Re: emacs-21.1.94 crash in gnus on Windows

[Eli Zaretskii]

> From: Chong Yidong <address@hidden>
> Cc: address@hidden, address@hidden, address@hidden
> Date: Thu, 25 Mar 2010 12:35:05 -0400
> 
> Eli Zaretskii <address@hidden> writes:
> 
> > But the part that you left in the code also uses it.glyph_row, which
> > is garbage after a call to move_it_to.  There's a single use of
> > it.glyph_row->x before the call to PRODUCE_GLYPHS (&it2), and I think
> > that needs to be fixed as well, because it potentially dereferences a
> > bad pointer.
> >
> > Or did I miss something?
> 
> I don't see the failure condition.  This branch handles the special case
> of charpos == 1 or charpos above the top of the window.  Since
> start_display initialized the iterator using the desired matrix of the
> window, it.glyph_row should always be valid.
> 
> Or am I confused?

it.glyph_row is indeed initialized in start_display.  But are you sure
it.glyph_row->x is set correctly?  move_it_to does not produce glyphs
in glyph_row, so who computes glyph_row->x ?