NTLMSSP fix for source4

[Jeremy Allison]

On Wed, May 19, 2010 at 11:20:57AM +1000, Andrew Bartlett wrote:
> On Tue, 2010-05-18 at 17:41 -0700, Jeremy Allison wrote:
> > On Wed, May 19, 2010 at 10:44:18AM +1000, Andrew Bartlett wrote:
> > > On Tue, 2010-05-18 at 16:55 -0700, Jeremy Allison wrote:
> > > > Andrew, please review and push. Same fix that
> > > > fixes the Microsoft torture tester I put into
> > > > source3.
> > > 
> > > Don't we have a better explanation for the version than just a series of
> > > bytes?  If not, we should get them from dochelp before we put yet
> > > another magic constant into the code. 
> > 
> > It's in the MS-NTLM doc. 
> > 
> > > Also, is this handled by the IDL parser?  We want to move to it soon, so
> > > we need that updated as well.
> > 
> > Yes, VERSION is handled in our idl.
> 
> I think the correct fix is to then declare a struct VERSION, fill in the
> elements with the constants from ntlmssp.idl and then do an
> ndr_push_struct_blob.  If VERSION isn't negotiated, then we should, as
> far as I can tell, be able to use the same msprc_gen call, but with a
> data_blob_null. 
> 
> You should be able to do the same in the source3 code, as we are trying
> hard to keep them identical, and they both have access to the same IDL.
> 
> I hope this helps.  I'm sorry for the extra work, but I really dislike
> unexplained magic constant byte arrays.

+1 on this. You are correct - thanks. I'll fix this up properly tomorrow
(other Microsoft raised bugs pending...).

Jeremy