On Sat, Mar 1, 2014 at 4:53 PM, ian o firstname.lastname@example.org wrote:
A bridge like this would allow us to move now, and that would allow pressure on the module providers.
Think about what the bridge really means. It means that you move now, yes, but it also means that the existing module is working fine. That means there is *no* pressure on the module provider. The pain of maintaining the bridge is all yours. You want to put pressure on the module provider? *Don't* have the bridge. Then you have an incentive to get that module ported, because then you could migrate. Maybe that means you lean on the developers; maybe that means you run the code through 2to3 and then start dealing with test suite failures, and then submit the code back to them and say "Here's Python 3.3 support, these are the downsides, will you accept it?". That's the sort of pressure that's likely to work. Saying "Hey, we can use your module in Python 3 now, thanks to a Python-in-C-in-Python bridge" will evoke the response "Okay fine, no problem then!".