07:31 <aka_[m]> Stupid question inc

07:31 <aka_[m]> Can halium distros run mesa tests?

07:31 <aka_[m]> And if so would they use android proprietary drivers to do so

07:36 <DavidHeidelberg> aka_[m]: piglit and DEQP yes (since they can run on anything), the tests from `meson test` - never thought about it

07:45 <aka_[m]> I wanted to try run tests and check how much more broken is mesa than android drivers

07:45 <aka_[m]> State of 5xx is so bad 😑

08:26 <konradybcio> Pull requests welcome!

08:56 <DavidHeidelberg> aka_[m]: if youre on alpine, check aports piglit MR and you can try fix some failing tests, Rob Clark could know about low hanging fruits on a5xx (in our CI it's also pretty unstable)

09:53 <aka_[m]> konradybcio: open issue is best I could do

09:59 <DavidHeidelberg> aka_[m]: sorry, that's not enough for a5xx when we have noone actively working on it. If you open issue with biswcted commit of regression, that would a give a value

10:00 <DavidHeidelberg> *bisected

10:03 <z3ntu> Honestly from postmarketOS perspective we could definitely pay someone some money to work on fixing mesa issues, we have a good chunk of money on opencollective now. Obviously not 5 people full time for many years but one-off things shouldn't be too impossible? But I'd also need to check how we could handle such a thing - if even someone's interested in working on this (for money)

10:05 <DavidHeidelberg> Maybe some bounties on specific fixes? Ofc depend on cpmpleyity of issue, because the entry level is pretty high

10:05 <DavidHeidelberg> *complexity

11:40 <konradybcio> aka_: opening issues is already contribution

23:13 <lumag> z3ntu, I think the biggest bounty should be to make konradybcio finish the CPR work.

23:13 <lumag> including the open-loop support for GPUs ;-)

23:30 <konradybcio> if you also pay for therapy..