On 9/25/07, Kevin Oberman <[EMAIL PROTECTED]> wrote:
> > Date: Tue, 25 Sep 2007 02:44:33 +0300
> > From: "Abdullah Ibn Hamad Al-Marri" <[EMAIL PROTECTED]>
> > On 9/25/07, Mel <[EMAIL PROTECTED]> wrote:
> > > On Monday 24 September 2007 22:47:59 Abdullah Ibn Hamad Al-Marri wrote:
> > >
> > > > Could you please put it somewhere where so that I could fetch it?
> > >
> > > Probably the safest way:
> > > pkg_add -r gcc-4.2.2_20070905.tbz
> > >
> > > Then set CC and CXX in /etc/make.conf to point to the installed gcc
> > > in /usr/local. This should get you through the build-tools stage. If it
> > > creates issues later on, you should have a libgcc_* in /usr/obj. In fact,
> > > I
> > > highly recommend pressing ctrl-c after gcc has been built and copy the
> > > missing library from /usr/obj to /usr/lib then unset CC and CXX
> > > in /etc/make.conf and re-run buildworld.
> > >
> > > I just looked on freebsd ftp servers, there's a package for amd64 arch and
> > > 7-current dated Sep 15.
> > > --
> > > Mel
> > Hello,
> > I installed that.
> > And it did the trick.
> > I just don't want to take the risk now.
> > I did these steps
> > rm -r /usr/obj/*
> > cd /usr/src
> > make cleandir
> > make cleanworld
> > make -j7 buildworld
> > I'm too worried, I may get broken world now.
> > Shall I recopy libgcc_* from /obj again and rebuild the world again?
> > since current libgcc_* is made by the gcc42 latest port.
> If you use gcc 4.2.2 to buildworld, it does a two stage build. It starts
> by building the base gcc (4.2.1) withe the default compiler and then
> builds it again using the just built compiler. You should be fine if you
> define CC as gcc422. It will only be used to build the gcc4.2.1
> compiler. Once that is done, the make system will use only the newly
> built version.
> If you have removed /usr/obj/*, you can speed the build with -DNO_CLEAN.
> (There is nothing to clean, but make will still try.) Since there is
> nothing to clean, making cleandir and cleanworld looks unnecessary,
> Unless you have a 6 core system, -j7 is probably excessive. Both my own
> tests and those of others show that having one more build thread than
> there are processors seems to be the sweet spot. (I don't recall if
> anyone has tested at over 6 cores, though.)
> I think I understand the build system, but I am far from a make(1) guru,
> so. if I misunderstand any of it, hopefully those who are will chime in
> with the right information.
> R. Kevin Oberman, Network Engineer
> Energy Sciences Network (ESnet)
> Ernest O. Lawrence Berkeley National Laboratory (Berkeley Lab)
> E-mail: [EMAIL PROTECTED] Phone: +1 510 486-8634
> Key fingerprint:059B 2DDF 031C 9BA3 14A4 EADA 927D EBB3 987B 3751
Thanks I fixed by coping it from obj
-Abdullah Ibn Hamad Al-Marri
firstname.lastname@example.org mailing list
To unsubscribe, send any mail to "[EMAIL PROTECTED]"