bug#24706: 26.0.50; Minor mode functions should do strict argument type checking

[Drew Adams]

>> Why would you change the behavior, instead of just clarifying
>> the doc?
>
> I generally prefer the behavior to be as strict as possible.
> Consider (foo-mode 'disable)  If you read such code, do you
> assume that this enables foo-mode?

If I understand the function, then yes.  But I agree that the
doc could make the function behavior clearer.

That example is akin to this:

(deconst two 42)
(+ 3 two)

There are plenty of ways to write code that is perverse in
terms of readability (whether or not it has the right behavior).

> However, in this case I guess it's too late, and fixing the
> documentation is indeed more appropriate.

It would be a backward-incompatible change.  (And I do not
see any advantage to it.  But others might disagree with
me and agree with you.)