[Ffmpeg-devel-irc] ffmpeg-devel.log.20141003

burek burek021 at gmail.com
Sat Oct 4 02:05:02 CEST 2014


[01:58] <akira4> ubitux,  I changed the  get_pts function. http://pastebin.com/VKbYHBPf . It still doesnt work though. 
[02:19] <cone-21> ffmpeg.git 03Michael Niedermayer 07master:77b3e110d2de: avcodec: bump micro version for av_lockmgr_register() changes in a950edb472e8823e34832c7313ba447b2db76f27
[02:19] <cone-21> ffmpeg.git 03Michael Niedermayer 07master:06e7d58410a1: avcodec/jpeglsdec: Check run value more completely in ls_decode_line()
[02:19] <cone-21> ffmpeg.git 03Michael Niedermayer 07master:5c378d6a6df8: avcodec/mjpegdec: check bits per pixel for changes similar to dimensions
[03:13] <cone-21> ffmpeg.git 03James Almer 07master:ddb813b0ef8f: mlpdec: move rematrix_channels code to output_data()
[03:13] <cone-21> ffmpeg.git 03James Almer 07master:0de1d6287ec3: x86/mlpdec: add ff_mlp_rematrix_channel_{sse4,avx2}
[03:47] <cone-21> ffmpeg.git 03Christophe Gisquet 07release/2.3:f74206cb40ee: wavpack: report if there is no bits left
[03:47] <cone-21> ffmpeg.git 03Christophe Gisquet 07release/2.3:25864cf562c3: wavpackenc: proper buffer allocation
[03:47] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:29ee8b72c40b: avcodec: fix aac/ac3 parser bitstream buffer size
[03:47] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:8a2022405964: avcodec/utils: add GBRP16 to avcodec_align_dimensions2()
[03:47] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:17a6ca7d3179: avcodec/vc1dec: fix null pointer dereference
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:d3bfb66a6665: ffmpeg: Fix missing {} in Stream to Packet side data code
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:e13e928baa6b: avcodec/snow: check coeffs for validity
[03:48] <cone-21> ffmpeg.git 03wm4 07release/2.3:9e43d92d6a16: oggdec: fix invalid free on error
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:9a840d5e1745: avcodec/mjpegdec: fix green vertical line at the right with upscale h
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:f903147f2d17: avcodec/mjpegdec: fix green line at the bottom with upscale v
[03:48] <cone-21> ffmpeg.git 03Mark Harris 07release/2.3:08f56b846c23: doc/filters.texi: fix filter name in examples
[03:48] <cone-21> ffmpeg.git 03Mika Raento 07release/2.3:8df77c37584a: segment: don't access outside seg->frames array
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:d18d48def647: avformat/swfdec: Use side data to communicate w/h changes to the decoder
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:eefb6b654dbd: avformat/swfdec: Do not change the pixel format
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:f68395f7fcab: avcodec/h264: Allow partial escaping
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:77bb6b5bcc84: avcodec/mpegvideo: Use "goto fail" for all error paths in ff_mpv_common_frame_size_change()
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:80aec733ada5: avcodec/mpegvideo: check that the context is initialized in ff_mpv_common_frame_size_change()
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:beb55b398134: avcodec/mpegvideo: Set err on failure in ff_mpv_common_frame_size_change()
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:6359be67517f: avformat/m4vdec: Check for non startcode 00 00 00 sequences in probe
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:6952f6f39b5e: avcodec/cabac_functions: fix "warning: UNCHECKED_BITSTREAM_READER is not defined, evaluates to 0"
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:9c57328b8192: ffmpeg: Do av_buffersink_set_frame_size() when reconfiguring the filtergraph not just when changing audio resample parameters
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:be94d15a03e7: tools/crypto_bench: fix build when AV_READ_TIME is unavailable
[03:48] <cone-21> ffmpeg.git 03Katerina Barone-Adesi 07release/2.3:5c3bc127ca68: apetag: Fix APE tag size check
[03:48] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:1c733a440aeb: avcodec/x86/vp9lpf: Always include x86util.asm
[03:48] <cone-21> ffmpeg.git 03Pascal Massimino 07release/2.3:4a4e30a6d87a: libavcodec/webp: treat out-of-bound palette index as translucent black
[03:48] <cone-21> ffmpeg.git 03Gianluigi Tiesi 07release/2.3:8e276fc96a30: avcodec/libilbc: support for latest git of libilbc
[03:48] <cone-21> ffmpeg.git 03Pascal Massimino 07release/2.3:aa40bbb492af: avcodec/webp: fix default palette color 0xff000000 -> 0x00000000
[03:48] <cone-21> ffmpeg.git 03Benoit Fouet 07release/2.3:ed9c6529f032: avformat/riffenc: Filter out "BottomUp" in ff_put_bmp_header()
[03:48] <cone-21> ffmpeg.git 03Philip DeCamp 07release/2.3:f913da3e15e0: libavutil/opt: fix av_opt_set_channel_layout() to access correct memory address
[03:49] <cone-21> ffmpeg.git 03lvqcl 07release/2.3:437f6fb48841: avutil/x86/cpu: fix cpuid sub-leaf selection
[03:49] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:ca2ccd85d7d1: avcodec/ac3enc_template: fix out of array read
[03:49] <cone-21> ffmpeg.git 03Reimar Döffinger 07release/2.3:45ca270ec9f4: configure: add noexecstack to linker options if supported.
[03:49] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:5e34dded10f4: avcodec/jpeglsdec: Check run value more completely in ls_decode_line()
[03:49] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:57bdb3f3dde3: avcodec/mjpegdec: check bits per pixel for changes similar to dimensions
[03:49] <cone-21> ffmpeg.git 03Michael Niedermayer 07release/2.3:d005e2ecce5c: update for FFmpeg 2.3.4
[04:01] <cone-21> ffmpeg.git 03James Almer 07fatal: ambiguous argument 'refs/tags/n2.3.4': unknown revision or path not in the working tree.
[04:01] <cone-21> Use '--' to separate paths from revisions
[04:01] <cone-21> refs/tags/n2.3.4:HEAD: x86/mlpdec: add ff_mlp_rematrix_channel_{sse4,avx2}
[04:19] <BBB> wm4: altref frames
[04:20] <BBB> ubitux: I believe there are indeed security concerns, its basically just whatever chrome links to but isnt shipped by them is a red flag
[04:20] <BBB> ubitux: since they cant fix it themselves, so the fate of a secure chrome lies with the user; that sounds awesome for 31337 h4x0rs, but most chrome users arent that, so they dont want it
[04:21] <BBB> ubitux: but obviously if debian wants to maintain such a patch and take care of updating and call it something else (since chrome name is protected for google binaries), they dont mind; its opensource, after all
[05:06] <cone-21> ffmpeg.git 03Michael Niedermayer 07master:105654e376a7: avcodec/utils: Add case for jv to avcodec_align_dimensions2()
[10:37] <j-b> 'morning
[10:43] <Rathann> hi j-b 
[12:23] <Irish_Galway> Hi anyone there?
[12:24] <Irish_Galway> has anyone noticed a problem with iOS8 playing native video mode with video off the net, and could it be ffmpeg related if that video was converted by it?
[12:48] <thardin> been fuzzing for 25 hours no. no results
[12:48] <thardin> I refuse to believe mxfdec.c is this stable
[12:50] <alpa_astero> Maybe you need a smarter fuzzer?
[12:51] <thardin> I'm testing american-fuzzy-lop. it's supposed to be smart
[12:51] <thardin> though to be fair it hasn't run through the entire first batch of fuzzing. it's perhaps too methodical
[12:52] <thardin> it slips each bit in the file, then each adjascent pair, each nybble etc
[12:52] <thardin> flips
[13:09] <cone-286> ffmpeg.git 03Michael Niedermayer 07master:d225b0f7aaa6: avcodec/h263dec: Fix decoding messenger.h263
[14:02] <thardin> there, got a crash. but I using the generated file doesn't reproduce it
[14:13] <knoch> hello, how to fix a pts and dts negative value in an AVPacket returned by av_read_frame() ?
[14:53] <Compn> thardin : probably set up some printfs and see if the values in that file are correct
[14:53] <Compn> if you cant get it to crash
[14:53] <Compn> i dunno
[14:53] <Compn> by file i mean source code
[14:57] <thardin> Compn: I'll let it run some more. I suspect it'll generate more test files similar to that one
[14:57] <thardin> plus I made a much smaller file to test with
[14:58] <Daemon404> thardin, valgrind doesnt show anything?
[14:59] <thardin> didn't check yet. at another machine though
[15:09] <cone-286> ffmpeg.git 03Michael Niedermayer 07master:97cb9b6cbb0e: Revert "avformat/img2dec: reduce bmppipe probe score"
[15:09] <cone-286> ffmpeg.git 03Michael Niedermayer 07master:8b0e96e1f21b: avcodec/mmvideo: Bounds check 2nd line of HHV Intra blocks
[15:46] <ubitux> so ffmpeg doesn't open all fifo at startup? interesting
[15:48] <ubitux> it seems it waits for at least getting some input into one
[15:53] <ubitux> that could become a bit troublesome but well
[15:53] <TimNich> thardin:  Will be interested to see how you get on with AFL (sounds like a breed of rabbit), the readme confused me when it said "The fuzzer is optimized for compact binary data formats, such as images and other multimedia" and then suggests " the fuzzer requires one or more input file& Whenever possible, the file should be reasonably small; under 1 kB is ideal". What size is your starting file as I guess it has an exponential impact on
[15:53] <TimNich> run time.
[15:55] <thardin> TimNich: yeah, considering it tests *every* bitflip the input needs to be small
[15:55] <thardin> I'm considering making it test every bitflip in say 1kbit ranges
[15:56] <thardin> so it goes "flip bit 0, skip bits 1..1023, flip 1024, skip 1025..2047"
[15:56] <thardin> then moves one bit forward for the next iteration
[15:58] <TimNich> Ideally skip from one K to the next, then try the L's then the V's&..
[16:46] <cone-286> ffmpeg.git 03Michael Niedermayer 07master:e1c0cfaa419a: avcodec/tiff: more completely check bpp/bppcount
[18:29] <cone-492> ffmpeg.git 03Michael Niedermayer 07master:e830902934a2: avcodec/pngdec: Calculate MPNG bytewidth more defensively
[19:44] <cone-492> ffmpeg.git 03Michael Niedermayer 07master:e7e5114c5069: avcodec/cinepak: fix integer underflow
[20:08] <llogan> Daemon404: want to take a look at my libx265 "documentation" patch?
[20:22] <omero> hello :D
[20:22] <omero> I'm trying to build mplayer but it stops with an error building ffmpeg
[20:23] <JEEB> that's an mplayer issue, they bundle a random version of ffmpeg there
[20:23] <JEEB> meanwhile you should take a look at mpv, you might like it
[20:24] <omero> I got the mplayer source trunk via svn
[20:24] <omero> during the --configure, it will check out the git ffmpeg
[20:25] <JEEB> it will check something out, nobody knows what
[20:25] <omero> rofl
[20:25] <omero> any suggestions to get over this impasse?
[20:25] <JEEB> take it to the people who maintain what mplayer checks out, in other words the mplayer project
[20:26] <JEEB> or you give up on the old mplayer and test out mpv
[20:28] <cone-492> ffmpeg.git 03Michael Niedermayer 07master:8f1457864be8: avcodec/gifdec: factorize interleave end handling out
[20:30] <omero> ok JEEB ... thanks for the hints
[20:30] <omero> I'm trying to manually checkout the git ffmpeg and give another try before giving up
[20:31] <wm4> try a ffmpeg version that matches the data of the latest mplayer commit
[20:31] <wm4> I don't know why mplayer keeps insisting on its utterly broken build system and ffmpeg usage
[20:32] <wm4> almost all of these breakages are avoidable
[20:33] <omero> wm4, yes.. every time is a hard struggle... that's why I only try upgrading mplayer very rarley
[20:34] <omero> how would I know which ffmpeg version matches?
[20:34] <JEEB> by asking the mplayer developers
[20:34] <JEEB> or looking at the date of your mplayer commit or whatever
[20:35] <omero> ah ok
[20:35] <JEEB> in any case, take this to #mplayer if you want to build mplayer
[20:35] <JEEB> because this most definitely does not relate to FFmpeg as such
[20:35] <wm4> yeah, it's offtopic here
[20:35] <wm4> mplayer doesn't even use ffmpeg in a way that's supported
[20:36] <wm4> because it partially replaces its build system
[20:40] <omero> ok, sorry ... 
[20:57] <omero> http://pastebin.com/AEBrVdBm
[20:57] <omero> maybe it can be ffmpeg related
[20:58] <kepstin-laptop> omero: your libx264 might just be too old. The ffmpeg configure checks for that, but mplayer's might not.
[20:59] <omero> ok, that's a good hint, thanks kepstin-laptop 
[21:18] <omero> yup! that solved it, thanks kepstin-laptop 
[21:19] <omero> my x264 was 3 years old :P
[21:32] <cone-492> ffmpeg.git 03Michael Niedermayer 07master:dd3bfe3cc1ca: avcodec/qpeg: fix off by 1 error in MV bounds check
[21:54] <cone-492> ffmpeg.git 03Joakim Roubert 07master:55f03d872640: hlsenc: Add parameter -hls_allow_cache
[21:54] <cone-492> ffmpeg.git 03Michael Niedermayer 07master:883bc7d2e7c7: Merge commit '55f03d872640175a00bfa262da4652b7312b905f'
[23:03] <ubitux> #4002, call for flames
[23:07] <J_Darnley> What's wrong with the name FFmpeg?
[23:07] <J_Darnley> What does git have to do with source sode?
[23:24] <cone-492> ffmpeg.git 03Michael Niedermayer 07master:c727401aa9d6: avcodec/smc: fix off by 1 error
[23:28] <llogan> so he likes bugzilla better than trac?
[23:29] <wm4> I wonder if he made the post on libav too
[23:30] <llogan> i'll be in Nome and Barrow of all places for a week or so starting Sunday, so I probably won't be able to do ML stuff
[23:34] <wm4> indeed he did https://bugzilla.libav.org/show_bug.cgi?id=756
[23:34] <wm4> oh he even wrote that
[23:56] <akira4> ubitux, would you mind having a look? I dont think I understood the putting buffer before timestamp part. Here is the link--http://pastebin.com/7d144DP3 
[23:57] <ubitux> you're suppose to shift *buf
[23:57] <ubitux> you can for example use %n in the sscanf
[23:57] <ubitux> which you assign to a int pos, then do *buf += pos
[23:57] <ubitux> that's one solution
[23:58] <ubitux> look how it's done in vplayerdec.c
[00:00] --- Sat Oct  4 2014


More information about the Ffmpeg-devel-irc mailing list