Re: config.guess and $CC

[Peter O'Gorman]

Paul Eggert wrote:
> Bruno Haible <address@hidden> writes:
> 
>> On a bi-arch system (i.e. a system that can execute 32-bit mode binaries and
>> 64-bit mode binaries), what is the expected behaviour of config.guess?
>> Should it return different triplets for CC="gcc -m32" and CC="gcc -m64",
>> or the same?
> 
> My assumption all along has been that it should return different triplets.
> 
> It now strikes me, though: what about "fat binaries" and Mac OS X,
> where the same program is compiled for multiple architectures all at
> one go?  Should these get a special triplet, as compared to thin
> binaries?

I worry that changing this will break something. I would guess that a
number of configure scripts currently only check for ppc*darwin* and
i386*darwin* and make decisions based on that. You could have someone
doing something like

        case $host in
                i?86*-darwin*) mac=yes; macintel=yes ;;
                *-darwin*) mac=yes; macintel=no ;;
        esac

Returning anything but i?86-apple-darwin9.1.0 on an intel Mac OS X
10.5.1 for config.guess would break this. I doubt that people would
expect updating config.guess to break their configure scripts in this
manner.

Libtool HEAD will be affected by a change like this, for example.

Perhaps config.guess could take an argument to return a more specific
target alias? Hmm, seems yucky.

Peter
-- 
Peter O'Gorman
http://pogma.com