Re: feature/icomplete-vertical

[Ergus]

On Fri, Oct 02, 2020 at 12:45:41PM +0800, jixiuf wrote:
> I have tried your branch and it worked perfectly.
>
> 1. And Is there any plan to implements  a completion style like helm or  
> orderless(https://github.com/oantolin/orderless 
> <https://github.com/oantolin/orderless>).
> Divides the pattern into space-separated components, and matches candidates 
> that match all of the components in any order.
Not that I am aware of. In principle we already have some completions
styles you could try...

C-h v: completion-styles-alist

AFAIR flex was added recently and it is similar to what you mention
(just a bit different requiring the order).

It is pretty straightforward to create and customize new completions and
differently to icomplete-vertical; orderless does not need hacks (like
advices, minibuffer resize or extra hooks) to do it's work fine. So I
don't see a need to add more completions to vanilla (I am not opposed to
that either; specially if the code is simple).

If you want more completion styles you could make a feature request
and/or contact oantolin to add part of his code to vanilla OR at least
add ordeless to elpa to make the package more "official".

> 2. And Is there any plan to implements async-completing-read for icomplete ,
> one of my use cases involve an external command to generate the list of 
> candidates,
> and that command may take some time to run( like conusel-rg).

Not from me ATM. I just implemented this in the core to avoid the need
of some hacks in icomplete-vertical. But I don't want to add excessive
complexity to icomplete. To do that I support adding ivy+counsel to
vanilla as mentioned in another thread some weeks ago.

> Hope those feature can be implemented in emacs core.    
>
>
> > BTW: If someone could give a look to the completions-highlight feature
> > branch too and make recommendations, report issues?
> >
> > Best,
> > Ergus