[Ffmpeg-devel-irc] ffmpeg-devel.log.20120906

burek burek021 at gmail.com
Fri Sep 7 02:05:02 CEST 2012


[00:44] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r79dcd58d83 10ffmpeg/libavutil/parseutils.c: 
[00:44] <CIA-56> ffmpeg: lavu/parseutils: make small_strptime() return a (non const) char *
[00:44] <CIA-56> ffmpeg: This is consistent with the standard definition of strptime().
[00:44] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * rfd5293d216 10ffmpeg/libavfilter/vf_tinterlace.c: 
[00:44] <CIA-56> ffmpeg: lavfi/tinterlace: set inlink->cur to NULL, since it is stored internally
[00:44] <CIA-56> ffmpeg: If not set to NULL, the reference is freed by ff_end_frame(), and later
[00:44] <CIA-56> ffmpeg: accessed in end_frame() by the filter code, causing a crash.
[00:44] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r77c991d8c5 10ffmpeg/tests/lavfi-regression.sh: 
[00:44] <CIA-56> ffmpeg: tests/lavfi-regressions: support user-specified test name in do_lavfi_pixfmts()
[00:44] <CIA-56> ffmpeg: Make do_lavfi_pixfmts() support an user-specified name for the test.
[00:44] <CIA-56> ffmpeg: This allows to specify two pixfmts tests for the same filter, e.g. to
[00:44] <CIA-56> ffmpeg: test a filter with different parameters. Useful for the pending
[00:44] <CIA-56> ffmpeg: tinterlace tests.
[00:44] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r29e972f67c 10ffmpeg/ (7 files in 4 dirs): 
[00:44] <CIA-56> ffmpeg: lavu/parseutils: add av_small_strptime()
[00:44] <CIA-56> ffmpeg: Make internal small_strptime() function public, and use it in place of
[00:44] <CIA-56> ffmpeg: strptime().
[00:44] <CIA-56> ffmpeg: This allows to avoid a dependency on strptime() on systems which do not
[00:44] <CIA-56> ffmpeg: support it.
[00:44] <CIA-56> ffmpeg: In particular, fix trac ticket #992.
[00:44] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r9a2daef53f 10ffmpeg/libavfilter/vf_tinterlace.c: 
[00:44] <CIA-56> ffmpeg: lavfi/tinterlace: declare source buffers in copy_picture_field() as const
[00:44] <CIA-56> ffmpeg: Increase performance (slightly).
[00:44] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r5780f9bbd5 10ffmpeg/tests/ (3 files in 2 dirs): 
[00:44] <CIA-56> ffmpeg: tests: add tinterlace filter tests
[00:44] <CIA-56> ffmpeg: Only the merge and pad modes are tested.
[00:45] <ubitux> yepee.
[01:15] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * rbdfffa66a0 10ffmpeg/tests/lavfi-regression.sh: 
[01:15] <CIA-56> ffmpeg: tests/lavfi-regression: name temporary files after $testname in do_lavfi_pixfmts()
[01:15] <CIA-56> ffmpeg: Fix a failure when running do_lavfi_pixfmts() for the
[01:15] <CIA-56> ffmpeg: lavfi-tinterlace_merge and lavfi-tinterlace_pad tests concurrently.
[01:15] <CIA-56> ffmpeg: Since they were using the same names for the temporary files, the first
[01:15] <CIA-56> ffmpeg: ending test was removing them, and the second test was failing at
[01:15] <CIA-56> ffmpeg: removing unexisting files.
[01:15] <saste> michaelni: should be fixed
[01:20] <michaelni> saste, fate works again, thanks
[01:23] Action: saste saste deserves to sleep
[03:42] Action: ubitux really loves the bprint API
[04:18] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r52fa5cea03 10ffmpeg/libavcodec/amrnbdec.c: 
[04:18] <CIA-56> ffmpeg: amrnbdec: silence a warning about unsigned cant be negative
[04:18] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[05:01] <CIA-56> ffmpeg: 03rogerdpack 07master * ref122ff507 10ffmpeg/libavformat/os_support.h: 
[05:01] <CIA-56> ffmpeg: eliminate some mingw warnings
[05:01] <CIA-56> ffmpeg: Signed-off-by: rogerdpack <rogerpack2005 at gmail.com>
[05:01] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[09:31] <xxthink> where can I find the spec for canopus HQ format?
[10:05] <j-b> good moroning
[11:31] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * rfb2e7ac3c4 10ffmpeg/ (5 files in 3 dirs): 
[11:31] <CIA-56> ffmpeg: lavfi/mp: drop rotate wrapper
[11:31] <CIA-56> ffmpeg: The native filter transpose is perfectly equivalent.
[12:14] <saste> I'm analyzing media files with desynchronized A/V, that is audio starts before (or after) video
[12:14] <saste> ffplay is not able to cope with that, vlc plays them fine
[12:15] <saste> mplayer has the same problem as ffplay
[12:34] <ubitux> saste: try mplayer2, it should handle that
[12:34] <ubitux> i had one (avi) sample with audio supposed to start something like 15 sec after the video
[12:34] <ubitux> and it was making mplayer crazy
[12:34] Action: saste wonders where's the level where it should be handled
[12:35] <ubitux> uau fixed that in mplayer2 a long while ago
[12:35] <saste> ubitux: why wasn't it "ported" to mplayer?
[12:35] <saste> i'm thinking about writing an audio/video pad filter, which fills the gap with content
[12:36] <ubitux> i'm not interested in porting stuff to mplayer, i'm satisfied with mplayer2, it has much more things i'm interested in, and afaict everything mplayer supports so..
[12:37] <ubitux> it's kind of hard to backport stuff from mplayer2 to mplayer given the difference of codebase if you're not familiar with both codes now
[12:37] <j-b> I should start vlc2
[12:37] <ubitux> @_@
[12:38] <saste> jbvlc
[12:38] <ubitux> j-b: any news for the rtmp-win-vm?
[12:38] <j-b> ubitux: I have still not emptied my mail-stack
[12:38] <j-b> ubitux: you know there was FOMS after VDD :)
[12:38] <ubitux> ah right indeed
[12:40] <thresh> j-b: vlc2 is already released...
[12:41] <j-b> shit
[12:41] <ubitux> fork it, and call it "vlc<3"
[12:42] <j-b> libav<ode<
[12:47] <saste> uhm ffmpeg -acodec copy -copyts is not preserving timestamps...
[12:50] <ubitux> saste: do you have the samples with a/v desync?
[12:50] <ubitux> +available in direct download
[12:50] <saste> ubitux: yes, but I can't share them
[12:51] <ubitux> ok :(
[12:51] <saste> what's the ffmpeg equivalent of asyncts?
[12:52] <ubitux> -async ?
[12:53] <saste>  the parameter is the maximum samples per second by which the audio is changed.
[12:53] <saste> i guess it could be explained better
[12:53] Action: ubitux still doesn't get async
[12:54] <saste> better to check the code
[12:58] Action: saste blames again the splitting of ffmpeg.c in many tiny files, hard to look for
[12:58] <ubitux> i don't like that very much either :p
[13:00] Action: ubitux is wondering about sending a 3rd WIP
[13:00] <ubitux> ...and starts a 4th.
[13:19] <ubitux> mmh we have kind of a ASS parser actually
[13:19] <ubitux> could be somehow easily extended for the AVSubtitles2 API
[13:21] <saste> i suspect av_write_interleaved_frame() is messing up with the ts
[13:49] <saste> cool, is a flvenc problem...
[14:38] <saste> michaelni: in flv_write_packet(): avio_wb24(pb, pkt->pts - pkt->dts);
[14:38] <saste> if the stream starts with a non-zero PTS, DTS is set to PTS, and we get 0
[14:38] <saste> so we lose the information in the original stream
[14:39] <saste> this happens also when remuxing
[14:39] <saste> same happens with MP3, but I'm not sure what happens in that case since PTS is handled in the encoder
[14:39] <saste> so maybe there is no way to deal with it differently
[14:40] <saste> what I'm not sure, is if we should fix the DTS at the in lavf, or fix the PTS output computation in flvenc
[15:26] <ubitux> saste: maybe add then "(-absf ... with ffmpeg)"
[15:26] <saste> yes i was thinking the same
[15:26] <ubitux> maybe a bit verbose but i believe it's more explicit for most of the users
[15:51] <michaelni> saste, about ffplay sync issues, best you talk with cus (marton balint)
[15:54] <saste> michaelni: another issue:
[15:54] <saste> [nut @ 0x23ffff0] No codec tag defined for stream 1
[15:54] <saste> does this mean that I can't copy nellymoser streams to NUT?
[15:55] <michaelni> try adding a fourcc for nellymoser to riff.c
[15:56] <saste> ok
[15:57] <saste> also I discovered that FLV muxer discard timestamp information for audio packets
[15:57] <saste> apparently because of a limitation in the format?
[15:59] <JEEB> http://download.macromedia.com/f4v/video_file_format_spec_v10_1.pdf
[15:59] <JEEB> (has f4v and flv)
[16:02] <Compn> not enca and encv tho :P
[16:02] <Compn> punkass adobe
[16:12] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r5e997688f8 10ffmpeg/libavcodec/h264.c: 
[16:12] <CIA-56> ffmpeg: h264: Remove an assert on current_picture_ptr being null
[16:12] <CIA-56> ffmpeg: It is possible in various error paths as well as gap handling
[16:12] <CIA-56> ffmpeg: that this has already been allocated. It is not clear why that
[16:12] <CIA-56> ffmpeg: would be a problem with the current code, thus disable the
[16:12] <CIA-56> ffmpeg: assert to avoid a common assert failure when asserts are enabled.
[16:12] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:12] <CIA-56> ffmpeg: 03Martin Storsjö 07master * r6d65496990 10ffmpeg/libavcodec/parser.c: 
[16:12] <CIA-56> ffmpeg: parser: Don't use pc as context for av_dlog
[16:12] <CIA-56> ffmpeg: The ParserContext class doesn't have an AVClass, required for
[16:12] <CIA-56> ffmpeg: using it as a logging class.
[16:12] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:13] <CIA-56> ffmpeg: 03Mans Rullgard 07master * rb797f3c4d0 10ffmpeg/configure: 
[16:13] <CIA-56> ffmpeg: configure: add section for libc-specific hacks
[16:13] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:13] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r376b16d7cc 10ffmpeg/: (log message trimmed)
[16:13] <CIA-56> ffmpeg: Merge commit '0db2d94280e260af5f3ad7993c5a6357462f17c9'
[16:13] <CIA-56> ffmpeg: * commit '0db2d94280e260af5f3ad7993c5a6357462f17c9':
[16:13] <CIA-56> ffmpeg:  dsputil: workaround __VA_ARGS__ missing tokenization for MSVC
[16:13] <CIA-56> ffmpeg:  configure: add section for libc-specific hacks
[16:13] <CIA-56> ffmpeg:  build: disable ranlib on mingw
[16:13] <CIA-56> ffmpeg:  parser: Don't use pc as context for av_dlog
[16:13] <CIA-56> ffmpeg: 03Mans Rullgard 07master * ra3222470de 10ffmpeg/configure: 
[16:13] <CIA-56> ffmpeg: build: disable ranlib on mingw
[16:13] <CIA-56> ffmpeg: Using ranlib is not required but prevents using the libraries with
[16:13] <CIA-56> ffmpeg: msvc.
[16:13] <CIA-56> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[16:13] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:13] <CIA-56> ffmpeg: dsputil: workaround __VA_ARGS__ missing tokenization for MSVC
[16:13] <CIA-56> ffmpeg: A second expansion forces the preprocessor to tokenize properly.
[16:13] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:21] <Compn> wait
[16:21] <Compn> what does ranlib do if its 'not required' why was it added ?
[16:21] <Compn> msvc taking over :P
[16:21] <Daemon404> it never did anything afaik in mingw
[16:22] <Daemon404> inb4 breakage elsewhere
[16:22] <Daemon404> but it certainly esplodes anything with msvc
[16:25] <nevcairiel> it got merged now eh. wonder how much needs fixing
[16:27] <Daemon404> hehe
[16:28] <CIA-56> ffmpeg: 03Ronald S. Bultje 07master * rdfa559bcbd 10ffmpeg/configure: 
[16:28] <CIA-56> ffmpeg: configure: Add support for MSVC cl.exe/link.exe
[16:28] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:28] <CIA-56> ffmpeg: 03Martin Storsjö 07master * r4a185ed2d4 10ffmpeg/configure: (log message trimmed)
[16:28] <CIA-56> ffmpeg: configure: Disable some warnings in MSVC
[16:28] <CIA-56> ffmpeg: This disables the following warnings:
[16:28] <CIA-56> ffmpeg: C4100: unreferenced formal parameter
[16:28] <CIA-56> ffmpeg:  (1035 occurrances)
[16:28] <CIA-56> ffmpeg: C4214: nonstandard extension used : bit field types other than int
[16:28] <CIA-56> ffmpeg:  (609 occurances)
[16:28] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r2bb62455c8 10ffmpeg/ (6 files in 3 dirs): (log message trimmed)
[16:28] <CIA-56> ffmpeg: Revert "lavu: add snprintf(), vsnprint() and strtod() replacements for MS runtime."
[16:28] <CIA-56> ffmpeg: This reverts commit ba537202801d71711f1ef9306a6248dd2d426fa0.
[16:28] <CIA-56> ffmpeg: A better implementation has been commited by the same author to qatar
[16:28] <CIA-56> ffmpeg: Conflicts:
[16:28] <CIA-56> ffmpeg:  configure
[16:28] <CIA-56> ffmpeg: Found-by: jamal <jamrial at gmail.com>
[16:28] <CIA-56> ffmpeg: 03Hendrik Leppkes 07master * rd914ea6fd8 10ffmpeg/libavcodec/x86/vp56_arith.h: 
[16:28] <CIA-56> ffmpeg: x86: vp56: cmov version of vp56_rac_get_prob requires inline asm
[16:29] <nevcairiel> well i need shared anyway
[16:29] <JEEB> support for that should come up next
[16:29] <Daemon404> [10:28] < CIA-56> ffmpeg: A better implementation has been commited by the same author to qatar <-- iirc it had some issues compared to this one
[16:29] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:29] <CIA-56> ffmpeg: 03Ronald S. Bultje 07master * r4fa661a857 10ffmpeg/ (4 files in 4 dirs): 
[16:29] <CIA-56> ffmpeg: lavu: add snprintf(), vsnprint() and strtod() replacements for MS runtime.
[16:29] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:29] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r42e63ec8dc 10ffmpeg/libavutil/opt.h: 
[16:29] <CIA-56> ffmpeg: avopt: fix examples to match the same style about default values as the actual code.
[16:29] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[16:29] <Daemon404> some stuff was fixed by wbs and co before pushing it
[16:30] <nevcairiel> shared is problematic because you need a list of symbols to export
[16:30] <Daemon404> its possible to do via scrip
[16:30] <Daemon404> t
[16:30] <nevcairiel> gcc generates that from the .ver file
[16:30] <Daemon404> im thinking of the least vad way
[16:30] <Daemon404> bad*
[16:31] <j-b> michaelni: did thresh fix the repos?
[16:31] <nevcairiel> dlltool can generate a list of *all* symbols. a script that can apply the .ver file against that would be useful
[16:32] <Daemon404> let me gues
[16:32] <Daemon404> the .ver is some gcc specific language
[16:32] <nevcairiel> pretty much. but its not too complicated
[16:33] <michaelni> j-b, git fetch libpostproc \n ERROR:gitosis.serve.main:Repository read access denied
[16:34] <michaelni> thresh, ping ^
[16:35] <nevcairiel> its called a gnu linker version script
[16:39] <Daemon404> nevcairiel, first things first: try building ffmpeg
[16:39] <Daemon404> with msvc
[16:39] <Daemon404> i suspect it will need a few changes here and there.. maybe
[16:41] <nevcairiel> plenty of variable declarations mixed with code in some placea
[16:41] <ubitux> is cl.exe/link.exe freely available?
[16:41] <nevcairiel> yes
[16:42] <nevcairiel> there is a visual studio express which is free
[16:42] <JEEB> only MFC etc. stuff is not available from the free version
[16:42] <ubitux> if anyone gives me some hint on where to get them, as well as the c89toc99 usage, i'm willing to setup a ffmpeg/wine/msvc box
[16:43] <Daemon404> [10:41] < nevcairiel> plenty of variable declarations mixed with code in some placea <-- those should be fixed anyway
[16:43] <Daemon404> bad practise
[16:43] <nevcairiel> yea
[16:43] <nevcairiel> should really turn that back into an error
[16:43] <ubitux> wbs: do you mind sharing your libav-wine-msvc config?
[16:44] <nevcairiel> ill setup a real windows fate when i get around to it
[16:44] <Compn> thats what everyone says
[16:44] <Compn> :P
[16:45] <Daemon404> i have one sitting in my office!
[16:45] <nevcairiel>  i already prepares the windows vm for it
[16:45] <ubitux> oh ok great
[16:45] <wbs> ubitux: the config is pretty much what you can see on fate, then I've got a shellscript wrapper around cl.exe and link.exe, and iirc I've lowercased the msvc include files to get around wine case mangling issues, I can share the wrapper shellscripts
[16:45] <Daemon404> wbs, is your c99wrap native?
[16:45] <Daemon404> or wine?
[16:45] <wbs> Daemon404: it's native
[16:46] <Daemon404> nice
[16:46] <ubitux> wbs: ok, but since nevcairiel is going to do it& :)
[16:46] <ubitux> thanks
[16:46] <Daemon404> nevcairiel needs a few copy of teh converter :P
[16:46] <Daemon404> soon
[16:46] <nevcairiel> i'll try to get to it on the weekend
[16:46] <wbs> Daemon404: and for a native/non-msvc clang, you need to add "-fms-extensions -target i386-pc-win32" to the clang args
[16:47] <wbs> ubitux: ok
[16:47] <Daemon404> wbs, ah
[16:47] <nevcairiel> but its my first fate station. so it may be rough
[16:48] <Daemon404> nevcairiel, will you send patches for teh mixed code / decl
[16:48] <Daemon404> or shall someone else
[16:49] <nevcairiel> i will when i get to it. still at work and goinf out this evening, so tomorrow evening or sat for me
[16:49] <Daemon404> oic
[16:49] <nevcairiel> if someone beats me to it... :)
[16:49] Action: Daemon404 may
[16:50] <wbs> http://albin.abo.fi/~mstorsjo/msvc-wine/ - if someone wants to give it a shot
[16:51] <j-b> what is it ?
[16:51] <Daemon404> wrappers
[16:52] <av500> or trojans
[16:52] <wbs> one will never know
[16:53] <Daemon404> it uses bash
[16:53] <Daemon404> so probably trojans
[16:53] <av500> I guess the .exe detects if one compiles  ffmpeg.c or avconvc.c
[17:00] Action: Daemon404 waits for configure to run
[17:00] <Daemon404> as if msys wasnt slow enough
[17:13] <Daemon404> c:\msys\1.0\devel\c\ffmpeg\libavdevice\dshow_capture.h(182) : error C2081: 'WINBOOL' : name in formal parameter list illegal
[17:13] <Daemon404> well that failed quickly
[17:15] <Daemon404> inb4 mingw-specific
[17:17] <Daemon404> should probably be BOOL on both systems
[17:17] <Daemon404> (mingw and msvc)
[17:17] <Daemon404> iirc mingw #define WINBOOL BOOL
[17:18] <Daemon404> or rather, #define BOOL WINBOOL
[17:24] <Daemon404> hummm
[17:24] <Daemon404> michaelni, ping, re: libmpcodecs
[17:27] <michaelni> Daemon404, libmpcodecs ?
[17:27] <Daemon404> yes
[17:27] <Daemon404> it has unprotected inline asm
[17:27] <Daemon404> (HAVE_INLINE_ASM)
[17:28] <Daemon404> but i see a) n oway to disable libmpcodecs
[17:28] <Daemon404> also i know you dont want to modify the code in there
[17:28] <michaelni> send patch to mplayer, and merge back to ffmpeg 
[17:28] <Daemon404> >_>
[17:28] <Daemon404> that would requrie adding HAVE_INLINE_ASM to mplayer
[17:28] <ubitux> Daemon404: --disable-filter=mp ?
[17:28] <Daemon404> which is somethign i dont want to do at all ;)
[17:29] <Daemon404> ubitux, it should be disabled if inline asm is disabled.
[17:29] <Daemon404> not manually
[17:30] <ubitux> maybe mp_filter_deps="inline_asm" or something?
[17:30] <Daemon404> i dont know configure well enough
[17:31] <ubitux> there is a huge list
[17:31] <Daemon404> ok
[17:31] <ubitux> with dependencies
[17:31] <Daemon404> will do
[17:31] <Daemon404> ill send it as part of my msvc patch set
[17:31] <ubitux> look for "# filters" in the configure
[17:31] <ubitux> and try adding that entry
[17:31] <ubitux> but that's assuming disabling mp filter won't build any libmpcodecs stuff and would fix your issue
[17:32] <Daemon404> ill try explicitly disabling first
[17:32] <Daemon404> time to wait 10 mins for configure!
[17:32] <ubitux> mingw doesn't provide dash?
[17:32] <Daemon404> ?
[17:32] <ubitux> dash ./configure ...
[17:32] <Daemon404> configure on msys is SLOW
[17:33] <Daemon404> forks
[17:33] <Daemon404> etc
[17:33] <Daemon404> its not a matter of what shell is use.d
[17:33] <ubitux> i got a huge perf boost on my laptop when switching from bash to dash, but well ok if it's the fork problem on windows yeah i guess that won't help :)
[17:34] <ubitux> (but maybe bash is forking randomly)
[17:41] <Daemon404> ubitux, yeah that did it
[17:41] <ubitux> great :)
[17:58] <thresh> j-b: michaelni: sorry for delay, I will do it today
[18:03] <Daemon404> michaelni, ping_v2
[18:10] <Daemon404> nvm then
[18:22] <Daemon404> anyone know who Martin Sliwka is?
[18:30] <av500> not me
[18:31] <Daemon404> C:\Msys\1.0\devel\c\ffmpeg\config.h(9) : warning C4005: 'av_restrict' : macro redefinition c:\msys\1.0\devel\c\ffmpeg\libavutil\attributes.h(66) : see previous definition of 'av_restrict'
[18:31] <Daemon404> ^ this is super annoying
[18:31] <Daemon404> on every single file
[18:40] <michaelni> Daemon404, #ifdef #undef ... or remove it from attributed but the later will possibly break something
[18:49] <ubitux> saste: we can't use t parameter in overlay? :(
[18:50] <ubitux> it's too bad, i wanted to try to reproduce this: http://www.youtube.com/watch?v=zdW7PvGZ0uM using the picz he provides and ffmpeg+overlay :))
[19:05] Action: Daemon404 watches as changes to ffmpeg stack up
[19:06] <Daemon404> all valid non-msvc fixes tho
[19:11] <ubitux> you have a hard time getting ffmpeg working with msvc?
[19:15] <Compn> so uh
[19:15] <Compn> what kind of apps are building against ffmpeg in msvc anyways ?
[19:15] <Daemon404> ubitux, there are a few unprotected inline asm places
[19:15] <Daemon404> and some some VLAs
[19:15] <Compn> a bunch of vbscript frontends ? :P
[19:15] <Daemon404> Compn, chrome
[19:15] <Daemon404> steam
[19:15] <Daemon404> games
[19:15] <Daemon404> nothing big ;)
[19:17] <Compn> fair enough
[19:32] <Daemon404> oh SHIT
[19:32] <Daemon404> i forgot postproc was enabled by default with gpl
[19:40] <saste> ubitux: "t parameter in overlay"
[19:41] <saste> it's on my todo since a long time
[19:41] <saste> while at it we should make it reconfigurable
[19:41] <saste> (e.g. through sendcmd)
[19:41] <saste> so people can do very nasty things
[19:43] <ubitux> i'm not sure sendcmd would allow that
[19:43] <ubitux> i mean, not friendly at least
[19:46] <saste> ubitux: suppose you want to get a logo moving from r->l, stay on the screen, and then get back
[19:47] <saste> you can do it with a fancy expression, but it should be easier to reset the expression at given moment
[19:49] <ubitux> ah yeah sure
[19:49] <ubitux> but that's assuming you have an expression using the time and that the filter is supporting it
[19:49] <ubitux> i wasn't even considering the loop at first
[19:50] <ubitux> just trying to get the overlay moving
[19:54] <Daemon404> postproc compile with msvc...
[19:54] <Daemon404> i must be a wizard
[19:55] <ubitux> :)
[20:00] Action: Compn dons his robe and wizard hat
[20:02] <gnafu> Compn: Oh, snap!
[20:04] <Compn> actually need to buy a robe and wizard hat :P
[20:40] <Daemon404> ffmpeg version N-44233-gf186ed5
[20:40] <Daemon404> built on Sep  6 2012 14:22:52 with Microsoft (R) 32-bit C/C++ Optimizing Compiler Version 16.00.40219.01 for 80x86
[20:40] <Daemon404> derp derp
[20:40] <Daemon404> patches start flowing in tomorrow
[20:40] <Daemon404> there are a lot-ish
[20:40] <Daemon404> mostly unprotected inline as,
[20:42] <Daemon404> # Your branch is ahead of 'origin/master' by 17 commits.
[20:46] <ubitux> :)
[20:50] <michaelni> :)
[21:07] <gnafu> "robe and wizard hat", "unprotected"...  What kind of channel have I stumbled into.
[21:07] <gnafu> ;D
[23:28] <saste> ubitux: why ";" in stead of ":"?
[23:28] <saste> usually we have:
[23:28] <saste> "error description: what to do to fix it"
[23:28] <saste> s/usually/sometimes
[23:29] <ubitux> it looked better to me in a sense that's an additionnal, but really feel free to ignore
[23:29] <ubitux> +info
[23:30] <ubitux> don't worry about such consideration if you don't like it :P
[23:32] <saste> also just discovered that "malformated" is not an English word
[23:32] <ubitux> misformated maybe?
[23:32] <saste> indeed the first results of google are related to that ffmpeg message
[23:33] <saste> malformed
[23:33] <ubitux> (seems not either..)
[23:33] <ubitux> heh ok
[23:33] <ubitux> we are the spelling issues reference
[23:33] <ubitux> fear
[23:36] <saste> ubitux: Malformed AAC bitstream detected: use audio bistream filter 'aac_adtstoasc' to fix it ('-bsf:a aac_adtstoasc' option with ffmpeg)
[23:37] <ubitux> looks ok
[23:37] <saste> will push it soon
[23:37] <ubitux> btw, the exact same message is in movenc as well
[23:37] <ubitux> feel free to change it
[23:43] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r1bd6b0c927 10ffmpeg/libavformat/flvenc.c: lavf/flvenc: return meaningful error codes
[23:43] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * rda2ea7c33c 10ffmpeg/libavformat/flvenc.c: 
[23:43] <CIA-56> ffmpeg: lavf/flvenc: apply various log fixes/clarifications
[23:43] <CIA-56> ffmpeg: Should improve user feedback in case of errors.
[23:43] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r144caa7280 10ffmpeg/libavcodec/raw.c: (log message trimmed)
[23:43] <CIA-56> ffmpeg: lavc/raw-test: invert the order of key/value when printing the fourcc->pix_fmt map
[23:43] <CIA-56> ffmpeg: Print
[23:43] <CIA-56> ffmpeg: FOURCC: PIX_FMT
[23:43] <CIA-56> ffmpeg: rather than
[23:43] <CIA-56> ffmpeg: PIX_FMT: FOURCC
[23:43] <CIA-56> ffmpeg: This seems more consistent with the help message:
[00:00] --- Fri Sep  7 2012


More information about the Ffmpeg-devel-irc mailing list