Re: [Jruby-devel] [PATCHES] Working with edge rails

2006-07-07 Thread Thomas E Enebo
On Thu, 06 Jul 2006, Charles O Nutter defenestrated me: > >It's possible since I made that change that we've switched to the >usually-more-correct CRef class, which could easily make this change >unnecessary (and incorrect). We may want to do an examination across >the board to see

Re: [Jruby-devel] [PATCHES] Working with edge rails

2006-07-06 Thread Charles O Nutter
It's possible since I made that change that we've switched to the usually-more-correct CRef class, which could easily make this change unnecessary (and incorrect). We may want to do an examination across the board to see if the RubyClass stack can be eliminated completely...it seems to be mucking u

Re: [Jruby-devel] [PATCHES] Working with edge rails

2006-07-06 Thread Thomas E Enebo
I applied both patches in new codehaus subversion repository. The eval patch is a little scary, but I am all for code removal. The patch does not seem to break WEBRick, gems, or running a rails app. Unit tests work. Charlie originally put that in there but perhaps we made a change which now mak

Re: [Jruby-devel] [PATCHES] Working with edge rails

2006-07-03 Thread Thomas E Enebo
I will take a look at this soon. The eval fix makes me scratch my head a bit, but it is possible we do not need to push a class in that case anymore. -Tom On Sun, 02 Jul 2006, Nick Sieger defenestrated me: > >The new, rewritten routes implementation in edge Rails gives JRuby a >run fo

[Jruby-devel] [PATCHES] Working with edge rails

2006-07-02 Thread Nick Sieger
The new, rewritten routes implementation in edge Rails gives JRuby a run for its money.1.  It exposes a missing Regexp.union method in the current implementation.  Patch and tests attached for that.2.  It uses Object#instance_eval to create special ActionController::Routing::Route#generate and #rec