no name services! rpc_ns_binding_xxx() is blank!

[Luke Kenneth Casson Leighton]

On Fri, Oct 19, 2001 at 10:47:48AM +1000, Luke Howard wrote:
> 
> >having looked at this a bit more, the rpc_ns_binding* api
> >is basically like a remote, distributed version
> >of the endpoint mapper.
> 
> Ah, I didn't realize this was separate from the endpoint
> mapper. 

yes.

in fact, what i have to do in the "hacked-up" implementation
of rpc_ns_binding_begin & next is to compare the
service uuid (e.g. the uuid for lsarpc or netlogon) to the
incoming uuid, and then call the endpoint "constructor / helper"
routines to create a binding, along with resolving the
interface .

function that has to be called inside rpc_ns_binding_next()
is rpc_ep_resolve_binding().

basically, all the setup "crap" that you have to do
in "explicit" handles is moved into rpc_ns_binding_next().

that's all, really.

> The endpoint mapper is more analogous to ONC RPC's
> portmap, then? 
 
yes.


> I can definitely take a look at the LDAP RpcNS stuff soon,
> but would working on the GSS-API authentication mechanism
> be more useful?

GSS-API.

i have a basic hack-me-up bit of code to keep me going,
and enough info to be able to do a decent job, later.

i want to get a freedce-native version of check_domain_security()
out as fast as possible, it's like at the bottom of
the critical path dependencies, scraping the barrel...

still can't believe quite how much all of this stuff just
hangs together, and all of a sudden baaam, it'll just
all work :)

lkcl