Section 6.12 Problems (Snakes and ladders)
ken at linuxfromscratch.org
Mon Nov 19 08:04:24 PST 2007
On Mon, Nov 19, 2007 at 02:26:54PM +0000, John Whitmore wrote:
> Sorry it's me again. I haven't gotten too much further I'm afraid.
> This LFS is a bit like Snakes and ladders, I've landed on a square
> with a snake in it and I've got to go back but I've no idea how far.
> I'm stuck with problems with include files not been found in the make
> of gcc-4.1.2 in Section 6.12 of Version 6.3 of the book. I've had a
> look at the web and found a few references to this problem. There seem
> to be a few suggestions as to the resolution but I plumbed for going
> back and installing the Linux headers (Section 6.7) Problem is that
> when I remade these they didn't seem to do too much, I assume that's
> because make looks at things and says that it's all up to date.
John, you're making anybody who wants to reply get out a crystal
ball! Just talking about 'include files not found' is not very
explicit. A _few_ lines of error messages always beats
generalisations. Please re-read section 1.5.3.
If any package fails to build, after you have attempted to fix the
problem you should *always* remove the source directory and any
build directory, and re-extract them.
Maybe I'm reading too much into your choice of words, but "going
back and installing the Linux headers" almost suggests you hadn't
bothered to do that. I know we all make mistakes, and sometimes
accidentally go forward two pages if we are reading in a browser, so
I'll assume you already understand that the LFS book is intended to
be followed in order and not omitting anything.
> Is the answer to my problem to remove all the directories created in
> Section 6 in the chroot environment and restart the whole section?
After you have re-adjusted the toolchain (6.10) that is probably
not a good idea.
> Another question is that when I get to the gcc section again I've
> already tampered with it. Actually I've already tampered via sed with
> the Linux directory so should I untar these directories or where these
> directories made via Section 5? I know the actual first version of gcc
> and all the rest where created in section 5 but did it make changes to
> the source directories which mean that you can't simply re-untar the
> source directories and execute the Section 6 instructions?
Sounds as if you are leaving the source and build directories
around after each build. You need _freshly_ extracted tarballs (and
empty build directories where used) each time you run configure.
I would point you to the text in the book, but at the moment I can't
find it ;)
das eine Mal als Tragödie, das andere Mal als Farce
More information about the lfs-support