Results for gcc-2.95 19990524 (prerelease) testsuite on i386-unknown-openbsd2.5

[craig@jcb-sc.com]

>FAIL: g77.f-torture/compile/19990502-0.f,  -O2  

It's a mystery to me why this fails on just *that* particular i386 system.
Could someone investigate?  On other i386 systems, e.g. mine, it *used*
to fail, and the bug wasn't fixed, but was accidentally "papered over"
by my flag_errno_math patch (which had, as its only goal, the improved
performance of square-root when compiled by g77).

        tq vm, (burley)