GetWindowDC(NULL) failing

I've a function that calls GetWindowDC(NULL). and that function is called by different threads. sometimes I see GetWindowDC(NULL) returning 0x0 Why ? and How to resolve ? the function is mutex gaurded properly. So thats not a Problem. Do I need to sleep() few ms after each call ? and Why ?

http://msdn.microsoft.com/en-us/library/dd144947%28v=vs.85%29.aspx

msdn Says

Return Value

If the function succeeds, the return value is a handle to a device context for the specified window. If the function fails, the return value is NULL, indicating an error or an invalid hWnd parameter.

but What 开发者_运维问答this indicating an error means I couldn't understand

Edit:

and Ya I am releasing all the DCs and also Deleting the DCs properly

Giving an immediate call to GetLastError retrns 0 tried.

if(hdc == 0x0)
    qDebug() << GetLastError()

Here goes my source code. It Works but after working more than a 100 times (35-40 secs) it refuses to work and GetWindowDC(NULL) returns 0

QPixmap Util::grabScreen(const DG::Rect* rect){
    mutex.lock();

    HDC hdc=GetWindowDC(NULL);
    HWND win=WindowFromDC(hdc);

    HDC cdc=CreateCompatibleDC(hdc);
    HBITMAP temp=CreateCompatibleBitmap(hdc,rect->width,rect->height);
    PAINTSTRUCT ps;

    hdc=BeginPaint(win,&ps);
    HBITMAP oldb=(HBITMAP)SelectObject(cdc,temp);
    BitBlt(cdc,0,0,rect->width,rect->height,hdc,rect->top,rect->left,SRCCOPY);
    SelectObject(cdc,oldb);
    EndPaint(win,&ps);

    char* buff;
    buff = new char[rect->size()];
    GetBitmapBits(temp,rect->size(),buff);
    qDebug() << "temp" << temp;
    if(temp == 0x0){
        qDebug() << "hdc" << hdc;
    }

    DeleteDC(cdc);
    ReleaseDC(NULL, hdc);
    DeleteDC(hdc);

    QPixmap pixmap = QPixmap::fromWinHBITMAP(temp);
    //QPixmap pixmap = QPixmap::grabWidget(desktopWidget,rect->toQRect());
    mutex.unlock();
    return pixmap;

}

---

According to MSDN you shouldn't call DeleteDC on an handle obtained using GetDC. You should just call ReleaseDC instead.

---

Without know more details on what is going on; the scenario sounds a bit odd to me. However...

Is it possible that code is also processing a WM_PAINT for the same window at the same time you are trying to get the WindowDC ?

I'm not sure if you can obtain a window DC for the same window that is being actively painted.