delay render thread for now to prevent race condition on startup

render.c

@@ -28,6 +28,8 @@ BOOL detect_cutscene();
 
 DWORD WINAPI render_main(void)
 {
+    Sleep(500);
+    
     int i,j;
     HGLRC hRC;
 

render_soft.c

@@ -51,6 +51,8 @@ BOOL detect_cutscene()
 
 DWORD WINAPI render_soft_main(void)
 {
+    Sleep(500);
+    
     PBITMAPINFO bmi = HeapAlloc(GetProcessHeap(), HEAP_ZERO_MEMORY, sizeof(BITMAPINFOHEADER) + sizeof(RGBQUAD) * 256);
 
     bmi->bmiHeader.biSize = sizeof(BITMAPINFOHEADER);

surface.c

@@ -121,9 +121,8 @@ HRESULT __stdcall ddraw_surface_Blt(IDirectDrawSurfaceImpl *This, LPRECT lpDestR
             for(y=y0; y<y1; ++y, to+=This->width, from+=Source->width)
                 memcpy(to, from, s); 
             
-            LeaveCriticalSection(&ddraw->cs);
-            
             ReleaseSemaphore(ddraw->render.sem, 1, NULL);
+            LeaveCriticalSection(&ddraw->cs);
         }
         else
         {