On Thu, Jan 05, 2012 at 06:15:36PM +0100, Nicolas Cellier wrote:
> 2012/1/5 David T. Lewis <lewis at mail.msen.com>:
> > CompiledMethodTest>>testPerformInSuperclassCanExecutelongMethodWithTemps fails
> > on the interpreter VM, but passes on Cog. The test was introduced to the image
> > in KernelTests-nice.180 with the update notice "Pick a few tests from Pharo".
> >
> > I am assuming that this test documents a preexisting limitation of the
> > interpreter VM that has been addressed in Cog. Is that right?
> >
> 
> Yes http://code.google.com/p/pharo/issues/detail?id=3468
> We could call this a "limitation" rather than a "bug" if you prefer
> and declare it as an expectedFailure (conditionaly vm-wise).

Thanks, that's exactly what I needed to know.

> 
> I think that it is good to document limitations, so it was good to
> introduce this test.
> To me, an undocumented limitation is a bug.
> 

+1

Dave