Proposed solution for COMCTL32 subclassing problems

[Filip Navara]

Robert Shearman wrote:

>Can you not just use stack->stacknum instead of 0 if the stack is inverted?
>  
>
Of course I can (now that I understand the code ;-).

>I'm not arguing against it and I'm all for it if it makes it easier to code,
>but I'm just trying to see what stops you from using the same stack
>direction as at the moment.
>  
>
If there is enough interest, I can recode it with using the previous 
stack layout. When I started the patch I just thought it would be easier 
this way, but now I can reasonably say that it wouldn't be much different.

- Filip

P.S.
In the last patch I sent there was still one bug:
+         stack->stacknum--;
+         if (n < stack->stackpos)
             stack->stackpos--;
should be
+         stack->stacknum--;
+         if (n < stack->stackpos && stack->wndprocrecursion)
             stack->stackpos--;