So today i decided to use DirectInput. I wrote the code,compiled and ran.. the game image appeared for 1 second,then a WM_QUIT msg was sent and the app just closed.
How is this possible? I checked for null pointers in the new DirectInput code and I have none.
What could trigger something like this?!
First: DirectInput is legacy code. For keyboard and mouse input, you should use Windows messaging.
What error checking do you do? Did you compile with _Debug? Do you test every call* (not just DirectInput but all API calls) for either an error or proper response? If there's an invalid call or error return, how do you handle it? Have you tried using OutputDebugString at places in the code to trace the progress of your app? Do you initialize every variable before it's used? If you're using Visual Studio, are there any pertinent messages in the output window?
*Every API call provides some indication of success or failure. Your app should check for and handle "with grace" every error indication.
EDIT: It may be that you're getting a bad return from a DirectInput call, the result of which is needed elsewhere. That "elsewhere" fails and is set up to close the window on failure.
From Braynzar:
You may wonder why we are not going to use Win32 API to get input from the user, after all, it can find which keys were pressed on the keyboard, get input from the mouse and even from a joystick. Well, the Win32 API was designed as a means for keyboard entry applications, not for speed real time input. When getting input with the Win32 API, it needs to process the keys, convert them to ascii and there is a lot of extra processing for special windows keys pressed like "alt". So much processesing that games don't need, and remember, we can't render a frame until our message queue is empty. It's actually very very slow. Another thing, is the Win32 API can only take input from 3 buttons, x and y axis, and a mouse wheel from a mouse. Many mouses now days have much more function than that. Direct Input can support pretty much any device as long as it has a Direct Input driver written for it, which is pretty much all of them. With Direct Input, We can use a feature called force feedback, thats when a joystick can rumble and get input from the application.
And I fixed it,I was passing an invalid argument to the acquire mouse function.
What makes you say direct input is legacy? The fact that a newer version wasn't released? Why would there be a new one if the old one works...
"Why would there be a new one if the old one works..."
If may work today, but not tomorrow.
See DirectInput (legacy). The implication is that DirectInput will not be revised to reflect any changes in other APIs. Conversely, any change to other software will not take into consideration its impact on DirectInput. If, for instance, the Windows API is revised or updated (new version of Windows, a service pack, etc.,) and that change would affect the operation of DirectInput, you're on your own. Microsoft isn't going to ensure the revision is backwardly compatible with respect to DirectInput.
From Braynzar:
You may wonder why we are not going to use Win32 API to get input from the user, after all, it can find which keys were pressed on the keyboard, get input from the mouse and even from a joystick. Well, the Win32 API was designed as a means for keyboard entry applications, not for speed real time input. When getting input with the Win32 API, it needs to process the keys, convert them to ascii and there is a lot of extra processing for special windows keys pressed like "alt". So much processesing that games don't need, and remember, we can't render a frame until our message queue is empty. It's actually very very slow. Another thing, is the Win32 API can only take input from 3 buttons, x and y axis, and a mouse wheel from a mouse. Many mouses now days have much more function than that. Direct Input can support pretty much any device as long as it has a Direct Input driver written for it, which is pretty much all of them. With Direct Input, We can use a feature called force feedback, thats when a joystick can rumble and get input from the application.
This is basically nonsense. Trust me - the "very very slow" "so much processing" won't even register on any performance graph, it's down in the noise. The person who wrote that is obviously someone who's never actually done any profiling but instead lets their own personal opinions (and prejudices) rule.
Despite that, there are still some valid reasons to continue using DirectInput, for mouse and mouse only.
First is that it automatically bypasses pointer acceleration, which may be desired behaviour for certain types of games (e.g. FPS).
Second is that it automatically handles focus loss and gain, mouse pointer visibility, and clamping to window bounds for you.
Third is that it's far better documented than Raw Input - tried using GetRawInputBuffer lately?
if these reasons apply to you and to the type of game you're making, then using DirectInput remains a viable choice - despite what Microsoft say.
"Why would there be a new one if the old one works..."
If may work today, but not tomorrow.
See DirectInput (legacy). The implication is that DirectInput will not be revised to reflect any changes in other APIs. Conversely, any change to other software will not take into consideration its impact on DirectInput. If, for instance, the Windows API is revised or updated (new version of Windows, a service pack, etc.,) and that change would affect the operation of DirectInput, you're on your own. Microsoft isn't going to ensure the revision is backwardly compatible with respect to DirectInput.
And then when everybody's legacy software breaks and they decide to not upgrade to the latest version of Windows because of that, it will be Microsoft's fault. Nice.
I think the only real implication is that new input devices (touch, voice, mind control) are not going to be supported via Direct Input. But, there's not a chance that Microsoft will 'break' any existing code that depends on.
