ur5us has joined #jruby
nirvdrum has quit [Ping timeout: 268 seconds]
nirvdrum has joined #jruby
ur5us has quit [Ping timeout: 240 seconds]
ur5us has joined #jruby
nirvdrum has quit [Ping timeout: 258 seconds]
henry_bone has joined #jruby
<henry_bone> I find myself in need of websocket support for my rails app.
<henry_bone> I am somewhat tied to deploying my app as a war file
<henry_bone> and so I have been exploring the jruby-rack code to see if I can implement rack-hijack in order to have actioncable (rails websockets) available to my app
<henry_bone> My questions are about jruby-rack and rack-hijack. Is there a preferred design/approach to adding rack-hijack support to jruby-rack?
<henry_bone> My understanding of Rails' ActionCable is that the webserver needs to support rack-hijack, and my current understanding is that jruby-rack could be modified to implement it and thus allow ActionCable connections to work through Jetty
<henry_bone> If my assumptions are incorrect re: rack-hijack, then please let me know.
<henry_bone> Is there an existing implementation, or method for getting ActionCable to work for a Rails app served with embedded Jetty?
<henry_bone> If not, then I'm happy to have a crack at it, so if there's any guidance/hints regarding the implementation I'd love to hear them.
ur5us has quit [Ping timeout: 240 seconds]
jeff_ has joined #jruby
jeff_ is now known as Guest62232
henry_bone has quit [Quit: henry_bone]
Guest62232 is now known as henry_bone
henry_bone has left #jruby ["Good Bye"]
henry_bone has joined #jruby
henry_bone has quit [Client Quit]
nirvdrum has joined #jruby
nirvdrum has quit [Ping timeout: 256 seconds]
ChrisBr has quit [*.net *.split]
rebelwarrior[m] has quit [*.net *.split]
slackfan[m] has quit [*.net *.split]
klobuczek[m] has quit [*.net *.split]
RomainManni-Buca has quit [*.net *.split]
aemadrid[m] has quit [*.net *.split]
headius[m] has quit [*.net *.split]
lopex has quit [*.net *.split]
justinmcp_ has quit [*.net *.split]
ang-st has quit [*.net *.split]
knu has quit [*.net *.split]
Osho has quit [*.net *.split]
Caerus has quit [*.net *.split]
adam12 has quit [*.net *.split]
claudiuinberlin[ has quit [*.net *.split]
JulesIvanicGitte has quit [*.net *.split]
KarolBucekGitter has quit [*.net *.split]
xardion[m] has quit [*.net *.split]
kares[m] has quit [*.net *.split]
BlaneDabneyGitte has quit [*.net *.split]
enebo[m] has quit [*.net *.split]
TimGitter[m]1 has quit [*.net *.split]
ebarrett has quit [*.net *.split]
_whitelogger has joined #jruby
nirvdrum has quit [Ping timeout: 256 seconds]
lopex[m] has joined #jruby
rtyler1 has joined #jruby
drbobbeaty has joined #jruby
rdubya[m] has joined #jruby
TimGitter[m] has joined #jruby
cshupp[m] has joined #jruby
annette[m] has joined #jruby
MattPattersonGit has joined #jruby
ThomasEEneboGitt has joined #jruby
rwilliams[m] has joined #jruby
nirvdrum has joined #jruby
postmodern has quit [Quit: Leaving]
shellac has joined #jruby
nirvdrum has quit [Ping timeout: 256 seconds]
shellac has quit [Quit: Computer has gone to sleep.]
nirvdrum has joined #jruby
nirvdrum has quit [Remote host closed the connection]
shellac has joined #jruby
nirvdrum has joined #jruby
<headius[m]> good morning! 🌅
lucasb has joined #jruby
<headius[m]> enebo: doesn't seem to be a ton of parser failures in language specs, but this is one of them
<headius[m]> rescue in lambda literal
<headius[m]> Ahh this also appears to be one because it's parsing this as a method call rather than a constant access:
<headius[m]> so out of the four language spec failures on ruby-2.6 branch three are parser
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:7cd084d by Charles Oliver Nutter): The build is still failing. https://travis-ci.org/jruby/jruby/builds/658701572 [160 min 46 sec]
travis-ci has left #jruby [#jruby]
<enebo[m]> headius: ok. I thought I got em all but I will get those fixed today
<headius[m]> I have not looked over core specs for parser issues yet
<headius[m]> so this is just a statement about the language specs
<headius[m]> I'm going to peck at the spec failures for a bit today and try to get closer to green
<enebo[m]> LOL
<enebo[m]> ruby_version_is "2.5" do
<enebo[m]> So I likely did fix all 2.6 error
<headius[m]> that's one of the features
<headius[m]> I don't see the lambda change in NEWS
<enebo[m]> so I am guessing we missed this for 2.5 and when I did a complete diff from 2.5->2.6 grammar I did not see it as a diff
<enebo[m]> ok wrong spec...this is about lambda allowing do
<enebo[m]> -> do ... end
travis-ci has joined #jruby
<travis-ci> jruby/jruby (master:8c07f77 by Charles Oliver Nutter): The build was broken. https://travis-ci.org/jruby/jruby/builds/658723801 [171 min 19 sec]
travis-ci has left #jruby [#jruby]
<enebo[m]> stabby fixed
<enebo[m]> Character.isUpperCase(StringSupport.preciseCodePoint(getEncoding(), tempVal.unsafeBytes(), tempVal.begin(), tempVal.begin() + 1)
<enebo[m]> ok + 1 has to be the problem here for constant
<enebo[m]> result = tCONSTANT;/* assume provisionally */
<enebo[m]> fun stuff...
<enebo[m]> Sort of looks like Foo! will end up as a method name
<enebo[m]> I updated rvm so the suspense will be longer
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:06e5377 by Charles Oliver Nutter): The build is still failing. https://travis-ci.org/jruby/jruby/builds/658728678 [186 min 19 sec]
travis-ci has left #jruby [#jruby]
<headius[m]> hah ok
<headius[m]> yeah +1 because it's assuming first char is always ASCII
<enebo[m]> and I believe it did until 2.7 on MRI too
<enebo[m]> this code is all rearranged to assume all identifiers are constants and to set back to identifier later
<enebo[m]> ruby -e 'Foo!'
<enebo[m]> lol
<enebo[m]> So I suppose that makes sense?
<enebo[m]> hmm it always did this
<enebo[m]> oh I see how it works for us
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:babbd92 by Thomas E. Enebo): The build has errored. https://travis-ci.org/jruby/jruby/builds/658731863 [160 min 16 sec]
travis-ci has left #jruby [#jruby]
<enebo[m]> Interesting they track mbc on identifiers as they build up the potential identifier but then skip keyword search on it if it is not 7bit
<enebo[m]> nirvdrum: ^
<enebo[m]> I could see this preventing the search but at the cost of calling !ISASCII on every byte read
<enebo[m]> most identifiers tend to be 7bit so as an optimization it seems like it would not work
<nirvdrum> enebo[m]: I think I missed some of the context here. It sounds like you're saying the identifier rules changed in the MRI parser. Is that right?
<enebo[m]> nirvdrum: I was changing ident code because constant values were relaxed
<enebo[m]> nirvdrum: or the check allows more constant values than the past
<enebo[m]> nirvdrum: but I noticed they ISASCII all bytes in an ident (and I thought of you) but then noticed they literally only use it to short circuit and not do keyword lookup
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:410d52b by Charles Oliver Nutter): The build failed. https://travis-ci.org/jruby/jruby/builds/658749148 [149 min 10 sec]
travis-ci has left #jruby [#jruby]
<nirvdrum> Interesting. I haven't looked at the parser in a while. But, both JRuby and TruffleRuby already track the code range of identifiers, I believe. So, I don't know how much is to be gained here.
<nirvdrum> Maybe I have that wrong. I should fire up IntelliJ.
<enebo[m]> I do not think we do for idents
<enebo[m]> we do for strings and regexps
nirvdrum has quit [Ping timeout: 260 seconds]
<enebo[m]> headius: constant capital thing fixed...may revisit layer but you said 3 syntax errors what was the third?
<headius[m]> ok if you have parser stuff fixed I got the other one done so 2.6 language specs should be 👍️
<headius[m]> there were two for lambda rescue
<enebo[m]> headius: ah ok those are fixed then
<headius[m]> progress!
<headius[m]> I'm working my way down NEWS checklist now
<enebo[m]> SyntaxError: /home/travis/build/jruby/jruby/lib/ruby/stdlib/bigdecimal/ludcmp.rb:40: syntax error, unexpected '='
<enebo[m]> if pividx!=k then
<headius[m]> wat
<enebo[m]> hmm ok need to figure this out
<headius[m]> oh it's treating it as pividx! =
<enebo[m]> I was just looking at my CI run so this is definitely fallout from that
travis-ci has joined #jruby
travis-ci has left #jruby [#jruby]
<travis-ci> jruby/jruby (ruby-2.6:b4cc0c0 by Thomas E. Enebo): The build is still failing. https://travis-ci.org/jruby/jruby/builds/658769221 [126 min 3 sec]
<enebo[m]> it should see = and not do that
<enebo[m]> I am missing something
<enebo[m]> heh I think I see it!
shellac has quit [Quit: Computer has gone to sleep.]
<headius[m]> do you see what I see
<headius[m]> jk I see nothing
<headius[m]> there's one failing ArithmeticSequence spec that made me aware of rb_undef_alloc_func
<enebo[m]> I fixed it...I need some parens to group an ||
<headius[m]> it sets the allocator to something wacky that raises TypeError instead of just undefining #allocate
<headius[m]> used for a bunch of types and apparently nobody ever tested this
<headius[m]> except specs for ArithmeticSequence do
<enebo[m]> heh wackiness
<enebo[m]> is that a global function?
<enebo[m]> and in Ruby?
<headius[m]> yeah
<headius[m]> since 1.8
<headius[m]> .0
<enebo[m]> geez...
<headius[m]> seems to set allocator to -1 and I assume somewhere else it checks that and raises TypeError
<headius[m]> I'm just going to set it to a lambda that raises TypeError
<headius[m]> this is apparently how they undefine alloc for Nil, True, False, and a bunch of others
<headius[m]> I believe we just undefine the method itself
<enebo[m]> historical oddity I guess
<headius[m]> yah Ruby archaeology
<headius[m]> I suppose I should support reverse backtraces one of these days
<headius[m]> I still don't like them
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:9ae6dc8 by Thomas E. Enebo): The build is still failing. https://travis-ci.org/jruby/jruby/builds/658778708 [152 min 6 sec]
travis-ci has left #jruby [#jruby]
<enebo[m]> headius: I notice no ruby/stdlib/prime.rb. Do you know what stdlib is here?
subbu is now known as subbu|away
<enebo[m]> looking at endless range since they are the most annoying error messages atm
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:b344999 by Charles Oliver Nutter): The build is still failing. https://travis-ci.org/jruby/jruby/builds/658784887 [150 min 15 sec]
travis-ci has left #jruby [#jruby]
subbu|away is now known as subbu
nirvdrum has joined #jruby
<enebo[m]> kares: d7c198151aedf00c5853e420b3149097292311a3 causes some infinite loops
<enebo[m]> kares: jruby --dev -e '("a"..."e").step { |x| p x }'
<enebo[m]> I can easily see why but I am not clear on what should happen in call with IRubyObject[]
<enebo[m]> I think some consumers work because it is a RangeCallBack but in others it still just does .call which is what sets it off the rails
<enebo[m]> maybe I just doCall in main call method?
<enebo[m]> kares: pushed this as a fix: 6fe36b95fb93f788bc77a649354fb4f738ae8c3c
<enebo[m]> kares: not sure if I undid something which was preventing boxing or not
rusk has quit [Remote host closed the connection]
travis-ci has joined #jruby
travis-ci has left #jruby [#jruby]
<travis-ci> jruby/jruby (ruby-2.6:6fe36b9 by Thomas E. Enebo): The build is still failing. https://travis-ci.org/jruby/jruby/builds/658799472 [149 min 31 sec]
<headius[m]> enebo: we could turn off verbose output for most of these suites to quiet down CI output
<headius[m]> I turned those on some years ago because we had a lot of flaky specs that hung occasionally but we've addressed that since
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:aa36a70 by Charles Oliver Nutter): The build is still failing. https://travis-ci.org/jruby/jruby/builds/658802667 [150 min 44 sec]
travis-ci has left #jruby [#jruby]
<headius[m]> hey specs under 300 F/E
<headius[m]> 89 failures, 174 errors
<headius[m]> no problem
<headius[m]> lopex: do we have access to unicode version as a constant somewhere?
<headius[m]> RbConfig::CONFIG['UNICODE_VERSION'] is 12.1.0 for Ruby 2.6.3+ and Ruby 2.7 FAILED
<headius[m]> Expected nil == "12.1.0"
<headius[m]> to be truthy but was false
<headius[m]> easy couple specs to make green
<headius[m]> library/rbconfig/unicode_version_spec.rb and library/rbconfig/unicode_emoji_version_spec.rb
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:ca37a39 by Thomas E. Enebo): The build has errored. https://travis-ci.org/jruby/jruby/builds/658806277 [204 min 42 sec]
travis-ci has left #jruby [#jruby]
<headius[m]> huh
<headius[m]> Ruby 2.6 ships the prime stdlib as a default gem but the gem doesn't appear to have been released
<headius[m]> so that's like a dozen failures in specs
<headius[m]> hmm
<headius[m]> I wonder if mvn clean for lib/ should remove default gem stuff
<headius[m]> maybe more hassle than it's worth
Antiarc has quit [Quit: ZNC 1.7.4+deb7 - https://znc.in]
<headius[m]> sweet, backports had Method and Proc #<< and #>> so that's done
Antiarc has joined #jruby
<lopex> headius[m]: oh I forgot https://github.com/jruby/jcodings/issues/29
<headius[m]> oh there we go
<lopex> we need to update script generator
<headius[m]> seems reasonable
travis-ci has joined #jruby
travis-ci has left #jruby [#jruby]
<travis-ci> jruby/jruby (ruby-2.6:3174664 by Charles Oliver Nutter): The build has errored. https://travis-ci.org/jruby/jruby/builds/658828729 [203 min 42 sec]
henry_bone has joined #jruby
ur5us has joined #jruby
<headius[m]> work it work it
<headius[m]> enebo: I had an idea about prepping release notes: start a draft release long before we actually release
<enebo[m]> headius: yeah I thought we even talked about that
<headius[m]> enebo: maybe? I remember talking about an open issue or wiki page
<enebo[m]> as we did anything significant we would add an entry more or less
<headius[m]> but I mean an actual Github release we edit until it's time to tag
<headius[m]> now I suppose I should see if that's possible
<headius[m]> ah yeah
<enebo[m]> yeah I guess. Whereever it is possible to edit a page
<headius[m]> it won't tag until publish
<headius[m]> you should be able to see 9.3 at https://github.com/jruby/jruby/releases
<enebo[m]> yeah I have no idea if you can do that or not on github
<headius[m]> as a draft
<enebo[m]> ah a draft ok
<enebo[m]> sure this is good
<headius[m]> confirmed it does not show up for plebs
<enebo[m]> so you need to be project signin to see it
<headius[m]> hmmm make more coffee or not
<headius[m]> yeah
<headius[m]> I wish the CBD place next door had a barista
<headius[m]> lopex: ohhh yeah
<headius[m]> that explains why other non-allocatable types were passing
<headius[m]> 15 years is too long to work on one project
<lopex> author headius, year 2007
<headius[m]> ok 13 years
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:97f40b4 by Charles Oliver Nutter): The build has errored. https://travis-ci.org/jruby/jruby/builds/658879840 [202 min 16 sec]
travis-ci has left #jruby [#jruby]
henry_bone has quit [Quit: henry_bone]
<headius[m]> lopex: it is!
<headius[m]> I love deleting code
<headius[m]> the answer is no more coffee
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:480eb69 by Charles Oliver Nutter): The build has errored. https://travis-ci.org/jruby/jruby/builds/658885199 [205 min 2 sec]
travis-ci has left #jruby [#jruby]
<lopex> o well we do scan for defines anyways
<headius[m]> changed formatting?
<lopex> no just that it's a #define
<headius[m]> isn't there a good library for programmatically parsing C files?
<lopex> ooh we do `cpp #{name2ctype_h}...` also
<lopex> I just cpp on c files
<headius[m]> ah
<lopex> cpp #{name2ctype_h} -DUSE_UNICODE_PROPERTIES -DUSE_UNICODE_AGE_PROPERTIES | grep "^[^#;]"
<lopex> bleh
<lopex> well, it works
<lopex> I hope that perfect hash generated code doesnt change often
<headius[m]> haha
<headius[m]> yeah
<lopex> since we reverse all of that
<lopex> and combining it with binary reads wasnt easy
<headius[m]> I suppose eregon filed that because they also use jcodings... they must be hardcoding it for now
<lopex> for now I think the way the generator works is the best one, since mri might change their own commited official unicode defs
<lopex> and we skip all the ruby generation logic
<lopex> that reads that and generated c code
<lopex> *generates
<lopex> the other option would be to provide our own ERB templates
<lopex> which have a lot of intertwined erb/c logic
<lopex> :/
sagax has quit [Read error: Connection reset by peer]
<headius[m]> yeah that sounds pretty icky
<headius[m]> as long as this stil works it's only a once yearly hassle to keep it working
<lopex> btw an interesting look at unicode http://www.lubutu.com/soso/write-out-unicode-in-octal
<lopex> er, utf8 rather
<headius[m]> ok range checkboxes are done...seems to be some changes in first, step, and bsearch though
<headius[m]> ahh infinite range support for bsearch
lucasb has quit [Quit: Connection closed for inactivity]
<headius[m]> ok I think I'm calling it a day
<headius[m]> there's a handful of items from NEWS and plenty of specs to look at but good progress
<headius[m]> huh there's a ton of stdlib specs failing, I wonder why
<headius[m]> should be using the same gems
travis-ci has joined #jruby
<travis-ci> jruby/jruby (ruby-2.6:21358d3 by Charles Oliver Nutter): The build has errored. https://travis-ci.org/jruby/jruby/builds/658924072 [203 min 29 sec]
travis-ci has left #jruby [#jruby]
nirvdrum has quit [Ping timeout: 258 seconds]
henry_bone has joined #jruby