Brown-Paper-Bag bugs in the winbind ntlm auth_helper

Since I've put live my winbindd setup, I've discovered two embarassing
truths about that code:

1) It doesn't like NTLMv2 at all, refusing access to anybody who dares
provide NTLMv2 credentials
2) (much more embarassing) due to an one-liner it ends up ignoring NT
credentals, always falling back to LM even when the NThash is provided.

Here's the patch. Please apply to the -PRE tree before -STABLE release.

--- wb_ntlm_auth.c.orig Tue Sep 17 09:49:40 2002
+++ wb_ntlm_auth.c Tue Sep 17 17:08:22 2002
@@ -219,21 +219,24 @@
        authfail(domain, user, "Broken LM hash response");
        return;
     }
-
+ have_nthash=0;
     nthash = ntlm_fetch_string((char *) auth, auth_length, &auth->ntresponse);
     switch (nthash.l) {
     case 0:
        debug("no nthash\n");
        request.data.auth_crap.nt_resp_len = 0;
+ have_nthash=0;
        break;
     case 24:
        memcpy(request.data.auth_crap.nt_resp, nthash.str, 24);
        request.data.auth_crap.nt_resp_len = 24;
+ have_nthash=1;
        break;
     default:
- debug("nthash len = %d\n", nthash.l);
- authfail(domain, user, "Broken NT hash response");
- return;
+ debug("nthash len=%d. Ignoring it.\n",nthash.l);
+ request.data.auth_crap.nt_resp_len = 0;
+ have_nthash=0;
+ break;
     }

     debug("Checking user '%s\\%s' lmhash len =%d, have_nthash=%d, "

-- 
	kinkie (kinkie-ml [at] libero [dot] it)
	Random fortune, unrelated to the message:
He is the best of men who dislikes power.
		-- Mohammed