On Friday, February 03, 2006, at 04:46PM, Bob Ippolito <bob@...> wrote:
>On Feb 3, 2006, at 6:15 AM, Ben Artin wrote:
>>> My primary concern is getting a reproduceable universal build of
>>> Python 2.4, everything else should be fairly shallow. The best
>>> thing you can do right now is trying to build you're application
>>> on Intel and reporting any problems you run into.
>> The problem that I have right now is that I need a Universal-savvy
>> py2app and Universal versions of all the extensions that I depend
>> on -- most of which are framework glue that comes with PyObjC. I am
>> using stock Apple python at this time.
>You shouldn't use stock Apple Python, for much of anything...
>Especially on i386, because your app will break when Leopard comes
>out, and there's lots and lots of bugs in Apple's 2.3.5 for i386
>(anything that takes a four character code is endian-busted).
That sucks. Python's unittests even test some of the four character code support code!
Using anything else than a stock python is also bothersome at the moment, there is no production-quality universal build of python 2.4 at the moment. But that will change soon of course :-)
>>> Depending on what you want to get done either to me or Bob. Bob is
>>> the primary author of py2app and is the only one that can perform
>>> checkins on its repository, I wrote most of PyObjC.
>> Ideally, what I would like to get is a PyObjC installer (for at
>> least the stock Apple python) that installs universal binaries for
>> framework glue and includes a Universal-savvy py2app. It sounds
>> like that would involve both you and Bob. I can't tell which of the
>> two of you would benefit more from hardware, funding, or my time,
>Just because I'm the only one with commit access to py2app doesn't
>mean that I'm the only one who could do the work! I'd be happy to
>commit non-broken patches from anyone else. Personally I don't have
>a whole lot of time and hardware isn't the issue. The only thing I'm
>missing is access to 10.3, but you're doing widgets so you could care
>less of course.
I have access to required hardware, and slightly more time than you it seems. I'm still recovering from a sports injury and therefore have some time left to work on this.
I'll try to build universal binary versions of some PyObjC examples, that should at the very least give me a good indication of how much work is needed.