13:11 <zmike> kusma anholt: does one of you want to ack docs https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/19205

13:15 <kusma> done

13:16 <zmike> thx

13:19 <kusma> I'm generally super happy with more insight into what's going on, since it seems like I'll be back to Zink work for a while now! So thanks a lot!

13:20 <zmike> I'll be back before the next branchpoint hits, but I'm hoping I can leave enough breadcrumbs that nothing gets blocked waiting for me

13:22 <kusma> Yeah, I think we'll manage. It just makes it easier :)

16:42 <zmike> anholt: any other areas you think are in dire need of docs

16:43 <anholt> descriptors were the place I kept getting really lost. calling vk images zink_surfaces was another trap (especially since pipe_surface already means something that's not zink_surface).

16:44 <zmike> hm ok surfaces this afternoon then

16:44 <zmike> descriptors should be good now

16:45 <zmike> or gooder, at least

18:37 <zmike> surface docs done

22:34 <zmike> anholt: are you able to repro the crash on the counterstrike trace from !19306 ?

22:35 <zmike> I tested a bunch of times here on my a630 with the same config and it seems fine

22:40 <zmike> and it passed on retry so...I guess it's a flake?

22:40 <zmike> bizarre

23:24 <anholt> link to job would help a ton

23:25 <zmike> oh sorry

23:25 <zmike> uhhh I'm not at a computer, will look when I get home

23:26 <anholt> https://gitlab.freedesktop.org/mesa/mesa/-/jobs/30426498

23:27 <anholt> all in zink_dri.so, but no useful backtrace :/

23:30 <zmike> yeah...

23:35 <anholt> wonder if we should stop stripping artifacts in general at this point. we're less desperate about storage and bandwidth on packet these days.

23:45 <daniels> far less desperate

23:45 <daniels> go for it; if it causes a problem it can be reverted