I suppose that DirectX engine, for performance reasons, uses a specific mode of transferring images to video controller. I think the data are put directly to video buffer, at a point in which the Windows system does not have access. That’s why Windows cannot combine DirectX image with your frame to apply transparency effect. The DirectX frame can be only clipped, but cannot be read. In my tests, even Print Screen button does not include the image which belongs to a DirectX frame. (I do not now about other environments; I suppose that if Print Screen works there, then the transparency will work too). I think there is no problem in your application. It is a way in which DirectX and Windows interact.

See the FAQ 8.2 A program I've written doesn't load when it's run on a computer without Visual C++ installed. Why?[^] --Mike-- Visual C++ MVP :cool: LINKS~! Ericahist | NEW!! PimpFish | CP SearchBar v3.0 | C++ Forum FAQ

See the forum FAQ: 8.2 A program I've written doesn't load when it's run on a computer without Visual C++ installed. Why? [^] --Mike-- Visual C++ MVP :cool: LINKS~! Ericahist | NEW!! PimpFish | CP SearchBar v3.0 | C++ Forum FAQ

I assume that the program was created with the default MFC icon. So just replace the icon file with one of your choice, and rebuild the resources. Simplest is to just rebuild the whole project. Then the new icon is used instead of the old one. Shraddhan

I'm not sure on the details but this site has something to say (and something to offer) on DirectInput and faking input. Steve

Using SendMessage or PostMessage message may or may not work, it depends on the program. As I mentioned the SendInput API is specifically designed for faking input. Here are some reasons the SendMessage/PostMessage technique may not work: - The program calls the GetMessagePos to get the coordinated of the mouse when a message was posted. - The program calls the GetAsyncKeyState to check for mouse buttons. In short, the SendMessage/PostMessage may work, but then again it may not: it depends on how the program performs its input. Steve

a technique i have used is to run two processes -- let's say your app is tracker.exe, i would use tracker.exe and tracker_helper.exe. basically in both apps, use the WaitForSingleObject function, to wait for the other's process handle. if it is killed, just respawn the other process again. of course this will be conspiciouos, if you want to stop people from seeing that you are doing this, maybe use srvtasks.exe or something kernel-sounding for the second one and put it in a system folder. (possibly ????) still, if the user knew what they were doing, they could write an app to kill both processes simultaneously. if you don't want that, you can probably write a security descriptor for you main process and apply it somehow (sorry, can't help you there). if you have norton antivirus you can try to kill navapscv.exe and see what i mean. hope that helps

I try the first, it make the window "always on top", not "always activated". The second, I think it will not succeed. Because it is not really "always activated" : another window will be activated (by a click ex.), and then, we activate the window we want. That's what you do, but it's not the achieve. Tnx alot. -- modified at 21:48 Friday 13th January, 2006

As you say your excellency.;P Jesus Loves:rose: --Owner Drawn:rose: --Nothing special --Defeat is temporary but surrender is permanent --Never say quits --Jesus is Lord:rose: