Re: highlight current window/modeline after switching to it (solved)

[Clemens Fischer]

* 2003-11-18 Kevin Rodgers:

>> ;(defadvice other-window (around other-window-flash activate)
>> ;  (interactive "p")
>> ;  (setq display-time-string-forms
>> ;        '( day "/" month "/" (substring year -2)
>> ;               " " 24-hours ":" minutes
>> ;               (if time-zone " (") time-zone (if time-zone ")")))
>> ;  ad-do-it
>> ;  (setq display-time-string-forms
>> ;        '( "* " day "/" month "/" (substring year -2)
>> ;           " " 24-hours ":" minutes
>> ;           (if time-zone " (") time-zone (if time-zone ")")))
>> ;  (force-mode-line-update t)
>> ;)
>> this version leaves me with the little `*' i wanted to signify the
>> active window, but in all and everyone of them:  no help.
>
> Try making it buffer local, and simplify your code at the same time:
>
> (set (make-local-variable 'display-time-string-forms)
>       (default-value 'display-time-string-forms))
> ad-do-it
> (set (make-local-variable 'display-time-string-forms)
>       (cons "* " (default-value 'display-time-string-forms)))

your proposal(s) looked entirely logical, but i couldn't get them to
work.  so i returned to a version using hooks.  this one isn't
perfect, because at times, when eg. M-x gnus manages my buffers, i
have to press the key 'other-window is bound to once to see the mark
indicating the M-x selected-window.

;; -ino: 221103-1133
(setq cf-model-selected '(
      "-"
      mode-line-mule-info
      mode-line-modified
      mode-line-frame-identification
      mode-line-buffer-identification
      " * "
      global-mode-string
      "   %[(" mode-name
      mode-line-process minor-mode-alist "%n"
      ")%]--"
      (which-func-mode ("" which-func-format "--"))
      (line-number-mode "L%L--")
      (column-number-mode "C%c--")
      (-3 . "%p")
      "-%-"
))

(setq cf-model-egal '(
      "-"
      mode-line-mule-info
      mode-line-modified
      mode-line-frame-identification
      mode-line-buffer-identification
      "   "
      global-mode-string
      "   %[(" mode-name
      mode-line-process minor-mode-alist "%n"
      ")%]--"
      (which-func-mode ("" which-func-format "--"))
      (line-number-mode "L%L--")
      (column-number-mode "C%c--")
      (-3 . "%p")
      "-%-"
))

(defun cf-before-other-win nil
  (if (or (eq this-command 'other-window)
          (eq this-command 'select-window))
      (setq mode-line-format cf-model-egal)))

(defun cf-after-other-win nil
  (if (or (eq this-command 'other-window)
          (eq this-command 'select-window))
      (progn
        (setq mode-line-format cf-model-selected)
        (force-mode-line-update t))))

(add-hook 'pre-command-hook 'cf-before-other-win)
(add-hook 'post-command-hook 'cf-after-other-win)

> Crufty!  Does setting deactivate-mark temporarily while a pre- or
> post-command-hook function is being evaluated actually have an
> effect?

i didn't check this at all, it was part of some code pasted there.  i
don't use it.

> And the local com variable is completely unnecessary: just test (eq
> last-command 'other-window)

thanks for this tip!  the previous version had M-x string= instead,
because it's doc-string indicated it would compare symbols using their
print-string.

question:  can i be sure that elisp will always keep only one copy of
any symbol, so that i can effectively compare the pointers to them,
which is the cheapest form of comparison?

> [making elisp re-read anything it prints and get the same structure]
> It's just too ugly.

too bad.  so i have to remember that elisp gets text properties, but
there are exceptions in certain other cases.

  clemens