When sfi calls CreateSliver, it uses Resolve beforehand to make up the
authorThierry Parmentelat <thierry.parmentelat@inria.fr>
Fri, 21 Sep 2012 14:36:20 +0000 (16:36 +0200)
committerThierry Parmentelat <thierry.parmentelat@inria.fr>
Fri, 21 Sep 2012 14:36:20 +0000 (16:36 +0200)
'users' arg
This change reverts to using Resolve (details=True) so that
testbed-dependant data gets added to the result of Resolve
It seems like our implementation is not yet testbed-neutral enough, so
for now we need this setting

sfa/client/sfi.py

index f16a234..5d44cf3 100644 (file)
@@ -1134,7 +1134,10 @@ or with an slice hrn, shows currently provisioned resources
         #    keys: [<ssh key A>, <ssh key B>]
         #  }]
         users = []
-        slice_records = self.registry().Resolve(slice_urn, [self.my_credential_string])
+        # xxx Thierry 2012 sept. 21
+        # contrary to what I was first thinking, calling Resolve with details=False does not yet work properly here
+        # I am turning details=True on again on a - hopefully - temporary basis, just to get this whole thing to work again
+        slice_records = self.registry().Resolve(slice_urn, [self.my_credential_string], {'details':True})
         if slice_records and 'reg-researchers' in slice_records[0] and slice_records[0]['reg-researchers']:
             slice_record = slice_records[0]
             user_hrns = slice_record['reg-researchers']