<byteit101[m]> headius: I keep finding eve more issues about become_java! being implemented. Just closed this one: https://github.com/jruby/jruby/issues/3366
andrea[m] has quit [Quit: You have been kicked for being idle]
<headius> byteit101: nice, good to get those cleaned up
<headius> enebo: I will continue trying to clean up suites today
<headius> master now uses strscan and stringio gems... just working out remaining testing and release stuff with @kou
<enebo[m]> headius cool. kwarg life
<headius> 😎
<headius> wow Random changed to have a base class with most of the stuff in it at some point
<headius> didn't I just port this a couple months ago? Perhaps this was a 3.1 change
<headius> Random < Random::Base, and most of the important instance methods are on Base now
<headius> Base includes and extends Random::Formatter which has the various typed random things like random_number
<headius> source of a few failures in test:jruby because I added Formatter but did not adjust Random ancestors
<headius> enebo: ongoing work here but test:jruby is nearly green so far: https://github.com/jruby/jruby/pull/7179
<headius> the remaining two failures are due to a change in how it calculates relative vs absolute paths which broke our support for URLs when we updated pathname.rb
<headius> I never filed an issue for it but this comment mentions the change: https://github.com/jruby/jruby/pull/7124#issuecomment-1064121691
<headius> basically we had a regexp that was used to detect an absolute path or URL, and now they flipped it so that absolute is anything !relative and that regexp does not work right with URLs
<headius> er my bad
<headius> we used to do !relative = absolute, but now absolute = match this regexp
<headius> and now relative = !absolute
<headius> so we need to figure out how to handle URL in that logic
<headius> I have to run for a while
drbobbeaty has quit [Ping timeout: 248 seconds]
drbobbeaty has joined #jruby
<headius> fixing a couple in test:jruby I missed... one JI thing trying to call top-level `import` and one socket error I broke while adding connect timeout to TCPSocket
<enebo[m]> I am a little past where I was when I started to redo this. I am tackling broken ruby2_keyword tests
<headius> nice
<enebo[m]> receiveKeyword is a little sloppy still but it should get clean by the time it is all done
<enebo[m]> ruby2_keyword is not 100% for me. I thought it was in the past but it is like I forgot everything about it :)
<headius> heheh
<headius> yeah mostly I know it allows using normal hashes for kwargs in more places
<enebo[m]> you receive the hash and it gets marked as ruby2_keyword
<enebo[m]> I think the design is weird
<enebo[m]> then if you look at it this field now exists on it
<enebo[m]> but if you then call a method which accepts kwargs is nukes the field and makes it a kwarg
<enebo[m]> this is probably extremely broken for concurrent calls
<enebo[m]> I make that sound scary but I think this would be a very rare thing
<headius> <"Connection refused - connect(2) for \"0.0.0.0\" port 42"> expected but was
<headius> <"Connection refused - connect(2) for 0.0.0.0 port 42">
<enebo[m]> oh hahah
<enebo[m]> Seeing the " is quite hard
<headius> yeah
<headius> without that test:jruby is down to just two Pathname relative pathing failures
<enebo[m]> here is my favorite error we have...
<enebo[m]> Expected ArgumentError ((?-mix:unknown keywords: foo))
<enebo[m]> but got: ArgumentError (unknown keywords: :foo)
<headius> hah
<headius> yeah those ones that throw some regexp inspecting in are weird
<enebo[m]> so is it possible that is not actually a single string?
<enebo[m]> or perhaps are we inspecting the regexp in expected wrong?
<enebo[m]> It looks so unimportant I have not even looked at it
<headius> is this from accessing a group with a non-existing name?
<headius> oh hahahah
<headius> foo vs :foo
<headius> that's all it is
<headius> it wants the keyword inspected
<headius> the ?-mix is just it printing out the regexp it expected the error to match
<enebo[m]> oh man
<headius> ok I fixed that TCPSocket message formatting... need to decide what to do with pathname but test:jruby is nearly there
<enebo[m]> headius: I think travis was hacked recently...if we have any oauth stuff we should remove it
<enebo[m]> headius: Looking at our project page I am not even sure where that could be
<headius> figures
<headius> enebo: probably has my credentials for deploying to sonatype so I'll sort that out
<headius> changed my password and updated secrets for GHA
<headius> according to GH network stats no untrusted forks of JRuby have been pushed recently, and we have not had travis running for some time, so we should be safe