On 26 January 2014 11:57, Eric Snow wrote:

On Sat, Jan 25, 2014 at 6:28 PM, Antoine Pitrou wrote:

...

Is there a reason to have separate "importlib" (toplevel) and "importlib.util" namespaces?

As to why they are separate, you'll need to ask Brett. I believe it's meant to keep the top namespace as small as possible.

Correct - the top level namespace is meant for people to use normally for dynamic imports, reloading modules and clearing the internal caches when dynamically creating modules. By contrast, importlib.util is for when people are poking around at import system internals more directly, writing custom importers and loaders, replicating parts of the import system, etc. It's a "if you don't know why you might want these operations, you probably don't need them" kind of module, so we can keep the top level namespace simple and relatively easy to learn. find_loader()/find_spec() are both borderline as to which category they fall into, but as Eric noted, there was a dependency issue in this case which meant there were practical benefits to putting find_spec() in the "import specialist" category. Cheers, Nick. -- Nick Coghlan | ncoghlan@gmail.com | Brisbane, Australia