[Ffmpeg-devel-irc] ffmpeg-devel.log.20160516

burek burek021 at gmail.com
Tue May 17 02:05:03 CEST 2016


[00:10:29 CEST] <durandal_1707> the only one virtual reality is hologram
[00:12:47 CEST] <durandal_170> anybody against pushing loudnorm?
[00:18:22 CEST] <atomnuker> durandal_170: yes, I'd still rather see libebur128 integrated rather than being liked
[00:18:29 CEST] <atomnuker> *linked
[00:19:32 CEST] <atomnuker> but I guess if it really starts to bother me I'd just do the work myself sometimes later and integrate it
[00:20:00 CEST] <atomnuker> so yeah, might be okay for the time being
[00:21:26 CEST] <atomnuker> at least unless someone else doesn't like it (ubitux said something about adding options to another filter)
[00:23:30 CEST] <cbsrobot> j-b: can you unban me on #videolan ?
[00:24:11 CEST] <cbsrobot> I guess I was banned because I was flooding the channel with join attempts
[00:31:41 CEST] <durandal_170> no, you were spamming
[00:42:02 CEST] <wm4> michaelni: thanks, confirmed fixed
[00:50:45 CEST] <cbsrobot> durandal_170: My ISP had a technical problem
[01:06:56 CEST] Action: michaelni tries to login to his apple hw after not seeing it on fate.ffmpeg anymore ... then looks in disbelive on the network cable that seemed so well attached before slightly pulling on it
[10:01:00 CEST] <j-b> cbsrobot: I can
[10:25:37 CEST] <cbsrobot> j-b: thanks
[11:45:03 CEST] <Compn> cbsrobot speaks!
[11:45:54 CEST] <cbsrobot> Compn: not all robots are inferior to humans
[13:53:14 CEST] <haasn> wm4: JEEB: https://0x0.st/qac.txt is this configure check needed? how old is the SMPTEST2084 entry?
[13:55:30 CEST] <haasn> hmm seems it was added in https://github.com/FFmpeg/FFmpeg/commit/c3cd6dd106b1381933e2f24898eeec0d8aa17746 on 28 aug 2015
[13:55:37 CEST] <haasn> is that new enough for ffmpeg/libav stable?
[13:55:50 CEST] <haasn> s/new/old/
[13:56:51 CEST] Action: haasn asks CI
[14:01:45 CEST] <haasn> yeah libav stable definitely doesn't have it
[14:11:13 CEST] <Daemon404> nevcairiel, ping re: h2645
[15:25:32 CEST] <cone-327> ffmpeg 03Michael Niedermayer 07master:feeb3a926163: swresample/resample: Fix division by 0 with tap_count=1
[15:25:32 CEST] <cone-327> ffmpeg 03Michael Niedermayer 07master:6085d6b2aeef: ffmpeg: Check that r_frame_rate is set before attempting to use it
[15:25:33 CEST] <cone-327> ffmpeg 03Michael Niedermayer 07master:635b2ec5f20d: avformat/utils: Do not compute the bitrate from duration == 0
[15:40:06 CEST] <cone-327> ffmpeg 03Paul B Mahol 07master:8767470fb572: avcodec/iff: check if there is enough bytes left
[16:26:22 CEST] <D`K_C> I have a question on how to use libswresample: I change the output sample rate using av_opt_set_int, and then call swr_init again as the doc says, but it looks like it's leaking memory when I do that, am I missing something?
[16:30:27 CEST] <DHE> you might find swr_alloc_set_opts easier. and yes you are making a new instance so the old one leaks
[16:30:48 CEST] <DHE> and technically API questions belong in #ffmpeg as well
[16:32:12 CEST] <D`K_C> oups, my bad for using the wrong chan
[16:58:22 CEST] <durandal_1707> stilll merging going?
[18:00:47 CEST] <durandal_170> ubitux: hey, you there?
[18:06:41 CEST] <ubitux> durandal_170: somehow
[18:10:12 CEST] <durandal_170> ubitux: so you against pushing loudnorm filter as it is now?
[18:11:04 CEST] <ubitux> i think that's a bad idea but i'm too busy & lazy to do anything about it 
[18:11:09 CEST] <pfelt> afternoon all.  can i ask a question about this discussion on adding AVClass that's going on on the mailing list?  I was wondering what this issue is.  i'm thinking it's related to comments like this in header files: "&.. New public fields should be added right above."  and the ABI?  just looking for some clarification on reasons, not the debate itself as i'm not sure what the reasons are.
[18:14:37 CEST] <pfelt> i suppose my question really boils down to a misunderstanding of abi, what it is and why ti's important to not change
[18:15:47 CEST] <durandal_170> because evil is out there
[18:16:53 CEST] <Daemon404> [15:58] <@durandal_1707> stilll merging going? <-- i am waiting on nevcairiel to help me with one commit.
[19:10:56 CEST] <Daemon404> Shiz, were you the one who wanted to be poked when i pushed my bindiff thing?
[19:11:18 CEST] <Shiz> ya
[19:11:22 CEST] <Shiz> well, one of the people at least
[19:11:32 CEST] <Daemon404> https://github.com/dwbuiten/bxd
[19:11:37 CEST] <Daemon404> enjoy some bad code and missing features
[19:11:52 CEST] <Shiz> nice
[19:11:54 CEST] <Shiz> looks good
[19:12:01 CEST] <Shiz> if i didn't enjoy bad code what i would be doing on freenode
[19:12:24 CEST] <Daemon404> laff
[19:14:10 CEST] <wm4> I guess you wouldn't be programming at all
[19:20:18 CEST] <durandal_170> Why you are using nut?
[19:22:12 CEST] <kierank> piping with timestamps I guess
[19:30:06 CEST] <JEEB> Daemon404: niice
[19:30:17 CEST] <JEEB> durandal_170: nut with raw video/audio is pretty common
[19:31:03 CEST] <JEEB> Daemon404: not golang?
[19:31:28 CEST] <Daemon404> didnt feel like it
[19:31:36 CEST] <JEEB> k
[19:31:41 CEST] <Daemon404> the only benefit of goland here is windows support
[19:31:45 CEST] <Daemon404> golang*
[19:32:03 CEST] <Shiz> and random panics
[19:32:26 CEST] <Daemon404> ive never had an issue with panics
[19:33:01 CEST] <Shiz> all the go software i've used was panicky as hell in the silliest circumstances
[19:33:16 CEST] <Shiz> like feeding it a JSON file that doesn't conform to its format, bam panic and stack trace
[19:33:30 CEST] <Shiz> or influxdb which would panic the entire DB if you did something like 'select (field) from table'
[19:33:37 CEST] <Shiz> (the parens killed it)
[19:33:59 CEST] <JEEB> panics are generally really rare and if something like JSON parsing fucks something up someone has fucked up with his code
[19:34:20 CEST] <Shiz> :p
[19:34:21 CEST] <Daemon404> [18:33] < Shiz> like feeding it a JSON file that doesn't conform to its format, bam panic and stack trace <-- theres no way normal could should do this
[19:34:34 CEST] <JEEB> Daemon404: people use riced JSON parsers
[19:34:35 CEST] <Daemon404> but... web people.
[19:34:36 CEST] <Shiz> well, tell terraform that
[19:34:59 CEST] <Daemon404> i have literally never written the 'panic' keyword in ym code in 3-4 years
[19:35:00 CEST] <Daemon404> so.
[19:35:17 CEST] <JEEB> why would you be writing it :P
[19:35:24 CEST] <JEEB> it's literally all's fucked up kind of case
[19:35:32 CEST] <Daemon404> some people can write java in any language, JEEB 
[19:35:41 CEST] <JEEB> (´4@)
[19:35:42 CEST] <Shiz> :3
[19:36:37 CEST] <Mavrik> Would writing Java in LISP be a special kind of insult?
[19:38:05 CEST] <Shiz> only if you're one of Those People
[22:39:47 CEST] <pfelt> would something like "make -j 40 fate" presumably not cause issues with the tests?
[22:40:02 CEST] <Compn> it shouldnt
[22:40:13 CEST] <Compn> might kill your disk / cpu
[22:40:47 CEST] <pfelt> i haven't tried it wondering if "did that error occur because of the -j or because of something i did in code"  kind of a situation
[22:41:02 CEST] <pfelt> that's ok.  my box is currently only dev/build/test
[22:46:40 CEST] <Shiz> lol
[22:46:49 CEST] <JEEB> uhh
[22:46:52 CEST] <Shiz> that ps2 fomrat looks just like psmf except with pcm in the ps1 packets
[22:47:04 CEST] <JEEB> yeah, wouldn't be surprised
[22:47:27 CEST] <JEEB> there's been raw PCM on the PSP too
[22:47:43 CEST] <JEEB> I think marcan extracted some delicious wakamoto GPS voices from a thing I bought
[22:47:44 CEST] <Shiz> inb4 part of the custom header
[22:47:46 CEST] <Shiz> D:
[22:49:50 CEST] <JEEB> meanwhile, just as we thought there weren't enough colour spaces: http://www.dolby.com/us/en/technologies/dolby-vision/ICtCp-white-paper.pdf
[22:52:51 CEST] Action: pfelt wishes he understood any of that paper
[22:54:22 CEST] <cone-653> ffmpeg 03Paul B Mahol 07master:72e33eec3a5e: avcodec/iff: add support for ANIMs with keyframes and concatenated ANIMs
[00:00:00 CEST] --- Tue May 17 2016


More information about the Ffmpeg-devel-irc mailing list