[Ffmpeg-devel-irc] ffmpeg-devel.log.20150317

[burek]

[00:15:44 CET] <cone-057> ffmpeg 03Clément BSsch 07master:b17620b81bc1: avcodec/libx264: use AVERROR_EXTERNAL instead of AVERROR_UNKNOWN
[00:29:03 CET] <Compn> Daemon404 : bad debian packaging still exists? :P
[00:29:11 CET] <Compn> >mplayer still not in debian
[00:29:17 CET] <Compn> oh yeah... deb-multimedia still required.
[00:46:26 CET] <Daemon404> ill propose addition of mpv-player
[00:46:30 CET] <Daemon404> just to spite you
[02:01:02 CET] <cone-057> ffmpeg 03Michael Niedermayer 07master:d24af7044d6e: avcodec/vc1_pred: Fix undefined shifts
[04:07:56 CET] <llllll> Hey, my ./configure tells me that my yasm/nasm is not found or too old
[04:15:13 CET] <lordkrondor> should ask that in #ffmpeg, but download a new yasm, compile it and install it.
[07:15:39 CET] <soulslicer> Hello is  Eugen Hoyos around? I have questions on GSOC
[08:59:00 CET] <gb27> hello,
[08:59:48 CET] <gb27> I have subscribed to ffmpeg-devel list,but not received any confirmation mail nor getting any mails of the group.
[08:59:59 CET] <gb27> I need to submit my patch,pls hepl me ot
[09:53:13 CET] <ninten> hi i was working on als decoder as my qualification task, it is compiling successfully but i need to test that i dont know how to do it !!! anyone can help !!
[10:29:03 CET] <gb27>  hello, I have subscribed to ffmpeg-devel list,but not received any confirmation mail nor getting any mails of the group.I need to submit my patch,pls help me out
[10:39:27 CET] <saste> ubitux, do you want to comment on my showwavespic patch?
[10:39:46 CET] <ubitux> i want to, but i'm a bit in a rush currently
[10:39:57 CET] <ubitux> i also leave at the end of the week and still haven't prepared shit
[10:40:00 CET] <ubitux> i'll try to review tonight
[10:40:10 CET] <saste> ubitux, ok
[10:40:23 CET] <saste> ubitux, it's basically your code, but i moved it to a separate filter
[10:40:39 CET] <ubitux> ok
[10:40:39 CET] <saste> i plan to add more extension to that filter, to enable more waves visualizations mode
[10:41:03 CET] <ubitux> you'll get a review in the next 12 hours
[10:41:25 CET] <saste> also my point is that having a separate filter for that is less confusing (from the POV of users and developers)
[10:41:41 CET] <ubitux> yeah, that's fine with me (at least in theory)
[10:41:54 CET] <saste> michaelni, do you want to review my lavfi/eq and lavf/segment patches?
[10:44:44 CET] <saste> ubitux, where are you leaving for?
[10:45:28 CET] <ubitux> far away, 1 month
[10:45:57 CET] <saste> ubitux, good travel :-)
[10:46:30 CET] <ubitux> yeah hopefully :p
[11:16:54 CET] <mateo`> hello there o/
[11:17:09 CET] <mateo`> would anyone be interested in reviewing my avfoundation patches ?
[13:00:30 CET] <poste9> guys what codec should I use to DECODE h264 packet?
[13:00:47 CET] <poste9> using av_codec_find_decoder(AV_CODEC_ID_H264) not working
[13:04:18 CET] <poste9> nvm... looks like I have to use av_codecc_register again wtf? oO
[13:22:22 CET] <sourabh_> Do anyone know what is the role of video encoder in converting videos format?
[14:34:25 CET] <Daemon404> it is sad that github makes firefox block for a good several seconds while trying to view mov.c
[14:35:58 CET] <wm4> Daemon404: be a neckbeard and disable js?
[14:36:53 CET] <Daemon404> p. sure github wouldnt even work
[14:48:47 CET] <nevcairiel> githubs repo browser sucks in many ways
[14:48:58 CET] <nevcairiel> tryng to view large diffs is also impossible
[14:49:05 CET] <Daemon404> yes
[14:49:07 CET] <nevcairiel> since its incapable of loading an individual files diffs
[14:49:11 CET] <nevcairiel> it loads all the files
[14:49:16 CET] <nevcairiel> and then is slow
[14:49:32 CET] <Daemon404> some files it will just say "cant display whole file, too many contributers"
[14:49:35 CET] <Daemon404> or something
[14:49:41 CET] <Daemon404> the future is now
[15:17:03 CET] <cone-595> ffmpeg 03Tobias Rapp 07master:5be04f978971: doc/protocols: Fixed missing example section marker
[15:17:04 CET] <cone-595> ffmpeg 03Andreas Cadhalpun 07master:39ddda12f17c: mxfenc: fix memleaks in mxf_write_footer
[16:21:50 CET] <wm4> lol
[16:22:02 CET] <wm4> so some months ago I changed the AVBuffer code to use mutexes
[16:22:14 CET] <wm4> now I look at ffmpeg's buffer.c...
[16:22:21 CET] <wm4> and the old atomics code is still in ifdefed!
[16:22:36 CET] <wm4> ffmpeg really is a rotten code garbage dump, isn't it
[16:24:13 CET] <nevcairiel> yes it is
[16:24:27 CET] <nevcairiel> apparently someone though there would be a build without threading, and it would need atomics then!
[16:26:28 CET] <wm4> yes, because OBVIOUSLY we can't require pthread (or a wrapper) being always present
[16:26:40 CET] <wm4> got to think of all these broken platforms and broken builds people want to make
[16:27:03 CET] <nevcairiel> the more curious question, how do you need atomics if you dont have threads? =p
[16:27:52 CET] <Daemon404> wm4, we need to add big endian amd64 support
[16:27:54 CET] <Daemon404> to match openssl
[16:28:20 CET] <wm4> we should follow openssl's successful project management
[16:28:49 CET] <nevcairiel> and people tell me i'm inconsiderate because i prefer the clean and neat looking libdcadec instead of joining into the effort to bully its developer into working on  avcodec instead =p
[16:29:09 CET] <wm4> when will you submit your libdcadec wrapper?
[16:29:16 CET] <nevcairiel> dunno, should i?
[16:29:27 CET] <wm4> I'd be interested
[16:29:40 CET] <nevcairiel> it doesnt exactly have a build-system, so its everyones own job to get it into place
[16:29:49 CET] <nevcairiel> at least it builds a .a library now
[16:30:53 CET] <nevcairiel> i guess i could just send it
[17:08:42 CET] <jamrial> nevcairiel: is this decoder xlldec renamed?
[17:08:52 CET] <nevcairiel> its xlldec converted to C
[17:08:54 CET] <nevcairiel> xlldec was D
[17:09:11 CET] <jamrial> ah, i see
[17:26:10 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:e6140e782270: avfilter/vf_field: Change enum to int, which is accessed via AVOption as int
[17:26:11 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:736388977189: avfilter/vf_fieldmatch: Change enums to int, which are accessed via AVOption as int
[17:26:12 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:6cd079a4e7ce: avfilter/vf_histeq: Change enums to int, which are accessed via AVOption as int
[17:27:09 CET] <nevcairiel> is it just me, or are 90% of all changes lately for some hypothetical weird system that may not behave like any other system on this world? :p
[17:28:30 CET] <Daemon404> more like to shutup compilers
[17:35:49 CET] <nevcairiel> jamrial: decorrelate? maybe I'm missing what that would really do, but what avcodec expects is audio in the MSBs (as it only has a 32-bit format), and the decoder outputs 24-bit in the LSBs of a 32-bit int
[17:35:57 CET] <nevcairiel> so it just shifts them up
[17:36:10 CET] <Daemon404> oh
[17:36:15 CET] <Daemon404> i misread too
[17:59:20 CET] <jamrial> nevcairiel: i meant planar -> interleave and vice versa. check flacdsp_template.c
[17:59:30 CET] <nevcairiel> but thats not what it does
[17:59:43 CET] <nevcairiel> we have planar output now
[17:59:46 CET] <nevcairiel> so it uses that :d
[18:36:31 CET] <michaelni> nevcairiel, the enum thing will even break with gcc on x86 with the right flags and these flags are default on some embeded system AFAIK
[18:36:45 CET] <michaelni> so its not exactly hypothetical, sadly ...
[18:36:48 CET] <michaelni> i wish it was
[18:45:15 CET] <Compn> dcadec ? has kurosu reviewed it for bugs?!?! :)
[18:50:23 CET] <adi_> hello ,  how can i sen you  project proposal for review....?
[19:12:05 CET] <Compn> doh
[19:17:55 CET] <wm4> Compn: it's only a wrapper for a new external lib
[19:21:17 CET] <jamrial> ok so, it doesn't build shared libraries, doesn't seem to have version defines, and make install does nothing
[19:21:30 CET] <jamrial> the decoder and the wrapper work nicely, though :p
[19:24:48 CET] <nevcairiel> yeah it lacks any kind of build system, two days ago it didnt even build a static library without modifications =p
[19:28:06 CET] <Daemon404> nevcairiel, still better than D
[19:44:33 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:13f8bbfb7051: avformat/mov: fix mis-detection of jpeg2000
[19:57:02 CET] <ruggles> i have a question about the recent slice threading in vf_colormatrix. is it possibly a problem that it's processing yuv420 slices starting/ending on odd rows? i'm seeing random lavfi crashes and so far that's the most likely culprit.
[19:59:22 CET] <ruggles> also, helgrind gives me warnings only when using an odd number of threads
[19:59:57 CET] <ruggles> (which happens to be default)
[20:00:08 CET] <ubitux> what is the resolution of your input?
[20:00:16 CET] <ubitux> helgrind or valgrind?
[20:00:22 CET] <ruggles> helgrind
[20:00:39 CET] <ubitux> how much threads for what height?
[20:00:42 CET] <ruggles> it's 568x320
[20:00:45 CET] <ruggles> 33 threads
[20:00:49 CET] <ubitux> ok
[20:00:50 CET] <ruggles> 32 cores
[20:02:45 CET] <ruggles> i added some debugging in process_slice_yuv420p()
[20:02:58 CET] <ruggles> first line is: [Parsed_colormatrix_0 @ 0xbcd9b80] jobnr=0 nb_jobs=33 slice_start=0 slice_end=9
[20:03:39 CET] <ruggles> next slice is: [Parsed_colormatrix_0 @ 0xbcd9b80] jobnr=1 nb_jobs=33 slice_start=9 slice_end=19
[20:04:49 CET] <ubitux> i wonder if something like this would help http://pastie.org/10033516
[20:06:13 CET] <cone-595> ffmpeg 03Diego Biurrun 07master:91894f27dce5: configure: Disable shift operator precedence warnings with MSVC
[20:06:14 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:33a79e443b86: Merge commit '91894f27dce531c13d7e1be20253877f2e1e9e66'
[20:06:35 CET] <ubitux> i can indeed reproduce the crash
[20:06:37 CET] <ruggles> well, yeah probably. when i specify an even number of threads all the slices are an even number of lines and helgrind doesn't complain.
[20:07:44 CET] <ubitux> ffmpeg -f lavfi -i testsrc=568x320 -threads 3 -vf format=yuv420p,colormatrix=bt709:smpte170m -f null -
[20:07:48 CET] <ubitux> reproducible with this
[20:08:03 CET] <ubitux> just checking if the 0 and 1 case are working
[20:08:12 CET] <ruggles> oh nice
[20:08:21 CET] <ruggles> yeah that's exactly the crash i get
[20:10:33 CET] <ubitux> my diff seems to fix it
[20:12:05 CET] <ubitux> mmh... maybe not exactly
[20:15:00 CET] <cone-595> ffmpeg 03Luca Barbato 07master:96dd04427fc4: isom: Add support for TSCC2
[20:15:00 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:2cfa1fdb08a9: Merge commit '96dd04427fc46e5e25a5ac1f00753207e3aedb0c'
[20:15:46 CET] <atomnuker> Hi, I'm thinking of applying for ffmpeg's GSOC projects, anyone here that can answer a few questions for me?
[20:17:29 CET] <kierank> yes
[20:18:53 CET] <atomnuker> so as far as I understand, I need to contact the person listed on one of the proposal, get a qualification tast, solve it by the 27th of March and if I pass I must apply through google by the 27th of March, right?
[20:19:25 CET] <kierank> pretty much
[20:20:00 CET] <cone-595> ffmpeg 03Clément BSsch 07master:09f057170e52: avfilter/vf_colormatrix: make sure the number of threads is even for yuv420p
[20:20:04 CET] <ubitux> ruggles: should be fixed
[20:21:07 CET] <michaelni> atomnuker, its recommanded to apply to google before finishing the qualification task
[20:21:11 CET] <ubitux> maybe +1 & ~1 would have been better, but oh well..
[20:21:31 CET] <ruggles> ubitux: great. thanks!
[20:22:48 CET] <atomnuker> michaelni: so I need to send e-mails to the project mentor, then apply in Google's site by the 27th
[20:23:06 CET] <atomnuker> do the qualification problems have a seperate deadline or?
[20:23:19 CET] <ubitux> this should probably be backported...
[20:23:59 CET] <michaelni> atomnuker, yes there should be a bit more time for qual tasks than googles deadline thats why you should apply before fiishing the qual task so you dont mis their deadline
[20:24:15 CET] <michaelni> googles application deadline that is
[20:25:50 CET] <atomnuker> alright, thanks, I'll go ahead and send e-mails then, might apply in google's site in a few days
[20:26:19 CET] <atomnuker> I can apply for 2 ffmpeg projects, right? I'm split between 2
[20:27:54 CET] <reimar> I expect you will have to discuss with the respective mentors. Since I don't think you want to do 2 qualification tasks
[20:33:22 CET] <cone-595> ffmpeg 03Luca Barbato 07master:48aef27f5232: x86: Put COPY3_IF_LT under HAVE_6REGS
[20:33:23 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:1eb28479dab3: Merge commit '48aef27f5232794e70ecef0d347b9f65e27a9bad'
[20:40:14 CET] <cone-595> ffmpeg 03Niels Möller 07master:c9ed48e80ef8: dca: Read params->pancABIT0[param_index] only if part0 is set
[20:40:15 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:6d57da9aa441: Merge commit 'c9ed48e80ef807ab0c1bb946ac8db5f34d83d9c9'
[20:48:43 CET] <cone-595> ffmpeg 03Carl Eugen Hoyos 07master:c0f504e947cc: libx264: Allow full-range yuv422 and yuv444 pixel formats
[20:48:44 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:182499be721e: Merge commit 'c0f504e947cc0a3c73770419bf7d1af98288fc59'
[20:56:28 CET] <cone-595> ffmpeg 03Vittorio Giovara 07master:02d76141d68e: riff: Add 'tshd' FourCC for H.264
[20:56:29 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:59f687b73ceb: Merge commit '02d76141d68e38c80f9a205a56b9af10d74f0995'
[21:06:09 CET] <cone-595> ffmpeg 03Paul B Mahol 07master:cf9861dea5d4: isom: Add X-Com Radvision FourCC
[21:06:10 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:de732a82b618: Merge commit 'cf9861dea5d42d0b5a6cf0609b9624525d96c1e0'
[21:18:58 CET] <cone-595> ffmpeg 03Carl Eugen Hoyos 07master:4326bc364b58: lavf: Do not list mov-only codecs in riff tags
[21:18:59 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:07ca090e814e: Merge commit '4326bc364b58e97fc3d89417054a6b7610179a00'
[21:26:53 CET] <cone-595> ffmpeg 03Vittorio Giovara 07master:7c60c124c8df: hqx: Drop unused offsets table
[21:26:54 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:66de11b064fc: Merge commit '7c60c124c8dfaebfafa2b3c11b316db549617732'
[21:37:48 CET] <cone-595> ffmpeg 03Vittorio Giovara 07master:cdce9e8025ff: aacsbr: Fix type for index variable
[21:37:49 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:45f62c74dd40: Merge commit 'cdce9e8025fff1dee2fd3d6bc28aebc0a330c5a0'
[21:56:39 CET] <cone-595> ffmpeg 03Vittorio Giovara 07master:423f5d50e499: mov: Fix little endian audio detection
[21:56:40 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:c7e1e4c48aee: Merge commit '423f5d50e499ab972f507115700c8f31ed8993e4'
[22:06:21 CET] <cone-595> ffmpeg 03Vittorio Giovara 07master:6289706b35d1: fate: Move tdsc to the appropriate screen capture file
[22:06:22 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:ae3d2afd4680: Merge commit '6289706b35d1b030258bba54cc27bb6d2c04bc9b'
[22:32:23 CET] <llllll> Can anyone help me with LMS algorithm implementation?
[22:48:15 CET] <llllll> nvm
[22:57:06 CET] <cone-595> ffmpeg 03Michael Niedermayer 07master:0b961423ee9f: avcodec/aacsbr: Replace impossible condition by assert
[23:26:39 CET] <cone-595> ffmpeg 03James Almer 07master:bf2bd5b807bc: avcodec/ac3dec_fixed: fix compilation when ac3dec is disabled
[00:00:00 CET] --- Wed Mar 18 2015