This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: introducing mips64{,le}-linux-gnu
On Jan 16, 2003, Richard Henderson <rth@redhat.com> wrote:
> On Wed, Jan 15, 2003 at 05:49:27PM -0800, Zack Weinberg wrote:
>> Wouldn't it be better to divorce "this is the default multilib" from
>> "this multilib gets installed loose in ${prefix}/lib"? Over in
>> VxWorks land I have a similar problem; I want *none* of the multilibs
>> installed loose in ${prefix}/lib...
> Almost certainly.
It surely would. But do you actually oppose any of the pieces of my
patch? E.g., do you actually find it unacceptable that we process
DRIVER_SELF_SPECS after reading in specs files? Would you bar the
possibility of installing a specs file different from that used at
build time? I don't really need approval for the latter, since it's a
configury issue, but I thought I'd give a chance for people to voice
opinions against it, or point out gotchas I might have missed. But
those are the changes I need enough for the whole thing to work. The
rest of the confusing plan is already in, and we should probably be
fixing the fact that MULTILIB_EXTRA_OPTS doesn't get used when it
should.
As for the issue of not installing libraries in ${prefix}/lib, I agree
that's a good idea and we should eventually do it, but it doesn't help
solve the problem I had to solve at all, and I don't have time atm to
do it just for the fun of it. The problem I have has to do with
SONAMing of shared libraries, and that's much trickier. I looked at
the code for a moment, and couldn't find a clean way to introduce the
change, so I left it alone.
--
Alexandre Oliva Enjoy Guarana', see http://www.ic.unicamp.br/~oliva/
Red Hat GCC Developer aoliva@{redhat.com, gcc.gnu.org}
CS PhD student at IC-Unicamp oliva@{lsd.ic.unicamp.br, gnu.org}
Free Software Evangelist Professional serial bug killer