If all else fails, hack it.

Chris Lingard chris at stockwith.co.uk
Mon Oct 7 11:18:27 PDT 2002

Gerard Beekmans wrote:

> On October 7, 2002 02:06 am, Greg Schafer wrote:
>> I'm afraid you are SOL if you want to use stock glibc 2.3 with those old
> <cut>
> Very nice.
> My vote goes for sed'ing a binary. I really don't like the idea of having
> to do things twice such as Glibc twice. Running sed on a binary might be
> considered a dirty hack, but personally I don't really care about that.
> Whatever is in $LFS/static is only temporary and I don't care what we do
> to the files in there. As long as it runs.
> The alternative as I see it would to re-write Bash and comment out the
> function calls to getpwuid. Write out own that returns a false like Glibc
> would do when NSS is unavailable. But that's a lot more work and is it
> worth the effort? I doubt it.

This is a bit rough, will tidy it up later

After building glibc-2.3, I rebuilt bash with the base system with the
following set

export CFLAGS=-I$LFS/usr/include
export LDFLAGS=-L$LFS/lib
 ./configure --prefix=$LFS/usr  --bindir=$LFS/bin  --libdir=$LFS/lib  

I copied libncurses.so.* into $LFS/lib

I could then do

chroot $LFS
export PATH=/static/bin

I then rebuilt fileutils inside chroot and installed

I could then do
export PATH=/bin:/usr/bin/:/static/bin
ls  -l

I then used MAKEDEV to build the devices

I think that I will be able to complete chapter 6 now.  I told you
this would be rough.

I think that we will have problems compiling ncurses _ctype_b_loc
not defined in headers


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

More information about the lfs-dev mailing list