This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: [PATCH]: Backport PR other/13906 for "fold" checking
> IMHO, most fold checking failures (like tree checking) would be
> platform independent.
Precisely. I think SPARC/Solaris is not the right platform to enable the
platform-independent fold checking because...
> So if you don't want to try it on solaris2, I don't think that's a big loss
> as long as long as somebody somewhere tries it out (which I just did).
...it is a big loss when the 2 only testers for a platform don't agree on the
testing results (and you test more regularly than me). The results you post
are essentially useless for me now, given that I cannot assert whether the
failures are real or not, since nobody else enables fold checking.
IMHO the numerous cycles burned by fold checking should be better used on
SPARC/Solaris, because they are a _very_ scarce resource nowadays. You could
for example enable multilib (-m32/-m64) testing or -fPIC testing again.
That's what you did a few years ago on your Solaris 7 box and that was quite
useful; thanks to it, -fPIC support is rock solid in the SPARC back-end.
> The 4.1 results look to be on par with mainline.
> http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00508.html
Look at the differences with mine:
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00545.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00544.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00543.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00542.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00541.html
> 4.0 is a little bit worse, but that may simply require indentifying
> and backporting the appropriate patch(es) from 4.1/mainline.
Here are mine:
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00482.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00481.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00480.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00479.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00478.html
http://gcc.gnu.org/ml/gcc-testresults/2006-02/msg00477.html
--
Eric Botcazou