[Ffmpeg-devel-irc] ffmpeg-devel.log.20171107

[burek]

[00:49:05 CET] <cone-857> ffmpeg 03Martin Vignali 07master:97ac8a47ddc8: libavcodec/blockdsp : fix comment. clear_block need 32 aligned now
[00:52:12 CET] <cone-857> ffmpeg 03Martin Vignali 07master:9c714731895a: libavcodec/utvideo : simplify decode_plane
[00:53:41 CET] <cone-857> ffmpeg 03Martin Vignali 07master:ad784e0cb68f: fate/utvideo : add test for decoding and encoding
[00:55:07 CET] <cone-857> ffmpeg 03Martin Vignali 07master:fbe91487797c: checkasm/llviddsp : add test for other dsp func
[00:56:46 CET] <cone-857> ffmpeg 03Martin Vignali 07master:da62128ea142: libavcodec/lossless_videodsp : add add_bytes avx2 version
[00:57:32 CET] <cone-857> ffmpeg 03Martin Vignali 07master:0380b72d35a0: libavcodec/lossless_video_dsp : cosmetic add better separator for each function, in order to make reading of the asm file easier
[01:17:32 CET] <jamrial> michaelni: did you push the above set?
[01:19:07 CET] <michaelni> no, i assume martin pushed them
[01:20:28 CET] <jamrial> he has push access?
[01:20:45 CET] <michaelni> yes, he is in MAINTAINERS
[01:21:45 CET] <jamrial> mmh, ok
[01:24:03 CET] <cone-857> ffmpeg 03Martin Vignali 07master:7e1f21148dfd: libavcodec/mepgvideo : remove warning introduce by previous patch
[02:51:47 CET] <liyou> hi all, 
[02:52:16 CET] <liyou> how can i debug a big shell script file?
[03:08:48 CET] <iive> liyou: bash -x ./script  
[03:16:55 CET] <liyou> i cant compile aribb24, the extras/tools/bin/pkg-config cant find the libpng
[04:24:51 CET] <cone-905> ffmpeg 03Kaustubh Raste 07master:ecd69cade8ed: avcodec/mips: Improve avc avg mc 10, 30, 01 and 03 msa functions
[04:24:51 CET] <cone-905> ffmpeg 03Kaustubh Raste 07master:991eca0f8729: avcodec/mips: Improve hevc bi 4 tap hv mc msa functions
[06:38:59 CET] <liyou> where can i get libmp3lame
[07:54:49 CET] <liyou_> hi all, i find one bug.
[09:42:05 CET] <JEEB> > has_b_frames
[09:42:12 CET] <JEEB> this seems to be a really painful thing
[10:58:26 CET] <sfan5> is there no interest in LibreSSL support or why is the patch I sent to the ML getting ignored? >.>
[10:58:48 CET] <nevcairiel> we just dont like them for pretending to be openssl but majorily screwing it up
[11:00:56 CET] <JEEB> his patch IIRC seemed different
[11:01:11 CET] <JEEB> sfan5: I just think not many people are able to / want to test it
[11:01:14 CET] <nevcairiel> i know, but that doesnt change the feelings toward libressl
[11:01:17 CET] <JEEB> sure
[11:01:35 CET] <nevcairiel> if they wanted to be different, they should've started with that, not half-assed drop-in support for openssl
[11:01:36 CET] <JEEB> I'd just say that the bigger reason is that most people don't have libressl around
[11:02:19 CET] <sfan5> pretty sure the long-term plan is to eventually drop libssl and have just libtls
[11:02:28 CET] <sfan5> the current state is a mess though, I agree with that
[11:03:27 CET] <JEEB> and they kind of pissed people off with the initial compatibility and then boom-boom with the API I guess. but since there are actual users of libressl the patch makes sense
[11:03:37 CET] <JEEB> just not going to be testing on my non-BSD :)
[11:06:47 CET] <BtbN> nevcairiel, I think libtls is different?
[11:06:58 CET] <BtbN> At least it does not collide with openssl library names anymore
[11:07:08 CET] <BtbN> No idea if the symbols collide, but why would you link both anyway
[11:07:27 CET] <sfan5> symbols can't collide because libtls is currently just a wrapper around libssl
[11:07:51 CET] <BtbN> That means they definitely do collide with openssl.
[11:10:09 CET] <sfan5> you're not supposed to install LibreSSL and OpenSSL alongside anyway, it's a "feature"
[11:16:19 CET] <BtbN> too bad that's highly unrealistic
[11:31:21 CET] <durandal_1707> nice,  breaking playing of olf files because of wrong security fix
[11:31:32 CET] <durandal_1707> *old
[13:50:13 CET] <cone-960> ffmpeg 03Carl Eugen Hoyos 07master:15c632719a5f: configure: libvmaf is not GPLv2-compatible.
[15:05:46 CET] <wbs> fwiw, about openssl vs libressl, one way to reduce the amount of ifdefs but still support it, would be something like #ifdef libressl_version #undef openssl_version #define openssl_version 1.0.0 #endif
[15:05:50 CET] <wbs> at the top of the file
[15:07:16 CET] <BtbN> That only works until they decide to break something else with the API
[15:07:20 CET] <wbs> true
[15:07:36 CET] <BtbN> They are explicitly not supporting OpenSSL while claiming to be OpenSSL, which is just a mess.
[15:07:48 CET] <BtbN> Also, with libtls there is a much nicer and cleaner library available from them.
[15:07:51 CET] <wbs> but if their intent with the libssl part is to keep it compatible with openssl 1.0.0 api while doing new stuff in libtls, it should be doable
[15:10:26 CET] <BtbN> I just don't get why they report a version that indicated openssl 1.1 or even higher support, instead of just keeping that define on 1.0 level, and having a new one for the libressl version
[15:10:31 CET] <BtbN> This way it just messes with people
[15:10:44 CET] <wbs> they report openssl api 2.0.0
[15:11:04 CET] <wbs> the thing that actually messed things up for them, was when openssl released 1.1 which deprecated things
[15:11:05 CET] <BtbN> which is >=1.1
[15:11:20 CET] <BtbN> It should have stayed right at 1.0
[15:11:25 CET] <wbs> yes, that I can agree with
[15:11:35 CET] <wbs> but the actual break happened when openssl itself invented 1.1 with a lot of dropped api
[15:23:59 CET] <rcombs> they probably should've had an intermediate version that deprecated (but didn't remove) stuff
[20:33:38 CET] <cone-960> ffmpeg 03Vittorio Giovara 07master:969f215957b2: hevc: Add support for alternative transfer characterics SEI
[20:33:39 CET] <cone-960> ffmpeg 03Martin Storsjö 07master:61f589e31e84: lavf: Remove codec_tag from dashenc and smoothstreamingenc
[20:33:40 CET] <cone-960> ffmpeg 03Martin Storsjö 07master:1912973a2d0a: d3d11va: Check WINAPI_FAMILY instead of HAVE_LOADLIBRARY
[20:33:41 CET] <cone-960> ffmpeg 03Martin Storsjö 07master:4d330da006fe: os_support: Use HAVE_UWP instead of manually checking WINAPI_FAMILY
[20:33:42 CET] <cone-960> ffmpeg 03James Almer 07master:514dda60703a: Merge commit '4d330da006fe48178a4c8047f06270925eaedf63'
[20:51:36 CET] <cone-960> ffmpeg 03James Almer 07master:ce52e430b2df: compat/w32dlfcn: add missing config.h include
[21:08:50 CET] <atomnuker> jamrial: is the v3 of the gamma side data patch okay?
[21:10:30 CET] <jamrial> atomnuker: the patch adding code to png is. but the patch adding the actual side data type has a comment by nicolas
[21:10:44 CET] <jamrial> which may affect the png patch
[21:12:11 CET] <atomnuker> oh, right
[22:20:13 CET] <dff_> how can help me with rtsp?
[22:20:23 CET] <dff_> who*
[22:27:49 CET] <microchip_> dff_: support is in #ffmpeg
[22:28:21 CET] <dff_> is not devlopers here?
[22:28:33 CET] <dff_> i need developer scripts stream rtsp
[22:29:16 CET] <microchip_> dff_: read the topic. this is only for ffmpeg development
[22:40:32 CET] <cone-960> ffmpeg 03Tomas Härdin 07master:89e632de9b3a: Add Cinepak encoder
[22:40:33 CET] <cone-960> ffmpeg 03addr-see-the-website at aetey.se 07master:7da9f4523159: cinepakenc: fixes and improvements
[22:40:34 CET] <cone-960> ffmpeg 03addr-see-the-website at aetey.se 07master:158d2e15454a: cinepakenc: add option handling for flexibility
[22:40:35 CET] <cone-960> ffmpeg 03Diego Biurrun 07master:b40236d954c3: cinepakenc: Drop broken debug code
[22:40:36 CET] <cone-960> ffmpeg 03Timothy Gu 07master:1e12730e5df2: cinepakenc: Stop using AVPicture
[22:40:37 CET] <cone-960> ffmpeg 03Diego Biurrun 07master:47cde2eabc1a: cinepakenc: K&R formatting cosmetics
[22:40:38 CET] <cone-960> ffmpeg 03Diego Biurrun 07master:ed3b2a999d18: cinepakenc: misc small changes
[22:40:39 CET] <cone-960> ffmpeg 03Diego Biurrun 07master:5be9939b46a6: fate: Add cinepak encoder vsynth tests
[22:40:40 CET] <cone-960> ffmpeg 03James Almer 07master:537127432901: Merge commit 'ed3b2a999d189771c9b0ada9a9200117595fa474'
[22:40:41 CET] <cone-960> ffmpeg 03James Almer 07master:0ff8f0b7b4d7: Merge commit '5be9939b46a6a4d3860a1b1a872aa50897095970'
[22:50:19 CET] <cone-960> ffmpeg 03Diego Biurrun 07master:fd92dafaff88: bink: Split read_dct_coeffs()
[22:50:20 CET] <cone-960> ffmpeg 03James Almer 07master:17fa37cf413a: Merge commit 'fd92dafaff8844b5fedf94679b93d953939a7f7b'
[22:55:01 CET] <cone-960> ffmpeg 03Luca Barbato 07master:79f64f7ebc2b: bitstream: Avoid undefined behavior in bitstream_skip()
[22:55:02 CET] <cone-960> ffmpeg 03Alexandra Hájková 07master:9c1e111406bd: flac: Convert to the new bitstream reader
[22:55:03 CET] <cone-960> ffmpeg 03James Almer 07master:e5841b22ca83: Merge commit '9c1e111406bd4cbb04d8e8611b71bebf203bec5e'
[23:03:13 CET] <cone-960> ffmpeg 03Luca Barbato 07master:15f1cc09a406: flac: Postpone unlikely condition checks
[23:03:14 CET] <cone-960> ffmpeg 03James Almer 07master:ff55b62a6547: Merge commit '15f1cc09a406cf6296818d475a256902235eefc4'
[23:12:29 CET] <cone-960> ffmpeg 03Luca Barbato 07master:0f5ad12ba2b5: flac: Use a local cache for decode_residual()
[23:12:30 CET] <cone-960> ffmpeg 03Diego Biurrun 07master:825e463a170c: build: Add feature test macros for glibc 2.19+
[23:12:31 CET] <cone-960> ffmpeg 03James Almer 07master:ec703910af66: Merge commit '0f5ad12ba2b538cb329c507ecc914e06bfa70194'
[23:12:32 CET] <cone-960> ffmpeg 03James Almer 07master:85b9fe958414: Merge commit '825e463a170c7004c63030dc484b2b2de869227b'
[23:18:24 CET] <cone-960> ffmpeg 03Sean McGovern 07master:95ce02b35d3d: rmdec: don't ignore the return value of av_get_packet()
[23:18:25 CET] <cone-960> ffmpeg 03James Almer 07master:fe421d4deecd: Merge commit '95ce02b35d3d1bb16111031df1d82a6e4d894d36'
[23:55:05 CET] <ubitux> heh, a patch from Aurelien Jacobs
[23:55:09 CET] <ubitux> that's a surprise
[23:58:10 CET] <ubitux> patchset* actually
[23:59:05 CET] <atomnuker> yeah, 6 years since his last patch
[00:00:00 CET] --- Wed Nov  8 2017