Re: Problem report #28

["Stuart D. Herring" <herring@lanl.gov>]

I don't have an answer for this one, but I've trimmed the issue down for
clarity.  We have the following if-else_if-else:

> 881  	      if (FRAMEP (window))
> 882  		{
> 883  		  f = XFRAME (window);
> 884  		  xpos = 0;
> 885  		  ypos = 0;
> 886  		}
>
> At conditional (12): "window & 7 == 4" taking false path
>
> 887  	      else if (WINDOWP (window))
> 888  		{
> 889  		  CHECK_LIVE_WINDOW (window);
> 890  		  f = XFRAME (WINDOW_FRAME (XWINDOW (window)));
> 891
> 892  		  xpos = WINDOW_LEFT_EDGE_X (XWINDOW (window));
> 893  		  ypos = WINDOW_TOP_EDGE_Y (XWINDOW (window));
> 894  		}
> 895  	      else
> 896  		/* ??? Not really clean; should be CHECK_WINDOW_OR_FRAME,
> 897  		   but I don't want to make one now.  */
>
> At conditional (13): "window & 7 != 4" taking true path
> At conditional (14): "0" taking false path
>
> 898  		CHECK_WINDOW (window);

f, guaranteed to be NULL before this block, is assigned in the first two
cases but not in the third (which is tagged with a "huh?" comment).  Later
f is used in a place Coverity claims requires non-nullity.

Hope this helps the next analyzer,
Davis

-- 
This product is sold by volume, not by mass.  If it appears too dense or
too sparse, it is because mass-energy conversion has occurred during
shipping.