<eregon>
the arguments that will be passed to the Layout constructor
<eregon>
so in your case, llocateObjectNode.allocate(rubyClass, someDefaultName);
<alex0ptr>
Which would be an DynamicObject again?
<eregon>
yes
<alex0ptr>
Alright
<eregon>
so you could give it nil(), since you know it should be properly initialized later, and nil() is also a DynamicObject
<alex0ptr>
Perfect. The DispatchHeadNode Example helps a lot thanks
<alex0ptr>
Maybe I should look at a few examples in IGV to see how stuff is build...
<eregon>
In general, it's nicer to call Java from Ruby than Ruby from Java in JRuby+Truffle, or at least it's less boilerplate. But it can be done if needed.
<eregon>
I would recommend using a simple Debugger to see the AST structure
<eregon>
the @CoreMethod stuff is a "magic" DSL to help extracting receiver and arguments
<alex0ptr>
Yeah I guess Ruby from Java is my current way to go.
<chrisseaton>
Java code should be minimal, single-purpose 'system calls' in most cases
<chrisseaton>
Deal with all the complexity in Ruby, then just call into Java for the final work if you need to
<eregon>
Ruby to Java can be done for instance with Truffle.invoke_primitive(:name, args...), which just calls a Node defined in Java, annotated with @Primitive.
<chrisseaton>
Ruby arrays are hard enough to handle in Java, hashes are nearly impossible to handle
<alex0ptr>
Well.. Maybe I should explain what my ultimate goal is: I'm tryin to evaluate Ruby + Truffle for functional and reactive programming. I'm especially interested in optimizing PIC. We assume this is where we would get the biggest benefits here.
<eregon>
DispatchHeadNode always includes a PIC mechanism, so maybe this already works?
<alex0ptr>
Don't ask me, I'm just starting to understand what is there already. Time will tell, I guess.
<eregon>
But if it's a PIC at the level of the graph created by connecting reactive streams, then it's different and depends how an event triggers the next one.
<alex0ptr>
Yeah I fear there has to be done something on the Truffle Level.
<eregon>
I think that project tried to guess the structure of the reactive graph, and optimize based on that, is that right?
<alex0ptr>
What do you mean by 'guess'?
<chrisseaton>
Profile and speculate, rather than just taking while guesses :)
<chrisseaton>
wild guesses, I mean
<eregon>
I think it's somehow detecting the structure and assuming it's not going to change, or reconsider when it changes?
<alex0ptr>
Yeah. At least that is what I understood.
<alex0ptr>
However it didn't work out as expected. So I'm like number two to try this.
<eregon>
Is there some document that got out of this except the code?
<alex0ptr>
Sure. The thesis is linked on the page.
<alex0ptr>
I guess one problem was the quite complex propagation logic that he chose. I will try to start with something much simpler and see where this takes me...
<eregon>
Oh right, somehow I missed the link
<eregon>
A lot of the code I looked at would be easier and faster to write in Ruby. So I think prototyping in Ruby might gain a lot of time.
<alex0ptr>
This is how I started
<alex0ptr>
I wrote something in Ruby first
<alex0ptr>
Wrote a few benchmarks
<alex0ptr>
And now I want to understand what Ruby + Truffle does with that.
<eregon>
Sonds like a good approach :)
<alex0ptr>
Yeah... Still feel kind of lost.
<eregon>
IGV is interesting, but it's a lot of details and needs some getting used to. It's usually best if you have a precise idea what to look for, and simple code to compare.
<eregon>
In any case, don't hesitate to ping me, chrisseaton or nirvdrum.
<eregon>
I have to go now though.
<alex0ptr>
Well I neither know what to look for (that's why I'm looking into how truffle works) neither will my code be simple(?).
<alex0ptr>
Thank you, eregon
<eregon>
if you can share some benchmark or code, it can help quite a bit
<alex0ptr>
I'll definitely come back to that.
<eregon>
I mean small and not doing much work by "simple"
<alex0ptr>
Okay I need to go sleep now. Good night. And thanks to everyone who is reading this.
<GitHub172>
jruby/master 6d09390 Charles Oliver Nutter: Add spec for Enumerable#chunk_while with single element....
<GitHub172>
jruby/master b003233 Charles Oliver Nutter: Fix chunk_while with single-element array. Fixes #4410
<GitHub57>
[jruby] nirvdrum pushed 4 new commits to truffle-head: https://git.io/vMZ6F
<GitHub57>
jruby/truffle-head c2a908e Kevin Menard: Improved command processing for exec/spawn....
<GitHub57>
jruby/truffle-head 261c709 Kevin Menard: [Truffle] Moved exec command processing out of the exec command, since it's useful to other process spawning operations.
<GitHub57>
jruby/truffle-head b0abffd Kevin Menard: [truffle] Fixed a typo in a class name.
jensnockert has quit [Remote host closed the connection]
<GitHub29>
[jruby] chrisseaton pushed 1 new commit to truffle-head: https://git.io/vMZXJ
<GitHub29>
jruby/truffle-head 340eca1 Chris Seaton: [Truffle] @Child fields should normally be private, and in that order.
<brocktimus>
headius: Just figured I'd let you know about https://github.com/rails/rails/pull/27363 since I've seen you mention about issues with big projects using refinements because of JRuby optimisations.
bga57 has quit [Ping timeout: 245 seconds]
jensnockert has joined #jruby
prasun has joined #jruby
marciol has joined #jruby
marciol has quit [Ping timeout: 272 seconds]
jensnockert has quit [Remote host closed the connection]
jensnockert has joined #jruby
jensnockert has quit [Ping timeout: 240 seconds]
pawnbox has joined #jruby
prasun has quit [Ping timeout: 255 seconds]
rsim1 has quit [Read error: Connection reset by peer]
rsim has joined #jruby
rsim1 has joined #jruby
rsim has quit [Read error: Connection reset by peer]
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
pawnbox has quit [Read error: No route to host]
pawnbox has joined #jruby
<GitHub1>
[jruby] bjfish force-pushed truffle-string-scrub from 9803c8d to 09b8c50: https://git.io/vMJ6n
<eonwe>
TO me the whole stack seems very odd as it's supposedly a tomcat thread. But still, I have had similar stack four times: thrice with org.jruby.RubyBasicObject$INVOKER$i$send19.call and once with org.jruby.internal.runtime.methods.JittedMethod.call
alex0ptr has joined #jruby
prasun has joined #jruby
bbrowning is now known as bbrowning_away
tcrawley-away is now known as tcrawley
<kares>
eonwe: quite ugly - well they make sense to someone. from a JRuby perspective it probably points to a JIT bug (not sure)
<kares>
what JRuby are you on?
<kares>
I am actually trying to isolate a JIT issue that ends up in a JVM segfault
<kares>
... on 9.1
<kares>
but I do have much more details available in the JVM (core) dump
pawnbox has quit [Remote host closed the connection]
jensnockert has joined #jruby
<eonwe>
kares: that's on 1.7.23.
<eonwe>
Do you mea JRuby JIT issue or JVM JIT issue?
<kares>
eonwe: not sure - I have 2 seemingly different cases
<kares>
but guessing the JIT produced by JRuby causes issues
<kares>
in one case I clearly see the JVM is trying to JIT a piece
<kares>
eonwe: your case seems weird considering how stable 1.7 seems to be
<kares>
would upgrade to latest 1.7.x to see if it happens again and than report (try to reproduce)
<eonwe>
The problem is that these happen very intermittenly so reproduction hasn't had much luck
<eonwe>
But the printed stack does look very odd to me: native frame, jruby (java) frame and then native frame. And that's it for a supposedly Tomcat thread so something's off there. Maybe the stack itself has been corrupted so whatever stack walking the crash reporter does, doesn't work
pawnbox has joined #jruby
<kares>
eonwe: it probably is JIT-ed code
<kares>
I am on Java 8 112 (Linux) and the generated reports include details
<kares>
but I've seen short reports such as yours in the past, really not sure what it is
<kares>
maybe look around to see if there's anything TC (JVM) specific (such as setting some command line properties) that can be done to help gather relevant info on core dumps
alex0ptr_ has joined #jruby
alex0pt__ has joined #jruby
alex0ptr_ has quit [Read error: Connection reset by peer]
alex0ptr_ has joined #jruby
alex0pt__ has quit [Read error: Connection reset by peer]
alex0ptr has quit [Ping timeout: 246 seconds]
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
lance|afk is now known as lanceball
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
marciol has quit [Remote host closed the connection]
enebo has joined #jruby
TheWhip has joined #jruby
marciol has joined #jruby
pawnbox has quit [Remote host closed the connection]
jensnockert has quit []
pawnbox has joined #jruby
pawnbox has quit [Ping timeout: 248 seconds]
vtunka has joined #jruby
bbrowning_away is now known as bbrowning
<GitHub119>
[jruby] chrisseaton pushed 2 new commits to truffle-head: https://git.io/vMnal
<GitHub119>
jruby/truffle-head 31b9253 Chris Seaton: [Truffle] Upgrade to Truffle 0.21.
<GitHub119>
jruby/truffle-head c91141a Chris Seaton: [Truffle] Upgrade to GraalVM 0.19.
prasun has quit [Ping timeout: 255 seconds]
pawnbox has joined #jruby
<GitHub121>
[jruby] chrisseaton pushed 1 new commit to truffle-head: https://git.io/vMnVi
<GitHub121>
jruby/truffle-head bea5212 Chris Seaton: [Truffle] Missing fixnum lower.
pawnbox has quit [Ping timeout: 248 seconds]
<GitHub89>
[jruby] chrisseaton pushed 1 new commit to master: https://git.io/vMnwq
<GitHub89>
jruby/master db207c1 Chris Seaton: Merge branch 'truffle-head'
<alex0ptr_>
However when I try to instantiate it I get:
<alex0ptr_>
RubyTruffleError: internal implementation error - AssertionError 0 arguments given but the factory takes 1: name com.oracle.truffle.object.ShapeImpl$DynamicObjectFactoryImpl.newInstance(ShapeImpl.java:959)
pilne has joined #jruby
thedarkone2 has joined #jruby
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
<headius>
hmm
<headius>
finally got a linux IRC client to hook up to my znc
<nirvdrum>
What'd you settle on?
<nirvdrum>
I like HexChat, but moved to a locally hosted version of The Lounge so I can still connect to IRC while on VPN.
<headius>
at the moment WeeChat
<headius>
we'll see if I can tolerate text-mode-only IRC
<headius>
/help
<headius>
hmm
<alex0ptr_>
nirvdrum: Could you take a look at my exception?
<headius>
gonna have to read some doco
knu has joined #jruby
subbu is now known as subbu|lunch
<nirvdrum>
alex0ptr_: Looking now.
<nirvdrum>
alex0ptr_: Generally, having the full trace is easier though.
<GitHub131>
[jruby] sumitmah opened pull request #4422: [4403] To be consistent with MRI, File.exists? for <path to file> end… (master...4403-fix-exists-method) https://git.io/vMcfd
<enebo>
headius: enjoy this message
<headius>
also beep
<headius>
that will do for now
<headius>
the thing I love about linux is everything has to be stitched together from half-true blog posts and seven-years-old mailing list threads
<nirvdrum>
I envy the luck you guys have had with Mac :-/ My wife's MBP has been a nightmare.
rsim has quit [Read error: Connection reset by peer]
<headius>
in what way?
<headius>
mac has definitely been the least hassle of any machine in my lifetime
<nirvdrum>
Wireless craps out all the time. Time Machine backups corrupt every few weeks. It comes and goes out of sleep as it pleases.
<headius>
mostly due to OS X being way more consistent internally than any other OS I'v eused
<alex0ptr_>
headius: totally agree... Some development tools even feel much better on mac than on linux (e.g. docker)
<headius>
yuck, that blows
<nirvdrum>
And searching for help is just a bunch of smug forum posts about how it must be end user fault because that person has never experienced it.
<nirvdrum>
I want to wipe the machine, but I can't get Time Machine to maintain backups :-/
<headius>
I never used TM because I didn't wany to buy a TimeCapsule and didn't feel like trusting my backups to jury-rigged clones
<nirvdrum>
Do you use something else?
<headius>
wifi I'd suspect an issue with your router
<headius>
I don't back anything up
<nirvdrum>
Gutsy :-)
<headius>
I don't have anything on local systems that isn't already in the cloud somewhere
<headius>
git repos, music, pictures, movies
<alex0ptr_>
I have coworkers that use 'carbon copy cloner' for backups
<nirvdrum>
The MBPs I had always had wireless issues, too. Maybe I just throw off some EM interference. I've gone through 6+ APs in that time, including official Apple ones.
<headius>
setting up a new device for me is only installing apps and updating settings
<nirvdrum>
I'm now using the Unifi ones.
<headius>
I wonder if your area is particularly noisy
<headius>
you might scan for other channels in use and try to force a better one
<nirvdrum>
Could be. We do have high voltage lines running not too far from here.
<headius>
there's a few nice tools that will show a map of channels and frequencies and how much chatter there is
<enebo>
I have a troubled corner where I sometimes sit and I can only get 802.11b there and only if I already have a wifi connection already. The lenovo and mbp both are identical at that spot
<headius>
enebo and I were both on 5-year-old macs too so there could be issues with newer hardware
<nirvdrum>
I've just set up a second Unifi AP, which should be way overkill for my house, but I'm hoping it helps.
<enebo>
the wifi antenna is too thin…a courageous choice
<enebo>
that joke never gets old
<nirvdrum>
Anyway, good luck with the Linux transition. I was ultimately pleased with it, but it did take some getting used to.
<headius>
I'm not sure those machines had ac...maybe not even n
<enebo>
nirvdrum: mostly for me it is just the communication apps
<nirvdrum>
alex0ptr_: Thanks. I'll check that out.
<headius>
yeah, if I could get trackpad working like I want and shortcut keys weren't different in EVERY app I'd be happy
<enebo>
and the trackpad but I am ok now with getting used to it
<enebo>
at least on fedora core
marciol has quit [Remote host closed the connection]
<alex0ptr_>
enebo: What have where you using before?
<nirvdrum>
alex0ptr_: I'm not seeing anything immediate. GreeterLayout shouldn't use RubyString, but that's not affecting anything.
<nirvdrum>
alex0ptr_: Is this code public by chance?
subbu|lunch is now known as subbu
rsim has joined #jruby
<alex0ptr_>
nirvdrum: No not yet. It's my first Ruby Truffle Code.
<enebo>
alex0ptr_: mbp ’11 and mbp back to 2006 with Sony Vaio (linux) before that for several years
<enebo>
so I was a linux user on a laptop first then macs now back to linux
Antiarc_ has quit [*.net *.split]
<enebo>
basic stuff like wifi is dirt simple now
<alex0ptr_>
enebo: that's an interesting journey...
<enebo>
basic installs works as expected
Antiarc has joined #jruby
<enebo>
alex0ptr_: I was a unix sysadmin several decades ago so I have a proclivity to using linux user space tools
<enebo>
heh I am now mixing linux with unix
<nirvdrum>
alex0ptr_: Can you paste the snippet for your entire registration please?
<GitHub138>
jruby/truffle-head 7442e5a Benoit Daloze: [Truffle] Write directly to $? in core.
yfeldblum has joined #jruby
drbobbeaty has quit [Ping timeout: 256 seconds]
marciol has joined #jruby
rsim has joined #jruby
rsim1 has quit [Read error: Connection reset by peer]
drbobbeaty has joined #jruby
Puffball has quit [Remote host closed the connection]
rsim1 has joined #jruby
rsim has quit [Read error: Connection reset by peer]
Puffball has joined #jruby
<GitHub3>
[jruby] headius opened pull request #4423: Kill the LocalJumpError logic for break in lambda. Fixes #4369. (master...lambda-break) https://git.io/vMcsE
<nirvdrum>
alex0ptr_: I took your gists and added them locally. I maybe missed something, but with the proper CoreLibrary set up, everything works fine.
<nirvdrum>
If you do pull from truffle-head, not that chrisseaton has been doing some work to avoid the need for explicit constructors, so you'll probably have to update your code a bit.
<alex0ptr_>
nirvdrum: That's weird. I just updated to current JDK/Graal/truffle-head/jvmci
lance|afk is now known as lanceball
<alex0ptr_>
nirvdrum: Maybe my test is bad?
<alex0ptr_>
But all it says is: include Frrb; Greeter.new('test')
<nirvdrum>
Well, your registration is definitely wrong.
<alex0ptr_>
nirvdrum: oh
<nirvdrum>
Define the module, then define the class in the module.
<nirvdrum>
And you're getting the same exact exception?
<alex0ptr_>
Still: assertionError 0 arguments given but the factory takes 1: name com.oracle.truffle.object.ShapeImpl$DynamicObjectFactoryImpl.newInstance(ShapeImpl.java:979)
<nirvdrum>
Have you registered your class with the CoreMethodNodeManager?
<alex0ptr_>
nirvdrum: Uhm no.
<nirvdrum>
See CoreLibrary#addCoreMethods
<nirvdrum>
Without that, all your @CoreMethods are just dead code.
rsim1 has quit [Read error: Connection reset by peer]
rsim has joined #jruby
<alex0ptr_>
I see. I thought this would be discovered.
<nirvdrum>
This isn't a Truffle annotation. It's a JRuby+Truffle thing.
<alex0ptr_>
Oh ok.
<alex0ptr_>
Great. Now I get a spec error for being stupid.
<alex0ptr_>
Thank you nirvdrum. Hope I can make some progress before stumbling again.
<nirvdrum>
No problem.
<nirvdrum>
So it's working now?
<alex0ptr_>
nirvdrum: Not quite but that is mspec complaining.
bbrowning is now known as bbrowning_away
TheWhip has joined #jruby
bbrowning_away is now known as bbrowning
mtoy has joined #jruby
<mtoy>
the sentry-raven gem doesn't parse 9000 errors correctly so i am making a patch ... one question though, looking through unparsed stack frames, i see traces lines of the file "at RUBY.block in method_name(uri:classloader:file_name.class:100)"
<mtoy>
... but i can't generate on of those on purpose. any clue what makes the output sometimes look like that?
TheWhip has quit []
<eregon>
alex0ptr_: The AssertionError sounds like there was a missing argument to allocate()
<alex0ptr_>
eregon: I got it fixed in the meantime
<eregon>
which was likely caused by the proper factory not being chosen, as the methods were not registered
<alex0ptr_>
eregon: The class and the factory were not registered correctly
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
rsim has quit [Quit: Leaving.]
enebo is now known as enebom
marciol has quit [Remote host closed the connection]