On Dec 20, 2015 11:01 AM, "Ken Moffat" <[email protected]> wrote: > > I'm getting close to making some specific suggestions to try to > reduce the number of errors that we see from new builders. One of > those is a script to be run as-desired in chapter 5 to look for > programs and libraries linked to the host system (see my past few > posts). > > Unfortunately, gcc pass 1 builds libcc1.so which is, of course > linked to the host. Ignoring pass 1 progs called -lfs- is easy, but > to decide whether or not to test libcc1 I need to determine if gcc > pass 2 has been installed. And unfortunately, the tests ought to be > runnable for all chapter 5 packages from glibc onwards. > > On x86_64, gcc pass 2 installs x86_64-unknown-linux-gnu-gcc. On my > previous 32-bit builds, it installed i686-pc-linux-gnu-gcc but I am > not certain if that was because of some minor way in which I might > have deviated (in the same way that I used to build glibc for i686), > or whether the 32-bit compiler will always be i686 ? I cannot > imagine anybody trying to build current LFS on i586 or i486, but who > knows ? > If we need somebody to test that... I have about 10 Pentium 75-233MMX systems. I also have a 486 coming in. Should I try it?
-- http://lists.linuxfromscratch.org/listinfo/lfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
