<GitHub28>
jruby/master c9e46b7 Charles Oliver Nutter: Use a defensive temp var when pulling possibly-undefined opt args....
<GitHub85>
[jruby] headius closed issue #3094: NoMethodError being reported as "UndefinedValue.java:52:in `undefinedOperation': java.lang.RuntimeException: IR compiler/interpreter bug: org.jruby.ir.operands” in jruby-9.0.0.0.pre2 https://git.io/vtMyE
<brocktimus>
Right, this is still a project I'm in planning phase of so I'm just checking early :-).
<brocktimus>
I couldn't find massive doco, but does AOT just work on single file or can I just say "GO DO EVERYTHING!" similar to what I read warbler would do?
<kares_>
jrubyc accepts a blog - just try it out e.g. `jrubyc lib/**/*.rb`
<kares_>
* glob
<brocktimus>
haha nice, I was confused by blog "wait do I need to write a blog about my idea for an answer!?"
<brocktimus>
Thanks kares!
shellac has joined #jruby
<kares_>
brocktimus: well ... blogs about jrubyc are always welcome :)
<brocktimus>
Yeah if project 1) comes off 2) uses right tech and 3) I'm allowed I'd want to write about it :-P
<brocktimus>
Also if ran into issues would love to help jrubyc / warbler
<brocktimus>
wanting to run rails, but reqs of probably obfuscation + windows
<brocktimus>
so idea of being able to compile to a few "binaries" which are just dependent on java seems like a perfect solution. Makes dependency management + deployment much easier too.
n00bdev has joined #jruby
<brocktimus>
Does that sound feasible? Just wondering about gems and bundler and stuff...
davidpgero has joined #jruby
n00bdev has quit []
mattwildig has joined #jruby
mattwildig has quit [Ping timeout: 276 seconds]
robbyoconnor has joined #jruby
vtunka has joined #jruby
donV has quit [Read error: Connection reset by peer]
donValentin has joined #jruby
robbyoconnor has quit [Ping timeout: 244 seconds]
vtunka has quit [Ping timeout: 248 seconds]
donValentin has quit [Ping timeout: 240 seconds]
vtunka has joined #jruby
<kares_>
brocktimus: yy, pretty much works - incidentally just helping out a startup with warbler like deployment
donV has joined #jruby
ITXpander1 has joined #jruby
ITXpander has quit [Ping timeout: 276 seconds]
vtunka has quit [Quit: Leaving]
_whitelogger has quit [Excess Flood]
_whitelogger has joined #jruby
ITXpander has joined #jruby
ITXpander1 has quit [Ping timeout: 248 seconds]
<GitHub92>
[jruby] chrisseaton pushed 1 new commit to master: https://git.io/v25Mt
<GitHub92>
jruby/master f125c8d Chris Seaton: [Truffle] Exclude Rails from the integration tests for now.
<GitHub4>
[jruby] chrisseaton pushed 1 new commit to truffle-head: https://git.io/v25MO
<GitHub4>
jruby/truffle-head 4b05809 Chris Seaton: Merge branch 'master' into truffle-head
<GitHub159>
jruby/master c427786 Petr Chalupa: [Truffle] add back test examples to PE tests when running all
<GitHub92>
[jruby] pitr-ch commented on commit 997e219: c427786 https://git.io/v25Du
<GitHub169>
[jruby] pitr-ch commented on commit 566807e: Because the superclass is changing, but that's flawed since it's `null` before which is never cached. Good catch. https://git.io/v25yc
bb010g has joined #jruby
vtunka has quit [Quit: Leaving]
<GitHub127>
[jruby] chrisseaton commented on commit f125c8d: @pitr-ch this test fails locally and on Travis. Sometimes throws an exception in J+T, sometimes in Rails, sometimes seems to hang. https://git.io/v25SK
camlow325 has quit [Read error: Connection reset by peer]
camlow325 has joined #jruby
camlow32_ has joined #jruby
camlow325 has quit [Read error: Connection reset by peer]
enebo has joined #jruby
mkristian has quit [Quit: This computer has gone to sleep]
donV has quit [Ping timeout: 248 seconds]
cremes has joined #jruby
_whitelogger has quit [Excess Flood]
_whitelogger has joined #jruby
camlow325 has joined #jruby
camlow32_ has quit [Read error: Connection reset by peer]
e_dub has quit [Read error: Connection reset by peer]
e_dub has joined #jruby
evenix has joined #jruby
<evenix>
Hi all
<evenix>
I really want to love jruby but its really hard to love. Whenever I use it I run into 242rf32f2 number of issues that drive me crazy
<evenix>
any tips and tricks you'd like to share to better debug and deal with jruby?
<headius>
evenix: what issues are you running into?
<evenix>
to be honest more than I can mention in one go.. memory leak.. openssl keystore issue.. etc.. I've been working on a client app and all these things keep popping
<evenix>
sometimes the server load corectly.. sometimes it doesnt
<evenix>
its just weird
<evenix>
gem compatibilities...
<evenix>
sometimes it doesn't detect my SessionController or ApplicationController..
<evenix>
it requests me to regenerate a secret_token that is already there...
<evenix>
new_relic is not working
<evenix>
and fail to be detected by newrelic
<evenix>
OpenSSL::SSL::SSLError: Socket closed
tjohnson has joined #jruby
<evenix>
on Rails 3.2.13 , jruby 1.7.12 (2.0.0p195) , java version "1.8.0_72"
<headius>
1.7.12 is really old
<headius>
and 2.0 mode is only experimental in that version
<evenix>
I know..thats the app of my client.. I'm not sure if it would be safe to just upgrade
<evenix>
what do you recommend?
<headius>
they're running with JRuby or you're trying to use JRuby with it?
<headius>
if you really need 2.0+ features, you want JRuby 9.x
<evenix>
the app been using jruby, I'm trying to use 2.0 with it because one of the gem i'm using `google-api-client` requires it (because of key parameters, it returns an error if i'm not using 2.0 )
<evenix>
@headius : which introduced quite a few issues into my app from what I remember. Maybe I should just try to upgrade step by step and I'll have less issues in the end.
<headius>
you're going to have much better luck with jruby 9k series then
<headius>
2.0 mode in JRuby 1.7 was just a few implemented APIs and some of the syntax
<headius>
very incomplete
<evenix>
ok makes sense. Thanks headius
<evenix>
I'll try to upgrade little by little it might just be better overall
<headius>
sounds good :-)
davidpgero has quit [Remote host closed the connection]
<rtyler>
I'm wondering if there's a better worker time gem which supports sub-rubies in MRI and something looking like threads in JRuby?
<headius>
hmmm
<headius>
yeah I remember hearing about one
<rtyler>
IIRC concurrent-ruby doesn't do any forking
lobner has joined #jruby
<lopex>
nunbers!
<evenix>
trinidad doens't seem to like Jruby 9.x
<headius>
unless they're using some special plugins, trinidad can be swapped for puma pretty easily
<headius>
trinidad hasn't been maintained in many years
lobner has quit [Remote host closed the connection]
mattwildig has quit [Remote host closed the connection]
_whitelogger has quit [Ping timeout: 260 seconds]
_whitelogger has joined #jruby
bb010g has quit [Quit: Connection closed for inactivity]
<lopex>
headius: hey, wrt #3632, that test you've come up with does involve threads ?
<headius>
yeah
<headius>
just that script I posted modified to be a unit test somewhere
<lopex>
headius: I'm thinking what if we could interrupt joni to leave it with broken state
<lopex>
headius: yeah, I tested it locally on simple executor without jruby
camlow32_ has joined #jruby
camlow325 has quit [Read error: Connection reset by peer]
<lopex>
headius: and that regexp doesnt use stack so we wont get stack threadlocal cache collision
brightball has quit [Quit: Leaving...]
<GitHub65>
[jruby] chrisseaton pushed 1 new commit to master: https://git.io/v2dA9
<GitHub65>
jruby/master 4892ade Chris Seaton: [Truffle] Storing the materialised frame in the backtrace was always an insane idea.
<GitHub80>
[jruby] chrisseaton pushed 1 new commit to truffle-head: https://git.io/v2dxF
<GitHub80>
jruby/truffle-head fd0865a Chris Seaton: Merge branch 'master' into truffle-head
<evenix>
I'm getting a cache error: undefined class/module Rack::Handler::Servlet::DefaultEnv ?
<evenix>
jruby 9.0.5.0 rack-cache-1.6.1 (on puma-3.0.2-java or trinidad)
tomjoro has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
<evenix>
jruby 9.0.5 and jruby-rack don't seem to work well together
<GitHub194>
[jruby] subbuss commented on commit c9e46b7: Is this extra tmp juggling necessary? Won't the checks in buildLocalAsgn and buildDAsgn fix this? https://git.io/v2FvQ
<bascule>
_____ ____ ___ ____ _ __ ___ _ _
<bascule>
| ___| _ \|_ _| _ \ / \\ \ / / | | |
<bascule>
| |_ | |_) || || | | |/ _ \\ V /| | | |
<bascule>
| _| | _ < | || |_| / ___ \| | |_|_|_|
<bascule>
|_| |_| \_\___|____/_/ \_\_| (_|_|_)
<bascule>
mattwildig has joined #jruby
<GitHub123>
[jruby] headius closed issue #3628: Command line flag '-p' doesn't reflect changes made to $_ https://git.io/vzAxJ
<GitHub54>
jruby/master 1a10a9a Charles Oliver Nutter: Fix ArrayDeref first-time indy call with no value for key.
<GitHub54>
jruby/master b7dc414 Charles Oliver Nutter: Fix -p printing happening before the body. Fixes #3628.
<GitHub54>
jruby/master e8ef0dd Charles Oliver Nutter: Untag working -p spec.
<GitHub83>
[jruby] headius closed issue #3627: Process#spawn error when redirecting error to stdout (jruby 9.0.4.0) https://git.io/vzA5D
<GitHub31>
[jruby] headius closed issue #3622: Regression in method calling on Java objects in 9.0.5.0 (field not calling getField for JDBC object) https://git.io/vz99S
<GitHub194>
[jruby] headius commented on commit c9e46b7: The problem was primarily that ReceiveOptArgInstr returns undefined for unprovided opt args, so assigning that to the target variable made it possible for undefined to leak out. My change does all opt arg logic with a temp var and assigns to the real var only after that has finished. https://git.io/v2Fkw
skade has joined #jruby
e_dub has quit [Read error: Connection reset by peer]
e_dub has joined #jruby
<GitHub74>
[jruby] ptepper opened issue #3711: Passing in command line arguments containing "%20n" causes jruby to abort with error "%n in writable segment detected" https://git.io/v2Fmr
<kares_>
evenix: seems like a marshalling issue - either JRuby one (most likely) or smt about rack-cache not liking Hash extensions
<kares_>
either way you should try to come up with a minimal reproduction case - preferably outside the server
skade has quit [Quit: Computer has gone to sleep.]
n00bdev has joined #jruby
subbu is now known as subbu|afk
n00bdev has quit []
<evenix>
Is ObjectSpace fixed in jruby 9.0.5.0 for delayedjob?
<headius>
evenix: sigh, what is delayedjob using ObjectSpace for?
<chrisseaton>
headius evenix I was looking at this today - doesn't it just use it for pre-forking stuff? And obviously the fork isn't going to work on JRuby anyway
<chrisseaton>
evenix: what I'm saying (and it's referred to in that answer as well) is that even if the use of each_object worked, the next thing I think it would do is try to fork, which isn't going to work either
<chrisseaton>
I thought there was an option to configure this not to fork, and so it wouldn't try to use each_object anyway
<evenix>
one guy suggested to run the worker as a rake task instead and daemonize it in the OS
<GitHub53>
[jruby] headius commented on commit c9e46b7: Yeah, leaning on the fact that we initialize variables to nil, so the first assignment of a variable to itself isn't a problem. Perhaps there's a better way to fix this? The alternatives I came up with were more complex. https://git.io/v2FQr
shellac has joined #jruby
<evenix>
Are you guys using delayedjob w/ jruby or resque?
<headius>
chrisseaton: I don't suppose you guys have fork working yet :-)
<chrisseaton>
headius: nope, and no bright ideas for it either. Potentially the SVM could allow it I guess?
<headius>
in theory yeah
shellac has quit [Client Quit]
evenix has quit [Remote host closed the connection]
e_dub has quit [Read error: Connection reset by peer]
<chrisseaton>
do people come and complain at you for not having fork? or has it never come up?
evenix has joined #jruby
<evenix>
did I miss anything? @headius @chrisseaton
e_dub has joined #jruby
<nirvdrum>
evenix: I've had fantastic results running with sidekiq.
<nirvdrum>
And it's compatible with resque's payloads.
e_dub has quit [Read error: Connection reset by peer]
e_dub has joined #jruby
<nirvdrum>
chrisseaton: Interesting. I don't know if that use of ObjectSpace is new or if I just never noticed it. mperham does support JRuby in Sidekiq, however.
<chrisseaton>
yeah I think it just doesn't take that code path at all if it's JRuby
<GitHub81>
[jruby] headius closed issue #3703: JRuby 9.0.5.0: uninitialized variables assigned to self as default method argument cause org.jruby.ir.operands.UndefinedValue https://git.io/v2aV4
<evenix>
nirvdrum: very very tempting
<evenix>
i love sidekiq
<nirvdrum>
Carpe diem.
<evenix>
Have you experienced any issue at all? @nirvdrum and what version of jruby are you running?
<evenix>
I'm dealing with scheduling email so I like the idea of having the job sitting in the db rather than in-memory
<nirvdrum>
evenix: I haven't tried with 9k yet. But I had very good results with 1.7.x.
<evenix>
I don't know what I would do without you guys and IRC <3
<nirvdrum>
Ahh. Yeah, that could be a problem. redis has persistence options, of course. But you do have a write hole.
<nirvdrum>
And someone ported redis to work with nvram.
<evenix>
any feedback?
<nirvdrum>
Feedback on what?
<nirvdrum>
There's a lot of other options in that area, too. You could just have a schedule sidekiq job that reads jobs in your own format out of the DB.
<nirvdrum>
It'd just be a batch processor.
<evenix>
nirvdrum: this is a really good alternative actually
<evenix>
but then you have the scheduled task > sidekiq > execute the job > ...etc..
<evenix>
I'll just use sidekiq.. I'm already storing thing in the db
<evenix>
ScheduledEmail.. and they are only deleted when the job is done.
<evenix>
lol I just ran `rake jobs:work` instead of `script/delayed_job start` and problem was solved
<nirvdrum>
Sounds like you can go enjoy your weekend now :-)
edub has joined #jruby
bbrowning is now known as bbrowning_away
e_dub has quit [Read error: Connection reset by peer]
e_dub has joined #jruby
norc has quit [Ping timeout: 252 seconds]
norc has joined #jruby
e_dub has quit [Quit: ZZZzzz…]
evenix has quit [Remote host closed the connection]
mkristian has joined #jruby
tcrawley is now known as tcrawley-away
<GitHub135>
[jruby] chrisseaton pushed 3 new commits to truffle-om-dsl: https://git.io/v2be3
<GitHub135>
jruby/truffle-om-dsl bea8aa7 Chris Seaton: Merge branch 'truffle-head' into truffle-om-dsl...
<GitHub135>
jruby/truffle-om-dsl 35d0aa8 Chris Seaton: [Truffle] We don't want implicitCastIntToDouble.
<GitHub135>
jruby/truffle-om-dsl 6972c05 Chris Seaton: [Truffle] OM-DSL processor jar is no more.
<lopex>
nirvdrum: hey, I dont see any dummy related branches in encoding loads
<lopex>
in MRI
<nirvdrum>
lopex: Is that new in 2.3.0?
<lopex>
nirvdrum: the loading ?
<lopex>
nirvdrum: I dont see anything new actually
<nirvdrum>
Okay.
<nirvdrum>
So was jcodings just wrong before?
<lopex>
nirvdrum: I see two routines for dummy creation and they both seem to set the base to ascii8bit
lanceball is now known as lance|afk
<lopex>
nirvdrum: it was wrong that it tried to load "something" that had a DUMMY field
<lopex>
and indeed ascii8bit field had that field before too
<lopex>
nirvdrum: oh and it couldnt have a chance to work since afaik that code was broken
<lopex>
that loadDummy thing
<lopex>
nirvdrum: but dont take my word too seriously though, I have a mess in my head from looking at too many MRI revisions
<lopex>
one missed diff entry, and youre done
<lopex>
nirvdrum: but is that related to the eager loading for Substrate still
<lopex>
?
<nirvdrum>
lopex: I think I need to just step in a debugger and look at the before and after.
<lopex>
nirvdrum: I'm still confused about the issues
<nirvdrum>
I never really understood what the old code was doing.
<lopex>
nirvdrum: it differs from MRI
<nirvdrum>
I think the problem was for an EncodingDB iterator, you could get both a real and a dummy encoding.
<lopex>
nirvdrum: I might made some bad assumptions, keep that in mind
<nirvdrum>
And if someone, from Ruby, specified a dummy encoding we needed to map that back to the real encoding somehow.
<lopex>
oh right
<lopex>
yes
<nirvdrum>
And we used the map built up from that iterator.
<lopex>
there's single table for entries
mkristian has quit [Quit: This computer has gone to sleep]
<lopex>
the entires can be either dummy or not
<nirvdrum>
So there's just one flat table now?
<lopex>
the entries are solely to postpone the loading
<lopex>
it alwasy was
<lopex>
same for transcodings
<lopex>
ah
<lopex>
I know where the confusion comes from
<lopex>
RubyEncoding
<nirvdrum>
Okay. I may have just been using the iterator wrong.
<lopex>
there's separate find*or*Alias things
<lopex>
nirvdrum: er, wrong there's separate lists
<lopex>
shit
<lopex>
it was so long ago :)
<nirvdrum>
Heh.
<nirvdrum>
I'll gist up what I was doing later.
<nirvdrum>
My source tree is a bit messy at the moment.
<lopex>
nirvdrum: doh, the evolution of this thing skewd it a bit it seems
<lopex>
nirvdrum: I was thinking from a pespective of a RubyEncoding api