> @Antoine, I agree that the tests for copy should be a proper unit test;
> that seems orthogonal to this patch though :)

Not really, since Lib/test/test_copy.py exists and contains tests for
deepcopy; you should add the new test there.

> I don't have a checkout of 3 at the moment, but do you think the test
> failure on 3 is shallow or deep?

What I mean is that Lib/copy.py fails without your changes under py3k.