<GitHub31>
[jruby] nirvdrum pushed 5 new commits to master: https://git.io/vVFCC
<GitHub31>
jruby/master 994615a Kevin Menard: [Truffle] Fixed some lingering references to MultiplyRope (renamed to RepeatingRope).
<GitHub31>
jruby/master b720338 Kevin Menard: [Truffle] Renamed MutableRope to RopeBuffer.
<GitHub31>
jruby/master 7d9467e Kevin Menard: [Truffle] Removed the Rubinius-specific @hash_value ivar from String.
pawnbox has quit [Ping timeout: 246 seconds]
<GitHub127>
[jruby] nirvdrum commented on commit 699ee62: @chrisseaton Changes like this are how I think we'll make it easier for long-lengthed ropes. We could have specialized various cases here, but doing that across the board for every operation is going to be messy. Whereas the Ruby implementation of `String#chop!` doesn't care about `int` vs `long` and doesn't need to fetch all the bytes from the rope. Note this also required an improvement in the `string_previ
<GitHub47>
jruby/master 9def32d Kevin Menard: [Truffle] Documented what the `string_previous_byte_index` primitive does.
pawnbox has joined #jruby
pawnbox has quit [Ping timeout: 248 seconds]
nirvdrum has quit [Ping timeout: 276 seconds]
bb010g has joined #jruby
johnsonch is now known as johnsonch_afk
enebo has joined #jruby
enebo has quit [Client Quit]
pawnbox has joined #jruby
thedarkone2 has quit [Quit: thedarkone2]
yfeldblum has quit [Ping timeout: 250 seconds]
mj12` has quit [Ping timeout: 240 seconds]
kwando has quit [Ping timeout: 260 seconds]
mj12` has joined #jruby
kwando has joined #jruby
skade has joined #jruby
skade has quit [Quit: Computer has gone to sleep.]
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
yfeldblum has joined #jruby
yfeldblu_ has joined #jruby
yfeldblum has quit [Ping timeout: 268 seconds]
yfeldblu_ has quit [Remote host closed the connection]
yfeldblum has joined #jruby
<GitHub25>
[jruby] kares opened issue #3794: improved Date/Time conversions on Java 8 https://git.io/vVFMr
shellac has joined #jruby
yfeldblum has quit [Ping timeout: 250 seconds]
pawnbox has quit [Remote host closed the connection]
shellac has quit [Quit: Computer has gone to sleep.]
CaptainHeavy has joined #jruby
olle has joined #jruby
vtunka has joined #jruby
<GitHub119>
[jruby] kares opened issue #3795: generix collections in Ruby land https://git.io/vVFdZ
<GitHub98>
[jruby] andy-twosticks opened issue #3796: Refinements causes "ArgumentError: tried to create Proc object without a block" https://git.io/vVFFP
brauliobo has joined #jruby
vtunka has quit [Quit: Leaving]
pawnbox has joined #jruby
olle has quit []
yfeldblum has joined #jruby
shellac has joined #jruby
vtunka has joined #jruby
<GitHub159>
[jruby] jtulach opened pull request #3797: Skeleton of a test to check Ruby Truffle debugger integration (truffle-head...RubyDebugTest) https://git.io/vVFxS
drbobbeaty has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
shellac has joined #jruby
jimbaker has quit [Ping timeout: 260 seconds]
jimbaker has joined #jruby
jimbaker has quit [Changing host]
jimbaker has joined #jruby
raeoks has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<GitHub142>
jruby/master 2876388 Karol Bucek: Merge pull request #3779 from jruby/test-become-java...
<GitHub182>
[jruby] kares deleted test-become-java at 27d8098: https://git.io/vVbjy
vtunka has quit [Quit: Leaving]
thedarkone2 has joined #jruby
enebo has joined #jruby
drbobbeaty has quit [Read error: Connection reset by peer]
drbobbeaty has joined #jruby
<nirvdrum>
enebo: I'm only looking at 9.0.5.0 right now, but a String#concat heavy benchmark is about 50% slower with indy than without.
<enebo>
nirvdrum: neat
<enebo>
nirvdrum: would you mind opening an issue on it
<nirvdrum>
Will do. Is a bench9000 benchmark fine?
<enebo>
nirvdrum: sure
<enebo>
nirvdrum: I was telling Charlie I was seeing lots of benches with indy be slower than non-indy but then did not notice any recently. I half wonder if I had not been running any string-heavy benches
<enebo>
but was before
<nirvdrum>
1.7 is faster than 9k on this benchmark, too.
<nirvdrum>
I really hope my CodeRangeable indirection isn't at fault.
<nirvdrum>
But they're on par for indy. So it may just be something with Java 6 JTI.
<nirvdrum>
enebo: I tried to look at a quick profile to see why non-indy was slower on 9K than 1.7, but virtually the entire profile is CachingCallSite#call.
norc has joined #jruby
eshea has quit [Remote host closed the connection]
bbrowning_away is now known as bbrowning
eshea has joined #jruby
<kares>
enebo: headius: idea JRuby license is running out - not that I like it that much but I'm not sure there's anything better with .rb + .java support
<kares>
will you guys get a new license for JRuby ?
<kares>
actually using it for JRuby only
<enebo>
kares: I pinged asarih on friday? and he emailed them about renewal
<enebo>
kares: you should be able to reenter evaluation mode once license runs out for a number of days though
<enebo>
kares: so you should not be left hung out to dry
tcrawley is now known as tcrawley-away
<kares>
enebo: ok thanks - still use netbeans as well but the JRuby project or the Ruby plugin is somehow freezing it hard on times :)
<enebo>
kares: :|
shellac has joined #jruby
<norc>
Hey guys, we just found something weird in MRI which jruby somehow does the same, which leads me to think that this must be intentional behavior:
<norc>
a = 'a\'b'; a.sub '\'', '\\\'' #=> 'abb'
<norc>
Would any of you from the top of their head know why this is?
<nirvdrum>
enebo: I've narrowed down the 4x slowdown from 1.7x -> 9K.
<enebo>
nirvdrum: as to what is causing it?
<nirvdrum>
enebo: Well, I guess I don't have that confidence, but I found two changes that get the performance back on par.
<nirvdrum>
The tail end of that benchmark is _erbout.force_encoding(__ENCODING__)
<enebo>
norc: geezus I cannot even read that :) can you gist that as a script
<nirvdrum>
For whatever reason, accessing __ENCODING__ is slow.
<enebo>
norc: Perhaps I can figure out what is up
<nirvdrum>
And force_encoding is doing an extra frozen check compared to 1.7, so that might be enough to account for its regression.
<enebo>
nirvdrum: ah so the bench was specifically sintrgs + erb
<enebo>
SINTRGS
<nirvdrum>
enebo: I ran ERB, generated the fragment it would eval, and wrapped up that fragment sans eval into a benchmark.
<nirvdrum>
The idea was to approximate as closely as possible a real world templating engine.
<enebo>
nirvdrum: perhaps we cache in 1.7 and in 9k we ask for a charset every time or something like that
<enebo>
norc: yeah I will mess a minute or two to see what I can see with this
djellemah_ has joined #jruby
<nirvdrum>
enebo: Another thing this benchmark highlights is ERB calls force_encoding with what is almost certainly the existing encoding on the sting. In JRuby+Truffle, we already recognize that an no-op. Non-Truffle JRuby forces ByteList invalidation.
<nirvdrum>
And clears the code range.
<norc>
enebo, oh! I just found it, this was actually missing from the documentation for a while but is on current MRI HEAD.
<enebo>
norc: oh you mean it was broken before mri 2.3?
<norc>
enebo, no, just never documented.
<enebo>
norc: oh wow you mean since 2.3 was out
<norc>
enebo, the feature has been around for at least since 1.9.3, but hasn't been documented until recently it seems.
<enebo>
norc: ok well so long as we behave the same I am happy
<enebo>
hmm that quote does not work well for this one
<enebo>
yeah at least it is being documented. Tough to fix some behavior since to much software relies on weird behavior over that much time
<norc>
enebo, there is enough in Ruby that I would deprecate in a heartbeat..
<enebo>
norc: yeah I think python3k has scared matz to death on big changes
<enebo>
norc: deprecation perhaps can not be so big? Tough to say
<norc>
enebo, well it begins with bizarre things like keyword_super which eventually lead to so much frustration and debugging until you learn about the subtle difference. Not sure whether keeping stuff like that in is wise either.
shellac has quit [Quit: Computer has gone to sleep.]
tcrawley-away is now known as tcrawley
<enebo>
norc: I agree that it would be nice to have some errata process with deprecation to fix stuff
djellemah has quit [Disconnected by services]
djellemah_ is now known as djellemah
djellemah_ has joined #jruby
subbu is now known as subbu|lunch
robbyoconnor has quit [Ping timeout: 244 seconds]
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
byteit101 has quit [Quit: No Ping reply in 180 seconds.]
byteit101 has joined #jruby
subbu|lunch is now known as subbu
shellac has joined #jruby
tcrawley is now known as tcrawley-away
yfeldblum has joined #jruby
tcrawley-away is now known as tcrawley
pawnbox has quit [Remote host closed the connection]
pawnbox has joined #jruby
pawnbox has quit [Remote host closed the connection]
yfeldblum has quit [Ping timeout: 268 seconds]
shellac has quit [Quit: Computer has gone to sleep.]