aemadrid_ has joined #jruby
chrisseaton_ has joined #jruby
yosafbridge` has joined #jruby
codefinger_ has joined #jruby
hoodow_ has joined #jruby
hoodow_ has joined #jruby
hoodow_ has quit [Changing host]
matthewd_ has joined #jruby
codefinger has quit [*.net *.split]
hoodow has quit [*.net *.split]
yosafbridge has quit [*.net *.split]
aemadrid has quit [*.net *.split]
matthewd has quit [*.net *.split]
chrisseaton has quit [*.net *.split]
codefinger_ is now known as codefinger
matthewd_ is now known as matthewd
chrisseaton_ is now known as chrisseaton
aemadrid_ is now known as aemadrid
adam12 is now known as adam
adam is now known as Guest62357
enebo has quit [Quit: Leaving.]
swills has quit [Remote host closed the connection]
swills has joined #jruby
swills has joined #jruby
swills has quit [Remote host closed the connection]
<kares> headius: could use your thoughts on https://github.com/jruby/jruby/pull/4718 ... if it looks good (there's one regression caused that I will fix) I'll put it on 2.4 so you do not have to deal with potential conflicts
swills has joined #jruby
swills has quit [Remote host closed the connection]
swills has joined #jruby
swills has joined #jruby
swills has quit [Remote host closed the connection]
prasun has joined #jruby
nicoulaj46 has quit [Quit: nicoulaj46]
nicoulaj46 has joined #jruby
nicoulaj46 is now known as nicoulaj
vtunka has joined #jruby
vtunka has quit [Client Quit]
vtunka has joined #jruby
shellac has joined #jruby
prasun has quit [Ping timeout: 258 seconds]
prasun has joined #jruby
prasun has quit [Ping timeout: 258 seconds]
prasun has joined #jruby
prasun has quit [Ping timeout: 258 seconds]
drbobbeaty has quit [Ping timeout: 246 seconds]
vtunka has quit [Quit: Leaving]
vtunka has joined #jruby
drbobbeaty has joined #jruby
shellac has quit [Remote host closed the connection]
swills has joined #jruby
clayton has quit [Ping timeout: 276 seconds]
clayton has joined #jruby
shellac has joined #jruby
bbrowning_away is now known as bbrowning
vtunka has quit [Quit: Leaving]
vtunka has joined #jruby
enebo has joined #jruby
vtunka has quit [Quit: Leaving]
shellac has quit [Quit: Leaving]
<GitHub87> [jruby] enebo pushed 2 new commits to symbol_love: https://git.io/vQxFX
<GitHub87> jruby/symbol_love 03c9565 Thomas E. Enebo: Remove no longer used constvardecl from IR
<GitHub87> jruby/symbol_love 61c7517 Thomas E. Enebo: Some deprecation warnings cleared up..
<GitHub95> [jruby] enebo pushed 1 new commit to ruby-2.4: https://git.io/vQxbT
<GitHub95> jruby/ruby-2.4 bfbc671 Thomas E. Enebo: Remove no longer used constvardecl from IR
bbrowning is now known as bbrowning_away
<headius> kares: I'll have a look
<headius> I saw you were working on this
<headius> kares: looks good, I had just a few items
<headius> 2.4 branch is close...I'm doing another pass over failures today and then I'll exclude and open an issue for the remainder
<enebo> I was correcting some of our tests for integer unification
<enebo> but test_integer_overflows is a bit crazy now
<enebo> hmm maybe it is only for formatting of errors
<enebo> vvs is hilarious in that file too...he uses his initials as var names
<enebo> ok yeah only for error strings...simplifying to only use Integer
<GitHub194> [jruby] enebo pushed 1 new commit to ruby-2.4: https://git.io/vQpcx
<GitHub194> jruby/ruby-2.4 b2cd9b8 Thomas E. Enebo: Remove some ::Fixnum/::Bignum deprecation warnings to de-noise this a bit
<headius> yay
<enebo> we still have a single one in date.rb to correct but that likely is in our other repo
deep-book-gk_ has joined #jruby
<headius> enebo: huh, I'll have a look
<headius> where is it?
<enebo> millis, @sub_millis = ((dt_or_ajd - UNIX_EPOCH_IN_AJD) * 86400000).divmod(1) raise ArgumentError, "Date out of range: millis=#{millis} (#{millis.class})" unless Fixnum === millis
<enebo> datr.rb has Fixnum check for overflow
<headius> right ok
<enebo> what is the check for this?
<headius> right, MRI moved most of this native
<enebo> or should I say how can it be done in Ruby now?
<headius> so our copy isn't getting updated anymore
<enebo> well our copy seems to match what we have in stdlib
<headius> we probably should just keep our copy in our repo only
<enebo> unless this is really native for us too
<headius> it's largely rewritten by eregon to use joda and nothing merges into it
<enebo> they do still have a file though
deep-book-gk_ has left #jruby [#jruby]
<enebo> it is only a small class though
<headius> right
<headius> looks mostly structural
<enebo> I am less concerned about merging that how we fix the warning
<headius> hmm yeah
<headius> I know the C call to check for fixnum range but not from Ruby
<headius> actually, just use java.lang.Long.MAX_VALUE
<enebo> I looked at a couple 'we unified things" posts but nothing which pointed out some libraries still need to know precision
<headius> it immediately goes into Java anyway
* headius skirts the issue
<headius> I'll remove date.rb from our ruby stdlib fork
<enebo> so raise if millis > MAX_VALUE or < 0?
<enebo> oh it is Ruby though
<enebo> so it will be fine for < 0
<enebo> so > MAX will be a bignum for us
<headius> well Fixnum === would be >= MIN and <= MAX
<enebo> true but this is a date
<headius> basically any long and it proceeds
<headius> right, it's just trying to ensure we can fit it in a long
<enebo> yeah
<enebo> ok I will fix in lib/ruby/stdlib
<enebo> but should I git rm from our branch or what?
<enebo> for ruby
<headius> yes
<headius> it has no correlation to anything in MRI lib anymore
<headius> nothing mergeable anyway
<enebo> ok
<headius> I dream of a future where travis loads immediately
<enebo> test_collect_concurrency : Detected invalid array contents due to unsynchronized modifications with concurrent users
<enebo> headius: do you know if array or collect itself changed on ruby-2.4
<enebo> This is the only fail for test:jruby
<headius> hmm
<enebo> get like 20 of those lines while it runs
<headius> there have been some changes but I thought they were mostly minor
<headius> yeah I think they might come from a concurrency test
<headius> it would be good to know though
<headius> in case something regressed
<GitHub16> [jruby] enebo pushed 1 new commit to ruby-2.4: https://git.io/vQplA
<GitHub16> jruby/ruby-2.4 f60985c Thomas E. Enebo: Change away from Fixnum for date.rb impl since Fixnum is Integergit add lib/ruby/stdlib/date.rbgit add lib/ruby/stdlib/date.rb
<enebo> HAHAHA
<enebo> I love you bash
Guest62357 is now known as adam12
<headius> enebo: you and subbu might be interested in this one: https://bugs.ruby-lang.org/issues/12705
<headius> one failure in test_lambda and there are some specs eregon wrote (already tagged)
<headius> hahaha
<headius> smooth
<headius> I guess I have a simple enough fix for that one
<enebo> that one == conc array or lambda?
<headius> assert_equal(mods, mods.map {m.owner.tap {m = m.super_method}})
<headius> the lambda thing
<headius> is fixed
<headius> this line is from test_method trying to check lineage of a prepended module
<headius> I bet nobu wrote this test
<headius> it calls tap there solely to return m.owner while also reassigning m
<headius> at least his fix looks simple
* subbu leaves it for enebo :)
<subbu> or headius maybe? :)
<enebo> hahaha
<enebo> that code
<enebo> I am a big fan of tap but as some weird corner case test it does not help aid in readability
<headius> enebo: it's pure nobu
<headius> 200 proof nobu
<enebo> I always wonder how much languages style infect other languages
<enebo> nobu uses lots of C and we use lots of Java
<enebo> fortunately for us Java and Ruby are very similar
<enebo> so multiple nested assignments with increments in an if statement is not so unusual in C (at least it seems that way)
<subbu> i remember seeing an if condition in gcc (when i was hacking on it) with conditionals that filled a full screen (or was it half a screen .. not that it makes it any better :)).
<enebo> I used to write C and I remember doing similar stuff but I think I have changed since I no longer uses languages which seems to encourage it
swills has quit [Remote host closed the connection]
<enebo> subbu: it makes it much better
<subbu> i was shocked seeing that.
<subbu> :)
<enebo> mix in some un-well-formed macros...FANTASTIC!
bbrowning_away is now known as bbrowning
<headius> hmmm
<headius> simple fix doesn't seem to do it here
swills has joined #jruby
<Aethenelle> I use tap all the time... great for debugging and logging. Also good is you need to change something before it gets used...
<headius> Aethenelle: yeah I was just thrown by the lack of any block arg
<Aethenelle> ? I miss something in 2.4?
<Aethenelle> ooo... try was finally addded in 2.3 ...
<Aethenelle> It's been a while since I looked at the new stuff...
<enebo> Aethenelle: I use tap all the time too
<Aethenelle> tap is good for you... just not for the people that have to read your code later...
<Aethenelle> fortunately, most of my code is run once...
<enebo> Aethenelle: oh I think it is really good for stuff like def gar; a = foo; a.bar(true); a; end
<enebo> Aethenelle: I even splurge and use do/end on full lines
<Aethenelle> I've been known to chain things off of end...
<enebo> heh
<enebo> no one judges on this channel
<enebo> except lopex
<enebo> but then he only does if it involves numbers
<lopex> judges what ?
<enebo> numbers
rdubya has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
bbrowning is now known as bbrowning_away
<GitHub193> [jruby] enebo pushed 2 new commits to symbol_love: https://git.io/vQpNa
<GitHub193> jruby/symbol_love 570eac5 Thomas E. Enebo: Nice thing we test AOT so we can see these. Missed reporting number of symbols...
<GitHub193> jruby/symbol_love d500788 Thomas E. Enebo: Merge branch 'ruby-2.4' into symbol_love
<headius> enebo: how's it going?
<enebo> going for Symbol push now
<enebo> so far only issues is some things are impossible to deprecate
<enebo> symbol should not be super hard but I expect it to knock on to m17n issues with method lookup
<enebo> but I will see I guess
<enebo> just realized it is really weird to eagerly get a RubySymbol and then make a SymbolNode
<enebo> Unfortunately I think the result will be that symbols are used in enough places we still need to make a node to wrap the value but I am half wondering if it is always needed
<enebo> :foo is a complete program and it will be in a block (AST) so it must be a node
<enebo> It sometimes makes me wish AST was all interfaces
<enebo> I have no idea how well that would work though
<GitHub110> [jruby] headius pushed 4 new commits to ruby-2.4: https://git.io/vQpxd
<GitHub110> jruby/ruby-2.4 40aca14 Charles Oliver Nutter: Guard step size against overflow into Bignum.
<GitHub110> jruby/ruby-2.4 0426982 Charles Oliver Nutter: Match MRI numeric remainder/positive/negative logic.
<GitHub110> jruby/ruby-2.4 0de8ebb Charles Oliver Nutter: If tap block has been procified, call that to get lambda logic....
<headius> enebo: if we didn't interpret it, interfaces would probably have been fine
<headius> ok so the enum thing is fixed and MRI test_numeric is 100% also
<enebo> yeah I guess the real issue woudl still be we want positioning for debugging
<enebo> It would still be possible but it would require something external to the nodes
<enebo> which may be better too I guess
<headius> ah yeah
<headius> positions could all be in a weak map somewhere :-)
<enebo> funny I realize we make LiteralNodes and we copy them in IR
<headius> a weak identity map
<enebo> which is not needed
<enebo> then realize that only alias and undef use them
<enebo> I guess it will kill a few more temps but very very few :)
<headius> heheh