Test suite assertion failures on NetBSD 3.0

[Andrew Black <ab...@roguewave.com>]

Greetings all.

I think this is the final note I have on the results for running the 
stdcxx test suite after building it on NetBSD 3.0.  The focus of this 
note is assertion failures that were observed.  When I looked at these 
assertions, I treated any assertions that also occurred on a SuSE 9 SP 2 
build as known and accepted.  These known assertions were 2 assertions 
plus a sefgault in 23.deque.special, a library assertion and segfault in 
25.partitions, and an assertion in 25.random.shuffle.

The new/unexpected assertions were 4 assertions in 25.libc that I 
suspect are cause by deficencies in the underlyting libc implementation, 
2 assertions in 26.c.math that appear to be caused by rounding problems 
on visual inspetion, and 2 assertions in 27.stringbuf.virtuals of an 
unknown origin.  It is my suspicion that the failures in 
27.stringbuf.virtuals are related to some of the other platform 
weirdness with input/output.  Attached is the run output for these tests.

--Andrew Black