Context

I often write utility "functions" as labels with parameters. Due to Python-style indentation I often forget to add a final return clause, as it looks obvious to me that the label ends with the indent.

This makes the script execution leak after the end of the label to the next label, which generally has at least one parameter, causing a runtime error:

TypeError: Missing required positional arguments: 'argument1'

Example

label say_hello_and_wait(duration):
    call _say_text_and_wait("hello", duration)
    # comment out to demonstrate bug
    # return

label say_goodbye(duration):
    call _say_text_and_wait("goodbye", duration)
    # comment out to demonstrate bug
    # return

label _say_text_and_wait(text, duration):
    "[text]"
    pause duration
    return

In this example, say_hello_and_wait will leak to say_goodbye where duration will not be defined, and say_goodbye will leak to _say_text_and_wait where text (and duration) is not defined.

Proposal

It would be nice if Renpy detected that we entered a parametered label directly from above (without a proper jump/call with parameters), so it can warn us about illegal parameter label "default" entrance. This seems the easiest way to warn the user against this case.

Alternatives

If this could be detected earlier via parsing before game start it would be even better, but I understand how difficult it can be to verify (because runtime code could enter branching differently depending on the context, therefore making it not obvious that all branches return, making the parser complain that a final return is missing even if reaching the end of the block was impossible at runtime).

Finally, Lint could be a more lightweight way to check such issues and only warning us if some code is suspicious (missing return statement) without causing a strict error. But in my case I hadn't even thought about linting before I fixed the issue manually, so YMMV - and again, parsing branches is not obvious.