Handling some things in WM_CREATE can cause undesirable results. Because if you pass the window handle to somewhere, it may fail because the window is not created yet. One good example, what can happen is there:
http://cboard.cprogramming.com/showthread.php?t=79130
"The Internet treats censorship as damage and routes around it." - John Gilmore
No. You used a global HWND, and tried to use it in createtask before createwindow returned to assign that window handle. That's your fault. If you used the hwnd passed to the window procedure, it would work.
Maybe yes, uninitialized too...
Once I had a problem with WM_CREATE when I was comparing some window handles when CreateWindowEx hadn't returned yet...
http://cboard.cprogramming.com/showthread.php?p=501964
Last edited by maxorator; 09-25-2006 at 05:26 AM.
"The Internet treats censorship as damage and routes around it." - John Gilmore
That still doesn't justify your assertion. Most people are aware that WM_CREATE is one of a small number of messages issued during system processing of CreateWindowEx but the window handle passed to the window procedure is nonetheless valid for that window and for all other messages for that window.
Any problems you may have encountered with the use of such handles is almost certainly of your own making.
CProgramming FAQ
Caution: this person may be a carrier of the misinformation virus.
From a system design perspective, what point would there be in calling a window procedure function when no window instance exists?
I can practically guarantee that at no time will Windows call a window procedure without supplying a window handle to use it with.
