hi Bruce, hi Ken. thanks for your answers. i actually started with 7.4
but s.t. went wrong and i thought starting with a not as young version
was a good idea. 7.0 is the newest from out of the history archives i
could find (in between i found 71..73, too). maybe more than for a
special version, i need a little help for general understanding.
i started from beginning and now i got up to 6.13 installing binutils
(some errors in test). first time 'expect -c "spawn ls"' i got the
error. but there was no obvious reason. i found that the tty-group-id
was 5 on host and 4 in chroot. i changed and error has gone.
when /dev is bind-mounted on $LFS/dev i know (now) why virtual fs like
shm:tmpfs and pts:devpts need to be mounted, but all the device-nodes
are available from inside chroot without further work. but why
/dev/console and /dev/null need to be created before? or do they
simply have to exists when in the end the system is actually booted ?
good school! thanks
2013/9/16 Ken Moffat <zarniwh...@ntlworld.com>
>
> On Mon, Sep 16, 2013 at 05:36:20PM +0200, andreas graeper wrote:
> >
> > book-version is 7.0, see subject.
>
> And that, in linux terms, is ancient history - we've had four more
> releases, a lot of things have changed.
>
> >
> > the testprogram is linked against /tools/lib/libgcc_s.so
> >
> > libgcc_s and libstdc++ are linked from /usr/lib to /tools/lib
> >
> That sounds correct
>
> > why libstdc++ was build ?
> > when it is build, is it done with c-compiler or with the wrong (host-)
> > g++-compiler ?
> >
>
> At that time, we built it with the C compiler. In 7.4, gcc-4.8.1
> is used : 4.8 needs c++ which is why Bruce mentioned.
>
> I don't understand why you are questioning why libstdc++ was built.
> The system needs a c++ compiler, and that needs libstdc++.
>
> > what is libgcc_s ? google shows only error messages ('not found' and so on)
> > .
>
> The most likely answer I can find is at
> http://unix.stackexchange.com/questions/1812/what-does-libgcc-s-so-contain
> i.e. it's a gcc runtime library.
>
> >
> > could this problem get solved by 're-adjusting the toolchain' (6.10) before
> > this test (glibc make install in 6.9) ?
> >
> > thanks in advance
> > andreas
>
> I guess the error message about cross-compilation means that
> somewhere you are still using the cross-compiler which was built
> in gcc pass 1, or else that something went seriously wrong with the
> instructions in gcc pass 2.
>
> I'm going to stick my neck out and suggest that you didn't delete
> the existing source and -build directories when you came to Pass 2 ?
> It is essential to use clean source, and empty -build directories
> where those are used, for *every* package.
>
> If that guess is correct you will be better to start again. Nobody
> remembers the details of a 2-year-old set of instructions, and the
> system will be out of date. So I suggest that you use the 7.4 book.
> Check the host requirements and update anything which is too old.
>
> ĸen
> --
> das eine Mal als Tragödie, dieses Mal als Farce
> --
> http://linuxfromscratch.org/mailman/listinfo/lfs-support
> FAQ: http://www.linuxfromscratch.org/lfs/faq.html
> Unsubscribe: See the above information page
--
http://linuxfromscratch.org/mailman/listinfo/lfs-support
FAQ: http://www.linuxfromscratch.org/lfs/faq.html
Unsubscribe: See the above information page
