<headius[m]>
donv has found a peculiar bug with JRuby on Android where for loops don't seem to be sharing their parent scope
<headius[m]>
kares: chatted with enebo briefly yesterday and we're going to put 9.2.8 without the load/require branch
<headius[m]>
it needs more work for me to be confident in it, even though it's like 99% there
<headius[m]>
and with Rails 6 still in beta we have some time before zeitwerk starts seeing a lot of use
<enebo[m]>
I am on riot
<enebo[m]>
headius: I commented on for loop thing. My main confusion there is why would it be different on android than master if he is in fact using master
sagax has quit [Remote host closed the connection]
<headius[m]>
it's quite odd
<headius[m]>
I'm looking at it now and the static scope for the for loop is clearly not the one for the surrounding scope
<enebo[m]>
isn't gitter being echoed here now?
<headius[m]>
no that was broken and spotty so we had to disable it
<rtyler>
sometimes the bridge works, sometimes it doesn't
<headius[m]>
messages were only going one way, etc
<headius[m]>
I'm pretty sure we just turned it off because of that
<headius[m]>
ok this is interesting
<headius[m]>
I think it's digging too far up
<enebo[m]>
ah counting
<headius[m]>
that would explain why it's nil and not an error...it seems to be getting the scope above the method body
<headius[m]>
which is an eval scope
<enebo[m]>
hmm it is possible something changed there possibly but I will have to check
<enebo[m]>
I would love to see the exact code if so
<kares[m]>
okay, looked good to me. how far do you estimate 9.2.8 is?
<kares[m]>
also we could try the gitter bridge again, maybe they fixed it
<enebo[m]>
kares: I am thinking Monday? I guess if you see something to do for 9.2.8 let me know
<enebo[m]>
kares: I still have not tried rails 6 rc2 with uit yet
<kares[m]>
not really, just thinking of helping out with some bugs but not sure how much time I'll have
<kares[m]>
... did not release AR-JDBC 60 for the new rc2 - as it needs some real Rails app testing
<kares[m]>
and I am not sure I know anything up-to-date
<enebo[m]>
kares: well I will build gem and try what would be rc2 with this and give it sanity test
<enebo[m]>
kares: but previous rc1 did not seem to generate a lot and rails tests were passing pretty good
<enebo[m]>
kares: I would almost be more worried about updates to the JDBC drivers themselves than local changes to make rc2 work
<kares[m]>
yy but that is pulled off pretty much
<kares[m]>
jdbc-mysql is on the wait list for the released 5X gems to get used more
<kares[m]>
and for jdbc-postres that was also updated but should be pretty stable
<enebo[m]>
kares: so what more involved testing can we do past making a simple app and seeing if it works?
<kares[m]>
but I did notice one new failure with updated PG server
<kares[m]>
well I was hoping there to be a Rails app, but that sounds naive now that I am wondering
<kares[m]>
since its pretty much just github and shopify testing out the rcs
<enebo[m]>
yeah. I think we can make sure an app boots and does some obvious stuff but we are somewhat at the whim of Rails internal testing and on that we are like 99.9\d+ passing
<kares[m]>
yy that is pbly as far as we can test
<kares[m]>
and if rc1 worked fine ... a quick look at the changeset might also tells us a lot
<enebo[m]>
kares: yeah true. It is interesting that each rc breaks arjdbc but it is seemingly a small thing
<kares[m]>
see some changes but some of it was definitely covered by dritz's work
<enebo[m]>
coolio
<kares[m]>
oh and they keep making the pool more complex :) ... started using weakref
<kares[m]>
that I never used under Ruby ... just used Java's refs with JRuby
<kares[m]>
but I guess its well and working
<kares[m]>
`::ObjectSpace::WeakMap.new` heh
<kares[m]>
its all implemented with a backing WeakValuedIdentityMap
<kares[m]>
thus should be fine
<enebo[m]>
kares: if it does make a significant difference we can entertain patching that out but at this point we should let it ride until we want to bite off supporting something we probably don't really want to support
<kares[m]>
yy just wasn't sure about the weakref library
<kares[m]>
they also implemented a NullPool ... well if that works its good news for JNDI
<kares[m]>
on the adapters its pretty much just `self.database_exists?` added
<enebo[m]>
oh yeah nullpool sounds like it would be the ticket
<kares[m]>
so I guess it should be good to go
<enebo[m]>
yeah I saw that was the new change
<enebo[m]>
kares: ok well I will make sure rails6 + master of arjdbc works on a simple app on windows and linux
<enebo[m]>
probably tomorrow morning
<kares[m]>
👍️
<enebo[m]>
but I am not too worried about arjdbc being broken since the changeset between rc1 and rc2 is pretty small
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
xardion has joined #jruby
shellac has joined #jruby
drbobbeaty has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
lucasb has quit [Quit: Connection closed for inactivity]