Contingent on #37. See any prior issues for more context. It's a sequence.
Perform limited processing of PKGBUILDs contents to prepare for edge kernel and GPU support
I've actually already done most of this work. (Not on github just yet.)
Let's use PKGBUILDs as the source of truth for which version of asahi tools we're building from. We pick one digest (say stable branch) and whether we're stable or edge kernel, build a corresponding kernel config, mesa configuration, etc... and they should all play well with each other. This is also where GPU support becomes configurable.
I already have stable GPU support running off this approach. :)
Trying out a new version of asahi should be synonymous with "updating the PKGBUILDs pin." Sometimes a bit more for hand tweaks, but generally just that.
I am reapplying my working changes to this new, more disciplined refactor. It shouldn't be more than a day or two.
I hope slogging through my sequence of improvements was worth it. This was the point.
onethirtyfive
commented
1 year ago
Contingent on #37. See any prior issues for more context. It's a sequence.
Perform limited processing of PKGBUILDs contents to prepare for edge kernel and GPU support
I've actually already done most of this work. (Not on github just yet.)
Let's use PKGBUILDs as the source of truth for which version of asahi tools we're building from. We pick one digest (say
stablebranch) and whether we'restableoredgekernel, build a corresponding kernel config, mesa configuration, etc... and they should all play well with each other. This is also where GPU support becomes configurable.I already have stable GPU support running off this approach. :)
Trying out a new version of asahi should be synonymous with "updating the PKGBUILDs pin." Sometimes a bit more for hand tweaks, but generally just that.
I am reapplying my working changes to this new, more disciplined refactor. It shouldn't be more than a day or two.
I hope slogging through my sequence of improvements was worth it. This was the point.