Thoughts about adopting community fork of intel-vaapi-driver

Intel halted development of intel-vaapi-driver and freedesktop.org SDK maintainers made the only sensible choice and dropped it from runtime 25.08.

Since then, a community fork has emerged, and some distributions are adopting it. The lack of intel-vaapi-driver in 25.08 did impact some kodi flatpak consumers, and we made the decision to rollback and delay migration to runtime 25.08 until 24.08 declared totally dead.

As a maintainer of kodi on flathub, my duty here is simple: is there any motivation at all to adopt this community maintained fork or should I prepare to advise users about alternative options that do support legacy intel hardware acceleration once we do make the jump from 24.08 to 25.08?

The vaapi extension comes from fdsdk, the question should go there.

As an aside, the hardware support that was lost after that was dropped is more than 10 years old at this point. Most of the cards were around 2008-2012 with a few being from 2017ish.

In general you shouldn’t expect everyone in the rest of the ecosystem to provide you unlimited backwards compatibility stretching decades.

So from my perspective asking your users to move on to newer hardware or alternatives is the sensible option.

and some distributions are adopting it.

I have never heard of LibreELEC as a general Linux distribution and looking at the readme it seems something specifically geared for Kodi.

LibreELEC is a ‘Just enough OS’ Linux distribution for the award-winning Kodi software on popular mediacentre hardware.

There’s nothing wrong in that but using it as an example here of “some distributions adopting it” feels wrong and roundabout when you are asking more general purpose things to adopt the fork.

Thanks for your thoughts, and please stop thinking I am being disingenuous, it was an honest and simple question only without any expectations. I am just trying to be a half decent maintainer here.

I didn’t think you were being disingenuous. My point is you should have some more convincing argument and examples if you want a platform to (re)adopt some legacy piece of software that has been dropped upstream in two places.