This is the mail archive of the
gcc@gcc.gnu.org
mailing list for the GCC project.
successfully built and installed GCC
- From: Bernd Jendrissek <berndj at prism dot co dot za>
- To: gcc at gcc dot gnu dot org
- Date: Wed, 24 Jul 2002 12:00:47 +0200
- Subject: successfully built and installed GCC
- Reply-to: berndfoobar at users dot sourceforge dot net
After months of trial and tribulation, I finally built and installed GCC 3
from pristine source. (See note at end to avoid the same pitfalls I made.)
So here's my celebratory note:
1. $ ./config.guess
i686-pc-linux-gnu
2. $ gcc-3.0.4 -v
Reading specs from /usr/lib/gcc-lib/i686-pc-linux-gnu/3.0.4/specs
Configured with: /usr/src/home/rpmtest/tmp/BUILD/gcc-3.0.1/configure --prefix=/usr --sysconfdir=/etc --localstatedir=/var --infodir=/usr/share/info --mandir=/usr/share/man --srcdir=/usr/src/home/rpmtest/tmp/BUILD/gcc-3.0.1 --with-slibdir=/lib --program-suffix=-3.0.4 --with-gnu-as --with-gnu-ld --enable-shared --enable-threads --enable-cpp --with-dwarf2 --with-system-zlib
Thread model: posix
gcc version 3.0.4
3. $ grep -ir 'red *hat.*6\.[12]' /etc
/etc/redhat-release.old:Red Hat Linux release 6.1 (Cartman)
4. $ uname -a
Linux xxxxx 2.4.18 #4 Sun Mar 3 16:50:20 SAST 2002 i686 unknown
5. $ rpm -q glibc
glibc-2.1.2-11
glibc-2.2.4-2
(Yes, two of them. Symlinks point to 2.2.4 libs; I haven't managed to
uninstall 2.1.2 completely yet - too much depends on stuff that shouldn't
be in the rpm. BTW 2.2.4 is *NOT* a Red Hat rpm - it's my own.)
Fantastic work by all the GCC guys! I would truly hate my day job if I
didn't have GCC to use or even to aspire for.
Bernd Jendrissek
P.S. Here is why I took so long (enjoy your meal, spiders!)
Trying to be smart, as part of my RPM spec file, I stripped *everything*
under /usr/lib/gcc-lib/<arch>/<version>/ - in particular, the crt*.o files
were some of my victims. Doh! So after barking up the libgcc_s.so tree
for a while, thinking it was some bad interaction with glibc-2.2.4 which I
*also* built and patched, I finally started doing the right thing: finding
out *why* _init jumped into the middle of call_gmon_start. Answer? Do not
strip crtbegin.o (or any other *.o for that matter). Stripping it caused
the crtbegin.o call to frame_dummy to call right into the middle of another
function. Result? Segfault on startup.
--
berndj@users.sourceforge.net is probably better to bookmark than any
employer-specific email address I may have appearing in the headers.