<donV>
Hullo! I am starting work on trying to make loading the JRuby compiler optional, but I get a build error when bootstrapping my development environment.
<donV>
Exception in thread "main" java.lang.NoSuchFieldError: classLoader
<JRubyGithub>
jruby/master e0916d1 Benoit Daloze: [Truffle] Introduce a SlowSpecsTagger chasing bad ruby_exe()'s!
<JRubyGithub>
jruby/master 964ca3a Benoit Daloze: [Truffle] JT: run specs if the argument is not recognized.
JRubyGithub has left #jruby [#jruby]
mister_solo has quit [Remote host closed the connection]
mister_solo has joined #jruby
<donV>
Hi all!
colinsurprenant has quit [Quit: colinsurprenant]
elia has joined #jruby
<donV>
headius enebo : Hi! I’ll be working on only loading the compiler if it is configured. I am thinking of copying the way truffle is loaded. Sounds good?
tenderlove has quit [Remote host closed the connection]
djbkd has joined #jruby
djbkd has quit [Remote host closed the connection]
djbkd has joined #jruby
vikaton has quit []
vtunka has quit [Quit: Leaving]
colinsurprenant has joined #jruby
<chrisseaton>
donV: do you actually want to move it to a separate jar?
<donV>
No, just the loading part.
<donV>
Have an interface as part of core JRuby.
<donV>
but have the implementation in a separate package that can be omitted in the packaging.
<chrisseaton>
i mean moving it to a separate jar would be cleanest break, and would be really helpful to you, but it might be more than the core team want to do for now
<donV>
Yeah, you are right. Be we can ask :)
<donV>
My first step is separating the class loading.
headius2 has joined #jruby
chamila has joined #jruby
shellac has joined #jruby
anaeem1_ has quit [Remote host closed the connection]
elia_ has joined #jruby
<enebo>
yeah I don’t see any point to separate packaging so long as loading can be reflective
<enebo>
if it is in a package and the package can be omitted wihout breaking compilation then that is preferred
drbobbeaty has quit [Read error: Connection reset by peer]
benlovell has quit [Ping timeout: 245 seconds]
enebo has joined #jruby
<donV>
enebo, headius, chrisseaton: Do you run “bin/jruby -S rake test:mri” ? It gives me 237 errors and then hangs on “PStoreTest#test_thread_safe”
<enebo>
donV: I routinely run it yeah
<enebo>
donV: it is also main run in ci
<donV>
:|
<donV>
mvn -Pclean
<donV>
“NotImplementedError: wait unsupported or native support failed to load”
benlovell has joined #jruby
bruceadams has quit []
bruceadams has joined #jruby
lanceball is now known as lance|afk
KevinCorcoran has joined #jruby
anaeem1_ has joined #jruby
erikhatcher has quit [Quit: erikhatcher]
camlow325 has quit [Remote host closed the connection]
KevinCorcoran has quit [Ping timeout: 246 seconds]
<chrisseaton>
donV: have you got some unusual system setup? you seem to get lots of unusual failures
camlow325 has joined #jruby
<donV>
Not that I am aware of :)
camlow325 has quit [Remote host closed the connection]
baroquebobcat has joined #jruby
camlow325 has joined #jruby
elia_ has quit [Quit: Computer has gone to sleep.]
<donV>
enebo: The org.jruby.ir.Compiler uses Ruby runtime. Is it OK to put Ruby runtime in the constructor and store it as an instance variable?
<enebo>
donV: Let me look quick
camlow325 has quit [Remote host closed the connection]
<enebo>
donV: It can be but what is the issue? You want it removed from execute() signature?
<donV>
Yes, and other methods that are belong to the compiler like tryCompile()
<enebo>
donV: I am missing something though…why do you need it removed?
benlovell has quit [Ping timeout: 276 seconds]
oko has joined #jruby
camlow325 has joined #jruby
<donV>
enebo: Well, it seems the compiler is created in the context of a runtime, so it could refer to it.
<donV>
enebo: Is the idea to use a compiler instance across different runtimes?
KevinCorcoran has joined #jruby
<enebo>
donV: well not for this class no so I don’t care but I was wondering if there was a reason related to the isolation itself. It sounds like you just want to refactor this
<enebo>
donV: This Compiler class is already a bit weird with the static reference
<donV>
Yes. I am exploring the code and refactor to make a clean interface between the runtime and the compiler.
oko has left #jruby ["peace ! i'm out !!"]
<enebo>
donV: ok. yeah runtime is find and just make sure you solve this weird CompilerHolder+static reference aspect
<donV>
:) working on a branch for complete safety :)
<enebo>
donV: if it stays static then you have to pass in runtime since the compiler wlil be tied to an individual runtime
<donV>
Thinking compiler should be a straight instance variable with getter.
<enebo>
donV: but we should not be storing things like this as static fields
<donV>
moving to instance fields.
<enebo>
donV: this is from me merging the old IR persistence GSoC
<enebo>
I even added a FIXME :)
<donV>
:)
<donV>
Do I need to add synchronization on creating the compiler instance? I think I will try without, and refine on the review of the pull request.
<enebo>
donV: it depends on lifecycle of when you make it
<enebo>
donV: if it is made as part of runtime.init then you do not
marr has quit [Ping timeout: 272 seconds]
<enebo>
err Ruby.initXXX
<donV>
enebo: Sounds good.
skade has quit [Quit: Computer has gone to sleep.]
<donV>
enebo: Something completely different: some IntelliJ IDEA help: Where can I see all compilation errors?
<enebo>
donV: I am also wondering if we can maybe eventually avoid reflective load
<donV>
enebo: OK, how?
<enebo>
donV: like we could maybe have two .java files with same name which loads these full vs nothing things
<donV>
OK, done during packaging, then?
<enebo>
donV: ruboto could load a noop init() in a class and compile with that and normal builds could call an init() which loads this up
<enebo>
donV: yeah it would load a little faster since a reflective call is measurable
<enebo>
donV: but hey if you isolate this stuff reflectively we can also entertain that later
<enebo>
donV: it is just that ruboto already has package-time activities
<donV>
Yup.
<enebo>
donV: food for thought
<donV>
Yes. Now, where are all the compiler errors listed in IDEA…?
asarih has quit []
jeregrine has quit []
asarih has joined #jruby
jeregrine has joined #jruby
<enebo>
donV: oh I don’t know…I generally do a find usages what I am changing and change all usages…then compile outside idea
<donV>
OK
<enebo>
donV: I find intellij less reactive to error reporting than netbeans or eclipse
<enebo>
donV: but it may be me not knowing something
thsig has quit [Ping timeout: 245 seconds]
anaeem1_ has quit [Quit: Leaving...]
camlow325 has quit [Remote host closed the connection]
rafaelfranca has joined #jruby
bbrowning_away is now known as bbrowning
<donV>
enebo: Any idea why I would get “NotImplementedError: wait unsupported or native support failed to load” when running “bin/jruby -S rake test:mri” ?
KevinCorcoran has quit [Remote host closed the connection]
<enebo>
donV: perhaps try running -Xnative.verbose=true on regular command
<enebo>
donV: perhaps on your platform some recent changes and causing something to not load native support properly
<donV>
trying it now
<donV>
Failed to load native POSIX impl; falling back on Java impl. Stacktrace follows.
<enebo>
headius2: well it used/uses ant integration
<enebo>
headius2: although it is possible that is no longer valid
<headius2>
right, I just mean that perhaps our build/bootstrap should be pulling ant somewhere appropriate
deobalds has joined #jruby
<headius2>
rather than this hidden prereq
<enebo>
headius2: yeah maybe if we still need ant
<headius2>
we're using some tasks from our test targets
<headius2>
they could probably be rewritten to not use ant but that's yakky
<enebo>
headius2: ah yeah. I guess it would make sense to add it as a dev dep on maven
<headius2>
not on my critical path anyway
djbkd has quit [Remote host closed the connection]
<kares>
oh yeah, sqlite jdbc 3.8.10.1 has no issues despite its .jar size doubled ... time to push a gem // cc donV
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius opened issue #2965: Rakefile/rakelib depends on Ant, will not boot if not present http://git.io/vTzK3
JRubyGithub has left #jruby [#jruby]
<donV>
huh?
<kares>
donV: actually nevermind you're not using it :) .. just realized
<donV>
:)
<kares>
was afraid if it's fine on Android
<headius2>
enebo: testing out a fix for \r\n translation on Windows. Since we can't do O_TEXT currently I'm having the IO transcode logic always add UNIVERSAL_NEWLINE if we're in text mode
<donV>
We use the SQLite part with the sqldroid jdbc driver.
<headius2>
MRI dodges this because they can use O_TEXT to do the translation
<headius2>
it will mean all text reads on Windows have to run transcoding but I don't have another solution right now
tenderlove has joined #jruby
<kares>
donV: yy I know ... just a momentary outage on the transmitter :)
<headius2>
yay, rake -T works on Windows again
tenderlove has quit [Remote host closed the connection]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] donv created optional_compiler (+1 new commit): http://git.io/vTzio
<JRubyGithub>
jruby/optional_compiler 4c5c563 Uwe Kubosch: Only load the IR compiler if it will be needed
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
jruby/master 8a9f17c Charles Oliver Nutter: Use transcoder to do textmode crlf translation on Windows....
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/vTzih
JRubyGithub has left #jruby [#jruby]
baroquebobcat has joined #jruby
benlovell has quit [Ping timeout: 252 seconds]
subbu is now known as subbu|afk
djbkd has joined #jruby
<headius2>
and test:jruby runs...only 6 failures
<headius2>
that's a good sign
<nirvdrum>
headius2: Awesome.
<nirvdrum>
headius2: I mentioned this before, but you might be able to remove some platform guards for native calls now, too. I've backfilled a decent chunk of the missing POSIX calls for Windows.
<headius2>
that's cool...I might look at that in a few places
<headius2>
I'm trying to hit the big items on Windows because we're probably not going to get tests 100% for RC1
<headius2>
we haven't run tests on Windows in a loooong time
tcrawley is now known as tcrawley-away
n1ftyn8 has quit []
n1ftyn8 has joined #jruby
JRubyGithub has joined #jruby
JRubyGithub has left #jruby [#jruby]
<JRubyGithub>
[jruby] mkristian opened issue #2966: some File methods do not treat uri like paths consistently http://git.io/vTzFe
vikaton has joined #jruby
mitchellhenke has quit [Quit: Computer has gone to sleep.]
x1337807x has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
bbrowning_away is now known as bbrowning
pjammer has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
subbu|afk is now known as subbu
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] cshupp1 closed issue #2964: In windows ScriptingContainerTest fails http://git.io/vTzcZ
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] cshupp1 reopened issue #2964: In windows ScriptingContainerTest fails http://git.io/vTzcZ
JRubyGithub has left #jruby [#jruby]
elia has joined #jruby
pjammer has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
JRubyGithub has joined #jruby
<JRubyGithub>
jruby/master d7812fe Brandon Fish: [Truffle] Add pstore to stdlib.
JRubyGithub has left #jruby [#jruby]
<JRubyGithub>
[jruby] bjfish pushed 1 new commit to master: http://git.io/vTzjj
havenwood has quit []
shellac has joined #jruby
bbrowning is now known as bbrowning_away
pjammer has quit [Quit: My Mac has gone to sleep. ZZZzzz…]