alternate clang driver

Imagine you are an Engineering Manager at a large corporation, considering adopting clang. You probably don't want to require your C++ developers to know Python, just so they can build their code.

Also, concerning Reed's claim that all solutions will converge to his, unless he's got a mathematical proof, I'm not convinced. Just because you were unable to discover a "better" solution, doesn't mean someone else won't. Or rather, because you failed to accept a simpler solution, doesn't mean that others wouldn't be okay with that.

Ultimately

Sorry, stupid phone sent prematurely.

Ultimately, anyone can write whatever scripts they want. I doubt most people would consider that an acceptable solution for clang itself.