Mozilla's xptcall

[mi+mx at aldan.algebra.com]

On Thu, 2003-11-20 at 15:11, mi+mx at aldan.algebra.com wrote:
=> On Thu, 2003-11-20 at 14:54, mi+mx at aldan.algebra.com wrote:
 
=> =mozilla and mozilla-devel build fine with -O2.
=> 
=> regxpcom usually crashes for me unless built with -O. But I'm using
=> -march=pentium4 and -fomit-frame-pointer.

=It's not the optimization level, but rather -march=pentium4 that causes
=problems with Mozilla.  This is a known issue.

The patch I submitted will make this knowledge usefull to others and
save everyone with the CPUTYPE defined the few hours I wasted on this
yesterday. Please, consider. Thank you,

=Everyone that has brought up compilation issues or crashers with Mozilla
=has been pointed back to /etc/defaults/make.conf to read the section on
=default CFLAGS.

I think, committing my patch (with modification if need be) is much better.
It lets people better optimize the rest of mozilla for their hardware and
saves frustration and time spent dealing with the problem and e-mailing for
support. The patch will definetly not do harm, as it will not change the
default case (even bumping PORTREVISION is not needed).

=However, -O2 has been tested, and does work.

I did not use -O2...

=There is even an option to enable it in the Mozilla Makefiles.

[BTW, how about an option to use -lthr instead of -lc_r? I built Qt/KDE
with that and it works terrific on a dual CPU machine).]

=However, I do not think this was ever tested on mozilla-vendor, so you
=may have something there. In that case, this would need to be applied
=to OpenOffice as they have the last remaining mozilla-vendor bits.

Yours,

	-mi