CVS - where should I build chapter 6 ?

Adam Trilling agt10 at
Tue Jun 4 11:10:26 PDT 2002

On Tue, 4 Jun 2002, Paulo Andre' wrote:

> >  The first place it makes a difference is the symlinks at the end of
> > gcc. I'm not sure I fully understand these, but I certainly don't have
> > ../usr from where I'm building, and ../bin is /usr/bin - I'm not
> > immediately clear whether /bin or /usr/bin is intended. Is there a
> > reason why these symlinks have relative paths ?

The way I understand it, they don't HAVE to have relative paths.  But the
reason they exist is that gcc lives in /bin in case you don't have /usr
mounted when you need it.  That's much more likely on an LFS system than
on another system, so it makes sense to do it that way.

However, /usr/bin/cc is very widely assumed to exist.  I think FHS
requires it, in fact.  The other parts of gcc are similar.  Therefore, you
make symlinks.

> I guess you mean these...
> ln -sf ../usr/bin/cpp /lib &&
> ln -sf ../bin/cpp /usr/lib &&
> ln -sf gcc /usr/bin/cc
> Well, I don't think the above are entirely correct, so it confuses a
> bit. I'd say it's something along these lines (where $LFS is the root of
> your LFS partition)...
> ln -sf $LFS/usr/bin/cpp /lib &&
> ln -sf $LFS/bin/cpp /usr/lib &&
> ln -sf gcc /usr/bin/cc
> However, I may be mistaken, don't trust heavily on it. Seems alright to
> me but you'd want to have a second opinion :)

This is in chapter 6, after we've chrooted.  Therefore, $LFS doesn't exist
and is assumed to be /

> >  Oh well, back to rebuilding gcc, my timing script only gives me a
> >  time if
> > the package and associated commands runs to completion. Of course,
> > this might mean I'm compiling with files already cached. Anybody got
> > any good suggestions for how I can flush the cache ?
> make clean? make distclean?

make distclean should do it, and a lot of configure scripts have an option
to ignore their cache file.  try ./configure --help

Adam Trilling
agt10 at

char m[9999],*n[99],*r=m,*p=m+5000,**s=n,d,c;main(){for(read(0,r,4000);c=*r;

Unsubscribe: send email to listar at
and put 'unsubscribe lfs-dev' in the subject header of the message

More information about the lfs-dev mailing list