Proposed solution for COMCTL32 subclassing problems

[Robert Shearman]

Filip Navara wrote:
>
> Hi,
>
> here is my proposed solution for crashes in COMCTL32 observed in
> specific situations while calling the window procedure subclassing
> functions. If there were removed all subclassed window procedures during
> the execution of SubclassWndProc there happened a crash, because the
> structure that holds the subclassing stack was freed in DefSubclassProc.

Yes, that was my fault. Sorry.

> Also if there were some window procedure subclassings removed during the
> execution of SubclassWndProc, the stack position counting could have
> gone wrong. My solution is to let the subclassing stack grow the
> opposite way, so that item 0 is the most nested one. I also removed the
> stacknew handling which I strongly believe is redundant now.

It looks much simpler now, but I don't understand why reversing the
direction of the stack should make any difference.
I don't think you submitted the required patch to comctl32.h.

Rob