Why are they generic in typeshed?

Why are they generic in typeshed?

DictReader has been generic for a while. DictWriter was made generic in python/typeshed#5366.

The classes are slightly complicated to type, so I am not sure the stubs are 100% correct at the moment (I have a PR open at the moment to improve them). But I think it is correct that they are both generic. Copying what I wrote in python/typeshed#7787 (comment):

DictReader seems pretty clearly like it should be generic -- the type that's returned from DictReader.__next__ depends very clearly on the type that you stick into DictReader.__init__.

In fact, DictReader seems like it could possibly be typed more correctly if it took another parameter: see python/typeshed#5994 (comment). (But that's irrelevant to the runtime; adding __class_getitem__ will allow it to take any number of parameters at runtime.)

I was confused about the implicit conversion to string in DictWriter, which made me wonder if DictWriter should be generic. But DictWriter does some runtime typechecking in the _dict_to_list method before it does the implicit conversion to string. So I think it makes sense for DictWriter to be generic as well (as it is in the stub currently).

I am not sure that they were designed this way. Why not return just dict[Any, Any] in DictReader.__next__? The type of the result depends not only on the type of field names (which can have different types), but on types and values of other parameters and on the content of the file.

The type of the result depends not only on the type of field names... but on types and values of other parameters and on the content of the file.

Can you give an example? If the stubs are incorrect, I would be interested in improving them.

The type of keys: the type of any item in fieldnames (str | float if fieldnames is None), or the type of restkey.
The type of values: str | float, list[str|float], or the type of restval.