<ebbflowgo>
does that need to be run using jruby or can i run it using ruby... my misunderstanding stems from java requiring ruby 2.1.0 or higher (gem install) but this code only running properly when i use jruby
<ebbflowgo>
headius: one other thing, trying to install your jruby-pg... if i run it through bundle install i get "Could not find gem 'jruby-pg (>= 0) java' in the gems available on this machine."
<a5i>
Have you guys seen Crystal?
<headius>
ebbflowgo: oops, I guess that's because we haven't released it :-\
<headius>
a5i: yeah I looked at it a while back
<headius>
looked interesting
nirvdrum has quit [Quit: Leaving]
<ebbflowgo>
headius: oh, for the time being what should i use then?
<a5i>
Yeah its gaining some hype on HN for about 1 or 2 weeks now
<headius>
ebbflowgo: is this for a Rails app?
<ebbflowgo>
no
<headius>
and you need pg itself...ok
<headius>
for now you might just have to build the jruby-pg gem yourself
<headius>
you could file an issue to get a release of it out and I can try to do that tomorrow (or the other maintainer)
<a5i>
would be nice to see Crystal vs JRuby 9.0.0.0 final benchmarks
<headius>
well, Crystal statically optimizes, so they're probably well ahead of us
<headius>
but of course that also limits how Ruby-like it can be
<lopex>
rafaelfranca: btw what 2.x features do they use ?
<rafaelfranca>
actually, none yet, but we officially don't support MRI 1.9 anymore (since it is not maintained anymore). But I can change it to ">= 1.9.3" it that would fix jruby problem
<rafaelfranca>
codefinger: good question, I thought it would be lang version
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] bjfish pushed 1 new commit to master: http://git.io/ve7Ik
<JRubyGithub>
jruby/master 971b87c Brandon Fish: [Truffle] Adding NoMethodError to exception.rb.
JRubyGithub has left #jruby [#jruby]
rcvalle has joined #jruby
baroquebobcat has joined #jruby
pjammer has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<codefinger>
do you want me to request access from sonatype, or just have one of you guys do it? I'm anticipating that this will be the last time ever because i want to rework this so that the jetty version is configurable (and the jar is built on the fly)
camlow32_ has quit [Read error: Connection reset by peer]
camlow32_ has joined #jruby
<enebo>
codefinger: well I can push this if you want. I could give you access to publishing but I would need to remember how to do that :)
<codefinger>
enebo: i know. it would take me forever to remember. i think i just need to open a ticket and have you comment on it.
<codefinger>
either way is fine. whatever you want
<enebo>
yeah I can do it. So I just need to use Java 7? and deploy from this ext dir?
<codefinger>
i guess so :) only nick knows for sure
camlow325 has quit [Ping timeout: 252 seconds]
lanceball is now known as lance|afk
temporalfox has joined #jruby
<enebo>
ah ok well I will try that
<enebo>
codefinger: I don’t think nick ever did any sonatype releases either
<enebo>
just the supplemental stuff associated with that artifact
<codefinger>
will that sync to repo1?
<enebo>
codefinger: that look like correct name and version?
<codefinger>
yea
<enebo>
codefinger: I have to connect up and close then release on their website
<codefinger>
ah
<enebo>
oh crud
camlow325 has joined #jruby
<enebo>
codefinger: this pushed to codehaus
<codefinger>
where should it go?
<enebo>
codefinger: I am guessing this will make it to maven central at some ponit but it is not quick
<codefinger>
yea, nothing to maven central is quick
<codefinger>
i'll wait a few hours/days and check back with you
<enebo>
codefinger: yeah. if not we should revv and release to sonatype which means figuring out how to change this pom
<enebo>
codefinger: but we should do that anyways I guess.
<codefinger>
i'll figure that out...but hopefully i'll be able to get rid of it altogether and build the jar on the fly
<enebo>
codefinger: ok. WellI I am happy to help if you need it later
rafaelfranca has quit []
codefinger has quit [Remote host closed the connection]
shellac has quit [Ping timeout: 256 seconds]
dinfuehr has quit [Remote host closed the connection]
camlow325 has quit [Remote host closed the connection]
camlow325 has joined #jruby
bbrowning_away is now known as bbrowning
ebbflowgo has quit [Read error: Connection reset by peer]
ebbflowgo has joined #jruby
dinfuehr has joined #jruby
dumdedum has quit [Quit: foo]
zorak8 has joined #jruby
dinfuehr has quit [Remote host closed the connection]
temporalfox has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
dinfuehr has joined #jruby
camlow325 has quit [Remote host closed the connection]
camlow325 has joined #jruby
shellac has joined #jruby
dinfuehr has quit [Ping timeout: 264 seconds]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] enebo pushed 1 new commit to master: http://git.io/ve7yV
<JRubyGithub>
jruby/master af2e4db Thomas E. Enebo: Fixed remaining mri kwargs issues sans parsing bug (did not untag because of some JIT failures with kwargs)
JRubyGithub has left #jruby [#jruby]
erikhatcher has joined #jruby
Aethenelle_ has joined #jruby
Aethenelle has quit [Ping timeout: 264 seconds]
Aethenelle_ is now known as Aethenelle
<Aethenelle>
I've been looking at https://github.com/ffi/ffi/issues/427 seems these specs are bad since GC can't be relied on in this manner. Any ideas? Would manually calling finalize be valid enough?
vtunka has quit [Quit: Leaving]
lance|afk is now known as lanceball
elia has quit [Quit: Computer has gone to sleep.]
elia has joined #jruby
e_dub has quit [Quit: ZZZzzz…]
<headius>
Aethenelle: for similar tests I've modified then to loop over calling something to trigger GC until it passes or a time limit expires
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] bjfish pushed 1 new commit to master: http://git.io/ve7pR
<Aethenelle>
i have them partially working by only trying to get it called once, allocating a 1g string and forcing gc every iteration for 1000 tries...
shellac has quit [Quit: Computer has gone to sleep.]
<Aethenelle>
that doesn't get all of them to pass either
e_dub has joined #jruby
mkristian has quit [Ping timeout: 244 seconds]
rcvalle has quit [Quit: rcvalle]
subbu is now known as subbu|lunch
chamila has quit [Quit: Page closed]
rcvalle has joined #jruby
djbkd has quit [Remote host closed the connection]
tcrawley is now known as tcrawley-away
Aethenelle_ has joined #jruby
<Aethenelle_>
and now i can't get it to repeat...
Aethenelle has quit [Ping timeout: 240 seconds]
Aethenelle_ is now known as Aethenelle
djbkd has joined #jruby
mkristian has joined #jruby
codefinger has joined #jruby
<headius>
Aethenelle: :-\
<Aethenelle>
unfortunately, I can't get it to repeat the success... not the failure; that's working fine
rcvalle has quit [Quit: rcvalle]
x1337807x has joined #jruby
dinfuehr has joined #jruby
dinfuehr has quit [Remote host closed the connection]
rcvalle has joined #jruby
dinfuehr has joined #jruby
pjammer has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 6 new commits to master: http://git.io/ve5al
<JRubyGithub>
jruby/master 20e0a16 Charles Oliver Nutter: Mark as used by JIT.
<JRubyGithub>
jruby/master 870aa4a Charles Oliver Nutter: All IR blocks need this treatment, including jitted.
<JRubyGithub>
jruby/master fd67247 Charles Oliver Nutter: Never return null from ReceivePostReqd since nobody wants it.
JRubyGithub has left #jruby [#jruby]
yfeldblum has joined #jruby
tcrawley-away is now known as tcrawley
<codefinger>
headius: enebo: can you check me on this. want to make sure there won't be any unintended consequences of sett `--dev` for all `bundle install` and `assest:precompile` calls: https://github.com/heroku/heroku-buildpack-ruby/pull/361
<codefinger>
s/sett/setting/
<headius>
absolutely none
<codefinger>
sweet!
<headius>
other than it will have slower runtime, obviously
<headius>
I think several of us do all local dev with --dev just to keep things moving
<Antiarc>
--dev mostly just turns off JIT and optimizes for faster startup time, IIRC
<codefinger>
but that should be ok, as long as it's not used for the prod part, right?
<headius>
lucky me, I'm usually working on JIT so I can't do it
<headius>
codefinger: right, no problem
<headius>
and it should speed up instance booting, which I would guess is your goal
<codefinger>
yup
rsim has joined #jruby
kfpratt has quit [Read error: Connection reset by peer]
kfpratt has joined #jruby
rsim1 has quit [Read error: Connection reset by peer]
pietr0 has joined #jruby
dumdedum has quit [Quit: foo]
<headius>
enebo: so we're looking at release early next week, yes?
<headius>
chrisseaton: ^^ fair warning...pre2 is nigh
<headius>
wanted to do it this week but hit a few more stumbling blocks
<headius>
Antiarc: how are your apps looking?
<Antiarc>
headius: I haven't tested in a couple days, I'll push a fresh head build shortly
<Antiarc>
But last I looked all green :)
<headius>
that's great, thank you
<headius>
idaelly we only ever get better, right? :-D
<Antiarc>
indeed!
<headius>
bjfish2: I'm cancelling you travis builds so enebo and I can see that all kwargs stuff is passing
<headius>
enebo: anything else you need me to look at before pre2?
<Antiarc>
There's some super rough benchmarks for our test suite. Not sure if that's useful info or not
<Antiarc>
But that's a suite that involves lots of really intensive CPU stuff, and bigass jars
<Antiarc>
I don't know that we'll hit the heavy parts enough to really invoke the jit though
<Antiarc>
But if I can get it into prod, I'll definitely get some good Newrelic charts out of it - we've got a fleet of about 60 AWS instances that manage to saturate most cores right now :)
<Antiarc>
Suites are definitely running a bit slower, though. Not enough to make me nervous, but it looks like there's an extra 10 seconds of overhead in both suites somewhere, which doesn't appear to be in rubyland
rsim1 has joined #jruby
mkristian has joined #jruby
<Antiarc>
Startup stuff, I guess
rsim has quit [Ping timeout: 265 seconds]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] mkristian closed issue #1754: JRuby 1.7.11 (bundled with LogStash v. 1.4.1) fails on FreeBSD 10.0 with NotImplementedError: stat.st_dev unsupported or native support failed to load http://git.io/ve5pO
JRubyGithub has left #jruby [#jruby]
<headius>
Antiarc: ok, yeah, startup is probably a component there, or unoptimized bits of the non-indy JIT
<headius>
we'll be iterating on perf once we're ready for final
<Antiarc>
yup yup
<Antiarc>
I'm not complaining so much as just hoping to provide some real-world benchmarks! :)
havenwood has quit []
<headius>
yeah we need those!
<headius>
definitely
<lopex>
numbers ?
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
lanceball is now known as lance|afk
bbrowning is now known as bbrowning_away
<Antiarc>
hrm, on head, bundler refuses to install manticore-0.4.1, while it happily works on 1.7.19. Not sure that head is to blame just yet, but it's a discrepancy.
elia has joined #jruby
<Antiarc>
gem install works fine
<kares>
headius: Antiarc: was actually thinking about getting a jruby-openssl release ready for pre2 and 1.7.20
<mjc_>
be07196 is slightly faster than 1.7.19 on a blank rails app with torquebox 3.1.1 for me (2068 req/s vs 1773)
<mjc_>
er sorry
<mjc_>
with puma, not torquebox
<kares>
not sure if mkristian will be around to push snapshots for me to make sure nothing broke on JRuby's suite
mkristian_ has joined #jruby
codefinger has quit [Read error: Connection reset by peer]
mkristian has quit [Ping timeout: 252 seconds]
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
erikhatcher has quit [Quit: erikhatcher]
<kares>
mkristian_: seems you just popped in?
<mkristian_>
kares, well my network is not too stable - so I might pop out and in again
<kares>
np ... you're on vacation next week or so right?
<kares>
mkristian_: was thinking about getting jruby-openssl 0.9.7 release ready for 9K.pre2 and 1.7.20
<mkristian_>
I am in India - had a few days vacation last week. next week I am around
subbu|lunch is now known as subbu|busy
<kares>
mkristian_: oh just read about issues with optics in India due monkeys chewing the cables :)
<mkristian_>
kares, so we should start using the 0.9.7-SNAPSHOT on both branch
<headius>
Antiarc: that's weird
<kares>
so I see what you mean that you might pop out
<headius>
kares: Tom or I should be able to push those too, I think
<Antiarc>
headius: after switching to 1.7.19 and bundling, then switching back to head and bundling it works.
<mkristian_>
kares, and there are frequent power cuts - at least once a night
<headius>
a release for pre2 would be excellent
<Antiarc>
Very odd. Probably some oddity in the lockfile.
<headius>
we want to do 1.7.20 shortly after pre2
<mjc_>
bah sorry, disregard my 1.7.19 vs be07196 comment, JIT wasn't warm
<kares>
mkristian_: woow
iamjarvo has joined #jruby
<headius>
mjc_: what's the actual result?
iamjarvo has quit [Max SendQ exceeded]
<mjc_>
trying it again now
<kares>
mkristian_: ok - for now just a snapshot off a test-branch ... based off jruby-1_7
<mjc_>
going to give it 10 minutes instead of 1 minute
<kares>
I'm still trying to make sure TLS 1.2 works
<headius>
kares, mkristian_: get 0.9.7-SNAPSHOT on both branches asap so they can bake
<headius>
we will want to know if they're good in travis etc before releases next week
iamjarvo has joined #jruby
pjammer has joined #jruby
<kares>
headius: well yeah but in case there are issues we do not want to hold the release back :)
<headius>
right
iamjarvo has quit [Max SendQ exceeded]
<headius>
we don't want that either :-)
<kares>
I can not do snapshot deploys so I'll poke around mkristian_or enebo
iamjarvo has joined #jruby
<mkristian_>
headius, kares will do the test-jossl-0.9.7-SNAPSHOT and see how travis takes it
<headius>
ok
<mkristian_>
kares I do the deploy
<mkristian_>
now
<kares>
mkristian_: yy - great!
<headius>
I'm in SF this weekend but I'll keep in touch
<headius>
enebo will be around I believe
<headius>
I'll be back late Monday
<Antiarc>
btw, if any of ya'll are going to be at I/O, I'd love to buy you a beer :)
<enebo>
yeah I will be here
<headius>
I wish...I never get IO tickets in time
erikhatcher has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] mkristian pushed 1 new commit to jruby-1_7: http://git.io/vedLM
<JRubyGithub>
jruby/jruby-1_7 206d9d6 Christian Meier: moved packaging to 'jar' again...
<headius>
I wonder how much of that is just IR load/compile overhead
<Antiarc>
I can do some profiling to get some rough numbers
<mkristian_>
kares, deployed. maybe we can remove the ".dev" then it will works as default gems. rubygems ignores defaults gems with prerelease versions :(
<mkristian_>
and this will fail a test or two
<kares>
mkristian_: oh ;(
<headius>
installing freebsd reminds me why I switched to OS X
<headius>
I'm getting too old to fiddle around in text mode to configure X
<mjc_>
heh
<mjc_>
last time I had to deal with configuring X I just learned to spend all my time in text mode
<headius>
yeah, that's what I'll be doing
<headius>
for the moment anyway
<Antiarc>
Ubuntu and Mint have really impressed me - there's almost none of that required anymore
<Antiarc>
err, Ubuntu and Fedora
<headius>
yeah I usually go that way for a desktop
<kares>
mkristian_: you did take the latest from master for the snapshot, right?
<kares>
... if so I'll setup a test-branch on jruby to see how it's doing, for now
<mkristian_>
kares, yes, last commit 24d4b5b7577fd11be948d26b50e162e053af6cb1
<kares>
mkristian_: ok, thx
<headius>
mkristian_: ahh yeah I forgot about that issue
<headius>
mkristian_: what about timestamp?
lance|afk is now known as lanceball
<headius>
I'm trying to think of anything we can guarantee to be unique, and that seems like the only one other than random numbers
<mkristian_>
timestamp combined runtime.number. sounds more general then just separating classloaders and runtimes
<Antiarc>
Well, I'm looking at hprof traces, and what's interesting is that the bulk of the tests is iowait (which isn't surprising), but I wouldn't expect substantially divergent performance numbers in those cases
<headius>
mkristian_: yeah let's do that then
<headius>
mkristian_: maybe a date stamp so it's meaningful
<mkristian_>
ok
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] enebo opened issue #2817: JavaMethod* needs appropriate getParameters info http://git.io/vedGn
JRubyGithub has left #jruby [#jruby]
<headius>
20150409145201
<headius>
second granularity at least
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] enebo closed issue #2816: UnboundMethod#parameters for an assignment method wrongly returns blank array http://git.io/veQhX
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] enebo pushed 1 new commit to jruby-1_7: http://git.io/vedGH
<JRubyGithub>
jruby/jruby-1_7 7861be0 Thomas E. Enebo: Fixes #2816. UnboundMethod#parameters for an assignment method wrongly returns blank array
JRubyGithub has left #jruby [#jruby]
bbrowning_away is now known as bbrowning
pjammer has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
subbu|busy is now known as subbu
pjammer has joined #jruby
x1337807x has joined #jruby
shellac has joined #jruby
robbyoconnor has quit [Ping timeout: 246 seconds]
rcvalle has joined #jruby
elia has quit [Ping timeout: 244 seconds]
e_dub has quit [Quit: ZZZzzz…]
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
elia has joined #jruby
elia has quit [Client Quit]
rcvalle has quit [Quit: rcvalle]
cremes has quit [Remote host closed the connection]
codefinger has joined #jruby
cremes has joined #jruby
djbkd has quit [Remote host closed the connection]
rcvalle has joined #jruby
dinfuehr has quit [Remote host closed the connection]
cremes has quit [Remote host closed the connection]
erikhatcher has quit [Ping timeout: 245 seconds]
erikhatcher_ has joined #jruby
rcvalle has quit [Client Quit]
rsim1 has quit [Quit: Leaving.]
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
shellac has quit [Quit: Computer has gone to sleep.]
iamjarvo has joined #jruby
x1337807x has joined #jruby
djbkd has quit [Remote host closed the connection]
skade has joined #jruby
robbyoconnor has joined #jruby
mrmargolis has quit []
Jamo__ has joined #jruby
<headius>
nirvdrum: oh yeah, in xerces
djbkd has joined #jruby
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
bbrowning is now known as bbrowning_away
havenwood has quit [Ping timeout: 250 seconds]
Aethenelle has quit [Quit: Aethenelle]
mkristian_ has quit [Ping timeout: 264 seconds]
x1337807x has joined #jruby
yfeldblum has quit [Remote host closed the connection]
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
dinfuehr has joined #jruby
lanceball is now known as lance|afk
yfeldblum has joined #jruby
nateberkopec has quit [Quit: Leaving...]
dinfuehr has quit [Ping timeout: 245 seconds]
robbyoconnor has quit [Read error: Connection reset by peer]
erikhatcher_ has quit [Quit: erikhatcher_]
havenwood has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] mkristian pushed 1 new commit to jruby-1_7: http://git.io/veFO9
<JRubyGithub>
jruby/jruby-1_7 f959d12 Christian Meier: ensure MBean names are unique by using a timestamp...
JRubyGithub has left #jruby [#jruby]
skade has quit [Quit: Computer has gone to sleep.]
<nirvdrum>
headius: Ahh, yeah, that was it. Not commons.
<nirvdrum>
lopex, headius: Is there any compelling reason to have ByteList be final?
<lopex>
the same as String is I suppose ?
<lopex>
no more performance reason I think
<lopex>
nirvdrum: but why ?
mitchellhenke has quit [Quit: Computer has gone to sleep.]
<nirvdrum>
Okay. I wanted to try out an ImmutableByteList, but not being able to slot it into the type hierarchy means a fair bit of rewriting.
<lopex>
ah, you think in terms of truffle
<lopex>
?
<nirvdrum>
I'm not entirely sure yet. It could be one way to look into ropes, as well.
<headius>
nirvdrum: right, no real benefit
<lopex>
nirvdrum: and you'd virtualize methods
<lopex>
nirvdrum: did I tell you they enable lots of perl features in ruby regexps ?
<lopex>
er, headius ^^ ?
<nirvdrum>
lopex: Fair enough. I think that could be mitigated a bit, but maybe not.
<lopex>
it there trimorphic inliing in hotspot yet ?
<nirvdrum>
No idea. But lazy loading the class should avoid the penalty of two implementations if you never use the other.
<lopex>
that's what I was in mind implementing jcodings all the time
<lopex>
*er, I had ?
stepheneb has joined #jruby
temporalfox has joined #jruby
<nirvdrum>
Yeah.
<lopex>
nirvdrum: there's still a check at runtime afaik, or the deopt just invalidates the code on classloading
<nirvdrum>
It'd deopt on class loading.
<nirvdrum>
But I'm hardly an expert on HotSpot. I've been trying to get up to speed by reading papers and such.
<nirvdrum>
Having said that, some of the stuff I'm looking at could potentially dominate whatever overhead a virtual call would induce.
<lopex>
nirvdrum: but I wonder, what about the thereads that are already executing that code ?
<nirvdrum>
An immutable ByteList could be a potential gain for IR's frozen string literals, too.
<lopex>
yeah
iamjarvo has joined #jruby
<mjc_>
that sounds neat
<lopex>
maybe ultimately not ropes, but some kind of persistent data astructure like clojure has ?
<lopex>
that would work for Strings
<nirvdrum>
They hit a safepoint and deopt as I understand it, then go back to running interpreted code.
<nirvdrum>
lopex: Part of what I want to look into too is a ByteList cache.
<lopex>
but savepoint might stop the world and deopt ?
<nirvdrum>
I don't think it's quite stop the world. My understanding is the code cache is partially rewritten with a jmp back into the interpreter and the vtable entry from the method entry to the code cache is removed.
<lopex>
nirvdrum: but I could imagine there might be more semantics of doing that
<nirvdrum>
Semantically it should be allowed to continue execution in the method until it exits, too, I'd imagine.
<lopex>
I mean, different ones
elia has joined #jruby
<nirvdrum>
As the method was entered prior to the class loading anyway. But I suppose inlining complicates matters.
<lopex>
yeah, that's a heystack of maps I guess
<nirvdrum>
headius probably knows more ;-)
<nirvdrum>
lopex: As a quick example, rack has a very common set of HTTP response headers and some of those are now frozen to save object allocations on MRI. IR looks like it can optimize frozen strings in some capacity, but if there were also an LRU cache of immutable bytelists, it might be able to save a whole bunch of Java object allocations as well.
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<lopex>
nirvdrum: would that be a weak cache ?
<nirvdrum>
The whole thing is in the very formative stages. It might not pan out to anything. But won't know until I try.
<lopex>
nirvdrum: truffle/grall should make it easier since they might reuse some java String facilities
<lopex>
since the literals are collected now ?
<lopex>
or for some time
<nirvdrum>
I don't think graal does anything special with String quite yet.
<nirvdrum>
But I could be wrong.
<lopex>
I'm speculating
<nirvdrum>
HotSpot does intrinsify some String operations.
enebo has quit [Quit: enebo]
<lopex>
nirvdrum: but afaik, now ldc does not just load the string from the pool
<lopex>
there's some additional activity if Strings are collected
<nirvdrum>
But where I think some things could be gained is if you have an immutable ByteList some operations can be reduced to reference comparisons.
<lopex>
nirvdrum: you're lucky too, since fixnums are frozen as well now
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
travis-ci has joined #jruby
<lopex>
till that mri maintained non collectable map for fixnum ivars