Fixed bug, which was so obvious that I am wondering how none of the test ran into that, or at least triggered it

Author: Byron

git/objects/submodule/base.py

@@ -905,7 +905,7 @@ def iter_items(cls, repo, parent_commit='HEAD'):
 				# try the index, maybe it was just added
 				try:
 					entry = index.entries[index.entry_key(p, 0)]
-					sm = cls(repo, entry.binsha, entry.mode, entry.path)
+					sm = Submodule(repo, entry.binsha, entry.mode, entry.path)
 				except KeyError:
 					raise InvalidGitRepositoryError("Gitmodule path %r did not exist in revision of parent commit %s" % (p, parent_commit))
 				# END handle keyerror

git/objects/submodule/root.py

@@ -109,7 +109,7 @@ def update(self, previous_commit=None, recursive=True, force_remove=False, init=
 
 
 		psms = self.list_items(repo, parent_commit=previous_commit)
-		sms = self.list_items(self.module())
+		sms = self.list_items(repo)
 		spsms = set(psms)
 		ssms = set(sms)