swills has quit [Remote host closed the connection]
bga57 has joined #jruby
_whitelogger has joined #jruby
<GitHub146> [jruby-openssl] kares pushed 2 new commits to master: https://git.io/vFLxe
<GitHub146> jruby-openssl/master 89e9e18 kares: [travis-ci] setup some tests with JDK 9 + adjust allowe failures
<GitHub146> jruby-openssl/master 66b4da3 kares: [refactor] work-around on BN#to_s impl which has been due JRuby 1.6
damnski has quit [Ping timeout: 246 seconds]
<kares> headius: jossl seems good to go however that PKCS#12 failure we're seeing on JRuby's CI seems at least interesting
<kares> would like to find some time to investigate - but I am not sure how this week will go ... yet
olle has joined #jruby
<kares> also some folks reached out due AR-JDBC 5.x ...
<kares> maybe this one will go through and I will be able to make plans to work a week or two exclusively on finishing up 5 support, we'll see
olle has quit [Client Quit]
nirvdrum has quit [Ping timeout: 264 seconds]
scgdzhaszh has joined #jruby
scgdzhaszh has quit [Remote host closed the connection]
olle has joined #jruby
<GitHub116> [jruby-openssl] kares pushed 1 new commit to master: https://git.io/vFtex
<GitHub116> jruby-openssl/master 1ace795 kares: [travis-ci] let's try some more (integration) Java 9 testing
vtunka has joined #jruby
claudiuinberlin has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
vtunka has quit [Quit: Leaving]
vtunka has joined #jruby
damnski has joined #jruby
drbobbeaty has joined #jruby
drbobbeaty has quit [Ping timeout: 248 seconds]
lance|afk is now known as lanceball
vtunka has quit [Quit: Leaving]
<headius> kares: yeah I haven't started looking into the pkcs12 error yet
<headius> enebo knows status of AR-JDBC for 5, I think he was hoping to release something soon
<headius> ah-ha
<headius> no wonder that test just up and started failing
<GitHub7> [jruby] headius pushed 2 new commits to jruby-9.1: https://git.io/vFtEy
<GitHub7> jruby/jruby-9.1 f6fd6e7 Charles Oliver Nutter: Disable this test on Travis, which has no IP6 now....
<GitHub7> jruby/jruby-9.1 a1beea2 Charles Oliver Nutter: Sync this code up with master.
danielg__ has joined #jruby
vtunka has joined #jruby
swills has joined #jruby
swills has joined #jruby
<GitHub20> [jruby] headius closed issue #4805: can not require openssl with JDK9 https://git.io/vdBOn
bga57 has quit [Ping timeout: 255 seconds]
bga57 has joined #jruby
drbobbeaty has joined #jruby
vtunka has quit [Quit: Leaving]
<GitHub113> [jruby] headius pushed 1 new commit to master: https://git.io/vFtVz
<GitHub113> jruby/master 0b72b03 Charles Oliver Nutter: Merge branch 'jruby-9.1'
<GitHub41> [jruby] headius closed issue #4824: FFI not working on Java 9 MacOS https://git.io/vd5Cj
<GitHub30> [jruby] enebo pushed 1 new commit to master: https://git.io/vFtoL
<GitHub30> jruby/master e2669a6 Thomas E. Enebo: Basic constant assign specs for &&=
vtunka has joined #jruby
<GitHub188> [jruby] headius closed issue #4801: "gem install bundler" times out instead of installing normally https://git.io/vdnBp
shellac has joined #jruby
danielg__ has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
danielgl_ has joined #jruby
danielgl_ has quit [Client Quit]
danielgl_ has joined #jruby
danielgl_ has quit [Client Quit]
<olle> Hello! I'm trying real hard to learn about the "Config.USE_UNICODE_PROPERTIES" used in Joni.
<olle> Can I... enable that, in some way?
danielgl_ has joined #jruby
danielgl_ has quit [Client Quit]
danielgl_ has joined #jruby
danielgl_ has quit [Client Quit]
<olle> yokolet and lopex seems to be the two people on the Web who've discussed this, at all.
<olle> (Sorry, jcodings, would be where USE_UNICODE_PROPERTIES would be defined. And set to true)
danielgl_ has joined #jruby
danielgl_ has quit [Client Quit]
danielgl_ has joined #jruby
danielgl_ has quit [Client Quit]
<headius> olle: hello there
<headius> olle: is it not working?
<olle> It ain't working for me, at least. MRI Ruby 2.4.0 works, while JRuby 9.1.13.0 does not work.
<olle> I should offer a description of what my issue is.
<GitHub119> [jruby] headius closed issue #4789: Downloads from http://jruby.org/download fail. https://git.io/v574Y
<headius> heh yeah that doesn't look right
<olle> Swedish letters have a hard time being picked up as "full graphemes"
<headius> yeah I wonder if this is related to unicode updates MRI did when they forked oniguruma to "onigmo"
<headius> lopex: maybe?
<headius> the unicode_normalize code is also pretty dense if I remember right
<headius> olle: smallest repro possible would help
<headius> mmm unicode_normalize output *seems* ok
danielgl_ has joined #jruby
danielgl_ has quit [Client Quit]
<headius> olle: so I guess that's why you figured this is a regexp issue
<olle> headius: Trying to get to a composed character to share a unicode_normalize-less example. Thinking.
danielgl_ has joined #jruby
danielgl_ has quit [Client Quit]
<headius> oh huh
<headius> without u_n it scans properly
<headius> so maybe it IS a problem with u_n
<headius> so u_n would normalize to a standard form, probably the expanded form that uses composition
<headius> scan then might be failing to pick up composed characters
<headius> irb(main):005:0> 'åäöÅÄÖ'.unicode_normalize(:nfd).bytes
<headius> => [97, 204, 138, 97, 204, 136, 111, 204, 136, 65, 204, 138, 65, 204, 136, 79, 204, 136]
<headius> irb(main):006:0> 'åäöÅÄÖ'.bytes
<headius> => [195, 165, 195, 164, 195, 182, 195, 133, 195, 132, 195, 150]
<headius> 97 is 'a' so yeah, that appears to be what's happening
<headius> I'd wager u_n is fine here but scan is missing logic to pick up composed characters
<olle> Yes, u_n returns the same on MRI and JRuby.
<olle> But for some reason all the "modifiers" get dropped in the /\X/ regexp match.
<olle> (modifier, in this case: the circle above the "a")
<GitHub117> [jruby] enebo pushed 1 new commit to master: https://git.io/vFt9t
<GitHub117> jruby/master e701f62 Thomas E. Enebo: Fixes #4792. Unhandled Java exception in Ripper on inline comments....
<GitHub192> [jruby] enebo closed issue #4792: Unhandled Java exception in Ripper on inline comments https://git.io/v5heU
<GitHub83> [jruby] enebo pushed 1 new commit to jruby-9.1: https://git.io/vFt9s
<GitHub83> jruby/jruby-9.1 eb1151a Thomas E. Enebo: Fixes #4792. Unhandled Java exception in Ripper on inline comments....
<headius> olle: well you may be right that the unicode mode isn't working right
<headius> it does seem odd to me that \X could be this broken
<headius> olle: I'd start by comparing our scan logic with MRI's, but I'm leaning toward a joni issue too
<headius> [--dev] ~/projects/jruby $ /usr/bin/ruby -e "p 'åäöÅÄÖ'.unicode_normalize(:nfd).match /(\X)/"
<headius> #<MatchData "å" 1:"å">
<headius> [--dev] ~/projects/jruby $ jruby -e "p 'åäöÅÄÖ'.unicode_normalize(:nfd).match /(\X)/"
<headius> #<MatchData "a" 1:"a">
<headius> olle: if this doesn't have an issue already, open one
<olle> headius: a joni issue, right?
<headius> file it with JRuby for now
<headius> it's likely a joni issue but we'll file that when we know for sure, plus there will be testing work to do on JRuby side
<GitHub86> [jruby] olleolleolle opened issue #4832: Modifiers are dropped in \X regular expression matches https://git.io/vFt5A
<GitHub43> [jruby] headius pushed 1 new commit to jruby-9.1: https://git.io/vFtd5
<GitHub43> jruby/jruby-9.1 6e2088a Dennis Marinus: switch ShellLauncher to use getAbsolutePath
<olle> I hope the title of that is legible, it may be only me in the whole world who calls those compound codepoints "modifiers".
<GitHub196> [jruby] headius pushed 1 new commit to master: https://git.io/vFtdx
<GitHub196> jruby/master d2c3370 Charles Oliver Nutter: Merge branch 'jruby-9.1'
<GitHub16> [jruby] headius closed pull request #4821: switch ShellLauncher to use getAbsolutePath (master...canonical-to-absolute) https://git.io/vd9hW
<headius> olle: I can never remember the right terminology
<headius> "combining characters" I guess
<headius> reading about it now since I don't know how it's supposed to know to keep going after seeing 97
<GitHub77> [jruby] enebo pushed 1 new commit to jruby-9.1: https://git.io/vFtFK
<GitHub77> jruby/jruby-9.1 2d078e3 Thomas E. Enebo: Fixes #4787 Ripper fails to lexical analyse around escaped newline on JRuby....
<GitHub155> [jruby] enebo pushed 1 new commit to master: https://git.io/vFtFM
<GitHub155> jruby/master 3b6f308 Thomas E. Enebo: Fixes #4787 Ripper fails to lexical analyse around escaped newline on JRuby....
<GitHub198> [jruby] enebo closed issue #4787: Ripper fails to lexical analyse around escaped newline on JRuby https://git.io/v5i1d
<headius> I'll move my comments to the bug
vtunka has quit [Quit: Leaving]
enebo has joined #jruby
<lopex> olle: USE_UNICODE_PROPERTIES is on by default
<olle> lopex: Thanks for answering. It was there, in the jencodings library, set to a final field with "true".
<lopex> olle: yeah, that's recent changes wrt grapheme clusters in 2.4
<headius> lopex: ahh right, that's the grapheme stuff
<lopex> headius: yeah that \X expands for this http://unicode.org/reports/tr29/#Regex_Definitions
<headius> lopex: but on my system MRI 2.3 works properly too
<lopex> I've got how it works though
<lopex> hmm
<lopex> 2.3 is since ?
<headius> perhaps it was backported? my system is using 2.3.3
<lopex> the newer one is completely rewritten
<lopex> that's theire first commit related to that
<headius> heh yeah, bit of work
<headius> maybe we can get chrisseaton's team to do it :-D
<lopex> it's not that bad
<lopex> headius: 90% of that code is that regexp expansion by hand
<lopex> case map is worse :P
<headius> kares: have you looked into https://github.com/jruby/jruby/issues/4802 much lately?
olle has left #jruby [#jruby]
olle has joined #jruby
<headius> lopex: 2.2.7 also works properly so this must be older
<lopex> 2.2.7 is march 17
<headius> so they are backporting this
<lopex> "backporting" by crosscommiting
<lopex> they essentially cross-merge
olle has quit [Quit: olle]
<kares> headius: nope but I think it should go away with a JDK 8 update
<headius> kares: that's what I'm suspecting too
<headius> the debug output seems to show a GeoTrust cert issued by Equifax, which is causing the verification failure on Ubuntu
<kares> well, we can't control the built SSL engine completely ... do you happen to have latest Java 8?
<kares> should be 152 report was for 131 or so
<headius> well I confirmed the bug with the same latest jdk8 on ubuntu
<headius> yeah 131
<headius> I'm trying newer jdk on same ubuntu now
<kares> ok I wonder if it goes away if not its worth looking into deeper ...
<headius> hmmm
<headius> still fails on 151
<kares> ;( ... that lookup path of jossl is not the cleanest piece of logic although it worked so far
<kares> might take a look at some point later but I was planning on looking on that Java 9 'boot' PKCS#12 failure
<headius> well I'm stumped but I posted my findings
<headius> it seems like we're doing the right thing, verification wise, but somehow getting a GeoTrust cert that's issued by Equifax
<GitHub66> [jruby] headius closed issue #4794: conda-forge package for jruby https://git.io/vdfWH
subbu is now known as subbu|lunch
<GitHub48> [jruby] MrBerg opened pull request #4833: Added specs for IO#each(sep,limit) (master...io-each-separator-and-limit-specs) https://git.io/vFqZe
eregon_ has joined #jruby
<GitHub180> [jruby] headius pushed 3 new commits to master: https://git.io/vFqZw
<GitHub180> jruby/master 1d47d52 Jonas Berg: Make BigDecimal::limit work for subtraction and division. Fix issue #1615.
<GitHub180> jruby/master eba3bf5 Jonas Berg: Use instanceof
<GitHub180> jruby/master 850a14b Charles Oliver Nutter: Merge pull request #4830 from MrBerg/enforce-bigdecimal-limit...
<GitHub63> [jruby] headius closed pull request #4830: Make BigDecimal::limit work for subtraction and division. Fix issue #1615 (master...enforce-bigdecimal-limit) https://git.io/vFT0M
<eregon_> headius: Hi
<headius> eregon_: hello there!
<GitHub85> [jruby] headius pushed 3 new commits to jruby-9.1: https://git.io/vFqn1
<GitHub85> jruby/jruby-9.1 28eb942 Jonas Berg: Use instanceof
<GitHub85> jruby/jruby-9.1 f386054 kares: cleanup BigDecimal from the dual 1.8/1.9 Ruby support heritage...
<GitHub85> jruby/jruby-9.1 dc7b498 Jonas Berg: Make BigDecimal::limit work for subtraction and division. Fix issue #1615.
<eregon_> headius: Could you make a PR with the changes for Date? Then I should be able to take a look at the other bug
<headius> yeah sure
<headius> are you sure MRI/unix time is using GJ and not ISO?
<eregon_> Time is ISO
<eregon_> but Date is GJ
<headius> that seems to match what we have
<eregon_> However Time#to_date creates a Date with GJ I think
<eregon_> So if you did (time.to_date+1000 days).to_time might be different than time+1000days since they use different chronologies
<eregon_> DateTime#to_time magically using the local timezone is not helping this, 2.4 is much saner in this regard
<eregon_> There is no guarantee it round-trips in the bug report due to changing timezones
shellac has quit [Ping timeout: 240 seconds]
<GitHub26> [jruby] headius closed issue #1615: BigDecimal::limit has not effect https://git.io/vFqcB
<headius> yeah
<eregon_> Well, at least their DateTime#ajd match in this case
<headius> I guess what confuses me is that if I change date.rb to use ISO instead of GJ it matches MRI then
subbu|lunch is now known as subbu
<GitHub156> [jruby] headius pushed 1 new commit to master: https://git.io/vFqCm
<GitHub156> jruby/master 854d7eb Charles Oliver Nutter: Merge branch 'jruby-9.1'
<GitHub140> [jruby] headius pushed 1 new commit to jruby-9.1: https://git.io/vFqCY
<GitHub140> jruby/jruby-9.1 31f25a9 Charles Oliver Nutter: Remove line removed on master.
<headius> kares: you could cherrypick some of those other BigDecimal tweaks from master to 9.1
<GitHub18> [jruby] headius pushed 1 new commit to master: https://git.io/vFqWU
<GitHub18> jruby/master c3b71d5 Charles Oliver Nutter: Merge branch 'jruby-9.1'
<GitHub187> [jruby] enebo pushed 1 new commit to master: https://git.io/vFqWo
<GitHub187> jruby/master 9d7737c Thomas E. Enebo: Whoops...2.4 ripper seems to have changed since we made jruby-9.1? Trivial...
<GitHub127> [jruby] headius pushed 1 new commit to master: https://git.io/vFql7
<GitHub127> jruby/master 510e937 Charles Oliver Nutter: Replace Fixnum with Integer to eliminate warnings.
eregon_ has quit [Ping timeout: 240 seconds]
<GitHub33> [jruby] enebo pushed 1 new commit to jruby-9.1: https://git.io/vFqBb
<GitHub33> jruby/jruby-9.1 10d6654 Thomas E. Enebo: Fix errors from my &&= and ||= work. I should have defaulted all other const...
<GitHub73> [jruby] enebo pushed 1 new commit to master: https://git.io/vFqBx
<GitHub73> jruby/master 94484f7 Thomas E. Enebo: Fix errors from my &&= and ||= work. I should have defaulted all other const...
eregon_ has joined #jruby
Puffball has quit [Remote host closed the connection]
Puffball has joined #jruby
drbobbeaty has quit [Quit: My MacBook Pro has gone to sleep. ZZZzzz…]
claudiuinberlin has quit [Quit: My MacBook has gone to sleep. ZZZzzz…]
claudiuinberlin has joined #jruby
eregon_ has quit [Quit: Leaving]
claudiuinberlin has quit [Quit: Textual IRC Client: www.textualapp.com]