[EMAIL PROTECTED] wrote: > Author: randy > Date: 2008-10-12 09:09:40 -0600 (Sun, 12 Oct 2008) > New Revision: 8649 > > Modified: > trunk/BOOK/appendices/dependencies.xml > trunk/BOOK/chapter01/changelog.xml > trunk/BOOK/chapter05/chapter05.xml > Log: > Moved the Chapter 5 M4 installation to before GCC Pass1 so that the internal > GCC build of GMP will not fail in case M4 doesn't exist on the host
I know this is coming late, and I'm speaking here without having tested anything, but this doesn't feel like it should work. What happens in chapter 6? I can't see how m4 would even work, because it would be linked against the host's libs since it is built before we build glibc in chapter 5. Is there something I'm missing? When I get a chance, I'll see if I can run a testcase, but there's one of two things happening here: 1. I'm way off and need to be thrown a major cluebat. 2. gmp doesn't really use m4 at all in chroot, in which case it can be dropped from chapter 5 entirely. Hmm. We build glibc early in chapter 6. Maybe once it's built and we move on to building our final gcc, /tools/bin/m4 just happens to work by pure luck because now we have a proper library in /lib Whichever way I look at it, it feels dirty. :/ -- JH -- http://linuxfromscratch.org/mailman/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
