bug/minor oddities

[blast-hardcheese]

Why

None of these changes impact production codepaths, so we've presumably been skating by without issue up until now, but let's make sure we don't run into trouble in the future.

What changed

• f"x: '{foo}' -> f"x: {repr(foo)}". If it's a string, it'll be a no-op. If it's not, it'll be an error.
• RiverNotType has never worked. Rip it out.
• Address incorrect init vs input bindings
• The result type of transport messages will be None in the case where the whole function also should return None, so remove some confusing metaprogramming about optional return statements.

Test plan

Describe what you did to test this change to a level of detail that allows your reviewer to test it