bug#48545: 28.0.50; `icomplete-vertical-mode` does not support the `group-function`

[Kévin Le Gouguec]

João Távora <joaotavora@gmail.com> writes:

>  FWIW (I didn't take the time to understand the ins and outs of the
>  completion system nor the changes that are being made; sorry about
>  that), 
>
> Don't beat yourself over it, it's not an easy or pleasent thing to understand.

(It's not for lack of trying either *cough*bug#40152*cough*)

>  with this tentative patch, Icomplete seems to honour
>  read-char-by-name-sort.  With icomplete enabled and
>  read-char-by-name-sort set to 'code, try
>
>      C-x 8 RET CLOCK FACE
>
>  - with the current master branch, the completions are sorted
>    alphabetically,
>  - with the tentative patch applied, the completions are sorted by code
>    point.
>
> Yes I think that's the idea. But do you think that's good or bad?

Apologies, my message was fairly clinical and lacked emphasis.
Icomplete heeding read-char-by-name-sort?  Yes please!  👍 for this
tentative patch.

>  This seems to compose well with completions-group AFAICT (see e.g. C-x 8
>  RET ROMAN).
>
>  This is good, right?

Yes, also good 👍