Your recent change

[Mike Stump]

> From: Mark Mitchell <mark@codesourcery.com>
> Date: Sun, 26 Mar 2000 09:58:50 -0800

> I plan on mandating that no tests should have unexpected failures,
> except when they have just broken.  That will make it easy for
> people to know what's a failure and what's not.

Gosh, you mean to try and use the testsuite the way we had designed
it?  :-)  I fully endorse the idea.  Though, in practice there may be a
few things that are hard to set as expected, due to the lack of
precision in expectedness setting.

For example, in the C testsuite I think it is hard to say this case at
-O3 fail on this list of hosts, but at -O2 they work....  (as I
recall).  So the choice then becomes, do you want unexpectedness on
the pass over unexpectedness on the fail, or do you want to improve
the testsuite driver?  The g++ testsuite doesn't suffer from this
problem as I recall.

Anyway, glad to hear it.