On 21 Nov, 2007, at 20:59, Ronald Oussoren wrote:
Does anyone know if there is a good reason for Python running on 10.5 to reject binaries built on 10.4, and if not what would be required to make this version check pass?This seems to be caused by distutils.util.get_platform() returning the wrong value (macosx-10.5-i386 instead of macosx-10.5-fat). This is caused by what seems to be a rather dumb bug in that function:| if (release + '.') < '10.4.' and \| get_config_vars().get('UNIVERSALSDK', '').strip(): | # The universal build will build fat binaries, but not on| # systems before 10.4 | machine = 'fat'The first part of the test seems wrong to me and causes machine not to be set to 'fat' on 10.5.
BTW. The first part of the fix is changing '<' to '>=' in the first test. That doesn't fix the issue though, the config/Makefile in Apple's Python.framework isn't configured for building universal binaries.
And to make matters even worse: I'm pretty sure that setuptools used to know that 'fat' builds are compatible with 'i386' and 'ppc' architectures (at least on OSX), but that code no longer seems to be there.
I'll see if I can check in a fix for the bogus if-statement tomorrow morning (both in the trunk and the 2.5 branch), that way python 2.5.2 will at least behave correctly. I don't know if have time to work on a patch for setuptools though.
I'll have to contact PJE about universal builds vs. setuptools anyway (through distutils-sig) to discuss how to deal with 4-way universal eggs.
Ronald
smime.p7s
Description: S/MIME cryptographic signature
_______________________________________________ Pythonmac-SIG maillist - [email protected] http://mail.python.org/mailman/listinfo/pythonmac-sig
