skade has quit [Read error: Connection reset by peer]
skade has joined #jruby
benlovell has quit [Ping timeout: 272 seconds]
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
x1337807x has joined #jruby
mcclurmc has joined #jruby
calavera has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
havenn has quit [Remote host closed the connection]
yfeldblum has quit [Remote host closed the connection]
yfeldblum has joined #jruby
<headius>
nirvdrum_: possibly, yes...I need to see how far back that change was made
<headius>
if you wanna just do it, go ahead...dinner and tree decorating for me the rest of the night
pietr0 has quit [Quit: pietr0]
elux has joined #jruby
yfeldblum has quit [Ping timeout: 250 seconds]
josh-k has joined #jruby
guilleiguaran has quit [Ping timeout: 272 seconds]
elux has quit [Quit: Leaving...]
cpuguy83 has quit [Ping timeout: 272 seconds]
guilleiguaran has joined #jruby
elux has joined #jruby
cpuguy83 has joined #jruby
elux has quit [Client Quit]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed issue #2231: Was the --indy cli option removed? (1.7.16.1) http://git.io/vvj8sQ
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to jruby-1_7: http://git.io/RzDJgg
<JRubyGithub>
jruby/jruby-1_7 a30c537 Charles Oliver Nutter: Remove never-implemented flag --indy. Fixes #2231.
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/KIHNmA
<JRubyGithub>
jruby/master da5b603 Charles Oliver Nutter: Remove never-implemented flag --indy. Fixes #2231.
JRubyGithub has left #jruby [#jruby]
skade has quit [Quit: Computer has gone to sleep.]
mcclurmc has quit [Remote host closed the connection]
bbrowning_away is now known as bbrowning
baroquebobcat has joined #jruby
yfeldblum has joined #jruby
<headius>
chrisseaton, eregon, nirvdrum_: btw, hopefully if you guys fix something in the truffle side that's also broken on the plain JVM side, you'll either fix both or file an issue informing us of the change needed
<headius>
it would be silly to go through these and re-fix them a few weeks from now or something
<headius>
I'd try to do the same but I guess I don't really touch anything that is in truffle-side yet
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
triple_b has joined #jruby
bbrowning_away is now known as bbrowning
josh-k has joined #jruby
josh-k_ has quit [Read error: Connection reset by peer]
shellac has quit [Quit: Computer has gone to sleep.]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius opened issue #2261: Thread poll may raise exceptions, but IR doesn't like it http://git.io/S8B3BQ
JRubyGithub has left #jruby [#jruby]
<headius>
eregon: submodule workflow never really made sense to me...we kept getting weird effects like local repo looking dirty when submodule upstream changed, stuff like that
<headius>
I'd be game to make it a submodule if that workflow did not have weird effects or make a big hassle for new devs
benlovell has joined #jruby
calavera has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed issue #2208: JIT fix needed for keyword args used in super http://git.io/kQIfcA
<JRubyGithub>
[jruby] chrisseaton pushed 1 new commit to truffle-head: http://git.io/8EozwA
<JRubyGithub>
jruby/truffle-head 7a21fc4 Chris Seaton: Merge branch 'master' into truffle-head
JRubyGithub has left #jruby [#jruby]
anaeem1_ has quit [Remote host closed the connection]
mcclurmc has quit [Ping timeout: 255 seconds]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius opened issue #2262: Explore raw park/unpark for fiber handoff http://git.io/5gbqpA
JRubyGithub has left #jruby [#jruby]
<headius>
chrisseaton: what's the status of fibers in jruby+truffle?
<headius>
I'm just curious since I was mucking about with that code yesterday in jruby proper
<chrisseaton>
headius: they're very basic - using a full thread for each one
<headius>
ahh, ok
<headius>
I did hear something about truffle or graal-based continuations, didn't I?
<chrisseaton>
headius: we have a coroutine patch for the JVM thanks to Lukas, but we're not using it for fibers at the moment
<headius>
yeah, the old coro patch is out there but needs to be refreshed periodically
<chrisseaton>
headius: oh - there is also generators in Python, which Zippy implements on pure Truffle
<headius>
I suppose coro support requires a lot more than just switching to graal
<chrisseaton>
headius: it compiles multiple entry points into methods, by having copies of the AST that start from various nodes - it's the same mechanism we use for goto in C
<headius>
python generators are lexical...so you can even do it with bytecode weaving
<headius>
jython does that I believe
<chrisseaton>
but basically there's no Truffle magic to solve fibres at the moment
<headius>
boo
<headius>
I guess it will be worth exploring fiber hand-off efficiency
<headius>
concurrent live fiber limits will still be a problem though :-(
<chrisseaton>
eregon mentioned the other day an idea for ObjectSpace in normal JRuby - why don't you add objects to your map of id->object only when you actually request an id for the first time? couldn't that always be enabled? if you haven't requested an ID, you wouldn't have an id to look the object up with.
<headius>
we do that
<headius>
or rather, we do that when id2ref is enabled
<chrisseaton>
but that doesn't have an overhead does it? so why do you need +O?
<headius>
the problem is that object_id is called for many other reasons that have nothing to do with id2ref
<chrisseaton>
ah ok
<headius>
like for every ActiveRecord model
<chrisseaton>
ahha
<headius>
we also got an official ruling from matz that id2ref was always intended to be an impl-specific API
shellac has joined #jruby
fidothe_ has quit [Ping timeout: 258 seconds]
fidothe_ has joined #jruby
ahadding1 has quit [Ping timeout: 258 seconds]
lance|afk is now known as lanceball
yfeldblum has joined #jruby
mkristian has joined #jruby
havenwood has joined #jruby
ahadding1 has joined #jruby
<nirvdrum_>
Yes. Rails loves to call object_id.
yfeldblum has quit [Ping timeout: 264 seconds]
<nirvdrum_>
This was a big performance issue in JRuby for a little while due to a broad lock.
nirvdrum_ is now known as nirvdrum
<headius>
rails was definitely the prime motivator to make id2ref off by default
<nirvdrum>
So, at least two other critical pieces of Rails that aren't ActiveRecord.
sferik has joined #jruby
johnmuhl has joined #jruby
sferik has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
sferik has joined #jruby
sferik has quit [Client Quit]
mcclurmc has joined #jruby
sferik has joined #jruby
anaeem1_ has joined #jruby
bbrowning has quit [Quit: Leaving]
postmodern has quit [Read error: Connection reset by peer]
bbrowning has joined #jruby
noop has quit [Ping timeout: 250 seconds]
calavera has joined #jruby
brettporter has joined #jruby
brettporter has quit [Ping timeout: 244 seconds]
tenderlove has joined #jruby
yfeldblum has joined #jruby
rsim has quit [Quit: Leaving.]
mitchellhenke has joined #jruby
bbrowning is now known as bbrowning_away
<headius>
tenderlove: hey...we have no MRI representation for the FOSDEM Ruby room...I pinged matz and ko1 but I never know if I'll get a reply from them
<tenderlove>
headius: hey
<tenderlove>
what are the dates?
yfeldblum has quit [Ping timeout: 240 seconds]
<chrisseaton>
Weekend of 31st of Januar
<headius>
tenderlove: yeah, two days in Brussels with a bunch of other fine Rubyists
fivebats_ has joined #jruby
fivebats has quit [Read error: Connection reset by peer]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] eregon pushed 1 new commit to master: http://git.io/tKCegw
<JRubyGithub>
jruby/master bcfa723 Benoit Daloze: [Truffle] Add unary and binary core method nodes for expressivity....
JRubyGithub has left #jruby [#jruby]
<tenderlove>
headius, chrisseaton: let me get back to you later today or tomorrow. I have RubyFuza in South Africa the following week, so I need to figure out the flights for that first.
Aethenelle has joined #jruby
mister_solo has quit [Ping timeout: 272 seconds]
sferik has quit [Ping timeout: 250 seconds]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] eregon pushed 1 new commit to master: http://git.io/eDNyVQ
<JRubyGithub>
jruby/master 7e82292 Benoit Daloze: [Truffle] Add missed files to last commit.
JRubyGithub has left #jruby [#jruby]
<headius>
tenderlove: I might be doing that as well
<headius>
it would be three conferences in three countries in just over a week, though :-\
<tenderlove>
ya....
benlovell has joined #jruby
<headius>
oh neat, rubyfuza CFP is still open
<chrisseaton>
headius: are you going to that Ruby Caribe conference? That looks fun
<rtyler>
towards the bottom of the gist is the same test run with scala 2.10
<headius>
I want something in the middle
<rtyler>
something is really fucked going from JRuby into Scala pre-2.10
<headius>
well, and I like mirah too
<headius>
rtyler: singleton object I guess?
<rtyler>
there are three different kinds of failures there
<headius>
woah, npe
<rtyler>
one: failed to coerce java.util.Properties to kafka.utils.VerifiableProperties
<rtyler>
the NPE
<rtyler>
then the Log4j thing
x1337807x has joined #jruby
erikhatcher has joined #jruby
<headius>
the properties issues are probably an API change that they plaster over with implicit conversion
<headius>
you might have to actually pass an instance of whatever that is rather than a Properties
<headius>
that's my guess on those
<rtyler>
don't happen to know anybody else tying scala and jruby together?
<headius>
implicit conversions are like the refinements of scala :-( magically turn one object into another when you pass it or call a method on it that isn't there
<headius>
rtyler: dpp originally did patches for us
<headius>
also mario...something
<rtyler>
mkristian: I might need to enlist your help :-!
<headius>
mario camou, mcamou
<lopex>
might also be the implicit class and it's implicit method being generated by different convention
<headius>
very likely you'd need to have different client code for each, if they've changed what we see through reflection across scala revs
<rtyler>
what do Scala traits turn into
<headius>
you don't want to know
<rtyler>
well, that Logging class is a trait
<rtyler>
dafuq
<headius>
I believe they're duplicated into the target hierarchy as an actual ancestor
<headius>
I'm not sure how they handle sharing the code and recompiling and such
<rtyler>
this makes me want to smash my keyboard to bits
<headius>
we have never done anything special to support traits
<headius>
so I wouldn't be surprised if there's gaps
<headius>
rtyler: it may be simpler for you to write a small java wrapper utility around the troublesome methods, assuming they can be called from Java
<headius>
or if it has a Java API, use that
<headius>
scala is even more magic than Ruby in many ways :-(
<rtyler>
this is their java API :/
<lopex>
they need recompilation
<lopex>
but the traits are quite simple, just an interface and a static methods for impl
<headius>
lopex: no super?
<headius>
or fields
<lopex>
any trait usage rewrites the impls call
<lopex>
*calls
<lopex>
headius: fields are in metadata afaik
<lopex>
headius: they're being recreated for end user
<rtyler>
good grief that's terrifying
<headius>
I can imagine how that would be implemented to share a single trait method body across all uses, but it makes my head hurt
<lopex>
headius: but given conventions lots can be done via scala integration
<headius>
super would have to force a bridge method or something
<lopex>
headius: super is all compile time, given trait implementation order
<headius>
ah, which is why it has to be recompiled if trait or consumers change
<headius>
oh, not consumer I guess
<lopex>
they call it linearization
<headius>
just callers
<headius>
hmm
<headius>
they do insert a synthetic class into the target hierarchy, right?
<headius>
I fixed a bug yesterday that was responsible for many flaky tests
<johnmuhl>
headius: the offending path is "dir/dir with ! mark/file with % sign.ext". If I remove the % from the file name it works or if I remove the ! from the directory name it also works.
drbobbeaty has quit [Read error: Connection reset by peer]
benlovell has quit [Ping timeout: 272 seconds]
drbobbeaty_ has joined #jruby
<asarih>
headius: high five
<headius>
asarih: indeed! *five*
<headius>
I hope you guys are testing stuff on jruby-head and will let us know if there's problems
<dfr|work>
headius, hmm. I think I'm fine with your proposed changes, since if it's not a URL I'd say all bets are off whether we can treat it as stuff inside of jar anyway.
<headius>
johnmuhl: if that fix works for your case I'll push it...and rely on you to add a test to rubyspec or some other suite
<headius>
bleh, spec run taking too long again
<headius>
death of a thousand subprocesses
<dfr|work>
headius, so +1 on your change (although I'd word the commend stating that if path has illegal UTF-8 symbols, it's likely not URI and we require jar paths to be URI paths)
<headius>
dfr|work: can do
<headius>
I will mention the bug as well
<dfr|work>
headius, and you should probably be able to reproduce it easily by trying to lookup "foo.jar!/bar % zeta" :)
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to jruby-1_7: http://git.io/PbCokA
<JRubyGithub>
jruby/jruby-1_7 f0aaf44 Charles Oliver Nutter: If presumed URI does not decode, it's not a URI. Fixes #2264.
<Aethenelle>
i hit that a few times in my prepend work...
<johnmuhl>
headius: Your fix works. I'm confused about what should be added to rubyspec though…
<asarih>
headius: I'll look into improving it when I have more time.
e_dub has joined #jruby
rimenes has joined #jruby
<headius>
johnmuhl: on second thought this may be too jruby-specific for rubyspec
<headius>
it's a special case that happened to fall into functionality we have that no other impls have
<headius>
dunno...might ask brixen for his thoughts on whether it deserves a rubyspec
<headius>
in the interim you could add a spec to our little regression suite in spec/regression
<headius>
basically just what you provided for repro in the bug
yfeldblum has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 3 new commits to jruby-1_7: http://git.io/RIJH1A
<JRubyGithub>
jruby/jruby-1_7 179811e Charles Oliver Nutter: Add a spec:ruby:fast target that avoids subprocess-spawning specs....
<JRubyGithub>
jruby/jruby-1_7 74f8381 Benoit Daloze: Update list of slow language specs.
<JRubyGithub>
jruby/jruby-1_7 0167f3f Charles Oliver Nutter: Add fast rubyspec for 1.8 mode and make travis use them.
JRubyGithub has left #jruby [#jruby]
<johnmuhl>
headius: I'll open an issue on rubyspec referencing the jruby one and in the meantime look at adding a regression test for you. I'm a bit novice so might have a few more questions.
<headius>
ok
tcrawley is now known as tcrawley-away
<headius>
no problem
<headius>
the regression specs are just rspec, have a look at what's there
<johnmuhl>
will do, not much else happening this afternoon :)
<headius>
asarih: I was going to file issues for each jruby-head build people report failing
<headius>
and then once we figure out what all the failures are that issue just becomes a parent issue
<headius>
checklist etc
<asarih>
headius: sounds reasonable.
pietr0 has quit [Quit: pietr0]
pietr0 has joined #jruby
anaeem1_ has quit [Remote host closed the connection]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius opened issue #2265: Improper access of Psych::Syck in travis-tasks http://git.io/hSXlcA
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius opened issue #2266: OutOfMemoryError running travis-core specs on jruby-head http://git.io/oSNsFA
<JRubyGithub>
[jruby] headius closed pull request #1668: Open3#popen3 should use leading Hash as additional environment variables (jruby-1_7...popen3_should_handle_env_hash) http://git.io/49LW0g
JRubyGithub has left #jruby [#jruby]
brettporter has joined #jruby
brettporter has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to jruby-1_7: http://git.io/gNUlww
<JRubyGithub>
jruby/jruby-1_7 66b2905 Charles Oliver Nutter: Peel off options hash passed to Open3.popen3....
triple_b has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<headius>
tenderlove: which mri make target would run test/test_open3.rb?
<tenderlove>
make test-all
<tenderlove>
AFAIK
<headius>
ok
<headius>
it doesn't want to run nicely for me outside the suite
<tenderlove>
make test-all TESTS=test/test_open3.rb
Hobogrammer_ has joined #jruby
pietr0 has quit [Ping timeout: 250 seconds]
<tenderlove>
headius: ^^^ that will run it for me
<headius>
tenderlove: awesome, thanks
pietr0 has joined #jruby
<headius>
we just use runner directly in jruby so I'm not familiar with the MRI make targets
Hobogrammer has quit [Ping timeout: 258 seconds]
<nirvdrum>
I don't suppose anyone else is seeing "[ERROR] Failed to execute goal on project jruby-noasm: Could not resolve dependencies for project org.jruby:jruby-noasm:jar:9.0.0.0-SNAPSHOT: Could not find artifact org.jruby:jruby-core:jar:noasm:9.0.0.0-SNAPSHOT in sonatype (https://oss.sonatype.org/content/repositories/snapshots/)" ?
<headius>
hmm
<headius>
I haven't done a master build in a little while, but I wasn't seeing it then
<nirvdrum>
It looks like someone already updated the old definition to use the new URLs.
mitchellhenke has quit [Quit: Computer has gone to sleep.]
<nirvdrum>
You're probably safe to remove the old ones. The only people it would impact are those that haven't updated ruby-build and also haven't already installed 9k.
<nirvdrum>
And also want to try out 9k.
<nirvdrum>
Not to be crass, but that's probably a really, really small set of users.
<headius>
ugh...1.7 spec failures are due to my instance_eval change
lance|afk is now known as lanceball
<headius>
nirvdrum: yeah, I'm a paranoid
<headius>
I updated those definitions, I had just forgotten rbenv uses ruby-build
cprice is now known as cprice404
<headius>
or rather rbenv users use ruby-build
multibot_ has quit [Ping timeout: 264 seconds]
<nirvdrum>
Strictly speaking, it doesn't need to. But virtually everyone uses them as a pair.
elia has joined #jruby
<nirvdrum>
I generally don't like breaking things on people. To the point that most of my own gems still support 1.8 and Rails 2.3. But, you're talking about running an old dev release. It seems suspect.
<nirvdrum>
Oh, and that old dev release was never guaranteed to be any particular revision.
<headius>
well fudge
<headius>
this fix breaks stupid backref and lastline
<nirvdrum>
My guess is anyone really interested in 9k, will just install 9.0.0.0.
<headius>
I shake my fist at you backref and lastline
<nirvdrum>
Heh.
viking has quit [Remote host closed the connection]
<headius>
sigh
<headius>
I ignored those failures because I thought I had a bad local env
DomKM has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 2 new commits to jruby-1_7: http://git.io/yylV9g
<JRubyGithub>
jruby/jruby-1_7 cdba7c8 Charles Oliver Nutter: Revert "Fully clone block + frame for instance_eval forms."...
<JRubyGithub>
jruby/jruby-1_7 15f651b Charles Oliver Nutter: Less broken but more grosser fix for #1290 and #1547....