Failures in the conform tests removed from arch-independent failure list for release.

["Carlos O'Donell" <carlos at redhat dot com>]

Joseph,

It was my expectation, an expectation completely unfounded in any
real or substantive data, that the conform test failures I was seeing
were merely the same failures I'd seen for years. It is true that with
this mindset I marked them as generic failures in the list of failures
for 2.20 because I see them on all arches that I test with.

Given that I have not investigated any further I've removed them from
the architecture-independent list and moved them back into all the
tests I've done.

I will endeavour to review them in 2.21 to determine why they are there
and how to fix them.

What is the most common cause of such a failure?

For example:
FAIL: conform/POSIX/semaphore.h/conform

Is caused by:
Testing <semaphore.h>
---------------------
  Checking whether <semaphore.h> is available... FAIL
    Header <semaphore.h> not available  Compiler message:
    -----------------------------------------------------------------------
    /home/carlos/build/glibc/conform/POSIX/semaphore.h/scratch/semaphore.h-test.c:1:23: fatal error: semaphore.h: No such file or directory
     #include <semaphore.h>
                           ^
    compilation terminated.
    -----------------------------------------------------------------------

  Checking the namespace of "semaphore.h"... SKIP


----------------------------------------------------------------------------
  Total number of tests   :    1
  Number of failed tests  :    1 (100%)
  Number of skipped tests :    1 (100%)

Which seems odd.

I don't see any easy way to reproduce exactly what the test was doing
without hacking the perl script, so this will have to wait until I have
more time.

Cheers,
Carlos.