Ok, since we got 3 different confirmations that things work on OS X, I'm going to restart the release vote process.
alex On Sun, Oct 4, 2009 at 7:31 PM, Rhett Sutphin <[email protected]>wrote: > Hi, > > On Oct 4, 2009, at 6:55 PM, Assaf Arkin wrote: > > On Sun, Oct 4, 2009 at 4:18 PM, Alex Boisvert <[email protected] >> >wrote: >> >> On Sun, Oct 4, 2009 at 3:49 PM, Rhett Sutphin <[email protected] >>> >>>> wrote: >>>> >>> >>> I'm seeing 5 failures on OS X 10.6.1 in "Hash ::from_java_properties". >>>> >>> All >>> >>>> the specs fail on the same line with the same error: >>>> >>>> can't coerce to char array >>>> /Users/rsutphin/buildr/svn/trunk/lib/buildr/core/util.rb:263:in >>>> `method_missing' >>>> /Users/rsutphin/buildr/svn/trunk/lib/buildr/core/util.rb:263:in >>>> `from_java_properties' >>>> ./spec/core/common_spec.rb:606: >>>> >>>> >>> Ok, thank for reporting. I think I'll let the other committers with OS >>> X >>> handle this one... wouldn't be very productive to debug this via email. >>> >>> >> Fixed and tested on JRuby 1.3.1 and Ruby 1.9.1/OS X 1.6.1. Need to double >> check against Ruby 1.8.x on a different platform. And with that, I >> believe >> we're done with OS X issues. >> > > Verified with stock MRI 1.8.7 on OS X 10.6.1 and with stock MRI 1.8.6 on OS > X 10.5.8. No other specs failing. > > Rhett > > > >> Assaf >> >> >> >> >>> It is possible there are other failures; it looks like the spec run is >>> >>>> aborting as it starts "Buildr::Packaging sources". >>>> >>>> >>> I've seen this on occasion on my system and I've determine it's due to >>> other >>> in which specs are executed. They used to be run in >>> reverse-chronological >>> order (last modified) which means it was pretty much non-deterministic. >>> I've committed a change such that they are run in directory order. >>> >>> alex >>> >>> >
