glibc-2.3.4 make check error

Brandon Peirce brandon_peirce at hotmail.com
Fri Aug 4 09:38:24 PDT 2006


Danny wrote:
>Although I could go on, I'm thinking of starting all over now with 6.2 (as 
>the 6.1.1 book isn't available online anymore (am I correct in this??))

http://archive.linuxfromscratch.org/lfs-museum/6.1.1/LFS-BOOK-6.1.1-HTML/

All old versions are available there, in all the usual formats.
If you wish to continue with 6.1.1 though, I suggest you make sure to
download everything you need before you go on. I had to abandon a build
of LFS-5.0 a couple of years ago because I got half way and then one of
the critical patches wasn't available anymore...


>Any chance that 6.2 will lead to more timeout events 'cause of the old 
>hardware?

Possibly.  Occasionally some code gets optimized, but as a general rule
new versions of the same package will be slightly slower because there
is more code. Usually the increase in time is so little that it's not 
noticable.

If you have a test case that is now finishing just under the timeout limit
then there is a good chance that it might now be just over the limit.
However, with such cases, you should now already observe that they
pass or fail "randomly" according to how heavily your system is loaded.

Remeber also that one of the big differences in LFS 6.2 is a major new
version of GCC which could lead to some unpredictable timing differences.

Go ahead and try 6.2 if you like. The major toolchain packages at the
beginning of the process Gcc and Glibc and possibly binutils are the
most likely to give you problems. If you get past about chapter 5.11 or
5.12 you should be OK for the rest.

BTW, there was a posting on lfs-chat from someone who completed a
successful LFS build on a 80386 16MHz 24MB. He didn't say what he
did with the testsuites, but I think the whole build took 3-4 months.
Glibc took ~7 days to compile and another ~5 days for 'make install'.
By comparison, your P-II should compile it in under 12hrs. I'm sure
6.2 is doable on your P-II, just a question of how much time to spend
investigating test failures.

Brandon





More information about the lfs-support mailing list