[PATCH 07/13] dix: reshuffle WindowOptRec to fill a hole on LP64

[Matt Turner]

On Tue, Nov 23, 2010 at 7:45 PM, Adam Jackson <ajax at redhat.com> wrote:
> Signed-off-by: Adam Jackson <ajax at redhat.com>
> ---
>  include/windowstr.h |    2 +-
>  1 files changed, 1 insertions(+), 1 deletions(-)
>
> diff --git a/include/windowstr.h b/include/windowstr.h
> index 29b95fc..0b66ebb 100644
> --- a/include/windowstr.h
> +++ b/include/windowstr.h
> @@ -79,8 +79,8 @@ typedef struct _DevCursorNode {
>  } DevCursNodeRec, *DevCursNodePtr, *DevCursorList;
>
>  typedef struct _WindowOpt {
> -    VisualID           visual;            /* default: same as parent */
>     CursorPtr          cursor;            /* default: window.cursorNone */
> +    VisualID           visual;            /* default: same as parent */
>     Colormap           colormap;          /* default: same as parent */
>     Mask               dontPropagateMask; /* default: window.dontPropagate */
>     Mask               otherEventMasks;   /* default: 0 */
> --
> 1.7.3.1

Probably just me not spotting the obvious, but I see

xserver $ git grep 'VisualID' | grep typedef
hw/dmx/dmxclient.h:typedef unsigned long VisualID64;
hw/xnest/Xnest.h:typedef unsigned long VisualID64;

So if CursorPtr is a pointer, it seems like VisualID will always be
the same size as CursorPtr. I guess they'd be different for ABIs like
MIPS/n32. I don't guess it hurts either way, as I don't know any
systems where sizeof(pointer) < sizeof(long).

Maybe I'm not finding the correct definition of VisualID?

Matt