bug#19267: 25.0.50; Enhance cycle-spacing with a state where only whitespace after point is deleted

[Tassilo Horn]

Lars Ingebrigtsen <larsi@gnus.org> writes:

>> I'd say just create a cycle-spacing-actions defcustom and use that in
>> the existing cycle-spacing instead of the hard-coded sequence.  And
>> bind M-SPC to cycle-spacing by default because it's simply superior
>> (a super-set) of just-one-space.
>
> It's two actions, and one which can't be done via Customize...

I don't understand.  What can't be done via Customize?  Users who want
back the old behavior could customize cycle-spacing-actions to
(just-one-space) and have the same behavior as if M-SPC was still bound
to just-one-space, no?  Or they can re-bind M-SPC back to just-one-space
(ok, that can't be done via customize).

>> I mean, how would people discover the possibility of customizing
>> cycle-spacing-actions if not by checking the info docs and/or NEWS
>> (where cycle-spacing is already mentioned as a more powerful
>> just-one-space alternative)?  And nobody would look in there if the
>> default behavior is just-one-space.  We need a disruptive change! :-)
>
> If we do this change, they'll find out about the variable by doing
> `C-h M-SPC'.  Eventually.

Yeah, eventually maybe.  My point is that if the behavior doesn't
change, most people won't check.  I mean, I don't `C-h k' bindings I
think I know.  I do that when I don't get what I expected where the
actual behavior may very well be better than what I expected.

Bye,
Tassilo