[Ffmpeg-devel-irc] ffmpeg-devel.log.20121016

burek burek021 at gmail.com
Wed Oct 17 02:05:02 CEST 2012


[00:14] <ubitux> michaelni: any neat trick to check for the overflow in swfdec?
[00:14] <ubitux> given w and h being int (at most 16 bits), an >= 0 check added
[00:14] <ubitux> and*
[00:15] <ubitux> (and linesize with max w * 4)
[00:16] <ubitux> mmh maybe av_check_image_size()..
[00:18] <ubitux> av_image_check_size*
[00:25] <ubitux> my brain doesn't work anymore
[00:25] <ubitux> 'night ppl :)
[01:16] <cone-200> ffmpeg.git 03Michael Niedermayer 07db3a0aae9fdc: msvideo1enc: fix initial keyframe value * 03http://tinyurl.com/8qubk8e03
[01:16] <cone-200> ffmpeg.git 03Michael Niedermayer 07f5581266f767: ra288: assert order to be withinn supported range in do_hybrid_window() * 03http://tinyurl.com/8vkvw2p03
[01:16] <cone-200> ffmpeg.git 03Michael Niedermayer 07901f4bb6fc70: utils: consider mpeg4 in mp4/mov to have a unreliable timebase * 03http://tinyurl.com/8t22sks03
[01:16] <cone-200> ffmpeg.git 03Michael Niedermayer 07ba31e59e83ca: ffmpeg: Override r_frame_rate by corrected timebase only for interlaced streams * 03http://tinyurl.com/8ah7zst03
[02:25] <cone-200> ffmpeg.git 03Michael Niedermayer 071fe45903b81f: msvideo1enc: fix interframe encoding * 03http://tinyurl.com/c4y7de203
[03:15] <cone-200> ffmpeg.git 03Michael Niedermayer 073896cd11a107: mxfenc: explicitly truncate continuity count. * 03http://tinyurl.com/brad5pq03
[03:15] <cone-200> ffmpeg.git 03Michael Niedermayer 079c669672c7fd: x86/motion_est: widen before multiply. * 03http://tinyurl.com/cy5opwz03
[04:06] <cone-200> ffmpeg.git 03Michael Niedermayer 074cc4ca584736: mandelbrot: fix inner=period coloring routine * 03http://tinyurl.com/bszbv6c03
[04:17] <cone-200> ffmpeg.git 03Andrew Euell 0736b3b2376dd9: gitignore: add *.dylib for those of us on Darwin/OS X systems. * 03http://tinyurl.com/8sgr8ux03
[10:50] <ubitux> michaelni: question about a very old commit :)
[10:50] <ubitux> c45388b1
[10:50] <ubitux> "if(s->streams[i]->codec->codec_tag == 5){"
[10:50] <ubitux> is this a special case for nellymoser?
[10:51] <ubitux> ah, i missed that this chunk was already present..
[10:53] <ubitux> 11a8e4257 then :)
[10:55] <ubitux> anyway, looks like a nellymoser 8khz special case to me
[10:56] <ubitux> mmh or maybe vp6...
[10:56] <ubitux> yes vp6a actually :p
[11:18] <ubitux> nyuhu: there is no virtualdub repository?
[11:18] <ubitux> only tarballs?
[11:50] <nyuhu> ubitux : theres the source code on sourceforge&
[11:53] <ubitux> sourcecode right
[11:53] <ubitux> but what about the history?
[11:57] <nyuhu> theres a changelog with a brief of whats new since previous version :/
[11:58] <nyuhu> +the
[12:51] <ubitux> nyuhu: ok :(
[12:51] <ubitux> it's too bad because it could have given some hints about the constants
[13:03] <nyuhu> yes&
[13:21] <ubitux> nyuhu: it looks like random to me
[13:21] <ubitux> or at least "designed" for random
[13:22] <ubitux> and "jran" sounds like j random
[13:22] <ubitux> http://sprott.physics.wisc.edu/phys505/lect07.htm
[13:22] <ubitux> Example:  A = 1366, B = 150889, C = 714025
[13:22] <ubitux> looks similar to your constants
[13:24] <ubitux> nyuhu: and they looks totally constant
[13:24] <cone-412> ffmpeg.git 03Diego Biurrun 077e68c91e247b: rmdec: Move SIPR code shared with Matroska demuxer to a separate file * 03http://tinyurl.com/8hv6tec03
[13:24] <cone-412> ffmpeg.git 03Luca Barbato 070fbb62a8e6c6: build: support asan and tsan toolchain shortcuts * 03http://tinyurl.com/9hhcz5703
[13:24] <cone-412> ffmpeg.git 03Mans Rullgard 074c995fafd861: configure: simplify get_version() function * 03http://tinyurl.com/9p8xsgv03
[13:24] <cone-412> ffmpeg.git 03Michael Niedermayer 079270a2b374ea: Merge commit '4c995fafd861f537360b3717901cdbed6a6844e7' * 03http://tinyurl.com/9cn7fqp03
[13:24] <ubitux> nyuhu: so you could remove them from the context
[13:30] <cone-412> ffmpeg.git 03Mans Rullgard 0766a1ccd7467a: configure: simplify argument handling in check_ld * 03http://tinyurl.com/crpncb703
[13:30] <cone-412> ffmpeg.git 03Michael Niedermayer 07d15be9108e6a: Merge commit '66a1ccd7467ab1913cd8877114c6d4c2588bb12f' * 03http://tinyurl.com/c5yu9bh03
[13:35] <cone-412> ffmpeg.git 03Stefano Sabatini 077ca102a7d7bd: examples: add README file with simple compilation instructions * 03http://tinyurl.com/ch82q5f03
[13:52] <cone-412> ffmpeg.git 03Stefano Sabatini 0725f3827e1daf: tools/ffeval: do not use UNIX-specific /dev/std{in,out} files * 03http://tinyurl.com/8w6c56g03
[13:56] Action: ubitux just founds something pretty funny with vp6 and flv
[14:01] <cone-412> ffmpeg.git 03Mans Rullgard 0784e65c763d8f: build: simplify enabling of compat objects * 03http://tinyurl.com/9ogafhq03
[14:01] <cone-412> ffmpeg.git 03Diego Biurrun 07c1fcfdec7546: rangecoder-test: Return in case of an error * 03http://tinyurl.com/9nnmaxu03
[14:01] <cone-412> ffmpeg.git 03Michael Niedermayer 075717562c789f: Merge commit 'c1fcfdec75468009dc7de29a5d1c6adf3b2ef77d' * 03http://tinyurl.com/9jxyv6j03
[14:02] <ubitux> ./ffmpeg -i ~/fate-samples/ea-vp6/g36.vp6 -c copy -y out.flv  output is rotated %)
[14:04] <JEEB> yeah, I think VP62F or whatever was actually upside down or whatever
[14:05] <ubitux> it works fine when muxed in mkv
[14:07] <cone-412> ffmpeg.git 03Diego Biurrun 07f1be514540dc: rangecoder-test: Drop timer output that clutters stderr * 03http://tinyurl.com/ccl3zmt03
[14:07] <cone-412> ffmpeg.git 03Mans Rullgard 07099294577c8c: fate: improve md5sum utility selection * 03http://tinyurl.com/dy2f3f403
[14:07] <cone-412> ffmpeg.git 03Mans Rullgard 073dc06b6972cf: tiny_psnr: check for specified sample size less than 1 * 03http://tinyurl.com/cj9cosd03
[14:07] <cone-412> ffmpeg.git 03Michael Niedermayer 0783962004f727: Merge commit '3dc06b6972cf389269e9c36ff0a4373f80f7149b' * 03http://tinyurl.com/bt4eqq703
[14:09] <ubitux> and same in mov
[14:13] <JEEB> well, there was some weird thing with muxing vp6X into flv, that's all I remember
[14:13] <JEEB> which I remember including upside-downess
[14:18] <ubitux> :(
[14:19] <cone-412> ffmpeg.git 03Mans Rullgard 07bf868c4a9bc6: tiny_psnr: fix range calculation for sample size of 32 bits * 03http://tinyurl.com/bn8839n03
[14:19] <cone-412> ffmpeg.git 03Mans Rullgard 0707b3790d3626: build: simplify linking tools with cmdutils.o * 03http://tinyurl.com/ck8hlzl03
[14:19] <cone-412> ffmpeg.git 03Michael Niedermayer 078227d36bd356: Merge remote-tracking branch 'qatar/master' * 03http://tinyurl.com/d9z5tfs03
[14:29] <burek> I'm sorry to bother you guys again, but is there any way I could modify the flv muxer to be _really_ verbose with its err output log, so that I can see what doesn't work in vlc, when it is being used through the libav linkage
[14:30] <burek> it seems it doesn't mux well the raw h264 stream (coming from the webcam, encoded already, no need to re-encode) with alsa input (encoded at the same time with vlc)
[14:31] <burek> described in ffmpeg's vocabulary, it would be like: ffmpeg -f h264 -i /dev/video0 -f alsa -i hw:0 -c:v copy -c:a libaacplus output.flv
[14:32] <burek> that's basically what vlc is supposed to be doing prior to http streaming
[16:36] <cone-412> ffmpeg.git 03Anton Khirnov 071846f3b5b15f: avconv: fix -force_key_frames * 03Error03
[16:36] <cone-412> ffmpeg.git 03Reimar Döffinger 078efae4cbbf51: avconv: fix parsing of -force_key_frames option. * 03http://tinyurl.com/co7nb5303
[16:36] <cone-412> ffmpeg.git 03Kostya Shishkov 0702b723946279: vc1dec: add flush function for WMV9 and VC-1 decoders * 03http://tinyurl.com/c67hjwo03
[16:36] <cone-412> ffmpeg.git 03Reimar Döffinger 07e9ac06160f45: sipr: fall back to setting mode based on bit_rate. * 03http://tinyurl.com/bllgnts03
[16:36] <cone-412> ffmpeg.git 03Anton Khirnov 07bed5847563d1: bmpdec: only initialize palette for pal8. * 03http://tinyurl.com/dx7q6o903
[16:36] <cone-412> ffmpeg.git 03Janne Grunau 07fdb708078174: Revert "nuv: check per-frame header for validity." * 03http://tinyurl.com/bmgqeyr03
[16:36] <cone-412> ffmpeg.git 03Janne Grunau 076704522ca9dd: nuv: check RTjpeg header for validity * 03http://tinyurl.com/d6t2ctq03
[16:36] <cone-412> ffmpeg.git 03Max Lazarov 0725a1a5b1b38b: eval: fix swapping of lt() and lte() * 03http://tinyurl.com/bt52k7b03
[16:36] <cone-412> ffmpeg.git 03Janne Grunau 077a7229b52d19: imgconvert: avoid undefined left shift in avcodec_find_best_pix_fmt * 03http://tinyurl.com/d6xpmor03
[16:36] <cone-412> ffmpeg.git 03Michael Niedermayer 07da0c45766347: mpegvideo: Don't use ff_mspel_motion() for vc1 * 03http://tinyurl.com/cfjwvtb03
[16:36] <cone-412> ffmpeg.git 03Anton Khirnov 077124fa5d3640: lavf: don't segfault when a NULL filename is passed to avformat_open_input() * 03http://tinyurl.com/cwoswto03
[16:36] <cone-412> ffmpeg.git 03Justin Ruggles 07d9ffa2aca1e4: golomb: check remaining bits during unary decoding in get_ur_golomb_jpegls() * 03http://tinyurl.com/czcvgcg03
[16:36] <cone-412> ffmpeg.git 03Michael Niedermayer 07a1b127515bb7: alsdec: check opt_order. * 03http://tinyurl.com/d4mzppa03
[16:36] <cone-412> ffmpeg.git 03Martin Storsjö 076d1b91324c30: h263: Add ff_ prefix to nonstatic symbols * 03http://tinyurl.com/cne5wp603
[16:36] <cone-412> ffmpeg.git 03Martin Storsjö 078c0bbe51561b: vlc/rl: Add ff_ prefix to the nonstatic symbols * 03http://tinyurl.com/ctscotf03
[16:36] <cone-412> ffmpeg.git 03Janne Grunau 07f695bd601640: rv34: use AVERROR return values in ff_rv34_decode_frame() * 03http://tinyurl.com/d4jy28603
[16:36] <cone-412> ffmpeg.git 03Janne Grunau 0790575bd7dd1e: rv34: Handle only complete frames in frame-mt. * 03http://tinyurl.com/dyq2hpy03
[16:36] <cone-412> ffmpeg.git 03Janne Grunau 07b1ad5a21da7a: rv34: error out on size changes with frame threading * 03http://tinyurl.com/bwlxpj903
[16:36] <cone-412> ffmpeg.git 03Mina Nagy Zaki 07e5f4e2494228: lavfi: avfilter_merge_formats: handle case where inputs are same * 03http://tinyurl.com/dxgkwhj03
[16:36] <cone-412> ffmpeg.git 03Kostya Shishkov 07d4f3abca6a76: indeo3: validate new frame size before resetting decoder * 03http://tinyurl.com/clgh5ex03
[16:36] <cone-412> ffmpeg.git 03Justin Ruggles 07e46cf805b100: vorbisenc: check all allocations for failure * 03http://tinyurl.com/c8np5qz03
[16:36] <cone-412> ffmpeg.git 03Alex Converse 079aaaeba45c41: vorbis: Validate that the floor 1 X values contain no duplicates. * 03http://tinyurl.com/ccj2rfu03
[16:36] <cone-412> ffmpeg.git 03Martin Storsjö 0731bc3fb563b1: snow: Check mallocs at init * 03http://tinyurl.com/cf4vuur03
[16:36] <cone-412> ffmpeg.git 03Michael Niedermayer 07be2dd2559f54: Merge remote-tracking branch 'qatar/release/0.8' into release/0.10 * 03http://tinyurl.com/cq2nks603
[17:12] <burek> michaelni, what is the usual way of sending patches for ffmpeg-web
[17:12] <burek> ml/pastebin/..?
[17:22] <ubitux> burek: ml
[17:22] <ubitux> classic ffmpeg-devel
[17:22] <burek> ok thx :)
[17:27] <burek> btw, should I attach patch or just paste it into the mail body?
[17:28] <ubitux> ideally git send-email
[17:28] <burek> oh :$
[17:28] <ubitux> and no, never paste it
[17:28] <ubitux> as an attachment is fine, but it's nice if it's inlined
[17:30] <burek> hm, i dont have git-send-email ?
[17:30] <burek> is that normal?
[17:30] <nevcairiel> git send-email, no - after git
[17:30] <burek> git: 'send-email' is not a git command. See 'git --help'.
[17:31] <nevcairiel> how old is your git? =P
[17:31] <burek> there is an entry in man git about git-send-email(1)
[17:31] <burek> 1.7.10
[17:32] <burek> solved
[17:32] <burek> debian provides it in separate package it seems
[17:35] <nevcairiel> how i hate debian and the practice of breaking everything up into meaningless packages =p
[17:35] <burek> it is easier for them to maintain it, i guess
[17:35] <nevcairiel> how so, all these git tools get updated as one package in upstream
[17:35] <burek> don't know, im no debian maintainer :)
[17:36] <burek> just guessing
[17:36] <burek> nice :) still no git-send-email :D
[17:36] <burek> how more stupid this can be
[17:36] <Compn> nevcairiel : you mean splitting up -dev headers from libs ?
[17:37] <nevcairiel> no, apparently splitting git in multiple packages
[17:37] <Compn> 'why wont xxx compile?' - 'because debian'
[17:37] <nevcairiel> i can understand splitting off dev headers
[17:37] <nevcairiel> normal users dont need them
[17:37] <burek> i had to do apt-get install git-email to be able to use git-send-email, but through the use of: git send-email :)
[17:37] <burek> shortly :D
[17:38] <burek> so, git send-email --to=... --subject=... 0001-blabla.patch
[17:38] <burek> right?
[17:38] <nevcairiel> you give it the revisions to send, not a patch
[17:39] <burek> wait, I don't need git format-patch?
[17:39] <nevcairiel> no
[17:45] <burek> f it.. i dont have time to setup everything including smtp
[17:46] <burek> i added it as attachment and sent it manually
[17:49] <ubitux> < nevcairiel> you give it the revisions to send, not a patch // you can as well
[17:58] <cone-412> ffmpeg.git 03Mans Rullgard 07b102d5d97dae: h264: allow cropping to AVCodecContext.width/height * 03http://tinyurl.com/cdxflrt03
[17:58] <cone-412> ffmpeg.git 03Mans Rullgard 070054d70f23ed: mov: set AVCodecContext.width/height for h264 * 03http://tinyurl.com/cwa2v7n03
[17:58] <cone-412> ffmpeg.git 03Anton Khirnov 07a60eb6ef12df: ffmpeg: fix -force_key_frames * 03http://tinyurl.com/bqgpvdr03
[17:58] <cone-412> ffmpeg.git 03Kostya Shishkov 070173a7966b33: vc1dec: add flush function for WMV9 and VC-1 decoders * 03http://tinyurl.com/c9c7e6p03
[17:58] <cone-412> ffmpeg.git 03Janne Grunau 07f31170d4e7f9: nuv: check RTjpeg header for validity * 03http://tinyurl.com/buc3k3603
[17:58] <cone-412> ffmpeg.git 03Janne Grunau 078812b5f16410: imgconvert: avoid undefined left shift in avcodec_find_best_pix_fmt * 03http://tinyurl.com/bomlrjp03
[17:58] <cone-412> ffmpeg.git 03Michael Niedermayer 07899d95efe12f: mpegvideo: Don't use ff_mspel_motion() for vc1 * 03http://tinyurl.com/crnx9ug03
[17:58] <cone-412> ffmpeg.git 03Anton Khirnov 0777d43bf42d76: lavf: don't segfault when a NULL filename is passed to avformat_open_input() * 03http://tinyurl.com/c68lp5803
[17:58] <cone-412> ffmpeg.git 03Michael Niedermayer 07b6ba39f931a8: alsdec: check opt_order. * 03http://tinyurl.com/d8rscdg03
[17:58] <cone-412> ffmpeg.git 03Mina Nagy Zaki 07b6c5848a1f8f: lavfi: avfilter_merge_formats: handle case where inputs are same * 03http://tinyurl.com/cr8sesx03
[17:58] <cone-412> ffmpeg.git 03Justin Ruggles 0761ece4137298: vorbisenc: check all allocations for failure * 03http://tinyurl.com/brmmq4503
[17:58] <cone-412> ffmpeg.git 03Alex Converse 07d6e250abfc36: vorbis: Validate that the floor 1 X values contain no duplicates. * 03http://tinyurl.com/d6u6z2a03
[17:58] <cone-412> ffmpeg.git 03Michael Niedermayer 07e28814e0e16c: Merge remote-tracking branch 'qatar/release/0.7' into release/0.8 * 03http://tinyurl.com/ce6cgz703
[18:03] <cone-412> ffmpeg.git 03Mans Rullgard 074eea330a2afa: h264: allow cropping to AVCodecContext.width/height * 03http://tinyurl.com/c5s7kdw03
[18:03] <cone-412> ffmpeg.git 03Mans Rullgard 0744e6cf3f75f0: mov: set AVCodecContext.width/height for h264 * 03http://tinyurl.com/cnsdbjb03
[18:03] <cone-412> ffmpeg.git 03Kostya Shishkov 07aa4121276777: vc1dec: add flush function for WMV9 and VC-1 decoders * 03http://tinyurl.com/cpzwot903
[18:03] <cone-412> ffmpeg.git 03Janne Grunau 07459feb7cce03: nuv: check RTjpeg header for validity * 03http://tinyurl.com/c5l56co03
[18:03] <cone-412> ffmpeg.git 03Janne Grunau 07fd7426ed8985: imgconvert: avoid undefined left shift in avcodec_find_best_pix_fmt * 03http://tinyurl.com/c7bojmm03
[18:03] <cone-412> ffmpeg.git 03Michael Niedermayer 07c82ae85a8a78: mpegvideo: Don't use ff_mspel_motion() for vc1 * 03http://tinyurl.com/c9ukn4j03
[18:03] <cone-412> ffmpeg.git 03Mina Nagy Zaki 07a4e277312cac: lavfi: avfilter_merge_formats: handle case where inputs are same * 03http://tinyurl.com/dxjcxub03
[18:03] <cone-412> ffmpeg.git 03Alex Converse 070e2f415adf5d: vorbis: Validate that the floor 1 X values contain no duplicates. * 03http://tinyurl.com/d29ve7903
[18:03] <cone-412> ffmpeg.git 03Michael Niedermayer 07f48d1fb16720: Merge remote-tracking branch 'qatar/release/0.6' into release/0.6 * 03http://tinyurl.com/coj95b503
[18:17] <cone-412> ffmpeg.git 03Kostya Shishkov 079125aa9218c3: vc1dec: add flush function for WMV9 and VC-1 decoders * 03http://tinyurl.com/dya8op403
[18:17] <cone-412> ffmpeg.git 03Janne Grunau 07f695be22d89a: nuv: check RTjpeg header for validity * 03http://tinyurl.com/cafb5r503
[18:17] <cone-412> ffmpeg.git 03Janne Grunau 077296a6b5e9c2: imgconvert: avoid undefined left shift in avcodec_find_best_pix_fmt * 03http://tinyurl.com/c95tkjn03
[18:17] <cone-412> ffmpeg.git 03Michael Niedermayer 076d6373dc6441: mpegvideo: Don't use ff_mspel_motion() for vc1 * 03http://tinyurl.com/csycyuy03
[18:17] <cone-412> ffmpeg.git 03Mina Nagy Zaki 070dfcbe5285f0: lavfi: avfilter_merge_formats: handle case where inputs are same * 03http://tinyurl.com/bwoltvy03
[18:17] <cone-412> ffmpeg.git 03Michael Niedermayer 07776fb2e10d4a: Merge remote-tracking branch 'qatar/release/0.5' into release/0.5 * 03http://tinyurl.com/d6665sl03
[19:04] <creep> fried potatoes or beans?
[19:46] <cone-412> ffmpeg.git 03Michael Niedermayer 07dd3e5baa5996: swr-test: check pointers before use * 03http://tinyurl.com/8saxjkw03
[19:46] <cone-412> ffmpeg.git 03Michael Niedermayer 079de5b11d58ed: swr-test: fix division by 0 * 03http://tinyurl.com/8doxm2g03
[20:12] <cone-412> ffmpeg.git 03Michael Niedermayer 07340305646a82: riff: fix division by zero for G726 * 03http://tinyurl.com/9jh7fzw03
[20:12] <cone-412> ffmpeg.git 03Michael Niedermayer 07de6c15044415: nut: fix int32 overflow * 03http://tinyurl.com/8ggtx8803
[20:12] <cone-412> ffmpeg.git 03Michael Niedermayer 0724d6af0f7e5e: sierravmd: flip reduce arg order. * 03http://tinyurl.com/9dvadjm03
[20:31] <cbsrobot> burek: you could also use the gmail smtp server
[21:47] <saste> michaelni: ping on "lavfi/scale: accept named option, make parsing more robust"
[21:57] <michaelni> saste, does it still work with changing input resolution after the patch ?
[21:57] <saste> I think so
[21:58] <saste> how can i test it?
[22:00] <michaelni> that is a good question
[22:03] <michaelni> you can try to concatenate 2 mpegs with different size and chec per avlog that the scale filter gets differenz input sizes while the whole filter graphs is NOT rebuild
[22:04] <saste> I believed we implemented normalization at some point, going to test it now
[22:05] <ubitux> yaaargl
[22:05] Action: ubitux realized something is wrong with the lavfi metadata patch :(
[22:06] <ubitux> we dup the dict in the AVFrame in the copy props
[22:06] <ubitux> but by default apps won't manually destroy the dict
[22:06] <ubitux> and actually, nothing will
[22:06] <ubitux> (except in case of the lavfi device with what we did)
[22:07] <ubitux> any suggestion? :(
[22:10] <saste> ubitux: why do you need to destroy the dict?
[22:11] <ubitux> because there is one allocated per frame?
[22:11] <ubitux> which will leak
[22:11] <michaelni> ubitux, "decode_frame" free last, dup current, "codec close" free last
[22:14] <ubitux> av_dict_free(&avctx->coded_frame->metadata);  so we would move this from raw decoders up to the general decode frame?
[22:15] <ubitux> ok let's try to activate my brain for once.
[22:25] <saste> michaelni, same behavior as current (that is broken)
[22:25] <saste> but somehow works
[22:25] <ubitux> michaelni: codec close can only access to "coded frame", which is likely no pointing on the last decoded frame most of the time afaict
[22:26] <saste> it is broken because is not doing the expected thing
[22:26] <michaelni> saste, how can i reproduce the "broken" ?
[22:26] <saste> when the size changes, it assumes the new size
[22:27] <michaelni> ubitux, you can store a pointer to the dict in the codec context
[22:27] <saste> for example if you have scale=2*iw:2*ih it will just keep the new "changed" size
[22:27] <ubitux> michaelni: right sure, ok
[22:27] <saste> i'd rather expect scale to resize the video accordingly
[22:27] <saste> but => unrelated to the patch
[22:30] <michaelni> saste, but its working if you use a fixed size like 400:300 ?
[22:35] <saste> michaelni, my patch is correct in the sense that it doesn't change the current behavior
[22:36] <saste> which is: when the size changes, assume the new input size as output, whatever expression the user set
[22:39] <saste> michaelni, BTW can you have a look at the examples/muxing issue?
[22:39] <saste> it is disturbing that we ship a broken example, and I don't know which is the best way to fix it
[22:43] <michaelni> about the scale patch, id like to look at things before its applied, if theres a bug it should be fixed before changing the code
[22:46] <michaelni> about the muxing issue, how can i reproduce it ?
[22:47] <saste> michaelni, muxing out.mov
[22:50] <saste> michaelni, i don't want to end up fixing the world before committing a simple change (which I have in my local repo since a few weeks)
[22:50] <saste> so if the patches doesn't change the current behavior, i'm for applying it
[23:16] <ubitux> puting a metadata in the codec context was actually so simple..
[23:16] <saste> life is a mess
[23:17] <saste> (^^ in case you need some useful and constructive feedback)
[23:18] <ubitux> :D
[23:35] <ubitux> yepee!
[23:35] <ubitux> it works, and no leak
[23:36] <ubitux> thx a lot michaelni, it was actually the simplest thing to do..
[00:00] --- Wed Oct 17 2012


More information about the Ffmpeg-devel-irc mailing list