Re: Edebug corrupting point in buffers.

[Alan Mackenzie]

Hello, Eli.

On Thu, Nov 03, 2022 at 20:15:08 +0200, Eli Zaretskii wrote:
> > Date: Thu, 3 Nov 2022 18:07:16 +0000
> > Cc: emacs-devel@gnu.org
> > From: Alan Mackenzie <acm@muc.de>

> > > The node you added is very short, barely a dozen lines.  It makes
> > > little sense to have it separate from where edebug-save-windows is
> > > described.  So I think you should move it there.  The location of the
> > > node inside the manual's hierarchy is much less important than to have
> > > the information pertaining to edebug-save-windows in a single place,
> > > because no one reads the ELisp reference manual in its entirety.  The
> > > only thing we need to facilitate people finding this place is add good
> > > index entries there.

> > So you're proposing leaving the "The outside context" node incomplete,
> > according to its clearly defined purpose, and therefore wrong?  Why?

> If you want, you can add a short sentence there about the issue, with
> a cross-reference to where the issue is described in full.

"There"?  There is no suitable place to put such a link, other than my
new node.  Such a strategy would unbalance "The Outside Context" by
having most of its contents in subsubsections, and the bit about point
corruption at the other end of a link, in some random page.

As a matter of interest, one of the other nodes under "The Outside
Context", namely "Checking Whether to Stop" has just 13 lines.

> This is how we organize our manuals: when some topic could be relevant
> to more than one situation, we describe it in full in one place, and
> have short references in all the others.

We should describe it in the PRIMARY relevant place.

> > Remember, this patch is not about edebug-save-windows.  It's about point
> > getting corrupted.

> The index entries and the cross-references should solve this.  And the
> issue _is_ related to edebug-save-windows ....

It is only tangentially related to edebug-save-windows.  It is about
point getting corrupted.  An angry victim of this bug should be be able
to find the description by searching for "corrupt".

> .... and to the other similar option described in the same node.  So
> having all of this info there makes the description more
> comprehensive.

Yes, stuff about options belongs in the "Options" page.  Stuff about
point getting corrupted does not, except at the other end of a link.

-- 
Alan Mackenzie (Nuremberg, Germany).