Re: [PATCH] `completing-read`: Add `group-function` support to completion metadata (REVISED PATCH VERSION 2)

[Eli Zaretskii <eliz@gnu.org>]

> From: Daniel Mendler <mail@daniel-mendler.de>
> Date: Sun, 2 May 2021 02:43:20 +0200
> Cc: Gregory Heytings <gregory@heytings.org>, Dmitry Gutov <dgutov@yandex.ru>,
>  Stefan Monnier <monnier@iro.umontreal.ca>,
>  "emacs-devel@gnu.org" <emacs-devel@gnu.org>
> 
> > Maybe this choice could be handled by an option?
> 
> Of course, we can add options to allow fine tuning of every behavior.
> While I like the tuneability of Emacs I am not sure if it is advised to
> make every tiny bit configurable. The way I wrote the current patch
> allows the addition of custom `completions--insert-*` functions.
> Furthermore there is always the advice mechanism, the possibility to
> override functions etc, so I don't feel the need to add configurations
> for every detail.

We don't provide knobs for every behavior, indeed.  But where the
"right" behavior is a matter of personal preferences, and there are
large enough groups of people who may want either of the possible
behaviors, offering an option is TRT.  Advice is not a valid
replacement for a user option, because writing an advice is orders of
magnitude harder than flipping an option, and requires the user to be
proficient in ELisp.