> I'm partial to
>     for i in 1...10: pass
> myself.

You have to handle step size != 1 in the loop.

I'm +0 on the whole idea of repurposing/creating syntax for range(). On the one hand, I think the syntactic sugar can be kind of handy, and runtime optimization of loop constructs is likely to be a bit better. OTOH, I doubt the presence of a call to range() vs. a fixed chunk of syntax is a big impediment to the PyPy folks and others interested in optimizing Python performance.