bug in vertical-motion?

bug in vertical-motion?

[Luc Teirlinck]

Do `emacs -q' and then hold the `q' (or your favorite letter) key down
until you get a continuation line.

Put cursor on the next to last `q' on the screen line that is
continued.  Then M-: (vertical-motion 0) carries you to the beginning
of the current screen line, as one would expect.  But then put cursor
on the very last `q' of the continued screen line.  Then
M-: (vertical-motion 0) carries you to the beginning of the _next_
screen line.  This appears to be a bug.

Sincerely,

Luc.

bug in vertical-motion?

[Luc Teirlinck]

I should add that the described bug does not occur in Emacs-21.3.

Sincerely,

Luc.

Re: bug in vertical-motion?

[Richard M. Stallman]

    Put cursor on the next to last `q' on the screen line that is
    continued.  Then M-: (vertical-motion 0) carries you to the beginning
    of the current screen line, as one would expect.  But then put cursor
    on the very last `q' of the continued screen line.  Then
    M-: (vertical-motion 0) carries you to the beginning of the _next_
    screen line.  This appears to be a bug.

Do people get good results from this patch?

*** xdisp.c	27 Aug 2005 22:41:25 -0400	1.1045
--- xdisp.c	29 Aug 2005 20:14:40 -0400	
***************
*** 6012,6017 ****
--- 6012,6019 ----
  	     glyphs have the same width.  */
  	  int single_glyph_width = it->pixel_width / it->nglyphs;
  	  int new_x;
+ 	  int x_before_this_char = x;
+ 	  int hpos_before_this_char = it->hpos;
  
  	  for (i = 0; i < it->nglyphs; ++i, x = new_x)
  	    {
***************
*** 6043,6050 ****
--- 6045,6066 ----
  		    {
  		      ++it->hpos;
  		      it->current_x = new_x;
+ 
+ 		      /* The character's last glyph just barely fits
+ 			 in this row.  */
  		      if (i == it->nglyphs - 1)
  			{
+ 			  /* If this is the destination position,
+ 			     return a position *before* it in this row,
+ 			     now that we know it fits in this row.  */
+ 			  if (BUFFER_POS_REACHED_P ())
+ 			    {
+ 			      it->hpos = hpos_before_this_char;
+ 			      it->current_x = x_before_this_char;
+ 			      result = MOVE_POS_MATCH_OR_ZV;
+ 			      break;
+ 			    }
+ 
  			  set_iterator_to_next (it, 1);
  #ifdef HAVE_WINDOW_SYSTEM
  			  if (IT_OVERFLOW_NEWLINE_INTO_FRINGE (it))

Re: bug in vertical-motion?

[Luc Teirlinck]

Richard Stallman wrote:       

       Put cursor on the next to last `q' on the screen line that is
       continued.  Then M-: (vertical-motion 0) carries you to the beginning
       of the current screen line, as one would expect.  But then put cursor
       on the very last `q' of the continued screen line.  Then
       M-: (vertical-motion 0) carries you to the beginning of the _next_
       screen line.  This appears to be a bug.

   Do people get good results from this patch?

   *** xdisp.c	27 Aug 2005 22:41:25 -0400	1.1045
   --- xdisp.c	29 Aug 2005 20:14:40 -0400	

It appears to work.

Sincerely,

Luc.