[Distutils] [tg-trunk] Re: [ANN] EggFreezer

Diez B. Roggisch deets at web.de
Thu Aug 7 14:01:58 CEST 2008

On Thursday 07 August 2008 00:48:06 Christopher Arndt wrote:
> zooko schrieb:
> > On Aug 5, 2008, at 10:07 AM, Ian Bicking wrote:
> >> No... which makes binary eggs unusable on Linux.  I feel like there
> >> was something else that made binary packages on a Mac unreliable, but
> >> I can't remember.
> >
> > Perhaps it was http://bugs.python.org/setuptools/issue19 (Will
> > setuptools on Mac Python accept fat eggs?).
> Yes, that's what I meant. setuptools on Mac OS X 10.5 does not accept
> 'fat' binary eggs build on 10.4.

I understood your post as if fat-binaries as whole weren't supported under 
10.5, which they are.

> Diez. R. Roggisch schrieb:
>  > Why shouldn't they?
> Or can you install all of the following on Leopard?
> http://files.turbogears.org/eggs/simplejson-1.9.1-py2.5-macosx-10.3-fat.egg
> http://files.turbogears.org/eggs/RuleDispatch-0.5a0.dev_r2306-py2.5-macosx-
> http://files.turbogears.org/eggs/PyProtocols-1.0a0dev_r2302-py2.5-macosx-10
> http://files.turbogears.org/eggs/Cheetah-2.0rc8-py2.5-macosx-10.3-fat.egg

Dunno, most probably not. But there is no problem making 10.3, 10.4 and 10.5 
available, isn't there? Inconvenient - for sure. 

But I understood the linux-problem to be that there actually is no way to 
determine if UCS2 or UCS4 is used, or at least not supported to generate 
different eggs for that - so you end up with one egg for two platforms, and 
on which it works is essentially luck.


More information about the Distutils-SIG mailing list