<headius> Good morning!
<byteit101[m]> Morning!
<byteit101[m]> Oh, did the fresh windows master checkout spec failures get fixed yet?
<headius> not that I did
<headius> enebo: I figured out what CRuby does for this array deref optimization in the presence of frozen_string_literal: they don't
<headius> so I'm just duplicating that which will fix this last spec
<enebo[m]> cool
<headius> if it's pre-frozen it will just use normal call path
<headius> if not, then it will cache a frozen string for hash#[] and otherwise just send the normal str into the call
<headius> assuming this coverage CI run is good I'm going to merge it and get back to some mainline work
<headius> enebo: 9.4.2 next week?
<headius> or following
<headius> I'm gone the week after that for family vacay
<headius> I'll look into this windows issue now
<enebo[m]> We could but we should examine what he pushed off onto .2
<enebo[m]> but it is also only Weds
<headius> yeah
<enebo[m]> I believe you had some spicier PRs which needed bake time too
<headius> I will fix the windows thing and then review stuff I punted to .22
<headius> .2
<headius> this is first run where it started: https://github.com/jruby/jruby/actions/runs/4134380658
<headius> of course it doesn't reproduce on my virtual instance
<enebo[m]> yeah quite weird
<headius> yeah I don't have any explanation other than something changing in GHA
<enebo[m]> revert it and see if it works
<enebo[m]> Perhaps we run our own install as part of the build infra and this is blowing up some finalization thing
<headius> revert changed nothing
<headius> I'm seriously going to have to look at maven code
<headius> FML
<enebo[m]> Is there something in GHA which tells us what changed?
<headius> I tried to find something like a change log
<enebo[m]> This is not even the 5th time something broke behind us
<enebo[m]> but I suppose a lot of actions are built on actions
<enebo[m]> byteit101: can you run mvn -Ptest on windows?
<headius> wiping out cache
<headius> no help
<headius> maven 3.9.0 released jan 31
<headius> error is in maven core
<headius> trying that locally on mac
<headius> aha
<headius> fails
<headius> more info
<headius> I'll see if there's a way to force an earlier maven version on GHA
<headius> got it
<enebo[m]> nice
<headius> we'll add it to other jobs if they update the other maven versions to 3.9.0
<enebo[m]> leave it there forever
<headius> heh yeah
<headius> annoying
<enebo[m]> So I made @@cvar warning static during build time
<headius> nice
<enebo[m]> it will not warn in cases of binding being passed with toplevel binding
<enebo[m]> err error
<enebo[m]> but I think this is a really unimportant error
<enebo[m]> unfortunately it is a little squirrelly because I need to pass in whether an eval is explicit binding or default
<headius> also pushed a PR to update the GHA cache action, since the v2 version uses a deprecated feature that's going away
<headius> once these run clean I'll merge everything forward
<headius> should be good everywhere now
<headius> bbl
<byteit101[m]> enebo: will try after work
<enebo[m]> byteit101: it was something with maven change so the culprit was found
subbu has joined #jruby
<enebo[m]> test:jruby was erroring because of some MyArray< Array tests failing because results were still MyArray.
<enebo[m]> I deleted those because that behavior is wrong but I am confused why they suddenly are failing now
<enebo[m]> results from calling methods like slice
<enebo[m]> That commit will land with my now passing work for cvar error on top-self stuff
<headius> Yeah I don't think that changed recently
<enebo[m]> It is weird
<enebo[m]> but I removed them since we have tons of coverage on it in spec already
<enebo[m]> we also run test:jruby I think 6 times in our suite
<enebo[m]> It does run pretty fast so perhaps not a big deal
subbu has quit [Quit: Leaving]
<headius> yeah we can keep chipping away at general ruby tests in there though
<headius> most of the cases in there are copied from old MRI tests
<headius> the general Ruby stuff anyway
adam12 has quit [Quit: Ping timeout (120 seconds)]
adam12 has joined #jruby