ahorek has joined #jruby
ahorek has quit [Client Quit]
akp has joined #jruby
akp has quit [Client Quit]
xardion has quit [Ping timeout: 264 seconds]
xardion has joined #jruby
claudiuinberlin has joined #jruby
ahorek has joined #jruby
ahorek has quit [Client Quit]
shellac has joined #jruby
shellac has quit [Quit: Computer has gone to sleep.]
ahorek has joined #jruby
drbobbeaty has joined #jruby
<kares> enebo: headius: my bouncer has gotten a bit behind and only delivered your messages later
<kares> anyway - haven't been putting big stuff on master
<kares> put some interesting pieces in PRs (as noted prev) - would be interesting to get your thoughts on those
chasm has joined #jruby
chasm has left #jruby [#jruby]
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
ahorek has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
rdubya has quit [Ping timeout: 245 seconds]
shellac has joined #jruby
drbobbeaty has joined #jruby
ahorek has joined #jruby
ahorek has quit [Client Quit]
Puffball has quit [Ping timeout: 252 seconds]
jrafanie has joined #jruby
rdubya has joined #jruby
jrafanie has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jrafanie has joined #jruby
jrafanie has quit [Client Quit]
rdubya has quit [Ping timeout: 276 seconds]
rdubya has joined #jruby
ahorek has joined #jruby
ahorek has quit [Client Quit]
Puffball has joined #jruby
Puffball has quit [Quit: Puffball]
shellac has quit [Quit: Computer has gone to sleep.]
shellac has joined #jruby
shellac has quit [Read error: Connection reset by peer]
jrafanie has joined #jruby
<travis-ci> kares/jruby (date-time_to-java:f4eb948 by kares): The build was fixed. (https://travis-ci.org/kares/jruby/builds/382512204)
jrafanie has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jrafanie has joined #jruby
jrafanie has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
jrafanie has joined #jruby
<headius> kares: yeah no problem
<kares> greetings
<kares> its mostly JI stuff
xardion has quit [Remote host closed the connection]
claudiuinberlin has quit [Quit: Textual IRC Client: www.textualapp.com]
xardion has joined #jruby
<headius> well things are looking pretty ok
<kares> headius: did you get a chance to review some of the recent PRs?
<kares> + I think I have also seen some older ones marked for 9.2
<headius> recent, I'll have a look
<GitHub165> [jruby] headius closed issue #5184: `true.dup` and `false.dup` blow up (Inconsistent with MRI) https://git.io/vhJ4v
jrafanie has quit [Quit: Textual IRC Client: www.textualapp.com]
<kares> recent 3 from me shall do
ahorek has joined #jruby
<GitHub158> [jruby] enebo pushed 1 new commit to master: https://git.io/vhTBC
<GitHub158> jruby/master daeb255 Thomas E. Enebo: Add FFI-based resolv for windows
<headius> kares: I reviewed
<headius> enebo should look at them too
<kares> thanks - looking
<kares> name suggestions for JRuby.subclasses_of would be simply JRuby.subclasses right?
<headius> heheh
<headius> probably
<headius> or all_subclasses or raw_subclasses or something?
<kares> whatever you prefer - have no opinions there, really wasn't sure about naming
<kares> well whether they're all is a side effect
<kares> used to be a boolean now its a all: keyword
<headius> ah right
<kares> JRuby.subclasses only return immediate ones not all
<kares> (like Class.subclasses does)
<GitHub109> [jruby] enebo pushed 1 new commit to master: https://git.io/vhTBH
<GitHub109> jruby/master da36c92 Thomas E. Enebo: Revert "resolv ffi"...
<headius> subclasses would be fine to me
<enebo> java_subclasses?
<kares> okay will change that
<enebo> not really following the discussion
<enebo> :)
<headius> enebo: they're ruby subclasses, but getting them via a java API
<headius> there's just no standard Class.subclasses, so Rails has one that conflicts with jruby/util's version
<kares> Class#subclasses replacement ... since that collides with Rails
<headius> so moving that to JRuby.subclasses_of or subclasses or whatever
<headius> descendents?
<headius> aligns with ancestors
<kares> headius: that is smt else :)
<headius> so this is just immediate subclasses
<enebo> ok I don't care much but generic names may collide in future
<kares> they actually do have that in Rails
<kares> descendants as well as subclasses
<enebo> descendants_of
<kares> enebo: but its now on the JRuby (internal) module
<enebo> I kid. I don't have an issue with the name
<headius> heh yeah, modules might gain subclasses as a method some day :-)
<kares> should be fine right?\
<kares> oh that
<headius> I think we'll be ok
<headius> JRuby.subclasses is fine with me
<headius> BOLD
<kares> descendants is bad since they use it already in Rails -> would get confusing
<headius> COURAGE
<enebo> kares: subclasses is also fine by me
<headius> it was subclasses for us before, and we're transplanting it to a better namespace
<headius> seems fine
<enebo> LAND IT
<kares> THX
<headius> enebo: one down
<kares> well that is another one for enebo :) https://github.com/jruby/jruby/pull/5180
<headius> people always like the "jruby runs rails" tweets
<headius> I feel like we don't do enough promo
<kares> interestingly, we haven't been using RubyString#length() our selves
<headius> yeah I noticed, only a couple internal uses
<kares> which actually surprised me -> so while its a breaking change on the Java API it shouldn't be bad
<headius> and RubySymbol called length19
<enebo> kares: my question is what performance tradeoffs exist here
<kares> + we only get a chance for that every major version :)
<enebo> Not converting seems nice but is it actually faster
<enebo> It will remove an alloc but we will be looping through m17n codepoint logic
<kares> why wouldn't it be?
<headius> test a loop getting a system property
<headius> wait no that's string
<headius> charsequence is the red-headed stepchild of java.lang
<kares> but you will avoid the common case such as having debug(str) around while having debug turned off
<enebo> kares: oh yeah I can see how it can help but I was wondering how it helps if you use it
<kares> in that case now we won't need to allocate since its never used
<enebo> we trade an alloc and an internal transcode in C++ for our Java code walking it via jcodings
<kares> honestly, haven't measured yet
<enebo> kares: so yeah if we never use the string it is great
<enebo> kares: It probably is also great if we do use it but I was wondering if we could observe a difference
<kares> if we're at the logging case I doubt we'll notice much there even as it gets printed in terms of degrading
<kares> simply cause most of the time logger sync upon a stream/file
<kares> but yeah can give it a spin with log4j2 (tomorrow)
<kares> or I will just invent a case for the loop and micro bench
<headius> nearly everyone does charseq.toString anyway
<enebo> We may put this in 9.2.1 I think we can still have a mild signature change like this but I would like to know trade offs if there are any
<enebo> headius: yeah that is a good point I guess
<enebo> I never use CharSequence
<enebo> Hmm
<headius> the CharSequnce interface is so poor
<headius> but if they're just passing them around it would save alloc for some time
<GitHub154> [jruby] ahorek closed pull request #5158: fix resolv on windows (master...resolvfix) https://git.io/vprJb
<enebo> yeah so if people specify charsequence as an API they gain a benefit here potentially by not having blind conversion
<enebo> If they call toString this is basically a noop on use over what we do today
<enebo> If they actually use charsequence then there may be a difference
<headius> we can also use a cached string to impl charAt if we want to improve that
<kares> hmm in that case I am not sure if its not to be dropped ... you want to put in a Java API change in a minor?
<enebo> So in third case I guess if this is so unlikely to actually be used then I will rescind wanting to see numbers of some kind
<enebo> kares: I bet 0% of our users will switch to .0
<kares> but still :)
<enebo> kares: but I talked myself into not caring about this optimization
<headius> the API change worries me less because 9.1+ (or earlier) still have strLength
<enebo> kares: you can land it
<enebo> headius: yeah
<headius> so there's a forward and backward-compat option
<GitHub170> [jruby] ahorek closed issue #4755: Implement get_dns_server_list https://git.io/v5sIs
<enebo> did anyone call length()?
<kares> nope
<enebo> ahorek: you see my patch? Perhaps you can review it but I optimistically committed it
<kares> not even AR-JDBC/JOSSL
<enebo> kares: you search github or just our projects?
<enebo> it may be a tough search :P
<headius> psych does not either
<enebo> I guess we need to consider how most native exts work with our APIs
<enebo> They probably work with RubyString directly via helps like cat/cat19 or break out as a bytelist
<headius> racc does not
<kares> checked manually some exts
<kares> outside of JRuby -- like 2 :) since search was tricky
<headius> concurrent ruby does not
<headius> yeah seems ok
<enebo> ok I think we can cave if fallout ends up being severe
<enebo> does not sound like we will have any
<headius> yeah worst case it reverts
<enebo> I guess native ext wants char length of the string they will be able to change it to something else
<headius> but we can PR any ext that fails
<enebo> we can just say strLength() will work on all versions
<enebo> right?
<headius> yes
<kares> yep
<enebo> yeah so a) tons of users we revert b) handful we tell them to update their gem
<kares> recall checking nokogiri's ext and psych
<enebo> having an alternative which is backwards compat makes this a bit easier to cope with
<headius> 6 years is what, JRuby 1.5?
<enebo> headius: long enough where an old ext is probably already broken with strings
<headius> yeah pre-encodings
<headius> they would probably just be calling size still
<enebo> yeah
<GitHub141> [jruby] kares pushed 1 new commit to deprecate-subclasses: https://git.io/vhTEk
<GitHub141> jruby/deprecate-subclasses 5f45b43 kares: rename JRuby.subclasses_of method to simply JRuby.subclasses
<headius> and I'm +1 on datetime change if those APIs changed were all new things
<headius> I wasn't sure
<kares> headius: great, thanks - that should make more sense as some users already complained on RubyTime's Java API being cumbersome
<kares> no public method changes there
<kares> so let's ship that one ... unless enebo can think of why not:
<enebo> kares: land it
<kares> enebo: which one?
<enebo> datetime and charsequence
<kares> okay
<enebo> actual liklihood of length() plus actual consumers of charsequence makes it seem reasonably low risk
<kares> subclasses after the rename should be also good, right?
<enebo> I just like to challenge changes over the guise of risk and benefit
<enebo> kares: yeah I agreed there too
<kares> yy much appreciated
<kares> a conservative view is always a good view point :)
<GitHub15> [jruby] kares closed pull request #5183: [ji] support Java time APIs with Time/Date/DateTime (master...date-time_to-java) https://git.io/vhJfK
<GitHub127> jruby/master b21bc4e kares: [ji] conversion to Java time types for Date/DateTime (similar to Time)
<GitHub127> jruby/master dbc867c kares: review Time#to_java and support Java time types + tune returns
<GitHub127> jruby/master fa0ab44 kares: normalize RubyTime Java API a bit + review JavaDoc on confusing parts
<GitHub127> [jruby] kares pushed 5 new commits to master: https://git.io/vhTE9
<GitHub81> [jruby] kares closed pull request #5180: [ji] RubyString implements CharSequence (master...rubystring-impl-charsequence) https://git.io/vpjgQ
<GitHub163> [jruby] kares pushed 5 new commits to master: https://git.io/vhTEA
<GitHub163> jruby/master 2b33311 kares: fix RubyString subSequence impl -> as gets start/end not beg/length
<GitHub163> jruby/master 9331814 kares: change RubyString to return itself seeing a CharSequence target type
<GitHub163> jruby/master 5110380 kares: [ji] RubyString implements CharSequence draft (based on GH-3937)
<GitHub177> [jruby] kares closed pull request #5181: deprecate jruby/core_ext Class#subclasses (master...deprecate-subclasses) https://git.io/vpj5q
<GitHub79> [jruby] kares pushed 1 new commit to master: https://git.io/vhTuq
<GitHub79> jruby/master f574f9d Karol Bucek: Merge pull request #5181 from deprecate-subclasses...
<GitHub165> [jruby] kares deleted deprecate-subclasses at 5f45b43: https://git.io/vhTum
<GitHub174> [jruby] kares closed issue #4760: deprecate Class#subclasses from jruby/core_ext.rb https://git.io/v5Gu5
joast has quit [Remote host closed the connection]
subbu is now known as subbu|away
<GitHub70> [jruby] greghuc opened issue #5185: Question: how to aggregate results from multiple profile calls (e.g. web requests) https://git.io/vhTwO
NightMonkey has quit [Quit: ZNC - http://znc.in]
NightMonkey has joined #jruby
joast has joined #jruby
ahorek has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
joast has quit [Quit: Leaving.]
<headius> woot
joast has joined #jruby
subbu|away is now known as subbu
subbu is now known as subbu|away
subbu|away is now known as subbu
<GitHub34> [jruby] headius pushed 1 new commit to master: https://git.io/vhTdP
<GitHub34> jruby/master d6b5623 Charles Oliver Nutter: Remove rogue bundler.rb. Leave it as a preinstalled gem.
<headius> heh the things we find immediately before release
<enebo> well it is fine we found this anyways because kares merged part of those PRs after I spun last thing
<enebo> I had thought I had waited long enough so I will rerun shit
<enebo> but hopefully for last time
ahorek has joined #jruby
ahorek has quit [Quit: http://www.kiwiirc.com/ - A hand crafted IRC client]
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
Liothen has quit [Changing host]
Liothen has joined #jruby
<GitHub157> [jruby] headius pushed 1 new commit to master: https://git.io/vhTj3
<GitHub157> jruby/master 4e1c9dd Charles Oliver Nutter: Limit generated reification to 10 since accessors etc are static....
<GitHub111> [jruby] headius opened issue #5186: Finish reification https://git.io/vhkem
<GitHub112> [jruby] headius opened pull request #5187: Non-volatile instance variables (master...nonvolatile_ivars) https://git.io/vhkeX
<GitHub176> [jruby] headius pushed 1 new commit to master: https://git.io/vhkeM
<GitHub176> jruby/master feb11a1 Charles Oliver Nutter: Force use of one of the VariableAccessor subclasses.
rdubya has quit [Ping timeout: 245 seconds]
<travis-ci> jruby/jruby (master:feb11a1 by Charles Oliver Nutter): The build was fixed. (https://travis-ci.org/jruby/jruby/builds/382927224)