bug#48493: 28.0.50; quit-window doesn't work

[pillule]

Eli Zaretskii <eliz@gnu.org> writes:

> > From: pillule <pillule@riseup.net>
> > Date: Tue, 08 Jun 2021 01:23:11 +0200
> > Cc: pillule <pillule@riseup.net>, Sujith Manoharan 
> > <sujith.wall@gmail.com>,
> >  48493@debbugs.gnu.org
> >
> > > We have to document it in the Elisp manual.
> >
> > Here another draft with the info manual changes:
>
> Thanks.  Once again, I leave it to Martin and others to comment 
> on
> most of the essence of the patch, and provide here only few 
> minor
> nits:
>
> > * doc/lispref/windows.texi
> >   (Buffers and Windows): mention the exception of side windows 
> >   and
> > add a reference.
> >   (Buffer Display Action Alists): explicit that
> > `display-buffer-in-side-window' is dedicating to side by 
> > default.
> >   (Dedicated Windows): add the case (4) and explicit its 
> >   meaning,
> > add a reference.
> >   (Displaying Buffers in Side Windows): move the
> > switch-to-(prev|next)-buffer paragraph into a new item to 
> > emphasize
> > the special meaning of dedication for side windows.
>
> Again, this log message is not formatted according to our rules. 
> It
> should look like this:
>
> * doc/lispref/windows.texi (Buffers and Windows): Mention the
> exception of side windows and add a reference.
> (Buffer Display Action Alists): Say explicitly that
> 'display-buffer-in-side-window' is dedicating to side by 
> default.
> (Dedicated Windows): Add case (4) and explain its meaning, add
> a reference.
> (Displaying Buffers in Side Windows): Move the paragraph about
> 'switch-to-(prev|next)-buffer' into a new item to emphasize the
> special meaning of dedication for side windows.
>
> Note that I also fixed the wording a bit, and that our 
> conventions for
> quoting in log entries is 'like this'.
>
> >  The replacement buffer in each window is chosen via
> > -@code{switch-to-prev-buffer} (@pxref{Window History}).  Any 
> > dedicated
> > -window displaying @var{buffer-or-name} is deleted if possible
> > -(@pxref{Dedicated Windows}).  If such a window is the only 
> > window on its
> > -frame and there are other frames on the same terminal, the 
> > frame is
> > -deleted as well.  If the dedicated window is the only window 
> > on the only
> > -frame on its terminal, the buffer is replaced anyway.
> > +@code{switch-to-prev-buffer} (@pxref{Window History}).  With 
> > the
> > +exception of side windows, any dedicated window displaying
>    ^^^^^^^^^^^^^^^^^^^^^^^^^
> Here' it is quite important that the reader understands what are 
> "side
> windows", otherwise he/she will not understand the exception.
> However, "side window" was not yet described in the manual by 
> this
> point, and its description is not in this section.  In these 
> cases,
> always include a cross-reference to where the term is described, 
> like
> this:
>
>   With the exception of side windows (@pxref{Side Windows}), any 
>   ...
>
> >     In particular, @code{delete-windows-on} (@pxref{Deleting 
> >     Windows})
> > -handles case (2) by deleting the associated frame and case (3) 
> > by
> > -showing another buffer in that frame's only window.  The 
> > function
> > +handles case (2) by deleting the associated frame and case (3) 
> > and (4)
>                                                          ^^^^
> "cases", plural.

Here note that I suppressed the comment

@c FIXME: Does replace-buffer-in-windows _delete_ a window in case 
(1)?

Because yes, it is the case.

> > +@item dedicated
> > +The dedicated flag is not reserved to this function but have a
>                                                       ^    ^^^^
> A comma is missing before "but".  Also, please use "has" instead 
> of
> "have".
>
> > +slightly different meaning for side windows.  They receive it 
> > upon
> > +creation with the value @code{side}; it serves to prevent
> > +@code{display-buffer} to uses these windows with others action
>                          ^^^^^^^
> "to use"
>
> > +functions, and it persists across invocations of 
> > @code{quit-window},
> > +@code{kill-buffer}, @code{previous-buffer} and 
> > @code{next-buffer}
> > +(@pxref{note Window History}).  In particular, these commands 
> > will
>            ^^^^^^^^^^^^^^^^^^^
> No need for "note" here.

Thanks.  Got them.

martin rudalics <rudalics@gmx.at> writes:

> Thanks for the patches.  Please follow ELi's suggestions first, 
> I'll
> then comment on the manual changes.
>
> -       (current (eq (window-buffer window) (current-buffer))))
> +       (current (eq (window-buffer window) (current-buffer)))
> +        (dedicated (window-dedicated-p window)))
>      (set-window-buffer window buffer)
> +    ;; restore the dedicated side flag
> +    (when (eq dedicated 'side)
> +      (set-window-dedicated-p window 'side))
>
> Here you could use the 'dedicated-side' solution you used below 
> in
> `replace-buffer-in-windows'.

ok

> +    (or ;; first try to delete dedicated windows that are not 
> side windows
> +     (and dedicated (not (eq dedicated 'side))
> +          (window--delete window 'dedicated (eq bury-or-kill 
> 'kill)))
> +     (cond
> +       ((and (not prev-buffer)
> +             (eq (nth 1 quit-restore) 'tab)
> +             (eq (nth 3 quit-restore) buffer))
> +        (tab-bar-close-tab)
> +        ;; If the previously selected window is still alive, 
> select it.
> +        (when (window-live-p (nth 2 quit-restore))
> +          (select-window (nth 2 quit-restore))))
>
> What's wrong with putting the first disjunct into the 
> conditional as in
> the below?  In general, always try to avoid larger indentation 
> changes -
> they can make forensics cumbersome while bisecting.
>
>      (cond
>       ;; First try to delete dedicated windows that are not side 
>       windows
>       ((and dedicated (not (eq dedicated 'side))
>             (window--delete window 'dedicated (eq bury-or-kill 
>             'kill))))
>       ((and (not prev-buffer)
>              (eq (nth 1 quit-restore) 'tab)
>              (eq (nth 3 quit-restore) buffer))

The difference is a window dedicated with flag t may not be 
deletable, and in this case, we want it
to pass through the others conditionals branch of 
quit-restore-window, so it can try to use the
'quit-restore parameter, close the tab or to fallback in t, etc.
Explaining it makes me thing I could use 'window-deletable-p' in 
its conditional and ...
I guess, problem solved

In this revision I also restored the use of (select-window (nth 2 
quit-restore)) on the branch t.

> BTW, how's your paperwork process proceeding?

I am currently exchanging with the clerk to complete it.

0003-Improve-handling-of-side-dedicated-flag.patch
Description: respect the indentation

--