On Wed, 16 Mar 2005 06:34:44 +0100
Joerg Sonnenberger <joerg@britannica.bec.de> wrote:
> On Tue, Mar 15, 2005 at 09:21:36PM -0800, Chris Pressey wrote:
> > > I think that should be
> > >
> > > splim = (u_short *)sa->sa_data +
> > > (sa->sa_len - offsetof(sa->sa_data)) / sizeof(u_short);
> > >
> > > but that'd be the same as the thing I committed, just longer.
> >
> > (With the exception of the "+ 1", which AFAICT is there so that the
> > division by sizeof(u_short) (or ">> 1" in the old version) doesn't cut
> > off the last short if sa_len is odd...)
>
> You are right with the versions being functionally identical.
> The compiler should create the same code, the difference is
> which the limit is based on.
>
> What I meant about the u_short usage is, that the hex printing should be
> done byte wise, not word wise.
That simplifies things a lot. Thanks.
-Chris