🚀 go-pugleaf

I recently reported that parse-partial-sexp returns the wrong result
when it restarts a parse (using a saved STATE value) when the previous
parse ended after the first character of a 2-character comment-end
sequence.  I discovered this because font-lock wasn't fontifying the
string in this simple program:-

----------------------------------------------------------------------

int main (int argc, char *argv[])
{
  /***************/

  printf ("Hello world\n");
}

----------------------------------------------------------------------

Here is a patch for font-lock.el that works around this problem.  I've
done this because I don't fancy working out a fix for parse-partial-sexp!


--
Kevin

*** font-lock.el.sav	Fri Sep 30 14:22:40 1994
--- font-lock.el	Fri Oct  7 11:14:12 1994
***************
*** 194,200 ****
  		  ;; Verify the state at that place
  		  ;; so we don't get fooled by \" or \;.
  		  (setq state (parse-partial-sexp prev (point)
! 						  nil nil state)))
  	(let ((here (point)))
  	  (if (or (nth 4 state) (nth 7 state))
  	      ;; We found a real comment start.
--- 194,200 ----
  		  ;; Verify the state at that place
  		  ;; so we don't get fooled by \" or \;.
  		  (setq state (parse-partial-sexp prev (point)
! 						  nil nil prevstate)))
  	(let ((here (point)))
  	  (if (or (nth 4 state) (nth 7 state))
  	      ;; We found a real comment start.
***************
*** 210,216 ****
  			(skip-chars-backward " \t"))
  		    (error (goto-char end))))
  		(put-text-property beg (point) 'face font-lock-comment-face)
! 		(setq state (parse-partial-sexp here (point) nil nil state)))
  	    (if (nth 3 state)
  		(let ((beg (match-beginning 0)))
  		  (while (and (re-search-forward "\\s\"" end 'move)
--- 210,223 ----
  			(skip-chars-backward " \t"))
  		    (error (goto-char end))))
  		(put-text-property beg (point) 'face font-lock-comment-face)
! 		;; Find the state at end of the comment.  Parse from `prev'
! 		;; (which is before the comment start) rather than `here'
! 		;; (which is inside the comment) because parse-partial-sexp
! 		;; gives the wrong result if `here' is between the characters
! 		;; of a 2-character comment-end sequence (such as "*/" in a
! 		;; c-mode buffer).  KMB
! 		(setq state (parse-partial-sexp prev (point) nil nil
! 						prevstate)))
  	    (if (nth 3 state)
  		(let ((beg (match-beginning 0)))
  		  (while (and (re-search-forward "\\s\"" end 'move)