Minor update to the PEP

[lysnikolaou]

I did a first pass and I have some comments (I still need to do another two three passes for a full-on review):

• I'm not sure about the types module. It seems that it mostly provides types that are the result of calling type on some objects. It would be the first exported symbol that does not with the suffix Type AFAIU.
• We should add to the specification of the format_spec that in case there's interpolations, they're eagerly evaluated.
• I'm not sure about the comment that the debug = symbol behaves differently to f-strings. F-strings actually behave the exact same way. If someone ast.parses an f-string, they'll get an AST tree very similar to our Template.args.

[davepeck]

We should add to the specification of the format_spec that in case there's interpolations, they're eagerly evaluated.

@lysnikolaou Good catch. Just added a change for this:

Format specifications in f-strings can themselves contain interpolations. This
is permitted in template strings as well; format_spec is set to the eagerly
evaluated result:

value = 42
precision = 2
template = t"Value: {value:.{precision}f}"
assert template.args[1].format_spec == ".2f"