[Python-Dev] Implementing PEP 382, Namespace Packages

[P.J. Eby]

At 01:19 PM 5/31/2010 -0700, Brett Cannon wrote:
>But as long as whatever mechanism gets exposed allows people to work
>from a module name that will be enough. The path connection is not
>required as load_module is the end-all-be-all method. If we have a
>similar API added for .pth files that works off of module names then
>those loaders that don't want to work from file paths don't have to.

Right - that's why I suggested that a high-level request like 
get_pth_contents() would give the implementer the most 
flexibility.  Then they don't have to fake a filesystem if they don't 
actually work that way.

For example, a database that maps module names to code objects has no 
need for paths at all, and could just return either ['*'] or None 
depending on whether the package was marked as a namespace package in 
the database...  without needing to fake up the existence of a .pth 
file in a virtual file system.

(Of course, since lots of implementations *do* use filesystem-like 
backends, giving them some utility functions they can use to 
implement the API on top of filesystem operations gives us the best 
of both worlds.)