Twist Guru
Joined: 03 Jan 2003 Posts: 414 Location: San Diego
|
Posted: Sun Jan 05, 2003 12:02 am Post subject: Xfree ebuild overly conservative? |
|
|
$cd /usr/portage/x11-base/xfree
b$ grep "pentium4" *
xfree-4.2.0-r12.ebuild: # should fix bug #4189. gcc-3.x have problems with -march=pentium4
xfree-4.2.0-r12.ebuild: echo "#define OptimizedCDebugFlags ${CFLAGS/pentium4/pentium4 -mno-mmx} -fno-merge-constants" >> \
xfree-4.2.1-r1.ebuild: # should fix bug #4189. gcc-3.x have problems with -march=pentium4
xfree-4.2.1-r1.ebuild:# export CFLAGS="${CFLAGS/pentium4/pentium3}"
xfree-4.2.1-r1.ebuild:# export CXXFLAGS="${CXXFLAGS/pentium4/pentium3}"
xfree-4.2.1-r2.ebuild: # Should fix bug #4189. gcc-3.x have problems with -march=pentium4
xfree-4.2.1-r2.ebuild: replace-flags "-march=pentium4" "-march=pentium3"
xfree-4.2.1.ebuild: # should fix bug #4189. gcc-3.x have problems with -march=pentium4
xfree-4.2.1.ebuild: export CFLAGS="${CFLAGS/pentium4/pentium3}"
xfree-4.2.1.ebuild: export CXXFLAGS="${CXXFLAGS/pentium4/pentium3}"
xfree-4.2.99.3-r1.ebuild: # Should fix bug #4189. gcc-3.x have problems with -march=pentium4
xfree-4.2.99.3-r1.ebuild: replace-flags "-march=pentium4" "-march=pentium3"
------------------------------------------------------------------
Is this a legacy fix that is persisting past it's point of use? It seems rather prophetic to claim in earnest that ALL gcc-3.X variants for all time will have issues with -march=pentium4. The bug in question existed in gcc 3.1.
Reason I ask is that I am running Xfree-4.2.1 built -march=pentium4 -mcpu=pentium4 with no problems whatsoever (and further, on a kernel built same parameters). It's stable as a rock.
I'm going to update the bug on this with the info and see what happens, shrug. Would hate to see all Xfrees for all time unable to optimize for a P4 processor. |
|