drbobbeaty has quit [Remote host closed the connection]
drbobbeaty has joined #jruby
drbobbeaty_ has joined #jruby
drbobbeaty has quit [Ping timeout: 256 seconds]
PavanNambi[m] has quit [Remote host closed the connection]
skumarg[m] has quit [Remote host closed the connection]
AlenSebastian[m] has joined #jruby
<headius> enebo: let me know when joni release is out and I can update the PR I made for it
<headius> I'm fixing dependency thing on jnr branch and then I'll merge it and fix the irb test thing
<enebo[m]> ok
<headius> PRs are down, I just noticed it myself because jnr update PR did not see my push
<headius> yeesh, lots of stuff breaking on GH now
<enebo[m]> hmm
<enebo[m]> my git-commit doing release:prepare is failing
<enebo[m]> It could be pushing but I half feel this is unrelated
<enebo[m]> [ERROR] /bin/sh: _module_raw: line 1: syntax error: unexpected end of file
<enebo[m]> headius: can you release joni from head right now?
<enebo[m]> I think something with bash + FC is messing up the scripting
<enebo[m]> I am only using this laptop another week or so and don't want to deal with this
<headius> enebo: sure
<enebo[m]> I merged so it should be good to go
<headius> not sure what to do about these PRs lagging right now
<enebo[m]> you mean actions?
<headius> enebo: 92d391e5625bf2b113f150a26b322b438234b2b9?
<headius> according to the status, PRs are slow, webhooks are slow, actions are slow
<headius> so it's not even running my PR to update irb tests and did not see my force push to jnr
<enebo[m]> ok. yeah that one
<headius> ok
<enebo[m]> if you diff you will see Java 8 and the backoff fix to not use fast //i search
<enebo[m]> I think lopex has realized that the logic in the method is wrong and picking fast in cases it shouldn't
<enebo[m]> but it is something which will get figured out later
<headius> yeah funny that IOOBE did not have a numeric constructor
<enebo[m]> no doubt message on 8 will just be "1" :)
<enebo[m]> newer API probably builds a nicer string but it is pretty obvious what 1 means with IIOBE
<headius> yeah
<headius> just pushed the release button on sonatype
<enebo[m]> I was like ok not much output it must be fixed
<enebo[m]> looks like rake just exited?
<enebo[m]> /home/runner/work/jruby/jruby/lib/ruby/stdlib/irb/workspace.rb:93: warning: previous definition of exit was here
<enebo[m]> maybe tests are causing rake to exit
<enebo[m]> you may want to use runner locally just to rule out GHA being responsible
<headius> yeah that's odd
<headius> I will check on linux machine
<headius> ugh something in the tests does terminate the run
<enebo[m]> that warning could be a clue
<enebo[m]> since it is about exit
<headius> heh ya thing
<headius> ya think
<headius> yeah probably related
<enebo[m]> ya thing
<headius> make your bets on whether it's refinement related
<enebo[m]> 💯
<headius> there's a lot of failures in the new tests looking for specifically CRuby's load path layout
<headius> lame
<enebo[m]> ouch
<enebo[m]> just assuming index of X for test or something more behavioral?
<headius> [ 19/231] TestIRB::EditTest#test_edit_with_class_method = 0.05 s... (full message at <https://libera.ems.host/_matrix/media/v3/download/libera.chat/1e4b8812057524625a41960e0eea440ce85d1e94>)
<headius> literally trying to match how CRuby does stdlib path
<headius> [ 21/231] TestIRB::LocaleTestCase#test_find = 0.02 s
<headius> TestIRB::LocaleTestCase#test_find [/home/headius/work/jruby/test/mri/irb/test_locale.rb:111]:
<headius> 7) Failure:
<headius> Expected "/home/headius/work/jruby/lib/ruby/stdlib/irb/lc/ja/error.rb" to include "/lib/irb/lc/ja/error.rb".
<enebo[m]> ah
<headius> there's several of these
<enebo[m]> stdlib has an rbconfig value or no?
<headius> I don't think so
<headius> it could capture the path of irb.rb at load though and use that to infer other paths
<headius> the tests need to be fixed though
<headius> I pushed updated excludes for the 15 "new" failures (some were renamed or respelled) but there's still some seeds that lead to early exit
<enebo[m]> I was poking but I see nothing
<headius> going to try to catch that
<enebo[m]> feels like an oversight but ultimately the test should test if that is anywhere in loadpath
<headius> narrowed down to test_ruby_lex that seems to be terminating early
<enebo[m]> mice
<headius> ok I haven't narrowed anything down
<headius> something this is setting up early in testing is terminating things later
<headius> ok I have a lead
<headius> I think something in did_you_mean is triggering and causing a stack overflow
<headius> like a method is missing and then DYM tries to do its thing but the method is missing so it raises again
<headius> that eventually bubbles out and the stack overflow is swallowed and it exits
<enebo[m]> I saw the respond_to_missing? missing from BasicObject...that was in earlier run before it exited
<headius> enebo: I know we are under a time crunch here so I propose two options
<headius> I can roll back IRB and try to get a version that does not have tests that terminate our run, or not update it at all and hope it works with newer reline
<headius> I can also just remove irb tests from stdlib testing and we fix this after release
<headius> it seems to be just a test thing
<enebo[m]> I think we could try this combo of irb+reline informally and see if they work on linux and windows
<headius> it was failing but not terminating with old tests so I am assuming something in new tests is hitting a bad spot for us
<enebo[m]> I do always try and start it using cmoplete on windows too
<enebo[m]> as it stands windows gets that TSPC signal crap which was fixed
<headius> it looks fine to me on Linux
<headius> yeah and that was in reline yes?
<enebo[m]> I think so
<headius> I have had runs of IRB tests complete too so it's something with particular sequence of tests that causes it to bail all the way out
<headius> I see it execute the last "exit" in runner.rb so it bubbles out weirdly
<headius> silently terminates the test run
<enebo[m]> so we could try and roll with irb + reline or just update reline and leave irb for next point
<headius> that's about the size of it
<enebo[m]> reline update is likely safer since it is just a few point releases and we are using first in a major
<headius> FWIW this is way ahead of 3.1's default IRB but we were already ahead
<enebo[m]> yeah let's just reline and I will make sure I can run it
<enebo[m]> we know M1 is not working
<headius> I can roll back to CRuby 3.1.4 version of IRB or whatever we had before, whichever is newer
<headius> ok
<headius> I will ditch this IRB test PR for now then and make one to roll back IRB
<enebo[m]> ok
<headius> ok pushing PR in a moment
<enebo[m]> coolio
<headius> enebo: ok we were on 1.4.2 but there's a 1.4.3: https://github.com/ruby/irb/releases/tag/v1.4.3
<headius> actually nevermind... I'll just do 1.4.2 rather than try to wrestle with any test changes from 1.4.3
<enebo[m]> A lot of changes
<headius> we'll do a proper update after release
<enebo[m]> ok
<headius> 3.1.4 is still on 1.4.1 so we're already ahead on that one
<enebo[m]> yay
<headius> JNR PR was down to just IRB failures so I merged it
<enebo[m]> likely we use 1.7 for next release and not 1.4.3. but I guess it depends on time spent to correct tests
<enebo[m]> ok cool
<headius> are you pushing joni update or shall I?
<enebo[m]> I can't
<enebo[m]> I get that weird error
<headius> I mean updating JRuby... the release is out and should be propagated now
<enebo[m]> oh I see. you have a PR already for that right?
<enebo[m]> you just need to update that?
<headius> yeah I can do that
<headius> I'll update so it reflects 2.2.1
<enebo[m]> yeah that is fixed by this joni release
<enebo[m]> I marked it 9.4.3.0 earlier
<headius> ok
<enebo[m]> That issue was the final nail for needing something changed since it was 2 more people
<headius> yeah good to get it in this release
<enebo[m]> this should not affect log4j since that was no //i
<enebo[m]> and //i being slow again is ok since it was slow for more than a year except for last point release
<enebo[m]> I think there may be other errors from updating but none look important
<headius> there are
<headius> I'm looking into updating those tests, probably tests added for fixes in these minor releases
<headius> some of these may have cve updates so I'd like to keep them
<enebo[m]> ENVIRONMENT_VARIABLE_IS_MULTIUSER_SAFE this is the only one I wonder about
<enebo[m]> that test open BADF has been happening for a while
<headius> I'm on it.. there's only a few
<headius> yeah
<headius> silly me thinking I could update these without tests
<headius> updating tests is looking pretty clean
<headius> oops pushed test updates and excludes to master
<headius> I will merge irb thing because that is no longer failing tests
<headius> that's the job to watch.. joni merged, updated tests and excludes, and irb downgrade all on master now
<enebo[m]> once merged I will spin and start testing
<headius> enebo: that build was completely green
<headius> ship it
<enebo[m]> yeah I am spinning now
<enebo[m]> perhaps write some release notes optimistically :)
<headius> oh yeah
<headius> I'll get on that
<headius> enebo: you have that issue list for me
<headius> ?
<enebo[m]> oh I can make it but I think I need to re-review whether all things were marked
<enebo[m]> headius: looks good
<enebo[m]> windows irb is "weird" but it is just tab completion of tokens is displaying oddly
<headius> notes coming along
<headius> lots of small fixes
<enebo[m]> I think the workaround on that is to disable completion or use swing irb
<enebo[m]> but perhaps newer irb will fix that
<enebo[m]> in any case it is better than what is was
<enebo[m]> I am going to finalize maven and push gem now
<enebo[m]> cool
<headius> I did not see a lot of other multi-issue/multi-PR major items
<enebo[m]> yeah I think a lot of small fixes for things like ripper crashing irb
<enebo[m]> the regexp fix (or unfix for //i)
<enebo[m]> doesn't matter too much
<enebo[m]> ok looks like it is up on maven
<enebo[m]> headius: going to push site. Can you do the tweet/toot/social + image thingies you have been doing
<enebo[m]> I will do GH release and update GH
<enebo[m]> yikes web site is taking a while
<enebo[m]> ok website is propd
enebo has joined #jruby
enebo has left #jruby [#jruby]
<headius> ok
<headius> I'll start doing my wrap up checklist
<enebo[m]> headius: I believe I am done with the items I typically do
<enebo[m]> message me on telegram if you need me
<headius> enebo: no problem
<headius> installers are updated, socials are posted, just doing docker now
<enebo[m]> cool cool