[Ffmpeg-devel-irc] ffmpeg-devel.log.20150429

burek burek021 at gmail.com
Thu Apr 30 02:05:03 CEST 2015


[00:08:37 CEST] <drwx> 12:41 < drwx> hi, could anyone look at this bug? https://trac.ffmpeg.org/ticket/2975
[00:08:40 CEST] <drwx> 12:41 < drwx> also, https://trac.ffmpeg.org/ticket/4487 looks like a duplicate
[01:14:47 CEST] <penthief> Do you want a bug report containing something like the following http://paste.lisp.org/display/147622 ? (Incorrect resolution with x11grab causes a segfault when switching windows.) 
[01:16:15 CEST] <penthief> I can also chuck you a 7 meg core dump if you tell me where to put it.
[01:18:12 CEST] <penthief> Possibly introduced with a6674d2e7771dbf7a4a5556f5e126be83cadac96 (Aug 24 2014).
[03:41:24 CEST] <cone-188> ffmpeg 03Vittorio Giovara 07master:f5ba67ee1342: flacenc: Move a scratch buffer to struct used by the function
[03:41:24 CEST] <cone-188> ffmpeg 03Michael Niedermayer 07master:77010a41fef0: Merge commit 'f5ba67ee1342b7741200ff637fc3ea3387b68a1b'
[03:56:27 CEST] <cone-188> ffmpeg 03Vittorio Giovara 07master:3919a4572690: jpeglsenc: Check memory allocations
[03:56:28 CEST] <cone-188> ffmpeg 03Michael Niedermayer 07master:27801500eed3: Merge commit '3919a4572690894d0a7ad4170d699c52b4748194'
[04:03:25 CEST] <cone-188> ffmpeg 03Vittorio Giovara 07master:c3418201307c: jpeglsenc: Mark codec as init-thread-safe and init-cleanup
[04:03:26 CEST] <cone-188> ffmpeg 03Michael Niedermayer 07master:966a57b6e918: Merge commit 'c3418201307c6b7a2b3e3be10f33ab4a20a32c3b'
[04:11:08 CEST] <cone-188> ffmpeg 03Vittorio Giovara 07master:59dfc2b0c89f: h263: Drop commented-out code and a related otherwise unused function
[04:11:09 CEST] <cone-188> ffmpeg 03Michael Niedermayer 07master:78f50e7eb694: Merge commit '59dfc2b0c89f35c5805abee72c2ae4ee7af8e98d'
[04:22:49 CEST] <cone-188> ffmpeg 03Vittorio Giovara 07master:0c69164f451c: h263: Convert function to macro
[04:22:50 CEST] <cone-188> ffmpeg 03Michael Niedermayer 07master:f6c524a9399f: Merge commit '0c69164f451cc9ca6ce9d6e7568083e2776bc845'
[05:46:40 CEST] <cone-188> ffmpeg 03John Högberg 07master:82de8d71118f: mpegts: Update the PSI/SI table only if the version change
[05:46:41 CEST] <cone-188> ffmpeg 03Michael Niedermayer 07master:12f7253d0ea1: Merge commit '82de8d71118f4eafd6a43e9ea9169bd411793798'
[05:46:42 CEST] <cone-188> ffmpeg 03Michael Niedermayer 07master:639781492684: avformat/mpegts: reset last_version on seeking
[05:56:12 CEST] <cone-188> ffmpeg 03Luca Barbato 07master:0ac8ff618c5e: avresample: Reallocate the internal buffer to the correct size
[05:56:13 CEST] <cone-188> ffmpeg 03Michael Niedermayer 07master:215430e049b7: Merge commit '0ac8ff618c5e6d878c547a8877e714ed728950ce'
[11:37:09 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:27b0e6ebfd47: h264: drop needs_realloc
[11:37:10 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:0d5e615c1058: Merge commit '27b0e6ebfd47b0c11156c18b90fa8c571f0f60c3'
[11:46:04 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:9d33bab583a8: h264: drop H264Context.ouputed_poc
[11:46:05 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:e8a460672f7e: Merge commit '9d33bab583a82cf12286c65258a29c6888e1ff98'
[11:55:52 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:a6cd154463be: h264: move the DPB init/uninit to init_context()/free_context()
[11:55:53 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:cd63252f1ff6: Merge commit 'a6cd154463bea7eb56d28192db4c8c6d83f67fd7'
[12:20:08 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:54986d6db911: h264: move context reinit lower down in update_thread_context()
[12:20:09 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:580c86925ddf: Merge commit '54986d6db9116ef9704b1ce8414ffdb7f1ca127f'
[13:24:36 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:ee62b364be0c: h264: eliminate ff_h264_set_parameter_from_sps()
[13:24:37 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:ab7ad5e6a1b8: Merge commit 'ee62b364be0c30cba83b5ff10a3ca8c3e866ade6'
[13:33:36 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:11f024ef0ab9: h264: move freeing the escaped RBSP buffer to free_context()
[13:33:37 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:d3a5dba81b3a: Merge commit '11f024ef0ab923ed8680fc35a087d576e549c849'
[13:41:23 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:440e8dd374b7: h264: drop a comment that carries no useful information
[13:41:24 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:1aa3a0f457c3: Merge commit '440e8dd374b732c48d564d9f1bb0ec3b1b786fb9'
[13:48:31 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:29dcde2b8755: h264: copy nal_length_size in update_thread_context
[13:48:32 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:16b90d0bcbca: Merge commit '29dcde2b87559b72957454a7e9bcd5227d8057ba'
[13:55:53 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:c9ccbc7333ed: h264: update the current bit depth/chroma type in init_slice_header()
[13:55:54 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:51f7ba9ceec0: Merge commit 'c9ccbc7333eddd025ebbde5cc4f27d68a950c623'
[14:02:52 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:8a66fd40260b: h264: drop the reinit parameter from init_slice_header()
[14:02:53 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:cc87493f0654: Merge commit '8a66fd40260b7aae6226d68c4dbad43b05a8e524'
[15:22:26 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:88c612e3a4d5: h264: merge the init and reinit paths in update_thread_context()
[15:22:27 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:bd7dce594ea7: Merge commit '88c612e3a4d5b584e2d9f6e2541de78d67bdfb9f'
[15:22:28 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:4a95cd5d9b4b: avcodec/h264_slice: Also copy x264_build in ff_h264_update_thread_context()
[15:36:36 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:7a4f74eed51f: h264: embed the DPB in the context
[15:36:37 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:07a79cf8694a: Merge commit '7a4f74eed51f914e9bbfebaffd4a92ac6791f819'
[16:16:02 CEST] <saste> kierank, michaelni, do you know why there is no multi-program support in the mpeg-ts muxer?
[16:16:33 CEST] <saste> see http://trac.ffmpeg.org/ticket/4525
[16:48:26 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:a0f2946068c6: h264: use properly allocated AVFrames
[16:48:27 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:36f862e04c2a: Merge commit 'a0f2946068c62e18cb05ac25c0df3d86077251a6'
[17:00:15 CEST] <cone-503> ffmpeg 03Anton Khirnov 07master:a09c499a2e25: rawenc: properly allocate coded_frame
[17:00:16 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:50e46b7f0dca: Merge commit 'a09c499a2e25ded575cb9f37528384402be86ad8'
[17:19:32 CEST] <cone-503> ffmpeg 03foo86 07master:e80b2b9c8171: avcodec/dca_parser: Extend DTS core sync word and fix existing check
[17:26:07 CEST] <j-b> So, what's the status/mess with DCA?
[17:26:43 CEST] <ubitux> stall game?
[17:26:43 CEST] <Compn> i think they are arguing about some encoder patch 
[17:27:33 CEST] <Daemon404> "just use foo86's library"
[17:28:03 CEST] <j-b> I mean, there is a new DTS-HD decoder in libav, but was it merged here? There was different code here for DTS-HD too, right? And there is the foo86 decoder, too, now?
[17:28:23 CEST] <Daemon404> j-b, the foo86 one is the only one with a fixed point dca core implementation
[17:28:27 CEST] <Daemon404> i.e. actually lossless
[17:29:22 CEST] <j-b> Daemon404: ok. The libav one was merged in FFmpeg or?
[17:29:57 CEST] <Daemon404> hell if know
[17:30:18 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:8a710ddc7c11: avcodec/libx264: Remove cleanup code, its uneeded since FF_CODEC_CAP_INIT_CLEANUP
[17:32:24 CEST] <Compn> its chaos! 
[17:32:25 CEST] <Compn> :P
[17:44:09 CEST] <wm4> j-b: it was merged, and disabled by default
[17:44:31 CEST] <wm4> at least AFAIK
[17:44:53 CEST] <wm4> and the libdcadec wrapper exists in both libs too now
[17:45:16 CEST] <j-b> wm4: disabled, because of innaccuracy?
[17:45:38 CEST] <j-b> what about X96 ?
[17:45:39 CEST] <wm4> and I think there were other problems with it too? maybe nevcairiel knows details
[17:45:48 CEST] <wm4> no idea
[17:46:07 CEST] <wm4> if you want some feature, I'd say foo86 is likely to implement it in his lib
[18:27:53 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:14e9a20083c9: avformat/mpegts: Also parse the FMC descriptor if the codec has not been identified yet
[18:29:44 CEST] <nevcairiel> j-b: wm4: libav disabled it as well by default on my suggestion, to avoid regressions in decoding, which do exist, and X96 is not supported
[18:30:12 CEST] <nevcairiel> dcadec from foo86 supports all of the things in perfect accuracy however, the only missing feature is DTS Express, aka. LBR
[18:32:43 CEST] <kierank> That is a separate codec
[18:34:19 CEST] <nevcairiel> well it uses the same extension header, and presumably shares similar coding techniques
[18:35:57 CEST] <kierank> It's different totally
[18:36:11 CEST] <kierank> It's like e-ac3 vs ac-3
[18:36:20 CEST] <kierank> But it is marketed like it is similar
[18:36:31 CEST] <nevcairiel> which share a lot of code in avcodec? :d
[18:36:49 CEST] <kierank> Yes share code because of headers etc
[18:36:56 CEST] <kierank> But it's a bigger gap
[18:37:01 CEST] <j-b> nevcairiel: regressions? how can there be regressions?
[18:37:12 CEST] <j-b> I mean, the core is the core, no?
[18:37:30 CEST] <nevcairiel> j-b: trying to parse HD parts, failing, causing odd behavior, like switching between 5.1 and 7.1 decoding every other audio frame
[18:39:51 CEST] <j-b> nevcairiel: do you have samples for such fails?
[18:40:19 CEST] <nevcairiel> somewhere, i gave them to the libav people when it was first merged
[18:54:04 CEST] <rcombs> any particular reason why libavformat/tls.c doesn't verify certs by default?
[19:01:19 CEST] <wm4> because it just breaks shit
[19:01:48 CEST] <rcombs> what kind of shit
[19:01:52 CEST] <rcombs> (maybe someone should fix that)
[19:02:43 CEST] <BtbN> I think enablding it would break a bunch of stuff.
[19:03:17 CEST] <BtbN> And cause a bunch of angry people, because their self-signed certs suddenly don't work anymore.
[19:03:51 CEST] <rcombs> there's an AVOption to disable it
[19:04:04 CEST] <BtbN> Yes, but it breaks existing setups.
[19:04:17 CEST] <BtbN> changing defaults allways does
[19:04:32 CEST] <rcombs> biggest issue I see is that it doesn't seem to pull in the system cert bundle by default
[19:05:23 CEST] <wm4> rcombs: maybe the main reason is because nobody wants to fight these crappy tls libs
[19:06:10 CEST] <rcombs> it's not like self-signed cert users haven't had to disable verification in every other piece of software they use (or actually add their certs to their cert bundle, in which case this isn't a problem if that's used [which it isn't but should be])
[19:10:41 CEST] <rcombs> hmm, curl (which does this pretty reliably cross-platform) uses OSX's and Windows's built-in TLS facilities (Secure Transport and Schannel), and on Linux the default CA bundle for OpenSSL/GNUTLS and all the other libs they support is a configure option
[19:16:36 CEST] <wm4> make libavformat use libcurl
[19:18:20 CEST] <rcombs> I've seen that done, actually
[19:18:53 CEST] <rcombs> quite some time ago, before lavf's protocol support was as stable
[19:19:09 CEST] <rcombs> (wait a sec why is avio part of lavfnever mind)
[19:20:37 CEST] <nevcairiel> i dont bundle any tls/ssl support in my windows builds right now because openssl is a license nightmare, and gnutls is terrible to build on windows
[19:20:54 CEST] <nevcairiel> i was meaning to see if building a schannel native-windows th ing would be as terrible as people claim..
[19:25:19 CEST] <rcombs> Secure Transport does let you provide your own read/write callbacks, so tls.c could continue to use lavf's tcp interface as usual
[19:25:24 CEST] <rcombs> hmm, I might actually do this
[20:00:55 CEST] <ramiro> are aac frames always standalone? (as in they don't depend on data from previous frames)
[20:22:10 CEST] <cone-503> ffmpeg 03Michael Niedermayer 07master:828759ed9afa: avcodec/cook: Remove unneeded #if
[20:25:36 CEST] <michaelni> about ssl certs, everything should check them by default IMHO, its more secure
[20:28:34 CEST] <wm4> lol
[20:28:36 CEST] <wm4> "secure"
[20:28:48 CEST] <wm4> a remote server can make ffmpeg connect anywhere you want
[20:34:16 CEST] <rcombs> wm4: http://suptg.thisisnotatrueending.com/archive/24232601/images/1365921834628.jpg
[20:36:59 CEST] <wm4> at least http redirects and things like MOV reference files and HLS (which open further requests)
[20:44:53 CEST] <rcombs> well that's the server you asked for (which could be verified) directing you somewhere else
[23:23:53 CEST] <cone-837> ffmpeg 03Michael Niedermayer 07master:f8fa392a2edb: avfilter/vf_vignette: force per frame evaluation if per frame variables are used
[00:00:00 CEST] --- Thu Apr 30 2015


More information about the Ffmpeg-devel-irc mailing list