<GitHub50>
jruby/master 078e459 Benoit Daloze: Do not leak a FD with File.open without block in lib/pom.rb...
<GitHub2>
[jruby] eregon commented on commit e498859: Why is `context` last? https://git.io/v2cjJ
<GitHub30>
[jruby] eregon commented on commit 870a9e9: I think we should use the pattern `getReadFrameSlotNode()` so side-effects are hidden in the specialization. https://git.io/v2cj9
skade has quit [Quit: Computer has gone to sleep.]
drbobbeaty has joined #jruby
<GitHub111>
[jruby] eregon commented on commit 1db32d2: We should have guards on Shapes like `isRubySymbol(Shape)` or as in ShapeCachingGuards.... https://git.io/v2CUC
<GitHub149>
[jruby] eregon commented on commit 1db32d2: The cache seems incorrect here, it should be either on identity or Shape (a Shape with a singleton class, so that means a singleton Shape so identity seems more useful).... https://git.io/v2CTq
vtunka has joined #jruby
<GitHub116>
[jruby] eregon commented on commit 149bdf9: I prefer no empty line here, it's already enough boilerplate IMHO. What do @jruby/truffle think? https://git.io/v2Ctf
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<GitHub51>
[jruby] eregon commented on commit d349235: We could have a global Assumption `only one Thread yet` to make it a bit nicer.... https://git.io/v2CtX
<GitHub77>
[jruby] eregon commented on commit db96418: Seems it should be `readOrNullNode` here. Please use an helper method for this or it just becomes unreadable and bug-friendly. https://git.io/v2CqV
<GitHub92>
jruby/master cf89d2f Benoit Daloze: [Truffle] More explicit name since the logic only makes sense for class variables.
<GitHub15>
[jruby] eregon commented on commit 0228c08: :+1: Always wanted to do this! https://git.io/v2CYm
mattwildig has joined #jruby
mattwildig has quit [Read error: Connection reset by peer]
mattwildig has joined #jruby
<GitHub129>
[jruby] eregon commented on commit 1b884f8: This cache should be smaller because the generic case is only 3 reads instead of 1 + compare.... https://git.io/v2C3k
<GitHub5>
[jruby] eregon commented on commit 2663f15: The purpose was to avoid N*N nodes for instance in the dispatch chain. But we probably want something more specialized like CheckMetaClassNode or so with a constant metaClass/shape anyway. https://git.io/v2Cs8
jensnockert has quit [Remote host closed the connection]
jensnockert has joined #jruby
n00bdev has quit []
<GitHub186>
[jruby] eregon commented on commit 6ee231d: This was used for the output in OSR compilation name, has this changed? https://git.io/v2CCY
vtunka has quit [Quit: Leaving]
<GitHub195>
[jruby] eregon commented on commit 6e3f361: AndInvalidate() ? It would make it clearer that value cannot be compiled as null. https://git.io/v2CCy
<GitHub88>
jruby/master 78d478b Benoit Daloze: [Truffle] No useless casts for RaiseException.getException().
skade has quit [Quit: Computer has gone to sleep.]
<GitHub18>
[jruby] eregon commented on commit 88da599: I think it would be more intuitive to have the `!` before the profile. Also the profile value would match with the condition value. https://git.io/v2Clc
<GitHub143>
[jruby] pitr-ch commented on commit 149bdf9: I prefer the empty line which separates two logical units, special case for null from the return value. https://git.io/v2ClC
skade has joined #jruby
<GitHub134>
[jruby] eregon commented on commit c18b79f: It's already a JavaDoc comment. This will make it less clear which one should be used. 99% of the cases it's this and not Truffle.getRuntime().getCallerFrame() for instance. https://git.io/v2C8u
yfeldblum has quit [Remote host closed the connection]
jensnockert has quit [Read error: Connection reset by peer]
jensnockert has joined #jruby
mattwild_ has joined #jruby
mattwildig has quit [Ping timeout: 244 seconds]
drbobbeaty has joined #jruby
shellac has joined #jruby
pawnbox_ has quit [Remote host closed the connection]
jensnockert has quit [Remote host closed the connection]
chi6rag has joined #jruby
jensnockert has joined #jruby
jensnockert has quit [Remote host closed the connection]
jensnockert has joined #jruby
<GitHub160>
[jruby] chrisseaton commented on commit e498859: Fixed. https://git.io/v2CPR
blaxter has joined #jruby
mattwildig has joined #jruby
mattwildig has quit [Remote host closed the connection]
johnsonch_afk is now known as johnsonch
<GitHub77>
[jruby] chrisseaton commented on commit 149bdf9: I think a blank line after `}` on a line on its own is a common part of many Java style guides. https://git.io/v2CXy
<GitHub170>
[jruby] chrisseaton commented on commit d349235: I don't think we want a global assumption - much better to let code paths that happen to just see one thread, no matter how many there are, specialise on their own. I've switched to using the DSL here. It's a pity I can't see a way to hoist the thread guard to be once per method. This probably has extremely limited practical benefit anyway. https://git.io/v2CHM
<GitHub78>
[jruby] chrisseaton commented on commit 1b884f8: I don't think it's important to tune this option now. Any argument either of us made would really just be guessing. All these numbers are arbitrary and what we need to do is run some serious code and do some machine learning to find the best combination for startup, for peak, and a balance between the two. https://git.io/v2CQv
<GitHub142>
[jruby] chrisseaton pushed 4 new commits to master: https://git.io/v2C7P
<GitHub142>
jruby/master e842fbe Chris Seaton: [Truffle] Make the context the first parameter to FlipFlopNode.
<GitHub142>
jruby/master bf2436d Chris Seaton: [Truffle] Move negation outside profile in UnlessNode.
<GitHub142>
jruby/master 2c8e164 Chris Seaton: [Truffle] Use DSL for ThreadLocalObjectNode
skade has quit [Ping timeout: 248 seconds]
<GitHub135>
[jruby] chrisseaton commented on commit 870a9e9: Done. I'll go back and look at other nodes' lazy allocation that I've refactored as well. https://git.io/v2Cdi
<GitHub195>
[jruby] eregon opened issue #3696: JRuby 1.7.20 causes EBAD while installing default gems https://git.io/v2Wfs
vtunka has quit [Quit: Leaving]
headius2 has joined #jruby
blaxter has joined #jruby
vtunka has joined #jruby
blaxter has quit [Client Quit]
mattwildig has joined #jruby
subbu is now known as subbu|away
mattwildig has quit [Ping timeout: 248 seconds]
tenderlove has joined #jruby
<GitHub75>
[jruby] chrisseaton commented on commit 6e3f361: Technically that's done by the assumed value node, but I think you're arguing for it for code clarity. I disagree - the only thing that's required here is the transfer. If someone else has a @CompilationFinal and so needs to invalidate, that's their business to manage. https://git.io/v2Wto
camlow325 has joined #jruby
<GitHub127>
[jruby] CompilationFinal commented on commit 6e3f361: Do you need me for something? https://git.io/v2Wqr
camlow325 has quit [Read error: Connection reset by peer]
rsim has quit [Read error: Connection reset by peer]
pawnbox has quit [Ping timeout: 244 seconds]
subbu|away is now known as subbu
rsim has joined #jruby
<kares>
how do I get a core-complete .jar to be build on 9K ?
<GitHub149>
[jruby] chrisseaton commented on commit 1db32d2: Well one thing we can easily do is run the guards on `cachedObject` rather than `object`. https://git.io/v2Wcp
rsim2 has quit [Ping timeout: 240 seconds]
blandflakes has joined #jruby
robbyoconnor has quit [Read error: Connection reset by peer]
e_dub has quit [Read error: Connection reset by peer]
e_dub has joined #jruby
mattwild_ has quit [Ping timeout: 252 seconds]
mattwildig has joined #jruby
rsim1 has joined #jruby
rsim has quit [Ping timeout: 240 seconds]
bbrowning is now known as bbrowning_away
mattwildig has quit [Ping timeout: 250 seconds]
mattwildig has joined #jruby
rsim has joined #jruby
rsim1 has quit [Read error: Connection reset by peer]
m4rCsi has quit [Ping timeout: 240 seconds]
<yopp>
chrisseaton, is it possible to make a "code heat map" in runtime somehow?
<chrisseaton>
op
<yopp>
basically, the how often particular line of code is executed?
<yopp>
I suppose that JIT is already somehow keeping track of that
<chrisseaton>
yopp: there's a research group at King's working on something like that
<chrisseaton>
But basic coverage works as well
<chrisseaton>
-Xtruffle.coverage.global=true
jeremyevans has quit [Read error: Connection reset by peer]
jeremyevans has joined #jruby
nirvdrum has quit [Ping timeout: 244 seconds]
m4rCsi has joined #jruby
headius2 has joined #jruby
<yopp>
chrisseaton, I was thinking more about API of some kind. To be able to export stats to external storage for further analysis.
<yopp>
Simple case: I have a very mature app that is running in production, and there defo some dead code. I'd like to get metrics from production servers, to find that dead code
<yopp>
or to get a better idea of hot parts of app, to check if there any room for improvements
tcrawley is now known as tcrawley-away
<GitHub19>
[jruby] headius pushed 2 new commits to ruby-2.3: https://git.io/v2lA7
<GitHub19>
jruby/ruby-2.3 3992617 Charles Oliver Nutter: Merge remote-tracking branch 'origin/master' into ruby-2.3...
<GitHub19>
jruby/ruby-2.3 e18220f Charles Oliver Nutter: Update to preinstalled rake version.
nirvdrum has joined #jruby
<lopex>
chrisseaton: *LayoutImpl are also generated by APT right ? for some reason I can't find them under eclipse .apt_generated
<lopex>
maven generated-sources is ok
<lopex>
trying to force eclipse to work and following eregon's wiki