<JRubyGithub>
[jruby] mkristian pushed 2 new commits to master: http://git.io/vJvvZ
<JRubyGithub>
jruby/master ff10216 Christian Meier: putting the jruby-classloader as context classloader means that all attached "jars" can be found via...
<JRubyGithub>
jruby/master 83767d5 Christian Meier: back to SNAPSHOT version
<JRubyGithub>
[jruby] torcido opened issue #2887: jrubyc failing - TypeError: failed to coerce org.objectweb.asm.ClassWriter to org.jruby.org.objectweb.asm.ClassVisitor http://git.io/vJvjh
JRubyGithub has left #jruby [#jruby]
tcrawley-away is now known as tcrawley
deobalds_ has joined #jruby
deobalds has quit [Ping timeout: 250 seconds]
deobalds_ is now known as deobalds
skade has joined #jruby
six has joined #jruby
vikaton has joined #jruby
<six>
hi guys. I'm trying to install mysql2 gem with jruby. "jruby: warning: unknown property jruby.cext.enabled" - the build fails because native extentions are not supported. Has the way to enable them changed in 9000?
drbobbeaty has joined #jruby
drbobbeaty has quit [Read error: Connection reset by peer]
<sluukkonen>
that's what the http response looks like
<nirvdrum>
That looks like a Ruby-level exception (ArgumentError), but there were issues with proper exception handling at some point. JRuby 1.7.3 is really old. You might very well not see the issue with a newer version.
<nirvdrum>
Beyond that, if you can't catch the exception you're limited to what you can do in logging tools. Something like logstash would help you filter it out.
<sluukkonen>
the point is not to print it at all
elia_ has joined #jruby
<sluukkonen>
it's an information leak
elia has quit [Ping timeout: 252 seconds]
deobalds has joined #jruby
<sluukkonen>
anyhow, I'll try 1.7.18. we're running some apps on 1.7.3 due to a bug in jnr-netdb that triggers on debian 6 x86 (which some production boxes still use)
lance|afk is now known as lanceball
<sluukkonen>
(jruby 1.7.3 had an old enough version of jnr-netdb that doesn't crash the vm)
<enebo>
mkristian_: so mvn clean did not get rid of this?
<enebo>
mkristian_: perhaps I needed to mvn clean -Pall?
<enebo>
mkristian_: still find it weird it is picking this up and not using a specific version though
<mkristian_>
enebo, maven/jruby-core is not part of jruby-1_7
<enebo>
mkristian_: actually I see this in a number of target artifacts
colinsurprenant has quit [Ping timeout: 256 seconds]
colinsurprenant_ has joined #jruby
<enebo>
mkristian_: but yeah what can you do if you dev master and jruby-1_7 from the same dir (which I don’t do on macos but I do on windows)
<enebo>
mkristian_: so I will just wipe out all target dirs
<mkristian_>
enebo, wipe out all target dirs is good. run mvn -Pclean before switching the branch - hope the profile still works
arturaz has quit [Ping timeout: 255 seconds]
<enebo>
mkristian_: since I ran into this I will likely figure this out the next time it happens and I guess most people who contribute are not jumping back and forth between branches so perhaps this is just a little learning curve for the few people this can happen to
pjammer has joined #jruby
<enebo>
mkristian_: Also had we not changed API of jnr-posix this would have normally worked anyways
dumdedum has joined #jruby
<mkristian_>
enebo, well somehow things can mixed on the classloader. the ruby script in lib/pom.rb should just work for jruby-1.7.3 as well for newer versions. t
<mkristian_>
that the execution is not isolated is not clear to me
<enebo>
mkristian_: I guess I find it weird 1.7.3 did not use a version of jnr-posix it was shipped with
<mkristian_>
enebo, I do remember there were versions ranges in those jnr-* artifacts before - maybe that is why
<enebo>
mkristian_: yeah perhaps
<mkristian_>
enebo, but the next ruby-polyglot-plugin should get rid of those JRUBY_HOME or jruby.home configurations - that is too close to environment
subbu has joined #jruby
elia_ has quit [Read error: Connection reset by peer]
elux has joined #jruby
<elux>
hey guys
<elux>
congrats on 9kpre2!
<elux>
just wondering, have you done any benchmarks comparing older versions of jruby...? would be cool to see some of that stuff
<enebo>
elux: we are still a bit slower on 9k than 1.7 but we have been addressing bottlenecks as we find them
<enebo>
elux: as we make it to final we should be faster for JIT
<enebo>
elux: but really our goal of .0 was parity with huge potential
<elux>
that makes sense. it's impressive what you guys have done and the new backend is really exciting
<elux>
keep it going and best of luck :)
<enebo>
elux: we have some infrastructure to inline methods and blocks and some plans to address some shortcomings we could never address in 1.7 so each point release of 9k will keep getting faster and faster
<enebo>
elux: IR was really about making a new runway for bigger optimizations
<enebo>
elux: unfortunately it has taked longer than anticipated to get the new runway written
elia has joined #jruby
<elux>
dont sweat it, every project takes longer than anticipiated
<enebo>
aint that the truth :)
<enebo>
I am really happy with our progress though. Pretty amazing how many perf fixes end up being above actual JIT and interpreter impl now during generation of IR instructions or during semantic analsysis
<elux>
the best part is all the potential after once you guys reach parity
<elux>
and the persistence of this project, its been constantly active and supported by a small group, that dedication is impressive
<enebo>
elux: thanks we do our best and have been happy to get paid to work on this
<enebo>
mkristian_: man…very weird. I am still getting the 1.7.3 with all those target dirs wiped out
<enebo>
mkristian_: I am going on a .jar hunt
<enebo>
mkristian_: Somehow something has a jruby 1.7.3 jar somewhere and it is getting slurped in on windows
<enebo>
heh just found a ./build dir
<mkristian_>
enebo, you should look for jruby as well for jruby-core or jruby-complete jar - everything is possible
baroquebobcat has joined #jruby
<enebo>
mkristian_: yeah I am systematically removing jar sources
<enebo>
mkristian_: I should probably just reclone and wipe this dir out
<enebo>
mkristian_: but if I can remove one thing like set of gems and it works then I will know where it came from at least
<enebo>
mkristian_: OH yeah I also just remembered I sent you a PR last night
<enebo>
mkristian_: for jar-dependencies
<mkristian_>
enebo, yes - I did see it. will look/merge it later
baroquebobcat has quit [Ping timeout: 250 seconds]
skade has quit [Quit: Computer has gone to sleep.]
<enebo>
It worked and I don’t know what I deleted :P
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
camlow325 has joined #jruby
erikhatcher has joined #jruby
<nirvdrum>
Did the psych gem/jar final location get worked out?
<mkristian_>
nirvdrum, there were some issue with copy thing in the wrong phase in ./lib/pom.rb - but the location never really changed.
elia_ has joined #jruby
<nirvdrum>
mkristian_: I mean the last I knew it was being sourced from a branch on a fork of yours.
<nirvdrum>
mkristian_: If it's going to be part of the JRuby build, we should probably get that under the JRuby org in a canonical location.
<nirvdrum>
I think headius was looking at that. But I didn't keep up.
<mkristian_>
nirvdrum, this is still the case. I think we want the psych released with java gem as well. but
<mkristian_>
whether it is my fork for the PR or some jruby org fork. both are not ideal
<nirvdrum>
Fair enough. It'd just be good to limit the bus factor.
elia has quit [Ping timeout: 264 seconds]
<mkristian_>
enebo, it looks like your are seeing 1.7.20 is working on windows ? is the release close ?
gregorsc5 has joined #jruby
KevinCorcoran has joined #jruby
elux has quit [Quit: Bye!]
baroquebobcat has joined #jruby
<enebo>
mkristian_: planning for monday if the rg upgrade is ok
tcrawley is now known as tcrawley-away
mrmargol_ has joined #jruby
<mkristian_>
enebo, anything I can do for checking the RG upgrade. run the torquebox testsuite with it ? bbrowning ?
mrmargolis has quit [Ping timeout: 256 seconds]
<enebo>
mkristian_: seems like RG is working so far but I am not sure what we can test to exercise LOAD_PATH changes
<enebo>
mkristian_: tb integs obviously is a good test
<bbrowning>
RG upgrade where? in 1.7 branch?
<enebo>
bbrowning: yeah it was made yesterday
<mkristian_>
bbrowning, yes, it is now 2.4.6
<bbrowning>
that's been tried various times before and failed - I'd have to go back and see why it kept getting reverted
<enebo>
bbrowning: well if it works for you that is a pretty good seal of approval
<bbrowning>
it never has before :)
donV has joined #jruby
<mkristian_>
bbrowning, but 9k.pre2 works ? there are some patched ported from 9k back to 1_7 - maybe things did improve
<bbrowning>
it had to do with some things getting loaded twice iirc
<bbrowning>
mkristian_: well the large TB3 integ suite doesn't run under 9k currently - I have to port a lot of the tests to work with Ruby 2.2
<bbrowning>
lots of app and their deps that have to be upgraded
<Antiarc>
They also announced ObjC for Windows apps, and Android apps will run on Windows runtimes. Hell of a keynote sofar.
KevinCorcoran has joined #jruby
<Antiarc>
Really making a strong play to be the platform that developers want to be on. Now if only they could have a POSIX-compliant Windows...
<enebo>
Antiarc: get rid of the weird mark/copy/paste console stuff too
<Antiarc>
That's gone in win10
<chrisseaton>
if they'd announce their own high quality line of laptops I'd be excited
gazay has quit [Ping timeout: 244 seconds]
gazay_ has joined #jruby
<Antiarc>
Yeah, the Apple hardware is still just unbeatable, from a consumer perspective
bbrowning is now known as bbrowning_away
elux has joined #jruby
elux has left #jruby [#jruby]
erikhatcher has joined #jruby
KevinCorcoran has quit [Remote host closed the connection]
<Antiarc>
reinstalling jruby-head via rvm is giving me readline missing errors - is that something I've screwed up?
elia has joined #jruby
<headius>
Antiarc: it should get installed during the build
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<Antiarc>
That's what I thought, but something's gone wrong then. Will debug in a bit
<chrisseaton>
Is anyone doing anything complicated or radical to the code base in general or the build system? I have a 30,000 line diff I want to merge over the next few hours
<headius>
I'm not...quiet day
<chrisseaton>
Are you going to JavaOne this year?
<headius>
most likely
<chrisseaton>
Should be some Graal people there, including me probably
erikhatcher has quit [Quit: erikhatcher]
mister_solo has joined #jruby
<headius>
excellent!
nirvdrum has quit [Ping timeout: 265 seconds]
<chrisseaton>
Remember we moved Truffle to Maven central? I'm going to move it back to our server for a couple of days while we do the central release. Shouldn't make a difference to anyone, as the reason for it going on Central was for people using JRuby as a library, and now Truffle is separate nobody will accidentally do that
<headius>
chrisseaton: that's fine
<enebo>
asarih: how long does it take for travis to update for 9.0.0.0.pre2 for jruby-9000? Or is it already done?
<headius>
I think that needs rvm update right?
subbu|lunch is now known as subbu
yfeldblum has joined #jruby
KevinCorcoran has joined #jruby
mrmargolis has quit [Ping timeout: 256 seconds]
bbrowning_away is now known as bbrowning
skade has quit [Quit: Computer has gone to sleep.]
<enebo>
headius: my question? I think travis is all on top of rvm so definitely
<enebo>
headius: but I don’t know how long it takes to make the changeover on travis
zorak8 has quit [Read error: Connection reset by peer]
zorak8 has joined #jruby
pjammer has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
mkristian_ has joined #jruby
lanceball is now known as lance|afk
skade has joined #jruby
elia has quit [Quit: Computer has gone to sleep.]
tcrawley-away is now known as tcrawley
djbkd has quit [Remote host closed the connection]
<nirvdrum>
headius: So, you do have access to the RubyGems page for json?
skade has quit [Read error: Connection reset by peer]
skade has joined #jruby
mkristian_ has quit [Ping timeout: 248 seconds]
lance|afk is now known as lanceball
pjammer has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 2 new commits to truffle-head: http://git.io/vJk37
<JRubyGithub>
jruby/truffle-head 33fb8c5 Chris Seaton: Merge branch 'master' into truffle-head...
<JRubyGithub>
jruby/truffle-head c28f4ce Chris Seaton: [Truffle] Update to 0.7
JRubyGithub has left #jruby [#jruby]
djbkd has joined #jruby
pjammer has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
bzb has joined #jruby
rsim has quit [Quit: Leaving.]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 1 new commit to master: http://git.io/vJknB
<JRubyGithub>
jruby/master ad4ba4b Chris Seaton: Merge branch 'truffle-head'
JRubyGithub has left #jruby [#jruby]
<headius>
yay
<headius>
nirvdrum: I might why
<headius>
?
<nirvdrum>
The source code URL metadata is wrong. But it's not part of the gemspec . . . it's just something you set on the RubyGems site if you're listed as a gem owner.
<nirvdrum>
It points to Flori's documentation page, which isn't quite right.
x1337807x has joined #jruby
<chrisseaton>
monster merge
pjammer has joined #jruby
bbrowning has quit [Quit: Leaving]
tenderlove has quit [Quit: Leaving...]
bbrowning has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 1 new commit to truffle-pack: http://git.io/vJk86
<JRubyGithub>
jruby/truffle-pack be3b19c Chris Seaton: Merge branch 'master' into truffle-pack...
JRubyGithub has left #jruby [#jruby]
tenderlove has joined #jruby
mister_solo has quit [Ping timeout: 256 seconds]
chamila has quit [Ping timeout: 246 seconds]
chamila has joined #jruby
<chrisseaton>
monster merge 2
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton merged truffle-pack into master: http://git.io/vJk0A
JRubyGithub has left #jruby [#jruby]
skade has quit [Quit: Computer has gone to sleep.]
<headius>
chrisseaton: have you guys gotten to BigDecimal yet?
<chrisseaton>
no - but it's blocking some stdlib so we might tackle it soon
<chrisseaton>
should be an easy one for beginners
<headius>
*maybe*
<chrisseaton>
We can just wrap Java's BigDecimal right?
<headius>
right, and that's what we do
<chrisseaton>
No doubt with some funky divide semantics
<headius>
if that's an option for you it's definitely the easiest way to go
<headius>
the original lib is a big ball of C
<headius>
that might mean we'd need to share more Numeric stuff since we have all of that implemented
<nirvdrum>
I discovered when MRI tests weren't processing excludes that wrapping Java's BigDecimal has some cases that run exponentially and MRI returns almost instantaneously.
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton closed pull request #2855: [Truffle] Complete, correct and optimised implementation of Array#pack (truffle-head...truffle-pack) http://git.io/vfkSv
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton deleted truffle-io at cb890b1: http://git.io/vJkzT
JRubyGithub has left #jruby [#jruby]
<headius>
nirvdrum: yeah Java's BigDecimal makes some trade-offs
<headius>
we've thought about replacing it with a better one, like from jscience
<headius>
also BigInteger sucks
<nirvdrum>
I thought I killed the process and realized I just detached the test runner. I forcibly killed it 3 days later.
<headius>
there's a number of numeric algorithms that are slower on JRuby solely because BigInteger and BigDecimal aren't very good
<chrisseaton>
BigInteger is bad
<chrisseaton>
I've not seen jscience
<headius>
one of the two *used* to convert to a string and back for one of the math operations
camlow325 has quit [Remote host closed the connection]
camlow325 has joined #jruby
chamila has quit [Quit: Page closed]
mrmargolis has quit [Remote host closed the connection]
fuzzyhorns1 has joined #jruby
<fuzzyhorns1>
having terrible memory leak issues using torquebox and really stuck, anyone with experience in that around to bounce ideas off of?
<fuzzyhorns1>
what i see when i do a histo on the jvm is a huge bunch of [C
<fuzzyhorns1>
closely followed by [B
<nirvdrum>
fuzzyhorns1: You may get help here, but there's a #torquebox channel, too.
rsim has joined #jruby
<nirvdrum>
bbrowning_away is usually the guy to bug.
<headius>
I need someone who has never seen the code for JavaSecuredFile.java in jnr-posix to clean room impl it for me
<headius>
it's (C) Google and we need to get it out of there, but it's a trivial implementation
<headius>
it's mostly just extending java.io.File and returning sane defaults for SecurityException cases
<fuzzyhorns1>
thanks nirvdrum
<fuzzyhorns1>
do you have any idea what is represented in char arrays in java via jruby in ruby?
<fuzzyhorns1>
like what kind of ruby objects are getting translated into char arrays?
<nirvdrum>
fuzzyhorns1: Generally, I've found when there's a memory leak with TorqueBox, something in your stack has spawned a daemon Thread and didn't register an at_exit handler to stop it.
<nirvdrum>
Some gems that have only been tested on MRI do this because the lifecycle is tied to the process and they just expect the thread to go away when the process does.
<fuzzyhorns1>
any common suspsects?
vikaton has quit []
<nirvdrum>
TorqueBox does hot reloading, so you'll have multiple runtimes available per process. A running thread is enough to keep an entire runtime around.
insidious has joined #jruby
<nirvdrum>
Run jstack or take a heap dump and look in jvisualvm/yourkit. You should see RubyThread and the file:lineno
fuzzyhorns1 has quit [Quit: Leaving.]
camlow325 has quit [Remote host closed the connection]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] ArthurUbuntu opened issue #2891: NameError: class not found in packages http://git.io/vJkp2
<Antiarc>
I rvm implode'd my install, made sure readline and readline-devel are installed, and reinstalled rvm and jruby-head, and I'm still getting "Readline was unable to be required, if you need completion or history install readline then reinstall the ruby."
<headius>
hmmm
<headius>
are the files where they ought to be in stdlib?
<headius>
they should get installed under stdlib during the build