<lopex> headius: but dont take my word for any thing I've said :P
<headius> ha
<headius> hey, enebo and I are going to BOTH be at GeeCON next spring
<headius> would be good to finally buy you a beer
<lopex> I'm in a jail remember ?
<headius> a gulag
<lopex> in russia
<lopex> in poland we have german camps
<lopex> ah, it
<lopex> it's a sensitive topic
<headius> yes I would bet it is
<headius> in the western world gulag is often synonymous with "bad prison far away from everything"
<headius> well US world I guess
<headius> we're both in the west in some way
<lopex> lots of poles were sent to gulags, so it's different here
<lopex> headius: did you run jcodings scripts for recent mri ?
<headius> no I have not yet
<headius> there might have been one or two tests that needed updates but there were very few encoding failures in 2.3 to 2.4
<enebo> lopex: WTF
<enebo> lopex: if you do not buy me a decent beer and bring your brothers beer to me I am not sure what I will think...and why are you still awake
<lopex> enebo: just give me yout address
<enebo> lopex: I can of course but I would rather see you
<lopex> enebo: and I'm just an owl
<enebo> lopex: you are just some miyazaki bot who happens to be on our channel
<enebo> lopex: however I have seen photos of you
<lopex> ha
<enebo> lopex: however if you want a beer exchange we can try...how bad can customs bee
<enebo> I do it in US and so far it works even though it is not allowed
<enebo> lopex: regardless of exchange I think we can have a fun time in person
<lopex> haha
<lopex> always
<enebo> lopex: I feel like we have a cosmic connection now...you cannot escape me
<lopex> :P
<enebo> lopex: seriously though you are still up?
<lopex> almost
<enebo> lopex: not that I was joking about the connection!
<enebo> lopex: also you see I ported the MRI goofy logic in magic comments
<lopex> yeah
<lopex> I've seen that
<enebo> I probably spent too much perf on char creation bs byte
<lopex> enebo: you had no choice
<enebo> lopex: yeah probably so but there are many places where MRI just accepts bytes in places
<lopex> it's mri
<enebo> yes
<lopex> what can you d\o
<enebo> we follow happily
<enebo> lopex: truth is we ride on back of what Ruby is...it is difficult to complain
<enebo> lopex: If I made a language it would be different but probably not nearly as popular
<enebo> lopex: truth is making the right mix is difficult
<enebo> lopex: Ruby has many problems but somehow those problems work out well
m4rCsi has quit [Quit: No Ping reply in 180 seconds.]
m4rCsi has joined #jruby
vifino has quit [Ping timeout: 248 seconds]
vifino has joined #jruby
_whitelogger has joined #jruby
anaeem1 has joined #jruby
anaeem1__ has joined #jruby
anaeem1 has quit [Ping timeout: 276 seconds]
jeremyevans has quit [Quit: leaving]
<travis-ci> kares/jruby (test-convert-java-pp3:dc3127a by kares): The build passed. (https://travis-ci.org/kares/jruby/builds/262382832)
<GitHub158> [jruby] kares pushed 3 new commits to master: https://git.io/v7XOE
<GitHub158> jruby/master 8dfbf42 kares: remove unused Java call-site site field
<GitHub158> jruby/master 648dbce kares: use Java call-site caching with (shared) to-integer coversion
<GitHub158> jruby/master 580c79e kares: prefer passing context down for RubyInteger's op_xxx overloads
anaeem1 has joined #jruby
anaeem1__ has quit [Ping timeout: 276 seconds]
<travis-ci> kares/jruby (test-call-site-2:d628422 by kares): The build passed. (https://travis-ci.org/kares/jruby/builds/262544816)
<travis-ci> kares/jruby (test-call-site-2:d9f1822 by kares): The build passed. (https://travis-ci.org/kares/jruby/builds/262545547)
vtunka has joined #jruby
chrisarcand has quit [*.net *.split]
lance|afk has quit [*.net *.split]
reto__ has quit [*.net *.split]
jhass has quit [*.net *.split]
michael_mbp has quit [*.net *.split]
mberg has quit [*.net *.split]
drbobbeaty has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
chrisarcand has joined #jruby
reto_ has joined #jruby
mberg has joined #jruby
jhass has joined #jruby
mberg is now known as Guest39259
lanceball has joined #jruby
michael_mbp has joined #jruby
anaeem1 has quit [Quit: Leaving...]
drbobbeaty has joined #jruby
<travis-ci> kares/jruby (test-call-site-2:41b2879 by kares): The build passed. (https://travis-ci.org/kares/jruby/builds/262592579)
<GitHub15> [jruby] kares opened pull request #4736: Fixnum/Float 'fast-ops' working even when re-opened (master...test-call-site-2) https://git.io/v7XK5
bbrowning_away is now known as bbrowning
vtunka has quit [Quit: Leaving]
<GitHub32> [jruby] kares closed issue #4679: SSL connection fails due to no available cipher on IBM JRE https://git.io/vQfZJ
<travis-ci> jruby/jruby (master:8dfbf42 by kares): The build passed. (https://travis-ci.org/jruby/jruby/builds/262545743)
ratnikov_ has quit [Ping timeout: 255 seconds]
ratnikov_ has joined #jruby
vtunka has joined #jruby
olle has joined #jruby
<travis-ci> kares/jruby (test-cache-numbers:63cfdbe by kares): The build passed. (https://travis-ci.org/kares/jruby/builds/262647376)
<subbu> headius, got it .. yes, more explicit instructions => more work during interpretation. build + compile time increase is probably marginal. if kwarg handling is not going to part of all calls, then, extracting it out into its own instructions make that explicit and can reduce the bloat of base / common calls.
<travis-ci> kares/jruby (test-cache-numbers:fda49ba by kares): The build passed. (https://travis-ci.org/kares/jruby/builds/262666014)
olle_ has joined #jruby
olle has quit [Ping timeout: 240 seconds]
olle_ is now known as olle
olle has quit [Ping timeout: 260 seconds]
drbobbeaty has quit [Ping timeout: 260 seconds]
drbobbeaty has joined #jruby
nowhereFast has joined #jruby
shellac has quit [Quit: Leaving]
<enebo> subbu: headius: I emailed you with kwargs writeup
<travis-ci> kares/jruby (test-cache-numbers:2e75db8 by kares): The build passed. (https://travis-ci.org/kares/jruby/builds/262718710)
<headius> enebo: got it, seems like we're on the same page
<headius> option 1 would be easy to do right now; add the appropriate info to CallBase when the last argument to the CallNode is a literal HashNode
<enebo> headius: I also think 1 is easier to implement
<enebo> headius: and it could still use 2 for out of band passing of key data
<headius> the benefit of 1 over 2 is that I can hide all of this inside an indy call site and not increase code size
<headius> it just shoves everything on the stack and dispatches with a mapping of keys to positions
<headius> inside the indy call site it would restructure stack args or use the out of band way
<enebo> so my only real question is making this part of IR itself (e.g. a #2 sort of soln) would we gain something from that
<enebo> subbu: I am hoping has an opinion on it
<headius> yeah I'm not sure either
<headius> inlining would be a big one, since there's nothing to tie passed kwargs with received kwargs once we've hidden it inside the call
<headius> but you can also reproduce it from the info we put in the call
<enebo> headius: but we have the info in Java APIs so we can do it in the inliner
<headius> yes
<headius> it just would be more to do by hand
<enebo> headius: but inliner for params is just a table and you register what holds the value from the callsite side
<headius> we could start on option 1 now and see what we can get out of that
<enebo> so it would fit well with that
<headius> if it looks like we need to be more IR-friendly later it can change
<enebo> true...right now this can be entirely encapsulated in callbase
<enebo> oh wait no it can't :)
<enebo> ok well it is not a big deal regardless
<enebo> IRBuilder will need to looksee and the emit a little differently and pass more crap into call
<enebo> but the .create() helpers can just have a variant for kwargs scenario
<enebo> First swipe could even be to make the hash immediately if it was to be done in pieces
<headius> yeah that would be fine
<headius> so instead of getting a final operand that's the load of the hash already created, it should be constructed with a list of key/operand pairs
<subbu> enebo, headius i'll take a look tonight and get back .. at a conference (wikimania) .. so, cannot focus on that right now.
<headius> WIKIMANIA
nowhereFast has left #jruby [#jruby]
swills has quit [Ping timeout: 248 seconds]
swills has joined #jruby
swills has joined #jruby
swills has quit [Changing host]
<headius> subbu: oh nice, Montreal must be nice in summer
<headius> I've only ever been there in February
joast has quit [Ping timeout: 246 seconds]
<subbu> ya .. not bad .. except I think i'll be stuck in the conference venue for the most part.
bbrowning is now known as bbrowning_away
<GitHub65> [jruby] enebo pushed 2 new commits to jruby-9.1.13.0: https://git.io/v71Qa
<GitHub65> jruby/jruby-9.1.13.0 8b35d30 Thomas E. Enebo: [noci] Update docs on args node handling.
<GitHub65> jruby/jruby-9.1.13.0 299c306 Thomas E. Enebo: allocating labels for lonely operator support even when the call was not part...
swills has quit [Ping timeout: 260 seconds]
swills has joined #jruby
<GitHub16> jruby/jruby-9.1.13.0 c1020af Thomas E. Enebo: Some more extraneous label creation removed.
<GitHub16> [jruby] enebo pushed 1 new commit to jruby-9.1.13.0: https://git.io/v71j8
<GitHub156> [jruby] enebo pushed 4 new commits to master: https://git.io/v71jV
<GitHub156> jruby/master d9566c6 Thomas E. Enebo: allocating labels for lonely operator support even when the call was not part...
<GitHub156> jruby/master 5afbd30 Thomas E. Enebo: [noci] Update docs on args node handling.
<GitHub156> jruby/master 5973a6a Thomas E. Enebo: Make literal hashes keep track of whether their keys are all symbols....
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<GitHub35> [jruby] headius pushed 2 new commits to master: https://git.io/v7MUs
<GitHub35> jruby/master ca55782 Charles Oliver Nutter: Simplify case/when eqq logic and enlist in normal invocation....
<GitHub35> jruby/master f11b413 Charles Oliver Nutter: Move JVM6 case/when eqq logic into JVM6 adapter.
bbrowning_away is now known as bbrowning