bug#11464: 24.1.50; pos-visible-in-window-p returns a false positive with bidi text

[Eli Zaretskii <eliz@gnu.org>]

> From: Ari Roponen <ari.roponen@gmail.com>
> Cc: 11464@debbugs.gnu.org
> Date: Thu, 17 May 2012 07:52:11 +0300
> 
> bottom_y = 270
> it.last_visible_y = 256
> top_y = 255
> window_top_y = 0

Now I'm totally bewildered.  I don't understand how this case is at
all relevant to the bug.  Please bear with me while I explain what
puzzles me, and please point out what I missed.

Here's the relevant code fragment:

      int top_x = it.current_x;
      int top_y = it.current_y;
      /* Calling line_bottom_y may change it.method, it.position, etc.  */
      enum it_method it_method = it.method;
      int bottom_y = (last_height = 0, line_bottom_y (&it));
      int window_top_y = WINDOW_HEADER_LINE_HEIGHT (w);

      if (top_y < window_top_y)
	visible_p = bottom_y > window_top_y;
      else if (top_y < it.last_visible_y)
	visible_p = 1;
      if (bottom_y <= it.last_visible_y
	  && it.bidi_p && it.bidi_it.scan_dir == -1
	  && IT_CHARPOS (it) < charpos)
	{

The original problem was that the "else if" clause would incorrectly
set visible_p to 1.  The "if" clause I added after that, viz.:

      if (bottom_y <= it.last_visible_y
	  && it.bidi_p && it.bidi_it.scan_dir == -1
	  && IT_CHARPOS (it) < charpos)

attempts to correct that, by eventually resetting visible_p to zero.

Now, if bottom_y = 270, it.last_visible_y = 256, and top_y = 255, then
the condition in the above "else if" clause is false, and visible_p
could not possibly be set to 1.  The preceding "if" clause is also
false, since window_top_y = 0.  Therefore, visible_p should have
stayed zero for this situation, and I don't understand why you needed
to change

      if (bottom_y >= it.last_visible_y

into

      if (bottom_y <= it.last_visible_y

Can you tell where I'm wrong?

Thanks in advance.