[Ffmpeg-devel-irc] ffmpeg.log.20170709

[burek]

[01:48:07 CEST] <hiihiii> hello
[01:50:27 CEST] <hiihiii> running this command to stream to youtube
[01:50:54 CEST] <hiihiii> ffmpeg -re -i input.mp4 -threads 0 -c copy -f mp4 rtmp://a.rtmp.youtube.com/live2/key-key-key-key
[01:52:26 CEST] <hiihiii> gives me the following error  https://thepasteb.in/p/Vmh0QNBm0gNs4
[01:53:25 CEST] <hiihiii> I'm getting it's due to seeking but I'm not sure how to enable it
[01:53:45 CEST] <hiihiii> with ftp it's : -ftp-write-seekable 1
[01:53:58 CEST] <hiihiii> with rtmp there's no option for it
[01:54:13 CEST] <furq> -f flv
[01:54:21 CEST] <nicolas17> "streaming" and "seekable" are fundamentally contradictory :P
[01:55:10 CEST] <hiihiii> wouldn't it encode the input?
[01:55:17 CEST] <nicolas17> no because you said -c copy
[01:55:17 CEST] <furq> no
[01:55:27 CEST] <nicolas17> I wonder why you are streaming a static video file though... why not just... upload it to youtube?
[01:55:27 CEST] <hiihiii> or just muxing itno flv
[01:55:41 CEST] <hiihiii> just testing
[01:55:47 CEST] <nicolas17> it was already demuxing the mp4 and remuxing into mp4 anyway
[01:55:59 CEST] <nicolas17> now it will demux the mp4 and remux into flv
[01:56:11 CEST] <nicolas17> in either case it won't reencode the actual video stream because you used -c copy
[01:56:46 CEST] <hiihiii> thanks. yes it makes sense but why doesn't it work with -f mp4
[01:57:16 CEST] <nicolas17> I think when you output to an mp4 file, it has to rewrite the header at the beginning of the file, when it *finishes* encoding
[01:57:59 CEST] <hiihiii> oh ok
[01:58:20 CEST] <nicolas17> if you output to an mp4 live stream, it can't go back to the beginning of the "file" and write some data when it finishes :P
[01:58:31 CEST] <hiihiii> I noticed that when I output FLVs, even if the encoding hasn't finished I can open the output
[01:58:36 CEST] <nicolas17> exactly
[01:58:43 CEST] <hiihiii> with MP4 it gives an error
[01:59:17 CEST] <hiihiii> okay thx that makes much sense
[02:40:11 CEST] <thebombzen> hiihiii: mp4 is not a format that you can record to a nonseekable device, like a stream
[02:40:55 CEST] <thebombzen> mp4 files have a table of contents, which by default is written at the end of the file after you've written the rest, but that means anyone reading it has to be able to seek to the end to have it, which isn't possible if you're streaming it
[02:41:37 CEST] <thebombzen> when you're done writing the file, you can move the table of contents (called a moov atom) to the beginning, but again, doesn't work with streaming.
[03:02:31 CEST] <hiihiii> thebombzen: I read that, thx
[03:13:26 CEST] <Anthony_> hello
[03:13:40 CEST] <Anthony_> having problems with ffmpeg
[03:14:01 CEST] <Anthony_> something about a missing AVFilterGraph
[04:05:02 CEST] <behael> 'Ello guys
[04:05:23 CEST] <behael> I need some help to understand a weird behavior of ffmpeg
[04:11:55 CEST] <dystopia_> what weird behavior
[04:14:48 CEST] <behael> god, cant use the pastebin https://pastebin.com/post.php  ERR_SPDY_PROTOCOL_ERROR
[04:14:53 CEST] <behael> sec, let me try again
[04:17:11 CEST] <behael> I managed to paste in another site, let me know if thats allowed to show here
[04:17:30 CEST] <dystopia_> just paste the link
[04:17:43 CEST] <behael> http://pasted.co/6fa9bb2d
[04:18:11 CEST] <behael> if you look at the paste - at my phys host ffprobe returns mpeg video for the "sample.m4v"
[04:18:43 CEST] <behael> but inside the lxc container, it returns GSM
[04:19:03 CEST] <behael> same ffmpeg, same ffprobe, same debian - diff is one is phys the other is a lxc container
[04:20:01 CEST] <behael> at the phys host I am able to convert from m4v to mp4 - but inside the container it gens a huge mp4 (about 50MB) with weird sounds and tons of "missing gsm magic" errors
[04:36:22 CEST] <behael> the "file" cmd returns (host or LXC) the same result: sample.m4v: MPEG sequence, v4, video, advanced simple @ L1
[04:45:12 CEST] <furq> pastebin the command and full output
[04:45:20 CEST] <nicolas17> he did
[04:45:29 CEST] <nicolas17> http://pasted.co/6fa9bb2d
[04:45:59 CEST] <furq> there's no command or output there
[04:46:42 CEST] <nicolas17> ...
[04:46:59 CEST] <nicolas17> "ffprobe sample.m4v" is the command, and it gives different output on the same file
[04:47:10 CEST] <furq> 03:20:01 ( behael) at the phys host I am able to convert from m4v to mp4 - but inside the container it gens a huge mp4 (about 50MB) with weird sounds and tons of "missing gsm magic" errors
[04:47:14 CEST] <furq> i meant this command
[04:47:27 CEST] <nicolas17> not sure if it matters?
[04:47:34 CEST] <nicolas17> if ffprobe alone gives different results on the input file
[04:47:43 CEST] <furq> well it might at least give us something to go off
[04:48:03 CEST] <nicolas17> behael: are you *sure* the .m4v is the same in both systems? run sha1sum on both
[04:48:06 CEST] <furq> yeah
[04:48:37 CEST] <furq> if they're the same then i can only imagine there's some major issue with your container
[04:57:27 CEST] <FurretUber> Hi, I am trying to encode a video with 10-bit depth with libx264. I tried to capture screen and tried to encode other videos. There is a warning about incompatible pixel format and the resulting videos loses a lot of its colours, here is the output of ffmpeg when capturing the screen: https://pastebin.com/di2yj11x I have built from the ffmpeg snapshot of 05-07-2017
[04:59:10 CEST] <behael> the cmd for converting ? The default one ffmpeg -i sample.m4v sample.mp4
[04:59:16 CEST] <behael> and the hashs matchs
[05:00:29 CEST] <nicolas17> FurretUber: I don't know why the x264 encoder refuses to use 10bit because I never dealt with 10bit myself
[05:00:54 CEST] <nicolas17> but I really wonder if the x11grab input stream will support 10 bit...
[05:01:04 CEST] <FurretUber> What I am actually using is 8-bit and produces, according to https://trac.ffmpeg.org/wiki/Encode/H.264#LosslessH.264 is lossless, is https://pastebin.com/wSWwg5fw
[05:02:08 CEST] <nicolas17> the input stream (x11grab) is bgr0, pretty sure that's 8 bit
[05:02:28 CEST] <nicolas17> so if you manage to get x264 to accept 10 bit... you'll be upscaling from 8 bit to 10 bit, which seems somewhat pointless
[05:04:43 CEST] <FurretUber> I believe this is true, but from what I have searched, the resulting 10-bit video would be smaller. Unless the Google bias to my search history is showing me lies
[05:05:43 CEST] <nicolas17> heh, I wonder if someone compared same crf with 8 bit and 10 bit to come to that conclusion :P
[05:05:43 CEST] <furq> FurretUber: you need a separate libx264 with 10-bit support
[05:06:01 CEST] <nicolas17> (the x264 crf has a different range in 10 bit)
[05:06:13 CEST] <thebombzen_> FurretUber it depends on how it's deepened, but yes
[05:06:41 CEST] <FurretUber> crf 0 would still be lossless with 10-bit libx264?
[05:06:48 CEST] <furq> -qp 0 is lossless
[05:07:03 CEST] <FurretUber> Is https://trac.ffmpeg.org/wiki/Encode/H.264#LosslessH.264 wrong?
[05:07:14 CEST] <nicolas17> define "lossless"
[05:07:18 CEST] <furq> no
[05:07:23 CEST] <thebombzen_> You can get lossless with -crf 0, but you should just set -qp 0 in that case
[05:07:24 CEST] <furq> -crf 0 is lossless with 8-bit x264
[05:07:27 CEST] <furq> -qp 0 is lossless with both
[05:07:36 CEST] <thebombzen_> Really? Huh
[05:08:02 CEST] <nicolas17> if you have RGB input and you convert it to 4:2:0 YUV, it won't be "lossless" whether it's 8bit or 10bit ;)
[05:08:06 CEST] <furq> also yeah
[05:08:13 CEST] <furq> i doubt you'll get a noticeable compressibility difference
[05:08:30 CEST] <furq> and the colours will be equally wrong either way
[05:08:32 CEST] <thebombzen_> nicolas17, true, but irrelevant
[05:08:39 CEST] <FurretUber> -crf 0 -qp 0 will work for both?
[05:08:52 CEST] <thebombzen_> Use -qp 0 for lossless
[05:09:25 CEST] <thebombzen_> Note that if your version of libx264 is 10bit, then libx264rgb will not work correctly.
[05:09:56 CEST] <thebombzen_> Or might not. In this case it's better to convert to yuv444p10le and encode that
[05:11:01 CEST] <thebombzen_> I'm not sure, but I think yuv444p is slightly more compressible that gbrp anyway. It might work better with a non-lavc decoder as well
[05:11:32 CEST] <nicolas17> gbrp?
[05:11:44 CEST] <thebombzen_> Green Blue Red Planar
[05:12:05 CEST] <FurretUber> My inputs are, generally, RGB. I was trying to make the encoded videos both 10-bit and don't lose the colors as they done with that command
[05:12:17 CEST] <thebombzen_> it's rgb24 but green/blue/red, and planar rather than interleaved
[05:13:10 CEST] <FurretUber> I could make it have the rgb colors correctly, but couldn't make any 10-bit video yet.
[05:13:13 CEST] <thebombzen_> Upping 8 to 10 has basically no benefit in lossless mode.
[05:13:37 CEST] <FurretUber> I have to go now, tomorrow I will build libx264 with support and will try again. Thank you
[05:14:12 CEST] <thebombzen_> Upping 8 to 10 can be more efficient because it allows more room for low-bit quantization error
[05:14:18 CEST] <thebombzen_> err, and gone.
[05:15:16 CEST] <thebombzen_> nicolas17: gbrp is planar RGB sorted by green, blue, red
[05:16:11 CEST] <thebombzen_> gbr is the order for parallel with yuv. Y is mostly based on the green channel. U and V are blue and red chroma respectively
[05:39:13 CEST] <behael> I really cant fix it... I just moved to a nfs share and a few scripts to do a "remote convert from m4v to mp4"
[05:39:33 CEST] <behael> I tried diff containers, with diff versions of debian - all showing the same behavior
[05:40:36 CEST] <behael> since the files are small and the host is the host of the containers, its good for now - but I will investigate it
[05:40:46 CEST] <behael> thanks guys
[05:41:07 CEST] <nicolas17> weird as hell
[05:42:03 CEST] <behael> Yeah I never saw it, file systems are ok no fault drive or anything
[05:42:23 CEST] <behael> I also tried to host a ubuntu container, centos container all the same behavior
[05:43:08 CEST] <furq> i guess you want strace and diff
[05:43:15 CEST] <furq> also fyi m4v is mp4, you can probably just rename it
[05:43:33 CEST] <behael> probably some lxc underground conf
[05:43:51 CEST] <nicolas17> memory_corruption=off
[05:43:53 CEST] <nicolas17> (?)
[05:44:15 CEST] <behael> yeah I tried, but the m4v are generated by the guacenc (guacamole encoder) from RDP session recordings. And they need to be played over html5
[05:44:21 CEST] <furq> oh
[05:44:24 CEST] <furq> yeah you'll need to convert it then
[05:44:28 CEST] <behael> So, html5 is refusing to play m4v or mp4 (renamed)
[05:44:41 CEST] <furq> yeah it's mpeg4, not h264
[05:45:09 CEST] <furq> but yeah ffprobe inside the container is somehow misdetecting the container and the codec
[05:45:24 CEST] <furq> and if the checksums match then something is seriously wrong there
[05:45:54 CEST] <furq> especially if it's happening with multiple different storage types
[05:46:46 CEST] <behael> yupe, I cant say the old "it used to work" b/c I never did it - so I will try a few diff confs tomorow - if I find something I will let you guys know
[05:46:59 CEST] <behael> thanks guys - almost 2am here time to hit the bed
[05:47:03 CEST] <behael> cya around
[05:48:06 CEST] <nicolas17> I'm removing the C++ wrappers I made around ffmpeg
[05:48:19 CEST] <nicolas17> so I can get a plain C program again and post it here and get help on why the hell it's misbehaving
[05:48:23 CEST] <nicolas17> -_-
[18:28:22 CEST] <lindylex> I am trying to concat two videos and I get a longer video than the sum of all the videos and I only see the first video.  This is my command :  ffmpeg -f concat -i l.txt -i cornerFinalBlack.png -codec:v libx264 -profile:v baseline -preset slow -pix_fmt yuv420p -b:v 3500k -threads 0 -an -y out.mp4   This is a link to all my assets http://mo-de.net/d/fh/
[18:31:06 CEST] <c_14> the input videos are different fps
[18:31:29 CEST] <c_14> so ffmpeg assumes the first video's fps for the other video
[18:31:50 CEST] <lindylex> Ok match the fps and that will solve it?
[18:32:01 CEST] <c_14> presumably
[18:32:40 CEST] <c_14> compressed/full range might also cause issues but not length related
[18:34:39 CEST] <lindylex> I do not understand you last sentence.
[18:35:11 CEST] <lindylex> One sec let me set the frame rate now and test.
[18:36:20 CEST] <c_14> the first video is yuv420p and the second is yuvj420p which means that the first video has mpeg (compressed) range and the second jpeg (full) range
[18:36:26 CEST] <c_14> this may cause bleaching of colors etc
[18:36:42 CEST] <lindylex> I see.
[18:37:13 CEST] <lindylex> So it is combination of framerate and compressed range?
[18:37:44 CEST] <c_14> the length issue is probably just the framerate, the range should only cause color bleaching etc
[18:37:57 CEST] <lindylex> One sec let em test.
[18:42:40 CEST] <lindylex> I am try to make the image into a video setting the frame rate to 29.97 like this:  ffmpeg -loop 1 -i cornerFinalBlack.png -framerate 30000/1001 -c:v libx264 -t 5 -pix_fmt yuv420p c1.mp4   I get a video with 25 fps instead
[18:42:57 CEST] <c_14> -framerate before -i
[18:43:03 CEST] <lindylex> ok
[18:43:42 CEST] <lindylex> Thanks that worked!
[18:44:55 CEST] <lindylex> Ok I got the duration to make sense.  Now I need to have the second video show up.
[18:48:14 CEST] <lindylex> I did this:  ffmpeg -loop 1 -framerate 30000/1001 -i cornerFinalBlack.png -c:v libx264 -t 5 -pix_fmt yuvj420p c1.mp4  to create the first video the second video does not display.
[18:53:27 CEST] <c_14> try adding -profile:v baseline
[18:58:38 CEST] <zyclonicz> How do I add a thumbnail in the beginning of my webm?
[19:06:32 CEST] <lindylex> c_14: the out.mp4 http://mo-de.net/d/fh/  is so strange.  I looks like it is skipping frames but it does show up.
[19:07:03 CEST] <lindylex> c_14: this is what I used to create the first video :  ffmpeg -loop 1 -framerate 30000/1001 -i cornerFinalBlack.png -profile:v baseline -c:v libx264 -t 5 -pix_fmt yuvj420p c1.mp4
[19:09:10 CEST] <The_8472> zyclonicz: mkvpropedit foo.webm --add-attachment "thumb.jpg"
[19:11:01 CEST] <The_8472> although that might be a non-compliant webm
[19:11:12 CEST] <The_8472> mkv is a superset
[19:11:39 CEST] <nicolas17> wouldn't webm players ignore any mkv thing they don't support?
[19:11:54 CEST] <The_8472> yeah
[19:12:15 CEST] <The_8472> but it also means they won't show the thumbnail he wants
[19:17:48 CEST] <atomnuker> I just insert 1-frame video packets to serve as cover art
[19:18:09 CEST] <atomnuker> any proper player will correctly display the 1 frame and still continue to play the audio
[19:18:12 CEST] <The_8472> that might be annoying when they loop. you would need VFR to make it zero-duration
[19:18:34 CEST] <atomnuker> mpv doesn't loop
[19:18:56 CEST] <atomnuker> it would only loop once it reaches the last stream's packet
[19:19:31 CEST] <The_8472> just saying that solution is not universal. he asked about webm, so i assume it's for... well... web use.
[19:23:00 CEST] <atomnuker> all web browsers handle it correctly
[19:23:16 CEST] <atomnuker> and I really do mean it, the only player that won't handle it correctly is vlc
[19:23:25 CEST] <atomnuker> which doesn't handle many things correctly like colorspace
[19:23:56 CEST] <atomnuker> (its because vlc nih their demuxers rather than use libavformat)
[19:24:12 CEST] <atomnuker> so its pretty universally supported
[19:27:01 CEST] <The_8472> I mean... what do they do when you loop it? do they flash that one frame?
[19:27:28 CEST] <The_8472> that's why I meant you need to make it 0-duration
[19:31:32 CEST] <lindylex> c_14: the solution I have to convert them like this ffmpeg -i c2.mov -c copy -bsf:v h264_mp4toannexb -f mpegts i2.ts and concat the .ts files.
[19:33:06 CEST] <atomnuker> The_8472: no, you don't get it
[19:33:43 CEST] <atomnuker> the decode video, they present it, they run out of packets of video to present so they don't decode it again, they don't draw it again, they do nothing to the video surface
[19:34:30 CEST] <atomnuker> they just leave the surface as is once they decode the last video frame, so it sticks
[19:34:57 CEST] <The_8472> that only works if you want to play audio with cover art. not have a video with a specific thumbnail
[19:35:12 CEST] <atomnuker> what do you mean, a specific thumbnail?
[19:35:27 CEST] <The_8472> websites thumbnail videos
[19:35:48 CEST] <atomnuker> yes, they take the first frame
[19:35:54 CEST] <atomnuker> so they take the cover art
[19:36:02 CEST] <The_8472> well, if you want to customize it you need to do something else
[19:36:10 CEST] <The_8472> cover art is not the same thing
[19:36:22 CEST] <The_8472> cover art is dedicated metadata for that sole purpose
[19:37:02 CEST] <atomnuker> nothing on the internet supports cover arts muxed in webms
[19:37:11 CEST] <atomnuker> so that's why the hack exists
[19:38:12 CEST] <The_8472> yes, but that's cover art. not a video thumbnail. different things.
[19:38:49 CEST] <nicolas17> cover art is for music
[19:40:05 CEST] <atomnuker> it'll appear as a thumbnail if muxed this way
[21:24:16 CEST] <lilibox> hi, i would like to make conversion and archiving some files grabbed to .mpg from DV source, could anybody recomende me, please, which format and which setting is optimal these days? I would like to have standard 2k resolution as output, decent color correction and maybe deinterlacing, here are input format parameters
[21:24:21 CEST] <lilibox> https://pastebin.com/zsnEznEA
[21:24:57 CEST] <lilibox> pretty thank you for any advice points me to success :)
[21:33:53 CEST] <DHE> well that's not 2k content. why upscale it?
[21:35:16 CEST] <lilibox> upscale is no needed then
[21:47:40 CEST] <kantlivelong> is there some issue with recording from pulse device? i get a crazy amount of stuttering
[21:47:59 CEST] <kantlivelong> seems okay with some devices but my 48k optical its throwing up wiht
[21:49:46 CEST] <nicolas17> okay!
[21:51:58 CEST] <nicolas17> https://paste.kde.org/pt8uuzaey here's my source code
[21:52:26 CEST] <nicolas17> https://paste.kde.org/pbmytvmit the output video is much larger than what I get with ffmpeg.c despite using the same x264 crf
[21:54:09 CEST] <nicolas17> clearly I'm doing something wrong with the ffmpeg API to get such different output...
[21:59:09 CEST] <nicolas17> omg I think I just figured it out, if I make outCodecCtx->time_base and outputVideoStream->time_base the same, the file size becomes correct
[22:00:53 CEST] <nicolas17> but why? I'm not even using bitrate, I'm using crf
[22:38:09 CEST] <thebombzen> nicolas17: timebase refers to the timestamps of frames, not the bitrate
[22:38:32 CEST] <thebombzen> my guess is that libavformat is duplicating or dropping frames if your timebase is incorrect
[22:38:42 CEST] <thebombzen> causing the actual amount of encoded data to be different
[22:40:51 CEST] <nicolas17> thebombzen: as you can see in the paste, the size of *each* frame is bigger
[22:41:16 CEST] <nicolas17> pkt_pts=0 pict_type=I pkt_size=30895 with ffmpeg.c
[22:41:29 CEST] <nicolas17> pkt_pts=0 pict_type=I pkt_size=80392 with my program not setting that time_base properly
[22:42:04 CEST] <thebombzen> That's not a surprise, because x264 is receiving different data and has quality-based ratecontrol
[22:42:54 CEST] <thebombzen> with dups, bframes work better with accurate references so using mroe data on an iframe makes sense
[22:43:52 CEST] <nicolas17> both videos have the same number of frames and packets (confirmed with ffprobe -show_frames -show_packets)
[22:44:15 CEST] <nicolas17> only their sizes differ
[22:46:17 CEST] <thebombzen> possibly relevant is the difference between AVC1 and Annex-B formatting
[22:46:38 CEST] <thebombzen> er, AVCC and Annex-B
[22:47:16 CEST] <nicolas17> yeah, I can't figure that one out either... how do I make my code output avcc like ffmpeg.c is? and why does it matter? :/
[22:47:18 CEST] <thebombzen> also notice the set framerate is wrong in your (incorrect) program encode
[22:47:25 CEST] <thebombzen> it says 30.15
[22:47:38 CEST] <thebombzen> the solution is something you already figured out though. which is to make sure the timebase is correct
[22:48:24 CEST] <nicolas17> which of the three timebases? there's AVStream.time_base, AVCodecContext.time_base for the encoder I create, and AVStream.codec.time_base which is deprecated but still causes yet another difference if I tweak it
[22:49:17 CEST] <nicolas17> I get correct fps in ffprobe if I set the AVStream.codec.time_base to {1,30}
[22:50:21 CEST] <nicolas17> if I set all three timebases to {1,15360}, ffprobe reports 30.30fps
[22:54:46 CEST] <thebombzen> I would recommend against changing the deprecated one
[22:54:54 CEST] <thebombzen> but otherwise, I don't know
[22:56:36 CEST] <nicolas17> the extradata difference also makes me think I'm doing something different from ffmpeg.c, but despite lots of source code grepping I can't figure out what it is
[23:01:44 CEST] <CoJaBo> VP9 apparently confuses the hell out of me :/
[23:02:32 CEST] <CoJaBo> Does the 2-pass encoding actually do anything (and if so, what) when using -crf without a bitrate?
[23:03:14 CEST] <CoJaBo> Also, -speed -8 is supposed to be "slowest possible encoding"; why is it encoding faster than x264 veryfast? >_>
[23:03:17 CEST] <thebombzen> CoJaBo: when using crf, 2-pass shouldn't do anything
[23:03:20 CEST] <iive> 2 pass encoding should not do anything without target size
[23:03:56 CEST] <thebombzen> 2-pass more accurately fits into target size. crf just uses "bits as needed" to get the right quality. doing that again doesn't help
[23:09:20 CEST] <CoJaBo> thebombzen: That's what I'd thought.. a LOT of examples/guides I'm finding are listing it tho.. can someone suggest one that isn't stupid
[23:11:14 CEST] <CoJaBo> (What I'm trying to do, BTW, is batch reencode a ton of videos; trying to optimize for smallest possible size with roughly-equiv quality, encoding time does not matter)
[23:13:09 CEST] <CoJaBo> And current command line I'm trying, if anyone cares: -i in.wmv -c:v libvpx-vp9 -crf 50 -b:v 0 -speed -8 -c:a libopus -b:a 32000 out.webm
[23:19:35 CEST] <The_8472> I think crf50 would result in pretty poor quality.
[23:26:52 CEST] <Ingersol> Hello.
[23:27:56 CEST] <Ingersol> what is right way to read packets from the media file - fread to buffer or something else?
[23:28:28 CEST] <Ingersol> i got different examples of different versions of ffmpeg so a bit confused
[23:37:49 CEST] <nicolas17> Ingersol: there are versions of ffmpeg where the API changed and the examples/documentation still used the old deprecated API, so you should look at the latest version
[23:38:00 CEST] <nicolas17> Ingersol: are you reading from a file?
[23:44:49 CEST] <CoJaBo> The_8472: I'm trying to find a CRF approximation for MovieMaker's default ("barf") quality setting.
[23:53:43 CEST] <BtbN> barf? Sounds like something bad
[23:59:08 CEST] <Ingersol> nicolas17, yep
[00:00:00 CEST] --- Mon Jul 10 2017