Xtensa: skip g++.pt/static11 test for xtensa-*-elf* targets

[Bob Wilson]

Hans-Peter Nilsson wrote:
> On 12 Jun 2002, Alexandre Oliva wrote:
> 
>>On Jun 11, 2002, Bob Wilson <bwilson@tensilica.com> wrote:
>>
>>
>>>The expected exit status in this test comes from a destructor calling
>>>_exit. With the dejagnu wrappers, however, the status that gets
>>>printed is the argument to exit -- before the destructor is even run.
>>
>>I hear this is fixed in some newer version of dejagnu, that wraps
>>_exit instead of exit.
> 
> 
> It wraps both, not *instead* of.

That's right.  The status that gets printed comes from the first call to either 
exit or _exit.  So for this testcase, the status is from exit being called at 
the end of main, not from _exit called from the destructor as expected by the test.

(I had been ignoring this failure for a long time because I had an old dejagnu 
snapshot.  I just grabbed the latest version, including some patches from the 
CVS tree at sources.redhat.com, and I still get the same failure.)

--Bob