Hector Martin on Nostr: Update on M3 and Asahi: Due to a bunch of reasons, I think we're going to be focusing ...
Update on M3 and Asahi: Due to a bunch of reasons, I think we're going to be focusing on existing M1 and M2 series chips for a little while, so M3 support will probably be delayed (at least from my side).
The main reasons are no M3 Mac Mini yet (means we can't CI it), plus I've been taking a break due to personal stuff that has gone on longer than expected, and we have a backlog of things that are honestly more pressing for existing machines, and a lot of really cool stuff that makes sense to invest time on instead of the new shiny machines. I need to get us CI ASAP, there's upstreaming work to be pushed forward, all the "fun" DP/Thunderbolt work that will come together this year, and of course I can't keep putting off spending a bunch of time on power management and sleep to fix the stupid sleep power drain issue.
I definitely expect M3 work to move forward this year, but I won't be prioritizing it over the other stuff at least personally. We *are* ticking off a lot of tasks off for M1/M2 though, so this is a finite issue. And I expect M4 probably won't diverge as much from M3 in important areas (GPU stuff, nonsense like mismatched timer timebases on CPUs), so if we run into that release it probably won't take as long to catch up M3 and M4 together.
The main reasons are no M3 Mac Mini yet (means we can't CI it), plus I've been taking a break due to personal stuff that has gone on longer than expected, and we have a backlog of things that are honestly more pressing for existing machines, and a lot of really cool stuff that makes sense to invest time on instead of the new shiny machines. I need to get us CI ASAP, there's upstreaming work to be pushed forward, all the "fun" DP/Thunderbolt work that will come together this year, and of course I can't keep putting off spending a bunch of time on power management and sleep to fix the stupid sleep power drain issue.
I definitely expect M3 work to move forward this year, but I won't be prioritizing it over the other stuff at least personally. We *are* ticking off a lot of tasks off for M1/M2 though, so this is a finite issue. And I expect M4 probably won't diverge as much from M3 in important areas (GPU stuff, nonsense like mismatched timer timebases on CPUs), so if we run into that release it probably won't take as long to catch up M3 and M4 together.