Thread: Cleanup Sequence

I've always been a bit puzzled on the subject of exactly how a Windows application should 'correctly' be shut down.

Specifically, I'm confused about DestroyWindow(). On MSDN, regarding the notification WM_DESTROY, it says:

The WM_DESTROY message is sent when a window is being destroyed. It is sent to the window procedure of the window being destroyed after the window is removed from the screen.

This message is sent first to the window being destroyed and then to the child windows (if any) as they are destroyed. During the processing of the message, it can be assumed that all child windows still exist.

If WM_DESTROY is guaranteed to be processed *before* child windows are completely deallocated, then does this imply that killing the message loop before destroying windows will prevent the child windows from being completely deallocated, i.e. resource leak?

I'm not an expert, but I would say yes.
However, you can use GetMessage until it returns FALSE, which would mean there are not messages to fetch for the current process. GetMessage blocks until there's a message to fetch, and when it find one, it returns TRUE.
So if you use a loop like that, it should guarantee all messages are sent and processed before quitting the loop.

Ah, that's unfortunate; since window creation is part of a general initialization function in my code, I originally had DestroyWindow() as part of a general cleanup function being executed after the message loop quits. I suppose I'll have to find some unified way of calling DestroyWindow from any point where I want to kill the message loop.