[Ffmpeg-devel-irc] ffmpeg-devel.log.20120920

burek burek021 at gmail.com
Fri Sep 21 02:05:02 CEST 2012


[00:51] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r36c2694d83 10ffmpeg/libavcodec/dcadec.c: 
[00:51] <CIA-56> ffmpeg: dcadec: fix "set but not used" variable
[00:51] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:51] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r168bc5f2c5 10ffmpeg/libavformat/mov.c: 
[00:51] <CIA-56> ffmpeg: mov: fix 32/64 format type
[00:51] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:06] <burek> there should be one additional item in the ffmpeg faq, that just says: -sameq does not equal the "same quality"
[01:45] <llogan> burek: you can make a patch. web repo info is on the download page.
[01:46] <burek> +<font size="48">-sameq does NOT mean "same quality"</font>
[01:46] <burek> :)
[01:46] <llogan> LGTM
[01:53] <Daemon404> wut... suddenly libfdk-aacenc encodes everything at 5.2 kbit/s
[01:57] <Daemon404> hmm nvm
[02:41] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r161c6d1682 10ffmpeg/configure: 
[02:41] <CIA-56> ffmpeg: configure: Detect old git domain and inform the user about it.
[02:41] <CIA-56> ffmpeg: Based on code by: Clément BSsch <ubitux at gmail.com>
[02:41] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[02:41] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * rd1974e05d1 10ffmpeg/libavcodec/ (j2k.c j2k_dwt.c): 
[02:41] <CIA-56> ffmpeg: j2k: fix "static is not at beginning of declaration" warning
[02:41] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[02:41] <CIA-56> ffmpeg: 03Matthieu Bouron 07master * r906a263813 10ffmpeg/libavformat/mxfenc.c: 
[02:41] <CIA-56> ffmpeg: mxfenc: factorize samples per frame code
[02:41] <CIA-56> ffmpeg: Reviewed-by: Tomas Härdin <tomas.hardin at codemill.se>
[02:41] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[02:41] <CIA-56> ffmpeg: 03Matthieu Bouron 07master * rd57ca5e5a8 10ffmpeg/libavformat/mxfenc.c: 
[02:41] <CIA-56> ffmpeg: mxfenc: support 23.976 and 24 frame rates
[02:41] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:44] <CIA-56> ffmpeg: 03Janne Grunau 07master * r01fc5d6609 10ffmpeg/libavcodec/mpegvideo.c: mpegvideo: check ff_find_unused_picture() return value for errors
[04:44] <CIA-56> ffmpeg: 03Mans Rullgard 07master * r3dbc777c7f 10ffmpeg/ (configure library.mak): 
[04:44] <CIA-56> ffmpeg: build: support some non-standard ar variants
[04:44] <CIA-56> ffmpeg: This adds support for the TI and Microsoft (lib.exe) variants of
[04:44] <CIA-56> ffmpeg: the ar utility.
[04:44] <CIA-56> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[04:44] <CIA-56> ffmpeg: 03Alex Converse 07master * rdf6c3f9fb3 10ffmpeg/libavcodec/tiff.c: 
[04:44] <CIA-56> ffmpeg: tiffdec: Add support for GRAY16LE.
[04:44] <CIA-56> ffmpeg: Tested with the GraphicsMagick TIFF archive and Libav generated files.
[04:44] <CIA-56> ffmpeg: 03Janne Grunau 07master * r8701f4f8e8 10ffmpeg/libavcodec/ (h263dec.c mpeg4videodec.c mpegvideo.c mpegvideo.h): 
[04:44] <CIA-56> ffmpeg: mpeg4: support frame parameter changes with frame-mt
[04:44] <CIA-56> ffmpeg: Adds a flag context_reinit to MpegEncContext to relieable keep track
[04:44] <CIA-56> ffmpeg: of frame parameter changes which require a context reinitialization.
[04:44] <CIA-56> ffmpeg: This is required for broken inputs which change the frame size but
[04:44] <CIA-56> ffmpeg: error out before the context can be reinitialized.
[04:44] <CIA-56> ffmpeg: 03Mans Rullgard 07master * rac3a9b51f7 10ffmpeg/configure: 
[04:44] <CIA-56> ffmpeg: configure: msvc: default to 'lib' as 'ar' tool
[04:44] <CIA-56> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[04:45] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * rd2981b8ef1 10ffmpeg/libavcodec/h263dec.c: 
[04:45] <CIA-56> ffmpeg: h263dec: disable coded dimensions = 0 handling
[04:45] <CIA-56> ffmpeg: This code causes infinite loops
[04:45] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[05:55] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r16e52c86ba 10ffmpeg/libavcodec/error_resilience.c: 
[05:55] <CIA-56> ffmpeg: error_resilience: guess_dc: check malloc failure
[05:55] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:46] <Tjoppen> michaelni: nb_index_tables  <= 0 sounds fine, but I'll take a quick look once I get to the office
[08:40] <Tjoppen> michaelni: ok, I checked. go ahead and change it to nb_index_tables :)
[08:56] <t355u5> Hmm, quite interesting that 1.11.2 does not support libopus. I believe that someone forgot to update configure.... Unknown option "--enable-libopus".
[09:46] <nevcairiel> isnt libopus rather new, 0.11 is already somewhat old now
[09:47] <nevcairiel> if support would be in the code, it would be in configure
[09:47] <nevcairiel> those are usually done in the same commit
[09:54] <ubitux> funny watermarking: http://www.ownedcore.com/forums/world-of-warcraft/world-of-warcraft-general/375573-looking-inside-your-screenshots.html
[12:48] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r98f753ec51 10ffmpeg/libavfilter/ (version.h vsrc_testsrc.c): 
[12:48] <CIA-56> ffmpeg: lavfi/testsrc: increase precision of the duration parameter
[12:48] <CIA-56> ffmpeg: Compute duration in microseconds, rather than in timebase units. Decrease
[12:48] <CIA-56> ffmpeg: approximation errors.
[12:48] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r5bd6a4800d 10ffmpeg/tests/ (9 files in 3 dirs): tests: add fate-ffprobe test
[12:48] <CIA-56> ffmpeg: 03Stefano Sabatini 07master * r53d71335cb 10ffmpeg/libavdevice/ (lavfi.c version.h): 
[12:48] <CIA-56> ffmpeg: lavd/lavfi: push frames until the last sink is EOF
[12:48] <CIA-56> ffmpeg: Previously the device was returning EOF when the first sink was ending,
[12:48] <CIA-56> ffmpeg: with the current change the device will continue to return frames until
[12:48] <CIA-56> ffmpeg: all the sinks are EOF, which seems the most expected behavior.
[12:58] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * rdb0087dc1d 10ffmpeg/libavformat/mxfdec.c: 
[12:58] <CIA-56> ffmpeg: mxfdec: fix index_tables check in mxf_read_seek()
[12:58] <CIA-56> ffmpeg: Approved-by: Tjoppen
[12:58] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:58] <CIA-56> ffmpeg: 03jamal 07master * r33f92a3ea2 10ffmpeg/configure: 
[12:58] <CIA-56> ffmpeg: configure: fix git domain check with out of tree build
[12:58] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[12:58] <Tjoppen> woop
[13:37] <CIA-56> ffmpeg: 03Luca Barbato 07master * r8d4dd55c37 10ffmpeg/libavformat/matroskadec.c: 
[13:37] <CIA-56> ffmpeg: matroskadec: refactor matroska_decode_buffer
[13:37] <CIA-56> ffmpeg: Make MATROSKA_TRACK_ENCODING_COMP_HEADERSTRIP behave like
[13:37] <CIA-56> ffmpeg: the other encodings and spare few lines of boilerplate code.
[13:37] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r1df2e3c707 10ffmpeg/: 
[13:37] <CIA-56> ffmpeg: Merge commit '8d4dd55c37b14ecea596f3477b35bf6c20688d8b'
[13:37] <CIA-56> ffmpeg: * commit '8d4dd55c37b14ecea596f3477b35bf6c20688d8b':
[13:37] <CIA-56> ffmpeg:  matroskadec: refactor matroska_decode_buffer
[13:37] <CIA-56> ffmpeg: Conflicts:
[13:37] <CIA-56> ffmpeg:  libavformat/matroskadec.c
[13:37] <CIA-56> ffmpeg: Merged-by: Michael Niedermayer <michaelni at gmx.at>
[13:44] <CIA-56> ffmpeg: 03Dale Curtis 07master * rdf1d84121b 10ffmpeg/libavformat/matroskadec.c: matroskadec: fix incorrect unsigned->signed conversion
[13:44] <CIA-56> ffmpeg: 03Luca Barbato 07master * rc9a39cec70 10ffmpeg/libavformat/matroskadec.c: matroskadec: return meaningful errors in matroska_decode_buffer
[13:44] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * ra16c512374 10ffmpeg/: (log message trimmed)
[13:44] <CIA-56> ffmpeg: Merge commit 'c9a39cec70603f662f4c326b21b11c4f0112079a'
[13:44] <CIA-56> ffmpeg: * commit 'c9a39cec70603f662f4c326b21b11c4f0112079a':
[13:44] <CIA-56> ffmpeg:  matroskadec: return meaningful errors in matroska_decode_buffer
[13:44] <CIA-56> ffmpeg:  matroskadec: fix incorrect unsigned->signed conversion
[13:44] <CIA-56> ffmpeg: Conflicts:
[13:44] <CIA-56> ffmpeg:  libavformat/matroskadec.c
[14:20] <mateo`> hi, the dnxhd encoders seems buggy with 8bits profile, artefacts appear in zone with high luma differences, see sample: http://samples.lolito.fr/DNXHD/test_dnxhd_120M_yuv422p.mxf
[14:22] <mateo`> i generated this sample using this picture http://samples.lolito.fr/DNXHD/test.png: ffmpeg -i test.png -vcodec dnxhd -b 120M -pix_fmt yuv422p test_dnxhd_120M_yuv422p.mxf
[14:24] <mateo`> could it be an overflow issue ?
[15:48] <ubitux> seems like a regression in a1bcc76e6036e78f25cbb7323c145056cfca9d93
[16:00] <ubitux> it looks like it's not in the merge commit but one of the fork commit in it
[16:10] <mateo`> ubitux: can't wait to figure out what is happening here :)
[16:12] <ubitux> first bad commit: d211547dddf9deeb34b7dda9cd80b86f3f5cd07a
[16:29] <mateo`> seems like the encoder is broken before a long time before this commit with --disable-asm ...
[16:29] <ubitux> let's bisect moar!
[16:34] <ubitux> yep it's broken way before with --disable-asm
[16:34] <mateo`> even in n0.11 :'(
[16:34] <ubitux> the commit from diego might have actually broken the use of the current code (asm one i guess) and always use the broken c version
[16:36] <nevcairiel> c is broken but asm worked?
[16:36] <nevcairiel> how funny :p
[16:36] <ubitux> yeah
[16:36] <ubitux> so, after d211547dddf9deeb34b7dda9cd80b86f3f5cd07a --disable-asm and normal are broken
[16:36] <ubitux> before d211547dddf9deeb34b7dda9cd80b86f3f5cd07a the normal one is fine but --disable-asm is still broken
[16:37] <ubitux> and it seems to be broken from a looong time ago
[16:49] <ubitux> we are unable to build ffmpeg 0.5 :(
[16:49] <ubitux> it's broken in 0.6
[16:50] <Daemon404> oh boy
[16:50] <nevcairiel> maybe it never worked? :D
[16:50] <Daemon404> every single ffprobe test fails on msvc
[16:50] <nevcairiel> lol
[16:50] <nevcairiel> wth changed now?
[16:51] <Daemon404> these tests were just added
[16:51] <Daemon404> http://fate.ffmpeg.org/report.cgi?time=20120920123859&slot=x86_32-msvc10-windows-native
[16:51] <nevcairiel> shouldnt it have catched new refs
[16:51] <Daemon404> i see 1) unicode fail that is windows
[16:51] <Daemon404> 2) i dont know how some other stuff differs
[16:51] <nevcairiel> eols maybe?
[16:51] <Daemon404> looking at json
[16:51] <Daemon404> its telling
[16:52] <Daemon404> also maybe EOLs
[16:52] <ubitux> the pos is wrong
[16:52] <ubitux> looks like a sscanf with a %n not pointing to the same position or something :))
[16:52] <ubitux> < nevcairiel> maybe it never worked? :D // possible but... it works with the asm!
[16:53] <Daemon404> ubitux, unicode fail too
[16:53] <Daemon404> and bitrate
[16:53] <ubitux> bitrate might be because of floating point rounding or something
[16:53] <ubitux> haha unicode nice
[16:54] <Daemon404> that is territory that makes me go derp on window
[16:54] <Daemon404> s
[16:54] <ubitux> with the new section stuff from saste, it might be possible at some point to filter out some values
[16:54] <Daemon404> well pos should be fixed
[16:55] <ubitux> yup
[16:55] <Daemon404> and i should fix bprint on mingw (update my patch)
[16:55] <ubitux> yup
[16:55] <ubitux> and utvideo
[16:55] <ubitux> seems the ffprobe test showed up some issues with nut too :)
[16:57] <Daemon404> ubitux, utvideo?
[16:57] <Daemon404> that valgrind / mmx2 thing?
[16:57] <ubitux> yes
[16:58] <Daemon404> that one is over my head
[16:58] <Daemon404> i cant figure it out
[16:58] <Daemon404> i tried.
[17:13] <burek> one completely off-topic question
[17:14] <burek> why is this notation used
[17:14] <burek> static int f(a, b, c, d)
[17:14] <burek> 	int a; int b; char c; int d; {
[17:14] <burek> instead of
[17:14] <burek> static int f(int a; int b; char c; int d) {
[17:16] <Tjoppen> because K&R C
[17:19] <burek> thanks, that was enough to find a wiki artcile that explains it kinda: http://en.wikipedia.org/wiki/Indent_style
[17:19] <burek> :beer: :)
[17:25] <ubitux> burek: it's the old C
[17:25] <ubitux> (your first definition)
[17:26] <burek> yes, I figured :) I'm reading wiki article :) it's interesting :)
[17:26] <burek> although I never really thought about importance of indentation, since today's IDEs all have source code formating shortcuts
[17:26] <burek> so, it kinda doesn't matter who uses what, you can always reformat it to your own style
[17:37] <ubitux> well it's not really a style issue
[17:37] <ubitux> it's a language evolution in this case
[18:28] <burek> ubitux, i agree, i used wrong terminology
[18:46] <CIA-56> ffmpeg: 03Clément BSsch 07master * r2f39d7ff3f 10ffmpeg/ (configure libavfilter/version.h libavfilter/vf_select.c): lavfi/select: make lavc dependency optional.
[18:58] <Ctut_> The Bitgravity stream is down. Can you guys restart?
[18:59] <Ctut_> sorry, wrong chat window.
[19:16] <CIA-56> ffmpeg: 03Luca Barbato 07master * r870e75524a 10ffmpeg/libavformat/matroskadec.c: 
[19:16] <CIA-56> ffmpeg: matroskadec: validate lace_size when parsed
[19:16] <CIA-56> ffmpeg: Stricter validation, explicitly exit on misparsing and some error
[19:16] <CIA-56> ffmpeg: forwarding from the ebml parsing functions used.
[19:16] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r0da50055eb 10ffmpeg/libavformat/matroskadec.c: 
[19:16] <CIA-56> ffmpeg: matroskadec: fix typo
[19:16] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[19:16] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r6902c3acb0 10ffmpeg/: 
[19:16] <CIA-56> ffmpeg: Merge commit '870e75524aa0d00ebcd1d15589c8d29b84af1565'
[19:16] <CIA-56> ffmpeg: * commit '870e75524aa0d00ebcd1d15589c8d29b84af1565':
[19:16] <CIA-56> ffmpeg:  matroskadec: validate lace_size when parsed
[19:16] <CIA-56> ffmpeg: Merged-by: Michael Niedermayer <michaelni at gmx.at>
[19:19] <ubitux> haha the typo :)
[19:54] <Daemon404> hey ubitux ive been meaning to ask: why does ffprobe print nb_frames as a string
[19:54] <Daemon404> i always have to cast it to an int when i load the json
[19:55] <ubitux> because some formats might not be able to present a uint64
[19:55] <ubitux> typically iirc in json you are limited to 63 bits
[19:59] <CIA-56> ffmpeg: 03Moritz Bunkus 07master * r8071dca3d5 10ffmpeg/libavformat/ (matroska.c matroskadec.c matroskaenc.c): (log message trimmed)
[19:59] <CIA-56> ffmpeg: matroska: implement support for ALAC
[19:59] <CIA-56> ffmpeg: Support Matroska native formatting.
[19:59] <CIA-56> ffmpeg: On demuxing reconstruct the 36-bytes QuickTime atom that the ALAC
[19:59] <CIA-56> ffmpeg: decoder expects by prepending the "atom size", "tag" and
[19:59] <CIA-56> ffmpeg: "tag version" fields missing from the Matroska's CodecPrivate
[19:59] <CIA-56> ffmpeg: element.
[19:59] <CIA-56> ffmpeg: 03Luca Barbato 07master * r117d8c6d1f 10ffmpeg/libavformat/ (matroska.c matroskadec.c matroskaenc.c): 
[19:59] <CIA-56> ffmpeg: matroska: implement support for ProRes
[19:59] <CIA-56> ffmpeg: Support Matroska native formatting.
[19:59] <CIA-56> ffmpeg: On demuxing prepend a Frame container atom (32bit big endian encoded
[19:59] <CIA-56> ffmpeg: frame size and 'icpf' string).
[19:59] <CIA-56> ffmpeg: On muxing remove it.
[19:59] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * rb8044972c4 10ffmpeg/: (log message trimmed)
[19:59] <CIA-56> ffmpeg: Merge commit '117d8c6d1f1c187ffc6098d9618457e00534e013'
[19:59] <CIA-56> ffmpeg: * commit '117d8c6d1f1c187ffc6098d9618457e00534e013':
[19:59] <CIA-56> ffmpeg:  matroska: implement support for ProRes
[19:59] <CIA-56> ffmpeg:  matroska: implement support for ALAC
[20:05] <burek> is it too much of a job to maintain the latest git version as a deb packet
[20:05] <burek> I'm willing to setup a cron job that will generate .deb packet every day if needed
[20:05] <burek> just to end this debian/ubuntu stupidity
[20:05] <ubitux> the problem are the libraries i guess
[20:05] <ubitux> debian usually splits everything in thousands of packets and dependencies
[20:06] <ubitux> the packaging kind of requires such "design"
[20:06] <burek> can I do, like I did for static builds
[20:06] <ubitux> you can also make a static build deb of course
[20:06] <burek> auto-compile all the dependencies and ffmpeg
[20:06] <burek> and just -> deb
[20:06] <ubitux> but that's kind of ugly to do thtat
[20:06] <JEEB> it's ugly but not much you can do about that
[20:07] <ubitux> it will cause some kind of sym conflicts etc
[20:07] <JEEB> ffmpeg ppa that installs into /usr/local or something
[20:07] <ubitux> also it will likely be a pain to re-generate a new package each time
[20:08] <michaelni> also such a auto build cronjob, if done should run fate before updating the package
[20:08] <michaelni> ideally on a debian box
[20:08] <michaelni> to make sure the package is always working ...
[20:08] Action: ubitux still believes the best thing is too annoy the debian packager(s)
[20:08] <ubitux> to*
[20:08] <ubitux> that's his damn work after all :))
[20:09] <burek> well they dropped entire ffmpeg from the repo.. what's there to bug about..?
[20:09] <burek> if we offer people a deb package at the ffmpeg's download page, then people will have a chance to continue using ffmpeg "normally" like they used to before this avstupidity
[20:10] <mateo`> another crappy solution is to package a static build of ffmpeg :)
[20:10] <burek> but the static build doesn't have ffserver I think
[20:10] <ubitux> why not?
[20:11] <burek> network libs issues
[20:11] <burek> something about dns libs being compiled as shared intentionally
[20:11] <burek> I'll find you a link if you are interested
[20:11] <Daemon404> ubitux, you expect to have enough frames to overflow an int64_t?
[20:12] <JEEB> well, if you make a deb you can just have the needed library as dependency
[20:12] <JEEB> and that'd be it
[20:12] <burek> exactly
[20:12] <mateo`> burek: the problem appears at build time ?
[20:12] <burek> mateo` yes
[20:12] <ubitux> Daemon404: what about a 64-bit bitfield, or a high pts?
[20:12] <Daemon404> ...
[20:12] <Daemon404> nb_frames i neither of those
[20:12] <Daemon404> is*
[20:13] <burek> I couldn't resolve it anyway, so I had to do --disable-ffserver and later when I consulted relaxed static build, I noticed it also doesn't have ffserver
[20:13] <burek> so I figured it has to be something bigger
[20:13] <ubitux> Daemon404: ATM we have print_int or print_fmt/str
[20:13] <ubitux> it can likely overflow 32 bits
[20:14] <ubitux> so we choose print_fmt/str
[20:14] <mateo`> burek: i haven't faced this problems (i use a static build of ffmpeg/ffplay/ffprobe/ffserver at work)
[20:14] <Daemon404> print_int should use (u)int64_t
[20:14] <Daemon404> not native int
[20:14] <Daemon404> imho
[20:14] <ubitux> Daemon404: 'coz we don't have a print_int64 for the reasons we talked about
[20:14] <burek> mateo`, do you have any link to docs on how did you do it
[20:14] <Daemon404> what?
[20:14] <ubitux> Daemon404: yeah why not
[20:14] <Daemon404> ubitux, i cannot think of a reason not to use (u)int64_t
[20:15] <ubitux> feel free to send a patch
[20:15] <mateo`> burek: the packaging code is at work so for now i can't, you'll have to wait tomorrow :)
[20:15] <ubitux> nb_frames is uint64 btw iirc
[20:16] <Daemon404> i know
[20:16] <ubitux> and you'll limit it to 63 bits, and signed
[20:16] <Daemon404> ...
[20:16] <Daemon404> do you realize how many frames 63 bits is>
[20:16] <ubitux> sure
[20:16] <Daemon404> and how logn of a duration taht would be?
[20:16] <ubitux> yup, that's fine in that case
[20:16] <burek> mateo`, thanks, it's not a problem to wait, whenever you can, please send me that, so I can finally make it, because I couldn't make it myself in no way
[20:17] <ubitux> Daemon404: that might be the only case where you would need that print_int supporting 64 bits
[20:17] <ubitux> it might not be a bad idea as i said
[20:18] <Daemon404> ubitux, if nb_frames needs 63 bits, duration would have overflowed logn ago :P
[20:18] <Daemon404> and internally libavformat probably would explode
[20:18] <ubitux> i believe i said i was ok with a smaller signed 63 bit for the frame count
[20:19] <ubitux> my point was that it's not applicable to all (u)int64, that's all
[20:20] <Daemon404> k
[20:43] <CIA-56> ffmpeg: 03Luca Barbato 07master * r2d0e7713f9 10ffmpeg/libavformat/matroskadec.c: matroskadec: split laces parsing
[20:43] <CIA-56> ffmpeg: 03Luca Barbato 07master * rc831ebf616 10ffmpeg/libavformat/matroskadec.c: matroskadec: split frame parsing
[20:43] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r8c51ea5489 10ffmpeg/: (log message trimmed)
[20:43] <CIA-56> ffmpeg: Merge commit 'c831ebf61629d219ebcaa9f02d262e67aad09d83'
[20:43] <CIA-56> ffmpeg: * commit 'c831ebf61629d219ebcaa9f02d262e67aad09d83':
[20:43] <CIA-56> ffmpeg:  matroskadec: split frame parsing
[20:43] <CIA-56> ffmpeg:  matroskadec: split laces parsing
[20:43] <CIA-56> ffmpeg: Conflicts:
[20:43] <CIA-56> ffmpeg:  libavformat/matroskadec.c
[21:10] <llogan> default!
[21:37] <Daemon404> /usr/bin/ld: /usr/local/lib/libswscale.a(rgb2rgb.o): relocation R_X86_64_PC32 against symbol `ff_bgr2YCoeff' can not be used when making a shared object; recompile with -fPIC
[21:37] <Daemon404> ^ this is -with- --enable-pic
[21:50] <michaelni> Daemon404, wild guess: you are missing -Bsymbolic 
[21:51] <Daemon404> yup
[21:51] <Daemon404> thanks michaelni 
[21:59] <llogan> this time all submitted design(s) for this theme will be under one ticket.
[22:07] <CIA-56> ffmpeg: 03Yusuke Nakamura 07master * rd17d0ec8c2 10ffmpeg/libavformat/ (isom.h mov.c): 
[22:07] <CIA-56> ffmpeg: mov: support random access point grouping
[22:07] <CIA-56> ffmpeg: Frames described by this grouping are the starter of a closed or
[22:07] <CIA-56> ffmpeg: an open GOP.
[22:07] <CIA-56> ffmpeg: This is useful for open GOP of H.264 stream which is not described
[22:07] <CIA-56> ffmpeg: by sync sample atom.
[22:07] <CIA-56> ffmpeg: Signed-off-by: Luca Barbato <lu_zero at gentoo.org>
[22:07] <CIA-56> ffmpeg: 03Luca Barbato 07master * r7d8431004a 10ffmpeg/libavformat/matroskadec.c: matroskadec: properly support BlockDuration
[22:07] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r6d87781758 10ffmpeg/libavformat/mp3dec.c: 
[22:07] <CIA-56> ffmpeg: mp3dec: fix seeking without xing TOC
[22:07] <CIA-56> ffmpeg: fixes regression from the xing toc support
[22:07] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[22:07] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r5864ce13d1 10ffmpeg/: (log message trimmed)
[22:07] <CIA-56> ffmpeg: Merge remote-tracking branch 'qatar/master'
[22:07] <CIA-56> ffmpeg: * qatar/master:
[22:07] <CIA-56> ffmpeg:  mp3dec: read Xing frame TOC index
[22:08] <CIA-56> ffmpeg: 03Anton Khirnov 07master * r2d1a1a7f62 10ffmpeg/libavformat/mp3dec.c: mp3dec: use named constants for Xing header flags
[22:26] <burek> I made a circular buffer which delays udp packets for that samplicator program :)
[22:26] <burek> and what's the worst, it works :D
[22:27] <burek> I'll test it a little bit and post a wiki article, if someone needs it again
[22:40] <llogan> how overly complicated can i make it?
[22:40] <llogan> (artists don't follow instructions)
[22:41] <burek> how about we create a php page, with a html form, which will force people to follow instructions and automate it to submit all that to trac :)
[22:43] <Daemon404> llogan, s/artists/people/
[23:09] <Tjoppen> I just realized lavf ought to compute SAR for BMP based on print resolution
[23:09] <Tjoppen> like if I have a file with 36x72 dpi SAR ought to be 2:1
[00:00] --- Fri Sep 21 2012


More information about the Ffmpeg-devel-irc mailing list