michaelni changed the topic of #ffmpeg-devel to: Welcome to the FFmpeg development channel | Questions about using FFmpeg or developing with libav* libs should be asked in #ffmpeg | This channel is publicly logged | FFmpeg 7.1.1 has been released! | Please read ffmpeg.org/developer.html#Code-of-conduct
System_Error has quit [Remote host closed the connection]
System_Error has joined #ffmpeg-devel
Traneptora has quit [Quit: Quit]
jarthur has quit [Ping timeout: 260 seconds]
jarthur has joined #ffmpeg-devel
Kimapr has joined #ffmpeg-devel
Kimapr_ has quit [Read error: Connection reset by peer]
<BBB>
re the comments on ML, I created an account on forgejo, how do I work with this? can I set notifications for specific files I'd like to review that I consider myself knowledgeable on? (or maintainer of, if you like that terminology)
<BBB>
and do I need permissions to be able to approve/merge patches?
<kasper93>
I don't think there is currently a way to do notification per file
<kasper93>
This could be possible with a bot that would assign reviewers to patches based on MAINTAINERS, but nothing like that exists atm
<kasper93>
for permissions ask BtbN
jamrial has quit []
jarthur has quit [Quit: jarthur]
SuperFashi has quit [Quit: No Ping reply in 180 seconds.]
jarthur has joined #ffmpeg-devel
SuperFashi has joined #ffmpeg-devel
Yalda has quit [Remote host closed the connection]
<fjlogger>
[FFmpeg/FFmpeg] Issue edited: #20073 Same videos (frame-to-frame wise) written by different applications produce different PSNR/SSIM/VMAF scores. (https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20073) by DimkaTsv
<fjlogger>
[FFmpeg/FFmpeg] Issue edited: #20073 Same videos (frame-to-frame wise) written by different applications produce different PSNR/SSIM/VMAF scores. (https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20073) by DimkaTsv
<fjlogger>
[FFmpeg/FFmpeg] Issue edited: #20073 Same videos (frame-to-frame wise) written by different applications produce different PSNR/SSIM/VMAF scores. (https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20073) by DimkaTsv
<fjlogger>
[FFmpeg/FFmpeg] Issue edited: #20073 Same videos (frame-to-frame wise) written by different applications produce different PSNR/SSIM/VMAF scores. (https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20073) by DimkaTsv
<fjlogger>
[FFmpeg/FFmpeg] Pull request opened: #20084 hwcontext_vulkan: transfer exclusive images to correct queue family in drm->vulkan mapping (https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20084) by vytskalt
<fjlogger>
[FFmpeg/FFmpeg] Pull request merged: #20084 hwcontext_vulkan: transfer exclusive images to correct queue family in drm->vulkan mapping (https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20084) by Lynne
<fjlogger>
[FFmpeg/FFmpeg] Issue edited: #20073 Same videos (frame-to-frame wise) written by different applications produce different PSNR/SSIM/VMAF scores. (https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20073) by DimkaTsv
jarthur has quit [Quit: jarthur]
<fjlogger>
[FFmpeg/FFmpeg] Issue edited: #20073 Same videos (frame-to-frame wise) written by different applications produce different PSNR/SSIM/VMAF scores. (https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20073) by DimkaTsv
<haasn>
dereferencing such a pointer? sure, but merely *adding*?
<nevcairiel>
Its one of those idiosyncrasies in the C spec that may go back to obscure niche systems where pointers are more special
<kasper93>
I don't like this ub too. But at the same time adding something to NULL potentially makse non-null pointer that is invalid and could be easier to deref it later
<kasper93>
as for listing changes/pr you may be interested it
<kasper93>
in*
<kasper93>
forgejo gives api, so it is two possible to 1. get all pr 2. for each pr grep list of changed files 3. print based on your filter
<BtbN>
If that exists, Forgejo/Github/... auto-assign reviewers
<kasper93>
ah, you are right
<BtbN>
I think we should have that? We basically already have it in the maintainers file, just need to translate the Maintainers file to username on code.ffmpeg.org
<kasper93>
I never though this as flexible enough, for projects with lots of small parts
<kasper93>
but yes, if that's preoperly set, it would work
<BtbN>
In theory it's also a small utility job that can run, check which files were modified in the PR, find the last X people that modified it, and add them as assignees
<BtbN>
But that seems quite involved and would very often tag random one-off contributors
<TheVibeCoder>
mkver: looks like -max_streams works only for demux, for muxing its ignored
<BtbN>
I'm still inclined to convert that python script to a github-script btw., simply because it'd be much more efficient, and wouldn't need an arbitrary "fetching 500 commits should be enough"
<BtbN>
Or the Python script could also check if it's running in an Action, and consult the API instead of git
<kasper93>
I won't have time today, feel free to do it if you want
<BtbN>
It shouldn't be too hard
<BtbN>
one API call can get you all the commit messages of the current PR
<BtbN>
and there are well defined env vars to detect running in an action
<TheVibeCoder>
jamrial: see above max_streams issue
___nick___ has quit [Ping timeout: 240 seconds]
<kasper93>
BtbN: I don't like how forgejo shows ci jobs for approval
<kasper93>
it looks like they are running/waiting for worker
<BtbN>
Yeah, the Approval Mechanism is really clunky
<BtbN>
They at the very least should have their own icon, instead of the normal "Waiting" one
<kasper93>
they do?
<BtbN>
no?
<BtbN>
That's the point
<kasper93>
ah, I read you message wrong, sorry
kurosu has joined #ffmpeg-devel
<BtbN>
they should have their own status, and not be mixed in with "Waiting"
<kasper93>
yes, agreed
<BtbN>
I'd also kinda like if Approval wasn't permanent
<BtbN>
so just for that PR, or for a couple days or so
witchymary has quit [Remote host closed the connection]
witchymary has joined #ffmpeg-devel
Flat has quit [Quit: Rip internet]
<jamrial>
TheVibeCoder: that option is flagged as demuxing only in options_table.h
Flat has joined #ffmpeg-devel
<Lynne>
let's add a codeowners file
<TheVibeCoder>
jamrial: that is point, you can not mux more than 1000 streams but can demuxer more than 1000 streams if you have such file
<TheVibeCoder>
jamrial: but even if you flag it also as mux it will not fix the issue
<kasper93>
BtbN: what do you mean by "permament"?
<kasper93>
you mean it should remove approve after push?
minimal has joined #ffmpeg-devel
<fjlogger>
[FFmpeg/FFmpeg] Issue edited: #20073 Same videos (frame-to-frame wise) written by different applications produce different PSNR/SSIM/VMAF scores. (https://code.ffmpeg.org/FFmpeg/FFmpeg/issues/20073) by DimkaTsv
<BtbN>
kasper93: for non-members, it should not be forever, yeah
<BtbN>
or have an option to un-approve someone at least (without rummaging through the DB)
<jamrial>
TheVibeCoder: so the cli has a single dictionary for format avoptions, be it demuxer and muxer. it will apply it to the former first, not caring how the option is flagged, then remove it if applied
<fjlogger>
[FFmpeg/FFmpeg] Pull request opened: #20088 avformat/mov: move AVC-Intra extradata generation to earlier in the stsd parsing process (https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20088) by jamrial
<fjlogger>
[FFmpeg/FFmpeg] Pull request review requested: #20088 avformat/mov: move AVC-Intra extradata generation to earlier in the stsd parsing process (https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20088) by jamrial
chillfish8 has quit [Quit: Konversation terminated!]
mkver has quit [Ping timeout: 260 seconds]
fjlogger has quit [Remote host closed the connection]
<pross>
BtbN: I am new to forgejo way of doing things. To 'push' an approved PR to git, do i just click on the 'Rebase-then-fast-forward' button?
<fjlogger>
[FFmpeg/FFmpeg] Pull request #20055 closed: BIGSLEEP-433502298 (dashdec) and BIGSLEEP-434637586 (sanm) and a unrelated issue in subfile i stumbled across while testing (https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/20055) by michaelni
<BBB>
BtbN: done
<BBB>
so ... how do I "watch" certain files (or components - like tags) for changes to make review easier?
<BBB>
e.g. I'd like to get notified when "vp9*.c" has patch submissions
<fflogger>
[editedticket] Balling: Ticket #5413 ([avformat] Image size, pixel format cannot be detected ( 0x0,-1 ) for short h264 video) updated https://trac.ffmpeg.org/ticket/5413#comment:9
<BtbN>
I don't know a good way to auto-assign issues to you
<BtbN>
that'd be some LLM level of magic to pull off
<BBB>
the alternative is that we need to hire some kind of person to "manage" our bug tracker, in the same way that trac used to be monitored by e.g. carl
<BBB>
I'm not saying carl is the best candidate, but the general idea of someone being professionally responsible for bug queue management isn't crazy