a few short questions about the development cross-lfs book

Robert P. J. Day rpjday at mindspring.com
Tue Nov 1 10:37:22 PST 2005


On Tue, 1 Nov 2005, Ken Moffat wrote:

> On Tue, 1 Nov 2005, Robert P. J. Day wrote:
>
> >
> > i did a quick toolchain build with crosstool and, AFAICT, there is no
> > libiberty.h header file that was copied over.  perhaps this step is no
> > longer necessary?
> >
> > rday
> >
>
>  Possibly. Looking at my logs, for LFS-6.1 both which and gdb tested
> for libiberty.h, but in a more recent build gdb used its own
> included libiberty (this is all on x86_64-64 and I'm uncertain why
> the earlier build was different).
>
>  'which' tests libiberty.a to see if xmalloc is present before
> looking to see if it can use libiberty.h, so no doubt it will have a
> fallback plan. Both binutils and gcc use their own included
> libiberty.
>
>  Certainly, if it is really redundant, that will affect LFS as much
> as CLFS. Actually, some of the discussion is in
> http://bugs.linuxfromscratch.org/show_bug.cgi?id=565
>
>  - sounds as if "three known packages" needed it.  I've no idea what
> those packages were, or whether they still need it.  I suspect that
> I'd be perfectly happy to see it go, in the spirit of "any package
> that needs it should ship its own".  IMHO, this is a matter for LFS.
> Unless anybody knows what the packages are, or were, merely
> asserting that they are broken isn't going to be a great help for
> anybody who wants to build them.

well, i'm not going to pretend i followed all that, and i'll leave it
to people way smarter than me to decide how to resolve this.  i only
posted what i observed in the construction of my toolchain.

rday



More information about the cross-lfs mailing list