[Ffmpeg-devel-irc] ffmpeg-devel.log.20150315

[burek]

[02:14:14 CET] <cone-251> ffmpeg 03Michael Niedermayer 07master:b8535b273fcb: Revert "ppc: lavc: use LOCAL_ALIGNED instead of DECLARE_ALIGNED"
[03:07:56 CET] <cone-251> ffmpeg 03Michael Niedermayer 07master:3b63819dfbee: avcodec/vc1: Fix undefined shifts
[03:43:59 CET] <cone-251> ffmpeg 03Michael Niedermayer 07release/2.2:53239e50ed36: avformat/mov: Use sizeof(filename) instead of a literal number
[03:44:00 CET] <cone-251> ffmpeg 03Michael Niedermayer 07release/2.2:bf0d2ee92c33: avformat/mov: Check for string truncation in mov_open_dref()
[04:01:10 CET] <cone-251> ffmpeg 03Michael Niedermayer 07n2.2.14:HEAD: avcodec/vc1: Fix undefined shifts
[04:28:32 CET] <cone-251> ffmpeg 03Andreas Cadhalpun 07release/2.6:480e18e6ff93: ac3dec_fixed: always use the USE_FIXED=1 variant of the AC3DecodeContext
[04:28:33 CET] <cone-251> ffmpeg 03Christophe Gisquet 07release/2.6:26bed98d6470: ac3_fixed: fix out-of-bound read
[04:28:34 CET] <cone-251> ffmpeg 03Christophe Gisquet 07release/2.6:7edd38066845: ac3_fixed: fix computation of spx_noise_blend
[04:28:35 CET] <cone-251> ffmpeg 03Christophe Gisquet 07release/2.6:7ff46a20d5cc: eac3dec: fix scaling
[04:28:36 CET] <cone-251> ffmpeg 03Michael Niedermayer 07release/2.6:350054e8e216: avformat/mov: Use sizeof(filename) instead of a literal number
[04:28:37 CET] <cone-251> ffmpeg 03Michael Niedermayer 07release/2.6:37469af29454: avformat/mov: Check for string truncation in mov_open_dref()
[11:45:24 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:94ccbad488d1: avcodec/ac3dec: Fix undefined shifts
[11:45:24 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:e3f5b6f16d60: avcodec/ac3: Fix undefined shift in ff_ac3_bit_alloc_calc_mask()
[13:18:27 CET] <cone-415> ffmpeg 03zhaoxiu.zeng 07master:bf696e265a25: avformat/avidec: use avpriv_find_start_code in avi_read_packet()
[13:18:28 CET] <cone-415> ffmpeg 03zhaoxiu.zeng 07master:213ddcb0298b: avformat/avidec: cleanup include
[13:37:42 CET] <cone-415> ffmpeg 03zhaoxiu.zeng 07master:84d8b4fb812c: avformat/mpegvideodec: use avpriv_find_start_code in mpegvideo_probe()
[13:55:16 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:32c4b504b4ba: avformat/mpegvideodec: fix hypothetical pointer overflow in mpegvideo_probe()
[13:55:17 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:2cda1a16d0d3: avformat/dvbsub: Fix hypothetical pointer overflow in dvbsub_probe()
[14:37:20 CET] <Daemon404> same question here:
[14:37:20 CET] <Daemon404> [13:27] <@Daemon404> if CODEC_FLAG_GLOBAL_HEADER is set, must it be initialized in the encoder_init?
[14:37:24 CET] <Daemon404> [13:27] <@Daemon404> i.e. can i lazily create avctx->extradats when the first frame is encoded
[14:37:27 CET] <Daemon404> [13:32] <@Daemon404> :|
[14:37:29 CET] <Daemon404> [13:33] <@Daemon404> libx265 needs to know if the input is interlaced in init(), but we have absolutely no way of accomplishing this as far as i can tell...
[14:41:03 CET] <kierank> correct
[14:41:17 CET] <kierank> it's the same with x264cli
[14:41:58 CET] <Daemon404> so basically im fucked?
[14:43:29 CET] <Daemon404> the way libx264.c does it is very wrong.
[14:46:32 CET] <Daemon404> similarily, avctx->field_order isn't init'd until after encode_init() is already called. and chanegs every frame.
[14:58:40 CET] <cone-415> ffmpeg 03Stefano Sabatini 07master:6a63d0d826ed: lavfi/eq: clarify error message in case of expression parsing error
[15:23:31 CET] <iive> afair interlace in libx264 is controlled by ilme and ildct flags.
[15:23:38 CET] <iive> you want to make it automatic?
[15:24:18 CET] <Daemon404> yes, and thats wrong
[15:24:29 CET] <Daemon404> it's a hack
[15:56:23 CET] <cousin_luigi> Greetings.
[15:56:39 CET] <cousin_luigi> http://pastie.org/private/f1qhsykeafupxieayszw <- what could be causing this problem? I'm using ffmpeg 2.6 linked against x265 1.5.
[16:03:02 CET] <wm4> ubitux: who's this someone working on subrip?
[16:04:06 CET] <ubitux> a secret friend willing to contribute
[16:04:22 CET] <ubitux> just giving random tasks (such as this sami thing)
[16:04:41 CET] <Daemon404> ubitux, im worried
[16:04:44 CET] <Daemon404> have you seen a doctor?
[16:04:49 CET] <Daemon404> schizophrenia is no laughing matter.
[16:04:53 CET] <ubitux> :((
[16:05:12 CET] <wm4> lol
[16:05:22 CET] <wm4> I thought maybe something with gsoc
[16:05:35 CET] <wm4> or worse, certain mplayer/ffmpeg devs
[16:05:38 CET] <ubitux> nope, not OPW/GSOC
[16:05:47 CET] <michaelni> cousin_luigi, is this a regression ?
[16:05:55 CET] <wm4> (by which I mean a certain person)
[16:06:17 CET] <Daemon404> we all know who you mean
[16:08:51 CET] <ubitux> that's not nicolas, really, just someone shy who wants to start contributing 
[16:09:17 CET] <ubitux> that task sounded appropriate when it was raised :p
[16:09:36 CET] <wm4> shy? afraid to get eaten by us? (given the somewhat "rude" tone we tend to use it's probably a reasonable worry)
[16:10:13 CET] <ubitux> i suppose it's more about contributing to opensource in general
[16:10:20 CET] <ubitux> but i'm not a psychologist :p
[16:10:40 CET] <Daemon404> youd be surprised how many people would rather die than show you their code
[16:10:44 CET] <Daemon404> out of some form of embarassment
[16:10:48 CET] <ubitux> :D
[16:11:15 CET] <ubitux> what do you think is worse, these people or those who are proud to show their poop to the world?
[16:11:43 CET] <Daemon404> as long as said poop is not forced on me, /care
[16:35:17 CET] <cousin_luigi> michaelni: I wondered myself. How can I find out?
[16:35:30 CET] <michaelni> cousin_luigi, test with older versions
[16:35:48 CET] <cousin_luigi> michaelni: I will.
[16:36:06 CET] <cousin_luigi> michaelni: Also is the x265 from bitbucket good or should I use the videolan one instead?
[16:36:28 CET] <michaelni> i have no idea
[16:37:02 CET] <cousin_luigi> michaelni: which version of ffmpeg should I start with?
[16:37:56 CET] <michaelni> master, then the olderst which supported x265 if that builds
[16:40:49 CET] <Daemon404> what does x264 have to do with what you posted?
[16:40:52 CET] <Daemon404> thats the hevc *decoder*
[16:40:57 CET] <Daemon404> s/x264/x265/
[16:41:34 CET] <cousin_luigi> Daemon404: oh
[16:41:50 CET] <cousin_luigi> Daemon404: which is internal?
[16:41:54 CET] <Daemon404> yes.
[16:41:57 CET] <Daemon404> x265 is an encoder.
[16:42:02 CET] <cousin_luigi> I see.
[17:24:57 CET] <cone-415> ffmpeg 03Mark Reid 07master:8d54850f33fb: libavformat/mxfenc: write user comment metadata
[17:39:43 CET] <cousin_luigi> michaelni: Same behaviour with 2.6, master and 2.4. I encoded those files using ffmpeg linked against libx265 1.5. Could that be the problem?
[18:19:29 CET] <michaelni> cousin_luigi, how can this be reproduced ? i mean IIUC encode a file with ffmpeg +libx265 that then does not play with ffmpeg ?
[18:20:16 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:0a4808741e52: avcodec/h264dsp_template: Fix undefined shifts
[18:21:47 CET] <cousin_luigi> michaelni: all files I've encoded have this defect. But they play, only I see that error while seeking.
[18:22:53 CET] <cousin_luigi> michaelni: but yes
[19:10:07 CET] <cone-415> ffmpeg 03Clément BSsch 07master:84da9339c21f: avfilter/palettegen: make sure at least one frame was sent to the filter
[19:10:27 CET] <cone-415> ffmpeg 03Clément BSsch 07release/2.6:b80c486fb3b5: avfilter/palettegen: make sure at least one frame was sent to the filter
[19:56:33 CET] <cousin_luigi> michaelni: Do I need to run some other test? Do you need a sample?
[20:08:47 CET] <michaelni> cousin_luigi, patch on ML that silences the errors
[20:22:14 CET] <cousin_luigi> michaelni: thanks man!
[20:39:28 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:de41d5372faa: avcodec/dvbsub_parser: Fix potential pointer overflows
[21:22:35 CET] <cone-415> ffmpeg 03Diego Biurrun 07master:4da5aacc7eba: dca: Mark syncword constants unsigned; they do not fit in 32-bit signed int
[21:22:36 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:e46a7fdc8748: Merge commit '4da5aacc7eba274a4f18411120de539d39c5151e'
[22:40:20 CET] <cone-415> ffmpeg 03Niels Möller 07master:217e4ff4d1f8: dca: Support for XLL (lossless extension)
[22:40:21 CET] <cone-415> ffmpeg 03Diego Biurrun 07master:acbe15a99f15: fate: Add test for DCA XLL
[22:40:22 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:209001073adf: Merge commit '217e4ff4d1f845b76e44634e29371cd09313d1c2'
[22:40:23 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:96bd0321936b: Merge commit 'acbe15a99f158dbb0edb837fb6557171dc4376d4'
[22:53:46 CET] <cone-415> ffmpeg 03Luca Barbato 07master:a4c580107019: sctp: Update to match tcp
[22:53:47 CET] <cone-415> ffmpeg 03Michael Niedermayer 07master:54db7df34b6b: Merge commit 'a4c5801070198c1f2af32fa5cf850c330615ac24'
[23:30:24 CET] <rcombs> does ffmpeg currently have a way to remux DTS-MA or E-AC-3, stripping down to the "core" track (DTS, AC-3)? If not, seems like a useful thing to have; perhaps as a bitstream filter?
[23:31:12 CET] <jamrial> i think there's a trac ticket requesting dts core extraction
[23:31:35 CET] <rcombs> also, it seems like it'd be useful to export the channel layout of the "core" track if it's different than that of the full one, perhaps in the metadata dict
[23:43:41 CET] <rcombs> eAC3->AC3 could be a weird one, since those have separate AVCodecIDs
[23:48:56 CET] <rcombs> also, why is XLL decoding disabled by default now? Is it known to be buggy, or slow, or& ?
[23:52:43 CET] <jamrial> not bitexact
[23:53:10 CET] <jamrial> something that's kinda important on a lossless codec :p
[00:00:00 CET] --- Mon Mar 16 2015