[Ffmpeg-devel-irc] ffmpeg-devel.log.20171028

burek burek021 at gmail.com
Sun Oct 29 03:05:02 EEST 2017


[00:42:42 CEST] <cone-137> ffmpeg 03James Almer 07master:2d11218877a5: avcodec/audiotoolboxdec: fix ac3_parser header include
[00:44:23 CEST] <jamrial> michaelni: could you add the folder fmvc and its samples from libav's fate samples suite to ours?
[02:12:35 CEST] <michaelni> jamrial, uploaded
[02:12:42 CEST] <jamrial> michaelni: thanks
[02:27:46 CEST] <cone-137> ffmpeg 03Paul B Mahol 07master:95a8a03a1912: Add FM Screen Capture Codec decoder
[02:27:47 CEST] <cone-137> ffmpeg 03Diego Biurrun 07master:4d4d7cf9d539: fmvc: Add FATE tests
[02:27:48 CEST] <cone-137> ffmpeg 03James Almer 07master:b94eda7194c2: Merge commit '95a8a03a191204c3ca5c8cf6b69ab166765d5ff3'
[02:27:49 CEST] <cone-137> ffmpeg 03James Almer 07master:3c4c092774e3: Merge commit '4d4d7cf9d539a053f531f662a972b23d335738eb'
[03:22:23 CEST] <cone-137> ffmpeg 03Ganesh Ajjanagadde 07master:1fe858136b31: utvideodec: Prevent possible signed overflow
[03:22:24 CEST] <cone-137> ffmpeg 03James Almer 07master:7f9de7b416c9: Merge commit '1fe858136b315796dd5349f3b4448a29d1bd6fa1'
[05:14:32 CEST] <michaelni> jamrial, e0250cf3651e6417e0117486a7816b45fb2d34cd seems to break tools/probetest 256 4096 (segfault)
[05:15:33 CEST] Action: michaelni has to go to bed now, ping me tomorrow if you need more info / gdb whatever
[15:12:15 CEST] <cone-514> ffmpeg 03Carl Eugen Hoyos 07master:851829455889: lswr/swresample: Mention the actually supported formats when erroring out.
[15:52:39 CEST] <jamrial> michaelni: https://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavformat/ac3dec.c;h=8ea73824a684bcc45dccb3f7080322948ef86d96;hb=HEAD#l50
[15:52:51 CEST] <jamrial> you're increasing the buf2 pointer and not checking it afterwards in a loop where the condition to run is buf2 < end
[15:53:34 CEST] <jamrial> the change that started making probetest fail added a call to a function passing the amount of bytes that remain in the buffer using end - buf2
[15:55:26 CEST] <jamrial> i could change that "end - buf2" to a hardcoded 7 or 8, which is the max amount of bytes the function will read and seems to fix probetest
[15:55:46 CEST] <jamrial> but the correct thing to do i think is making sure buf2 is still < end after the code in the link above
[16:15:31 CEST] <michaelni> jamrial, ill add a check for buf2
[17:16:43 CEST] <cone-514> ffmpeg 03Bjorn Roche 07master:aba926e7d654: lavfi/paletteuse: fix to support transparency
[17:16:44 CEST] <cone-514> ffmpeg 03Clément BSsch 07master:5c8e904eb386: lavfi/paletteuse: move "new" option before debugging options
[17:42:07 CEST] <cone-514> ffmpeg 03Clément BSsch 07master:33e226f84d59: lavfi/paletteuse: simplify color get function prototypes
[17:56:34 CEST] <cone-514> ffmpeg 03Paul B Mahol 07master:3729ae659fb2: avcodec: add MagicYUV encoder
[17:58:19 CEST] <atomnuker> nice
[18:03:58 CEST] <cone-514> ffmpeg 03Clément BSsch 07master:3d547c19c211: lavfi/paletteuse: fix debug_accuracy after aba926e7d
[18:03:59 CEST] <cone-514> ffmpeg 03Clément BSsch 07master:157d08e971f4: lavfi/paletteuse: fix debug_mean_error after aba926e7d
[18:16:03 CEST] <cone-514> ffmpeg 03James Almer 07master:26ad6df1ced7: avcodec: bump minor after addition of MagicYUV encoder
[18:16:04 CEST] <cone-514> ffmpeg 03James Almer 07master:93a1aac44198: configure: add missing llvidencdsp dependency to MagicYUV encoder
[18:28:59 CEST] <durandal_1707> abi is still unstable so bumping things is futile
[18:30:04 CEST] <jamrial> perhaps, but it doesn't hurt either way
[18:50:01 CEST] <cone-514> ffmpeg 03Clément BSsch 07master:1d348dd03fb4: lavfi/paletteuse: fix debug_kdtree after aba926e7d
[19:03:26 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:19c5fa9eea47: golomb: Simplify get_ur_golomb_jpegls
[19:03:27 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:9e4a5eb51b9f: avformat: Free the internal codec context at the end
[19:03:28 CEST] <cone-514> ffmpeg 03James Almer 07master:f4bda7e30538: Merge commit '19c5fa9eea47d45b5eedce3bff4edeaf83fb3942'
[19:03:29 CEST] <cone-514> ffmpeg 03James Almer 07master:1e5669457f60: Merge commit '9e4a5eb51b9f3b2bff0ef08e0074b7fe4893075d'
[19:10:03 CEST] <atomnuker> how does video thread encoding work?
[19:10:20 CEST] <atomnuker> does each thread get a different priv_context or a different avctx or different both?
[19:14:23 CEST] <cone-514> ffmpeg 03Paul B Mahol 07master:c5230955640c: utvideodec: Support UQY2
[19:14:24 CEST] <cone-514> ffmpeg 03Paul B Mahol 07master:030c8be7a2c0: pixfmt: Add gbrap10 pixel format
[19:14:25 CEST] <cone-514> ffmpeg 03Paul B Mahol 07master:f6a9c20a52d6: swscale: Add input support for gbrap10 pixel format
[19:14:26 CEST] <cone-514> ffmpeg 03Michael Niedermayer 07master:ac29b82ec596: swscale: Add gbrap10 output
[19:14:27 CEST] <cone-514> ffmpeg 03Paul B Mahol 07master:a93faf30d688: utvideodec: Support UQRA and UQRG
[19:14:28 CEST] <cone-514> ffmpeg 03Paul B Mahol 07master:4f33d9d41a03: utvideodec: Support ULY4 and ULH4
[19:14:29 CEST] <cone-514> ffmpeg 03Paul B Mahol 07master:9227bd8ac2f2: utvideodec: Reuse the huffyuv add_left
[19:14:30 CEST] <cone-514> ffmpeg 03Paul B Mahol 07master:378460fef1e5: utvideodec: Support for gradient prediction
[19:14:31 CEST] <cone-514> ffmpeg 03James Almer 07master:4205e20c2222: Merge commit '378460fef1e563704bc5a307ba748a2b819cc09d'
[19:30:21 CEST] <jamrial> michaelni: do you prefer how f56fa95cd1 (libav) handles left shift of negative value, or how we're currently doing it?
[19:30:31 CEST] <jamrial> ours is cleaner looking, IMO
[19:30:56 CEST] <ubitux> i agree
[19:31:12 CEST] <ubitux> if we didn't forget any, i think this commit is better noop'ed
[19:31:57 CEST] <jamrial> ubitux: we even have a bunch of other cases this commit doesn't handle that i'd have to manually change, so yeah, i'm leaning towards nooping it
[19:32:58 CEST] <ubitux> SHIFT_LEFT not actually doing generating a shift instruction but a mult instead is absurd :p
[19:33:18 CEST] <ubitux> the shift is not even the meaning of what the code wants, it's the optimization
[19:33:35 CEST] <ubitux> it's a DO_OPTIMIZATION() ... // i lied
[19:34:04 CEST] <michaelni> jamrial, +1, ours as its cleaner
[19:34:05 CEST] <ubitux> and given we have these cases all over the codebase, i don't see why it should be local to that file
[19:36:12 CEST] <cone-514> ffmpeg 03dxfhgwet 07master:b43d13144bdb: avfilter/zscale: fix segfault on library error
[19:38:56 CEST] <cone-514> ffmpeg 03dxfhgwet 07master:ff763351e745: avfilter/zscale: fix memory leak
[19:39:13 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:f56fa95cd13f: swscale: Do not shift negative values directly
[19:39:14 CEST] <cone-514> ffmpeg 03James Almer 07master:f361f15f362e: Merge commit 'f56fa95cd13f627891a1bfb66bf61b971b9e0238'
[19:39:39 CEST] <jamrial> yikes, push race
[19:40:09 CEST] <jamrial> cone says it's pushed but it was not
[19:40:16 CEST] <ubitux> :D
[19:40:53 CEST] <jamrial> oh, they are in the tree, just not as part of master, haha
[19:41:04 CEST] <jamrial> nothing references them, so a git gc is in order
[19:41:27 CEST] <jamrial> see https://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=f361f15f362e
[19:41:47 CEST] <nevcairiel> it happens, it should clean itself eventually
[19:43:48 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:f56fa95cd13f: swscale: Do not shift negative values directly
[19:43:49 CEST] <cone-514> ffmpeg 03James Almer 07master:27050d4836b1: Merge commit 'f56fa95cd13f627891a1bfb66bf61b971b9e0238'
[19:52:50 CEST] <jamrial> michaelni, ubitux: 37f573543c yay or nay? we don't have the warnings mentioned there because we cast the argument
[19:54:09 CEST] <ubitux> i don't really see the improvement, no opinion
[19:58:17 CEST] <atomnuker> JEEB: are you going to push the EOF patch?
[20:00:08 CEST] <jamrial> atomnuker: he said he'd be mostly afk until tomorrow
[20:04:14 CEST] <michaelni> jamrial, functions are better/cleaner than macros
[20:04:20 CEST] <michaelni> IMHO that is
[20:07:34 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:37f573543c4f: swscale: Convert the check check_image_pointers helper to a macro
[20:07:35 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:5352802da81f: dca: Account for lfe when checking for the channel count
[20:07:36 CEST] <cone-514> ffmpeg 03James Almer 07master:dea0f2b36b98: Merge commit '37f573543c4fd7f44339e04d8d15b95118493ddd'
[20:07:37 CEST] <cone-514> ffmpeg 03James Almer 07master:7d829da3f919: Merge commit '5352802da81f2083e65d466612e639a4e6e5530e'
[20:09:11 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:4809781586d1: travis: Exclude gcc for the macOS target
[20:09:12 CEST] <cone-514> ffmpeg 03James Almer 07master:16a17fb6b259: Merge commit '4809781586d1c956005f72946a2aab5915eab350'
[20:28:32 CEST] <cone-514> ffmpeg 03Peter Große 07master:3606c114172e: dashenc: use av_dict_set_int instead of static function
[20:28:33 CEST] <cone-514> ffmpeg 03Peter Große 07master:8edb9d457251: dashenc: move UTCTiming element to the end of the manifest
[20:28:34 CEST] <cone-514> ffmpeg 03Sasi Inguva 07master:801375311395: lavf/mov.c: Fix parsing of edit list atoms with invalid elst entry count.
[20:28:35 CEST] <cone-514> ffmpeg 03Michael Niedermayer 07master:eb54efc1e1aa: avformat/ac3dec: Check buf2 before adding 16 in ac3_eac3_probe()
[20:30:09 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:279e3aaa14da: flv: Validate the packet size
[20:30:10 CEST] <cone-514> ffmpeg 03James Almer 07master:220603d0c83e: Merge commit '279e3aaa14daba6b7a37d75f3fb6e29c732d123f'
[20:56:27 CEST] <jamrial> michaelni: can't reproduce the failure f2c469b73f is trying to "fix"
[20:56:30 CEST] <jamrial> do we support what that commit claims is unsupported?
[20:57:35 CEST] <jamrial> for that matter, avi error messages are funny
[20:57:38 CEST] <jamrial> [avi @ 0000000002f27100] crazy start time, iam scared, giving up
[20:58:59 CEST] <Compn> lol
[21:56:54 CEST] <michaelni> jamrial, yes, we support lossless rgba 
[22:01:36 CEST] <cone-514> ffmpeg 03Marton Balint 07master:859144fb1fff: fate: change fate-ffmpeg-attached_pics to encode to pcm_s16le
[23:10:08 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:f2c469b73f8a: mjpeg: Report non-3 component rgb lossless as not supported
[23:10:09 CEST] <cone-514> ffmpeg 03James Almer 07master:1232b349175c: Merge commit 'f2c469b73f8ac3dd6a98d38281f61b68ea6dd336'
[23:18:09 CEST] <cone-514> ffmpeg 03Luca Barbato 07master:1731c3530bff: mm: Skip unexpected audio packets
[23:18:10 CEST] <cone-514> ffmpeg 03James Almer 07master:495d3b2cc3fa: Merge commit '1731c3530bffb876deb9e00dfffdf9841a8412cd'
[23:25:21 CEST] <cone-514> ffmpeg 03Martin Storsjö 07master:e00db9f78bb4: checkasm: hevc: Add a hevc_ prefix to the add_residual functions
[23:25:22 CEST] <cone-514> ffmpeg 03James Almer 07master:f568d9d0baff: Merge commit 'e00db9f78bb475ed5103364f61892f4e75ef89ba'
[23:32:24 CEST] <cone-514> ffmpeg 03Aaron Levinson 07master:b22094d74901: qsvenc: Use MFXVideoENCODE_Query() to update the parameters
[23:32:25 CEST] <cone-514> ffmpeg 03Aaron Levinson 07master:8fd8f91e47f3: qsvenc: Make sure the interlaced encoding works
[23:32:26 CEST] <cone-514> ffmpeg 03James Almer 07master:49fe0ecd405d: Merge commit '8fd8f91e47f33cd82371a97ac81afc476144964f'
[00:00:00 CEST] --- Sun Oct 29 2017


More information about the Ffmpeg-devel-irc mailing list