[Swan] an xauth logging message

[Paul Wouters]

On Fri, 1 Mar 2013, D. Hugh Redelmeier wrote:

> Are xauth_server and xauth_client mutually exclusive?
> Are modecfg_server and modecfg_client mutually exclusive?

Yes both are mutually exclusive. but the two are now. That is,
you can be an xauth server but a modecfg_client (it is called
pull or push mode but that term too is confusing when you think
that initiator/responder can flip around as well)

> If so, perhaps a PASSERT and a simpler message is in order.
>
> Assuming all are independent, I think that this generates a simpler
> message:
>
> 	    DBG(DBG_CONTROLMORE , DBG_log("this is a%s%s%s%s in state %s. "
> 		"Reply with UNSUPPORTED_EXCHANGE_TYPE"
> 		, st->st_connection->spd.this.xauth_server ? " xauthserver" : ""
> 		, st->st_connection->spd.this.xauth_client ? " xauthclient" : ""
> 		, st->st_connection->spd.this.modecfg_server ? " modecfgserver" : ""
> 		, st->st_connection->spd.this.modecfg_client  ? " modecfgclient" : ""
> 		, enum_name(&state_names, st->st_state)
> 		));
>
> Admittedly, this looks funny if none of the fields is true, but I
> would guess this never happens.  Am I wrong?

Nope, that's right. I'm fine with this change. I prefer it over assuming
the server-client is exclusive for logging, though I guess it _should_
never happen and in that sense a passert is justified.

Paul