05:27 <byteit101[m]> Hmm, java -jar complete -S rspec complains of no such file, while bundle exec rspec says jruby.exe isn't a command

23:36 <headius> I wonder sometimes about all these warnings in ruby/spec... I assume some are valid but it can't be all

23:36 <headius> byteit101: I forget if we still bundle rspec

23:36 <byteit101[m]> Ts his fteair a bunde linstall and a gem install rspec

23:36 <headius> I thought we removed it at some point

23:36 <byteit101[m]> This is after*

23:36 <headius> ah ok

23:36 <byteit101[m]> (I think this matrix client has multithreading input that has race conditions)

23:36 <headius> haha

23:37 <byteit101[m]> But yea, those errors shouldn't happen, I would have thought, given that rspec is installed

23:38 <headius> yeah not sure what's up

23:38 <byteit101[m]> Oh, and not sure if you noticed, but I went through old issues and assigned myself a bunch of old process spawn and related tickets

23:40 <byteit101[m]> I think subspawn should be able to fix most of them sooner rather than later. But I would like some opinions on my plan: 9.4 remains PTY-only, while 9.5 we remove most of the java Process.spawn impl in favor of subspawn.

23:42 <byteit101[m]> Oh and I found a compatibility issue with MRI: File.open("foo.txt").fileno returns a C FD on windows & linux, MRI & JRuby. IO.pipe.first.fileno returns a C FD on MRI, and Jruby Linux, but a handle on JRuby Windows: https://github.com/byteit101/subspawn/blob/master/subspawn-win32/lib/subspawn/win32.rb#L527

23:58 <headius> I have no objection to removing the Java code in favor of the Ruby version

23:58 <headius> It was a brute force attempt to match behavior and I've never been a fan of the code