Date: Thu, 20 Nov 2003 15:18:19 -0500 From: Joe Marcus Clarke <marcus@marcuscom.com> To: mi+mx@aldan.algebra.com Cc: FreeBSD GNOME Users <gnome@freebsd.org> Subject: Re: Mozilla's xptcall Message-ID: <1069359499.783.65.camel@gyros> In-Reply-To: <200311201511.08686@misha-mx.virtual-estates.net> References: <200311201454.56022@misha-mx.virtual-estates.net> <1069358484.783.51.camel@gyros> <200311201511.08686@misha-mx.virtual-estates.net>
next in thread | previous in thread | raw e-mail | index | archive | help
--=-627TtQcGIZDMHsJ82LJI Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Thu, 2003-11-20 at 15:11, mi+mx@aldan.algebra.com wrote: > On Thu, 2003-11-20 at 14:54, mi+mx@aldan.algebra.com wrote: > =3D> Hello! > =3D>=20 > =3D> In order to build mozilla (and hence, OpenOffice) I needed the > =3D> attached patch. It seems, that the xptcall files are very fragile > =3D> to the compiler optimization, so the patch is needed to replace > =3D> whatever the configured through the make.conf with the vanilla > =3D> ``-O''. Curiously, the file being patched already contains similar > =3D> hacks for other platforms. Yours, >=20 > =3Dmozilla and mozilla-devel build fine with -O2. >=20 > regxpcom usually crashes for me unless built with -O. But I'm using > -march=3Dpentium4 and -fomit-frame-pointer. It's not the optimization level, but rather -march=3Dpentium4 that causes problems with Mozilla. This is a known issue. >=20 > This time I went with a very elaborate: >=20 > -march=3Dpentium4 -O \ > -fomit-frame-pointer -fbuiltin -fstrength-reduce \ > -fthread-jumps -funroll-loops -fcse-follow-jumps \ > -fcse-skip-blocks -frerun-cse-after-loop \ > -fexpensive-optimizations -fschedule-insns -fschedule-insns2 >=20 > Most (all?) of the rest of the code is either C or C++ and it is reasonab= le > to expect the compiler to do the right thing. The directory in question, > however, contains mostly assembly code, which is being mis-compiled with > more agressive optimization options. Everyone that has brought up compilation issues or crashers with Mozilla has been pointed back to /etc/defaults/make.conf to red the section on default CFLAGS. However, -O2 has been tested, and does work. There is even an option to enable it in the Mozilla Makefiles. 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. Joe >=20 > -mi >=20 --=20 PGP Key : http://www.marcuscom.com/pgp.asc --=-627TtQcGIZDMHsJ82LJI Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.3 (FreeBSD) iD8DBQA/vSGLb2iPiv4Uz4cRAjmtAJ9lW4izEzsn6i9ZpENN2lVgBiFp9gCfcSTJ 7CS3yiCnVdGTCL34YV+4kGo= =2AZG -----END PGP SIGNATURE----- --=-627TtQcGIZDMHsJ82LJI--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?1069359499.783.65.camel>