ChanServ changed the topic of #zink to: official development channel for the mesa3d zink driver || https://docs.mesa3d.org/drivers/zink.html
omegatron has quit [Quit: What happened? You quit!]
hch12907_ has joined #zink
hch12907 has quit [Ping timeout: 480 seconds]
omegatron has joined #zink
hch12907_ is now known as hch12907
<kusma> That seems to be the last detail needed to land zink + msvc on CI
<kusma> got one from Jesse :)
<zmike> kusma: your msvc job broke ci
<kusma> Huh?
<zmike> some dependency thing I guess
<zmike> dunno how the original MR didn't trigger this
<zmike> seems like that should be a bug
<kusma> Did you get a gitlab notification about it?
<zmike> about what?
<kusma> I got an email about something similar, but I assumed it was a fluke because the MR passed
<zmike> I didn't get a mail
<kusma> My mR didn't change any of those things that the error complained about
<kusma> I think this is a pre existing issue
<zmike> 🤔
<kusma> But you're now seeing it because the windows stuff builds for zink changes
<kusma> I'll take a look tomorrow
<zmike> it seems suspicious to me that the existing d3d12 job wouldn't have been adding the msvc build job that's required to run the tests
<kusma> All I'm saying is I didn't change any relationships between jobs, just a tag and a script
<zmike> the d3d12 job depends on the rules changed here, so now the d3d12 job is triggering from zink changes
<zmike> or something like that
<kusma> Yeah, this just triggers more often now
<zmike> oh you made it also run the d3d12 job?
<zmike> and that now runs both?
<zmike> I think I gotta revert that top patch for now and you can revise it when you get time
<kusma> Uhh, please don't.
<zmike> we can't have a non-functional ci
<zmike> disabling a broken job until tomorrow isn't going to kill you
<kusma> No, I just made the build also trigger on zink changes
<zmike> yes, I see that
<zmike> and it now can't run ci at all
<zmike> it's fine that you're off the clock now
<zmike> you can fix it tomorrow
<zmike> but not having ci able to run for however long is not an option
<kusma> Please don't revert to land more Zink changes in the meantime. You broke the msvc build in a new way every day i rebased it.
<kusma> Other mrs have landed in the meantime already
<zmike> whatever
<zmike> this is not cool though
omegatron has quit [Quit: What happened? You quit!]
<kusma> zmike: If you want to, you can cherry-pick that to your branch and see if it goes through.
<zmike> hm
<zmike> ci is beachballing me...
<zmike> I'll have to run an initial pipeline on it and then try triggering the issue to determine whether it fixes things, so it'll be a while
<zmike> thanks for looking into it today though, I really appreciate it
<kusma> TBH this seems like a bug in gitlab to me, it seems silly that we need to manually ensure all previous jobs were triggered.
<kusma> No problem, it was kinda obvious as soon as I sat down on the computer instead of scrolling on my phone
<zmike> now I can get back to my real job
<zmike> subtly breaking zink on windows
<kusma> Awesome
<kusma> The more subtle, the better
<zmike> I know how you like a challenge
<kusma> Even better if I never find out!
<zmike> might be too subtle then
<kusma> There's a bug in there that only triggers on Wednesdays if you're only wearing one shoe
<zmike> there wasn't before, but now that you've given me the idea...
<kusma> That's what I get for buying those fancy smart shoes
<zmike> smart shoes, now that's a nightmare scenario
<kusma> Google needs to know whenever you take a step
<kusma> Oh didn't they already buy fitbit? Nevermind
<zmike> I was more imagining that the smart somehow controls the laces
<zmike> so every time you software update suddenly your laces are tripping you all the time
<kusma> I'd buy that
<kusma> Kinda like in back to the future
<kusma> I wouldn't wear them, but I'd buy them for sure!
<zmike> yea that's what I was 🤔
<zmike> didn't nike make an actual pair of those?
<kusma> Yeah, but they didn't do that stuff
<zmike> someday
<kusma> I'm turning in for the night. Enjoy the evil scheming
<zmike> later
orbea1 has joined #zink
orbea1 has quit []
orbea1 has joined #zink
orbea1 has quit []
orbea1 has joined #zink
orbea has quit [Ping timeout: 480 seconds]
orbea1 has quit []
orbea has joined #zink