RE: ctl-x-map key binding conventions for new major/minor modes

[Tino Calancha]

On Fri, 12 May 2017, Drew Adams wrote:

> > (info "(elisp) Key Binding Conventions")
> > Do we need a recommendation in this node about bindings with
> > prefix 'C-x'?  There is no guidance in the node above about
> > this topic.  Just loading a new mode shouldn't shadow some
> > previous C-x bindings, right?
> >
> > For instance,
> > * timeclock.el suggests some C-x bindings in the comentary, that is,
> >    it doesn't define these bindings by default.
> > * dired-x.el has the user option `dired-bind-jump' to control if
> >    `dired-jump' must be bound or not to 'C-x C-j'.
> > * find-func.el has the autoloaded function `find-function-setup-keys',
> >    which binds some commands into `ctl-x-map'.
>
> Are you asking whether code that is distributed with Emacs
> should bind keys with prefix `C-x'?  That's usually discussed
> and decided case by case, e.g. on this list.
>
> Are you asking whether we should make particular recommendations
> in this regard for 3rd-party code?  If so, I think not.
I am asking for both.
For example, magit.el suggests the binding 'C-x g' for `magit-status'.  It
doesn't establish the binding automatically, though.

Imagine one user sets that binding in her .emacs file.  Now she loads
a lib `foo.el' which automatically sets a global binding
'C-x g' to `foo-whatever'.

In the examples above (timeclock, dired-x, etc) the user can easily
control such key binding collisions.  It's something good for people
loading many libs.
We might add a recommendation in that info node about 'C-x ...'
as we do with 'C-c ...'.

> Loading a 3rd-party library is optional.  If a library binds keys
> just by loading it then that should at least be made clear in the
> library doc/commentary.  A user can always override any bindings
> that might be made by a library, but s?he should be aware of what
> happens when the library is loaded or a mode is turned on.
>
> Have you noticed an actual problem in this regard?
No yet.