Testing - Chap 6 Glibc

Larry Lawrence larry at linuxfromscratch.org
Sun Jul 18 08:46:02 PDT 2004

On Sun, 18 Jul 2004 17:19:46 +0200, Miguel Bazdresch wrote:

> * Larry Lawrence <larry at linuxfromscratch.org> [2004-07-18 15:11]:
>> In June Bruce Dubbs wrote:
>> In the make check, I got two errors: wordexp-test.out and scanf4.out.
>> Does this sound normal?
>> There is no pointer in the book where to see what errors others have
>> found.  Its not on the wiki.  There needs to be better discussion in the
>> book on how to handle these errors.
>>   -- Bruce
>> P.S.  Using Unstable CVS 20040615
>> There was no response on the mailing list and no changes to the book or
>> the wiki. Obviously the LFS editors know the answer to this.  The way I
>> am reading the book, the two errors above are not on the 'approved
>> error' list and I should not proceed.
> I'm not an editor but I know that at least on x86 there should be *no*
> errors on Ch. 6 glibc make check.
> Some older non-intel, x86-compatible proocessors may have errors in the
> math tests due to limitations of their FPUs.
> Some older hardware might present time-outs in some tests. I have a P3-800
> and I've never had a timeout, though. There is more about it on the
> archives, including how to change the timeout value, IIRC.
> HTH,

It does work IF you have /dev/null and /dev/console.  The instructions
create them and then destroys them with the very next instruction and none
of the instructions after that created the /dev/null and /dev/console
again on my build.  Problem number 1 for me is I have no idea how udev
works, but the instructions imply the it's utilizing /sys to work and
there is a null and console device in /sys. What was suppose to happen?


More information about the lfs-dev mailing list