[Python-Dev] New encoding error in debug build

[Tim Peters]

This assert near the end of get_coding_spec() in tokenizer.c triggers when
running test_heapq in a debug build:

					assert(strlen(r) >= strlen(q));

It's a very new failure.  Note that heapq.py begins with the line

# -*- coding: Latin-1 -*-

I assume that's relevant, but Latin-1 is way beyond my personal experience
with strange character sets <wink>.