[Ffmpeg-devel-irc] ffmpeg-devel.log.20160626

burek burek021 at gmail.com
Mon Jun 27 02:05:03 CEST 2016


[00:14:28 CEST] <DSM_> michaelni: why unref infmts be a problem? http://ffmpeg.org/pipermail/ffmpeg-devel/2016-June/195444.html
[00:15:11 CEST] <DSM_> i thought, it should be fixed before release
[00:16:54 CEST] <michaelni> DSM_, there are 2 calls to ff_formats_ref(), if the first succeeds and the 2nd fails i think you would have a double free if you unref the first too
[00:22:34 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07master:fdd8aac7e12e: avfilter/src_movie: Add option to remove timestamp discontinuities
[01:07:14 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07master:9981b70da50f: avcodec/h264_sei: Do not skip subsequent SEIs on errors potentially caused by missing parameter sets.
[01:07:14 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07master:f1dabc958e09: avfilter/dualinput: Add support to skip initial unpaired frames
[02:52:10 CEST] <DSM_> michaelni: why unref infmts be a problem? http://ffmpeg.org/pipermail/ffmpeg-devel/2016-June/195444.html
[02:53:52 CEST] <michaelni> <michaelni> DSM_, there are 2 calls to ff_formats_ref(), if the first succeeds and the 2nd fails i think you would have a double free if you unref the first too
[02:54:20 CEST] Action: michaelni hopes DSM_ doesn get disconnected before reading it this time :)
[02:56:08 CEST] <DSM_> :D
[03:07:31 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07master:8baa1d220987: avfilter/vf_paletteuse: enable skip_initial_unpaired
[03:07:32 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07master:7b92fe8d3faa: doc/APIchanges: Fill in and correct more stuff
[03:31:50 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07master:074fdf4d12b0: doc/APIchanges: Add FFmpeg 3.1 cut marker
[03:32:52 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07release/3.1:HEAD: doc/APIchanges: Add FFmpeg 3.1 cut marker
[04:03:51 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07release/3.1:e5d434b84040: tests/checkasm/checkasm: Disable checkasm_check_pixblockdsp for ppc64be
[04:03:51 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07release/3.1:182cfe4832d8: release notes (based on release/3.0)
[04:19:09 CEST] <cone-423> ffmpeg 03Michael Niedermayer 07release/3.0:fbdf5ca763f0: tests/checkasm/checkasm: Disable checkasm_check_pixblockdsp for ppc64be
[09:34:18 CEST] <nevcairiel> disabling checkasm on some platform and only in releases sounds wrong, if anything the asm thats broken should be turned off <.<
[10:26:16 CEST] <michaelni> nevcairiel, fate passes except checkasm, also looking at the checkasm doesnt it test get/diff pixels with unaligned input ? these are documented in  libavcodec/pixblockdsp.h as requiring alignment 
[10:28:52 CEST] <cone-970> ffmpeg 03Paul B Mahol 07master:8fb92ea7e2a6: avcodec: add YUY2 Lossless Codec decoder
[13:11:08 CEST] <cone-970> ffmpeg 03Michael Niedermayer 07release/3.1:b2a74dd629c1: Set version to 3.1
[14:03:38 CEST] <ubitux> i wonder how to solve the problem with subtitles who have timing obtained at decoding process
[14:03:46 CEST] <ubitux> (hello dvdsubs)
[14:04:05 CEST] <ubitux> and xsub
[14:08:02 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:35c858066840: avconv: stop using AVStream.codec
[14:08:03 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:b8945c4e2a1b: Merge commit '35c858066840352d6d43385bbc728467c5150974'
[14:09:34 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:e2a3df1a4640: avconv: Use more precise deprecation ifdefs
[14:09:35 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:4531981a9b20: Merge commit 'e2a3df1a464091075040a81ffc419fa21a8a5fd8'
[14:13:47 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:e45a638f50cc: dump: Drop unused variable
[14:13:48 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:b23d4e52fd24: Merge commit 'e45a638f50cc1dbeb87b9792e68f57e77fc0c3b5'
[14:22:50 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:52567e819866: get_bits: Drop some TRACE-level debug code
[14:22:51 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:e2836397f11c: Merge commit '52567e8198669a1e7493c75771613f87a90466c3'
[14:32:25 CEST] <cone-970> ffmpeg 03Luca Barbato 07master:74b1bf632f12: mp3: Make the extrasize explicit
[14:32:27 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:314e6086315d: Merge commit '74b1bf632f125a795e66e5fd0a060b9c7c55b7a3'
[14:32:50 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:14634429b915: lavf: update muxing doxy
[14:32:50 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:97bdb94341ae: Merge commit '14634429b915333f3612eaf41db3954222dc4aaf'
[14:37:02 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:5f30ac27795f: lavc: deprecate avcodec_copy_context()
[14:37:03 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:04fc8e24a091: lavc: deprecate avcodec_get_context_defaults3()
[14:37:04 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:2ef6dab0a79a: lavc: document that avcodec_close() should not be used
[14:37:05 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:6988cf2969d7: Merge commit '5f30ac27795f9f98043e8582ccaad8813104adc4'
[14:37:06 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:69c38d64d745: Merge commit '04fc8e24a091ed1d77d7a3c0cbcfe60baec19a9f'
[14:37:07 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:f9a186c59b4c: Merge commit '2ef6dab0a79a9852a92ed80b07f9e32a37530d9e'
[14:40:41 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:ac84e618df47: avfiltergraph: check the query_formats() return value
[14:40:42 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:7c6ae9ca0489: Merge commit 'ac84e618df4765ba751327497994066d0931e6a8'
[14:49:29 CEST] <cone-970> ffmpeg 03Francois Cartegnie 07master:ad61da054bd8: jpeg2000: Fix profile define values
[14:49:30 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:e47b8bbf0b54: avcodec: Bump micro version after changing public JPEG 2000 defines
[14:49:31 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:46ff31d39c2e: Merge commit 'ad61da054bd8c74a5d5b38d80846228fc6147108'
[14:49:32 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:b95534bb674a: Merge commit 'e47b8bbf0b54599d44b9330eb4d68cdde4f6d298'
[14:51:51 CEST] <cone-970> ffmpeg 03Martin Storsjö 07master:3fdffc032e8e: rtsp: Use avcodec_descriptor_get instead of avcodec_find_decoder
[14:51:52 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:f972653ec198: Merge commit '3fdffc032e8ea5676bc0c2551b900c0dc887835b'
[15:01:58 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:07377fb55790: fate: More fine-grained dependencies for voice codec tests
[15:01:59 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:093e06f2d687: Merge commit '07377fb557909015338b4c215b87b9c73ea577ad'
[15:06:57 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:7ca4fdc046a9: fate: More fine-grained dependencies for demuxer tests
[15:06:58 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:324f0fbff124: Merge commit '7ca4fdc046a945949866bb6fc550e3cab1fc8c00'
[15:12:59 CEST] <ubitux> do we have xsub in sth else than avi?
[15:15:46 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:4024b566d664: golomb: Give svq3_get_se_golomb()/svq3_get_ue_golomb() better names
[15:15:47 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:b20fe650ef85: Merge commit '4024b566d664a4b161d677554be52f32e7ad4236'
[15:21:29 CEST] <cone-970> ffmpeg 03Andrey Turkin 07master:24b5cff01bba: lavc: handle hw_frames_ctx where necessary
[15:21:30 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:2c1d0f871f6a: Merge commit '24b5cff01bbac4e08acfd6d19c499e880988f520'
[15:23:36 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:1c9e8616c535: hwcontext: add a function for opening devices
[15:23:37 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:481f320aa388: Merge commit '1c9e8616c535ef496e7ee8a5cbc5e9e972a6977d'
[15:26:10 CEST] <nevcairiel> i cant really test most of these, but since they are new functions that apply cleanly, its going to be alright, isnt it :D
[15:26:17 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:38392b2af815: hwcontext_vdpau: implement device creation
[15:26:18 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:2e219b491bcc: hwcontext_cuda: implement device creation
[15:26:19 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:8e70385a8ed0: hwcontext_dxva2: implement device creation
[15:26:20 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:b8bf9194af60: hwcontext_vaapi: implement device creation
[15:26:21 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:613ac3bf82cb: Merge commit '38392b2af815898b8716826c4e29d95c04fb2647'
[15:26:22 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:c5a0c16850fc: Merge commit '2e219b491bcc0845248345fdad31231b081e06d1'
[15:26:23 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:162370b47320: Merge commit '8e70385a8ed06f96f1e9e35cf7b6788a5c56eded'
[15:26:25 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:ed5d03ff6f90: Merge commit 'b8bf9194af602cf3a4bcd19a5e278e3d6d69f8fa'
[15:30:50 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:f72db3f2f3a8: avconv_vdpau: use the hwcontext device creation API
[15:30:51 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:844256740d27: Merge commit 'f72db3f2f3a8c83a4f5dede8fa03434b2bf676c6'
[15:36:05 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:18c506e9e6e8: avconv_dxva2: use the hwcontext device creation API
[15:36:06 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:85da4f92580d: Merge commit '18c506e9e6e8df8b1d496d093077b8240ea68c28'
[15:36:07 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:06a0e184359b: hwcontext_dxva2: Use GetDesktopWindow instead of GetShellWindow
[15:36:34 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:9f7590f4e497: avconv_vaapi: use the hwcontext device creation API
[15:36:35 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:4e6f2dbcd3ca: Merge commit '9f7590f4e497c12693247d7f935a20d7773a64dd'
[15:39:35 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:80fb19bc234a: avconv: fix a check for av_bsf_get_by_name() return value
[15:39:36 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:fe7b21c8f148: avconv: fix parsing bitstream filters
[15:39:37 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:c9ee36e60ffb: Merge commit '80fb19bc234a3f2350d891adf39f3738a8e4849f'
[15:39:38 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:7fe7cdca3665: Merge commit 'fe7b21c8f148493c6fbceb7f887a77531dd1ae0e'
[15:41:53 CEST] <cone-970> ffmpeg 03Anton Khirnov 07master:f5c43884190b: FATE: drop the audio stream from the dxtory test
[15:41:53 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:cd52ec2337ae: Merge commit 'f5c43884190be094749acfba104e3eab98d8ff1e'
[15:43:25 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:6641819feedb: build: Ignore generated mapfile and remove it on distclean
[15:43:26 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:5a9134f5f142: Merge commit '6641819feedb086ebba3d2be89b8d33980f367e1'
[15:44:58 CEST] <cone-970> ffmpeg 03Mark Thompson 07master:2bfa067d0b63: vaapi_encode: Check config attributes before creating config
[15:44:59 CEST] <cone-970> ffmpeg 03Mark Thompson 07master:b51c7c6b8a5b: vaapi_h264: Fix frame_num after non-reference frames
[15:45:00 CEST] <cone-970> ffmpeg 03Mark Thompson 07master:a86aa16088ad: vaapi_h264: Add trivial support for low-power encoding
[15:45:01 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:18b1b8be6a2a: Merge commit '2bfa067d0b636e7b2004fb0ad5a53d0d48c6de32'
[15:45:02 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:d3327b7eda97: Merge commit 'b51c7c6b8a5b35cfd06cb9655f9ec4c9f0ddd81b'
[15:45:03 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:f2a3798ea459: Merge commit 'a86aa16088ad7f22a8918d71adb8c040d6033d84'
[15:47:13 CEST] <cone-970> ffmpeg 03Luca Barbato 07master:5264e7ba217b: ac3: Check the array bound before dereferencing
[15:47:14 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:5e418fd63a29: Merge commit '5264e7ba217b3c0ceae813917134e1ab52573141'
[15:53:15 CEST] <cone-970> ffmpeg 03Diego Biurrun 07master:dc40a70c5755: Drop unnecessary libavutil/x86/asm.h #includes
[15:53:16 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07master:c142dc203e25: Merge commit 'dc40a70c5755bccfb1a1349639943e1f408bea50'
[15:54:36 CEST] <ubitux> oh the timings in xsub are in the packets because there is only constant frame rate support in avi?
[15:54:49 CEST] <ubitux> meh.
[15:54:51 CEST] <nevcairiel> yes avi is strictly cfr
[15:54:57 CEST] <nevcairiel> codecs hack around that in various odd ways
[15:55:10 CEST] <nevcairiel> some video codecs have 0-packets to emulate some sort of vfr
[15:55:33 CEST] <ubitux> not a surprise there isn't much subtitles codecs in avi
[15:55:57 CEST] <nevcairiel> avi is very limited format, for some reason its just been hacked up to support all sorts of crazy crap
[15:56:28 CEST] <nevcairiel> they also have this special way to store subtitles in one big blob with in-band timings and let the demuxer just figure it out
[15:56:58 CEST] <ubitux> yeah that's what happened with the ssa file in the avi?
[15:57:21 CEST] <nevcairiel> i've seen it with ass and srt, i think
[15:57:27 CEST] <nevcairiel> but i dont really see any other crazy subtitle formats
[15:57:32 CEST] <nevcairiel> so that may not be limited to those
[15:58:13 CEST] <ubitux> we have an issue with xsub currently in avi
[15:58:28 CEST] <ubitux> they are muxed with a 0/0 timescale
[15:58:32 CEST] <nevcairiel> i honestly dont even know what that is
[15:58:39 CEST] <ubitux> so we fallback on a 1/25
[15:59:00 CEST] <ubitux> (so subtitles get demuxed approximately as pts = 1, 2, 3, ... )
[15:59:18 CEST] <ubitux> and then the decoder set the start/end time
[15:59:30 CEST] <ubitux> but it means we have a drift because the pts should be 0
[16:00:01 CEST] <ubitux> nevcairiel: it's the divx bitmap subtitles
[16:00:30 CEST] <ubitux> it's the only format along with dvdsub to use this start/end display time in the decoded AVSubtitle
[16:00:37 CEST] <nevcairiel> i see
[16:20:07 CEST] <jamrial> nevcairiel: i think 1c9e8616c535ef496e7ee8a5cbc5e9e972a6977d will break ABI for 3.2 seeing we cut 3.1 with this api but without this addition
[16:20:56 CEST] <jamrial> mmh, no wait, the struct is in the internal header
[16:21:01 CEST] <nevcairiel> its only a new function
[16:21:35 CEST] <jamrial> yeah, i was worried about the changes to the struct, but it's internal
[16:21:40 CEST] <jamrial> nevermind then
[16:30:26 CEST] <nevcairiel> it would've been nice nevertheless to get the new api in 3.1, but what can you do
[16:34:47 CEST] <BtbN> it isn't release yet? So can't you push it to the 3.1 branch?
[17:16:03 CEST] <jamrial> nevcairiel: what btbn said, we can still backport that commit or any other until the release is tagged
[17:53:11 CEST] <jamrial> michaelni: since you're doing the release, would you agree to include the latest merge batch from master into 3.1?
[17:53:26 CEST] <jamrial> or are you already in the middle of testing/packaging?
[17:54:28 CEST] <jamrial> cherry picking the above api alone would make the supposed branch cut point innacurate (there was a jpeg2k change before it), so i'm not sure that'd be a good idea
[18:07:14 CEST] <michaelni> jamrial, feel free to include the commits ill make the release in a few hours
[18:07:32 CEST] <michaelni> still wanted to fix a libopenjpeg issue i saw
[18:11:07 CEST] <cone-970> ffmpeg 03Michael Niedermayer 07master:e9394ca63dab: avcodec/libopenjpegenc: Set numresolutions by default to a value that is not too large
[18:20:41 CEST] <michaelni> btw i get on master "libavutil/hwcontext_vdpau.c:447:0: warning: "GET_CALLBACK" redefined [enabled by default]"
[18:21:34 CEST] <nevcairiel> does it tell you where the first one is?
[18:22:11 CEST] <nevcairiel> oh i see
[18:23:32 CEST] <nevcairiel> those two are identical, so either get rid of one or undef it after its use, i guess
[18:24:13 CEST] <nevcairiel> hm, almost identical
[18:30:07 CEST] <michaelni> jamrial, nevcairiel ill make the release from release/3.1 probably in 4-5 hours if you want to add or change anything in release/3.1. Of course if you need more time i can also wait and make the release tomorrow
[18:30:40 CEST] <jamrial> michaelni: https://github.com/jamrial/FFmpeg/commits/release/3.1 https://github.com/jamrial/FFmpeg/commits/master is this ok? i essentially updated the marker in master and merged everything into 3.1
[18:46:05 CEST] <michaelni> jamrial, iam testing this a bit will reply once iam done
[19:20:10 CEST] <cone-970> ffmpeg 03Marton Balint 07master:340cea9f22c1: avdevice/decklink_dec: fix bitrate calculations
[19:20:11 CEST] <cone-970> ffmpeg 03Marton Balint 07master:1fc85d8511fb: avdevice/decklink: factorize cleanup function to common code
[19:20:12 CEST] <cone-970> ffmpeg 03Marton Balint 07master:4ce0a77ec80f: avdevice/decklink: remove unused itermode iterators
[19:20:13 CEST] <cone-970> ffmpeg 03Marton Balint 07master:38d75fe90696: avdevice/decklink: factorize device finder function
[19:20:14 CEST] <cone-970> ffmpeg 03Marton Balint 07master:e22760aafd30: avdevice/decklink: always free decklink resources on error
[19:20:15 CEST] <cone-970> ffmpeg 03Marton Balint 07master:8f9fa49bd8bf: avdevice/decklink: add support for setting duplex mode
[19:20:16 CEST] <cone-970> ffmpeg 03Marton Balint 07master:da89c6e37cf2: avdevice/decklink: add support for audio and video input selection
[19:20:17 CEST] <cone-970> ffmpeg 03Marton Balint 07master:cf0a05668b99: avdevice/decklink: add support for setting input packet timestamp source
[19:20:52 CEST] <michaelni> jamrial, seems to still work
[19:20:59 CEST] <JEEB> so, uh, anyone know where some HLG test content can be grabbed?
[19:21:24 CEST] <JEEB> it was just implemented in mpv's opengl renderer and it'd be nice to get some actual test content for it
[19:58:50 CEST] <cone-970> ffmpeg 03James Almer 07master:069fd69662a1: Update FFmpeg 3.1 cut marker
[19:59:25 CEST] <jamrial> ugh, hook declined for the 3.1 branch
[19:59:53 CEST] <jamrial> it could have aborted the whole push in that case
[20:00:10 CEST] <jamrial> "remote: -Deny-          ERROR: update contains a merge."
[20:02:13 CEST] <nevcairiel> you need to add a Merged-by line
[20:03:29 CEST] <jamrial> nevcairiel: ok, let me try again
[20:03:52 CEST] <jamrial> i any case marton's push above kinda made things a bit more complex
[20:04:26 CEST] <nevcairiel> not really
[20:04:41 CEST] <nevcairiel> you dont need to re-do the merge, just amend the log message
[20:24:00 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:35c858066840: avconv: stop using AVStream.codec
[20:24:01 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:e2a3df1a4640: avconv: Use more precise deprecation ifdefs
[20:24:02 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:e45a638f50cc: dump: Drop unused variable
[20:24:03 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:52567e819866: get_bits: Drop some TRACE-level debug code
[20:24:04 CEST] <cone-970> ffmpeg 03Luca Barbato 07release/3.1:74b1bf632f12: mp3: Make the extrasize explicit
[20:24:05 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:14634429b915: lavf: update muxing doxy
[20:24:06 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:5f30ac27795f: lavc: deprecate avcodec_copy_context()
[20:24:07 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:04fc8e24a091: lavc: deprecate avcodec_get_context_defaults3()
[20:24:08 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:2ef6dab0a79a: lavc: document that avcodec_close() should not be used
[20:24:09 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:ac84e618df47: avfiltergraph: check the query_formats() return value
[20:24:10 CEST] <cone-970> ffmpeg 03Francois Cartegnie 07release/3.1:ad61da054bd8: jpeg2000: Fix profile define values
[20:24:11 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:e47b8bbf0b54: avcodec: Bump micro version after changing public JPEG 2000 defines
[20:24:12 CEST] <cone-970> ffmpeg 03Martin Storsjö 07release/3.1:3fdffc032e8e: rtsp: Use avcodec_descriptor_get instead of avcodec_find_decoder
[20:24:13 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:07377fb55790: fate: More fine-grained dependencies for voice codec tests
[20:24:14 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:7ca4fdc046a9: fate: More fine-grained dependencies for demuxer tests
[20:24:15 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:4024b566d664: golomb: Give svq3_get_se_golomb()/svq3_get_ue_golomb() better names
[20:24:16 CEST] <cone-970> ffmpeg 03Andrey Turkin 07release/3.1:24b5cff01bba: lavc: handle hw_frames_ctx where necessary
[20:24:17 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:1c9e8616c535: hwcontext: add a function for opening devices
[20:24:18 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:38392b2af815: hwcontext_vdpau: implement device creation
[20:24:19 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:2e219b491bcc: hwcontext_cuda: implement device creation
[20:24:20 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:8e70385a8ed0: hwcontext_dxva2: implement device creation
[20:24:21 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:b8bf9194af60: hwcontext_vaapi: implement device creation
[20:24:22 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:f72db3f2f3a8: avconv_vdpau: use the hwcontext device creation API
[20:24:23 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:18c506e9e6e8: avconv_dxva2: use the hwcontext device creation API
[20:24:24 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:9f7590f4e497: avconv_vaapi: use the hwcontext device creation API
[20:24:25 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:80fb19bc234a: avconv: fix a check for av_bsf_get_by_name() return value
[20:24:26 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:fe7b21c8f148: avconv: fix parsing bitstream filters
[20:24:27 CEST] <cone-970> ffmpeg 03Anton Khirnov 07release/3.1:f5c43884190b: FATE: drop the audio stream from the dxtory test
[20:24:28 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:6641819feedb: build: Ignore generated mapfile and remove it on distclean
[20:24:29 CEST] <cone-970> ffmpeg 03Mark Thompson 07release/3.1:2bfa067d0b63: vaapi_encode: Check config attributes before creating config
[20:24:30 CEST] <cone-970> ffmpeg 03Mark Thompson 07release/3.1:b51c7c6b8a5b: vaapi_h264: Fix frame_num after non-reference frames
[20:24:31 CEST] <cone-970> ffmpeg 03Mark Thompson 07release/3.1:a86aa16088ad: vaapi_h264: Add trivial support for low-power encoding
[20:24:32 CEST] <cone-970> ffmpeg 03Luca Barbato 07release/3.1:5264e7ba217b: ac3: Check the array bound before dereferencing
[20:24:33 CEST] <cone-970> ffmpeg 03Diego Biurrun 07release/3.1:dc40a70c5755: Drop unnecessary libavutil/x86/asm.h #includes
[20:24:34 CEST] <cone-970> ffmpeg 03Paul B Mahol 07release/3.1:8fb92ea7e2a6: avcodec: add YUY2 Lossless Codec decoder
[20:24:35 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:b8945c4e2a1b: Merge commit '35c858066840352d6d43385bbc728467c5150974'
[20:24:36 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:4531981a9b20: Merge commit 'e2a3df1a464091075040a81ffc419fa21a8a5fd8'
[20:24:37 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:b23d4e52fd24: Merge commit 'e45a638f50cc1dbeb87b9792e68f57e77fc0c3b5'
[20:24:38 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:e2836397f11c: Merge commit '52567e8198669a1e7493c75771613f87a90466c3'
[20:24:39 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:314e6086315d: Merge commit '74b1bf632f125a795e66e5fd0a060b9c7c55b7a3'
[20:24:40 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:97bdb94341ae: Merge commit '14634429b915333f3612eaf41db3954222dc4aaf'
[20:24:41 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:6988cf2969d7: Merge commit '5f30ac27795f9f98043e8582ccaad8813104adc4'
[20:24:42 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:69c38d64d745: Merge commit '04fc8e24a091ed1d77d7a3c0cbcfe60baec19a9f'
[20:24:43 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:f9a186c59b4c: Merge commit '2ef6dab0a79a9852a92ed80b07f9e32a37530d9e'
[20:24:44 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:7c6ae9ca0489: Merge commit 'ac84e618df4765ba751327497994066d0931e6a8'
[20:24:45 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:46ff31d39c2e: Merge commit 'ad61da054bd8c74a5d5b38d80846228fc6147108'
[20:24:46 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:b95534bb674a: Merge commit 'e47b8bbf0b54599d44b9330eb4d68cdde4f6d298'
[20:24:47 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:f972653ec198: Merge commit '3fdffc032e8ea5676bc0c2551b900c0dc887835b'
[20:24:48 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:093e06f2d687: Merge commit '07377fb557909015338b4c215b87b9c73ea577ad'
[20:24:49 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:324f0fbff124: Merge commit '7ca4fdc046a945949866bb6fc550e3cab1fc8c00'
[20:24:50 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:b20fe650ef85: Merge commit '4024b566d664a4b161d677554be52f32e7ad4236'
[20:24:51 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:2c1d0f871f6a: Merge commit '24b5cff01bbac4e08acfd6d19c499e880988f520'
[20:24:52 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:481f320aa388: Merge commit '1c9e8616c535ef496e7ee8a5cbc5e9e972a6977d'
[20:24:53 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:613ac3bf82cb: Merge commit '38392b2af815898b8716826c4e29d95c04fb2647'
[20:24:54 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:c5a0c16850fc: Merge commit '2e219b491bcc0845248345fdad31231b081e06d1'
[20:24:55 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:162370b47320: Merge commit '8e70385a8ed06f96f1e9e35cf7b6788a5c56eded'
[20:24:57 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:ed5d03ff6f90: Merge commit 'b8bf9194af602cf3a4bcd19a5e278e3d6d69f8fa'
[20:24:57 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:844256740d27: Merge commit 'f72db3f2f3a8c83a4f5dede8fa03434b2bf676c6'
[20:24:58 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:85da4f92580d: Merge commit '18c506e9e6e8df8b1d496d093077b8240ea68c28'
[20:24:59 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:06a0e184359b: hwcontext_dxva2: Use GetDesktopWindow instead of GetShellWindow
[20:25:00 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:4e6f2dbcd3ca: Merge commit '9f7590f4e497c12693247d7f935a20d7773a64dd'
[20:25:01 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:c9ee36e60ffb: Merge commit '80fb19bc234a3f2350d891adf39f3738a8e4849f'
[20:25:02 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:7fe7cdca3665: Merge commit 'fe7b21c8f148493c6fbceb7f887a77531dd1ae0e'
[20:25:03 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:cd52ec2337ae: Merge commit 'f5c43884190be094749acfba104e3eab98d8ff1e'
[20:25:04 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:5a9134f5f142: Merge commit '6641819feedb086ebba3d2be89b8d33980f367e1'
[20:25:05 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:18b1b8be6a2a: Merge commit '2bfa067d0b636e7b2004fb0ad5a53d0d48c6de32'
[20:25:06 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:d3327b7eda97: Merge commit 'b51c7c6b8a5b35cfd06cb9655f9ec4c9f0ddd81b'
[20:25:07 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:f2a3798ea459: Merge commit 'a86aa16088ad7f22a8918d71adb8c040d6033d84'
[20:25:08 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:5e418fd63a29: Merge commit '5264e7ba217b3c0ceae813917134e1ab52573141'
[20:25:09 CEST] <cone-970> ffmpeg 03Hendrik Leppkes 07release/3.1:c142dc203e25: Merge commit 'dc40a70c5755bccfb1a1349639943e1f408bea50'
[20:25:10 CEST] <cone-970> ffmpeg 03Michael Niedermayer 07release/3.1:e9394ca63dab: avcodec/libopenjpegenc: Set numresolutions by default to a value that is not too large
[20:25:11 CEST] <cone-970> ffmpeg 03James Almer 07release/3.1:104c357b6a3e: Merge branch 'master' into release/3.1
[20:25:12 CEST] <cone-970> ffmpeg 03James Almer 07release/3.1:970f2ad966c2: Update FFmpeg 3.1 cut marker
[20:25:58 CEST] <jamrial> sorry for the spam
[20:31:27 CEST] <cone-970> ffmpeg 03James Almer 07master:c6f2d1a21f80: changelog: fix entry order
[20:31:29 CEST] <cone-970> ffmpeg 03James Almer 07release/3.1:cf09348b9e03: changelog: fix entry order
[20:48:30 CEST] <andrey_turkin_> a lot of merges done in a day. great job!
[21:15:18 CEST] <cone-970> ffmpeg 03Martin Vignali 07master:971ee09d32a4: fate/exr : add tests for file with half and float channels
[21:15:19 CEST] <cone-970> ffmpeg 03Martin Vignali 07master:bd2cec7021a0: fate/webp : add test for webp lossless decoding (rgb and rgba)
[21:30:57 CEST] <atomnuker> nevcairiel: was it C89->C99 converters that had trouble with for (int loops?
[21:33:15 CEST] <atomnuker> (or was it ancient broken MSVC compilers which shouldn't have been used to compile C in the first place)
[21:33:18 CEST] <rcombs> (you mean C89->C99?)
[21:33:29 CEST] <rcombs> (erm)
[21:33:31 CEST] <atomnuker> yes
[21:33:33 CEST] <rcombs> (C99->C89)
[21:40:17 CEST] <kierank> for( int is not c89 that was the problem in the past
[21:41:09 CEST] <rcombs> does anyone care about C99->C89 or ancient MSVCs anymore
[21:41:36 CEST] <JEEB> 2013 u2 and 2015 should now handle c99 well enough
[21:41:51 CEST] <JEEB> I'm not sure if we should try to maintain the older stuff with the conversion app
[21:42:07 CEST] <rcombs> sounds not worth it
[21:42:39 CEST] <rcombs> (is there any sane reason to be using a MSVC older than 2013?)
[21:42:57 CEST] <JEEB> nope, since even the XP aimers can use 2013
[21:42:58 CEST] <rcombs> ("building on WinXP" doesn't count as sane)
[21:43:05 CEST] <jamrial> 2012 still compiles with that tool and passes most fate tests
[21:43:47 CEST] <jamrial> and i think someone fixed a compilation error with 2010 not too long ago, so that also works, again with the tool
[21:44:12 CEST] <atomnuker> so no reason to keep not using for (int loops anymore? brb writing email
[21:44:19 CEST] <jamrial> 2013 is the minimum to get fate to fully pass anyway
[21:45:47 CEST] <jamrial> atomnuker: at this point it's done mainly for cosmetic reasons i guess
[21:45:50 CEST] <jamrial> constitency with the rest of the codebase
[21:52:56 CEST] <nevcairiel> breaking existing support for compilers, especially if they are tested by fate as well, is not something we should do for some silly cosmetic like moving the int
[21:53:11 CEST] <nevcairiel> (i dont know if thats something that breaks it, fwiw)
[21:53:31 CEST] <atomnuker> we have fate tests for those, don't we
[21:53:47 CEST] <atomnuker> if they get tripped we can always change them, it's a silly cosmetic tweak after all
[21:53:55 CEST] <nevcairiel> git master is not for experiementing with things :p
[21:54:44 CEST] <atomnuker> not an experiment if you know what you're doing
[21:55:03 CEST] <atomnuker> we can't live in the past forever you know
[21:55:12 CEST] <nevcairiel> if you knew what you were doing, you wouldnt need fate to confirm if it breaks :)
[21:56:41 CEST] <atomnuker> but what if the compiler's broken and doesn't know 1 << 1 + 1 != (1 << 1) + 1
[22:00:29 CEST] <andrey_turkin_> I don't think there is much reason to worry about MSVC < 2013. Main force to use particular MSVC version (apart from licensing) would be to link with some binary libraries, and most of the vendors provide at least 2013 compatible binaries these days.
[22:00:58 CEST] <nevcairiel> the main point is that we dont decide to drop support for a compiler version on a whim like this
[22:00:58 CEST] <rcombs> does that even matter when it's a C lib
[22:01:18 CEST] <nevcairiel> especially for silly cosmetics
[22:02:11 CEST] <andrey_turkin_> if only thing to keep older MSVC happy is to not use declarations in for loop - then maybe
[22:02:20 CEST] <andrey_turkin_> is it the only thing?
[22:03:57 CEST] <andrey_turkin_> I actually tried to build ffmpeg with MSVC 2010 some time ago - it was a giant pain to build and then some pain to use
[22:05:26 CEST] <andrey_turkin_> it's actually seemed easier to build it with mingw crosscompilation (and now then cv2pdb tool works better, you can even get some debugging information for cross-compiled ffmpeg)
[22:05:53 CEST] <nevcairiel> vs2012 seems to be fine with for (int i = ...) btw
[22:05:58 CEST] <nevcairiel> must have been some other crazy compiler
[22:06:14 CEST] <wbs> nevcairiel: no, c99toc89 has _rudimentary_ support for int declarations there
[22:06:39 CEST] <wbs> like, it works for trivial cases. not tested much, not intended to be used
[22:06:44 CEST] <nevcairiel> i see
[22:06:57 CEST] <andrey_turkin_> there was some compiler which bleeded declaration outside of scope. Maybe that was the problem
[22:07:32 CEST] <wbs> yes, there are such warnings/errors for gcc and/or clang, which are turned on intentionally to catch these issues, since the policy is to not use them
[22:08:01 CEST] <wbs> other than that, I really agree with nevcairiel. since all the code builds just fine right now, I dont see the hurry to drop something that works, unless you have a compelling case
[22:08:19 CEST] <wbs> every time you drop support for something like that, you rob a perfectly working setup from somebody
[22:08:28 CEST] <JEEB> true
[22:08:33 CEST] <wbs> forcing them to do extra boring upgrade work instead of something productive
[22:08:52 CEST] <wbs> and every time you do such an upgrade dance, you need to upgrade X and Y and Z, just because somebody said "I don't think you should be using that"
[22:09:13 CEST] <JEEB> the problem of course is if nobody is really checking it. if we have FATE setups then that's fine of course
[22:09:31 CEST] <JEEB> then we can't really say we support it in any way
[22:09:50 CEST] <JEEB> that said as I noted, we do have some FATE boxes for those compilers
[22:10:30 CEST] <atomnuker> fine, but if I catch any of you defending that in 5 years I'll bring a gallon of dr pepper and forcibly make you drink that
[22:10:59 CEST] <atomnuker> no one should be using 25+ year old compilers
[22:11:35 CEST] <wbs> if there's no technical reason in 5 years why you shouldn't be able to use MSVC 2010, I'd still use the same argument
[22:11:45 CEST] <wbs> "zomg old, I don't think you should use it" is not a valid argument
[22:11:47 CEST] <rcombs> atomnuker: is that supposed to be a threat or a promise
[22:12:02 CEST] <wbs> (people shouldn't say what others should or should not use)
[22:12:13 CEST] <wbs> if there's a decent reason why one can't support it any longer, fine
[22:12:14 CEST] <rcombs> wbs: forces developers to keep old compatibility layers in place
[22:12:18 CEST] <atomnuker> ^^!!
[22:13:16 CEST] <wbs> rcombs: sure. the c99conv setup stuff in configure, which afaik is the only thing you could drop, hasn't required any maintainance yet in the 4 years in the codebase
[22:13:39 CEST] <atomnuker> if things get really bad I'll say "no one MUST use X right now"
[22:13:48 CEST] <atomnuker> which isn't the same as saying they shouldn't
[22:13:58 CEST] <wbs> c99conv itself isn't maintained much though, so if one runs into issues with it that myself or someone else can't be arsed to fix, then sure
[22:14:08 CEST] <wbs> (I've done a few minor fixes on it so far)
[22:14:15 CEST] <jamrial> atomnuker: every other trac ticket has some user with something like gcc 4.2 or older on some obscure arch that replies "it works, we won't be changing the compiler" when asked to upgrade
[22:14:23 CEST] <jamrial> some people love their ancient setups :p
[22:14:33 CEST] <atomnuker> it will break some day
[22:14:54 CEST] <atomnuker> 4.2 isn't that ancient
[22:15:04 CEST] <jamrial> ten years old by now
[22:15:05 CEST] <wbs> gcc 4.2 is way way way more ancient than MSVC 2010
[22:15:08 CEST] <atomnuker> we still have a 12 year old gcc version on fate right now
[22:15:21 CEST] <fritsch> does ffmpeg still run on windows xp?
[22:15:22 CEST] <atomnuker> gcc 2 or gcc 3, can't remember
[22:15:29 CEST] <atomnuker> yeah it does
[22:15:32 CEST] <fritsch> hehe
[22:15:45 CEST] <fritsch> and you like to fix newest compiler bugs on ancient windows xp?
[22:15:52 CEST] <atomnuker> why shouldn't it, it's just tossing number around, not asking the kernel if there's a sound blaster installed
[22:16:03 CEST] <nevcairiel> there arent really that many features that would force a update to a newer windows
[22:16:13 CEST] <rcombs> hmm, I was going to point out threads and atomics
[22:16:21 CEST] <andrey_turkin_> using c11 <- not going to happen ever unless there is c11toc89?
[22:16:24 CEST] <fritsch> yeah, i was also after threads and atomics
[22:16:27 CEST] <rcombs> but apparently MSVC still doesn't support those in C
[22:16:34 CEST] <nevcairiel> we needed an emulation for pthread_once at one point, but that was really not that complicated
[22:16:36 CEST] <rcombs> 5 years later
[22:16:46 CEST] <fritsch> kodi has it until today
[22:16:58 CEST] <rcombs> why is MSVC so bad
[22:17:09 CEST] <nevcairiel> MS just doesnt care about C, they are a C++ shop
[22:17:16 CEST] <andrey_turkin_> looks like their (valuable) customers don't care about c99
[22:17:18 CEST] <nevcairiel> not that new C features are really that good
[22:17:27 CEST] <nevcairiel> the "modern" C features have awkward syntax
[22:17:31 CEST] <nevcairiel> like it wants to be C++ but cant
[22:17:37 CEST] <atomnuker> wut?
[22:17:40 CEST] <atomnuker> which features?
[22:17:44 CEST] <rcombs> well C threads are useless if you already have pthreads
[22:17:52 CEST] <rcombs> just, MSVC doesn't because ??????
[22:18:01 CEST] <nevcairiel> pthreads is not C :p
[22:18:03 CEST] <iive> will there ever be templetes for C ?
[22:18:14 CEST] <jamrial> fritsch: i think you can get ffmpeg working on windows 98 if you compile using and old mingw build
[22:18:23 CEST] <rcombs> so they added threads to C and I'm like "oh hey maybe MSVC will have standard threading now"
[22:18:24 CEST] <rcombs> nope
[22:18:42 CEST] <JEEB> win9x has djgpp
[22:18:47 CEST] <JEEB> or well, dos/win9x
[22:19:13 CEST] <nevcairiel> rcombs: thats C11 either way, noone comes even close to claiming msvc supports that
[22:19:45 CEST] <nevcairiel> atomnuker: those C11 things like _Generic, just looks weird for C
[22:19:56 CEST] <rcombs> nevcairiel: I'm just -_- because what was the point of adding it if it doesn't do anything pthreads don't and isn't supported on anything that didn't already have pthreads
[22:20:32 CEST] <nevcairiel> C threads are useless either way, the standardization comes decades too late to matter
[22:20:37 CEST] <atomnuker> nevcairiel: generics are basically a preprocessor macro - if variable is float use this bit of code, else...
[22:20:46 CEST] <nevcairiel> atomnuker: i know what they are, thats not the point
[22:20:53 CEST] <nevcairiel> i find the look stupid in C
[22:20:56 CEST] <nevcairiel> they*
[22:21:02 CEST] <rcombs> nevcairiel: they would've been useful if MSVC had added support
[22:21:15 CEST] <andrey_turkin_> frankly it would make more sense to switch to c++11 than to c11 though
[22:21:31 CEST] Action: andrey_turkin_ ducks
[22:21:35 CEST] <rcombs> then you could switch code to use C threads instead of pthreads and have support everywhere [with modern compilers]
[22:21:44 CEST] <atomnuker> but yeah, they're kinda weird, but that's what they could do to make it macro-like I suppose
[22:22:32 CEST] <andrey_turkin_> I think they had to call them these strange names just to make sure it doesn't conflict with already existing names. Basically everyone already has "Thread" in their namespace, right? )
[22:22:36 CEST] <atomnuker> anything else (that they haven't removed yet because they realized in C1X+1 it's a bad feature)
[22:23:15 CEST] <nevcairiel> the entire naming scheme is odd with _Identifier scheme, but it probably really is for silly reasons like conflict avoidance
[22:23:36 CEST] <rcombs> the user should never have to deal with _Stuff, though
[22:23:39 CEST] <andrey_turkin_> iirc that's the reserved portion of namespace so they had to
[22:23:52 CEST] <rcombs> since the headers will map normal names to them when included
[22:27:32 CEST] <rcombs> like how `<tgmath.h>` does most of the things you'd want to do with _Generic
[22:28:30 CEST] <nevcairiel> its an awkward way to implement overloading :p
[22:28:59 CEST] <rcombs> yeah, well, C
[22:29:25 CEST] <rcombs> _Generic is unusual in not having a macro; I guess because it's macro-like itself
[22:31:33 CEST] <RiCON> marton balint around here?
[22:59:28 CEST] <Compn> RiCON : i dont see him in this irc chan
[23:09:05 CEST] <RiCON> --enable-decklink building seems broken, i'll bisect before creating the ticket
[23:13:01 CEST] <rkern> RiCON, does the error have something to do with c++11?
[23:15:46 CEST] <RiCON> not sure? https://i.fsbn.eu/cUtp.txt
[23:17:06 CEST] <nevcairiel> rkern: nah he doesnt build on osx
[23:23:31 CEST] <RiCON> 8f9fa49 is the first bad commit
[23:28:06 CEST] <cone-970> ffmpeg 03Rick Kern 07master:729d82abae2f: configure: use c++98 for c++ files
[23:34:28 CEST] <RiCON> also fails with -std=c++98 btw
[23:40:08 CEST] <cone-970> ffmpeg 03Rick Kern 07release/3.1:18ce5a4d1b3f: configure: use c++98 for c++ files
[00:00:00 CEST] --- Mon Jun 27 2016



More information about the Ffmpeg-devel-irc mailing list