Should we modify lib2to3?

[liamcurry]

Since we are already bundling lib2to3 with this package, do you think we should modify it, or leave it as is? There's a lot of code in there that could use some cleaning up. Also, if we're looking to increase performance then we'll probably have to do this anyway.

[dustinmm80]

I'd say since we're vendorizing it, we might as well improve it. If it was being continually released, we would have to worry about compatibility, but since this lib isn't changing on 2.x I'd say we can fix it up.

[myint]

If necessary, then sure. Though we should keep in mind that things like python-modernize manage to do everything they need with the default lib2to3 (without vendorizing/modifying). And it even works on Python 2.6, 2.7, 3.2, and 3.3. (See the unwanted_fixes code for example.)

[liamcurry]

@myint True, but the goals of these projects seems to be different. python-modernize and lib2to3 have a large emphasis on automatic refactoring of existing code bases, whereas py3kwarn is basically just a syntax checker used to help developers write forward-compatible code.

[myint]

Right. I only mention python-modernize so you guys know that you can get pretty far (in terms of both functionality and Python compatibility) with the default lib2to3.