Re: Bugs #7240, #9133 - fixing them both at the same time.

[Wolfgang Jenkner <wjenkner@inode.at>]

Alan Mackenzie <acm@muc.de> writes:

> The cause of these failures was trying to make one variable, par-beg, do
> two different jobs, namely 1) beginning of the paragraph; 2) start of
> text in the paragraph.
>
> By separating these out into par-beg and text-beg the bug simply
> evaporates.

So part of your change is

>     (let ((opoint (point))
>           (sentence-end (sentence-end)))
>       (while (< arg 0)
> !       (let* ((pos (point))
> ! 	     (text-beg
> ! 	      (progn (start-of-paragraph-text)
> ! 		     (point)))
[...]
>   	    (goto-char (match-end 0))
> ! 	  (goto-char text-beg)))

and my whole patch is

 	   (goto-char (match-end 0))
-	 (goto-char par-beg)))
+	 (goto-char opoint)
+	 (start-of-paragraph-text)))

This is equivalent.  Of course, style matters.

The rest of your change (with respect to the code before rms patched it)
consists in adding the 4th and 5th line in

> ! 	     (par-beg
> ! 	      (progn
> ! 		(backward-paragraph)
> ! 		(if (looking-at paragraph-separate)
> ! 		    (forward-line))
> ! 		(point))))

That may be a good change, but it is not directly related to the bugs.

> Please try out the following patch and confirm to me that it's OK.  Then
> I will commit it.

Oh my.  I'm afraid you got a code review instead;-)

> Wolfgang:  Thanks very much for taking the trouble to report this bug,
> submitting a patch (even if it wasn't quite right),

Wolfgang