14:28 <zmike> kusma: do you have any further ideas on https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/15536

14:28 <kusma> I'm out sick right now, sorry

14:29 <zmike> dang sorry to hear it

14:29 <zmike> hope you feel better soon

14:30 <ajax> hm cubes

14:31 <zmike> indeed

14:35 <zmike> ajax: just pushed another copper update to fix some compile fails with zink_create_screen parameter count

14:40 <ajax> sweet

14:40 <ajax> i have a bunch of compile fixes locally too, will rebase atop yours

14:40 <zmike> cool

18:25 <ajax> zmike: did you see the build failure i got on debian-gallium? https://gitlab.freedesktop.org/ajax/mesa/-/jobs/20189239#L3345

18:26 <ajax> i'm reasonably sure that's a gcc parser bug, but i have this that maybe works around it, is it still right: https://paste.centos.org/view/raw/ef5818f7

18:27 <ajax> i guess the question is should we re-run update_swapchain for that goto. imma say... yes?

18:28 <zmike> ajax: the original label location was as intended

18:29 <ajax> ack

18:29 <zmike> it has to potentially try getting new images

18:29 <ajax> yeah, makes sense. will fix

18:29 <zmike> 👍

18:31 <ajax> i seem to have also accidentally broken egl for regular dri drivers in this one weird corner, but otherwise we might be ci-clean

18:31 <zmike> I hate weird corners

18:58 <ajax> was vkDriverAPI your doing or mine?

18:59 <zmike> uhhhhh

18:59 <zmike> no idea?

18:59 <zmike> I guess you could check your old copper branch to see

19:00 <ajax> the root cause of the break (above) is trying to get to zink through the old swrast init path

19:00 <ajax> which you'd think would be impossible now,

19:01 <zmike> it should be 🤔