Pane Manager bug in 17.2.0

Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic


   I had a resize bug, where if a splitter was moved, the growing pane would not redraw until some other event caused a repaint.  I tracked it down to the filtering of messages whilst in capture for the splitter move.  The code was calling ValidateRect on the window, which prevented repaints from happening.  I commented it out and now it's happy and so am I.  Even 2 paints would be better than none.  I haven't checked if it's getting 2, but it's smooth as silk.
BOOL CXTPDrawHelpers::ProcessPendingPaintMessages(HWND hWnd /*= NULL*/)
{
	BOOL bResult = TRUE;
	MSG msg;

	while (::PeekMessage(&msg, hWnd, WM_PAINT, WM_PAINT, PM_NOREMOVE)
		|| ::PeekMessage(&msg, hWnd, WM_ERASEBKGND, WM_ERASEBKGND, PM_NOREMOVE)
		|| ::PeekMessage(&msg, hWnd, WM_SYNCPAINT, WM_SYNCPAINT, PM_NOREMOVE)
		|| ::PeekMessage(&msg, hWnd, WM_NCPAINT, WM_NCPAINT, PM_NOREMOVE))
	{
		if (WM_QUIT != msg.message)
		{
			if (::PeekMessage(&msg, hWnd, msg.message, msg.message, PM_REMOVE))
			{
				if (WM_QUIT != msg.message) // WM_QUIT should never happen here but still extra verification won't harm
				{
					::DispatchMessage(&msg);

					// This ensures the window won't enter the second painting cycle.
//					::ValidateRect(msg.hwnd, NULL);
				}
				else
				{
					// Re-post WM_QUIT as it's been removed.
					::PostQuitMessage(static_cast<int>(msg.wParam));

					bResult = FALSE;
					break;
				}
			}
		}
		else
		{
			bResult = FALSE;
			break;
		}
	}

	return bResult;
}

Author	Message Topic Search Topic Options Post Reply Create New Topic Printable Version Translate Topic
adrien Members Profile Send Private Message Find Members Posts Add to Buddy List Senior Member Joined: 30 April 2007 Location: New Zealand Status: Offline Points: 449	Post Options Post Reply Quote adrien Report Post Thanks(0) Quote Reply Topic: Pane Manager bug in 17.2.0 Posted: 19 May 2016 at 10:20pm
	I had a resize bug, where if a splitter was moved, the growing pane would not redraw until some other event caused a repaint. I tracked it down to the filtering of messages whilst in capture for the splitter move. The code was calling ValidateRect on the window, which prevented repaints from happening. I commented it out and now it's happy and so am I. Even 2 paints would be better than none. I haven't checked if it's getting 2, but it's smooth as silk. BOOL CXTPDrawHelpers::ProcessPendingPaintMessages(HWND hWnd /= NULL/) { BOOL bResult = TRUE; MSG msg; while (::PeekMessage(&msg, hWnd, WM_PAINT, WM_PAINT, PM_NOREMOVE) \|\| ::PeekMessage(&msg, hWnd, WM_ERASEBKGND, WM_ERASEBKGND, PM_NOREMOVE) \|\| ::PeekMessage(&msg, hWnd, WM_SYNCPAINT, WM_SYNCPAINT, PM_NOREMOVE) \|\| ::PeekMessage(&msg, hWnd, WM_NCPAINT, WM_NCPAINT, PM_NOREMOVE)) { if (WM_QUIT != msg.message) { if (::PeekMessage(&msg, hWnd, msg.message, msg.message, PM_REMOVE)) { if (WM_QUIT != msg.message) // WM_QUIT should never happen here but still extra verification won't harm { ::DispatchMessage(&msg); // This ensures the window won't enter the second painting cycle. // ::ValidateRect(msg.hwnd, NULL); } else { // Re-post WM_QUIT as it's been removed. ::PostQuitMessage(static_cast<int>(msg.wParam)); bResult = FALSE; break; } } } else { bResult = FALSE; break; } } return bResult; }
	http://www.wingate.com