bug#21465: [PATCH] CC-modes hierarchy

[Alan Mackenzie <acm@muc.de>]

Hello, Lars.

On Mon, Sep 07, 2020 at 18:52:00 +0200, Lars Ingebrigtsen wrote:
> Stefan Monnier <monnier@iro.umontreal.ca> writes:

> > You already have this layer, with c-mode-common-hook and
> > c-mode-base-map.  My patch just makes this layering more standard and
> > then takes advantage of it to share some code.

> > Here's another version of the patch with the following changes:
> > - Got rid of c-derivative-mode (which I still use at other places in my
> >   local patches, but for now, I won't push for it any more).
> > - Keep c-make-inherited-keymap for backward compatibility.
> > - Fix the "c-mode-hook is run twice" bug, using the same hack as is used
> >   by define-globalized-minor-mode.
> > The third point also makes the c-mode-common "extra layer" more useful
> > since that hack is added in there.

> > Any objections to this version?

> This was four years ago, and there apparently wasn't any response here?
> Alan?

Five years on, I'm still not keen on this change.  It is a large patch.
It would fragment the currently coherent functions c-mode, c++-mode and
so on.  It would extend the stack of modes by inserting a new,
meaningless mode, c-mode-common; this might induce hackers to try to
derive directly from c-mode-common (which wouldn't work) rather than one
of the genuine CC Mode modes.  It would magnify differences between the
Emacs version of CC Mode and the upstream version, or cause a lot of
work integrating the patch into the latter.

Of the above, I think it is the artificial introduction of c-mode-common
I'm least keen on.

I would be happiest if this bug could just be closed with "won't fix".

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

-- 
Alan Mackenzie (Nuremberg, Germany).