obsproject
Upstreaming current ffmpeg patches
Operating System Info
Other
Other OS
Arch Linux
OBS Studio Version
32.0.2
OBS Studio Version (Other)
No response
OBS Studio Log URL
N/A
OBS Studio Crash Log URL
No response
Expected Behavior
pacman doesn't replace the system ffmpeg package
Current Behavior
pacman asks to replace the system ffmpeg package
Steps to Reproduce
- Build the obs-studio-tytan652 package in the AUR
Anything else we should know?
So there's still 2 tiny ffmpeg patches in this repo that aren't upstreamed (which is a problem because obs-studio-tytan652 AUR package requires a custom ffmpeg build that replaces the system one (which can lead to dependency hell sometimes) likely because of that reason alone)
So are these upstream-friendly enough to make merge requests or do they require some changes? One problem I see is that there's no authorship information for them (which might be problematic for upstreaming)
Even if those patches were upstreamed, OBS currently is built against FFmpeg 7.1 and afaik we haven't updated it to work with 8.x yet.
I don't know why they were never upstreamed (asinine mailing list process aside), but I believe both were written by us, so that shouldn't be a problem.
The deps.ffmpeg/patches/FFmpeg/0001-flvdec-handle-unknown.patch and deps.ffmpeg/patches/FFmpeg/0001-flvdec-handle-unknown-Windows.patch patches were both authored by @Lain-B to fix what appeared to be a regression in FFmpeg where some FLV files would not play/loop correctly in OBS Studio after https://github.com/FFmpeg/FFmpeg/commit/d7638d8dfc3c4ffd0dc18a64937a5a07ed67b354.
The deps.ffmpeg/patches/FFmpeg/0002-libaomenc-presets.patch and deps.ffmpeg/patches/FFmpeg/0002-libaomenc-presets-Windows.patch patches were both authored by @PatTheMav in #77 as far as I can tell.
Both patch sets are trivial one liners and can probably be independently submitted if they are actually viable for inclusion.
The
deps.ffmpeg/patches/FFmpeg/0002-libaomenc-presets.patchanddeps.ffmpeg/patches/FFmpeg/0002-libaomenc-presets-Windows.patchpatches were both authored by @PatTheMav in #77 as far as I can tell.
Those are also Lain's work.
I've already submitted the aom one upstream (updated since the rane now actually goes up to 12): https://code.ffmpeg.org/FFmpeg/FFmpeg/pulls/21041
The other one seems like something we might need to fix in media-playback?