SynrG has quit [Read error: Connection reset by peer]
SynrG has joined #jruby
_dabradley has quit [Ping timeout: 255 seconds]
lupine has quit [Ping timeout: 272 seconds]
lupine has joined #jruby
_dabradley has joined #jruby
calavera has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
calavera has joined #jruby
tenderlove has joined #jruby
tenderlove has quit [Remote host closed the connection]
phrinx has joined #jruby
phrinx_ has joined #jruby
zorak8 has quit [Ping timeout: 244 seconds]
phrinx has quit [Ping timeout: 256 seconds]
tlarevo has quit [Remote host closed the connection]
tlarevo has joined #jruby
tlarevo has quit [Remote host closed the connection]
tlarevo has joined #jruby
subbu has joined #jruby
projectodd-ci has quit [Remote host closed the connection]
deobalds has joined #jruby
projectodd-ci has joined #jruby
calavera has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
iamjarvo has joined #jruby
robbyoconnor has joined #jruby
aewffwea has joined #jruby
<aewffwea>
Is there any difference between ruby and jruby?
johnsonch is now known as johnsonch_afk
<Antiarc1>
Yes, ruby is written in C, jruby in Java :P
tlarevo has quit [Remote host closed the connection]
Antiarc1 is now known as Antiarc
tlarevo has joined #jruby
<Antiarc>
I suspect what you're asking is "are they interoperable", and yes, they are. You typically can't use gems with C extensions in jruby, and you can't use gems with Java extension in MRI ruby, but in terms of the language and API, it's all the same
towski has quit [Read error: Connection reset by peer]
<aewffwea>
Antiarc: That makes me want to use ruby :p
<Antiarc>
jruby has some support for c extensions, but in my experience it's fiddly and you should avoid gems with c extensions. For all the popular cases there are Java analogs anyhow
tenderlove has quit [Ping timeout: 256 seconds]
towski has joined #jruby
<Antiarc>
The reasons you might want to use jruby though are a) true concurrency, which you can't get in MRI Ruby due to the GIL, b) the JVM, which is extremely mature, particularly in regards to GC, whereas MRI is still quite young in that regard, and c) Java interop support - it's extremely easy to use Java code or packages from jruby.
<Antiarc>
The reason you might want to use MRI are basically "it's the reference implementation" and C extension support
<Antiarc>
And if you want 2.0/2.1 features, I guess
<Antiarc>
The current jruby stable targets 1.9, though jruby 9000 will bring it up to par
<Antiarc>
But given that some of the biggest features in 2.0/2.1 are GC-related, Jruby already has something of an edge there
benlovell has quit [Ping timeout: 272 seconds]
tlarevo has quit [Ping timeout: 258 seconds]
anaeem1_ has joined #jruby
havenwood has joined #jruby
erikhatcher has quit [Quit: erikhatcher]
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
subbu has quit [Ping timeout: 265 seconds]
tlarevo has joined #jruby
phrinx_ has quit [Remote host closed the connection]
JohnBat26 has joined #jruby
benlovell has joined #jruby
skade has joined #jruby
benlovell has quit [Ping timeout: 255 seconds]
skade has quit [Quit: Computer has gone to sleep.]
deobalds has quit [Quit: Computer has gone to sleep.]
purplefox has joined #jruby
deobalds has joined #jruby
<aewffwea>
If I modify a Java class from Ruby, let's say I modify a method... will that modification affect Java next time that method is called from Java?
deobalds has quit [Quit: Computer has gone to sleep.]
diegoviola has quit [Read error: Connection reset by peer]
benlovell has joined #jruby
benlovell has quit [Ping timeout: 245 seconds]
wojtczak has joined #jruby
benlovell has joined #jruby
mister_solo has joined #jruby
anaeem1_ has quit [Remote host closed the connection]
anaeem1 has joined #jruby
thsig has joined #jruby
mister_solo has quit [Ping timeout: 240 seconds]
wojtczak has quit [Quit: Leaving...]
thsig has quit [Ping timeout: 260 seconds]
jhe_ has joined #jruby
thsig has joined #jruby
tlarevo has quit [Read error: No route to host]
tlarevo has joined #jruby
thsig has quit [Ping timeout: 260 seconds]
SynrG has quit [Ping timeout: 244 seconds]
ephemerian has joined #jruby
noop has joined #jruby
SynrG has joined #jruby
Hobogrammer has quit [Ping timeout: 240 seconds]
jhe_ has quit [Ping timeout: 255 seconds]
elia has joined #jruby
thsig has joined #jruby
marr has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
drbobbeaty has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
deobalds has joined #jruby
rsim has joined #jruby
deobalds has quit [Quit: Computer has gone to sleep.]
yfeldblum has quit [Ping timeout: 265 seconds]
toshetm has joined #jruby
postmodern has quit [Quit: Leaving]
mister_solo has joined #jruby
DJGummikuh has quit [Remote host closed the connection]
benlovell has quit [Ping timeout: 245 seconds]
benlovell has joined #jruby
havenwood has quit [Remote host closed the connection]
drbobbeaty has joined #jruby
thsig has quit [Remote host closed the connection]
thsig has joined #jruby
skade has joined #jruby
etehtsea has joined #jruby
zorak8 has joined #jruby
tenderlove has joined #jruby
thsig_ has joined #jruby
tenderlove has quit [Ping timeout: 265 seconds]
thsig has quit [Ping timeout: 256 seconds]
zorak8 has quit [Ping timeout: 240 seconds]
thsig_ has quit [Remote host closed the connection]
tcrawley-away is now known as tcrawley
zorak8 has joined #jruby
iamjarvo has joined #jruby
erikhatcher has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
tcrawley is now known as tcrawley-away
benlovell has quit [Ping timeout: 250 seconds]
mister_solo has quit [Ping timeout: 272 seconds]
rsim has quit [Quit: Leaving.]
mberg has quit [Read error: Connection timed out]
Guest77449 has joined #jruby
benlovell has joined #jruby
deobalds has joined #jruby
<headius>
aewffwea: no, the change you make is only visible to Ruby
anaeem1 has quit [Remote host closed the connection]
mister_solo has joined #jruby
tlarevo has quit [Remote host closed the connection]
thsig has joined #jruby
vyorkin has joined #jruby
nirvdrum has joined #jruby
johnsonch_afk is now known as johnsonch
aewffwea has quit [Ping timeout: 250 seconds]
rsim has joined #jruby
zorak8 has quit [Ping timeout: 255 seconds]
rsim1 has joined #jruby
rsim has quit [Ping timeout: 265 seconds]
benlovell has quit [Ping timeout: 260 seconds]
imperator has joined #jruby
skade has joined #jruby
deobalds has quit [Quit: Computer has gone to sleep.]
<headius>
subbu: it's the move of current-scope/module instrs after arg receive apparently
<headius>
I'll check that reverting fixes it
<subbu>
weeeeiiird.
<Aethenelle>
headius: the loop is being caused by instanceSuper being supplied the delegate class... is there a way to turn off JIT to make debugging better?
<subbu>
Aethenelle, -X-C
<Aethenelle>
gotta switch tracks... thanks
Aethenelle has quit [Client Quit]
deobalds has joined #jruby
Aethenelle has joined #jruby
havenwood has joined #jruby
iamjarvo has joined #jruby
<headius>
subbu: yeah weird, but that is the one
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to master: http://git.io/Vxmgsg
<JRubyGithub>
jruby/master 74617f0 Charles Oliver Nutter: Formatting.
JRubyGithub has left #jruby [#jruby]
<subbu>
ok, for now, reverting is good. given that you still don't exploit that pattern.
<subbu>
headius, but, wil have to investigate what is causing that to fail later.
<subbu>
unless you or enebo get to it ... and it turns out to be a simple thing.
<headius>
yeah I will poke at it for a few minutes
<subbu>
ah .. i wonder if that arg construction requires those vars already set up.
<headius>
tries to look up constant before static scope is prepped
<subbu>
k.
<headius>
I will revert...that's not a trivial thing to fix
<subbu>
right.
<subbu>
maybe worth adding that reduced test somewhere :)
<headius>
good idea
<subbu>
it is jruby specific
<enebo>
so I only caught some of this but I can see what is wrong
Aethenelle has quit [Quit: Aethenelle]
<subbu>
they will be worth their weight in gold later whe nwe start fixing and opting things :)
<enebo>
Is it really a big deal to put scope at the top and still arity split?
<subbu>
headius found a broken test suite
<enebo>
subbu: yeah opt arg population instrs before recv_scope right?
<subbu>
and reduced it to 2fbbf6066d556dbcbf273be57a66b655d8347c65 where i had moved current-scope/module above args.
<subbu>
and that is the test case above that bombs it.
<enebo>
subbu: above you mean after?
<subbu>
before
<subbu>
see 2fbbf6066d556dbcbf273be57a66b655d8347c65
<subbu>
oh, after :) yes.
<subbu>
i am waking up.
<headius>
heheh
<subbu>
you are already quite awake.
<enebo>
HAHAHA what!?!?!
<headius>
enebo: I'm reverting and adding a test
<headius>
for what arity-fixing we're doing now it isn't important
<enebo>
This is weird. So opt-arg is not dependent on scope of method body?
<headius>
it may not be important in general if I create a generic mechanism to inform the JIT that we're in a specific arity, so it handles args right
<headius>
it was an opt arg that accesses a constant
<headius>
constant lookup goes through scope that wasn't loaded yet
<enebo>
yeah I guess that is what I was asking
<enebo>
I never thought about it
<enebo>
I find that actually mildly weird …
<enebo>
At least it breaks my mental model
skade has quit [Quit: Computer has gone to sleep.]
<headius>
optional args can have side effects and make calls, so they have to operate in a working context
Aethenelle has joined #jruby
<enebo>
headius: yeah I know I just thought it would be the methods
<subbu>
if we tweak some of our constant lookup logic ... earlier it was using implicit values from the stack .. and in the current ir mode, it uses explicit values setup ... so, if you go back to implicit, it'll work. but, it seems wrong.
<headius>
that's one reason I never split them in the old JIT...but emitting N versions as in the new JIT may be fine
<enebo>
headius: It would make more sense to me
<headius>
methods?
<enebo>
headius: but we also need to be aware that if we change this to the prologue style of binding we may ultimately want it needs to work
<enebo>
headius: methods scope
elia has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 2 new commits to master: http://git.io/gKqe7w
<JRubyGithub>
jruby/master 0d38e26 Charles Oliver Nutter: Revert "Add current-scope/module instrs after arg receive"...
<JRubyGithub>
jruby/master 579b8f2 Charles Oliver Nutter: Add test case for reverted scope-load instruction moves.
JRubyGithub has left #jruby [#jruby]
anaeem1 has joined #jruby
<enebo>
Could someone add a crystal clear comment on ordering
<headius>
I can
<enebo>
I think in 6 months I will look at it and think it should be moved before again
<headius>
hopefully this will get those last three suites green again
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to master: http://git.io/Di_S7w
<JRubyGithub>
jruby/master c93c7bb Charles Oliver Nutter: Add comment explaining why scope loads can't move later.
JRubyGithub has left #jruby [#jruby]
joast has joined #jruby
Aethenelle_ has joined #jruby
marr has quit [Ping timeout: 244 seconds]
Aethenelle has quit [Ping timeout: 256 seconds]
Aethenelle_ is now known as Aethenelle
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
anaeem1 has quit [Remote host closed the connection]
e_dub has joined #jruby
anaeem1_ has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci created jruby-1_7_16 from 1.7.16 (+0 new commits): http://git.io/PeGc7Q
JRubyGithub has left #jruby [#jruby]
anaeem1_ has quit [Remote host closed the connection]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
iamjarvo has quit [Max SendQ exceeded]
iamjarvo has joined #jruby
Aethenelle has quit [Read error: Connection reset by peer]
deobalds has quit [Quit: Computer has gone to sleep.]
anaeem1 has joined #jruby
Aethenelle has joined #jruby
Aethenelle has quit [Read error: Connection reset by peer]
iamjarvo has quit [Client Quit]
Aethenelle has joined #jruby
Aethenelle_ has joined #jruby
Aethenelle has quit [Ping timeout: 245 seconds]
Aethenelle_ is now known as Aethenelle
anaeem1 has quit [Remote host closed the connection]
triple_b has quit [Read error: Connection reset by peer]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to jruby-1_7_16: http://git.io/72IlnA
<JRubyGithub>
jruby/jruby-1_7_16 b8e1016 Charles Oliver Nutter: "keep the entity size within the limitation"...
<headius>
enebo: you can review 1.7.16.1 branch now
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to jruby-1_7: http://git.io/QHDNLQ
<JRubyGithub>
jruby/jruby-1_7 566b9d6 Charles Oliver Nutter: Update 1.8.7 rexml to latest.
JRubyGithub has left #jruby [#jruby]
<enebo>
headius: yeah
erikhatcher has joined #jruby
<headius>
eregon: it looks like you updated some date/time .rb files but a bunch of the indentations are off
<enebo>
headius: Is the branch name exactly that?
<headius>
I'm trying to update jruby-1_7 stdlibs from our ruby fork and most of the changes are whitespace... some sections of code are indented only one space
<headius>
enebo: should be jruby-1_7_16
<enebo>
ok
<headius>
eregon: or maybe you updated jruby and our ruby fork is wrong?
<enebo>
Am I the only person looks at last line returns in ruby methods and think it is a bug
<headius>
I use it for simple cases
<enebo>
def foo; return 1; end?
<headius>
oh, you mean explicit return?
<enebo>
My mind is so conditioned in Ruby to think return is only for early returns
<enebo>
yeah
<headius>
ahh right
<headius>
yeah I double-take too
<enebo>
That entity expansion path uses a return
<enebo>
patch
<enebo>
headius: Did you try all three of these?
<headius>
I tried all three modes with the tests I could run
<enebo>
headius: They look fine to me so far as I can tell
<headius>
something's odd about MRI's test_entity.rb...it doesn't seem to run
<headius>
on MRI either
<enebo>
headius: ok good enough for me.
<headius>
ok
thsig has joined #jruby
<headius>
enebo: updating stdlibs on 1.7 and master now
<enebo>
headius: not a bad idea but we might want to review changes on 1.7 to make sure we are not accidentally nuking jruby behavior
<enebo>
headius: I cannot be sure we have not accidentally committed to lib/ruby/*/*.rb
<headius>
enebo: I always review them
<enebo>
headius: of course you do
<enebo>
headius: just saying it just in case :)
<headius>
eregon: can you confirm pathname.rb is correct on jruby-1_7 branch? I need to update it into our fork
yfeldblum has joined #jruby
yfeldblum has quit [Remote host closed the connection]
yfeldblum has joined #jruby
elia has quit [Quit: Computer has gone to sleep.]
thsig has quit [Remote host closed the connection]
tcrawley is now known as tcrawley-away
tcrawley-away is now known as tcrawley
JRubyGithub has joined #jruby
<JRubyGithub>
jruby/jruby-1_7 2bd2879 Charles Oliver Nutter: "keep the entity size within the limitation"...
<JRubyGithub>
jruby/jruby-1_7 9ff5675 Charles Oliver Nutter: Update stdlib from ruby 1.9.3 HEAD.
<JRubyGithub>
[jruby] jrubyci pushed 4 new commits to jruby-1_7: http://git.io/3IajOg
<JRubyGithub>
jruby/jruby-1_7 8cfc303 Charles Oliver Nutter: Update stdlib from 1.8.7 HEAD.
JRubyGithub has left #jruby [#jruby]
<headius>
enebo: another pair of eyes on those updates might help, but they looked pretty mundane to me
triple_b has joined #jruby
marr has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
jruby/jruby-1_7 234acf6 Charles Oliver Nutter: Don't copy openssl files anymore; jruby-openssl gem is back.
JRubyGithub has left #jruby [#jruby]
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to jruby-1_7: http://git.io/oUoQ_g
thsig has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to jruby-1_7: http://git.io/bC7LTA
<JRubyGithub>
jruby/jruby-1_7 0732187 Charles Oliver Nutter: Oops, actually commit the openssl change to tools.
JRubyGithub has left #jruby [#jruby]
benlovell has joined #jruby
<headius>
enebo: I guess we should pul the trigger on 2.2 soon
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to master: http://git.io/uFWKZw
<JRubyGithub>
jruby/master bacbde8 Charles Oliver Nutter: Update stdlib from ruby 2.1 HEAD.
JRubyGithub has left #jruby [#jruby]
<enebo>
headius: oh updating the stdlib?
<enebo>
headius: yeah
<headius>
I'll have to apply our patch to 2.2 stdlib and see how it looks
benlovell has quit [Ping timeout: 272 seconds]
ludyte has quit [Quit: ludyte]
skade has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to master: http://git.io/wTQZRQ
<enebo>
nirvdrum: Did you verify we make it further yet on your app?
tenderlove has joined #jruby
<nirvdrum>
I'll do that momentarily.
<enebo>
nirvdrum: cool. I am hoping we make it past your super problem
<nirvdrum>
Announcing the shutdown of Mogotest and fighting with HireRight has kept me busy, unfortunately :-/
<enebo>
nirvdrum: np
<nirvdrum>
The good news is once Mogotest is gone, you don't have to fix any of these bugs!
<mpapis>
enebo, just commited change of all older checksums urls to https
<enebo>
mpapis: yeah great
dudleyf has joined #jruby
bbrowning has joined #jruby
<Aethenelle>
headius: I just tried to debug my stuff with JIT off... Stack overflow went away :(
<nirvdrum>
enebo: I can't get past loading gems now.
<nirvdrum>
I get a nice NPE.
<headius>
NPE?!
<headius>
Aethenelle: oh, that's good! Then it's something I can fix
<nirvdrum>
ERROR: Got exception java.lang.ClassCastException: org.jruby.RubyNil cannot be cast to org.jruby.parser.StaticScope but instr %v_0 = class_var_module(%v_1, %self) is not supposed to be raising exceptions!
<nirvdrum>
LoadError: load error: active_support -- java.lang.ClassCastException: org.jruby.RubyNil cannot be cast to org.jruby.parser.StaticScope
<nirvdrum>
require at org/jruby/RubyKernel.java:926
<headius>
I fixed that last night
<Aethenelle>
headius: then I'll focus on the other failures...
<headius>
but we need a green build...which should come today
<nirvdrum>
headius: Okay. That was a fresh install via ruby-build.
<enebo>
headius: well unfortunately we might need to toggle some of those tags I added off again
towski has quit [Ping timeout: 244 seconds]
<nirvdrum>
Okay. That is fixed on master.
<enebo>
headius: interpreted specs now takes >50 minutes from options processing
<nirvdrum>
Starting a rails console works. Starting rspec I get the NPE when loading things.
<nirvdrum>
I'll try to track that down.
<enebo>
nirvdrum: yeah we turned on ACP and a couple of things fell out…but good news I think it also fixed your problem
<enebo>
headius: also funny was my ensure block bug I fixed the other day was a virtually identical error
<nirvdrum>
Okay. Seems to actually come from running through bundle exec.
<nirvdrum>
And bundler eats backtraces.
<nirvdrum>
I wonder if it's too late to switch to polisci.
* subbu
switched in jruby to see nirvdrum wonder about polisci
<enebo>
nirvdrum: only if you are sociopath since hacking involves an axe
<nirvdrum>
I lack the constitution to put up with this stuff for another 35 years, I'm afraid.
<enebo>
the software problem will be solved by then
<enebo>
jetpacks and shit
<subbu>
i want what enebo has been smoking lately.
<enebo>
ibuprofin
<subbu>
:)
<nirvdrum>
For the life of me, I don't understand why bundler, a tool used almost exclusively by developers, finds it helpful to throw away backtraces.
<nirvdrum>
So, I had the same exact problem on Aug. 26.
<nirvdrum>
And apparently I couldn't type "txt" correctly.
subbu has joined #jruby
<nirvdrum>
And I'm pretty sure subbu fixed it in roughly that timeframe.
<subbu>
what did i fix?
<enebo>
nirvdrum: The method that is occurring in is mysterious to me
<enebo>
subbu: perhaps you just turned off ACP
<nirvdrum>
I don't recall which, but taking a quick look there was: 5d4636b6c2c01733a8fa655965038f8f7f311f8b, 5d4636b6c2c01733a8fa655965038f8f7f311f8b, f1eca7a9a4e8cdcdb2aee53d6d54e35d972d748f, and 4626537bec2c5899d7cd3880e7f925b7cab9286a
<chrisseaton>
can you create me bench9000 and give me push rights? we can rename it later if needs be
<subbu>
nirvdrum, can you rerun with -X-C (so JIT is off)?
<nirvdrum>
Same problem.
robbyoconnor has joined #jruby
pietr0 has joined #jruby
wojtczak has joined #jruby
<subbu>
ok .. i think enebo is figuring this out ... he may have a fix soon.
iamjarvo has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
<yopp>
chrisseaton, just out of curiosity, what about compativility? At least for 9k? I have massive ruby app, made from dozen of in-house gems. it's right now runs on 1.9, but I'm thinking about giving jruby another try.
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jrubyci pushed 1 new commit to master: http://git.io/m2rBqg
<JRubyGithub>
jruby/master ea47cab Thomas E. Enebo: Re-add important line? :)
JRubyGithub has left #jruby [#jruby]
<yopp>
•compatibility
wojtczak has left #jruby [#jruby]
<enebo>
nirvdrum: ^ Try it with that commit again
_elia has joined #jruby
<subbu>
nirvdrum, your memory of an old fix helped us nail this rapidly ... but we need to figure out a regression test for this .. and possibly clean up some code.
<chrisseaton>
yopp: not sure what you mean sorry? compatibility of the benchmark suite? or Truffle?
elia has quit [Ping timeout: 245 seconds]
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<nirvdrum>
I suppose it helps if I pull the change down.
<yopp>
chrisseaton, thing is, I might have some time to invest to port current MRI app, to jruby. Is there will be any need to "upgrade" of any sort to make it compatible with 9k?
<yopp>
Or it will be drop-in replacement?
<enebo>
chrisseaton: I think yopp wants to know which version of Ruby that Ruby truffle plans ot support. I am guessing 2.2 like our IR implementation
<yopp>
enebo, +/- yeah
<enebo>
yopp: 9000 is targeting 2.2 and final around end of year. It is unclear how far truffle backend will be by that release
<chrisseaton>
yopp: we're going to align with JRuby's policy, which is to keep moving with the latest Ruby
<yopp>
uh
<chrisseaton>
yopp: but it sounds like you just want to run Ruby code, so I wouldn't worry about Truffle for the next six months at least - it's not going to be useful for you yet I'm afraid
<nirvdrum>
enebo, subbu: That did it.
<enebo>
nirvdrum: cool. So no errors now?
<nirvdrum>
Not so fast there. I said it fixed the glitch.
<yopp>
ah, half year.
<enebo>
nirvdrum: I man can hope
<nirvdrum>
enebo: My StackOverflow is still present.
<subbu>
ok, that hasn't changed.
<nirvdrum>
Correct.
<subbu>
nirvdrum, did you ever give me a snippet to test it or was it for a different bug?
<nirvdrum>
Although, the stacktraces seem to be nicer.
<nirvdrum>
I'm guessing reification from JIT.
bbrowning is now known as bbrowning_away
<nirvdrum>
subbu: I haven't been able to reduce the problem. It only fails when I run every spec in this file. If I try running just one of the specs that's failing in isolation, it passes.
<subbu>
and is that a public codebase, i.e. can i just get the whole unreduced code?
<nirvdrum>
Negative.
<subbu>
ah, ok.
<enebo>
nirvdrum: if you can generate IR debug output we can see the method it is failing with
<subbu>
stack overflow :)
<subbu>
so, will be voluminous.
<Aethenelle>
nirvdrum: try running two of the failing specs?
<Aethenelle>
at the same time...
<subbu>
we really need a way to turn on debugging selectivley ..
<enebo>
subbu: yeah but it is probably less than 1500 frames
<subbu>
enebo, maybe your trick still works?
<enebo>
subbu: I am guessing the test output will be way larger than the overflow
<enebo>
subbu: If he can isolate it enough then perhaps
<enebo>
I guess I recomment using -X-C and -Xir.debug=true and redirect all stdout and stderr to a file
<nirvdrum>
I believe the super call ends up in an infinite loop.
<enebo>
nirvdrum: but it probably only runs a couple of thousand calls deep
<enebo>
nirvdrum: we will just look before the nonsense starts and what precedes the super death
<nirvdrum>
Before it overflows the stack?
<enebo>
nirvdrum: if we are saving to a file the history will be there and the overflow should kill the process
<enebo>
nirvdrum: Or you can just kill the process at that point and we can look right before the super recusion to see which source and instrs are around it
lanceball is now known as lance|afk
<subbu>
enebo, i wonder if it is just some bdining/frame not getting pushed causing the super to resolve back to the same frame.
<enebo>
subbu: makes sense…why though :)
<yopp>
chrisseaton, but if app will be compatible with MRI2.2, it will be compatible with 9k, right?
<subbu>
but that should be easily reproducible .. unlike what nirvdrum is reporting.
pietr0 has quit [Ping timeout: 264 seconds]
<chrisseaton>
yopp: for most purposes yes - there are some edge cases that most people never have to deal with (ObjectSpace and things)
<yopp>
can you ellaborate on "things"? :)
<yopp>
OS is not a case, but "things" easily can be
<chrisseaton>
ObjectSpace and set_trace_func are slower, C extensions aren't supported (there used to be some support)
<chrisseaton>
yopp: if you give it a go, we can help you workaround any problems here
<yopp>
ah. no C extensions -> not tinytds -> no fun :|
<subbu>
yopp are you asking about the non-truffle runtime in 9k or the truffle runtime in 9k?
<nirvdrum>
Aethenelle: Passes with 2.
<nirvdrum>
I'll just generate this massive file.
zorak8 has joined #jruby
<chrisseaton>
subbu: I think yopp is asking about normal JRuby now
<subbu>
ah, ok.
<yopp>
yep
<yopp>
because for me there not so much diffirence to port it to mri 2.x or to jruby. But from current point of time, jruby looks much better in short term
<Aethenelle>
nirvdrum: try going through the good ones, three to a trial: passing[x] failing failing ?
<yopp>
at least all this truffle stuff in 9k sounds very-very impressive, and if you will make at least half of ideas in production, this might give good performance boots almost for free
_elia has quit [Ping timeout: 256 seconds]
<nirvdrum>
Aethenelle: Unfortunately, I don't have the cycles to try all these combinations. If enebo is willing to dig through the IR log, I'll just generate that.
<yopp>
but it's allways a cost, in my case cost of "porting". it's kinda huge application from "enterprise" world, and it will require some time to properly test it
<yopp>
last time (~2 years ago) we had CI with rbx and jruby, but it never ever became green :)
<Aethenelle>
nirvdrum: k. that works too
<yopp>
and at some point we gave up on alternative implementations.
<enebo>
nirvdrum: yeah I will look. I think with a decent log of what is executing we can reduce it
<nirvdrum>
430 MB so far.
diegoviola has joined #jruby
elia has joined #jruby
johnsonch is now known as johnsonch_afk
tcrawley is now known as tcrawley-away
benlovell has joined #jruby
pietr0 has joined #jruby
<enebo>
nirvdrum: if it ends up too large and you can defer when to turn on output until later you can add a line like; org.jruby.RubyInstanceConfig.IR_DEBUG=true
<nirvdrum>
enebo: IIRC, it compresses pretty well.
<nirvdrum>
But I'm at 1.5 GB right now.
<enebo>
nirvdrum: This allows you to conditionally remove some. like if you know first test you can probably remove all bootstrapping of rails + app
<enebo>
nirvdrum: but I guess we shall see
benlovell has quit [Ping timeout: 256 seconds]
<Antiarc>
How far out is 9k now, approximately?
<enebo>
Antiarc: well we are hoping for a preview around Rubyconf and a first full release at end of year
<Antiarc>
Nice :D
<enebo>
Antiarc: but we need to temper that with the progress we make :)
<enebo>
Things are coming together though
<Antiarc>
I've basically committed to jruby for all my Ruby work for the forseeable future, so I'm pretty stoked to see the progress ya'll are making :)
<nirvdrum>
enebo: I can start over. A *lot* of this is bootstrapping rails.
<Antiarc>
yopp: There are myriad Java libs to talk to mssql though, aren't there?
<Antiarc>
So while you couldn't use tinytds directly, you could use one of a billion Java libs
<enebo>
nirvdrum: so if you know a nice safe point or perhaps you know order of most tests you can enable it much later in the test run
Aethenelle has quit [Quit: Aethenelle]
<yopp>
Antiarc, and 0 time to port it from tinytds to this libs.
<nirvdrum>
enebo: Well, it just finished. Want me to regen?
<Antiarc>
Well, yeah :)
calavera has joined #jruby
<yopp>
Antiarc, my approach is maximize benefit from minimal moves
<yopp>
•is to
<enebo>
nirvdrum: It might be worth a second file with you adding that somewhere later in the run
<enebo>
nirvdrum: If you know spec order and about which spec it starts at you can probably reduce the size a lot
<enebo>
nirvdrum: yeah it is the case Time which is boming
<enebo>
err when Time
<nirvdrum>
Well, let me know if you need me to try anything.
<enebo>
nirvdrum: It looks like module being passed to super is itself
<nirvdrum>
Back in ActiveSupport you mean?
<enebo>
nirvdrum: yeah but I am still grokking this
<enebo>
nirvdrum: so we get an instance of something which hits the super you highlight and it spins away
<enebo>
nirvdrum: subbu is also probably going to look at this tonight but if we cannot figure out what is wrong you could maybe print from info about the object in question like ancestors
subbu has joined #jruby
<nirvdrum>
enebo: Is it valid for super to ever be "false" ?
<nirvdrum>
Not the String but singleton of FalseClass.
<enebo>
nirvdrum: umm
<enebo>
nirvdrum: You mean return it?
<nirvdrum>
NoMethodError: undefined method `ancestors' for false:FalseClass
<nirvdrum>
When I print "super.ancestors" in that ActiveSupport class.
<enebo>
nirvdrum: That wasn’t where I mean exactly
<enebo>
nirvdrum: I was thinking in the sequel method before the case
<nirvdrum>
That's fine. I just found this to be odd.
<enebo>
but you probably actually want self.ancestors
<nirvdrum>
Well, I did that too.
<nirvdrum>
I was just surprised that super could evaluate to false.
<enebo>
nirvdrum: super can return any value and it is for ===
<enebo>
nirvdrum: so I would expect it to be true or false generally
<nirvdrum>
Ahh, right.
<nirvdrum>
Too much time in other langs.
<enebo>
yeah
<enebo>
nirvdrum: p anceostors should be good though
elia has quit [Ping timeout: 255 seconds]
thsig has joined #jruby
r0bby has joined #jruby
<nirvdrum>
enebo: The ancestors within the sequel timezones call are always [Sequel, Sequel::SQL::Constants]