
[11:17] <cone-453> ffmpeg.git 03Michael Niedermayer 07master:a696b0917d1d: Merge commit 'fb5cf145b6bcfa4f83af94398e5560c1132cc410'
[12:05] <cone-453> ffmpeg.git 03Vittorio Giovara 07master:e8e560f2a20e: fate: add a bmpparser test
[12:05] <cone-453> ffmpeg.git 03Michael Niedermayer 07master:e1d10017039d: Merge remote-tracking branch 'qatar/master'
[12:58] <cone-453> ffmpeg.git 03Michael Niedermayer 07master:92005c26360a: fate/aliaspix: Use standard test sample
[13:55] <J_Darnley> To whoever runs trac: ticket #2823 has a spam comment, see: https://trac.ffmpeg.org/ticket/2823#comment:10
[14:36] <j-b> good morning!
[17:40] <cone-453> ffmpeg.git 03Michael Niedermayer 07master:2b4543ff6968: cmdutils/filter_codec_opts: do not discard all options for CODEC_ID_NONE
[17:40] <cone-453> ffmpeg.git 03Michael Niedermayer 07master:de96e20be959: avfilter/lavfutils: call av_find_stream_info() before reading various information about the stream
[17:42] <ubitux> wm4: going to submit the patch?
[17:42] <wm4> ubitux: probably
[17:42] <wm4> just wondering what I should submit
[17:42] <ubitux> your patch
[17:42] <ubitux> obviously
[17:42] <ubitux> :)
[17:54] <wm4> ubitux: any modifications I should make before submitting?
[17:54] <ubitux> well if you're motivated, what i suggested a few hours ago
[17:55] <ubitux> otherwise if you feel like it's good enough for now then go for it
[17:55] <ubitux> a test would be welcome btw
[18:56] <jnvsor> Suppose you guys would know more - in filter evals is there any way to get key states? (Ie push to talk by using if() in volume filter with key states?)
[18:57] <wm4> no, ffmpeg is not an arcade game emulator
[19:00] <jnvsor> Hmm, don't suppose there's any way to do that without separately scripting a keybind or something huh?
[19:03] <ubitux> jnvsor: command inject probably
[19:03] <ubitux> with ømq
[19:04] <ubitux> look at http://ffmpeg.org/ffmpeg-filters.html#zmq_002c-azmq
[19:04] <ubitux> if the filter has the command you might be able to interface with it through that
[19:04] <ubitux> otherwise, directly the api
[19:04] <ubitux> for what it's worth
[19:20] <jnvsor> ubitux: Sweet! Looks like it will do the job nicely once I get it working
[19:22] <wm4> ubitux: int ff_get_line(AVIOContext *s, char *buf, int maxlen)
[19:23] <wm4> ubitux: so that uses a static buffer, not AVBPrint
[19:23] <wm4> I expect I'd see resistance if I replaced AVBPrint with a static buffer
[19:23] <ubitux> yes
[19:23] <ubitux> that's why you shoud do the other way around
[19:24] <jnvsor> Don't suppose you know anything that reacts to keyup as well as keydown?
[20:01] <jnvsor> What about asendcmd? Could that work?
[00:00] --- Sat Mar 29 2014