mcclurmc_ has quit [Remote host closed the connection]
mcclurmc has joined #jruby
mcclurmc has quit [Remote host closed the connection]
mcclurmc has joined #jruby
zorak8 has joined #jruby
x1337807x has joined #jruby
elia has joined #jruby
zorak8 has quit [Ping timeout: 245 seconds]
zorak8 has joined #jruby
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
skade has joined #jruby
zorak_ has joined #jruby
zorak8 has quit [Read error: Connection reset by peer]
subbu has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
<Antiarc>
https://github.com/jruby/jruby/issues/1500 - I'm running into this when attempting to debug a Java lib used in my app. Any recommendations for where to loko before I start poking around?
<Antiarc>
(I'm not quite sure what the error means)
x1337807x has joined #jruby
mitchellhenke has quit [Quit: Computer has gone to sleep.]
elia has quit [Quit: Computer has gone to sleep.]
nateberkopec has quit [Quit: Leaving...]
zorak_ has quit [Read error: Connection reset by peer]
zorak8 has joined #jruby
havenwood has joined #jruby
zorak8 has quit [Read error: Connection reset by peer]
zorak8 has joined #jruby
pgokeeffe has joined #jruby
subbu has quit [Ping timeout: 245 seconds]
mrmargolis has joined #jruby
zorak8 has quit [Ping timeout: 264 seconds]
subbu has joined #jruby
mrmargolis has quit [Remote host closed the connection]
pgokeeffe has quit [Quit: pgokeeffe]
pgokeeffe has joined #jruby
pgokeeffe has quit [Quit: pgokeeffe]
pgokeeffe has joined #jruby
<Antiarc>
auuuuuuuuuuuuuuuughhhhhhhhhhh academia has the worst commit discipline ever
<Antiarc>
Showing 119 changed files with 12,911 additions and 18,270 deletions.
<Antiarc>
*twitch*
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
subbu has quit [Ping timeout: 240 seconds]
x1337807x has joined #jruby
subbu has joined #jruby
pgokeeffe has quit [Quit: pgokeeffe]
pgokeeffe has joined #jruby
subbu has quit [Ping timeout: 264 seconds]
DrShoggoth has quit [Quit: Leaving]
metadave has joined #jruby
yfeldblu_ has joined #jruby
metadave has quit [Client Quit]
yfeldblum has quit [Ping timeout: 245 seconds]
<rtyler>
Antiarc: what's wrong with the commit message of 'changes' ^_^
<rtyler>
it's accurate!
subbu has joined #jruby
<Antiarc>
*twitch*
nirvdrum has joined #jruby
subbu has quit [Ping timeout: 255 seconds]
BadQuanta has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 1 new commit to truffle-break: http://git.io/nx479w
<JRubyGithub>
jruby/truffle-break 3eab260 Kevin Menard: [Truffle] Tag failing 'break' specs.
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 1 new commit to master: http://git.io/auQSQg
<JRubyGithub>
jruby/master 6428162 Kevin Menard: Merge branch 'truffle-break'
JRubyGithub has left #jruby [#jruby]
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum created truffle_enumerable (+4 new commits): http://git.io/ukycfw
<JRubyGithub>
jruby/truffle_enumerable 0cf0f61 Kevin Menard: [Truffle] Implemented Rubinius.single_block_arg.
<JRubyGithub>
jruby/truffle_enumerable 145ae71 Kevin Menard: [Truffle] Pulled in Rubinius's implementation of Enumerable now that Rubinius.single_block_arg works.
<JRubyGithub>
jruby/truffle_enumerable 1245eed Kevin Menard: [Truffle] Range implements Enumerable.
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum force-pushed truffle_enumerable from febbf0e to 8759803: http://git.io/rFoXxQ
<JRubyGithub>
jruby/truffle_enumerable 4080237 Kevin Menard: [Truffle] Implemented Rubinius.single_block_arg.
<JRubyGithub>
jruby/truffle_enumerable 5816c7a Kevin Menard: [Truffle] Pulled in Rubinius's implementation of Enumerable now that Rubinius.single_block_arg works.
<JRubyGithub>
jruby/truffle_enumerable e7e548c Kevin Menard: [Truffle] Range implements Enumerable.
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
jruby/truffle_enumerable 3a2ab4a Kevin Menard: [Truffle] Implemented Rubinius.single_block_arg.
<JRubyGithub>
jruby/truffle_enumerable e86ce74 Kevin Menard: [Truffle] Pulled in Rubinius's implementation of Enumerable now that Rubinius.single_block_arg works.
<JRubyGithub>
[jruby] nirvdrum force-pushed truffle_enumerable from 8759803 to 8c741bb: http://git.io/rFoXxQ
<JRubyGithub>
jruby/truffle_enumerable a2ce0ef Kevin Menard: [Truffle] Range includes Enumerable.
skade has quit [Remote host closed the connection]
skade has joined #jruby
bbrowning_away is now known as bbrowning
joast has joined #jruby
jmimi has quit [Ping timeout: 244 seconds]
jmimi has joined #jruby
kl__ has joined #jruby
<kl__>
Why does jruby support 1.9 mode by default, instead of 2.0 ?
graphex has quit [Read error: Connection reset by peer]
metadave has joined #jruby
metadave has quit [Client Quit]
benlovell has quit [Ping timeout: 255 seconds]
tcrawley-away is now known as tcrawley
yfeldblum has joined #jruby
Aethenelle has quit [Quit: Aethenelle]
yfeldblu_ has quit [Ping timeout: 245 seconds]
metadave has joined #jruby
benlovell has joined #jruby
<kares>
kl__: cause 2.0 is not fully supported ... 2.x will be polished in 9k
treehau55 has joined #jruby
mitchellhenke has joined #jruby
<treehau55>
hello, I am curious to see if someone has a workaround for the issue when you are debugging with pry in a jruby on rails app, you are unable to interact with stdin
graphex has joined #jruby
<kl__>
kares: I suspected as such - what kinda stuff isn't supported? Looking to evaluate whether I could/should enable that option
vtunka has quit [Quit: Leaving]
deobalds has quit [Quit: Computer has gone to sleep.]
<kares>
kl__: I do not know the exact details, I know there were issues with keyword arguments ... if you stick around till US wakes up you might get a better anwser
vtunka has joined #jruby
iamjarvo has joined #jruby
subbu has joined #jruby
temporalfox has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
temporalfox has joined #jruby
nateberkopec has joined #jruby
nirvdrum has joined #jruby
deobalds has joined #jruby
dumdedum has quit [Quit: foo]
treehau55 has quit [Quit: Page closed]
<cpuguy83>
Downloading the JRuby package from S3 has an SSL cert error when trying to use https, due to the "jruby.org" in the host name.
<cpuguy83>
Any way around this without ignoring ssl errors?
<headius>
Good morning!
<cpuguy83>
Good morning.
<headius>
cpuguy83: which package?
<cpuguy83>
headius: the main jruby-bin-<version>.tar.gz
<headius>
kl__: we'd actually love to have folks testing out the next version of JRuby, which supports 2.2 features
<headius>
cpuguy83: as in from the main page link?
<cpuguy83>
s3 is covered by `*.s3.amazonaws.com`, but `jruby.org.s3.amazonaws.com` violates that
<headius>
jruby.org shouldn't appear in the link if you're coming through our pages...if it does we need to fix that
<headius>
I mean shouldn't be in the address
<cpuguy83>
Ah, I pulled the URl awhile ago.
<headius>
ok
<cpuguy83>
Great, fantastic!
mitchellhenke has quit [Quit: Computer has gone to sleep.]
<cpuguy83>
Been pulling from HTTP and didn't even think about it!
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] Who828 opened pull request #2451: No need to synchronise a newly created local RubyArray (master...sync_thread_group_fix) http://git.io/VIX_YA
JRubyGithub has left #jruby [#jruby]
<headius>
cpuguy83: yeah I think someone brought this up over the summer and the URL juggling was an easy fix
benlovell has quit [Ping timeout: 245 seconds]
jrgarcia has joined #jruby
<cpuguy83>
Oh wonderful, and now rubygems is having issues with pulling bundler, invalid SSL cert.
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 6 new commits to truffle_enumerable: http://git.io/Ru5gMQ
<JRubyGithub>
jruby/truffle_enumerable 88ece10 Kevin Menard: [Truffle] Removed Array#all? since it should be defined in Enumerable.
<JRubyGithub>
jruby/truffle_enumerable 57d58cf Kevin Menard: [Truffle] Removed Array#max since it should be defined in Enumerable.
<JRubyGithub>
jruby/truffle_enumerable c36890a Kevin Menard: [Truffle] Removed Array#min since it should be defined in Enumerable.
JRubyGithub has left #jruby [#jruby]
lance|afk is now known as lanceball
<headius>
nirvdrum: some Enumerable methods are explicitly defined on Array for a reason, FYI
jrgarcia has quit [Client Quit]
<headius>
I don't remember if any of those three are in that category but I figured I should mention it :-)
temporal_ has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed issue #2450: GC Symbol http://git.io/FM4wgA
JRubyGithub has left #jruby [#jruby]
temporalfox has quit [Read error: Connection reset by peer]
graphex has quit [Read error: Connection reset by peer]
benlovell has joined #jruby
iamjarvo has joined #jruby
graphex has joined #jruby
cultureulterio-1 has quit [Read error: Connection reset by peer]
<mrmargolis>
I am soooooo excited to hear you all talking about a preview!
<subbu>
nirvdrum, that interp method that caused the crash will probably be removed from 9k-final .. that method exists only to flag problems that need fixing .. more like a sanity check .. the last one to fix was the splat warning.
Aethenelle has joined #jruby
<headius>
subbu: ahh interesting
<headius>
I suspect the problem is that the main script exits, which tears down the runtime, which closes IO streams, which are being used in the animation loop
Aethenelle has quit [Client Quit]
<headius>
I'm not sure how best to do that...main script exits, main script exits...the program is supposed to be done at that point
<subbu>
headius, but is that specific to 9k though?
_elia has joined #jruby
<headius>
1.7 did not close stdio streams on teardown
<headius>
yeah that's it...if I add a gets to his script it's fine
<subbu>
ah .. ok :)
jrgarcia has joined #jruby
mitchellhenke has joined #jruby
elia has quit [Ping timeout: 240 seconds]
deobalds has quit [Quit: Computer has gone to sleep.]
<nirvdrum>
To my script?
havenwood has joined #jruby
<subbu>
nirvdrum, script from issue 1917
<headius>
ya
<nirvdrum>
Gotcha.
<headius>
hmm
subbu is now known as subbu|away
<headius>
bugger, this is a tricky one to compromise on
<headius>
all streams register a finalizer that we run on teardown
<headius>
stdio streams are disabled but not really "closed" at descriptor level, but that's enough to make them unusable after main script exits
jrgarcia has quit [Read error: Connection reset by peer]
<headius>
I almost feel like the script should call something to say "when main script exits please don't assume I'm done
<headius>
JRuby.auto_teardown = false
<headius>
or something
<headius>
enebo: any thoughts?
<enebo>
headius: So they need them to be open for finalizers or just never close because other things may reconnect?
<headius>
the latter
<headius>
in this case it's a message pump thread that lives beyond end of script
<headius>
e.g. Swing
mcclurmc_ has joined #jruby
<enebo>
headius: well it is non-normal Ruby behavior so I guess we can solve it via extension
<enebo>
headius: but I wonder how specific we should be
<enebo>
headius: like JRuby::ShutdownHooks::close_stdio = false
<headius>
right
<headius>
my quick thought was to just disable the call to teardown when main script exits
<enebo>
headius: I think perhaps being a little more granular might be right since they may make an at shutdown hook for themselves and expect that to work
<headius>
since in this case, you really need to control teardown and do it at actual program exit
<enebo>
Author of processing script and author of processing-core are different people with different expectations
<headius>
and they may not expect that to work too
<headius>
I'm not sure we can make all the people happy on this one
<headius>
they may expect at_exit to be at program exit, not main script exit
<enebo>
but we could just expose the minimal amount needed for this to start working
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 1 new commit to truffle_enumerable: http://git.io/jYPhPQ
JRubyGithub has left #jruby [#jruby]
<JRubyGithub>
jruby/truffle_enumerable f41b133 Kevin Menard: [Truffle] Untagged a bunch of passing Enumerable specs.
mcclurmc has quit [Ping timeout: 245 seconds]
<enebo>
yeah I guess anything use supplied I would think they would expect to always run at some point
<enebo>
which point is open for debate I guess
<enebo>
but the stdio closing is an implicit action which happens at mani script closing
<headius>
hmmm
<headius>
we may have a compatible way to do this
<headius>
there's already a hardExit flag on RubyInstanceConfig
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 11 new commits to master: http://git.io/6oZccQ
<JRubyGithub>
jruby/master ebccec6 Kevin Menard: [Truffle] Implemented Rubinius.single_block_arg.
<JRubyGithub>
jruby/master f61c2c6 Kevin Menard: [Truffle] Pulled in Rubinius's implementation of Enumerable now that Rubinius.single_block_arg works.
<JRubyGithub>
jruby/master ab0256a Kevin Menard: [Truffle] Range includes Enumerable.
JRubyGithub has left #jruby [#jruby]
<headius>
hmmm
<headius>
maybe not
<headius>
it doesn't prevent tearDown at end of script, it just prevents registering a JVM-level teardown
<headius>
but it could (and maybe should) mean "don't teardown this runtime when my script exits"
<enebo>
headius: but we must have somethiing which does the right thing for embedding
<enebo>
headius: Or perhaps tht is because it bypasses main in embedding use
<headius>
embedding doesn't autoteardown
<enebo>
headius: which does beg a question
<enebo>
headius: Is processing an embedding application
<headius>
oh I remember another reason we wanted to close stdio on teardown: flushing streams
<headius>
we've had those bug reports before
subbu|away has quit [Ping timeout: 265 seconds]
<enebo>
headius: so I wonder if perhaps we can help them set this up through embedding for 9k
<headius>
enebo: yeah, I think it's basically an embedded app that's run from command line via a Ruby script
<headius>
hmm
<headius>
what do you mean? Like the Ruby script would launch a second runtime?
<headius>
there's no normal Java entry point for this thing
cultureulterior1 has quit [Read error: Connection reset by peer]
vtunka has quit [Quit: Leaving]
cultureulterior1 has joined #jruby
<headius>
I keep thinking of magic ways to avoid shutting down the streams, like checking if any non-daemon threads are running...but I don't like the magic part
<headius>
that works for 1917... JRuby.runtime.instance_config.hard_exit = false
<headius>
a bit wordy, unfortunately
<headius>
that could just be JRuby.hard_exit =
<enebo>
headius: I meant make something using ScriptingContainer and not have them call through Main
<headius>
it's a Ruby script
<enebo>
yeah
<enebo>
headius: but it is launching a org.jruby.MAin from jruby-jars right?
<headius>
I guess I am confused how you would run a Ruby script from command line without going through Main
<headius>
no
<headius>
it's a Ruby script :-)
<headius>
it just runs at command line
<headius>
at least that's how he showed the reproduction
<headius>
if they were going through ScriptingContainer in his actual deployment then it shouldn't have torn down
<enebo>
headius: but then I am confused why he needs things to still be alive then?
jmimi has joined #jruby
<headius>
because a backgrond thread for message pump of the app starts up
<headius>
the main script exits after it sets the graphics in motion
<enebo>
oh this is only because that thread is not done
<headius>
yeah
<headius>
or rather, because we make it be done by killing the streams on main exit
<headius>
YOU'RE DONE...*BOOM*
tenderlove has joined #jruby
<enebo>
hmmm thinking bout daemon threads
<enebo>
so the message pump is swing itself right?
<headius>
or something
<headius>
I assume it's Swing
<enebo>
hmm
diegoviola has joined #jruby
<headius>
I dislike users having to add something to existing working scripts to have them work in 9k but I don't have a better solution
<headius>
we are more consistent now wrt stream shutdown on script exit, but that consistency isn't necessarily backward-compat
subbu|away has quit [Ping timeout: 264 seconds]
<enebo>
I still don’t even understand the issue. Should the script be done and shutown is stumbling because something is closed before something else is done thinking it has finished?
<headius>
the main script exits, which triggers runtime teardown; because the script uses stdio in code that runs in message pump, message pump raises an exception and terminates
<enebo>
Or is this like a launcher which starts a second thread but then runs off the end of the world and shuts down even though there is still a live thread?
<headius>
it's basically that, yes
<headius>
but it's a thread started implicitly by swing or whatever
<enebo>
but the app is still running
<headius>
Ruby threads would end when the script ends, but non-daemon threads in the rest of the JVM can keep it running and have a torn-down runtime in hand
<headius>
right
<enebo>
It is just that main script finishes
<headius>
we did tear down before, just not stdio
<enebo>
yeah this to me feels like ramification of daemon threads
<headius>
so you ended up with a half-torn-down runtime
<headius>
now you end up with a fully torn down runtime after main exits
<enebo>
If we provided non-daemon option then perhaps shutdown could decide not to run because it knows something is still running
<headius>
except that we don't even start this thread...we could mine our registered RubyThread instances for non-daemon, but again that's really magic
<headius>
so you start a thread pool you're not using and suddenly we stop flushing streams on exit
<headius>
things like that make me want an explicit option
<enebo>
but how does Swing handle this?
<headius>
well, if you were to shut down streams in your Java main, it would have exactly the same effect
<enebo>
unless this is just something uncommon done in their code?
<headius>
the problem is that our lifecycle expects that when the main script exits, you're done with the runtime
<headius>
when run from Main
<headius>
that's a good point though
<enebo>
so basically it is us explicitly killing streams whereas Java lets JVM finalization deal with it
<headius>
or doesn't bother
<headius>
most IO channels don't finalize...I think only files do
<enebo>
perhaps when from Main we don’t bother killing them and let JVM try
<headius>
hmm
<headius>
that would require not running all finalizers, or not registering stdio for teardown
<headius>
JVM does not run finalizers on exit though, so we may end up with stdio not flushing again
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 2 new commits to master: http://git.io/AL5qdA
<JRubyGithub>
jruby/master 609d838 Kevin Menard: [Truffle] Dir and IO should include Enumerable.
<JRubyGithub>
jruby/master 09552e5 Kevin Menard: [Truffle] String and Symbol should include Comparable.
JRubyGithub has left #jruby [#jruby]
<enebo>
headius: I think that is simple to not run though
<headius>
well, all I have in the teardown loop is a Finalizable object
skade has quit [Quit: Computer has gone to sleep.]
<enebo>
headius: I am more concerned about behavior
<enebo>
headius: we can just add field to the finalizable object which runs it but does not close
<headius>
hmm
<enebo>
headius: but assume for the moment we can do that what could go right or wrong? :)
<headius>
I'm trying to remember all the reasons I made stdio shut down like MRI
<enebo>
headius: I assume Java does adequately shutdown properly written Java apps
<headius>
all I can remember offhand is flushing and consistency with the rest of teardown
<headius>
heh, actually it doesn't do much for shutdown
<enebo>
headius: Improper might be keeping some thread running which does not stop because IO is still open
<headius>
no finalization, no flushing streams
mitchellhenke has quit [Quit: Computer has gone to sleep.]
<headius>
enebo: what was that last thing? my machine rebooted
<enebo>
heh
<enebo>
enebo: headius: Improper might be keeping some thread running which does not stop because IO is still open
<headius>
it would be consistent with 1.7 behavior, but yes, that's a possibility
skade has joined #jruby
<headius>
we never had clear contract about what happens at script exit
<enebo>
headius: I guess for org.jruby.Main execution I can only think we explicitly kill stdio because we were coping with external RubyThreads in some way
<enebo>
but it does make me wonder how we even make daemons at this point as a Ruby script with this behavior
<headius>
yeah
mitchellhenke has joined #jruby
<headius>
bleh, I'm leaning toward keeping them open now too :-(
<enebo>
That last sentence has me really questioning our actions
<headius>
I don't like that it goes back to half teardown
<enebo>
headius: well it defers teardown to Java
<headius>
hmm
<headius>
actually, that's an interesting thought
<enebo>
headius: because we cannot know if some Java lib called by Ruby is still usnig stdio
<headius>
it could *just* register runtime exit hook
<headius>
rather than teardown in finally
<headius>
then as long as the JVM started for main is running, that runtime is alive
<enebo>
headius: I guess I was confused above because that was what I thought it was doing :)
<headius>
no, we have a finally in Main
<enebo>
headius: but you sid finalization so :)
<headius>
we do both
<headius>
we do all nine
<headius>
hmm
<headius>
this could work
<enebo>
yeah I knew some stuff was finalized but we tried to minimize that when we wanted to control when
anaeem1 has quit [Remote host closed the connection]
<headius>
that would appease my desire to have teardown fully tear down, but still allow command-line scripts to have a runtime that lives past the end
<headius>
only concern now would be how bad it is for Runtime exit hooks to be running full-on Ruby code
<headius>
as is done for some test frameworks
<headius>
if the runtime's still 100% I think it should be ok...
<enebo>
hmm
<enebo>
I guess we always run this one last :)
<enebo>
exit hooks are LIFO?
<headius>
JVMs or ours?
<headius>
I don't know if it's specified what order JVM runs exit hooks
<enebo>
outs
<enebo>
ours
mitchellhenke has quit [Quit: Computer has gone to sleep.]
<headius>
dfr|work: great! feeling good about a 9k preview very soon
<dfr|work>
headius, :D
<dfr|work>
headius, looking forward to it!
<headius>
me too...gotta kick this baby out the airlock
DrShoggoth has joined #jruby
balo has joined #jruby
mitchellhenke has quit [Quit: Computer has gone to sleep.]
mitchellhenke has joined #jruby
frobs has quit [Read error: Connection reset by peer]
<headius>
enebo: going with it
<headius>
it looks good and I feel ok
<enebo>
headius: ok
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/Zyr2qA
JRubyGithub has left #jruby [#jruby]
<JRubyGithub>
jruby/master 3a19c89 Charles Oliver Nutter: Do not allow finalization to close non-autoclose streams....
mrmargolis has joined #jruby
<headius>
chrisseaton: who should I talk to about trying out more advanced JIT features in graal like partial escape analysis?
<headius>
I feel like I'm not getting those features just running normally
<headius>
or something's still preventing EA even in simple cases
baroquebobcat has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed pull request #2405: Module#include is public for a while now. (master...include-should-be-public) http://git.io/x1UQTw
JRubyGithub has left #jruby [#jruby]
<chrisseaton>
headius: send me a test case, and I'll see if I can understand it, if not I'll pass it on
<chrisseaton>
we were going to start testing JRuby+Indy on Graal as well - what benchmarks were you running and seeing lower performance?
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius opened issue #2452: Add salt to Array#hash http://git.io/4cG4FQ
JRubyGithub has left #jruby [#jruby]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed pull request #2451: No need to synchronise a newly created local RubyArray (master...sync_thread_group_fix) http://git.io/VIX_YA
JRubyGithub has left #jruby [#jruby]
iamjarvo has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to jruby-1_7: http://git.io/UwmZnA
<JRubyGithub>
jruby/jruby-1_7 d99bfd9 Charles Oliver Nutter: Merge pull request #2451 from Who828/sync_thread_group_fix...
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] eregon pushed 1 new commit to master: http://git.io/KlgYZg
<JRubyGithub>
jruby/master 4c064f6 Benoit Daloze: Fix the build (missing imports).
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 2 new commits to master: http://git.io/wciDqg
<JRubyGithub>
jruby/master 01a832b Charles Oliver Nutter: Merge remote-tracking branch 'origin/jruby-1_7'...
<JRubyGithub>
jruby/master 6ffbfc9 Charles Oliver Nutter: Merge remote-tracking branch 'origin/master'
JRubyGithub has left #jruby [#jruby]
mjc__ has quit []
mjc_ has joined #jruby
eregon has quit [Remote host closed the connection]
eregon has quit [Remote host closed the connection]
eregon has joined #jruby
<mjc_>
wow, our test suite is 1min15s on 9k (745848c), vs 1 min 8 sec on MRI, used to be far slower on jruby (getting 1.7.18 installed right now to retest)
<headius>
hey that's great to hear
<headius>
enebo: ugh
<mjc_>
including booting the JVM and all that
<headius>
enebo: is that push failing?
<enebo>
headius: printing out the hash
<enebo>
headius: but just jruby lib/ruby/stdlib/erb.rb to confirm
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<headius>
hmm
<enebo>
headius: I am trying to land my rebased commits from this weekend
<enebo>
which of course makes sense if you consider the code
<enebo>
although I am mildly surprised nil is not byte[] {}
<headius>
bleh, I realized now there's other problems with that commit and I didn't review it well
<headius>
I will revert and work with him to fix it
<enebo>
headius: ok
<enebo>
mjc_: please verify we are actually running your whole test suite. You never know...
mitchellhenke has quit [Quit: Computer has gone to sleep.]
<mjc_>
enebo: hah, coverage report looks the same when I turn on --debug
<mjc_>
I'll double check though :)
<enebo>
mjc_: I have not timed long cold runs like major test suites so it is possible we end up warming up internally better over time than 1.7
kl__ has joined #jruby
<enebo>
headius: ah yeah that code is really messed up
<mjc_>
hah weird that wasn't it
<mjc_>
it's actually slower than 1.7.18
<mjc_>
I wonder what version I was on when it was 2x slower than MRI
<mjc_>
57 seconds on 1.7.18, 1 min 8 sec on MRI 2.2.0, and 1 min 15 sec on 9k
<enebo>
mjc_: we will work on improving 9k startup between preview and final but it is still slower than 1.7 in nearly everything I have ran
Specialist has quit [Ping timeout: 265 seconds]
<enebo>
mjc_: fwiw I have noticed that MRI keeps starting up slower and slower
<mjc_>
there was a point where this test suite was about 2 min 30 sec on 1.7.something
<mjc_>
could also be that the tests have improved
subbu has joined #jruby
<enebo>
still cool we are in the general ballpark in both releases re mri
mitchellhenke has joined #jruby
<mjc_>
indeed
<mjc_>
these tests are super ugly too, in terms of the kinds of things they do
<mjc_>
lots of objects created, both big and small
multibot_ has quit [Read error: Connection reset by peer]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 3 new commits to master: http://git.io/u7PAdQ
<JRubyGithub>
jruby/master e435c21 Teemu: Add message for implicit type conversion error
<JRubyGithub>
jruby/master c400f2d Teemu: Make String#start_with? and #end_with? raise implicit type conversion errors
<JRubyGithub>
jruby/master 0d4e9f7 Charles Oliver Nutter: Merge branch 'string_end_with_implicit_conversion' of https://github.com/lumeet/jruby into lumeet-string_end_with_implicit_conversion...
JRubyGithub has left #jruby [#jruby]
multibot_ has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/mr2t4A
<JRubyGithub>
jruby/master 5e46bfc Charles Oliver Nutter: Revert "Merge pull request #2443 from Who828/master"...
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius reopened issue #2437: Array hashing is in Ruby 1.8 mode http://git.io/xjjyBg
JRubyGithub has left #jruby [#jruby]
elia has joined #jruby
benlovell has quit [Ping timeout: 265 seconds]
elia has quit [Client Quit]
_elia has quit [Ping timeout: 256 seconds]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] enebo pushed 3 new commits to master: http://git.io/Bqf78Q
<JRubyGithub>
jruby/master c34f31c Thomas E. Enebo: Convert all instrs to store the operands and result to instr class....
<JRubyGithub>
jruby/master c9faa41 Thomas E. Enebo: Result instr has common base class to help save memory
<JRubyGithub>
jruby/master 6526643 Thomas E. Enebo: linearized BBs now a primitive array
JRubyGithub has left #jruby [#jruby]
<headius>
enebo: measurable improvement on anything?
<headius>
commits look good
<enebo>
headius: we use about 1.5M more now
<enebo>
perhaps more but no improvement
<enebo>
just simpler code
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 4 new commits to master: http://git.io/ycAitQ
<JRubyGithub>
jruby/master 3dc06a8 Teemu: Replace array flattening with manual handling in java_import
<JRubyGithub>
jruby/master 7d29f53 Charles Oliver Nutter: Merge branch '2324' of https://github.com/lumeet/jruby into lumeet-2324
<JRubyGithub>
jruby/master 732de48 Charles Oliver Nutter: Fix formatting.
JRubyGithub has left #jruby [#jruby]
<enebo>
I think some of our startup time is algo complexity and temp object creation is not it
<headius>
ok
<headius>
yeah, wouldn't be surprised at this point
<enebo>
opttempvars is like 0.14s for rails hoh of 5s
<headius>
so the risk of not doing it is having large methods with way too many tmps
<headius>
trade algo for alloc
<enebo>
but I replaced it so it did not allocate iterators or make a list for the iterable+iterator…after it tooks 0.14s
<enebo>
headius: well I think I can reduce temp vars a lot during IRBuild which should have very little cost
<headius>
well that's something :-)
<headius>
enebo: more explicit bounding of "really temporary" uses would help
<headius>
like explicitly saying "ok, I'm done with this one"
<headius>
I do that in 1.7 and 9k JITs
<headius>
less in 9k because less stack juggling at JIT level
camlow32_ has joined #jruby
<enebo>
headius: ah I meant to say that I think we can potentially not even have a opttempvars pass if we do it during IRBuild time since many uses I think can be determined at build time
<subbu>
headius, rb tree is about 50% slower in 9k compared to 1.7 with -X+C and compile.invokedynamic (with java 7 on my laptop)
<headius>
subbu: ok, worth profiling
<enebo>
I am just not positive all good cases can be done since we recurse through build and buildXXX methods
<headius>
subbu: throw it in an issue perhaps
<headius>
enebo: right, need some kind of temp var registry
<enebo>
headius: well not what I was thinking at least not as a first attempt
camlow325 has quit [Ping timeout: 256 seconds]
<subbu>
ok.
<enebo>
we make copy(temp, operand), instr(temp) but this can in several cases simple be built as instr(operand)
<enebo>
this is the first big loop in opttempvars
shellac has quit [Ping timeout: 252 seconds]
<enebo>
not sure 100% of those case be simplified during build time though
iamjarvo has joined #jruby
tenderlove has quit [Quit: Leaving...]
camlow32_ has quit [Remote host closed the connection]
<enebo>
lunch
Athenelle has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 2 new commits to jruby-1_7: http://git.io/-LZy-Q
<headius>
enebo: that looks like valid failures from your change
<headius>
I'm checking other builds since I'm on a PR quest today but travis is still behind
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] Who828 opened pull request #2453: Made changes to RubyArray hash to not use PerlHash/SipHash (master...master) http://git.io/BEfotg
<headius>
I'm cancelling a bunch of builds to get caught up
mister_solo has quit [Client Quit]
<headius>
there's half a dozen with the bad hash PR in there
<headius>
asarih: does "skip ci" apply to the push or just that commit?
<headius>
i.e. if some commits in a push have it and some don't, what happens?
anaeem1 has quit [Remote host closed the connection]
<Antiarc>
Is there an existing solution for providing a Ruby repl in the browser? ie, I want to just provide a repl into a running ruby instance through the browser. It's not a hard task, but it seems like something someone would have done already
<headius>
so anything that really wants a LocalVariable would fail there
<enebo>
yeah this is actually just broken in general
<headius>
enebo: do you still need anything from my memory-opto branch?
<enebo>
the instr is specifically for saving binding vars but we end up simplifying it to a templocalvar
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 2 new commits to master: http://git.io/HSndYg
<JRubyGithub>
jruby/master 021ce5a Kevin Menard: [Truffle] Initial implementation of MatchData#begin....
<JRubyGithub>
jruby/master bcbe444 Kevin Menard: [Truffle] Initial implementation of MatchData#end....
JRubyGithub has left #jruby [#jruby]
<enebo>
headius: I still only see a single commit to that branch
<headius>
yeah I mean have you taken everything you need
<enebo>
headius: sorry. I mean I thought you told me you have more than one commit on it
<enebo>
headius: but I only see one
<headius>
I was not going to merge it since you were working on stuff
<enebo>
headius: just making sure tht is all
<headius>
assuming you'd just pick up the changes
<headius>
yeah that's all
<enebo>
headius: leave it a bit longer
<headius>
just tidying up some local branches
<headius>
ok
<enebo>
headius: I never evaluated up the analysis cleansing
<headius>
ok
<enebo>
headius: I was trying to reduce space without touching that but I need to trace through it to see if we should not naturally be throwing some data out
nanoyak_ has joined #jruby
nanoyak has quit [Read error: Connection reset by peer]
<headius>
right, and I didn't consider that part of my patch to be pushable work anyway
<enebo>
headius: It is possible we do LVA or something earlier and then end up needing to redo LVA anyways
<headius>
nor the post-JIT nulling
<enebo>
headius: most of the gains was from changing memory of directed graph
<enebo>
headius: I can mostly only reduce GC churn there now but memory savings is done
<enebo>
CFG storage is a dirty tree I have which does some more elimination of collections
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius deleted test-mri-22-tests at eeae6ef: http://git.io/SruvHw
<enebo>
It is an interesting branch because I noticed we can do some bulk relocations instead of instr at a time
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum deleted truffle_enumerable at f41b133: http://git.io/Wm5uGg
JRubyGithub has left #jruby [#jruby]
<enebo>
hmmm I must have introduced this bug somewhere…there is no way this instr should ever have that value there
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 2 new commits to master: http://git.io/uI9lCQ
<JRubyGithub>
jruby/master 12abf73 Chris Seaton: [Truffle] Fix all medium findbugs warnings.
<JRubyGithub>
jruby/master 252fda5 Chris Seaton: [Truffle] Fix div by zero in Fixnum#%
JRubyGithub has left #jruby [#jruby]
<enebo>
ah I bet this is a special case for simplifyOperands
Specialist has joined #jruby
<headius>
enebo: where is it failing? I didn't get as far as a real stack
<enebo>
oh it is just that we are simplifying these operands but for this instr we never want them to
<enebo>
I solved load/storelocalvar
<enebo>
but I missed a few of these exceptions for specializeOperands which should not be. Especially the base case where we have specializeOperands be {}
<enebo>
by default it now runs in those
<headius>
ahhh ok
<enebo>
interpreter could have found all these problems as well had we run all these passes
<enebo>
and the debugging would have been a little bit easier
<headius>
enebo: I had another thought about clone...perhaps it should return self by default, and then a bunch of boilerplate in operands that are singletons would go away
<headius>
I mean, by default Operand has no mutable state, so clone should just return this
<headius>
if it does that, the other immutable operands don't need to override
<enebo>
headius: I think we defaulted so that new operands would not assume that
<enebo>
headius: but you are right in that we do not need to clone many things
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<headius>
just trying to find ways to reduce mostly-boilerplat operands and instrs
<enebo>
headius: fwiw most of them already don’t have clone methods
<headius>
there's several that should be singletons and aren't too
<enebo>
headius: ImmutableLiteral returns self
<headius>
well not several
<headius>
a few I've seen that are left
<enebo>
I know boolean is not
<enebo>
and should be but there is so few instances I did not bother yet
<headius>
any operand that has a singleton should have a private constructor too
<headius>
boolean has singleton references but still public constructor
<enebo>
headius: The bigger issue is operands which record scope
<headius>
bigger in what wat?
<headius>
way
<enebo>
headius: it must be cloned so it eats memory
<headius>
ahh, sure
<enebo>
headius: If we can eliminate that reference then they can all be shared
<headius>
we have boiled down direct scope references to CurrentScope, haven't we?
<enebo>
headius: Some of this is specifcally for inlining as well which makes removing it less clear
<enebo>
we have all those down to nothing now
<enebo>
we create 5 by default and rails console never needs more than 5
<enebo>
so 5 instances ever
<enebo>
same for the temp version
<enebo>
also for scopemodule + tempscopemodule
yfeldblum has joined #jruby
yfeldblum has quit [Remote host closed the connection]
<Athenelle>
[off] headius: fwiw, #2452 should be fixed quickly. I think there are some cases where a developer is playing with napalm while pointing a bazooka at their foot where this could be exploitable.
mitchellhenke has joined #jruby
kl__ has quit [Ping timeout: 264 seconds]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 1 new commit to master: http://git.io/FLvjFQ
<JRubyGithub>
jruby/master 0e46662 Kevin Menard: [Truffle] Removed compilation guards....
<JRubyGithub>
[jruby] headius closed pull request #2221: Preserve ThreadGroup#list order with WeakIdentityLinkedHashSet (master...preserve-thread-group-list-order) http://git.io/V_ud_Q
JRubyGithub has left #jruby [#jruby]
DomKM has quit [Read error: Connection reset by peer]
locks has quit [Read error: Connection reset by peer]
mje113__ has quit [Read error: Connection reset by peer]
n1ftyn8 has quit [Read error: Connection reset by peer]
lopex has quit [Write error: Connection reset by peer]
mswart has quit [Remote host closed the connection]
Guest81780 has quit [Ping timeout: 265 seconds]
mswart has joined #jruby
jmimi has joined #jruby
gazarsgo has quit [Ping timeout: 244 seconds]
dfr|work has quit [Ping timeout: 265 seconds]
havenwood has quit [Remote host closed the connection]
mje113__ has joined #jruby
mccraig has quit [Read error: Network is unreachable]
fidothe has quit [Read error: Connection reset by peer]
n1ftyn8 has joined #jruby
havenwood has joined #jruby
locks has joined #jruby
<enebo>
“21st Amendment Brewing in San Francisco has a beer called Hell Or High Watermelon, a seasonal watermelon-flavored wheat beer that is perfect in summer”….bleh
<enebo>
AHAHA jesus..there are so many good beers in CA
gazarsgo has joined #jruby
fidothe has joined #jruby
AckZ has quit [Ping timeout: 243 seconds]
mccraig has joined #jruby
<lopex>
enebo: what kinds ?
AckZ has joined #jruby
flavorjones has quit [Ping timeout: 244 seconds]
<enebo>
lopex: what kinds of good beer in California?
<lopex>
yes
dfr|work has joined #jruby
<lopex>
what type ?
<enebo>
lopex: all types. It is difficult to fathom how many breweries are in the US and how much experimentation
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] nirvdrum pushed 1 new commit to master: http://git.io/U4_INg
<JRubyGithub>
jruby/master 4e86fcd Kevin Menard: [Truffle] Implemented MatchData#to_s.
JRubyGithub has left #jruby [#jruby]
<enebo>
lopex: but mentioning Anchor and 21st Amendment (which I do actually like) and not Russian River or Stone is pretty weird. Perhaps it is just that those breweries already get too many accolaides
<JRubyGithub>
[jruby] nirvdrum pushed 1 new commit to master: http://git.io/pfW6KA
<JRubyGithub>
jruby/master 9fa7d9a Kevin Menard: [Truffle] Whoops. We need to make a copy here since Strings are mutable and changes shouldn't be shared.
JRubyGithub has left #jruby [#jruby]
<headius>
enebo: I confur
<enebo>
HAHA
<enebo>
headius: I will take that as agreement? :)
<headius>
it's better than concurring
electrical has joined #jruby
<enebo>
headius: I still like your story of being at the beer festival and 21st only had that there :)
<nirvdrum>
In your next talk you should go on about confurrent programming. Kinda like the game in Super Troopers. And just see if anyone calls you out on it.
<enebo>
if you search ipa beer you get more but then to level our lager is difficult because all beer articles explaining beer mention ale and lager so lager is inflated
<lopex>
enebo: ^
<jeregrine>
and it doesn't seem to send sentry errors
<enebo>
lopex: this seems reasonable to showing IPA’s growth. It would be cool to compare with Lager but lager is to general to work since it pulls in lots of descriptive articles about beer
<enebo>
lopex: stouts have been increasing due to increase in craft brewing but IPA is becoming a mandatory offering if you are a craft beer maker
<lopex>
Antiarc: but xIPAs can be vastly different
<Antiarc>
I know they can, I just really don't like bitterness as a feature
<Antiarc>
I like stouts because they tend to be really layered and complex, IPAs can have nice flavors but most just taste like bleach to me
<Antiarc>
And that's coming from a scotch drinker!
yfeldblu_ has quit [Remote host closed the connection]
elia has quit [Client Quit]
<Antiarc>
I'm not complaining that they exist, just that they consume the vast, vast majority of the popular beer market right now
<headius>
interesting...maybe you taste hops differently
<headius>
like some people think cilantro tastes like soap
<Antiarc>
It's entirely possible. I don't begrudge anyone else of their IPAs, but I've yet to find one that I've finished and thought "man, I'd like another of those"
<lopex>
headius: yeah it's very differentiated throughout population
<lopex>
most ppl dont like the bitterness
<lopex>
I never had enough
<lopex>
lol, maybe I dont taste it!
<lopex>
bitterblindness ?
<Antiarc>
hah. That would be quite funny.
<lopex>
well, I can tell the difference
<Antiarc>
I like bitter - I prefer my coffee black jet-fuel grade - but I don't like it as the dominant feature of a drink
<headius>
I don't like black coffee at all but I love IPAs
<lopex>
60ibus an 90ibus is a difference for me
<Antiarc>
And IPAs seem like this massive race to the bottom to see who can stuff to most IBUs into a bottle
<headius>
well, I guess that's not entirely true...a good cup of coffee isn't supposed to be bitter
<headius>
but they are 99% of the time
<lopex>
Antiarc: and I love that :)
<Antiarc>
Yeah, a good roast and a good brewing process can eliminate a *lot* of bitterness
<Antiarc>
If you haven't tried it, the aeropress produces the smoothest cup of coffee I've ever had
<headius>
yeah, the hoppiest beers stop tasting bitter to me, they just start to taste like hop fruit or something
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<headius>
Antiarc: I think the only cup of black coffee I've ever liked was from Blue Bottle in SF
<headius>
they're meticulous
<Antiarc>
Heh, that's world-class stuff
<headius>
ahh, you know it
<Antiarc>
Yup
<headius>
easily the best coffee I've ever had
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<headius>
I'm just not patient enough to buy one most days I'm in SF
<Antiarc>
Aeropress gets about 90% of the way there with almost no effort
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 1 new commit to master: http://git.io/9wmbxQ
<JRubyGithub>
jruby/master bc23c47 Chris Seaton: Revert "[Truffle] Removed Array#{inject,reduce} since they should be defined in Enumerable."...
JRubyGithub has left #jruby [#jruby]
<Antiarc>
The brewing method doesn't really give the water enough time to extract too much bitterness from the beans, so you can end up with a super nice cuppa
<headius>
nice
<headius>
enebo: yield perf regressed since last week
<headius>
I went from 8s down to 4
<headius>
and I'm seeing a LOT more bindings and frames being created
<Antiarc>
Is there any kind of formal speed regression testing in place?
<Antiarc>
Seems like that might be valuable
<enebo>
headius: sorry was on phone
<enebo>
headius: interesting re regression
<headius>
yeah I'm looking at alloc profile now
<headius>
Antiarc: chrisseaton has some he's running
<chrisseaton>
yeah we have about 41 medium sized benchmarks running on a server - we don't bench non-truffle very commit but I try to run it a few days a week
<chrisseaton>
we're a bit low on CI servers - you buy more but then you just create new jobs to fill them up!
<Antiarc>
I'm trying to think what the best way to formalize that would be. You'd need some kind of baseline benchmark to compare against to account for server load or general wonkiness, I guess?
<headius>
enebo: christ, nevermind...rvm switched me to 1.7.0 when I wasn't looking
<headius>
in other news, 1.7.18 is 2x faster at this bench than 1.7.0
<enebo>
headius: ok well good to know
<headius>
I always forget it switches immediately when you install
<headius>
mpapis: seems like a misfeatre to me
<enebo>
headius: I have seen some things that 8 is way better than 7 like instanceof chains
<chrisseaton>
Antiarc: that's easy enough to do - we do very extensive warmup and steady state detection and the lines on the graph are very stable in practice so it's easy to catch regressions
<headius>
not too surprising
<enebo>
headius: It is possible we are using some pattern 8 is doing well with an we can change that
<enebo>
headius: if we can figure out where :)
<Antiarc>
chrisseaton: nice. Where's that code? I'd love to help contribute to it when I can
<Antiarc>
(performance regressions are a particular annoyance of mine)
<chrisseaton>
it's all based on jruby/bench9000, and we run it using buildbot
DrShoggoth has quit [Quit: Leaving]
<enebo>
Antiarc: I respect your brave stance on IPA v Stout…You came out! (I am big fan of IPA but stouts/porters come in a somewhat close second)
<enebo>
Antiarc: IIPA’s also due lots of the same crazy shit big stouts do with barrel’s and extra adjuncts, but I have eeb shying away from them lately
<mpapis>
headius, make your prompt show you your ruby, something like: ${GEM_HOME##*/}
<headius>
mpapis: boo
<headius>
my prompt is already too big
<mpapis>
headius, there is unicode option in rvm-prompt
benlovell has joined #jruby
<mpapis>
:D
<headius>
I knew you were going to go there
<headius>
we need a Jay Ruby emoji
<chrisseaton>
has anyone seen mkristian recently? is he on holiday or something?
<mpapis>
headius, hard to tell why rvm picked different version
<headius>
mpapis: I just did install jruby-1.7.0 and forgot it switches
<headius>
it was my fault for not remembering, or rvm's fault for doing that when I won't remember
<mpapis>
headius, ah install switches, it's the default since it's what most of the time happens ... you could avoid it with: command rvm install
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed issue #2442: yield is slower than MRI in 9k http://git.io/w-2QQw
JRubyGithub has left #jruby [#jruby]
<enebo>
mpapis: Could (or maybe this has changed since I made my alias) rvm-prompt v show ‘system’ when system and not ‘’?
<mpapis>
enebo, "sv" to get system
nanoyak has quit [Quit: Computer has gone to sleep.]
<enebo>
oh yeah…cool then I can remove my _rvm_p method