pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
skade has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
robbyoconnor has quit [Ping timeout: 265 seconds]
vtunka has quit [Quit: Leaving]
bbrowning_away is now known as bbrowning
vtunka has joined #jruby
yfeldblum has quit [Ping timeout: 260 seconds]
etehtsea has quit [Quit: Computer has gone to sleep.]
vtunka has quit [Ping timeout: 240 seconds]
nirvdrum has joined #jruby
tcrawley-away is now known as tcrawley
vtunka has joined #jruby
lance|afk is now known as lanceball
vtunka has quit [Quit: Leaving]
vtunka has joined #jruby
robbyoconnor has joined #jruby
vtunka has quit [Quit: Leaving]
enebo has joined #jruby
vtunka has joined #jruby
<eregon>
headius: Hi, are you here?
<headius>
I am!
<eregon>
:)
<headius>
trip prep today but I'll be around
<eregon>
I have a question about @ivars in JRuby
<headius>
ok
<eregon>
(a few questions actually)
<eregon>
Is the indy logic for @ivars used currently?
<headius>
yes
<headius>
the non-indy logic just caches the VariableAccessor
enebo has quit [Quit: enebo]
<headius>
and calls through it
<headius>
indy logic builds a guard based on var accessor's class ID and inlines the access of the field back to use site
<eregon>
I don't find any caller for Bootstrap.ivarGet(), is it hidden?
<headius>
IRBytecodeAdapter7.putField and getField
<headius>
Bootstrap.ivar binds appropriately to set or get
<eregon>
I see :)
<eregon>
What about RubyObjectVar6 and friends, is that still used?
<headius>
they're used if you pass -Xreify.variables=true
<headius>
I've never spent time to make sure they're 100% safe to have on all the time, but they do reduce object overhead from IRubyObject[] ivar table
<headius>
they need to not be an inheritance hierarchy too...I guess most VMs don't cope with that well
<eregon>
I see, so not by default
<headius>
not yet
<headius>
could be a big help if I spent the time
<headius>
could also have some types that double fields and have primitives
camlow32_ has quit [Read error: Connection reset by peer]
camlow325 has joined #jruby
<headius>
chrisseaton: there was an ugly wart in the whitequark parser that made me think of truffle
<headius>
they used a hash to select the escape character for various incoming parsed tokens, but it wasn't in a constant...the literal hash was constructed every time
<chrisseaton>
headius: is the hash small?
camlow325 has quit [Read error: Connection reset by peer]
<chrisseaton>
sometimes literal and constructed every time is far better - it's not going to escape! a hash in a constant is a pain as you can't constant fold it (ironically)
camlow325 has joined #jruby
<chrisseaton>
maybe if it was frozen, but we don't take frozen into account at the moment
<headius>
chrisseaton: it was maybe ten entries, literal strings
<chrisseaton>
I think we stop unrolling hash access for literals at 4
<headius>
seems sensible
rsim has joined #jruby
<GitHub70>
[jruby] nirvdrum pushed 2 new commits to master: https://git.io/vz57g
<GitHub70>
jruby/master f7cb856 Kevin Menard: [Truffle] Avoid unnecessary byte copying in Array#pack.
<GitHub70>
jruby/master 018eeac Kevin Menard: [Truffle] Improved reuse of existing descendent byte[] calculations when flattening a rope.
robbyoconnor has quit [Quit: Konversation terminated!]
robbyoconnor has joined #jruby
<headius>
huh, well I fixed variable reification and things pass
<jensnockert>
Hello, I'm getting very weird backtraces from JRuby (line numbers are wrong), especially in ArgumentError exceptions, is this a known problem with a known workaround? My Googlefu is too weak.
<jensnockert>
I'm trying to run ActiveRecord, so it might be doing something silly.
<GitHub89>
jruby/master a447e47 Kevin Menard: [Truffle] Use a StringAppendPrimitiveNode directly in InterpolatedStringNode....
<jensnockert>
headius: jruby 9.0.4.0 (2.2.2) 2015-11-12 b9fb7aa Java HotSpot(TM) 64-Bit Server VM 24.79-b02 on 1.7.0_79-b15 +jit [darwin-x86_64] more specifically.
lanceball is now known as lance|afk
<jensnockert>
I'll spew binding.pry all over the code I guess and try to narrow it down.
<jensnockert>
Hm… it might just be that it takes line numbers for the last frame from previous ones.
dfr has joined #jruby
<jensnockert>
No, that doesn't seem to be the case either.
bbrowning is now known as bbrowning_away
<headius>
jensnockert: can you show me?
<headius>
it would be a pretty high priority item if normal Ruby errors were way off on line numbers
<jensnockert>
I can try my best.
<jensnockert>
I think it is only ArgumentError, or rather, that is the only one I have seen.
lance|afk is now known as lanceball
<headius>
tip: pass -Xbacktrace.style=full to show both Java and Ruby line numbers
<torrancew>
headius: I have, I guess I'm mostly wondering, do I need to actually embed a scriptingcontainer/runtime/etc as with typical redbridge stuff, or is there a JVM class that gives me access to the compiler, specifically?
skade has joined #jruby
cpuguy83 has joined #jruby
<headius>
well you need JRuby available, but how you embed it is up to you
<headius>
redbridge or the javax API
<torrancew>
got it, so I must embed to use the compiler
<torrancew>
(mostly looking to achieve the results of "jrubyc <file>" from Java, and not rely on jrubyc being in $PATH, etc
<headius>
ahh ok
<headius>
yeah
<torrancew>
thanks, headius!
subbu|lunch is now known as subbu
<headius>
subbu: I did some work to get unboxing going again
<headius>
mandelbrot doesn't box anything now...booleans, fixnums, and floats are all primitive
<headius>
no guards of course
<subbu>
nice.
hobodave_ has joined #jruby
<subbu>
as far as i remember, the current code doesn't work well when one path requests float and another request fixnum for the same value ... i had a plan for using a convert operation for those scenarios, to be done later.
yfeldblum has quit [Ping timeout: 240 seconds]
<subbu>
how did you work around that?
hobodave has quit [Ping timeout: 264 seconds]
<headius>
subbu: that doesn't happen in mandelbrot
<headius>
but yeah I ran into that case...we need all permutations of float/fixnum conversion and operations
jensnock_ has quit [Read error: Connection reset by peer]
jensnockert has joined #jruby
<subbu>
yes. i haven't had an opportunity to go back to that code ... i lost some work-in-progress code in some git snafus a few months back. anyway, will circle around to it later.
jensnockert has quit [Read error: Connection reset by peer]
jensnockert has joined #jruby
hobodave_ has quit [Quit: Computer has gone to sleep.]
pawnbox has joined #jruby
pawnbox has quit [Ping timeout: 245 seconds]
jensnock_ has joined #jruby
rsim has joined #jruby
jensnockert has quit [Ping timeout: 250 seconds]
bbrowning is now known as bbrowning_away
yfeldblum has joined #jruby
lanceball is now known as lance|afk
rsim has quit [Quit: Leaving.]
jensnockert has joined #jruby
tcrawley is now known as tcrawley-away
jensnock_ has quit [Ping timeout: 260 seconds]
<GitHub190>
[jruby] chrisseaton pushed 1 new commit to master: https://git.io/vzFLb
<GitHub190>
jruby/master 007e04e Chris Seaton: [Truffle] Use bin/jruby instead of a system ruby to run mspec.
<GitHub5>
[jruby] chrisseaton pushed 1 new commit to truffle-head: https://git.io/vzFtf
<GitHub5>
jruby/truffle-head f3460a2 Chris Seaton: Merge branch 'master' into truffle-head