[Ffmpeg-devel-irc] ffmpeg-devel.log.20111115

[burek]

[00:00] <ubitux> it's the style used over most of the codebase
[00:00] <ubitux> and for consistency&
[00:02] <ubitux> not much to change btw
[00:02] <ubitux> http://blackhole.ubitux.fr/0001-vda-use-K-R-style.patch
[00:06] <dilaroga> ubitux: thx for your patch i'll apply it asap
[00:06] <ubitux> i may have missed some, it's just a quick pass
[00:07] <dilaroga> ok
[00:07] <ubitux> i'll send a patchset in a few minutes
[00:08] <dilaroga> great :)
[00:17] <ubitux> i wonder if the struct initializers work with all the c++ compilers
[00:19] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r2b1a4c5b34 10ffmpeg/libavfilter/src_movie.c: 
[00:19] <CIA-41> ffmpeg: amovie: avoid crash in amovie_request_frame() if no samplesref is available
[00:19] <CIA-41> ffmpeg: The crash occurrs when a frame is successfully decoded, but no decoded
[00:19] <CIA-41> ffmpeg: data is immediately available (typically happens with ogg/vorbis).
[00:45] <dilaroga> ubitux: thanks to you for your work i pushed it
[00:45] <ubitux> np
[00:45] <dilaroga> michaelni: ^^ another cosmetic patch today that you can merge :)
[00:52] <CIA-41> ffmpeg: 03Clément BSsch 07master * redc26bfae5 10ffmpeg/libavcodec/ (vda.c vda.h): vda: use K&R style.
[00:53] <michaelni> done :)
[00:55] <dilaroga> thx
[01:29] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rbb4eabacea 10ffmpeg/libavformat/asfdec.c: 
[01:29] <CIA-41> ffmpeg: asfdec: restore timestamp reading code
[01:29] <CIA-41> ffmpeg: This reverts part of:
[01:29] <CIA-41> ffmpeg:  commit e65ab9d94f1c8d8893e32d90467d9525625d306a
[01:29] <CIA-41> ffmpeg:  Author: Mans Rullgard <mans at mansr.com>
[01:29] <CIA-41> ffmpeg:  Date: Wed Jun 1 17:26:27 2011 +0100
[01:29] <CIA-41> ffmpeg:  Remove unused variables
[01:29] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rba7f18876c 10ffmpeg/libavformat/asfdec.c: 
[01:29] <CIA-41> ffmpeg: asfdec: use dts for seeking.
[01:29] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[01:36] <ubitux> michaelni: maybe a av_unused should be added after "ts1" to avoid the warning?
[01:47] <pasteeater> michaelni: do you want the notify mails on instead?
[01:59] <michaelni> pasteeater, i would prefer it slightly, yes
[02:00] <michaelni> ubitux, ill fix it in a moment
[02:05] <teratorn> could anyone tell me why I have so many artifacts in this video? http://teratorn.org/code/mpg_overlay/ (see out.mpg) - I'm reading in.mpg, and overlaying a watermark on each frame before re-encoding it to out.mpg
[02:05] <pasteeater> michaelni: i assume you admin via email instead of the web interface. you want it on for all ML, or just devel?
[02:06] <teratorn> the watermark position is changed each frame, as you can see it bouncing around - but I don't understand why it looks so bad :(
[02:08] <michaelni> pasteeater, devel is enough
[02:08] <pasteeater> fixed
[02:08] <michaelni> thx
[02:10] <michaelni> teratorn, i guess the overlay is written into the decoders internal frames maybe
[02:11] <teratorn> michaelni: I get the frame from the decoder, and I change various pixels in that frame to do the overlay, before encoding and writing that frame out
[02:12] <teratorn> so is that really dumb, or what?
[02:15] <teratorn> I guess I should copy the frame before changing it, as the decode may need access to the current frame to properly decode the next frame?
[02:15] <michaelni> exactly
[02:16] <teratorn> yay, I understand video now
[02:16] <michaelni> theres a reference field in AVFrame that tells you if it might be used
[02:16] <teratorn> michaelni: thanks a lot I'll give that a try :)
[02:16] <michaelni> also there are buffer_hints
[02:16] <michaelni> that is if you want to do the copy only optionally if its really needed
[02:16] <michaelni> but it will be needed for all modern codecs
[02:35] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r6760bd0a19 10ffmpeg/libavformat/asfdec.c: 
[02:35] <CIA-41> ffmpeg: asfdec: avoid unused variable warning.
[02:35] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[02:35] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r66cd3c12ab 10ffmpeg/libavcodec/mpegaudiodata.c: 
[02:35] <CIA-41> ffmpeg: mpegaudio: prettify avpriv_mpa_bitrate_tab
[02:35] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[02:35] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rbe4c9a440f 10ffmpeg/libavcodec/mpeg4videodec.c: 
[02:35] <CIA-41> ffmpeg: mpeg4videodec: dont treat empty frames as errors for QMP4, it seems thats normal.
[02:35] <CIA-41> ffmpeg: Fixes remaining complaints in Ticket560 possibly.
[02:35] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:02] <teratorn> copying an AVFrame seems non-trivial :/
[03:27] <CIA-41> ffmpeg: 03Sebastien Zwickert 07master * rac3dbb4d58 10ffmpeg/ (14 files in 3 dirs): 
[03:27] <CIA-41> ffmpeg: hwaccel: OS X Video Decoder Acceleration (VDA) support.
[03:27] <CIA-41> ffmpeg: Signed-off-by: Luca Barbato <lu_zero at gentoo.org>
[03:27] <CIA-41> ffmpeg: 03Luca Barbato 07master * rcaf27e37b6 10ffmpeg/libavformat/options.c: 
[03:27] <CIA-41> ffmpeg: lavf: simplify format_child_class_next()
[03:27] <CIA-41> ffmpeg: And fix the error introduced when adding private option to avio.
[03:27] <CIA-41> ffmpeg: See 32caa7b13cecca59213c73fa94dd683c2b003bfd
[03:27] <CIA-41> ffmpeg: 03Janne Grunau 07master * rd8a091698a 10ffmpeg/doc/t2h.init: 
[03:27] <CIA-41> ffmpeg: doc: add support for an optional navigation bar in texi2html pages
[03:27] <CIA-41> ffmpeg: The navigation header for the web pages lives in a different repository.
[03:27] <CIA-41> ffmpeg: Read it during documentation regeneration to use the same navigation bar
[03:27] <CIA-41> ffmpeg: on all pages.
[03:27] <CIA-41> ffmpeg: 03Luca Barbato 07master * r2625b621ce 10ffmpeg/libavdevice/pulse.c: 
[03:27] <CIA-41> ffmpeg: pulse: compute frame_duration once and fix it
[03:27] <CIA-41> ffmpeg: The frame duration was calculated without taking in account the
[03:27] <CIA-41> ffmpeg: bytes per sample.
[03:27] <CIA-41> ffmpeg: Thanks to Lorenzo Pistone <blaffablaffa at gmail.com> for pointing
[03:27] <CIA-41> ffmpeg: the issue and providing an initial fix.
[03:27] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * ra72580fc9e 10ffmpeg/: (log message trimmed)
[03:27] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master'
[03:27] <CIA-41> ffmpeg: * qatar/master:
[03:27] <CIA-41> ffmpeg:  binkvideo: simplify and remove invalid shifts
[03:28] <CIA-41> ffmpeg:  pulse: compute frame_duration once and fix it
[03:28] <CIA-41> ffmpeg:  lavf: simplify format_child_class_next()
[03:28] <CIA-41> ffmpeg:  hwaccel: OS X Video Decoder Acceleration (VDA) support.
[03:28] <CIA-41> ffmpeg: 03Mans Rullgard 07master * r3f5d6a665a 10ffmpeg/libavcodec/bink.c: 
[03:28] <CIA-41> ffmpeg: binkvideo: simplify and remove invalid shifts
[03:28] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[03:30] <teratorn> does this function properly deep-copy an AVFrame? http://bpaste.net/show/20257/ It seems to work, but I'm slightly concerned about possibly alignment issues, as the docstrings for avpicture_layout() mentions something about alignment :/
[04:28] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r7ea3963817 10ffmpeg/libavformat/mpegts.c: 
[04:28] <CIA-41> ffmpeg: mpegts: Completely ignore H222 spec and consider reserved adaption
[04:28] <CIA-41> ffmpeg:  fields to be perfectly fine mpeg TS.
[04:28] <CIA-41> ffmpeg: Fixes Ticket580
[04:28] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[04:48] <michaelni> teratorn, see av_image_alloc()
[05:34] <teratorn> michaelni: mmmm, OK. but what to use for the `align' argument?
[05:44] <michaelni> teratorn, try 16
[05:45] Action: michaelni goes to bed
[05:45] <teratorn> michaelni: ty
[13:30] <uau> mail delivery from avserver.banki.hu to my isp seems to be broken so ffmpeg lists don't work
[13:30] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r53cb0de25b 10ffmpeg/libavformat/asfdec.c: 
[13:30] <CIA-41> ffmpeg: asfdec: mark only ts1 as av_unused
[13:30] <CIA-41> ffmpeg: Idea-by: ubitux
[13:30] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:30] <uau> has anyone else seen such problems?
[13:51] <cofol1986> "ffmpeg: relocation error: /usr/local/lib/libswresample.so.0: symbol av_get_default_channel_layout, version LIBAVUTIL_51 not defined in file libavutil.so.51 with link time reference"
[13:51] <cofol1986> hello everyone, I compile ffmpeg in Ubuntu 11.10.after compiling,I start up ffmpeg,and get an error "ffmpeg: relocation error: /usr/local/lib/libswresample.so.0: symbol av_get_default_channel_layout, version LIBAVUTIL_51 not defined in file libavutil.so.51 with link time reference"
[13:52] <cofol1986> can anyone help me ?
[13:55] <cofol1986> ?
[14:00] <ubitux> uau: i don't personnaly
[14:45] <michaelni> uau, since when does the problem exist ?
[14:47] <iive> cofol1986: would you check if you have more than one libavutil.so.51 file (e.g. in /usr/lib/ and /usr/local/lib )
[15:00] <uau> michaelni: since late friday/early saturday
[15:01] <uau> IIRC there was some temporary problem once earlier
[15:06] <royger> hello
[15:07] <royger> the fdct/idct process seems to scale the values of the coefficients, is there any field in DSPContext than contains the value of the scale factor (from what I see it seems to be 8)?
[15:44] <michaelni> royger, its always 8 for the 8bit fdct
[15:47] <royger> michaelni: thanks
[15:51] <michaelni> uau, delivery temporarily suspended: host mta.inet.fi[195.156.147.12] refused to talk to me: 554 5.7.1 rejected because sending IP on http://www.trustedsource.org DNSBL 192.190.173.45
[16:11] <uau> i wonder what exactly that "www.trustedsource.org DNSBL" is - the site mostly seems to redirect to mcafee.com
[16:14] <buzz_> check your mail server on http://www.mxtoolbox.com/blacklists.aspx
[16:16] <buzz_> dnsbl is a dns blackhole list that mailservers can be set up to use. however the one mentioned there doesnt seem to be operating as a dnsbl though as it used to, or maybe im just querying it wrong
[16:16] <buzz_> http://www.trustedsource.org/query/a.ip.address.here
[16:17] <uau> that mxtoolbox.com thing doesn't show anything blocking the ffmpeg server
[16:17] <uau> but it doesn't seem to have any mcafee/trustedsource item
[16:18] <michaelni> ive sent trustedsource/mcaffee some feedback, it says "Our email response charter is 24-48 hours (Monday-Friday)."
[17:30] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r4c52adad69 10ffmpeg/libavfilter/vsrc_mandelbrot.c: 
[17:30] <CIA-41> ffmpeg: mandelbrot: Fix "warning: passing argument 2 of draw_mandelbrot from incompatible pointer type"
[17:30] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:30] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r5582e557e6 10ffmpeg/libavfilter/vsrc_mandelbrot.c: 
[17:30] <CIA-41> ffmpeg: mandelbrot: add "fractal" to the description
[17:30] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[17:30] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r5a68b1ac3b 10ffmpeg/libavfilter/vsrc_mandelbrot.c: 
[17:30] <CIA-41> ffmpeg: mandelbrot: add inner drawing method to show converence time
[17:30] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:17] <CIA-41> ffmpeg: 03Panagiotis H.M. Issaris 07master * rbf0ca404c6 10ffmpeg/MAINTAINERS: 
[18:17] <CIA-41> ffmpeg: Update my GnuPG fingerprint.
[18:17] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[18:36] <CIA-41> ffmpeg: 03Panagiotis H.M. Issaris 07master * r6f20921dee 10ffmpeg/libavformat/applehttp.c: (log message trimmed)
[18:36] <CIA-41> ffmpeg: Fix seeking when using Apple HTTP Live Streaming
[18:36] <CIA-41> ffmpeg: The Apple HTTP Live Streaming demuxer's implementation of seeking searches for
[18:36] <CIA-41> ffmpeg: the MPEG TS segment which contains the requested timestamp. In its current
[18:36] <CIA-41> ffmpeg: implementation it assumes that the first segment will start from 0.
[18:36] <CIA-41> ffmpeg: But, MPEG TS streams do not necessarily start with timestamp (near) 0, causing
[18:36] <CIA-41> ffmpeg: seeking to fail for those streams.
[20:07] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r8d51cb4fb8 10ffmpeg/libavfilter/vsrc_mandelbrot.c: 
[20:07] <CIA-41> ffmpeg: mandelbrot: add mincol inner coloring method.
[20:07] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:07] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r7575980ba4 10ffmpeg/libavfilter/vsrc_mandelbrot.c: 
[20:07] <CIA-41> ffmpeg: mandelbrot: make mincol the default.
[20:07] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:44] <michaelni> uau, i got a reply from McAfee
[21:44] <michaelni> Thank you for your request.  The reputation of this IP will be adjusted.  Please allow some time for the update.
[21:47] <Compn> thats cool that mcafee actually responds to things like that
[21:47] Action: Compn does not have good luck with talking to large companies
[22:16] <dilaroga> http://pastie.org/2869052
[22:16] <dilaroga> ^^ is that a better workaround ?
[22:24] <michaelni> dilaroga, iam fine with whatever you prefer.
[22:42] <dilaroga> use the pragma gcc diagnostic appears to me safer instead of undefining something
[22:48] <michaelni> sure, #undef __* is somewhat risky
[22:49] <michaelni> but i dont know how well both the variants work with non gcc compilers,
[22:52] <dilaroga> i don't get warning or error message with llvm-gcc or clang using pragma gcc diagnostic
[22:53] <dilaroga> I guess it's ignored
[22:54] <dilaroga> #undef affects non gcc compilers too
[23:01] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r04bb26e3ee 10ffmpeg/libavfilter/vsrc_mandelbrot.c: 
[23:01] <CIA-41> ffmpeg: mandelbrot: add dither to convergence & mincol colorings.
[23:01] <CIA-41> ffmpeg: This avoids banding artifacts.
[23:01] <CIA-41> ffmpeg: Note, low end TFTs still show some banding.
[23:01] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:01] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rd5e1244ca7 10ffmpeg/libavutil/opt.c: 
[23:01] <CIA-41> ffmpeg: opt: Fix *av_opt_find2(NULL)
[23:01] <CIA-41> ffmpeg: Fixes Ticket650
[23:01] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:05] <dilaroga> I guess wrong :)
[23:05] <dilaroga> clang supports gcc's pragma for compatibility with existing source code
[23:23] <Compn> http://insidesearch.blogspot.com/2011/11/search-using-your-terms-verbatim.html
[23:23] <Compn> google disabled the +plus modifyer a few weeks ago
[23:23] <Compn> finally they have reenabled it
[23:23] <Compn> why must they toy with us peons?!
[23:25] <Compn> actually, i'm not sure if the verbatim search will do that :\
[23:25] <Compn> change bad 
[23:26] <Compn> "we found that users typed the “+” operator in less than half a percent of all searches" but then "we’ve received a lot of requests for a more deliberate way to tell Google to search using your exact terms."
[23:26] <Compn> so either no one used it, or it was the most important feature ever :P
[23:26] <Compn> whatever google
[23:26] <Compn> pengvado : hey, go tell google they are being stupid
[23:26] <Compn> :)
[23:29] <Compn> google needs an irc channel where i can berate them further.
[23:32] <kierank> why pengvado?
[23:32] <gnafu> Compn: Use #vp8 like I do ;D.
[23:33] <gnafu> kierank: Good point.  He probably meant peloverde.
[23:33] <Compn> oh no
[23:33] <Compn> i did mean peloverde
[23:33] <Compn> pengvado : sorry
[23:34] Action: Compn can barely remember the first letter of nicknames
[23:34] <Compn> it ... started with a p!
[23:34] <Compn> gnafu : isnt that just the on2 jerks tho?
[23:34] <Compn> just because google owns them now, doesnt really mean much
[23:35] <Compn> still on2 jerks :P
[23:35] <Compn> on2 still too lazy to opensouce all of their other codecs
[23:35] Action: Compn ignorant on copyright, licensing, contractual obligations when changing the license on commercial software
[23:36] <gnafu> Meh, I still will shout to the all-hearing Goog in there when I feel like grumbling :-D.
[23:36] <gnafu> BBB is in there, so sometimes I address him with my Google concerns.
[23:36] <Compn> cant hurt :)
[23:37] <Compn> gnafu : if you can pass my rants on to google , that would be cool :)
[00:00] --- Wed Nov 16 2011