Hi everyone,
I've been working on getting glibc-2.4 into the book over the past few
days. Everything is looking pretty promising so far. Here's a little
status report.
I'm using jhalfs, so I've already got the book diffs done. Built on
two hosts, one Athlon XP 2100+ and a Pentium III 500. One common test
failure, tst-cancel24 in glibc was due to me forgetting that I need to
add /tools/lib/libstdc++.so{,.6} to the Essential Symlinks.
The Athlon got two other glibc test failures, but it always does. It's
a server under load, so I attribute these failures to that. Wish I
knew more, but the randomness of which failures it is tells me that
the errors aren't due to the build.
On the PIII, I got one GCC failure due to a timeout (this box is dog
slow and I think I was using a GNOME desktop during the tests). Also
one failure in tar listed02.at that I got when I was building LFS-6.2,
too. I can't narrow this bastard down, but I don't attribute it to
glibc-2.4.
Next step is for me to boot up this baby at home, build a few more
packages, and make sure everything's running smoothly.
If anyone wants to see the diff as it stands, I've placed it here:
http://www.linuxfromscratch.org/~dnicholson/glibc-2.4.diff
A couple quick notes about the build. I used the iconv_fix patch from
CLFS, but not the localedef patch as it seems to be non-x86 specific.
I also didn't bother with the openat fix needed when using
--enable-kernel=2.6.17+. Next major change will be the kernel headers.
That's another discussion, though.
--
Dan
--
http://linuxfromscratch.org/mailman/listinfo/lfs-dev
FAQ: http://www.linuxfromscratch.org/faq/
Unsubscribe: See the above information page
