On 3/4/19 8:23 PM, Nikolay Sivov wrote:
On 3/4/19 7:25 PM, Gabriel Ivăncescu wrote:
@@ -772,7 +772,6 @@ static LRESULT CALLBACK BUTTON_WindowProc(HWND hWnd, UINT uMsg, WPARAM wParam, L case WM_KILLFOCUS: TRACE("WM_KILLFOCUS %p\n",hWnd); infoPtr->state &= ~BST_FOCUS; - paint_button( infoPtr, btn_type, ODA_FOCUS ); if ((infoPtr->state & BUTTON_BTNPRESSED) && GetCapture() == hWnd) ReleaseCapture();
That looks like an artifact of user32 behavior when it often repaints directly without going through invalidation. If it's not how it works, it's obviously good to change.
Just to be clear, I shouldn't make this change later in the user32 code, right?