Win98 connection - MS quality/compability

Christopher R. Hertel crh at
Fri Jan 4 10:08:02 GMT 2002

> > > Andrew Bartlett wrote:
> > > > but do you think anybody ever tests against a
> > > > machine thats intentationaly broken?
> > > 
> > > You have _no_ idea of the crazy stuff
> > > we have done with our mashines ;)
> > 
> > This was more of a reference to the 'if MS didn't test it (ie doesn't
> > happen with their client/server) it *will* be buggy' syndrome :-).
> Why should they ?
> They are following the RFC, aren't they ? ;)

As your wink suggests... No, Microsoft's implementation does *not* follow 
the RFCs.  There are implementation errors.

> But they will behave now after the DOJ verdic ? =D

The deadline for comments is January 26th.

Regarding this thread:  The fact that W/NT and W2K, etc, can connect to 
smbd when nmbd is not running is the result of NT and 2K being kludged to 
work.  Nmbd creates the NBT virtual LAN that should exist to support the

Knowing a good deal about the NBT layer, however, I would say that it
should be very easy to get the session service to work without nmbd.  In 
fact, SMB over native TCP/IP (port 445) does essentially that.

So, do another trace and watch particularly on ports UDP/137 and 
UDP/138.  My bet is that W/9x is doing as Andrew B suggests:  some sort 
of NetBIOS query (adapter status, most likely).  I'd further guess that 
W/9x is doing this to try to find the service name (looking for a <20> 
name in the name list).  When it doesn't get a response, it gives up.  
W/NT and W2K will try the "*SMBSERVER" name in the session setup.

Chris -)-----

Christopher R. Hertel -)-----                   University of Minnesota
crh at              Networking and Telecommunications Services

    Ideals are like stars; you will not succeed in touching them
    with your choose them as your guides, and following
    them you will reach your destiny.  --Carl Schultz

More information about the samba-technical mailing list