<MacRohard>
I have a question about reflection.. I have a thing I want to get that's protected.. I can see it by doing this; fenv = @channel.getClass().getSuperclass().getDeclaredField("env") fenv.setAccessible(true) but I can't figure out what to pass to (Field)fenv.get to retreive it.. nothing seems to work
brightball has quit [Quit: Leaving...]
<MacRohard>
the thing in question is a Map - actually a LinkedHashMap if that makes any difference
<chrisseaton>
Just checking that you have a good reason for reading encapsulated fields from inside a LinkedHashMap - can you explain your use case and maybe we can do something else instead? Or maybe you really do need it for whatever reason, which is also fine.
<MacRohard>
I can't find any other way to access it
<chrisseaton>
Yeah, but can you explain why you need to access the field?
<MacRohard>
well it's actually a field in mina ssh that contains the environment for a remote ssh connection. I want to know what SSH_AUTH_SOCK was assigned to the connection
<chrisseaton>
I'm not even sure there is an 'env' field in a LinkedHashMap
colinsurprenant has joined #jruby
<MacRohard>
no that's the name of the field in the class that i'm getting it out of
<chrisseaton>
oh right I thought you were trying to read fields from a LinkedHashMap object
pierreatkillbill has quit [Ping timeout: 276 seconds]
<chrisseaton>
does that translate to what you're doing?
<MacRohard>
I guess "foo".to_java is java.lang.String or something?
<chrisseaton>
yeah
<MacRohard>
I need a Map
<MacRohard>
I think
<chrisseaton>
This reads a field from a given Java object - it doesn't matter what class it is
<mkristian>
to_hash or to_h gives you a ruby hash
<MacRohard>
I think you'll find it only reads strings
<chrisseaton>
This gets a char[] from the object - if you want to get a LinkedHashMap it will be the same code
<chrisseaton>
MacRohard: well it will read anything with a value field - just change the name of the field if that's not right for your class
<MacRohard>
the objecT I'm getting is a Map, not a String
<chrisseaton>
that's fine - set o to your object, change "value" to be the name of the field you want from your object, and it should do the same thing
<MacRohard>
what's my object though?
anaeem1_ has joined #jruby
<chrisseaton>
@channel you said
dinfuehr has quit [Remote host closed the connection]
<chrisseaton>
so do o = @channel
<MacRohard>
hmm. I've tried that and it doesnt' give any error but I jsut got an empty Map back.. maybe it is actually empty i guess
<chrisseaton>
If you got an actual object, then you've read the field successfully
<MacRohard>
I'm also confused as to why I needed .getSuperClass - shouldn't all the fields be available in the class itself?
<MacRohard>
like this works fenv = @channel.getClass().getSuperclass().getDeclaredField("env") but fenv = @channel.getClass().getDeclaredField("env") doesn't. is that normal?
<chrisseaton>
getDeclared field means declared in this class - your field looks like it is inherited from another class
<MacRohard>
it is
brightball has joined #jruby
<MacRohard>
hmm i think the answer is that the map is just empty
<MacRohard>
I tried writing something to it befor getting it and that works
<MacRohard>
so I'm still confused but at least this is working :)
<MacRohard>
thanks for the help everyone
cprice404 has quit [Quit: Konversation terminated!]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] eregon pushed 2 new commits to master: http://git.io/vvyFN
<JRubyGithub>
jruby/master bf7ad69 Benoit Daloze: [Truffle] Identify a few slow specs.
<JRubyGithub>
jruby/master 099a5f7 Benoit Daloze: [Truffle] Fix Kernel#{freeze,frozen?} for nil and Bignum.
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] jmiettinen opened issue #2844: WEBrick randomly fails with "can't add a new key into hash during iteration" http://git.io/vvybG
<enebo>
This is commit where is stops working for us
<enebo>
oh haha
<enebo>
this is a new spec
<headius>
hah ok
<headius>
eregon: if you update specs at least give spec:ruby a run
<headius>
this is an easy fix though
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] enebo pushed 6 new commits to master: http://git.io/vvS9J
<JRubyGithub>
jruby/master 512b0a5 Thomas E. Enebo: StaticScope now has Signature instead of three individual fields. All other...
<JRubyGithub>
jruby/master bff01dd Thomas E. Enebo: 1. staticScope.setSignature now only happens when ArgsNode is created and nowhere else (well some dead code which needs to be removed yet). For AOT, IR persistence still needs to be corrected so that we can popular signature when we reify the StaticScope (to be done in a followup commit)....
<JRubyGithub>
jruby/master 3f09737 Thomas E. Enebo: Unbreak StaticScope persistence from previous commits. This also indirectly fixes proper...
JRubyGithub has left #jruby [#jruby]
<enebo>
alright. things are looking good locally so let’s see about travis
<headius>
I will patch Bignum
<enebo>
headius: great
rcvalle has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/vvSHW
<JRubyGithub>
jruby/master 073d772 Charles Oliver Nutter: Bignums are frozen at create now.
JRubyGithub has left #jruby [#jruby]
<chrisseaton>
headius: can I get PATH_MAX via jnr at all?
<headius>
I don't know what that is, but if it's just a #define then the answer is no
<headius>
max path length I guess?
<enebo>
jnr-constants update should be simpler
<chrisseaton>
yeah - looks like you hard coded it as 1024
rsim1 has joined #jruby
<chrisseaton>
I'll just do the same and make sure any error is obvious
<enebo>
by simpler I mean not having to find n people to update it
<headius>
enebo: it's possible to just do it manually
<headius>
just cumbersome
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] eregon pushed 1 new commit to master: http://git.io/vvSQe
<JRubyGithub>
jruby/master 7b6475d Benoit Daloze: [Truffle] Simplify constructors of RubyCallNode....
JRubyGithub has left #jruby [#jruby]
<enebo>
chrisseaton: jnr-constants can define things like this but it requires running once on each supported platform
<chrisseaton>
jnr-posix is a delight to work with BTW - implementing the POSIX layer of Rubinius and it is super easy
<headius>
excellent
<headius>
9k uses it heavily too
kares has quit [Ping timeout: 244 seconds]
<chrisseaton>
why don't you use use POSIX for all of IO? Is it because you want to be able to have a Ruby file with a Java channel underlying it if you get channels from some API?
rsim has quit [Ping timeout: 265 seconds]
pierreatkillbill has joined #jruby
camlow325 has joined #jruby
e_dub has quit [Quit: ZZZzzz…]
subbu|breakfast is now known as subbu
<headius>
chrisseaton: platforms where we don't or can't have native support
<headius>
all our native stuff has to have fallback semi-gracefully
<chrisseaton>
headius: our only concern is moving so far away from the underlying JVM IO layer - not sure if some future problems might emerge maybe in tooling or something
<headius>
it's definitely a problem
e_dub has joined #jruby
<headius>
for example, you can't mix enxio selectors with JDK channels
<chrisseaton>
The SVM simply provides an implementation of POSIX that is native to the VM
<chrisseaton>
You have a JEP for HotSpot to do that don't you?
<headius>
I started it but I have not had cycles to help
<headius>
they're working on it
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] chrisseaton pushed 4 new commits to master: http://git.io/vvSbv
<JRubyGithub>
jruby/master 87be3a7 Chris Seaton: [Truffle] Move chdir to Rubinius.
<JRubyGithub>
jruby/master bb94fb0 Chris Seaton: [Truffle] Move rmdir to Rubinius.
<JRubyGithub>
jruby/master 0484d31 Chris Seaton: [Truffle] Move Dir.exist? into Rubinius.
<enebo>
headius: but I copletely removed all tags files related to arity that I knew of
<headius>
enebo: ok
<headius>
maybe my mistake, running now
<enebo>
with that said I might not have found all arity specs or tests
<enebo>
The nice thing with how things work now is they all use the same method from signature on staticscope
<enebo>
so I just need to tweak one piece of code now
<enebo>
My only fear is my simplification of only setting Signature once during parsing or when rehydrating static scope in IR persistence is too simple
<enebo>
running spec:ruby now…I am hoping I do not see those errors :)
camlow32_ has quit [Read error: Connection reset by peer]
camlow32_ has joined #jruby
camlow32_ has quit [Read error: Connection reset by peer]
camlow32_ has joined #jruby
camlow325 has quit [Ping timeout: 250 seconds]
camlow32_ has quit [Remote host closed the connection]
mister_solo has quit [Ping timeout: 245 seconds]
<headius>
enebo: reviewing pre2 issues?
tcrawley is now known as tcrawley-away
<headius>
I'm mopping up a couple now
pietr0 has joined #jruby
<nirvdrum>
enebo: I liked that one. I guess he missed the word "up"
bbrowning_away is now known as bbrowning
<lopex>
how does pre2 look perf wise ?
<nirvdrum>
headius, enebo: Unfortunately, this regexp issue on Windows looks legitimate. I don't know if it's the lexer or the joni changes, but "a1".match(/\p{L}/).to_a raises a SyntaxError.
<nirvdrum>
SyntaxError: (RegexpError) invalid character property name <L>: /\p{L}/
<lopex>
nirvdrum: it's the literal
<lopex>
nirvdrum: and I suspect it's an uncorrect __ENCODING__ for that file
<headius>
lopex: better than pre1 :-)
<lopex>
nirvdrum: forcing the encoding as utf via /../u or coding: utf-8 will not error
<lopex>
headius: number!
<lopex>
*numbers
<headius>
numbers numbers numbers!
<headius>
indy mode is as fast or faster than 1.7 on most things
<headius>
non-indy mode is still a bit slower
tcrawley-away is now known as tcrawley
<chrisseaton>
I'm running a new big report on 54 benchmarks - but it is slanted towards computationally intensive stuff, small objects and metaprogramming
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius pushed 1 new commit to master: http://git.io/vv9Xw
<JRubyGithub>
jruby/master d2ea85f Charles Oliver Nutter: Propagate coderange out of StringLiteral through all consumers.
JRubyGithub has left #jruby [#jruby]
<chrisseaton>
no IO, no encoding etc
<headius>
chrisseaton: ok
<nirvdrum>
lopex: But it works on non-Windows and worked a month ago.
<enebo>
nirvdrum: It works?
<lopex>
nirvdrum: yeah, since windows encoding logic might have changed
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed issue #2163: StringLiteral consumers may not be CR-aware http://git.io/vv913
JRubyGithub has left #jruby [#jruby]
<enebo>
nirvdrum: The <L> thing?
<lopex>
nirvdrum: on most linuxes utf is the default
<lopex>
nirvdrum: and that L is available exactly for unicode encodings
<lopex>
nirvdrum: so no wonder
<nirvdrum>
enebo: Yeah, it works fine on Linux.
<nirvdrum>
lopex: Makes sense.
<enebo>
ok I remember seeing that in some case but I don’t remember where/when
<lopex>
if for some reason it's iso-8859-x the it will error
<lopex>
*thne
<nirvdrum>
Actually, this file looks somewhat new. A spec eregon backported maybe?
<nirvdrum>
In that case, I guess we should just set the encoding on it and call it a day.
<enebo>
lopex: do you think it is possible if it is present the regexp is converted to //u?
<headius>
eregon did update specs recently
<lopex>
enebo: if what's present ?
<enebo>
<L>
rsim1 has quit [Quit: Leaving.]
<nirvdrum>
eregon: Are you around?
camlow325 has joined #jruby
<lopex>
enebo: u will fix the issue
camlow325 has quit [Remote host closed the connection]
<enebo>
lopex: I am sure it wlil but I am wondering if perhaps it just existing in a regexp will convert the regexp to be unicode
<lopex>
enebo: it's too late at that time
<enebo>
(eating so I am not typing much :) )
camlow325 has joined #jruby
<lopex>
enebo: it already parsed the regexp using former encoding
e_dub has quit [Quit: ZZZzzz…]
<enebo>
lopex: and you know MRI has no special logic before that point?
<lopex>
enebo: mri will error out as jruby does
<enebo>
ah I just tried it
<lopex>
enebo: or there's some lexer issue
<lopex>
enebo: which makes that file utf-8 somehow
<enebo>
lopex: both MRI and JRuby behave the same here
<lopex>
in mri
<lopex>
yes, they should
<enebo>
lopex: so it must be default external encoding without an explicit encoding
<lopex>
yes
<lopex>
enebo: also, can different versions of windows/java report system encoding inconsistently ?
<nirvdrum>
lopex: However, MRI 2.1.5 is able to handle that character class just fine.
<lopex>
nirvdrum: on windows ?
<nirvdrum>
Yes.
<nirvdrum>
So even if it is a new spec, it looks like we're doing something wrong.
a5i has quit [Quit: Connection closed for inactivity]
<enebo>
nirvdrum: do -e ‘p __ENCODING__’
<lopex>
nirvdrum: yeah, try that
<lopex>
I get utf-8
<lopex>
on mingw 2.2.1
<nirvdrum>
MRI is UTF-8. JRuby is IBM437.
<lopex>
there you go
<nirvdrum>
MRI reports Encoding.default_external as IBM437.
<lopex>
I get #<Encoding:CP852>
<nirvdrum>
Up until 30s ago, I didn't know __ENCODING__ even existed. So I don't know what to make of the discrepancy.
<headius>
enebo: adding an rc2 milestone
<enebo>
headius: really?
<enebo>
headius: so we know rc1 won’t be enough?
<headius>
well we need something for post pre2 won't we?
<headius>
oops
<headius>
I meant rc1
<enebo>
hah ok yeah
<nirvdrum>
Heh.
<headius>
I have some muscle memory for rc2 for some reason
<lopex>
pre-rc2
<enebo>
rc9000
<nirvdrum>
enebo: So should our __ENCODING__ default to UTF-8 as well?
<nirvdrum>
Oh, shit.
<nirvdrum>
I take that back.
mister_solo has joined #jruby
<enebo>
nirvdrum: console is a weird one. I believe there is some transcoding happening for all things printed to stdout if ttty?
<nirvdrum>
If I run with -e, __ENCODING__ is IBM437. If I do it in a file, it switches to UTF-8.
<nirvdrum>
JRuby sticks with IBM437 in either case.
<enebo>
nirvdrum: yeah I wondered about that
<lopex>
so -e ?
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed issue #2510: Explicit require of autoloaded class triggers double load http://git.io/trBYSw
JRubyGithub has left #jruby [#jruby]
<enebo>
nirvdrum: because they cannot really console it as UTF-8 and have it print without a trancode
<nirvdrum>
lopex: -e is IBM437 in each.
<nirvdrum>
lopex: But when reading a file, MRI switches to UTF-8 and JRuby sticks with IBM437 (Encoding.default_external).
<enebo>
nirvdrum: so I think any tty? IO’s might be IBM437 and all others will be UTF-8
<lopex>
nirvdrum: yeah, I wondered if it works that way
<enebo>
nirvdrum: I remember now that somewhere there is magic to print out encoding as IBM437 somewhere for stdout
<enebo>
nirvdrum: there must be some weird logic for UTF-8 somewheres in MRI. I would guess in windows ifdef since it is UTF-8 by default everywhere else already
<nirvdrum>
enebo: But it looks like this is set by the parser?
<enebo>
travis always horks when I commit something big
subbu has quit [Ping timeout: 256 seconds]
<headius>
ok this is something different
<nirvdrum>
enebo: Have a few minutes?
<headius>
ack
<enebo>
nirvdrum: sure
<headius>
enebo: psych.jar is missing PsychService.class
<enebo>
Psychotic
<headius>
something wrong with psych gem build
<nirvdrum>
enebo: Cool. So, I'm trying to track this down. It looks to me like Ruby#setupSourceEncoding should default to UTF-8, not the locale encoding. I made that change. But it looks like the lexer ignores whatever ParserConfiguration has anyway.
<nirvdrum>
GetsLexerSource has this frobnicateEncoding method that does something.
<enebo>
heh
<enebo>
nirvdrum: yeah hmmm let me look at setupSourceencoding
<nirvdrum>
frobnicateEncoding sets to default_external. That'd be fine -- it's supposed to get overridden by RubyLexer#parser_prepare. But that reads the encoding from the first line of the file and that encoding is influenced by whatever is set when the line was read, I think.
pjammer has quit [Quit: My Mac has gone to sleep. ZZZzzz…]
<enebo>
yeah and I can see we basically ignore it
<nirvdrum>
I could just have the chain here wrong.
<enebo>
we == me when I switched lexers
<enebo>
This seems to be wrong in more than one place too
<enebo>
ByteList in = new ByteList(content, runtime.getDefaultExternalEncoding());
<enebo>
This is one line in Parser related to bytelistlexersource
<nirvdrum>
My guess is if you change your locale locally, you'll see problems similar to Windows.
<enebo>
this should probably be getting it from parserConfig
<enebo>
parserconfig will generally be default external which is why this is not manifesting more
<nirvdrum>
You can tell it's a frustrating day when you go to edit shit.rb and the file already exists :-/
<nirvdrum>
Yeah.
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius opened issue #2846: Psych dependency is installing non-JRuby version http://git.io/vvHr7
JRubyGithub has left #jruby [#jruby]
<enebo>
heh
<nirvdrum>
enebo: Write a file that just contains 'p __ENCODING__' and then run it setting LC_ALL=C
<enebo>
nirvdrum: I will see what I can change here. I am mildly confused in all cases here because sometimes we receive a bytelist and a bytelist already has an encoding
<nirvdrum>
JRuby prints US-ASCII. MRI prints UTF-8.
<headius>
oh you nkow what
<headius>
I wonder if we're copying the stdlib version
<enebo>
C?
<headius>
the .rb files appear to be the stock stdlib ones
<nirvdrum>
I think it's a short-hand for POSIX, which is some ancient holdover.
<headius>
but we have a bit of extra logic to load our ext
<nirvdrum>
enebo: Like I said, my analysis could be wrong. This is the first I've looked at MRI's parser, but I couldn't find anything that looked like it was win32 specific.
<enebo>
nirvdrum: you are likely correct here
<lopex>
enebo: same here
<enebo>
nirvdrum: all the paths coming in confuse me since several of these do not call this setupmethod to parserconfig
<kares>
codefinger: oh nice, did not expect it to be that simple ... but than on the other hand why not
<kares>
if it boots - great work, thanks!
e_dub has joined #jruby
<codefinger>
kares: yea, i was almost thinking "this is too easy, must be wrong"
<codefinger>
this will be a lot easier to extend
x1337807x has joined #jruby
<kares>
yy
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] headius closed issue #2846: Psych dependency is installing non-JRuby version http://git.io/vvHr7
JRubyGithub has left #jruby [#jruby]
<headius>
enebo: false alarm
<enebo>
lopex: nirvdrum: Do we even have LC_ALL awareness?
lance|afk is now known as lanceball
<enebo>
config.getEncoding() shows UTF-8
erikhatcher has quit [Quit: erikhatcher]
<enebo>
headius: hooray
<nirvdrum>
enebo: We do. I just need to find it again.
<lopex>
enebo: only one mention of LC_ALL in mri is langinfo.c
<nirvdrum>
It's used by some rubyspecs.
<enebo>
getEncodingService().getLocaleEncoding() should be what if I LC_ALL=C
<nirvdrum>
US-ASCII, I think.
<enebo>
yeah it is UTF-8 for us
<enebo>
so configuration was not passing it to parser but it is not set in the first place
<lopex>
but how come mri doesnt need it ?
<enebo>
THEY NEED IT
<nirvdrum>
See: spec/ruby/core/encoding/default_external_spec.rb for instance.
<lopex>
I cant find it
<nirvdrum>
But those are ruby_exe specs.
<enebo>
So this script I am running will change default_external to US-ASCII but it will leave __ENCODING__ as UTF-8
<nirvdrum>
I mean, the snippet I showed in the issue does change the output at least on my system.
<nirvdrum>
Maybe Mac ignores it?
<enebo>
yeah it does for me too
<enebo>
but __ENCODING__ which the parser receives is UTF-8
<enebo>
default_external is being messed with but not being used by the parser
<enebo>
My mind feels like this is backwards so I am not sure I still understand anything more than parser will not neccesarily use default_exterminal
<enebo>
EXTERMINAL
subbu has joined #jruby
<enebo>
lopex: I guess it uses something called nl_langinfo
erikhatcher has joined #jruby
<enebo>
langinfo.c and localeinit.c are two files which seems to do all this stuff
<headius>
bbl
mister_solo has quit [Ping timeout: 250 seconds]
skade has joined #jruby
<nirvdrum>
I get different values when I debug. So that's fun.
<enebo>
nirvdrum: make sure you specfy in an ENV var field in your debugger
<enebo>
nirvdrum: in intellij you need to enter in a different area
<nirvdrum>
I was trying a remote debugger, but yeah.
zorak8 has quit [Remote host closed the connection]
<nirvdrum>
I wasn't even debugging the right thing. Wrong runtime configuration.
<enebo>
yay
<nirvdrum>
enebo: In my case, frobnicateEncoding is returning US-ASCII.
<nirvdrum>
From this line: if (!io.respondsTo("encoding")) return io.getRuntime().getDefaultExternalEncoding();
<enebo>
yeah
<enebo>
nirvdrum: This line is actually only for ripper too
<enebo>
nirvdrum: it should be configuration.getEncoding()
<enebo>
nirvdrum: with that said I am always UTF-8 and it is ignoring LC_LANG altogether
<enebo>
nirvdrum: I mean once I see LC_LANG work I should see parser get worse in my case because it will use default external which will be LC_ALL
<enebo>
I meant LC_ALL for LC_LANG
<nirvdrum>
A lot of the core files seem to explicitly set the file encoding.
<nirvdrum>
And that goes down a different path.
<nirvdrum>
But maybe something else is at play.
<lopex>
what's the status of IR persitence ?
<enebo>
lopex: it works and is used for AOT only atm
<lopex>
enebo: how reliable is it ?
<lopex>
the AOT
<enebo>
lopex: I don’t know
<enebo>
lopex: as reliable as IR persistence
<enebo>
lopex: which can write and read in simple rails app and gem list
<enebo>
lopex: so it seems to be working…
<lopex>
oh, I guess rails is more than I'll ever need
<enebo>
lopex: It is an improvement over all AOT because it loads as startup interp and will JIT so it will generate a runnable .class 100% of the time
<enebo>
lopex: Also it means improvements to later versions of JRuby will end up getting applied to AOT compiled classes
_djbkd has quit [Remote host closed the connection]
<enebo>
nirvdrum: I am fixing another IR persist bug involving AOT. I will come back to this in a bit
pjammer has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
<nirvdrum>
enebo: No problem.
erikhatcher has quit [Ping timeout: 252 seconds]
_djbkd has joined #jruby
pierreatkillbill has quit [Ping timeout: 240 seconds]
<JRubyGithub>
[jruby] enebo pushed 1 new commit to master: http://git.io/vvQl1
<JRubyGithub>
jruby/master 1009f65 Thomas E. Enebo: Follow up fix to #2840. TraceInstr was missing decode(). String was encode/decode needed ability...
<Antiarc>
So jruby destroys MRI once the JIT warms up (no surprise there), but I was really shocked that `for` was so much faster in 1.7.19 than anywhere else, and it's *way* slower in jruby-head
<lopex>
numbers!
<Antiarc>
It's not particularly a problem, but I'm curious about the huge swing there
<lopex>
I'd guess it will be there once IR development settles for good
<lopex>
for still uses each right ?
<Antiarc>
I haven't looked at the implementation yet
<lopex>
Antiarc: I'd guess there's a lot of special casing for that in 1.x and there was no time to make special "matchers" on 9k side
<Antiarc>
The first three cases each losing about 10% of their performance from 1.7.19 -> jruby-head is what mostly stuck out to me, but the `for` swing was *huge* and tickled my curiosity bone.
<Antiarc>
Ahhh, okay.
Obeliks_ has joined #jruby
<lopex>
Antiarc: the point is that now optimizing 9k given it's IR will be a whole lot easier and more consistent
<lopex>
so I wouldnt worry about slowdowns
<Antiarc>
Excellent. I'm excited to see what comes out of it :)
<Antiarc>
Yeah, I'm not worried, it just stuck out to me
<Antiarc>
"make it work, then make it fast" :)
Obeliks has quit [Disconnected by services]
Obeliks_ is now known as Obeliks
<lopex>
right
<lopex>
some optz will be trivial, and werent possible on 1.x branch
<lopex>
I guess alias analysis might be one of them
<Antiarc>
Hehe, that'll be fun
<enebo>
I wonder if we are caching range in 1.7 and not 9k
<lopex>
literals ?
<Antiarc>
The early optimization passes where you get huge gains out of small change is probably the most enjoyable part of development, IMO
<nirvdrum>
Antiarc: I'm a big fan of when the compiler guys make things faster and all I need to do is upgrade.
<lopex>
the other point is that IR is being used in virtually all compilers
erikhatcher has joined #jruby
<Antiarc>
nirvdrum: Indeed!
_djbkd has quit [Remote host closed the connection]
<lopex>
nirvdrum: for jruby code is both sides, jruby itself and jdk
<lopex>
so dont forget to update both
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
<nirvdrum>
lopex: Indeed. A big part of why I started using JRuby in the first place.
_djbkd has joined #jruby
erikhatcher has quit [Quit: erikhatcher]
_djbkd has quit [Remote host closed the connection]
elia has quit [Quit: Computer has gone to sleep.]
e_dub has quit [Ping timeout: 240 seconds]
_djbkd has joined #jruby
_djbkd has quit [Remote host closed the connection]
DrShoggoth has quit [Ping timeout: 265 seconds]
codefinger has quit [Quit: Leaving...]
jsvd has joined #jruby
_djbkd has joined #jruby
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] enebo pushed 1 new commit to master: http://git.io/vvQwg
<JRubyGithub>
jruby/master 3735705 Thomas E. Enebo: Fix indy string regression
JRubyGithub has left #jruby [#jruby]
JRubyGithub has joined #jruby
<JRubyGithub>
[jruby] bjfish opened pull request #2848: [Truffle] Adding additional long handling for multiple specializations. (master...truffle_long_specializations) http://git.io/vvQrt
JRubyGithub has left #jruby [#jruby]
elia has joined #jruby
mister_solo has joined #jruby
_djbkd has quit [Remote host closed the connection]