[sent also to samba-technical mailing list]
hi,
i've just been alerted to the netbios work in wine (hooray!) as part of an effort to get the network neighbourhood going (yippee!).
there is a problem, which i can help overcome.
the problem is this:
imagine that there is no TCP/IP stack in linux.
imagine that TCP/IP runs in two separate user-space services, both of which are GPL'd: the TCP service is a fork-on-demand one that binds to all interfaces, and the UDP service is even worse, it's a single-process running a state machine.
let's pretend that neither of the processes understand anything but Web services, and now you want to add SMTP, or worse, HTTPS.
you have to consider hard-coding and linking into a paired monolith of programs, and you have a licensing and logistical nightmare that is totally unacceptable for practical purposes.
this is the situation that is faced by wine, because samba, which consists of two programs, smbd and nmbd, "controls" the NetBIOS transport, from both a server _and_ (due to some bugs in Windows) also a client perspective.
a flawed attempt to make it possible to cut-in to the NetBIOS equivalent of Dynamic-DNS-plus-DHCP (the NBNS system on port 137) was added in 2000 into nmbd, however from the Wine perspective this is totally useless because the code to access the bypass mechanism is GPL'd (tdb - trivial database).
happily there is a convenient way round the mess, because it is possible to create the equivalent of a "Network Address Translation" service, and working code along these lines, which maintained a state machine of NetBIOS connections, was created over three years ago.
nmbd needs to talk via this NAT service, so does nmblookup and, once the code is all there, so does Wine's Network Neighbourhood browsing code.
at present, all you have to do to get the Wine "Network Neighbourhood" code to work is to disable samba, which is going to be totally impractical for most people.
it is not _yet_ necessary for such a service to be created, but by the time the NetBIOS code in wine is released, it will _become_ necessary.
if anyone wants to sponsor me [pay money] to revive the experiments i made in late 1999 / early 2000 in order to implement a NetBIOS "NAT" service, please contact me.
l.