bug#193: Fix for bug#193.

[Alan Mackenzie]

Hi, Christoph,

On Fri, Jul 09, 2010 at 07:41:12PM -0600, Christoph wrote:
> Hi Alan,

> On 7/9/2010 12:51 PM, Alan Mackenzie wrote:
> >Hi, Deniz,

> >On Sun, Jun 27, 2010 at 02:47:52PM +0200, Deniz Dogan wrote:
> >>Did anyone ever agree on what needs to be done to fix the problem?

> >Would you try this patch, please, and then tell me whether or not it
> >works fully.  Thanks!

> M-q works fine for me with C as well as C++ style comments, both ways, 
> i.e. shorting long comments and extending short comments to fill the line.

> The only "quirk" I found:

> Using C-style comments, if the original comment looks like this:

> /* This is a short comment
>    which is extended                      */

> the resulting comment is:

> /* This is a short comment which is extended  */

> Notice the two spaces after the last word 'extended', before the close 
> comment marker.

The reason being that the filling function assumed that a comment ends
with a sentence, and should therefore leave two spaces after it.
However, if there's no full stop (etc.), the two spaces is silly.

> This might be nitpicking but I thought I bring it up.

We need nitpickers.  ;-)  Thanks for drawing it to my attention.  Would
you please try this amended patch: It should leave just one space after
"is extended", but two spaces after "is extended.".


*** orig/cc-cmds.el     2010-07-05 20:17:46.000000000 +0000
--- cc-cmds.el  2010-07-11 18:46:13.066679040 +0000
***************
*** 3975,3980 ****
--- 3975,3988 ----
                                       (goto-char ender-start)
                                       (current-column)))
                       (point-rel (- ender-start here))
+                      (sentence-ends-comment
+                       (save-excursion
+                         (goto-char ender-start)
+                         (and (search-backward-regexp
+                               (c-sentence-end) (c-point 'bol) t)
+                              (goto-char (match-end 0))
+                         (looking-at "[ \t]*")
+                         (= (match-end 0) ender-start))))
                       spaces)
  
                  (save-excursion
***************
*** 4017,4023 ****
                              (setq spaces
                                    (max
                                     (min spaces
!                                         (if sentence-end-double-space 2 1))
                                     1)))
                          ;; Insert the filler first to keep marks right.
                          (insert-char ?x spaces t)
--- 4025,4033 ----
                              (setq spaces
                                    (max
                                     (min spaces
!                                         (if (and sentence-ends-comment
!                                                  sentence-end-double-space)
!                                             2 1))
                                     1)))
                          ;; Insert the filler first to keep marks right.
                          (insert-char ?x spaces t)
***************
*** 4229,4236 ****
    (let ((fill-paragraph-function
         ;; Avoid infinite recursion.
         (if (not (eq fill-paragraph-function 'c-fill-paragraph))
!            fill-paragraph-function)))
!     (c-mask-paragraph t nil 'fill-paragraph arg))
    ;; Always return t.  This has the effect that if filling isn't done
    ;; above, it isn't done at all, and it's therefore effectively
    ;; disabled in normal code.
--- 4239,4249 ----
    (let ((fill-paragraph-function
         ;; Avoid infinite recursion.
         (if (not (eq fill-paragraph-function 'c-fill-paragraph))
!            fill-paragraph-function))
!       (start-point (point-marker)))
!     (c-mask-paragraph
!      t nil (lambda () (fill-region-as-paragraph (point-min) (point-max) arg)))
!     (goto-char start-point))
    ;; Always return t.  This has the effect that if filling isn't done
    ;; above, it isn't done at all, and it's therefore effectively
    ;; disabled in normal code.



> Christoph

Looking forward to hearing back from you,

-- 
Alan Mackenzie (Nuremberg, Germany).