Re: OLE32: fix gcc 4.0 -Wpointer-sign warnings

[Robert Shearman]

Mike McCormack wrote:
> ChangeLog:
> - fix gcc 4.0 -Wpointer-sign warnings
> Index: dlls/ole32/antimoniker.c
>
> RCS file: /home/wine/wine/dlls/ole32/antimoniker.c,v
> retrieving revision 1.25
> diff -u -p -r1.25 antimoniker.c
> --- dlls/ole32/antimoniker.c	6 Jun 2005 19:50:37 -0000	1.25
> +++ dlls/ole32/antimoniker.c	4 Jul 2005 14:48:34 -0000
> @@ -49,7 +49,7 @@ typedef struct AntiMonikerImpl{
>      */
>     const IROTDataVtbl*  lpvtbl2;  /* VTable relative to the IROTData interface.*/
> - ULONG ref; /* reference counter for this object */
> - LONG ref; /* reference counter for this object */
>
> } AntiMonikerImpl;

IUnknown_AddRef and IUnknown_Release both return ULONGs so it looks like 
you are just moving the problem elsewhere. I think you should be casting 
the values passed into Interlocked{Increment,Decrement} or find some 
other nicer way of preventing that warning, possibly by ignoring them or 
just turning off the -Wpointer-sign flag.

OLE32: fix gcc 4.0 -Wpointer-sign warnings

[Mike McCormack]

Robert Shearman wrote:
> > - ULONG ref; /* reference counter for this object */
> > - LONG ref; /* reference counter for this object */
> IUnknown_AddRef and IUnknown_Release both return ULONGs so it looks like 
> you are just moving the problem elsewhere. I think you should be casting 
> the values passed into Interlocked{Increment,Decrement} or find some 
> other nicer way of preventing that warning, possibly by ignoring them or 
> just turning off the -Wpointer-sign flag.

I don't think the reference count will get over MAXINT and it shouldn't 
go negative, so I can't see how this will cause a problem.

I already submitted a patch [1] to turn off -Wpointer-sign. Alexandre 
prefered that we fix the warnings.

Mike

[1] http://www.winehq.org/hypermail/wine-patches/2005/07/0003.html