[perl #23222] perl-5.8.1-RC4: test failures on Solaris 7 / Sparc

Some more info on this case (anybody out there with Solaris and
Forte 6.2 more familiar with the Perl guts than I am?)

* the same failures (signal handling!) happen with optimization (-O)
  switched off, and with debugging enabled (-g)
* these failures do NOT occur when compiled with gcc-3.3 and the
  rest of the environment unchanged

This makes me believe that Sun's Forte 6.2 compiler is the culprit -
but it compiled 5.8.0 with the same options ok. So there seems to
be a change near the signal handling code in Perl 5.8.1 that triggers
a bug in Forte, or the change itself relies on some property of gcc
that confuses Forte.

• [perl #23222] perl-5.8.1-RC4: test failures on Solaris 7 / Sparc by Marek Rouchal