Yep, everything works fine in 1.8 mode. I'm pretty sure that problem is related with object constructors because the "equal?" method is the same for 1.8 and 1.9 mode. Right now I'm working in the other problem, I've arrived to the jit layer, so I think I'm arriving to an end point because I don't have any idea on how jit works.
On Mon, Sep 28, 2009 at 11:26 PM, Vladimir Sizikov <[email protected]>wrote: > Hi David, folks, > > On Mon, Sep 28, 2009 at 6:00 PM, David Calavera > <[email protected]> wrote: > > Hi, > > > > I've seen some recurrent weird behaviour running specs with 1.9 mode that > I > > can't fix and perhaps you could give me some clues to resolve it. > > > > Some specs work fine in 1.8 mode, but, although the code is the same for > 1.9 > > mode, they fail in 1.9 mode. I've copied two of them in this gist: > > > > http://gist.github.com/195524 > > I looked into the following case: > > Array#replace replaces the elements with elements from other array > FAILED > Expected ["a", "b", "c"] > not to be identical to ["a", "b", "c"] > > Looks like there is a problem in JRuby so that Array.equal? works > incorrectly in 1.9 mode (at least, not like it works in MRI 1.9): > > # jruby --1.9 -ve "puts [1, 2, 3].equal?([1, 2, 3])" > jruby 1.4.0dev (ruby 1.9.2dev trunk 24787) (2009-09-28 b03c7b4) (Java > HotSpot(TM) Client VM 1.6.0_03) [i386-java] > true > > # jruby -ve "p [1, 2, 3].equal?([1, 2, 3])" > jruby 1.4.0dev (ruby 1.8.7 patchlevel 174) (2009-09-28 b03c7b4) (Java > HotSpot(TM) Client VM 1.6.0_03) [i386-java] > false > > # /opt/ruby19-dev/bin/ruby -ve "puts [1, 2, 3].equal?([1, 2, 3])" > ruby 1.9.2dev (2009-09-25) [i686-linux] > false > > # /opt/ruby18-dev/bin/ruby -ve "puts [1, 2, 3].equal?([1, 2, 3])" > ruby 1.8.8dev (2009-09-26) [i686-linux] > false > > As you can see, only JRuby in 1.9 mode returns true. So it seems that > rubyspecs did found a genuine issue in JRuby. > > Thanks, > --Vladimir > > --------------------------------------------------------------------- > To unsubscribe from this list, please visit: > > http://xircles.codehaus.org/manage_email > > > -- David Calavera http://www.thinkincode.net
