Re: [autoconf-conversion] Re: Toplevel configury, multilibs, new autoconf versions

[Phil Edwards <phil at jaj dot com>]

On Tue, Jul 29, 2003 at 12:29:42AM +0200, Paolo Bonzini wrote:
> > My local-changes tree happens to be a unified one, and I didn't have any
> > problem with the faked multilib build (for whatever that's worth).
> 
> That's good, but I think CC should be preserved too.

Actually, I think I /want/ $CC to break in libstdc++-v3.  It'll be a
warning flag, since it shouldn't be used.

(Right now, we use it twice, but only the driver, not the actual compiler,
and I'll be changing it to use $CXX instead.)


> You might multilib on -mrtd or -mregparm=3 if you're building for x86, which
> will definitely break things if they go wrong.  That's much less faked and,
> to some extent, it might even be useful (but not in an unified tree, as it
> will interact with assembly language files in newlib and BTW also in
> libffi -- so no Java as well).

Actually, we've got a couple of sparc tests going now, which are multilibbed
for real.


> > > Also, did you do this by modifying the configure script manually or by
> > > overriding an Autoconf macro definition?
> >
> > The latter.
> 
> Which is great.

Absolutely.  I'm crazy, but I'm not insane.  :-)


-- 
If ye love wealth greater than liberty, the tranquility of servitude greater
than the animating contest for freedom, go home and leave us in peace.  We seek
not your counsel, nor your arms.  Crouch down and lick the hand that feeds you;
and may posterity forget that ye were our countrymen.            - Samuel Adams