bug#54754: 28.1; Still requiring typing "yes" when read-short-answer is non-nil

[Lars Ingebrigtsen]

Damien Cassou <damien@cassou.me> writes:

> when killing an unsaved buffer, Emacs would ask me to type "yes" to
> express that I'm sure. Reading the documentation of `read-answer-short',
> I thought that setting this variable to t would allow me to type just
> "y" but it doesn't: I still have to type "yes".
>
> In the same documentation, I can read that if the variable's value is
> `auto' (which is the default but I changed it to t) then
> `use-short-answers' is used to decide if typing "y" is enough. I tried
> to set `use-short-answers' to t while keeping `read-answer-short' to t
> and it solved my problem.
>
> I suggest clarifying the documentation of `read-answer-short' to say
> that the value of `use-short-answers' might have an impact even if
> `read-answer-short' is t. I also suggest explaining when this is
> necessary.

Hm, yes, this is a bit confusing.

But as the documentation for the `read-short-answers' says -- that
variable only affects how the function `read-answer' prompts.
`read-answer' is not involved at all when killing an unsaved buffer --
that's `yes-and-no-p', which has the length of the answer controlled by
the `use-short-answers' variable.

But `use-short-answers' also affects `read-answer', which makes this
doubly confusing.

I've now clarified this in the Emacs 29 doc string, and also added links
from the yes-or-no-p doc string.

-- 
(domestic pets only, the antidote for overdose, milk.)
   bloggy blog: http://lars.ingebrigtsen.no