--- Ken Moffat <[EMAIL PROTECTED]> wrote:

> > >  I'm now almost certain this is a bash problem
> > 
> [...]
> > 
> >  So, it looks as if we need to apply the bash
> patches in chapter 6
> > (temporary system).  Unfortunately, proving that
> means starting over
> 
>  Testing has been somewhat delayed by my adventures
> with e2fprogs,
> but a build is running at the moment.  It's still
> building 64-bit
> glibc, but the test results from 32-bit glibc are
> not encouraging:
> an extra failure (one of the nptl tests, no big
> deal), and then
> c++-types-check.out failed (on the previous build,
> that only failed
> in 64-bit).  So, applying the patch to bash in the
> temporary system
> will NOT reliably fix this failure.
> 
>  I still think it's probably bash-related, but I
> have no way of
> proving that, nor of fixing it.  Maybe we should add
> a comment that
> c++-types-check often fails for unknown reasons ?  I
> don't know
> which is worse - not mentioning a failure which
> people are fairly
> likely to see, or saying we haven't a clue.

Ken,

I've done some more tests, and I can confirm that it's
a bash problem. But I can't point exactly what.
However, as I looked in bash's source code, I saw that
all the parsing is related to bison. However, bash can
be installed without bison, as it's done in chapter 6.
But I suspect it could be the problem, and that bison
should be installed before bash in 6.6. Unfortunately,
I have no machine to test it from the beginning, so
it's only hypothesis.
What do you think about this?

G. Moko


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 
_______________________________________________
Clfs-dev mailing list
[email protected]
http://lists.cross-lfs.org/cgi-bin/mailman/listinfo/clfs-dev

Reply via email to