[MERGE] Improve tests for the behaviour of Tree.iter_changes for missing paths that are only present in one tree, and fix found bugs. (Robert Collins)

[Andrew Bennetts]

Robert Collins wrote:
> On Thu, 2008-08-14 at 09:36 +1000, Andrew Bennetts wrote:
[...]
> > If we are genuinely concerned about compatibility other test runners
> 
> ALL the patches I've made to bzrlibs test support have been made with
> other runners in mind. I dunno about you, but supporting a walled-garden
> of tests is not a good idea in my book.

I phrased that sentence badly.  I am concerned about it, and I keep other
runners in mind too.  I really just meant to say that this isn't very hard to do
compatibly.  The fact that the current implementation isn't compatible is not a
strong argument against having or using the TestNotApplicable feature.

> > , it is
> > possible to change our TestCase to detect TestResults that aren't from
> > bzrlib.tests and adapt our behaviour (or write an adapter for stdlib TestResult
> > instances).  Twisted's Trial manages to do this for it's SkipTest and other
> > related extensions (so e.g. SkipTest is recorded as a success if it can't be
> > recorded as a skip).
> 
> That would make it less objectionable for me. But still - this is not an
> *exceptional* circumstance; using exceptions for situation-normal flow
> control is pretty ugly IMO.

It feels sufficiently exceptional to me that it doesn't feel unnatural.  I guess
the boundary between exceptional and normal is a matter of taste.

-Andrew.