Re: [Emacs-diffs] emacs-25 040e0d6: Fix 'toggle-save-place'

[Eli Zaretskii <eliz@gnu.org>]

> From: Stefan Monnier <monnier@iro.umontreal.ca>
> Cc: emacs-devel@gnu.org
> Date: Thu, 25 Feb 2016 12:05:06 -0500
> 
> > It makes no sense not to autoload it, because then there's no way for
> > users to turn save-place in a single buffer.  Previously, users needed
> > to load saveplace.el to have the feature, so an autoload was not
> > needed.
> 
> But that still works "as well as before"
> (i.e. as long as they (require 'saveplace)).

Except that I just told them in NEWS not to...

> > I have nothing against having a local mode, but I don't have time to
> > work on this now, so patches are very welcome.
> 
> I think I have something better, but I think some of the changes are too
> late for 25.1, so I suggest we install the patch below (which against
> forces users to use (require 'saveplace) before calling
> toggle-save-place, like before: it's not ideal, but it's not
> a regression).

Fine with me, thanks.

> -;;;###autoload

I'd like to keep the autoload, otherwise there's an asymmetry: the
global mode is autoloaded, the local one isn't.

>  (defun toggle-save-place (&optional parg) ;FIXME: save-place-local-mode!
>    "Toggle whether to save your place in this file between sessions.
>  If this mode is enabled, point is recorded when you kill the buffer
> @@ -167,17 +166,21 @@ toggle-save-place
>                           (> (prefix-numeric-value parg) 0)
>                         (not save-place)))
>      (cond
> -     (save-place
> +     (save-place-mode

Shouldn't we also replace save-place above this fragment?

> +      ;; FIXME: We should remove the hooks, but only in case this was the last
> +      ;; buffer using save-place-local-mode.  Is it worth the trouble, tho?

No, I don't think it's worth the trouble.  Chances are, the user will
promptly enable the mode in some other buffer.