Minor change in test matrix

[Jeffrey A Law]

  In message < 37540CCB.73989BD4@acm.org >you write:
  > Is this due to lack of 2.6 testers? If so, I volunteer, as I happen to be
  > one of the few running 2.6.
More because solaris-2.6's user base seems a lot smaller than solaris-2.5.

Getting results for solaris-2.6 is still a good thing to do though.  One
approach might be to just get gcc-2.95 testresults and verify that they are
not significantly different than the solaris-2.5 gcc-2.95 test results instead
of going through a full regression cycle.
jeff