[Ffmpeg-devel-irc] ffmpeg-devel.log.20150502

[burek]

[00:53:29 CEST] <alezandro> when running doc/example/demuxing_decoding is see output like this
[00:53:33 CEST] <alezandro> p     30537  d     27534
[00:53:33 CEST] <alezandro> p     36543  d     30537
[00:53:49 CEST] <alezandro> where is the printf that prints this? .... thx
[03:06:15 CEST] <Timothy_1u> alezandro: in pgm_save(): https://github.com/FFmpeg/FFmpeg/blob/master/doc/examples/decoding_encoding.c#L500
[03:06:34 CEST] <Timothy_1u> scratch that oops
[03:16:39 CEST] <cone-151> ffmpeg 03Michael Niedermayer 07master:61976727204e: avcodec/svq3: Print unknown fields at debug level
[08:04:18 CEST] <BBB> nevcairiel: bitdepth is profile 2, yes
[08:04:29 CEST] <BBB> nevcairiel: profile 1 is rgb support and non-420 chroma subsampling, thats supported now
[08:13:43 CEST] <rcombs> https://gist.github.com/d613e5029d78b6407676 <-- anyone here have any thoughts before I send this? (Secure Channel [OSX TLS] support)
[11:01:05 CEST] <rcombs> so, if anyone wants to add schannel support (Windows) then we could probably switch over to verify-by-default in the next minor version
[11:04:16 CEST] <nevcairiel> sounds more like a major version t hing
[11:04:45 CEST] <nevcairiel> I would work on schannel, but the schannel API in c++ is already terrible, and adding the C layer on top of that makes it not any better
[11:05:04 CEST] <rcombs> does ffmpeg have any particular rules for when major bumps happen?
[11:27:00 CEST] <wm4> nevcairiel: IMO SOMEOME should unmess the TLS code and separate each impl. to its own file
[11:27:10 CEST] <wm4> maybe we could memcpy(rcombs, SOMEONE)
[11:27:20 CEST] <wm4> or is it the other way around
[11:27:30 CEST] <nevcairiel> if rcombs is your target, then thats fine
[11:29:39 CEST] <rcombs> hey I don't like being overwritten
[11:29:45 CEST] <rcombs> rude
[11:30:40 CEST] <wm4> things happen when using C
[12:13:30 CEST] <cone-512> ffmpeg 03Rodger Combs 07master:71b69df62248: lavf/http: support HTTPS when listening
[13:41:58 CEST] <cone-512> ffmpeg 03Michael Niedermayer 07master:c1872172873c: avformat/matroskadec: Mark static tables as const
[13:54:13 CEST] <cone-512> ffmpeg 03Michael Niedermayer 07master:270e647adc3b: avcodec/x86/dct-test: Make static table const
[13:54:14 CEST] <cone-512> ffmpeg 03Michael Niedermayer 07master:dfc0708e233d: avcodec/x86/dct-test: Use uint8_t for idct_simple_mmx_perm
[14:58:24 CEST] <cone-512> ffmpeg 03Michael Niedermayer 07master:156f64bdf142: avutil/cast5: Mark key_bits as const
[14:58:25 CEST] <cone-512> ffmpeg 03Michael Niedermayer 07master:536b42e85841: avutil/cast5: Make iv array static
[15:14:25 CEST] <cone-512> ffmpeg 03Michael Niedermayer 07master:dc4a6e4d5a49: avfilter/vf_libopencv: Make ocv_filter_entries const
[16:00:11 CEST] <wm4> so what does --enable-hardcodec-tables do?
[16:13:17 CEST] <cone-512> ffmpeg 03Michael Niedermayer 07master:4e8ee7acb858: libavcodec/svq3: clarify unknown_flag by renaming to has_watermark
[17:41:51 CEST] <kierank> wm4: increases binary size by hardcoding tables
[17:42:09 CEST] <Daemon404> what was the original usecase anyway
[17:42:18 CEST] <Daemon404> very very slightly faster init?
[17:42:18 CEST] <wm4> does it make it faster?
[17:42:26 CEST] <wm4> the only thing I could guess
[17:42:27 CEST] <Daemon404> wm4, well the tables are generated only on init...
[17:42:59 CEST] <wm4> it broke recently (probably fixed by now?), so it caused some fallout
[17:43:05 CEST] <wm4> due to users cargo-culting it
[17:43:51 CEST] <Daemon404> thats so stupid
[17:44:01 CEST] <Daemon404> the only usecase is some embedded system probably
[21:24:12 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:94340e4b28cc: avfilter: add find_rect filter
[21:24:13 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:61b4ab18609a: avfilter: Add cover_rect filter
[21:48:44 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:94f4bdc4f146: configure: add forgotten avcodec/avformat deps for find/cover_rect
[22:13:29 CEST] <BtbN> I can't find anything on how other filters handle the pixel format matching. So i think it just works automaticaly, and the filter chain is adjusted once the first frame is output?
[22:13:50 CEST] <BtbN> That's also how it looks like in a debug log.
[22:18:35 CEST] <cone-666> ffmpeg 03Martin Storsjö 07master:21180b73239c: avplay: Factorize code for adding filters to the filter pipeline
[22:18:36 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:e874772f7090: Merge commit '21180b73239c6360aa28496d4879713b7ba4a8e5'
[22:38:13 CEST] <cone-666> ffmpeg 03Martin Storsjö 07master:e2d50fc2f5f3: avplay: Add support for rotated video
[22:38:14 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:deedf3e59002: Merge commit 'e2d50fc2f5f3600e13055acf1a10fec35e941f37'
[23:11:47 CEST] <jamrial> http://www.zdnet.com/article/rip-windows-media-center/
[23:13:00 CEST] <cone-666> ffmpeg 03Martin Storsjö 07master:16302246b1fc: avconv: Add an option for automatically rotating video according to display matrix
[23:13:01 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:f5b26fbc2f56: Merge commit '16302246b1fcb7ad4e6f7bd31c49956a455336d2'
[23:13:02 CEST] <cone-666> ffmpeg 03Michael Niedermayer 07master:a1acae040970: ffmpeg_opt: remove rotate metadata in case of autorotate
[00:00:00 CEST] --- Sun May  3 2015