configure.build_arch

[Jack Howarth]

On Tue, Sep 15, 2009 at 08:21:20PM -0400, Mike Alexander wrote:
> --On September 15, 2009 5:18:32 PM -0700 Toby Peterson  
> <toby at macports.org> wrote:
>
>> On Tue, Sep 15, 2009 at 17:05, Mike Alexander <mta at umich.edu> wrote:
>>> --On September 15, 2009 4:01:43 PM -0700 Toby Peterson
>>> <toby at macports.org> wrote:
>>>
>>>> For the few ports that actually care, only --build makes any sense.
>>>> We don't support cross-compiling, and very few configure scripts
>>>> get it right anyway.
>>>
>>> I think Jack's point is that in 10.6 on a 32 bit kernel you
>>> effectively are cross-compiling whether you want to or not.  You're
>>> building 64 bit binaries on a system that claims to be a 32 bit
>>> system.  That seems like cross-compiling to me.  Am I
>>> misunderstanding something?
>>
>> As I've noted numerous times now, config.guess still reports i386 on
>> K64, so this perceived problem exists regardless of the kernel
>> architecture.
>
> Ok, that just seems to make the problem worse.  Even after switching to  
> the 64 bit kernel you're still cross-compiling, right?
>
>          Mike

Mike,
    Yes. I reported this issue on radar several times during the
Snow Leopard development but it was never addressed. The solution
(which Ben Elliston who is the config.guess maintainer agrees with)
is to decouple the config.guess result on i386 darwin from the kernel and to
use the code generation of the compiler or the code execution in
the absence of a compiler. However even when config.guess is fixed
it will take a long time for all projects to update to a fixed
config.guess. In the meanwhile we have to be very careful to
make sure that configure is clued into the correct architecture.
In the case of the configure.build-arch, the fix is obvious and trivial
if 'configure.build-arch x86_64' is used "--target=x86_64-apple-darwin${os.major}"
needs to be added to the configure arguments by configure.build-arch
as well.
                 Jack