Re: Extensive ScrollDC investigation (resubmit)

[Alexandre Julliard]

Aric Stewart  writes:
> @@ -120,13 +122,58 @@ BOOL CDECL X11DRV_ScrollDC( HDC hdc, INT dx, INT dy, const RECT *lprcScroll,
>      if( TRACE_ON( scroll)) dump_region( "Destination scroll region: ", DstRgn);
>      /* if there are any, do it */
>      if( res > NULLREGION) {
> -
>    RECT rect ;
> -
>    RECT rect;
> -
>    HRGN UpdateClip = CreateRectRgn(0,0,0,0);
> -
>    /* Scroll Window does not scroll any already invalid areas */
> -
>    if (GetUpdateRgn(WindowFromDC(hdc),UpdateRgn,FALSE)!=NULLREGION)
> -
>    {
> -
>        HRGN SrcMask;
> -
>        /* Apparently special case.  If the update Rgn fills the entire
> -
>           DstRgn windows appears to ignore it */

This looks very wrong. Retrieving the update region from a DC doesn't
seem right in the first place, and it you have to ignore it in some
cases it shows that there's something else going on here.