10:50 <kares> enebo: great - good job!

10:51 <kares> enebo: one annoying thing really - which we already fixed once but than went wrong in the 50/51 release

10:51 <kares> https://github.com/jruby/activerecord-jdbc-adapter/issues/898

12:21 <jmalves> Hey, I created this PR https://github.com/jruby/jruby/pull/5262 that is broken. Should I create an issue for the underlying problem?

12:23 <jmalves> Not sure how far off my PR suggestion is, but might require some totally different approach. The parser is tricky to understand :X

14:06 <enebo> kares: oh yeah shoot...that will be a total rev of all releases

14:12 <enebo> jmalves: sorry I will try and look at that today. Your approach almost could have worked but $1 is ' $$ = Long.valueOf(lexer.getCmdArgumentState().getStack());'

14:13 <enebo> jmalves: I think I have an idea here

14:25 <jmalves> enebo: The idea behind the PR was that $-1 would be either `{` or `do`.

14:26 <enebo> jmalves: oh you know I missed it was -1

14:27 <enebo> jmalves: but I think 'jay' is not perfect implementation of yacc so I am unclear why there was a NPE

14:27 <jmalves> enebo: I did not get also. I guess I just don't know enough about this, long time since uni.

14:28 <enebo> jmalves: yeah and interestingly there has never been use of negative offsets in Ruby parser

14:28 <jmalves> enebo: also -1 is feels ugly. Your solution looks clear

14:28 <jmalves> enebo: Thx for looking at this!

14:28 <enebo> but MRI could definitely have used it since they use bison

14:28 <enebo> jmalves: sorry it took so long

14:29 <enebo> jmalves: your irc message made me realize it was forgotten about (I have been working on ancilliary projects more this summer)

14:30 <jmalves> enebo: No problem, I was also on vacations. This was not anything critical anyway.

14:31 <enebo> jmalves: yeah some day I look forward to no one finding more positioning issues in our AST

14:32 <jmalves> enebo: when looking at this I realised how tricky it must be to keep all this position information correct. Had no idea!

14:33 <enebo> jmalves: yeah and MRI has their own correction method we do not use and I don't remember why but I don't think we can use the same technique

18:27 <enebo> kares: I merged 50 -> 51 -> 52 -> master and it picked up some changes which I suspect are good (like updated jdbc version, money support in pg).