Recursion Desired.

[Allen, Michael B (RSCH)]

> -----Original Message-----
> From:	Christopher R. Hertel [SMTP:crh at ubiqx.mn.org]
> 
> > But anyway, in general I think you give
> > implementors too much credit. Why bother to have RD off if it doesn't
> > give you anything? I made RD a settable Constructor parameter to the
> > NameQueryRequest packet in jCIFS but if you look at NameServiceClient
> > it's hard coded to true. This is because I was simply never compelled to
> > turn it off. Was always puzzling to me as to why smbclient makes you
> > flip it on with -R.
> 
> nmblookup
> 
	Oh, yeah, nmblookup I mean. Not smbclient.

>  turns on RD if you are doing a broadcast query.  I had to write
> my own code to turn it off.
> 
	Yes, I see. This is just backwards. For broadcast it should by default be off
	and for unicast it should be on. Or like I mentioned before it could just
	always be on. That should really be fixed actually because obviously someone
	doing:

	$ nmblookup -U 1.2.3.4

	where 1.2.3.4 is the WINS server is going to get confused when they get
	nada because RD is off. I'm sure I was. Actually it would be nice if the WINS
	address was snarfed out of smb.conf too.

	Mike

> For Unicast queries, nmblookup assumes that you are querying an endstation
> rather than an NBNS.  You have to turn on RD in order to query the NBNS.
> 
	backwards is all I have to say :~)