[lxml-dev] lxml binary egg redux

[Stefan Behnel]

Stephan Eletzhofer wrote:
> Am 12.07.2009 um 21:29 schrieb Stefan Behnel:
>> Stephan Eletzhofer wrote:
>>> I've built 2.4, 2.5 and 2.6 versions of LXML 2.2.2 and uploaded them.
>>
>> Any reason why the 2.5 egg is so much bigger than the others?
> 
> Hmm:
> 
> $ ls -lah lib.macosx-10.5-i386-2.?/lxml/*.so
> -rwxr-xr-x 1 seletz staff 3,0M 12. Jul 11:09
> lib.macosx-10.5-i386-2.4/lxml/etree.so
> -rwxr-xr-x 1 seletz staff 1,7M 12. Jul 11:09
> lib.macosx-10.5-i386-2.4/lxml/objectify.so
> -rwxr-xr-x 1 seletz staff 5,5M 12. Jul 21:15
> lib.macosx-10.5-i386-2.5/lxml/etree.so
> -rwxr-xr-x 1 seletz staff 3,0M 12. Jul 21:16
> lib.macosx-10.5-i386-2.5/lxml/objectify.so
> -rwxr-xr-x 1 seletz staff 2,8M 12. Jul 21:20
> lib.macosx-10.5-i386-2.6/lxml/etree.so
> -rwxr-xr-x 1 seletz staff 1,5M 12. Jul 21:20
> lib.macosx-10.5-i386-2.6/lxml/objectify.so
> 
> I've built the 2.5 version using the Apple-supplied python 2.5, the
> other ones using the
> macports variants. I've no idea about the sizes, really. :p

I guess the macports builds are platform specific, while the Apple-builds
are fat eggs, i.e. they would work on both PPC and x86.

Given that Apple's platform Python is 2.5, I guess it's ok to provide the
other eggs only for macports. And I also guess that 'fatness' isn't really
a future-proof requirement, either...

Anyway, thanks for the contribution!

Stefan