confused with so many python package locations for imports

Ben Finney ben+python at benfinney.id.au
Mon Apr 27 00:28:18 CEST 2009


Scott David Daniels <Scott.Daniels at Acm.Org> writes:

> Ben Finney wrote:
> > Thank you for the correction. Please mentally substitute into my
> > argument the files that *are* compiled C extension modules; it
> > remains otherwise unchanged.
> 
> But the compiled C extension modules are called ".pyd" files and
> they are built separately from the execution time and thus do not
> suffer the faults of which you complain.

I didn't discuss build versus execution time at all.

My point was, for those who are still distracted: Whatever they're
named, whenever they're built, the architecture-dependent executable
module files are by default located and searched for in the same place
as architecture-independent executable module files. That is a mismatch
between Python and the FHS.

One which is being actively addressed, if I understand some enthusiastic
reports from PyCon 2009 correctly; but nevertheless, one which impacts
the ideal the original poster is hoping for.

-- 
 \        “I saw a sign: ‘Rest Area 25 Miles’. That's pretty big. Some |
  `\                      people must be really tired.” —Steven Wright |
_o__)                                                                  |
Ben Finney



More information about the Python-list mailing list