bug#12708: 24.1; M-x display-time-world "q" close window

[martin rudalics <rudalics@gmx.at>]

 >>> (defun occur-mode-goto-occurrence-kill-buffer ()
 >>>   "Go to the occurrence the current line describes, and kill the Occur buffer."
 >>>   (interactive)
 >>>   (let ((buf (current-buffer))
 >>>         (pos (occur-mode-find-occurrence)))
 >>>     (switch-to-buffer-other-window (marker-buffer pos))
 >>>     (goto-char pos)
 >>>     (kill-buffer buf)
 >>>     (run-hooks 'occur-mode-find-occurrence-hook)))
 >> Why don't you quit the *Occur* buffer window first and do a simple
 >> `pop-to-buffer' afterwards?
 >
 > I tried it to be closer to the workflow of
 > `occur-mode-goto-occurrence-other-window'.
 > But your recipe would work as well.

For <= two windows per frame users there's a difference: If the *Occur*
buffer appeared in a new window, the other window would be reused with
your approach.  If we delete the *Occur* window first, the other window
would be split.

 >> BTW, I find bindings like
 >>
 >>     (define-key map "\C-c\C-c" 'occur-mode-goto-occurrence)
 >>     (define-key map "\C-m" 'occur-mode-goto-occurrence)
 >>     (define-key map "o" 'occur-mode-goto-occurrence-other-window)
 >>
 >> very confusing.
 >
 > You might find it more confusing when I tell you that in ~/.emacs
 > I bound `occur-mode-goto-occurrence-kill-buffer' to C-RET ;-)

Not really.  I use M-RET for going to the tag around `point'.  It has
some kind of "come on, get there ..." connotation for me.

 > Shouldn't `display-buffer-at-bottom' create a window with the width of
 > the frame's full width?  In this case, left/center/right are not needed.

I'm not sure.  People with two side-by-side windows would probably like
to split-below their window on the right for calendar-like things that
take fairly few columns only.  We need some framework for specifying
such behavior in a user-friendly way.  Well, they could use side-windows
already ...

martin