[Ffmpeg-devel-irc] ffmpeg-devel.log.20171004

[burek]

[00:46:34 CEST] <jkqxz> tmm1:  I have an N3700, which fits between your J1800 and N3450 in generations (though Windows 7 rather than 10).  I've never seen any problems on it, but I will try with your sample.
[00:47:06 CEST] <tmm1> i was trying with -f lavfi -i color=black:640x480 and could not reproduce, nor with the test.avi that was on the ticket before
[00:47:13 CEST] <tmm1> but the sample i uploaded triggers it 
[00:47:20 CEST] <tmm1> definitely curious if you can repro
[00:49:23 CEST] <jkqxz> I tried with some cases provided by Zeranoe a while ago, but could never get it.
[01:07:29 CEST] <cone-950> ffmpeg 03Martin Vignali 07master:cbbec68847ed: libavcodec/blockdsp : add AVX version
[01:13:34 CEST] <cone-950> ffmpeg 03John Stebbins 07master:248dc5c1646d: h264dec: fix dropped initial SEI recovery point
[01:13:35 CEST] <cone-950> ffmpeg 03John Stebbins 07master:cd7a2e1502f1: asfdec: fix reading files larger than 2GB
[01:13:36 CEST] <cone-950> ffmpeg 03James Almer 07master:73c85d3af242: Merge commit '248dc5c1646dcdd96fe79761105c4ae889e711fd'
[01:13:37 CEST] <cone-950> ffmpeg 03James Almer 07master:1902bbbe7225: Merge commit 'cd7a2e1502f174c725c0de82711d2c7649057574'
[01:26:06 CEST] <cone-950> ffmpeg 03Lou Logan 07master:d17a00379e7a: doc: Add mailing list FAQ
[01:30:32 CEST] <cone-950> ffmpeg 03Anton Khirnov 07master:b2788fe9347c: svq3: fix the slice size check
[01:30:33 CEST] <cone-950> ffmpeg 03James Almer 07master:cb222d73225a: Merge commit 'b2788fe9347c02b1355574f3d28d60bfe1250ea7'
[01:32:43 CEST] <cone-950> ffmpeg 03James Almer 07master:3f258f5ee05c: apetag: fix flag value to signal footer presence
[01:32:44 CEST] <cone-950> ffmpeg 03James Almer 07master:4cc0227040ad: apetag: account for header size if present when returning the start position
[01:32:45 CEST] <cone-950> ffmpeg 03James Almer 07master:9266cc54aa5c: Merge commit '4cc0227040adb9efc63be6a5765e3214f5c6f662'
[01:35:55 CEST] <tmm1> jkqxz: just repro'd on a Win10 i5-6600k as well
[01:57:00 CEST] <cone-950> ffmpeg 03Anton Khirnov 07master:984736dd9e5b: lavc: make sure not to return EAGAIN from codecs
[01:57:01 CEST] <cone-950> ffmpeg 03James Almer 07master:5678b2a82074: Merge commit '984736dd9e5b50987a5910e22495304e4a6d975c'
[01:58:40 CEST] <cone-950> ffmpeg 03Diego Biurrun 07master:da8093f712d6: fate: Use bitexact optimizations in the svq3-2 test
[01:58:41 CEST] <cone-950> ffmpeg 03James Almer 07master:93d27e022fb1: Merge commit 'da8093f712d625db7ce4a2526fb52994e01921ec'
[02:19:18 CEST] <cone-950> ffmpeg 03Zhong Li 07master:fd7cb86468cc: mpegdec: fix redundant dummy frames issue of interlaced clips
[02:19:19 CEST] <cone-950> ffmpeg 03Michael Niedermayer 07master:dcf9bae4a93f: avcodec/aacdec_template: Clear tns present flag on error
[02:19:20 CEST] <cone-950> ffmpeg 03Michael Niedermayer 07master:44874b4f5ec2: avcodec/truemotion2: Fix integer overflows in tm2_high_chroma()
[02:19:21 CEST] <cone-950> ffmpeg 03Michael Niedermayer 07master:d662143f0646: avcodec/dxv: Check for end of input in dxv_decompress_dxt5()
[02:19:22 CEST] <cone-950> ffmpeg 03Ingo Brückl 07master:08c751309670: avformat/mp3dec: Fix definition of MIDDLE_BITS
[02:30:27 CEST] <cone-950> ffmpeg 03Diego Biurrun 07master:4141a5a240fb: Use modern avconv syntax for codec selection in documentation and tests
[02:30:28 CEST] <cone-950> ffmpeg 03James Almer 07master:a2b0602cdad8: Merge commit '4141a5a240fba44b4b4a1c488c279d7dd8a11ec7'
[02:59:14 CEST] <cone-950> ffmpeg 03Luca Barbato 07master:150e99d694f3: rtsp: Factor out packet reading
[02:59:15 CEST] <cone-950> ffmpeg 03Luca Barbato 07master:b9b82151a1aa: rtsp: Move the pollfd setup out of the for loop
[02:59:16 CEST] <cone-950> ffmpeg 03Luca Barbato 07master:5263f464db5f: rtsp: Lazily allocate the pollfd array
[02:59:17 CEST] <cone-950> ffmpeg 03James Almer 07master:c244d4af43a1: Merge commit '150e99d694f33ab9ad678834964909aa315d14a1'
[02:59:18 CEST] <cone-950> ffmpeg 03James Almer 07master:7c74efeaf891: Merge commit 'b9b82151a1aaa8bbf389853a6142c4e101d80b86'
[02:59:19 CEST] <cone-950> ffmpeg 03James Almer 07master:c2916564d8e4: Merge commit '5263f464db5f2df74ddf712f6d1221b24475fa8e'
[03:13:19 CEST] <cone-950> ffmpeg 03Diego Biurrun 07master:c483398bb7ef: build: Drop DOC_ prefix from EXAMPLES-related variables
[03:13:20 CEST] <cone-950> ffmpeg 03James Almer 07master:749b35799cc5: Merge commit 'c483398bb7ef66f61ed2dcb09f3d6160683da0eb'
[04:00:13 CEST] <cone-950> ffmpeg 03Diego Biurrun 07master:698ac8f9cabd: fate: Make null comparison method more useful
[04:00:14 CEST] <cone-950> ffmpeg 03James Almer 07master:e1f01ce8d20b: Merge commit '698ac8f9cabd053f2c19346a77b92f8eae4218fc'
[04:04:53 CEST] <cone-950> ffmpeg 03Ben Chang 07master:d8f36a6aa33e: nvenc: Fix the preset mapping list
[04:04:54 CEST] <cone-950> ffmpeg 03James Almer 07master:85fb45586804: Merge commit 'd8f36a6aa33e9f904fa47caa0329ddaac391cd7d'
[04:14:29 CEST] <memeka> hi, how can i change codec preferences?
[04:15:11 CEST] <memeka> e.g. h264 is used, even if h264_vdpau is available
[04:20:16 CEST] <cone-950> ffmpeg 03Luca Barbato 07master:79331df362fb: rtsp: Lazily set up the pollfd array once
[04:20:17 CEST] <cone-950> ffmpeg 03James Almer 07master:1e7b6e47d275: Merge commit '79331df362fb05a0d04ca9489c87e5b80077a3f4'
[04:41:05 CEST] <cone-950> ffmpeg 03Diego Biurrun 07master:e1a6d63c7eef: fate: Rename WMV8_DRM decoder tests to WMV3_DRM
[04:41:06 CEST] <cone-950> ffmpeg 03James Almer 07master:9e009ee312ad: Merge commit 'e1a6d63c7eeff2f0ec8173546357bfaa9deecea4'
[04:53:23 CEST] <philipl> google for: "vk_khr_external_memory" cuda
[04:53:37 CEST] <philipl> get 8 hits
[04:53:37 CEST] <philipl> what is wrong with them?
[06:05:33 CEST] <cone-950> ffmpeg 03Diego Biurrun 07master:21cca00dfeae: build: Explicitly disable external libraries when not explicitly enabled
[06:05:33 CEST] <cone-950> ffmpeg 03Michael Niedermayer 07master:d7b2bb5391bf: h264_sei: Check actual presence of picture timing SEI message
[06:05:34 CEST] <cone-950> ffmpeg 03James Almer 07master:63b9fd75db40: Merge commit '21cca00dfeaec08ca93cf94ed33f4311cf1d8c84'
[06:05:35 CEST] <cone-950> ffmpeg 03James Almer 07master:ade77f9c1395: Merge commit 'd7b2bb5391bf55e8f9421bff7feb4c1fddfac4bf'
[06:15:04 CEST] <cone-950> ffmpeg 03Diego Biurrun 07master:d154bdd3d053: configure: Simplify dlopen check
[06:15:05 CEST] <cone-950> ffmpeg 03James Almer 07master:84b3f53acadc: Merge commit 'd154bdd3d053128c908a994bb26e14bbc17f0e53'
[08:33:29 CEST] <wm4> philipl: wut?
[08:34:05 CEST] <wm4> BtbN: it checked the avcodeccontext class instead the avcodec one, broke our cuvid wrapper
[08:34:36 CEST] <wm4> so, we could just drop the cuvid AVHWAccels
[08:34:43 CEST] <wm4> and modify ff_get_format
[08:35:13 CEST] <wm4> nothing else required (except maybe an internal AVCodec cap to skip the check in ff_get_format)
[09:44:32 CEST] <ubitux> wm4: "While the AVFrame is within libavcodec
[09:44:39 CEST] <ubitux> you meant in libavutil, right?
[09:44:59 CEST] <wm4> no, libavcodec
[09:45:25 CEST] <wm4> as in while the AVCodecContext code holds a reference to an AVFrame instance
[09:45:49 CEST] <wm4> the owner of a reference gets to decide what opaque_ref means
[09:46:04 CEST] <ubitux> oh ok, my bad
[10:17:18 CEST] <durandal_1707> kode54: was atrac9 re-ed or its only shim over another lib?
[10:19:17 CEST] <JEEB> :o
[10:19:33 CEST] <JEEB> which reminds me I should totally finish work on the MPEG-PS mappings for that since a lot of samples are in MPEG-PS
[10:19:50 CEST] <JEEB> mappings/parser
[10:27:22 CEST] <JEEB> can I request a push on http://ffmpeg.org/pipermail/ffmpeg-devel/2017-September/217102.html ?
[10:32:11 CEST] <wm4> JEEB: LGTM
[10:37:28 CEST] <ldts> wm4: I have a fix for  https://trac.ffmpeg.org/ticket/6716; should I close the ticket after having validated the fix or after the fix is merged?
[10:38:04 CEST] <wm4> I don't know/care about the bug tracker
[10:38:04 CEST] <JEEB> I'd close it after commenting on it that it's validated, or if there are no replies
[10:38:19 CEST] <JEEB> as in, comment that the reporter can test it now
[10:38:21 CEST] <wm4> it's pretty much /dev/null to me
[10:38:31 CEST] <wm4> (too much noise in any case)
[10:38:49 CEST] <wm4> I think we usually reference the ticket number in the commit message
[10:38:50 CEST] <ldts> wm4: I see JEEB: yes I tested it with the poster
[10:38:58 CEST] <JEEB> ok, then it's validated?
[10:39:02 CEST] <ldts> yes
[10:39:04 CEST] <JEEB> close with a comment towards the hash
[10:39:06 CEST] <ldts> he tested on his board
[10:39:09 CEST] <JEEB> which fixes it on master
[10:39:33 CEST] <ldts> ok so will send the patch first and then wait to be merged and then closed
[10:40:15 CEST] <ldts> JEEB: yes this all happened in the #kodi channel this morning so I shared the patches with him
[10:40:32 CEST] <JEEB> yup
[12:59:54 CEST] <cone-950> ffmpeg 03Clément BSsch 07master:8be3fc88d3fa: build: make h264 VT encoder select the VT encoder dependency
[13:18:42 CEST] <cone-950> ffmpeg 03Jan Ekström 07master:ffc58b2ce27e: movenc: take packet dts shifting into mention in check_pkt
[13:19:10 CEST] <ubitux> (^ patch applied after irc request)
[13:19:54 CEST] <wm4> doesn't JEEB have push rights
[13:20:21 CEST] <JEEB> wm4: I haven't gotten to doing the key thing yet
[13:20:36 CEST] <wm4> ah
[13:20:39 CEST] <JEEB> and I'm not at home/machine that I'm doing FFmpeg dev at
[13:58:31 CEST] <JEEB> whew
[13:58:40 CEST] <JEEB> happy that FATE failures are not due to me :P
[13:59:04 CEST] <JEEB> `src/libavfilter/x86/af_afir.asm:22: fatal: unable to open include file `libavutil/x86/x86util.asm'`
[13:59:08 CEST] <JEEB> on FreeBSD things
[16:47:53 CEST] <philipl> wm4: Just my frustration that nvidia won't talk about cuda vulkan interop
[16:48:38 CEST] <philipl> It clearly doesn't exist: vk_khr_external_memory explicitly enumerates the external types it understands and those are dx and gl right now
[16:49:17 CEST] <nevcairiel> nvidias primary market for cuda is cuda and cuda and more cuda, why would they care? =p
[16:49:39 CEST] <BtbN> Because they have interop with everything else?
[16:50:53 CEST] <nevcairiel> thats not a reason
[16:51:28 CEST] <wm4> philipl: ah, but that's no nvidia specific issue?
[16:51:38 CEST] <wm4> or do other vendors provide interop?
[16:51:46 CEST] <wm4> I guess for vaapi interop is a non-issue
[16:51:51 CEST] <wm4> (DMA buf stuff)
[16:52:52 CEST] <wm4> so I guess what it'd need is some sort of cuda extension that exports raw memory buffers
[16:53:02 CEST] <philipl> cuvid is useless with vulkan without it
[16:53:30 CEST] <nevcairiel> i still doubt the big use for vulkan in video processing
[20:49:29 CEST] <wm4> öl
[20:49:31 CEST] <wm4> lol
[20:49:35 CEST] <wm4> that linked in message
[20:49:48 CEST] <wm4> oh, it's to ffmpeg-security
[20:49:51 CEST] <wm4> that's even more fun
[20:50:16 CEST] <wm4> (it's basically badly disguised spam apparently from linked in directly)
[20:53:19 CEST] <ubitux> wm4: ffmpeg-security is subject to a lot of spam
[20:53:46 CEST] <ubitux> you may not see it all, but since i don't have a spam filter (yet) i can tell
[20:54:05 CEST] <ubitux> i'd say one out of 20 is spam on that ml
[20:54:51 CEST] <wm4> I get a lot of spam on it, yes
[20:55:08 CEST] <wm4> probably more than real mail
[22:25:47 CEST] <kode54> durandal_1707: I think it was re-ed, but the source is still being documented
[22:50:01 CEST] <cone-895> ffmpeg 03Marton Balint 07master:cbf09f23b655: MAINTAINERS: change the decklink maintainer to myself
[22:50:01 CEST] <cone-895> ffmpeg 03Marton Balint 07master:3dc01223ef52: avdevice/decklink_dec: fix multipacket op47 decoding
[22:50:02 CEST] <cone-895> ffmpeg 03Devin Heitmueller 07master:b6782a192e18: avdevice/decklink: Fix segfault when running -list_devices on OSX
[00:00:00 CEST] --- Thu Oct  5 2017