[RfC] Saving narrowing information within window configuration

[RfC] Saving narrowing information within window configuration

[Tobias C. Rittweiler]

Hi,

`(info "(Elisp)Window Configurations")' says:

       ``A "window configuration" records the entire layout of one frame--all
    windows, their sizes, which buffers they contain, _what part of each
    buffer is displayed_, and the values of point and the mark.  You can
    bring back an entire previous layout by restoring a window
    configuration previously saved.'' (emphasis mine)

As far as I can see (Emacs v22.0.91.1), narrowed regions are not
restored by saved window configurations. I propose to change that.

Comments?

  -T.

Re: [RfC] Saving narrowing information within window configuration

[David Kastrup]

"Tobias C. Rittweiler" <tcr@freebits.de> writes:

> Hi,
>
> `(info "(Elisp)Window Configurations")' says:
>
>        ``A "window configuration" records the entire layout of one frame--all
>     windows, their sizes, which buffers they contain, _what part of each
>     buffer is displayed_, and the values of point and the mark.  You can
>     bring back an entire previous layout by restoring a window
>     configuration previously saved.'' (emphasis mine)
>
> As far as I can see (Emacs v22.0.91.1), narrowed regions are not
> restored by saved window configurations. I propose to change that.
>
> Comments?

Don't like it.  This is not about what parts of a buffer are
displayed, but which parts are accessible.

-- 
David Kastrup

Re: [RfC] Saving narrowing information within window configuration

[Tobias C. Rittweiler]

David Kastrup <dak@gnu.org> writes:

> "Tobias C. Rittweiler" <tcr@freebits.de> writes:
>
> > As far as I can see (Emacs v22.0.91.1), narrowed regions are not
> > restored by saved window configurations. I propose to change that.
> >
> > Comments?
>
> Don't like it.  This is not about what parts of a buffer are
> displayed, but which parts are accessible.

I think window configurations are mostly used for saving and restoring
the state as seen by the user, aren't they? So restoring narrowing just
seems to be the right thing. 

(If restoring narrowing by default is a change that is too invasive, I'd
have nothing against making it optionally.)

  -T.

Re: [RfC] Saving narrowing information within window configuration

[David Kastrup]

"Tobias C. Rittweiler" <tcr@freebits.de> writes:

> David Kastrup <dak@gnu.org> writes:
>
>> "Tobias C. Rittweiler" <tcr@freebits.de> writes:
>>
>> > As far as I can see (Emacs v22.0.91.1), narrowed regions are not
>> > restored by saved window configurations. I propose to change that.
>> >
>> > Comments?
>>
>> Don't like it.  This is not about what parts of a buffer are
>> displayed, but which parts are accessible.
>
> I think window configurations are mostly used for saving and
> restoring the state as seen by the user, aren't they? So restoring
> narrowing just seems to be the right thing.

Narrowing is not a part of the window configuration.  Strictly
speaking, I already find that restoring point and mark is a mistake
here: window-point is the only thing that should be restored in each
window (if there is a currently active window, this will incidentally
also restore point, but that can't be avoided).

> (If restoring narrowing by default is a change that is too invasive,
> I'd have nothing against making it optionally.)

It is simply a mistake to do so.  Narrowing is not a feature of the
display, but of buffer accessibility.  You would not think of doing
"undo" until the state of the window configuration is reached, even
though buffer changes affected the state as seen by the user, would
you?

In my opinion, narrowing is of a similar degree.

-- 
David Kastrup

Re: [RfC] Saving narrowing information within window configuration

[Tobias C. Rittweiler]

David Kastrup <dak@gnu.org> writes:

> "Tobias C. Rittweiler" <tcr@freebits.de> writes:
>
> > I think window configurations are mostly used for saving and
> > restoring the state as seen by the user, aren't they? So restoring
> > narrowing just seems to be the right thing.
>
> Narrowing is not a part of the window configuration.  Strictly
> speaking, I already find that restoring point and mark is a mistake
> here: window-point is the only thing that should be restored in each
> window (if there is a currently active window, this will incidentally
> also restore point, but that can't be avoided).

(Indeed, now coming of think of it, I can recall being "bitten" by the
automatic restoration of poin. "Bitten" in quotes because other times
this snapshot of state behaviour was exactly what I wanted.)


> > (If restoring narrowing by default is a change that is too invasive,
> > I'd have nothing against making it optionally.)
>
> It is simply a mistake to do so.  Narrowing is not a feature of the
> display, but of buffer accessibility. [...]

I can understand that point of view. Thanks for the clarification.

Maybe there should be a more leight-weight version of window
configurations, and a more heavy-weight snapshot feature?

  -T.

Re: [RfC] Saving narrowing information within window configuration

[Richard Stallman]

	   ``A "window configuration" records the entire layout of one frame--all
	windows, their sizes, which buffers they contain, _what part of each
	buffer is displayed_, and the values of point and the mark.  You can
	bring back an entire previous layout by restoring a window
	configuration previously saved.'' (emphasis mine)

    As far as I can see (Emacs v22.0.91.1), narrowed regions are not
    restored by saved window configurations. I propose to change that.

I think you've misunderstood that text in the manual.
That's partly the fault of the text.
I will make it clearer.