09:03 <GitHub168> [jruby] ChrisBr opened pull request #5026: Fix String#rpartition for RegEx on Strings with multibyte chars (master...bug/fix-string-rpartition) https://git.io/vNdVr

11:54 <headius> double blast...I can't regenerated it after loading it because it's a duplicate class then

12:18 <headius> yeah this is a conundrum

12:44 <headius> this change may have to be 9.2+

12:49 <cremes> Is there a way to disable the fiber thread pool and/or lock a fiber to a thread? Checked —properties but didn’t see anything obvious.

12:50 <headius> cremes: hmm

12:50 <headius> I don't believe there is a way right now

12:51 <cremes> ok. If I wanted to hack that in for testing purposes, where would you point me?

12:51 <headius> ThreadFiber.java in createThread()

12:51 <cremes> i assume there is some kind of pool checkin/checkout that i could tweak...

12:51 <cremes> great, I’ll have a look

12:52 <cremes> is 9.2 imminent? :)

12:55 <headius> cremes: not really...need to chat with enebo about whether we should wait for his symbol identifier work or push that to 9.3

12:55 <headius> 9.2 may absorb the Ruby 2.5 branch and skip 2.4

15:01 <GitHub102> [jruby] enebo closed pull request #5026: Fix String#rpartition for RegEx on Strings with multibyte chars (master...bug/fix-string-rpartition) https://git.io/vNdVr

15:02 <GitHub28> [jruby] enebo pushed 2 new commits to master: https://git.io/vNFOE

15:02 <GitHub28> jruby/master f8de9a0 Thomas E Enebo: Merge pull request #5026 from ChrisBr/bug/fix-string-rpartition...

15:02 <GitHub28> jruby/master 2c7b8f2 Christian Bruckmayer: Fix String#rpartition for RegEx on Strings with multibyte chars...

15:22 <olle> I read this PR, and it's damn clear: "remove the try-catch, and copy in the MRI test cases" https://github.com/jruby/jruby/pull/5025

15:38 <GitHub29> [jruby] enebo pushed 4 new commits to ruby-2.5: https://git.io/vNFnO

15:38 <GitHub29> jruby/ruby-2.5 e29816f Miguel Landaeta: No longer rescue exceptions from Integer#step when given a step value which cannot be compared with #> to 0...

15:38 <GitHub29> jruby/ruby-2.5 b729677 Miguel Landaeta: No longer rescue exceptions of #coerce for Numerical comparison operators (< ,<=, >=, >)...

15:38 <GitHub29> jruby/ruby-2.5 28100b3 Miguel Landaeta: Update MRI tests for methods that previously hid exceptions from coerce methods

15:38 <GitHub5> [jruby] enebo closed pull request #5025: No longer hide exceptions from coerce method (ruby-2.5...no-longer-hide-exceptions-from-coerce-method) https://git.io/vN5Kr

15:47 <cremes> headius: I know that fibers can be backed by different native threads but I didn’t think this would change the RubyThread. Is it expected that Thread.current calls from within a Fiber would return different values over time?

15:49 <GitHub84> [jruby] enebo pushed 1 new commit to ruby-2.5: https://git.io/vNFcb

15:49 <GitHub84> jruby/ruby-2.5 b51ad8b Thomas E. Enebo: Minor refactoring after last PR (I like early returns vs nesting) and unused...

16:48 <GitHub82> [jruby] enebo closed pull request #5021: Do modifyCheck for each pair in Array#sort! (master...bug/frozen_sort) https://git.io/vN9fl

16:48 <GitHub115> [jruby] enebo pushed 2 new commits to master: https://git.io/vNFEc

16:48 <GitHub115> jruby/master 68c886e Thomas E Enebo: Merge pull request #5021 from ChrisBr/bug/frozen_sort...

16:48 <GitHub115> jruby/master 2776815 Christian Bruckmayer: Do modifyCheck for each pair in Array#sort!...

20:37 <cremes> I know I can send ‘kill -QUIT <pid>’ to get the JVM to dump its *java* thread backtraces. Is there something similar for getting all of the Ruby thread & fiber backtraces?

21:49 <headius> cremes: this is a known issue/problem/bug in JRuby, Thread.current will reflect the thread backing the fiber rather than the parent thread

21:49 <headius> MRI actually had similar inconsistencies that led them to introduce fiber locals because thread locals were not always in the right place

21:49 <cremes> ah…

21:49 <headius> Thread.current from within a single fiber shouldn't change over time, but it won't reflect the parent thread

21:50 <cremes> ok, that’s good to know for certain. I had a few moments where I thought I was losing my mind.

21:50 <headius> I played with forcing it to become the current thread but synchronization of fiber results, resume/yield, etc got really difficult if it just became the other thread

21:50 <headius> thread-based fibers are just a major hassle...if I could have real stack-swapping fibers I'd do it in a heartbeat

21:51 <cremes> I have an open bug on MRI’s tracker requesting that they allow fibers to move freely between threads. They’ve already come back and said “no” but havne’t given reasons yet.

21:51 <cremes> yeah, real coroutines would rock

21:51 <cremes> maybe java11 ;)

21:53 <cremes> too bad quasar or whatever that patch was called never got accepted into mainline jvm

22:00 <cremes> shoot, a bunch of my code was written with the assumption that there would only ever be one parent Thread.current per Fiber and *all of its child fibers* within the same thread.

22:01 <cremes> I don’t see an easy work around for JRuby… maybe a singleton/global registry where we can track the “ancestry” of a fiber and its parent thread…

22:01 <cremes> ^^ I know that probably makes no sense to you but it does in the context of what I’m working on.

22:15 <cremes> headius: I’ll search for this issue on github. If I don’t see it, I’ll open one up to track it.

22:15 <cremes> Thanks for your responses.

22:36 <jrubyn00b> For this file https://gist.github.com/anonymous/b74cbed674240e5c9ee6b3df837b06ea.txt - why would jruby come up with this error?

22:36 <jrubyn00b> ThreadError: unable to create new native thread initialize at org/jruby/RubyThread.java:571 new at org/jruby/RubyThread.java:522 block in child.rb at child.rb:13 loop at org/jruby/RubyKernel.java:1292 <main> at child.rb:7

22:37 <jrubyn00b> Buh, ignore me.