Re: ws2_32/socket16.c:319

[Marcus Meissner]

On Thu, Oct 12, 2006 at 10:06:39AM +0200, Hans Leidekker wrote:
> On Thursday 12 October 2006 03:58, Michael [Plouj] Ploujnikov wrote:
> > INT16 WINAPI ioctlsocket16(SOCKET16 s, LONG cmd, ULONG *argp)
> > {
> >     return WS_ioctlsocket( s, cmd, argp ); /* <- line 319 */
> > }
> > Is it wrong "BSD-only" u_long in this function? Should WS_ioctlsocket
> > be fixed to use a ULONG pointer?
>
> No that would just move the problem, because WS_ioctlsocket is also the
> implementation for ioctlsocket, which takes a pointer to u_long. Maybe
> something like this will do?
>
> INT16 WINAPI ioctlsocket16(SOCKET16 s, LONG cmd, ULONG *argp)
> {
>     u_long arg = *argp;
>     return WS_ioctlsocket( s, cmd, &arg );
> }

argp is for in- and output.

I would just suggest to cast the warning away.

Ciao, Marcus