completing-read, empty collection, require match

completing-read, empty collection, require match

[Qiantan Hong]

(completing-read "test: " nil nil t)
will pop up a prompt in mini buffer, pretend asking you for something
and then inevitably fail.

Is there some rationale behind the current behavior?
Failing immediately seems strictly more efficient.

Best,
Qiantan

Re: completing-read, empty collection, require match

[Stefan Monnier]

Qiantan Hong [2021-12-04 14:30:09] wrote:
> (completing-read "test: " nil nil t)
> will pop up a prompt in mini buffer, pretend asking you for something
> and then inevitably fail.
>
> Is there some rationale behind the current behavior?
> Failing immediately seems strictly more efficient.

Yes, it would be more "efficient" in this case if `completing-read`
catches the situation and signals an error, but I don't think such an
optimization is worthwhile: the better optimization is to not make such
a call at all.


        Stefan

Re: completing-read, empty collection, require match

[Thierry Volpiatto]

Qiantan Hong <qthong@stanford.edu> writes:

> (completing-read "test: " nil nil t)
> will pop up a prompt in mini buffer, pretend asking you for something
> and then inevitably fail.
>
> Is there some rationale behind the current behavior?
> Failing immediately seems strictly more efficient.

Helm is now quitting when collection is empty and require-match is
strict (i.e. `t').

    (completing-read "test: " nil nil t)
=> Quit.
    (completing-read "test: " nil nil)
=> Returns input.
    (completing-read "test: " nil nil 'confirm)
=> Returns input after confirmation (press RET twice).

-- 
Thierry