[Pythonmac-SIG] macpython on future apple's intel processors?

[Bob Ippolito]

On Jun 7, 2005, at 11:14 AM, Christopher Petrilli wrote:

> On 6/7/05, Bob Ippolito <bob at redivi.com> wrote:
>
>> That's totally not true, though.  Basically, almost NONE of the Mac-
>> specific Python stuff works correctly on Mac OS X for Intel, and the
>> changes required to fix that are hard.
>>
>
> I'm not that familiar with the specific changes, but from reading your
> blog, a lot of it "sounds" like it's working outside the defined APIs
> and frameworks.  Given Apple's recent interests in Python, and the
> year lead-time before we have anything serious to worry about, it
> seems that a proposal to formalize the interface some, and add some
> stable APIs for what we need would be the best way to go.

Well, bgen is working with the published APIs, but it assumes four  
character codes are going to have a big endian representation.. which  
was true from when it was written many years ago until yesterday.

libffi is part of the gcc project, which is really Apple's  
responsibility to take care of, but since it's not used in anything  
they've written it might not be fixed by them.  Fortunately, it  
should already understand most of the calling convention for Mac OS X  
on Intel, but beating autoconf into making it do the right thing is  
non-trivial.  Also, compiling libffi fat is going to be excruciating  
(since ALL of the code in there is platform specific, and it's done  
with autoconf).

> Note that it took a while for the APIs in the kernel to stabilize, but
> that's happened to a large extent with Tiger. If we were to approach
> them with a proposal, then I think we might have a decent chance.
> This wouldn't eliminate the problems, just simply fomalize them and
> make them recognized by Apple. That's what this whole 1 year's notice
> is about.
>
> Or am I just on crack today?

I guess what you don't realize is that a lot of this work is going to  
end up on a few people outside of Apple (including myself), because  
nobody is going to fix the problem for us.  Fortunately, you don't  
have that problem.

-bob