Hello,,I encounter a problem,,

hmmmm,,if this topic you don't understand,,tell me,,I'll explain it..

The most accurate way to time things is using the high performance counter if your hardware has one. This is controlled and queried using QueryPerformanceFrequency() and QueryPerformanceCounter() which are part of . For example your code might look something like:

struct Button{  long m_frequency;  long m_startCount;  bool m_locked;   float m_coolDown;  void Initialise();  bool Push();  void Update();}void Button::Initialise(){  QueryPerformanceFrequency(&m_frequency);   m_locked = false;  m_coolDown = 1.0f;   // 1 second count before re-activation allowed}bool Button::Push(){  if(!m_locked)  {    QueryPerformanceCounter(&m_startCount);    m_locked = true;    return true;  }  return false;}void Button::Update(){  if(m_locked)  {    long currentCount;    QueryPerformanceCounter(&currentCount);    float timeElapsed = (float)(currentCount - startCount)/(float)m_frequency;    if(timeElapsed >= m_coolDown)    {      m_locked = false;    }  }}

Note that I just made that up to give you an idea, there's no error checking or anything like that so I wouldn't use it directly. Also not all hardware has a high performance counter. In this case QueryPerformanceFrequency() will give you a frequence of zero. If this happens you can always fall back on using timeGetTime() which returns the system time in milliseconds.

I pretty much agree, except to make one small point. You wouldn't want seperate timer code in every little thing that uses timed effects, but would probably have a game timer that your button would check against. (Unless QPC is really low overhead and doesn't need any error checking, I guess. But it's probably still a good idea to abstract it out in case a better timing model comes out or something.)

Quote:

Original post by Bob Janova I pretty much agree, except to make one small point. You wouldn't want seperate timer code in every little thing that uses timed effects, but would probably have a game timer that your button would check against. (Unless QPC is really low overhead and doesn't need any error checking, I guess. But it's probably still a good idea to abstract it out in case a better timing model comes out or something.)

Sound advice. In fact the way I do it is I've embedded my clock into the windows messaging loop and pass the time elapsed each frame as a parameter into the main game loop. This has several advantages. For a start the high performance counter doesn't have to be queried by each object that requires timing thus saving some cycles just like Bob says. It also ensures that every game object observes the same passage of time each frame. If each object is responsible for its own clock then they can start to diverge due to differing execution lengths of code between objects. Lastly when the clock is placed here it can also be used to fix your frame rate to prevent stuttering graphics.

Thanks ....but that's not the answers....

all I want to know just the effect..not compute the cooldownh time,,just want to know how can i do to make the weapon/magic button rounded itself ? should I make a lots of alpha image and definition there slope/(alpha area) and render them at right moment?

very thanks those guy,,it make me more clear about my poor English,, LOL

I'd just render one alpha image (the button) with a different colour value applied to it, and then interpolate the colour using code similar to Motorherp's (although I'd probably avoid the QueryPerformance* routines as I've heard there's multi-threading issues with them).