I'd rather not wait. If packaging is not able to reliably build Python itself, we shouldn't release Python 3.3 until it is (or withdraw packaging from the standard library altogether).
> I'd rather not wait. If packaging is not able to reliably build Python > itself, we shouldn't release Python 3.3 until it is (or withdraw > packaging from the standard library altogether). Indeed, this is exactly the reason I entered this enhancement request. Eating our own dogfood is a good way of detecting issues :)
I'm not willing to hold up 3.3 for this. On the other hand, it should not be done in the beta stage. And I don't assume it's a trivial matter (Éric?), so I'm inclined to move it to 3.4.
At the moment, it seems that this just won't happen. If a new attempt is made to integrate packaging with Python, this may or may not become an issue (e.g. whoever is doing the integration the next time might adapt setup.py from the beginning). So closing this as "won't fix".