Re: XFAILing V3 tests.

[Gabriel Dos Reis <gdr at integrable-solutions dot net>]

Benjamin Kosnik <bkoz at redhat dot com> writes:

| >Maybe we should find DejaGnu incantations to run tests involving
| >locales only if the locales we need are present. 
| 
| Of course, any help at teaching dejagnu to understand C++ exceptions
| directly would be most welcome. I can't figure that out.

I'll try to see whether this is a case where the hook ${tool}-dg-prune
would prove helpful. 

| I think you'll find that this is what Loren's patch does, abet in a
| different manner: basically, if __newlocale returns NULL, then an
| exception is thrown (locale bits), then in the testsuite this exception
| caught, and should be marked as pass since we know that the test
| environment can't deal with this kind of test. This is why the
| BSD/AIX/HPUX results have returned to normal (or more normal: there were
| previously covered up issues that are now exposed.)

Thanks for these info.

| I'm pretty sure Mark is seeing a different kind of failure than this.
| Perhaps he could provide some more info?

Mark?

-- Gaby