drbobbeaty has joined #jruby
dave___ has joined #jruby
dave___ has quit [Ping timeout: 248 seconds]
dave___ has joined #jruby
dave___ has quit [Ping timeout: 240 seconds]
<headius> jesus, finally figured out why my split frame broke CI
<headius> that took too long
dave___ has joined #jruby
dave___ has quit [Ping timeout: 260 seconds]
dave___ has joined #jruby
dave___ has quit [Ping timeout: 256 seconds]
mistergibson has joined #jruby
<GitHub41> [jruby] headius created fix_ci (+15 new commits): https://git.io/vNJGQ
<GitHub41> jruby/fix_ci 45e1b37 Charles Oliver Nutter: Cleanup and improve concurrency of frame field updates.
<GitHub41> jruby/fix_ci f7491ac Charles Oliver Nutter: Doco.
<GitHub41> jruby/fix_ci 6cd8f64 Charles Oliver Nutter: Another fixed typo in specs.
<GitHub107> [jruby] headius pushed 1 new commit to fix_ci: https://git.io/vNJnH
<GitHub107> jruby/fix_ci 747f40a Charles Oliver Nutter: Tag off remaining failures for now. #4687
dave___ has joined #jruby
dave___ has quit [Ping timeout: 264 seconds]
<headius> oif
<headius> done
<GitHub174> [jruby] headius pushed 1 new commit to fix_ci: https://git.io/vNJc7
<GitHub174> jruby/fix_ci 0c813de Charles Oliver Nutter: Last bit of tagging. #4687
claudiuinberlin has joined #jruby
dave___ has joined #jruby
vtunka has joined #jruby
dave___ has quit [Ping timeout: 250 seconds]
<GitHub24> [jruby] yui-knk opened pull request #4937: Fix "Complex#finite? returns false for NaN FAILED" (ruby-2.5...ruby-2.5_fix_test_complex_finite_p) https://git.io/vNJB1
<GitHub2> [jruby] yui-knk opened pull request #4938: Add `#test_singleton_method` to test targets (master...test_singleton_method) https://git.io/vNJBM
Puffball has joined #jruby
GitHub118 has joined #jruby
<GitHub118> [jcodings] lopex pushed 2 new commits to master: https://git.io/vNJRI
<GitHub118> jcodings/master d7aee22 Marcin.Mielzynski: fix a typo
<GitHub118> jcodings/master 403cdb9 Marcin.Mielzynski: deprecate maxLengthDistance
GitHub118 has left #jruby [#jruby]
<GitHub47> [jruby] boris-petrov opened issue #4939: webdrivers gem not consistent with MRI https://git.io/vNJzN
dave__ has joined #jruby
dave__ has quit [Ping timeout: 255 seconds]
olle has joined #jruby
<GitHub172> [jruby] nomadium opened pull request #4940: Implement KeyError#receiver and KeyError#key (ruby-2.5...add-ruby-2.5-methods-to-key-error) https://git.io/vNJD1
bbrowning_away is now known as bbrowning
dave__ has joined #jruby
dave__ has quit [Ping timeout: 256 seconds]
<headius> kares: ok so master is damn near green again, finally
<headius> but there's some weird failure, perhaps intermittent, where it claims SSL isn't available
<headius> I have no theories and will be looking into it today but if you have any thoughts let me know
<GitHub99> [jruby] headius pushed 1 new commit to fix_ci: https://git.io/vNJAJ
<GitHub99> jruby/fix_ci 3ac50c0 Charles Oliver Nutter: Try disabling compressed OOPs to pass the ripper rlimit test....
dave__ has joined #jruby
dave__ has quit [Ping timeout: 265 seconds]
ChrisBr has quit [Remote host closed the connection]
vtunka has quit [Quit: vtunka]
ChrisBr has joined #jruby
<GitHub34> [jruby] enebo closed pull request #4938: Add `#test_singleton_method` to test targets (master...test_singleton_method) https://git.io/vNJBM
<GitHub126> [jruby] enebo pushed 2 new commits to master: https://git.io/vNUUu
<GitHub126> jruby/master e2c8040 yui-knk: Add `#test_singleton_method` to test targets...
<GitHub126> jruby/master 9892aac Thomas E Enebo: Merge pull request #4938 from yui-knk/test_singleton_method...
olle has quit [Quit: olle]
<GitHub190> [jruby] enebo closed pull request #4936: Implement Dir.glob base option (ruby-2.5...ruby-2.5) https://git.io/vNfyh
<GitHub151> [jruby] enebo pushed 2 new commits to ruby-2.5: https://git.io/vNUTT
<GitHub151> jruby/ruby-2.5 4e6b330 Thomas E Enebo: Merge pull request #4936 from ChrisBr/ruby-2.5...
<GitHub151> jruby/ruby-2.5 6858258 Christian Bruckmayer: Implement Dir.glob base option...
<ChrisBr> enebo: thanks for merging, didn't expect that it gets merged without some review rounds ...
<headius> ChrisBr: thank you!
<ChrisBr> is it fine going forward with 2.5 issues or is it prefered to work on other issues first?
<enebo> ChrisBr: yeah 2.5 is great or even 2.4 on master if there is anything tagged
<enebo> ChrisBr: do what seems more enjoyable "for now" :)
<enebo> ChrisBr: the barrier to acceptance on 2.5 branch is probably lower since we have lots of time to fix mistakes or tweak if need be
<enebo> but a missing 2.4 feature will not swatted away either although since it will be nearer term it needs to fully work (ironically partially working is almost always worse than just not existing)
<enebo> The exception to that is missing a new addition to an old method
<ChrisBr> enebo: great thanks!
<ChrisBr> mostly picked the 2.5 because they seem "straight forward" for the beginning ...
<enebo> ChrisBr: yeah as it happens new features are almost always the first thing people try
<ChrisBr> hehe
<enebo> ChrisBr: unless you are an itch scratcher who is already using JRuby but we have not fixed a bug yet
<enebo> but those are the two main vectors into contribution
<ChrisBr> enebo: thanks!
<headius> kares: the SSL thing does not appear to be related to SSL
<headius> seems to be an IR Bug
<GitHub30> [jruby] headius pushed 1 new commit to master: https://git.io/vNUmA
<GitHub30> jruby/master 2d48da9 Charles Oliver Nutter: Merge branch 'fix_ci'
<GitHub72> [jruby] headius deleted fix_ci at 3ac50c0: https://git.io/vNUmj
<headius> ok, master will be really close to green now
<headius> couple intermittent things and oddities remain
<lopex> headius, enebo: Onigmo is does not blow on https://github.com/jruby/joni/issues/17 now
<lopex> because they switched length for encoding for parsing and regexp execution to this https://github.com/k-takata/Onigmo/blob/master/regenc.c#L55
<lopex> but they will still segfault on truncated chars
<lopex> so what's the point ?
<lopex> or am I missing something
<lopex> oh they dont blow on that either
claudiuinberlin has quit [Quit: Textual IRC Client: www.textualapp.com]
<GitHub187> [jruby] nomadium opened pull request #4941: Implement Dir.children (ruby-2.5...implement-dir-children) https://git.io/vNUsa
<headius> lopex: this only fails for that special character?
<lopex> no
<GitHub96> [jruby] enebo closed pull request #4940: Implement KeyError#receiver and KeyError#key (ruby-2.5...add-ruby-2.5-methods-to-key-error) https://git.io/vNJD1
<lopex> for any that is not a correct codepoint
<GitHub46> [jruby] enebo pushed 3 new commits to ruby-2.5: https://git.io/vNUsb
<GitHub46> jruby/ruby-2.5 0cc6193 Thomas E Enebo: Merge pull request #4940 from nomadium/add-ruby-2.5-methods-to-key-error...
<GitHub46> jruby/ruby-2.5 6a543f1 Miguel Landaeta: Implement KeyError#receiver and KeyError#key...
<GitHub46> jruby/ruby-2.5 0cf8356 Miguel Landaeta: Add MRI tests for KeyError#receiver and KeyError#key
<headius> what causes it to hang?
<headius> I've not been clear on that...shouldn't it either see a valid character or not but keep consuming?
<lopex> infinite loop, since the parser uses length which can return -1
<GitHub54> [jruby] enebo pushed 4 new commits to ruby-2.5: https://git.io/vNUGU
<GitHub54> jruby/ruby-2.5 f5a7ef6 Miguel Landaeta: Implement Dir.children...
<GitHub54> jruby/ruby-2.5 dd19fb3 Charles Oliver Nutter: Set source to Java 8 so we can use lambdas.
<GitHub54> jruby/ruby-2.5 5233dfd Miguel Landaeta: Add MRI tests for Dir.children
<GitHub80> [jruby] enebo closed pull request #4941: Implement Dir.children (ruby-2.5...implement-dir-children) https://git.io/vNUsa
<lopex> infinite loop, since the parser uses length which can return -1
<lopex> whoop
<lopex> s
<headius> so it goes backwards
<headius> can't just error on negative?
<lopex> no, it proceeds with 1 then
<lopex> no, it proceeds with 1 then
<lopex> er
<enebo> lopex: you are in an infinite loop too
<lopex> enebo: Onigmo isnt, we are
<enebo> lopex: you are
<lopex> we need to change the length globally
<lopex> but not for jruby string support
<enebo> lopex: so to do it there way is some large code change? like in lots of data changes?
<lopex> since in jruby code base we use different save function
<enebo> their
<lopex> *safe
<lopex> yeah, we need to change every joni parser / interpreter length invocation
<lopex> and add safeLength to encoding
<enebo> lopex: I think I use a variant of that from jcodings (or StringHelper) in lexer
<enebo> lopex: or is ours different in impl?
<lopex> you use one from StringSupport
<enebo> ah which is in JRuby and not joni
<enebo> :)
<lopex> and is different :P
<lopex> not sure it matters
<headius> enebo: one failure on 9.1 is an issue with precompiled target script...we are pushing a new scope rather than the TOP_LEVEL_BINDING scope
<headius> I think simplest might be to add a different binding instr for the root script...PushTopLevelBindingInstr or something
<headius> I'm not sure why it started failing...or how it passed before
<enebo> headius: hmm
<enebo> seems cruddy to need to make a new binding instr for root it we will always have to add it
<headius> this is only for the startup script
<enebo> I guess in theory maybe we want the instr for analysis to know a binding is there but in practice it should be a noop
<headius> the toplevel for load/require'ed files are not TOP_LEVEL_BINDING
<enebo> we can run passes on that though
<enebo> so we probably do want something but what we actually emit for JIT can be outside of that instr
<enebo> interpreter also needs to do something but we also know what IRScope it is
<enebo> I would hate to make an if/else on that though
<headius> interpreter works ok
<enebo> oh ok
<headius> this is just logic from the interpreter startup path that's not in the compiled startup path because compiled has its own call protocol
<enebo> headius: we should maybe just mark the scope if it is not already handled by scope type
<headius> but the protocol pushes a new scope instead of existing toplevel scope
<headius> hmmm
<enebo> then JIT should just emit differently in that case
<enebo> this is just the JIT right?
<headius> yeah
<headius> actually no...full probably has this too but it never compiles toplevel
<headius> so unsure if there's a vector
<enebo> it may happen for -X-C
<enebo> we JIT so it will be full
<enebo> unfortunate name there but you know what I mean
<headius> I think it's ok
<enebo> ok
<headius> toplevel doesn't have a JIT trigger at all
<headius> it only compiles in the main script and only to bytecode
<headius> it's AOT compiled, not via the jit mechanism
<lopex> numbers ?
<headius> lopex: yeah: 3...the number of failures left keeping 9.1 red :-)
<lopex> cool
<enebo> headius: so I think worst-case should be extend IRScriptBody or add boolean as rootScope and then check during JITing
<headius> enebo: I could also ignore this
<headius> yeah that's not bad
<enebo> headius: because we do want compiler passes to see a binding is present so we want some instr
<enebo> but adding a new one just for this seems like it will be pushing us past a byte of instrs (from persistence requirements :) )
<lopex> nirvdrum7160: afaik you were interested in unicode case mapping in jcodings https://github.com/jruby/jcodings/blob/master/test/org/jcodings/specific/TestUnicode.java#L52
<enebo> I would like to partially stuff the genie back in the bottle on instr count since the original specialization of calls was not intended to be new full instrs but just simple implementations of call
<lopex> headius: we need a glue for ^ too in String
<enebo> we should be able to reify a single operand call back into a specialized 1O call without needing an operation
<enebo> This was just done to make interpreter simpler but we should figure out a nicer way for that check
<enebo> anyways a different issue :)
yosafbridge has quit [Quit: Leaving]
<enebo> but reason I brought it up is we are getting close to needing more than a byte to emit some instrs identity and though I think I made it encode past a byte we probably don't want that
yosafbridge has joined #jruby
<headius> sure
<headius> the flag will work fine, I just have to tweak some of the boot pipeline
<headius> enebo: I found a simpler way
<enebo> headius: ok
<headius> enebo: e29f1a339a
<GitHub87> [jruby] headius pushed 4 new commits to jruby-9.1: https://git.io/vNU8k
<GitHub87> jruby/jruby-9.1 68c66a8 Charles Oliver Nutter: Tag RLIMIT-tweaking test that crashes.
<GitHub87> jruby/jruby-9.1 2e6bbb6 Charles Oliver Nutter: Align naming.
<GitHub87> jruby/jruby-9.1 e29f1a3 Charles Oliver Nutter: Script bodies should always push the provided dynscope.
<GitHub27> [jruby] headius pushed 1 new commit to jruby-9.1: https://git.io/vNU8X
<GitHub27> jruby/jruby-9.1 8131628 Charles Oliver Nutter: Check Module type before proceeding.
<headius> ok, fingers crossed that should be green 9.1 branch
<headius> bbl, going to buy myself a Switch
<headius> ugh
<GitHub155> [jruby] headius pushed 1 new commit to jruby-9.1: https://git.io/vNU08
<GitHub155> jruby/jruby-9.1 806d858 Charles Oliver Nutter: Refactoring.
mdrip has joined #jruby
<mdrip> hey, anyone use drip? Having an issue with not getting stdout when using a preloaded drip jvm
<mdrip> Seems related to https://github.com/jruby/jruby/issues/2690 perhaps, but that was merged in for 9.1.3, i'm on 9.1.15.0 atm
<headius> hey there
<headius> does the app do something peculiar with the streams?
<headius> drip works ok but the console experience takes a hit
<mdrip> hm, i can reproduce with
<mdrip> jruby -e 'puts "foobar"'
<mdrip> i'm running rbenv, not sure if that's causing something weird to happen
<mdrip> running jruby --version confirms it's on 9.1.15.0
<headius> and Java 8 or 9?
<headius> honestly I haven't tried drip recently, so it's possible we regressed on something simple
<mdrip> jruby 9.1.15.0 (2.3.3) 2017-12-07 929fde8 OpenJDK 64-Bit Server VM 24.95-b01 on 1.7.0_95-b00 +jit [linux-x86_64]
<headius> can you open an issue with all your env stuff please?
<mdrip> Yeah, for sure
<headius> I'm just walking out the door but will be back on later
<enebo> the irony of you irc handle
<mdrip> :)
<mdrip> thanks!
<enebo> :)
dave__ has joined #jruby
<GitHub108> [jruby] marvinTD opened issue #4942: No stdout with drip on jruby-9.1.15.0 https://git.io/vNUuc
dave__ has quit [Remote host closed the connection]
dave__ has joined #jruby
dave__ has quit [Remote host closed the connection]
dave__ has joined #jruby
bbrowning is now known as bbrowning_away
dave__ has quit [Ping timeout: 248 seconds]
dave__ has joined #jruby
dave__ has quit [Remote host closed the connection]
ruurd has joined #jruby
<lopex> enebo: btw the parser should raise on "\uD8000" too
<lopex> but the duplication of logic in onigmo and mri scares me
<lopex> I guess I need to understand why truncated chars work on mri
<lopex> one working case is convert to code point and then use codeToMbcLength
<lopex> but that causes some problems too like buffer overrun exploits
<lopex> which I cannot reproduce :P
ruurd has quit [Ping timeout: 248 seconds]
dave__ has joined #jruby
dave__ has quit [Ping timeout: 264 seconds]
bbrowning_away is now known as bbrowning
dave__ has joined #jruby
dave__ has quit [Ping timeout: 264 seconds]
<GitHub190> [jruby] abelsromero opened issue #4943: Open3.open2e returns `Errno::ENOENT` in jRuby but runs fine on Ruby https://git.io/vNUbF
mdrip has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
dave__ has joined #jruby
dave__ has quit [Ping timeout: 265 seconds]
<lopex> enebo: you're not bothered ?