This is the mail archive of the
mailing list for the GCC project.
Re: gcc/build-make wrongly disables fixinc
- To: flee at aracnet dot com
- Subject: Re: gcc/build-make wrongly disables fixinc
- From: Bruce Korb <bkorb at cruzio dot com>
- Date: Tue, 20 Feb 2001 19:38:08 -0800
- CC: GCC-Patches <gcc-patches at gcc dot gnu dot org>
- Reply-To: bkorb at allegronetworks dot com
> ok, the huge gaping flaw in what I was doing was
> right now, I'm thinking libiberty needs to be built 3 ways,
> for the build machine as well as for host and target. the
> alternative is kind of messy.
> or am I missing something?
No, I think this time you hit it square on the head. Except
maybe that all the alternatives are kind of messy. There are
occasional mumblings about building libiberty for build and
host machines. The real deal is that fixincl must run on
whatever platform has access to the header files. Perforce,
that must be the host machine, though it may also be the build.
An interesting project might even be to *install* fixincludes
along with gcc et al. That way you could do binary-only
installations (object RPMs instead of source RPMs) and OS
upgrades without having to rebuild everything for the sole
purpose of fixing the include files. Not today. :-)