How did you choose the categories? What do the stats look like?

They look quite tailored to the benchmark suite. Would it possible to generalize them a bit?
Certain builtin classes, like zip and enumerate are worth checking for. But beyond that it is probably only worth categorizing into broader categories like: "iterator for underlying sequence", "computed iterator", "implemented in C", "implemented in Python".
That sort of thing.

I chose the categories by running the test suite and pyperformance with a printf at the end of _PySpecialization_ClassifyIterator until there stopped being so many prints. I can pare down the number of cases to ones we care about, but here are some of the results with all of those details:

• pyperformance FOR_ITER stats by benchmarks

• pyperformance FOR_ITER stats in total

• python -m test FOR_ITER stats

Ok, we can always change the categories later, if we need to.
In the meantime, this is useful information.

Results from python -m test, now with the new ascii string iterator:

I think this is useful information, so I'll go ahead and merge. As was said, we can always adjust things more later.