mister_solo has quit [Quit: So long, and thanks for all the fish!]
Integral1st has quit [Ping timeout: 264 seconds]
sferik has joined #jruby
mister_solo has joined #jruby
metadave_ has joined #jruby
Who has joined #jruby
metadave has quit [Ping timeout: 246 seconds]
nateberkopec has joined #jruby
Integral2st has joined #jruby
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sferik has joined #jruby
kl has joined #jruby
Who has quit [Ping timeout: 246 seconds]
Who828 has joined #jruby
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
rcvalle has joined #jruby
AlHafoudh has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed pull request #2458: Fixed yielded arrays of 1 element bug when using Enumerator#first (master...2376_bugfix) http://git.io/SmT9uw
JRubyGithub has left #jruby [#jruby]
JohnBat26 has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed issue #2376: Enumerator#first when yielded arrays of 1 element bug. http://git.io/rhjPfA
JRubyGithub has left #jruby [#jruby]
deobalds has quit [Quit: Computer has gone to sleep.]
nirvdrum has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] mkristian pushed 1 new commit to jruby-1_7: http://git.io/Te4k_w
<JRubyGithub>
jruby/jruby-1_7 ce87497 Christian Meier: adds some rack appliction using the new ClassPathLayout of jruby-rack
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] eregon pushed 1 new commit to master: http://git.io/klo6yQ
JRubyGithub has left #jruby [#jruby]
<JRubyGithub>
jruby/master 65ce1eb Benoit Daloze: [Truffle] Tag "slow" core specs....
sferik has joined #jruby
Who828 has quit [Quit: Who828]
sferik has quit [Remote host closed the connection]
<bbrowning>
I think I've narrowed down an issue I'm seeing to one of these interrupt calls interrupting a fiber's Queue.pop
shellac has quit [Ping timeout: 240 seconds]
kl has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] mkristian pushed 1 new commit to jruby-1_7: http://git.io/tAX23w
<JRubyGithub>
jruby/jruby-1_7 612e797 Christian Meier: [build] travis behaves differently then the local build
JRubyGithub has left #jruby [#jruby]
<headius>
hmmm
<headius>
it used to do that because it was using JVM monitor-based signalling and could potentially need to be interrupted twice if still waiting for a value
<headius>
we do pool these threads...
<headius>
perhaps add a line to end of the Enumerator thread runnable that clears thread interrupt status
kl has quit [Ping timeout: 265 seconds]
vdmo has joined #jruby
<bbrowning>
calling .interrupt() should clear the interrupt status, I think
<headius>
bbrowning: interrupt would set the status...interrupted() would clear it
<headius>
let me know what you see
<bbrowning>
headius: hmm - I thought that when an InterruptedException gets created as a result of calling .interrupt() it ends up clearing the interrupted status
<headius>
yes, but if the interrupt is spurious it leaves thread set in interrupted state until that exception happens or the state is cleared
<headius>
if there's nothing to interrupt the status remains until there is
enebo has joined #jruby
<headius>
which could cause a subsequent use of the same pooled thread to immediately interrupt
<headius>
that's my current theory
<bbrowning>
ok - I'll give it a shot
<bbrowning>
I don't see a .interrupted() method though - just interrupt() and isInterrupted()
<bbrowning>
ahh I see it's a static method - Thread.interrupted()
mitchellhenke has quit [Ping timeout: 264 seconds]
mitchellhenke has joined #jruby
sferik has joined #jruby
Hobogrammer has joined #jruby
sferik has quit [Client Quit]
yfeldblum has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/yjp5Xw
<JRubyGithub>
jruby/master 9f4d9a4 Charles Oliver Nutter: Add Windows platform masks to get JRuby + RubyGems booting again.
JRubyGithub has left #jruby [#jruby]
<nirvdrum>
Nice.
Who has joined #jruby
metadave has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
yfeldblu_ has joined #jruby
<headius>
nirvdrum: it's a start
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 3 new commits to master: http://git.io/QW_Ygw
<JRubyGithub>
jruby/master dc30c92 Charles Oliver Nutter: Modify sleep logic to be zero-alloc for reduced race potential.
<JRubyGithub>
jruby/master 697b06a Charles Oliver Nutter: Ensure sleep permits are reacquired before returning.
<JRubyGithub>
jruby/master 47bc672 Charles Oliver Nutter: Add missing Fiddle files from MRI.
JRubyGithub has left #jruby [#jruby]
yfeldblum has quit [Ping timeout: 245 seconds]
<nirvdrum>
headius: I wasn't sure what you wanted to guard out and what you wanted to get into the native libs.
<headius>
I'm guarding the ones that are broken :-)
<nirvdrum>
I think dropping native delegate helped too.
<enebo>
headius: you got license?
<nirvdrum>
headius: It might be worthwhile to have some sort of wrapper method as a fencepost you can easily grep for later. Like Platform.notYetWorkingInWindows or whatever.
<headius>
enebo: EC2 and RDP
<headius>
I got tired of waiting and IT has not gotten back to me in over a week
<headius>
I'll bill them for the Windows time on EC2
cremes has quit [Read error: Connection reset by peer]
<headius>
our win32/importer used dl but dl is gone
mitchellhenke has quit [Ping timeout: 245 seconds]
<enebo>
Seems like ffi should work but it would be nice to just use stdlib code
mitchellhenke has joined #jruby
<headius>
yeah I will probably need to pull out the parts of dl we still need
<headius>
ah-ha, that's actually where Handle came from
<enebo>
yeah
mister_solo has quit [Ping timeout: 264 seconds]
<enebo>
we sort of cheated Fiddle by using our DL impl
<enebo>
I understand why DL was broken for 64 bits or whatever so Fiddle makes sense from that angle but FFI existed. I am confused about the overlap of these libraries
baroquebobcat has quit [Quit: baroquebobcat]
<enebo>
Seems any good FFI library worth it’s salt should be able to make a compat layer for the other FFI library :)
<enebo>
Aethenelle: ^
<Aethenelle>
I concurr… it’s on my list just pretty low on it...
<enebo>
I see
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/fKQ75A
<JRubyGithub>
jruby/master f129eab Charles Oliver Nutter: Transplant bits of 'dl' library we still need for our fiddle.
JRubyGithub has left #jruby [#jruby]
<enebo>
heh
<enebo>
Aethenelle: well if what headius just did will work then it may end up lower on our list as well :)
<headius>
require 'fiddle' works, at least
<enebo>
heh
<enebo>
ship it
<Aethenelle>
can i get the short version?
<enebo>
Aethenelle: win32/*.rb files need fiddle to work
<enebo>
Aethenelle: we have always hacked fiddle to use DL (which was removed) which was written using I think jffi
<Aethenelle>
ic
<enebo>
Aethenelle: headius just merged forward DL I am guessing
<enebo>
Aethenelle: So our bandaid may live on
<Aethenelle>
i thought dl/fiddle was impl’d with jruby’s ffi… i’ll take a look while I’m bored waiting for these embedded builds to work...
<enebo>
I think all parties in all rooms agrees DL should no longer exist
<enebo>
btu you are right DL may be Ruby ffi already and not jffi
<Aethenelle>
DL shouldn’t have existed as soon as FFI was released...
<enebo>
Aethenelle: Just think DL would still be in stdlib today if it had thought about 32/64 bit architectures
<enebo>
Aethenelle: but had it then fiddle may not have been written
<enebo>
who knows…I just wish there was one
<Aethenelle>
from what i’ve seen fiddle isn’t much better than dl… at least it has decent docs in english… never thought i’d use the japanese i took at work, just anime
<headius>
enebo: I'm trying to just pull from dl what is needed, and if it's just using FFI I just use that code in fiddle now
<enebo>
headius: yeah great call
<headius>
it looks like tenderlove basically just did the same thing with fiddle when dl was removed
<headius>
probably very recently
<enebo>
headius: I wonder how much non-stdlib code uses Fiddle
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/0Bsjmw
<JRubyGithub>
jruby/master 0e8e197 Charles Oliver Nutter: More missing bits transplanted from DL.
JRubyGithub has left #jruby [#jruby]
<headius>
enebo: me too
<headius>
oh hell, thread bits got into that commit
<enebo>
headius: Numeric. has singleton_method_added which returns a Type error if hit…Make something hit it
baroquebobcat has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 2 new commits to master: http://git.io/mAZi_A
<JRubyGithub>
jruby/master e7b1ba3 Charles Oliver Nutter: Revert "More missing bits transplanted from DL."...
<JRubyGithub>
jruby/master 7b15e95 Charles Oliver Nutter: More missing bits from dl+fiddle unification.
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/vGeLmg
<JRubyGithub>
jruby/master 9be5e6a Charles Oliver Nutter: Make rest of RubyThread state volatile and make sleep more robust.
JRubyGithub has left #jruby [#jruby]
<headius>
enebo: can't?
<enebo>
headius: I do in refinements code and notice we had a bug in our definition but then I was like how the fuck can I make this trigger
<enebo>
headius: unless it is added for some internal magic or some non-immediate Numeric needs to throw
<Aethenelle>
before i forget… i seem to remember jruby automagically handling attr_accessor :foo => getFoo when accessed by Java objects … is that correct or am I making up features?
<headius>
what do you mean "when accessed by Java objects"
<Who>
hey there
erikhatcher has quit [Quit: erikhatcher]
<Who>
enebo: around?
Integral2st has quit [Ping timeout: 272 seconds]
<enebo>
Who: I did not look at PR yet
<enebo>
Who: and I am starting to make lunch. I will look after lunch. If it does not fail any tests I will probably just apply it though
<Who>
Acutally I was going to ask if there an issue or task you wanted me to help with it
<enebo>
Who is that pr not associated with an issue? You can make one if not
<enebo>
Who: just put a snippet showing the failing case
<Who>
is*
<Who>
headius has merged the PR already IIRC I think
<enebo>
oh I see
<enebo>
Who: perhaps submit a test to MRI tests?
<Who>
ah, that I can do.
camlow32_ has joined #jruby
benlovell has quit [Ping timeout: 244 seconds]
yfeldblum has joined #jruby
diegoviola has quit [Read error: Connection reset by peer]
pietr0 has joined #jruby
<Who>
enebo: I meant if there was any other issues/tasks you wanted help with before JRuby 9k is released :)
yfeldblu_ has quit [Ping timeout: 264 seconds]
camlow32_ has quit [Remote host closed the connection]
camlow325 has quit [Ping timeout: 240 seconds]
nirvdrum has joined #jruby
tcrawley is now known as tcrawley-away
<nirvdrum>
Screw you, corporate firewall.
subbu has joined #jruby
vtunka has joined #jruby
<bbrowning>
nirvdrum: wait - you're inside a corporate office now?
<nirvdrum>
bbrowning: Once a week I head into an Oracle office.
<bbrowning>
ahh
<nirvdrum>
But it blocks IRC, SSH, LastPass, and a bunch of other stuff.
<nirvdrum>
All UDP actually, I think.
<bbrowning>
socks proxy on port 443!
<nirvdrum>
I ended up setting up OpenVPN on a DD-WRT router at home and miraculously I can connect to that. Then I route all my traffic over that link.
<nirvdrum>
It has the added benefit of allowing me to connect to my desktop at home where I do all of my primary dev.
mister_solo has joined #jruby
josh-k has joined #jruby
camlow325 has joined #jruby
camlow325 has quit [Read error: Connection reset by peer]
camlow325 has joined #jruby
<headius>
enebo: I have win32 working again it appears
<headius>
have moved on to a socket error on Windows
<headius>
another native call that's unsupported (fcntl CLOEXEC)
<nirvdrum>
Aethenelle: Is that a MacOS thing? They both print [] on Linux for me.
<nirvdrum>
I've only ever seen '**' for intermediate paths.
benlovell has joined #jruby
<Aethenelle>
if I use ** it returns [] in both
<nirvdrum>
Weird. Maybe my example path is wrong. When I use '**' I get the file in both.
<Aethenelle>
I know I've used /path/to/jar/**/*.jar before successfully... it's how we usually load a clients jar files Dir.glob("/path/**/*.jar").each{|x| require x};
yfeldblum has quit [Remote host closed the connection]
<nirvdrum>
Aethenelle: Do you get different results if you disable native?
<Aethenelle>
what's that switch again?
<nirvdrum>
-Xnative.enabled=false
<nirvdrum>
bin/jruby --properties | grep native
<nirvdrum>
That's how I track these things down.
<Aethenelle>
i keep forgetting about that...
<Aethenelle>
only */*.rb works
camlow325 has quit []
<nirvdrum>
Weird. I'm out of ideas.
benlovell has quit [Ping timeout: 265 seconds]
<Aethenelle>
here's a fun one... Dir.glob "/path/to/my/file.rb" ... doesn't work with or without native
<Aethenelle>
works in MRI
camlow325 has joined #jruby
diegoviola is now known as resentido
<nirvdrum>
That works for me on Linux.
<nirvdrum>
Although it blows up in Truffle, so that needs to be fixed.
<Aethenelle>
lemme try on my mt lion box...
<nirvdrum>
Any symlinks in between? That could maybe come into play.
<Aethenelle>
... i think so... lemme try another path without...
<chrisseaton>
nirvdrum: also bin/jruby -Xnative?
<chrisseaton>
bin/jruby -Xtrufle...
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/IbdXMQ
<JRubyGithub>
jruby/master 496c638 Charles Oliver Nutter: const_defined? should not trigger autoload.
JRubyGithub has left #jruby [#jruby]
<Aethenelle>
looks like it's the link that's killing it all of the above work fine w/out the symlink
<nirvdrum>
chrisseaton: Eh?
<nirvdrum>
chrisseaton: The Truffle globbing code requires a "*" to be present.
<nirvdrum>
No big deal. We have other cases to clean up anyway.
<chrisseaton>
nirvdrum: I'm saying you don't need to use grep: -Xfoo? searches for foo and -Xfoo... searches by prefix foo
<nirvdrum>
Ahh.
<nirvdrum>
I thought you were asking a question.
<nirvdrum>
Silly syntax.
<chrisseaton>
The idea is if you start typing an option, and then realise halfway through you have no idea what it is, you can just trail off ... or ? and press enter
<nirvdrum>
Makes sense.
resentido is now known as diegoviola
<nirvdrum>
The "..." doesn't seem to work for me.
<nirvdrum>
But the "?" does.
<chrisseaton>
Does bin/jruby -Xtruffle... not print Truffle options?
<nirvdrum>
I get "jruby: warning: unknown property jruby.truffle..."
<chrisseaton>
hm I'll look into it
diegoviola is now known as resentido
<nirvdrum>
I'm taking another crack at this indy paper. So many symbols to keep track of.
DrShoggoth has quit [Remote host closed the connection]
josh-k__ has joined #jruby
josh-k_ has quit [Ping timeout: 256 seconds]
mrmargolis has quit [Remote host closed the connection]
josh-k has joined #jruby
josh-k__ has quit [Ping timeout: 264 seconds]
mister_solo has joined #jruby
viking has quit [Remote host closed the connection]
enebo has quit [Quit: enebo]
<chrisseaton>
headius: idea for an IR optimisation - overflow check removal - we've been looking at this in Truffle and it looks relatively straightforward. Annotate Fixnum dataflow edges with a mask of possible bit values.
mister_solo has quit [Ping timeout: 264 seconds]
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]