[Ffmpeg-devel-irc] ffmpeg-devel.log.20170408

[burek]

[00:21:34 CEST] <ubitux> jamrial: niice :)
[00:21:57 CEST] <ubitux> jamrial: honest thanks, you're helping a lot
[00:28:11 CEST] <jamrial> ubitux: you're welcome :)
[00:28:15 CEST] <jamrial> the faster we get rid of the backlog the better
[01:17:16 CEST] <michaelni> atomnuker, can you take a look at CID 1400585 (opusenc) i suspect its a false positive but you know that code much better
[01:26:19 CEST] <atomnuker> michaelni: yeah, it looks like a false positive, even though that bit of code is 1:1 copied from the decoder
[01:27:15 CEST] <michaelni> ok, please mark it as false positive then
[01:30:36 CEST] <atomnuker> done
[01:57:33 CEST] <cone-368> ffmpeg 03Rostislav Pehlivanov 07master:db788de8ecde: mdct15: use existing complex multiplication macros
[01:57:33 CEST] <cone-368> ffmpeg 03Rostislav Pehlivanov 07master:6d0b62a12b46: opus_celt: fix some style issues
[01:57:33 CEST] <cone-368> ffmpeg 03Rostislav Pehlivanov 07master:a66121d96455: opus_rc: add macros for checkpoint rollback
[01:57:33 CEST] <cone-368> ffmpeg 03Rostislav Pehlivanov 07master:63744d8afd57: opusenc: remove unused header entries and simplify normalization
[01:57:33 CEST] <cone-368> ffmpeg 03Rostislav Pehlivanov 07master:42ffc67208b1: opusenc: minor style changes
[01:57:34 CEST] <cone-368> ffmpeg 03Rostislav Pehlivanov 07master:c39fba703bf3: opusenc: do not signal digital silence
[01:57:34 CEST] <cone-368> ffmpeg 03Rostislav Pehlivanov 07master:d0a3143193ff: opusenc: switch between intra/inter mode for coarse energy
[04:56:38 CEST] <philipl_> If you were holding it correctly, they wouldn't need to tell you how to do it.
[04:56:41 CEST] <philipl_> bah. sorry
[06:16:35 CEST] <cone-198> ffmpeg 03James Almer 07master:b438a7868c8a: ffmpeg: use av_stream_new_side_data() in init_output_stream
[13:52:40 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:d4f2a681cb6d: configure: MMAL-related decoders should depend on, not select, mmal
[13:52:40 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:09027c89777e: Merge commit 'd4f2a681cb6d4e913f1c879ad376313255d6cf4c'
[14:09:16 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:9254344e11f9: configure: Move hardware-accelerated codec deps out of hwaccel section
[14:09:17 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:8e0bea55158e: Merge commit '9254344e11f9b016088ec6250724f74377f5d7a0'
[14:11:54 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:8b56dbe7435d: configure: Do not add newlines in filter()/filter_out() functions
[14:11:55 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:f98ccbd6a06d: Merge commit '8b56dbe7435d8cfe3964f447fc45fe98db5d9042'
[14:15:07 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:c21d78a903da: configure: Integrate X11 checks into vaapi/vdpau checks
[14:15:08 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:8aaacbf7fd7e: Merge commit 'c21d78a903da378af48c9e5f971fe989592ddb65'
[14:17:43 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:0983f9117f31: metasound: Drop unused tables
[14:17:44 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:2d40fbe63985: Merge commit '0983f9117f31521643162cb85380672495a9de1b'
[14:18:17 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:5c89022542ce: hevc: Drop pointless av_unused attribute
[14:18:18 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:6e836052d8c8: Merge commit '5c89022542ce8521d89ef58858342a7bc1c3cd0d'
[14:21:17 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:e4382a4ab481: hevc: Eliminate pointless variable indirection
[14:21:19 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:f5fdf355af29: Merge commit 'e4382a4ab48138d43a19ea0da96f536a5e49b50c'
[14:25:25 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:c4c5f5386c83: vp9dsp: add DC only versions for idct/idct.
[14:25:26 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:6db8248a592e: Merge commit 'c4c5f5386c83bb8d66f8d67cd8533c8697f06d04'
[14:28:04 CEST] <cone-479> ffmpeg 03Martin Storsjö 07master:3c87039a404c: arm: vp9itxfm: Only reload the idct coeffs for the iadst_idct combination
[14:28:05 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:35f62a3b40b7: Merge commit '3c87039a404c5659ae9bf7454a04e186532eb40b'
[14:28:42 CEST] <cone-479> ffmpeg 03Martin Storsjö 07master:9c8bc74c2b40: arm: vp9itxfm: Skip empty slices in the first pass of idct_idct 16x16 and 32x32
[14:28:43 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:a30f5fd428ae: Merge commit '9c8bc74c2b40537b0997f646c87c008042d788c2'
[14:29:23 CEST] <cone-479> ffmpeg 03Martin Storsjö 07master:cad42fadcd2c: aarch64: vp9itxfm: Skip empty slices in the first pass of idct_idct 16x16 and 32x32
[14:29:24 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:0768aaec1d68: Merge commit 'cad42fadcd2c2ae1b3676bb398844a1f521a2d7b'
[14:32:50 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:0a35f128f3c6: cabac: x86: Give optimizations header a more meaningful name
[14:32:51 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:5be1440c74d0: Merge commit '0a35f128f3c6e0ae9a0a2236c557602c108da269'
[14:43:26 CEST] <cone-479> ffmpeg 03Vittorio Giovara 07master:ff9db5cfd145: lavc: Use a stricter check for the color properties values
[14:43:27 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:883fe17e3bef: Merge commit 'ff9db5cfd14558df9cfcc54d6c062bc34bf1f341'
[14:49:57 CEST] <cone-479> ffmpeg 03Vittorio Giovara 07master:5168026a0525: options_table: Do not rely on enum size as option bound
[14:49:58 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:97745b131a7f: Merge commit '5168026a05258537f1c48ca95c1776b1554997f4'
[14:50:36 CEST] <cone-479> ffmpeg 03Vittorio Giovara 07master:706af9227b58: lavu: Document the color properties enumeration values origin
[14:50:37 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:68c9a399420b: Merge commit '706af9227b58657c73e3a4df3689da734f010500'
[14:51:39 CEST] <cone-479> ffmpeg 03Aleksandr Slobodeniuk 07master:a91f1023bc06: examples: fix a typo in an error message
[14:51:40 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:3f06dfcdc6e9: Merge commit 'a91f1023bc06091ef84dce0f1e12b72d7f3ba3ca'
[14:52:20 CEST] <cone-479> ffmpeg 03Luca Barbato 07master:518a22471e40: configure: Be more helpful when requiring options set
[14:52:21 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:efb0ea2af4a8: Merge commit '518a22471e40604a158180a0660aabeeec65e07f'
[14:53:09 CEST] <ubitux> wm4: does d82d5379caca21005d8906829b35361c4a65408e look correct to you?
[14:53:23 CEST] <ubitux> it appears at first glance that this init can be called several times
[14:53:43 CEST] <wm4> looks correct to me
[14:53:49 CEST] <wm4> it's called on a fresh allocation
[14:57:06 CEST] <ubitux> ok
[14:57:21 CEST] <cone-479> ffmpeg 03Wan-Teh Chang 07master:d82d5379caca: mmaldec: initialize refcount using atomic_init()
[14:57:22 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:b853d6aa5b91: Merge commit 'd82d5379caca21005d8906829b35361c4a65408e'
[14:58:22 CEST] <cone-479> ffmpeg 03James Almer 07master:33a2b73b9837: mpeg4audio: correctly propagate meaningful error values
[14:58:23 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:ec10c543b8b5: Merge commit '33a2b73b98374de4781ae0497cf74b2ce07a9615'
[15:01:13 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:984e50e74ba7: build: Add -Wpedantic and -Wextra to extra warning flags
[15:01:14 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:3d0475ef387f: Merge commit '984e50e74ba71301720652e56232161a994bd0b5'
[15:12:36 CEST] Action: ubitux doesn't understand 2575b1917b2703fae41a4b4901ad85c888374dfd
[15:14:37 CEST] <wm4> seems about passing special cflags to checkheaders runs
[15:14:54 CEST] <Compn> to disable warnings
[15:14:59 CEST] <wm4> though I don't really understand the purpose either
[15:15:05 CEST] <wm4> seems pointless
[15:19:18 CEST] <Compn> maybe deigo has some scripts to monitor warnings
[15:19:41 CEST] <Compn> hes worked to minimize warnings for 10+ years
[15:25:55 CEST] <ubitux> i'll just merge
[15:34:57 CEST] <Compn> if you are looking for an OK, i'd say OK.
[15:35:14 CEST] <cone-479> ffmpeg 03Diego Biurrun 07master:2575b1917b27: build: Add separate flags for checkheaders targets
[15:35:15 CEST] <cone-479> ffmpeg 03Clément BSsch 07master:d14a1bd8e470: Merge commit '2575b1917b2703fae41a4b4901ad85c888374dfd'
[15:41:03 CEST] <ubitux> afk for a while, feel free to commits or merges
[15:41:44 CEST] <ubitux> 469 left, beginning of dec 2016
[16:57:54 CEST] <thresh> hi everyone
[16:58:12 CEST] <thresh> I plan to move git.videolan.org to a new server, because the old is obviously having troubles coping with the load
[16:58:31 CEST] <thresh> that means the devs will be presented with a new ssh host key when the dns propagates
[16:58:55 CEST] <thresh> and possibly some minor issues with hooks etc (although I'm now manually checking them)
[17:19:39 CEST] <ubitux> thresh: any downtime planned?
[17:25:01 CEST] <thresh> ubitux, only for the devs, I think.
[17:25:18 CEST] <thresh> I plan to disable ssh access for a few hours, until DNS is surely propagated everywhere
[17:25:29 CEST] <ubitux> when?
[17:27:03 CEST] <J_Darnley> Oh no.  Clearly a sign we should be using Github!  </sarcasm>
[17:27:14 CEST] <thresh> how does 7 PM UTC until the next morning sound for you?
[17:27:22 CEST] Action: kierank slaps J_Darnley
[17:27:35 CEST] Action: kierank slaps J_Darnley around a bit with a large ocean perch
[17:27:40 CEST] <kierank> hmm that command still works
[17:28:19 CEST] <ubitux> thresh: sounds a bit early but it's ok with me
[17:28:47 CEST] <thresh> ubitux, I'm in UTC+3 sooo...
[17:29:45 CEST] <thresh> I can maybe move it to 9/10 PM, but don't expect me to react fast :)
[17:30:07 CEST] <thresh> read-only access via git/http will still work though
[17:41:46 CEST] <thresh> michaelni, ^^^
[18:01:13 CEST] <jamrial> thresh: have you looked at the weird bug in git web i told you about some time ago? of how a given file's history shows every merge commit even if they didn't actually affect said file?
[18:01:23 CEST] <jamrial> see for example https://git.videolan.org/?p=ffmpeg.git;a=history;f=version.sh;h=edc4dd33c54207e7b7796c745a6f55234dc67068;hb=HEAD
[18:01:41 CEST] <jamrial> while a normal git log version.sh locally shows the actual history
[18:02:23 CEST] <nevcairiel> its not a "bug", its kind of intentional
[18:02:38 CEST] <jamrial> regardless, it's extremely annoying :p
[18:02:42 CEST] <nevcairiel> indeed
[18:02:51 CEST] <nevcairiel> i modified my local gitweb install to remvoe that flag
[18:03:45 CEST] <nevcairiel> somewhere in the gitweb.cgi is the log command it uses, and it passes --full-history to that
[18:03:50 CEST] <nevcairiel> that includes all unrelated merges
[18:05:15 CEST] <thresh> I welcome diffs :)
[18:17:40 CEST] <nevcairiel> thresh: https://pastebin.com/rAdp4iRa :D
[18:20:34 CEST] <thresh> jamrial, https://git2.videolan.org/?p=ffmpeg.git;a=history;f=version.sh;h=edc4dd33c54207e7b7796c745a6f55234dc67068;hb=HEAD looks better right ?
[18:20:39 CEST] <thresh> nevcairiel, thanks!
[18:21:06 CEST] <nevcairiel> definitely better
[18:21:30 CEST] <jamrial> thresh: absolutely :D
[18:22:21 CEST] <thresh> sweet
[18:22:40 CEST] <JEEB> gg
[18:38:27 CEST] <michaelni> thresh, as far as iam concerned, do the move whenever its most convenient to you, i assume everything will still work fine and the same as now, afterwards 
[18:40:22 CEST] <thresh> I sure hope it will;  ping me on IRC or mail if you see something odd
[20:46:17 CEST] <cone-479> ffmpeg 03James Almer 07master:a8b5f3750109: avformat/matroskaenc: don't try to update flac extradata if live streaming
[21:16:04 CEST] <thresh> fyi I've moved the DNS entries to point to a new machine.  SSHFP are accessible via `dig @ns0.videolan.org SSHFP git.videolan.org`
[21:21:10 CEST] <thresh> you can actually push to a new one.  it might a minor confusion for some not to get the links to open via http or something, but we can live wit hthat.
[21:23:04 CEST] <thresh> and if you happen to use DSS keys, well, you're screwed
[21:23:24 CEST] <BtbN> I don't understand what changed...?
[21:23:37 CEST] <BtbN> new ssh keys?
[21:23:45 CEST] <nevcairiel> moved to a new server
[21:23:48 CEST] <nevcairiel> so new ssh fingerprint
[21:23:53 CEST] <thresh> BtbN, I've moved git.videolan.org to a new server; technically, you should only receive a warning from SSH when trying to push
[21:24:13 CEST] <nevcairiel> (also people using DSA/DSS keys derve to be screwed)
[21:24:15 CEST] <thresh> new enough ssh clients will actually check SSHFP records
[21:24:30 CEST] <nevcairiel> *deserve
[21:24:47 CEST] <thresh> a side effect of moving to a new server is moving to a newer openssh-server; which forbids DSA keys
[21:25:17 CEST] Action: ubitux uses source.ffmpeg.org and can not access it anymore
[21:26:06 CEST] <thresh> ubitux, flush your dns cache maybe
[21:26:20 CEST] <thresh> I'm all ears
[21:26:42 CEST] <ubitux> what ip am i supposed to get?
[21:26:50 CEST] <ubitux> ah that's good now
[21:26:53 CEST] <thresh> 88.191.250.5 / 2a01:e0d:1:3:58bf:fa02:c0de:c01d
[21:27:45 CEST] <ubitux> new host detected but no read access (with ssh)
[21:28:23 CEST] <thresh> let's PM
[21:30:42 CEST] <durandal_1707>  no pm please
[21:33:43 CEST] <thresh> it iss still resolving to a wrong address for ubitux
[21:34:30 CEST] <ubitux> yep, that's working now
[21:34:48 CEST] <ubitux> thresh: write accesses are disabled until tomorrow?
[21:35:01 CEST] <thresh> no, it should work for a new server
[21:35:07 CEST] <ubitux> ok, cool
[21:35:09 CEST] <ubitux> thanks
[21:35:26 CEST] <thresh> please tell me if you encounter weird push logs (like failing hooks etc)
[21:35:44 CEST] <thresh> I've manually checked them and they did work for me (sent emails, IRC notifications), but...
[21:36:08 CEST] <ubitux> i could try a merge 
[21:36:43 CEST] <ubitux> but the next one is slightly sensible and need brain, which i don't seem to have currently so i'll pass
[21:37:09 CEST] <thresh> no problem.  PM / mail me if anything.
[21:37:36 CEST] <thresh> this new server has SSD and lots of RAM, so it'll be miles better than the old one from 2006.
[21:38:32 CEST] <durandal_1707> SSD means it will die next month
[21:38:44 CEST] <thresh> but those are enterprise SSD
[21:38:49 CEST] <thresh> (so next year)
[21:39:17 CEST] <durandal_1707> how much ram?
[21:39:40 CEST] <thresh> only 128G
[21:56:09 CEST] <durandal_1707> can someone review my 2 speed improvement patches?
[21:57:21 CEST] <atomnuker> durandal_1707: which ones?
[21:58:08 CEST] <durandal_1707> hqx and cllvc frame threading
[21:58:21 CEST] <durandal_1707> didnt you get them?
[22:32:14 CEST] <cone-479> ffmpeg 03Thierry Foucu 07release/3.3:c9c977be2761: libavformat/mov: Fix memory leak, need to free the chapter tracks array
[22:32:18 CEST] <cone-479> ffmpeg 03Michael Niedermayer 07release/3.3:1833ec53349c: avcodec/h264: Check weight values to be within the specs limits.
[22:32:21 CEST] <cone-479> ffmpeg 03Michael Niedermayer 07release/3.3:a430ba99251e: avcodec/dvdsubdec: Fixes 2 runtime error: left shift of 170 by 24 places cannot be represented in type 'int'
[22:32:24 CEST] <cone-479> ffmpeg 03Michael Niedermayer 07release/3.3:f4400a92f58f: avformat/mov: Check creation_time for overflow
[22:54:58 CEST] <cone-479> ffmpeg 03James Almer 07release/3.3:37fcf089b4a2: avformat/matroskaenc: don't try to update flac extradata if live streaming
[23:27:19 CEST] <atomnuker> michaelni: is it okay to apply the mjpeg patches with ljpeg optimized huffman disabled for now?
[23:28:08 CEST] <atomnuker> I can fix ljpeg later once I have time to see what needs to be done with LJpegEncContext to support those
[23:41:11 CEST] <jamrial> atomnuker: if you can't fix the valgrind failures before 3.3 is tagged, then just revert the whole thing in that branch
[23:45:33 CEST] <atomnuker> jamrial: I did fix them, one of the commits there does that
[23:46:02 CEST] <jamrial> ah, good
[00:00:00 CEST] --- Sun Apr  9 2017