Re: looking for: query-replace with a "yank-word" binding ala isearch

[Kevin Rodgers]

bigfaceworm wrote:

> Kevin's help worked wonderfully,
> *except* now I have some odd behavior in my elisp.
> When I debug the routine (via edebug) it works just
> fine, but when it's loaded outside of edebug, it 
> grabs the wrong characters:
>
> (defun bfw-insert-word ()
>   "insert word the cursor is on into minibuffer"
>   (interactive)
>   (let ((oldtxt (buffer-substring (save-excursion (beginning-of-line)
> (point)) (point))))
>     (insert
>      (save-excursion
>        (set-buffer (window-buffer minibuffer-scroll-window))
>        (buffer-substring (progn (forward-char (length oldtxt))
> (point)) (progn (forward-word 1) (point)))))))


Why should the contents of the minibuffer affect what text is inserted?

(defun bfw-insert-word ()
  "Insert the word after point from the selected buffer into the minibuffer.
This advances point in the selected buffer."
  (interactive)
  (insert (save-excursion
            (set-buffer (window-buffer minibuffer-scroll-window))
            (buffer-substring (point)
                              (progn (forward-word 1) (point))))))


> M-x global-set-key f9 bfw-insert-word
>
> Create a new buffer, with contents:
>
> 1 2 3 4 5 6 7 8 9 a b c d e f g h i j k l m n o p q r s t u v w x y z
>
> move the point to the left of the `1', and query-replace
>
>     M-% f9 f9 f9 f9 f9 f9
>
> gives you "12 4 7 b g"
>
> but if you debug the routine (M-x edebug-defun)
> and do the same test (hitting 'c' to continue)
> you get "1 2 3 4 5 6"
>
>
> Any idea why a defun would act differently when debugged versus 
> normal?


Perhaps the debugger is restoring point or some other aspect of the window/
buffer configuration.

--
<a href="mailto:&lt;kevin.rodgers&#64;ihs.com&gt;";>Kevin Rodgers</a>