josh-k has quit [Remote host closed the connection]
_whitelogger__ has quit [Excess Flood]
_whitelogger__ has joined #jruby
teamon has quit [Ping timeout: 260 seconds]
teamon has joined #jruby
ludyte has quit [Quit: ludyte]
ephemerian has quit [Quit: Leaving.]
_whitelogger__ has joined #jruby
_whitelogger__ has joined #jruby
zorak8 has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to master: http://git.io/Gc4CxQ
<JRubyGithub>
jruby/master 9075f66 Thomas E. Enebo: Add two more internal properties not caught be new property warning system
JRubyGithub has left #jruby [#jruby]
johnsonch_afk is now known as johnsonch
_whitelogger__ has quit [Excess Flood]
_whitelogger__ has joined #jruby
teamon has joined #jruby
enebo has quit [Quit: enebo]
johnsonch_afk is now known as johnsonch
e_dub has joined #jruby
e_dub has quit [Read error: Connection reset by peer]
_whitelogger__ has joined #jruby
_whitelogger__ has joined #jruby
jmimi has quit [Quit: Leaving.]
_whitelogger__ has quit [Ping timeout: 260 seconds]
_whitelogger__ has joined #jruby
ludyte has quit [Quit: ludyte]
_whitelogger__ has joined #jruby
ludyte has joined #jruby
zorak8 has joined #jruby
zorak8 has quit [Ping timeout: 250 seconds]
tylersmith has joined #jruby
_whitelogger__ has quit [Ping timeout: 260 seconds]
_whitelogger__ has joined #jruby
Aethenelle has joined #jruby
_whitelogger__ has quit [Excess Flood]
_whitelogger__ has joined #jruby
zorak8 has quit [Ping timeout: 265 seconds]
tlarevo has joined #jruby
SynrG has quit [Read error: Connection reset by peer]
havenwood has quit [Remote host closed the connection]
rcvalle has quit [Quit: rcvalle]
SynrG has joined #jruby
tcrawley is now known as tcrawley-away
_whitelogger___ has joined #jruby
elia has quit [Quit: Computer has gone to sleep.]
josh-k has quit [Read error: Connection reset by peer]
teamon has quit [Ping timeout: 260 seconds]
_whitelogger___ has joined #jruby
teamon has joined #jruby
robbyoconnor has joined #jruby
zorak8 has quit [Ping timeout: 260 seconds]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #jruby
Aethenelle has quit [Quit: Aethenelle]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #jruby
_whitelogger___ has joined #jruby
nipra has quit [Quit: Leaving.]
dmarcotte has quit [Quit: Leaving.]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #jruby
SynrG has quit [Read error: Connection reset by peer]
SynrG has joined #jruby
_whitelogger___ has joined #jruby
josh-k has joined #jruby
_whitelogger___ has joined #jruby
subbu has joined #jruby
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #jruby
tbrock has joined #jruby
dmarcotte has joined #jruby
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #jruby
teamon has quit [Ping timeout: 260 seconds]
teamon has joined #jruby
_whitelogger___ has joined #jruby
tlarevo has quit [Remote host closed the connection]
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #jruby
dmarcotte has quit [Quit: Leaving.]
jrhe_ has quit [Quit: Connection closed for inactivity]
teamon has quit [Ping timeout: 260 seconds]
teamon has joined #jruby
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #jruby
iamjarvo has joined #jruby
_whitelogger___ has quit [Excess Flood]
_whitelogger___ has joined #jruby
<r01>
can anyone answer a classpath jar conflict question for me?
<r01>
if a jruby gem includes a jar but a new version of that jar appears elsewhere on the classpath added by the hosting application context, is there well known way to resolve which jar is used for the jruby gems?
johnsonch is now known as johnsonch_afk
_whitelogger___ has quit [Excess Flood]
tbrock has quit [Quit: Computer has gone to sleep.]
_whitelogger___ has joined #jruby
tbrock has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to master: http://git.io/RtkKXQ
<JRubyGithub>
jruby/master cc5ae83 Chris Seaton: [Truffle] Remove NilPlaceholder and just use RubyNilClass.
JRubyGithub has left #jruby [#jruby]
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to pedantic: http://git.io/ZKkCOA
<JRubyGithub>
jruby/pedantic 1618457 Chris Seaton: Document -Xcompliance.strict as being about MRI, rather than RubySpec.
<mkristian>
enebo, FYI: I am about to fix the -Posgi and have a look at -Pj2ee as well
<r01>
replaying a question i submitted last night
<r01>
if a jruby gem includes a jar but a new version of that jar appears elsewhere on the classpath added by the hosting application context, is there well known way to resolve which jar is used for the jruby gems?
dmarcotte has quit [Quit: Leaving.]
<r01>
or is the only simple path to get the jar versions to match?
<mkristian>
jruby-9000 will not use the jar from the classpath but the one from the gem. otherwise is a matter which one gets loaded first
<mkristian>
so get the versions in sync ;)
<mkristian>
otherwise == jruby-1.7.x
<Aethenelle>
mkristian: how's that going to work? search for the jar, create a class loader for it and ... profit?
<Aethenelle>
r01: the short version of mkristian's anser for 1.7.11 is the same way you'd do it in Java: hope
<Aethenelle>
maybe not short but definitely less technical... :P
<mkristian>
Aethenelle, first question: the jruby-classloader is where the jars from the gems are loaded. and this classloader has a self-first-strategy to load classes. so parent classloader comes second
<headius>
howdy
<Aethenelle>
and anything loading in the jruby runtime is going to hit that classloader first?
mister_solo has quit [Ping timeout: 255 seconds]
<Aethenelle>
I'm still a bit fuzzy on classloader symantics.
<r01>
we've tried that approach with 1.7.13 and it didn't work
<mkristian>
if the runtime will not look into runtime.getJRubyClassLoader then runtime would be not separated - regarding required jars
<r01>
but that being said, the gem was probably not requiring the jar until that codepath was hit at runtime
<mkristian>
r01, which approach ?
<r01>
the approach where our code is first on the classpath
<r01>
to give a bit more background
<r01>
1) our gems are bundled in a single jar with jruby-complete
<r01>
2) our jar is first on the classpath
<r01>
3) the gson gem requires a bundled gson-2.2.2.jar file that ships with that gem
vyorkin has quit [Quit: WeeChat 1.0.1]
<Aethenelle>
r01: mkristian isn't just talking about the classpath... he's talking about a classloader hierarchy change going into 9k that provides the segregation.
<r01>
so i don't think that jar actually appears on the classpath until require "gson-2.2.2.jar" is executed
<mkristian>
r01, add you have another gson-2.2.2.jar somewhere ?
<r01>
no there was a gson-2.2.4.jar that appeared later in the classpath
<Aethenelle>
require 'something.jar' does not manipulate the classpath
<r01>
Aethenlle: I didn't know that
<r01>
good to know
<mkristian>
r01, and what exactly is the problem ? does the gson-2.2.2.jar once or twice ?
<Aethenelle>
it adds the jar to an internal URLClassloader. If one the classloaders up the hierarchy has already loaded the class, that one will get loaded. Otherwise the one in the URLClassloader will get picked up.
<r01>
the problem was that require 'gson' didn't function properly because at runtime it was picking up the 2.2.4 jar instead of the 2.2.2 jar that was shipped with the gem
ephemerian has quit [Quit: Leaving.]
<r01>
btw, is there any good resource about the status and timeframes for jruby 9000?
<mkristian>
r01, for jruby-1.7.x with different versions of the same jar like this - it will use the classes you packed along side jruby. (beside classes which only exist in 2.2.2)
<r01>
ok, let me try to repro that will a small test case
<Aethenelle>
ugh... almost had a heart attack there... thought I'd merged in my typed enums code to FFI/master ...
r01 has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
phrinx has joined #jruby
<headius>
r01 preview in about a month, hopefully final a month later
cremes has joined #jruby
<headius>
enebo, subbu: non-indy-jit branch completes mvn clean package now without disabling JIT or using a different JRuby
<headius>
so at least the gem install bits the build uses are working with JIT
<subbu>
nice!!
<subbu>
oh, although i think that part of the build has been turned off, enebo, right?
<subbu>
in master at least.
<Aethenelle>
headius: ugh... i need to get goin on that PR then, don't I?
<headius>
Aethenelle: well, it would be nice to have prepend ready for the preview :-)
<Aethenelle>
I'll be able to work on it a bunch in about a week or so... Right now I have a talk to prepare
<headius>
ok
<headius>
what's the talk?
<Aethenelle>
power analysis attacks against cheap hardware for software people
<Aethenelle>
i get to ditch chicago weather for a couple days
dmarcotte has joined #jruby
<headius>
wow, sounds like an interesting talk
<headius>
I wish I was in SD
calavera has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
havenwood has quit [Remote host closed the connection]
yfeldblum has joined #jruby
havenwood has joined #jruby
e_dub has quit [Quit: ZZZzzz…]
yfeldblum has quit [Ping timeout: 244 seconds]
kares has quit [Ping timeout: 272 seconds]
nanoyak has joined #jruby
<headius>
enebo: I need some help hooking up cloning stuff to JIT
<headius>
I'm getting an error calling prepareForCompilation: Should not clone interp context
<headius>
:-)
<headius>
or subbu
purplefox has joined #jruby
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
<subbu>
headius, hmm ... let me look for a few mins ... i have to prepare for a meeting soon .. if i dont see anything, might have to wait for enebo to get back.
<headius>
it looks like it's when it tries to clone a CallInstr
<headius>
easy to trigger error on branch... jit logging on, try to gem install anything
<headius>
I'm thinking prepareForCompilation is not aligned with cloning
<headius>
hmmm
<headius>
same scope object that interp will use should be ok for me?
<subbu>
the scope is ready to use without cloning since interp has cloned what it needs.
<headius>
or else I don't know what I'm supposed to call first
<subbu>
interp doesn't use the scope object at all now.
<headius>
hmm
<headius>
or is there an "original" scope object I should start from
<subbu>
it builds a "closure" of all state it needs in InterpreterContext and uses it.
<subbu>
you shouldn't have to clone anything.
<subbu>
right now, becuase of a small hack, it is in operands ... but it will get moved out of there soon into interpreter/
<headius>
ok, that's what I thought...but when I try to call prepareForCompilation, it goes down a cloning path and eventually tries to clone a CallInstr...that's when the error is triggered
<headius>
some operand there is the InterpreterContext, and it doesn't clone
<subbu>
ah, so, yes, prepareForCompilation hasn't been fixed up since he added that clone path for interp. checking.
<headius>
ok :-)
<headius>
does it sound right for me to make something like JITContext?
<headius>
I will look at interp logic and try to duplicate
<subbu>
yes, prepareInstructions (which is shared between interp and compiler) clones instrs.
noop has joined #jruby
<subbu>
it sets ipc, rpc, specializes for interp, etc.
<headius>
ok
<subbu>
so, perhaps prepareInstructions needs to be copied and fixed up for JIT ...
<subbu>
you can get rid of the cloning, ipc, rpc, specialize-for-interp.
<headius>
hmm yeah, what it's doing now is obviously not right
<headius>
that's the path under which it blows up
<subbu>
yup. that is now hugely interp-centric.
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<subbu>
alright, i have to switch back now ... but, hopefully enebo will be back soonish if there is anything else to fix up there.
<headius>
ok
thsig_ has joined #jruby
<subbu|busy>
you shouldn't need a JITContext since multiple JIT threads wont compile the same scope at the same time + once you compile, you don't need the scope anymore ... this might change in the future, but for now, this should work.
subbu is now known as subbu|busy
thsig has quit [Ping timeout: 272 seconds]
Hobogrammer has joined #jruby
Hobogrammer has quit [Read error: Connection reset by peer]
cremes has quit [Read error: Connection reset by peer]
enebo has quit [Read error: Connection reset by peer]
<mkristian>
headius, sonatype wants to include jruby in their nexus-server so they are digging deep into license.
lance|afk is now known as lanceball
anaeem1_ has joined #jruby
<mkristian>
headius, add https://github.com/jnr/jnr-posix is still CPL - did this slip when the change from CPL to EPL took place ? (BTW sorry for those license questions - can't help it right now)
anaeem1_ has quit [Read error: Connection reset by peer]
anaeem1_ has joined #jruby
elia has joined #jruby
<headius>
mkristian: ok, we should fix that too
<headius>
anything that says CPL can be changed to EPL without hassle
<mkristian>
ok
Aethenelle has quit [Quit: Aethenelle]
brixen has quit [Ping timeout: 250 seconds]
Aethenelle has joined #jruby
elia has quit [Ping timeout: 258 seconds]
brixen has joined #jruby
nanoyak has quit [Quit: Computer has gone to sleep.]
robbyoconnor has quit [Ping timeout: 255 seconds]
_ko1 has joined #jruby
nanoyak has joined #jruby
yfeldblum has quit [Ping timeout: 246 seconds]
anaeem1_ has quit [Remote host closed the connection]
e_dub has joined #jruby
dumdedum has quit [Ping timeout: 272 seconds]
JohnBat26 has joined #jruby
teamon has quit [Ping timeout: 260 seconds]
teamon has joined #jruby
purplefox has quit [Ping timeout: 250 seconds]
colinsurprenant has quit [Quit: colinsurprenant]
tenderlove has quit [Quit: Leaving...]
tenderlove has joined #jruby
Aethenelle has quit [Quit: Aethenelle]
anaeem1 has joined #jruby
nanoyak has quit [Quit: Computer has gone to sleep.]
yfeldblum has joined #jruby
Aethenelle has joined #jruby
teamon has quit [Ping timeout: 260 seconds]
teamon has joined #jruby
bbrowning is now known as bbrowning_away
purplefox has joined #jruby
havenwood has quit [Remote host closed the connection]
<enebo>
asarih: well even if there is a bug this is a reason for us to get it green. We need to know what is not working
<enebo>
asarih: cool to see it updated in any case :)
jmimi has joined #jruby
colinsurprenant has quit [Quit: colinsurprenant]
<enebo>
asarih: what is not passing?
<asarih>
enebo: in general. when it passes, a build is triggered.
colinsurprenant has joined #jruby
<asarih>
enebo: headius: I think I can manually build this, but it would be best if you could have passing builds. ;-)
<enebo>
asarih: yeah this is first time we have been close enough to get this green
<enebo>
asarih: we did not anticipate our cross over to new runtime would take so long to back back to this point
<enebo>
in retrospect keeping 2 extra branches from bitrotting may have been less painful but it would have been painful in either case I think
<enebo>
asarih: I will tell you a lesson I did learn…do not do major dev on master to it on a trunk because travis expects that
calavera has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<asarih>
enebo: master is the only branch with the magic, I think.
<enebo>
asarih: yeah there is a dichotomy to all this. Earlier in major dev we really don’t want the entire universe of apps depending on things working because we are breaking stuff
<enebo>
asarih: later in major dev we do want apps depending on it to help verify things are in a good state
jrhe_ has joined #jruby
<enebo>
asarih: From 3rd part perspective this also exists where early on they want latest stable non-major release but nearing a major release they want to also make sure the new version works