Optimization of str.format() for cases with unicode, long, and float

Optimization of str.format() for cases with unicode, long, and float
arguments.  This gives about 30% speed improvement for the simplest
(but most common) cases.  This patch skips the __format__ dispatch, and
also avoids creating an object to hold the format_spec.

Unfortunately, backporting this to 2.6 is going to be more challenging
due to str/unicode issues with format_spec.  I'll work on that next.

Then I'll spend some time profiling and see what that tells me.