<GitHub115>
[jruby] eregon commented on commit 5f631b9: I liked my comment, this one indicates that InternalMethod name can change, but this never does.... https://git.io/vrJZ9
<GitHub29>
[jruby] eregon commented on commit 5f63605: `kind` could be `core` here or `builtin`. https://git.io/vrJZb
skade has quit [Quit: Computer has gone to sleep.]
phlebas has joined #jruby
skade has joined #jruby
drbobbeaty has joined #jruby
fuzzyhorns has joined #jruby
Caerus is now known as Caerus|Away
<eonwe>
headius: falling asleep at that time. Regarding the locking + stack overflows, I think it's not worth to fight if there are other non-closeable resources JRuby uses that need to be handled in finally-blocks. If it were just locking, then using monitors should help. Or waiting for http://openjdk.java.net/jeps/270
grs has joined #jruby
<GitHub135>
[jruby] ahorek opened pull request #3866: expand_path should use the original encoding (master...encoding_issue) https://git.io/vrJVK
<GitHub151>
[jruby] eregon commented on commit 3eff054: cc @jruby/truffle building a RubyNode[] to fill with null is no longer necessary, the generated code handles it already. https://git.io/vrJrA
nirvdrum has joined #jruby
<GitHub60>
[jruby] chrisseaton commented on commit 3eff054: Ah that's amazing! Now we just need an override of `create` that takes no parameters and passes `null` for you. https://git.io/vrJoD
<GitHub106>
[jruby] nirvdrum commented on commit 3eff054: Great. I've been meaning to look into fixing that, but haven't been able to find the time. Now I don't need to! https://git.io/vrJ6Z
<GitHub49>
[jruby] eregon closed issue #3858: [Truffle] remove_const doesn't completely clear a module's name https://git.io/vwx20
phlebas___ is now known as phlebas
pawnbox has quit [Remote host closed the connection]
<eregon>
headius: there is some flaggy MRI tests, saw a couple EBADF and some race in some Queue#close test, bu the rest should be green
pawnbox has joined #jruby
<headius>
yeah there are a few tests in MRI that are a little sensitive...I need to spend a day and fix them
<GitHub56>
[jruby] chrisseaton commented on commit 5f63605: Does anyone ever use kind? I'm trying to simplify the features of source section we use to work out what is really useful and inform a redesign of the API with @mlvdv https://git.io/vrJ9z
blandflakes has joined #jruby
<GitHub140>
[jruby] chrisseaton commented on commit 5f631b9: I removed it because it's a JavaDoc comment, but attached to a private field.... https://git.io/vrJ99
pawnbox has quit [Remote host closed the connection]
<GitHub101>
[jruby] headius commented on commit 5f631b9: I don't see any reason not to have JavaDoc for private methods and fields. You can enable generating docs with private items included, and it gets picked up by IDEs. https://git.io/vrJQm
<GitHub78>
[jruby] chrisseaton commented on commit 5f631b9: But who would see them? If I'm using `SharedMethodInfo` then I don't even know this field exists. https://git.io/vrJQM
Aethenelle has joined #jruby
pawnbox has joined #jruby
donV has quit [Quit: donV]
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
<GitHub60>
[jruby] eregon commented on commit 5f631b9: I put them on fields so that when looking at the class implementation it clarifies what fields serves what purpose.... https://git.io/vrJdT
fuzzyhorns has joined #jruby
<GitHub13>
[jruby] eregon commented on commit 5f63605: It appears in TraceTruffleCompilation output AFAIK, which might be useful to easily differentiate methods implemented in Ruby or Java. https://git.io/vrJdb
<GitHub147>
[jruby] headius commented on commit 5f631b9: @chrisseaton Anyone working on SharedMethodInfo would see them. But also...why delete it completely? If the comment had value, it should be restored...and then the difference is in one `*` that will let IDEs show the doco. This seems like a philosophical debate at this point. https://git.io/vrJdh
<headius>
chrisseaton: it may be the most realistic idea but I have not found anyone doing that for any language
nirvdrum has joined #jruby
<chrisseaton>
Yeah it's a full research project (I think Evan said it wasn't but I disagree)
<headius>
seems a bit like the re-compiling profiling bits I believe they removed some years ago because they were half-baked
<chrisseaton>
I'd take it on if I were a postdoc looking for ideas
<headius>
they = llvm
blandflakes has joined #jruby
<headius>
sure, it would be interesting to see how far a surface-level llvm JIT with all IR available would see through C code
<headius>
sufficiently smart compiler framework
<chrisseaton>
Yes that would be an easier step - Evan was talking about full partial evaluation, and didn't discuss the problem of when to stop
<headius>
I'm sure it will all be fine
<headius>
yeah, there's the rub...most of the jit work I've seen for MRI glosses over when to JIT and when to back off
<headius>
some of them glossed over guards too and assumed last run would be the same as this one
Aethenelle has joined #jruby
enebo has joined #jruby
<headius>
ugh, I need to fix that circular require warning
<headius>
WE need to
<headius>
sure is a lot of weird output from test:jruby now
<headius>
like a puts got left in somewhere
camlow32_ has quit [Remote host closed the connection]
donV has joined #jruby
camlow325 has joined #jruby
bjfish2 has joined #jruby
vtunka has quit [Quit: Leaving]
blandflakes is now known as mid-afternoonwal
mid-afternoonwal is now known as middaywalker
<fuzzyhorns>
hey all, i am having an error that traces to /opt/jruby/jruby-9.0.5.0/lib/ruby/stdlib/jar_dependencies.rb:309 but that doesnt seem to exist at that place anymore in the current codebase on github
<fuzzyhorns>
can anyone give me a pointer where it might be?
<headius>
fuzzyhorns: those are installed at build time...jar_dependencies is just a gem
<codefinger>
getting "Could not generate DH keypair" no matter what i try
<codefinger>
it works if I make the call with javax.net.ssl.HttpsURLConnection directly
donV has quit [Quit: donV]
<GitHub195>
[jruby] chrisseaton pushed 2 new commits to master: https://git.io/vrURG
<GitHub195>
jruby/master bdb1225 Chris Seaton: Update MRI benchmarks.
<GitHub195>
jruby/master 3b2fc41 Chris Seaton: Rename 'yarv' to just 'mri' in benchmarks for simplicity.
brauliobo has quit [Ping timeout: 276 seconds]
<GitHub69>
[jruby] mohamedhafez opened issue #3868: 9.1.0.0 hangs on syswrite https://git.io/vrUR1
lanceball is now known as lance|afk
grs has quit [Ping timeout: 240 seconds]
grs has joined #jruby
hobodave has quit [Ping timeout: 265 seconds]
<codefinger>
kares: when i use JRuby it's using TLSv1.2 and TLS_DHE_RSA_WITH_AES_128_CBC_SHA, but when I use HttpsURLConnection it uses TLSv1.2 and TLS_ECDHE_RSA_WITH_AES_128_GCM_SHA256