Re: 21.2-b36 and 'make check'
>>>> "Y" == Yoshiki Hayashi
<yoshiki(a)xemacs.org> writes:
Y> Martin Buchholz <martin(a)xemacs.org> writes:
> >>>>> "Steve" == Steve Youngs
<youngs(a)xemacs.org> writes:
>
Steve> | symbol-tests.el: 262 of 264 (99%) tests successful.
Steve> | Testing /usr/local/src/xemacs-21.2/tests/automated/syntax-tests.el...
>
> The regression suite is supposed to pass 100%, and it does for me.
Y> Hmm... Then I think I have to debug this myself.
Y> Testing /src/xemacs/tests/automated/hash-table-tests.el...
Y> hash-table-tests.el: 9864 of 9866 (99%) tests successful.
I tried to track this down on Steve's computer. Steve is using
amazingly aggressive optimization, ie.
pgcc-2.95.3 -O3 -fomit-frame-pointer -Wall -mathlon -mcpu=athlon
-march=athlon -malign-functions=4 -funroll-loops
-fexpensive-optimizations -malign-double -fschedule-insns2
-mwide-multiply
With that degree of optimization, I expect optimizer bugs from the
gcc/pgcc guys. We've had them before. Of course, it *could* be a
stray pointer bug in XEmacs, but it smells like a compiler bug.
If you just use -O3 like configure wants you to, the world will be
happy, and your xemacs will be fast enough.
Yoshiki, are you also using pgcc or athlon-specific optimizations?
Martin