Re: Bidi reordering engine upgraded

[Eli Zaretskii]

> From: Jan Djärv <address@hidden>
> Date: Sat, 18 Oct 2014 14:34:06 +0200
> Cc: address@hidden,
>  address@hidden
> 
> >   /* Compute frame-relative coordinates for phys cursor.  */
> >   get_phys_cursor_geometry (w, row, cursor_glyph, &x, &y, &h);
> > -  wd = w->phys_cursor_width;
> > +  wd = w->phys_cursor_width - 1;
> 
> Are we sure this never becomes negative?

If it does, we already have that problem, since previously, the
subtraction of 1 was done in get_phys_cursor_geometry.

> I suggest you check it in so we can test it.

OK, will do.

> Do we have a test case?

Yes, the one posted by Dmitry here:

  http://lists.gnu.org/archive/html/emacs-devel/2014-10/msg00518.html