On Wed, Jul 29, 2009 at 12:21 AM, Ola Bini<[email protected]> wrote: > Charles Oliver Nutter wrote: >> >> 2009/7/28 Marcin Mielżyński <[email protected]>: >> >>> >>> There's very few places where new 1.8.7 features differ from 1.9 ones >>> (notably individual String/Enumerable methods, or slight difference for >>> deep >>> hashing like Hash#hash which shouldn't be detectable in 1.8.7/1.9). We >>> have >>> almost everything (if not all) in place. >>> We'd have to go by http://svn.ruby-lang.org/repos/ruby/tags/v1_8_7/NEWS >>> and >>> annotate them also as 1.8.7 ones. The only question is what protocol to >>> use >>> for run time disambiguation (like runtime.is19() etc.. an Enum ?) since >>> some >>> features are reused in both 1.8.6/1.8.7 and others in both 1.8.7/1.9 >>> modes. >>> >> >> I think we'd want to just make a hard switch to 1.8.7 and ditch 1.8.6 >> completely. It's not feasible for us to maintain *three* different >> compatibility modes, and I don't think there's any remaining reason >> not to move to 1.8.7. Every time I've heard about Matz asking people >> to report any remaining incompatibilities between 1.8.6 and 1.8.7, >> nobody has been able to point out a single thing. >> >> I vote we go straight-up 1.8.7. >> >> > > +1
Yeah only one 1.8 version and only one 1.9 version. I think at least during the last talk Nick, Charles, and I had in MN and we thought 1.8.7 was the thing to do. The reasoning was: 1. Multiple "I can't move to 1.8.7" threads with not a single case why 1.8.7 breaks anyone's apps. I think the original problem was the first release of 1.8.7 was booched. They fixed those initial problems (and Rails worked after that) and all was well after that. 2. Distributions now all ship 1.8.7. We get a steady stream of bugs against 1.8.7 behavior because we still support 1.8.6. My only open question about moving at this point is rubyspec? Does it include 1.8.7 tests? -Tom -- blog: http://blog.enebo.com twitter: tom_enebo mail: [email protected] --------------------------------------------------------------------- To unsubscribe from this list, please visit: http://xircles.codehaus.org/manage_email
