Summary

There's been some discussion how the "best" way to write if-clauses depending on a text parameter in combination with _api.check_in_list, e.g. #24486 (comment)

Proposed fix

There are three primary ways in the code base:

1. Explicit checking all arguments twice

_api.check_in_list(["foo", "bar"], param=param)
if param == "foo":
     ...
elif param == "bar":
    ...

2. Skip explicit checking of the last argument

_api.check_in_list(["foo", "bar"], param=param)
if param == "foo":
     ...
else: # "bar"
    ...

3. Use _api.check_in_list for generating an error message once we know that no option is matching.

if param == "foo":
     ...
elif param == "bar":
    ...
else:
    _api.check_in_list(["foo", "bar"], param=param)

Personally, I'm in favor of 2. Primarily as it skips a redundant check and that it is not possible to get 100% coverage in case 1. Also, it is often the case that the _api.check_in_list is done in another place in the code (like in the constructor), so option 3 may lead to a delayed error.

I know that @timhoffm prefers 1 as that is more likely to lead to errors if a case is missed in the if-part.

(In addition, this should really be called _api.check_in_tuple and tuples being used as this will have some performance benefits when it comes to compiled code. Parts of the code use tuples already though...)