LFS-5.1-pre2 glibc

Ken Moffat ken at kenmoffat.uklinux.net
Sun May 16 13:13:48 PDT 2004


On Fri, 14 May 2004, Ken Moffat wrote:

>
>  Hmm, I'm starting to wonder why I was so keen to get the "better"
> version of glibc.  On my i686 the scripted chapter 6 build appears to be
> hanging at the end of the gettext test suite : "minimal" locales, 5
> tests failed, which is par for the course with whatever I'm missing in
> my build method, a couple of Error indications from make, then nothing,
> and `ps' doesn't show any descendants of my build-script, except for the
> predictable filebuf-members.exe.  Anybody got any good results on x86 ?
>

 I'll have to chalk that one up to experience (looks like the sort of
thing that can happen with my KVM switch when it gets confused about
things, although I'm 99% sure it was on a different console).  After
restarting, it went through ok, and then built itself without any
difficulty.

 However, there is one new anomaly between the two builds with
glibc-2.3.3-lfs on i686 (LFS-4.1 building it all, then the new system
repeating the process).  In coreutils (chapter 6) there is a test in
configure

checking whether getcwd properly handles pathes longer than PATH_MAX

- first time through this got a result of no, second time got a yes.  As
a result, on one build getcwd.o was included in libfetish.a, the other
build it wasn't.  The following files have different sizes between the
two builds, I'm guessing this was why:

/bin/{df,install,mkdir,mv,pwd,rm}

Curious.

Ken
-- 
 das eine Mal als Tragödie, das andere Mal als Farce




More information about the lfs-dev mailing list