Re: master e4df903f6b5 2/2: ; Fix last change

> From: Richard Stallman <[email protected]>
> Cc: [email protected], [email protected]
> Date: Mon, 03 Aug 2026 00:05:56 -0400
> 
>   > I like this idea. We could add the following macro to ERT:
> 
>   > --8<---------------cut here---------------start------------->8---
>   > (defmacro ert-with-forced-function-value (name value &rest body)
>   >   "Bind function NAME to return always VALUE, and evaluate BODY."
>   >   (declare (indent 2) (debug (functionp sexp body)))
>   >   `(cl-letf (((symbol-function ',name) (lambda (&rest _) ,value)))
>   >      ,@body))
>   > --8<---------------cut here---------------end--------------->8---
> 
> What I had in mind is to use a special mechanism to force a certain function's
> value to VALUE, without ever altering a symbol's function definition.
> 
> It would be evidently unclean to use this mechanism in ordinary code,
> and the point is that we would not accept code to do so.

The function's mock might need some non-trivial logic or side effects,
in which case simply forcing the value will not be enough.

What we need is a solid infrastructure for mocking a function.  No
good test suite can do without something like that, so we should have
it as well.  IME, it is never as trivial as forcing the value.