Fix XML handling of member links (bug 1156594)

This is a was revealed as a blocker for the default_project_id scoping
tests.

Also improves test coverage of XML de/serialization invertibility. The
tests were previously sensitive to things like attribute ordering which
are not relevant to test results. The fix is to compare canonically
serialized XML instead of whatever lxml spews out.

Change-Id: I003583038421d35aadbc781144d4cafb09392db5

1 files changed, 18 insertions, 35 deletions

diff --git a/keystone/common/serializer.py b/keystone/common/serializer.py
index e5f21a73..9b9df418 100644
--- a/keystone/common/serializer.py
+++ b/keystone/common/serializer.py
@@ -71,35 +71,10 @@ class XmlDeserializer(object):
     def __call__(self, xml_str):
         """Returns a dictionary populated by decoding the given xml string."""
         dom = etree.fromstring(xml_str.strip(), PARSER)
-        links_json = self._find_and_remove_links_from_root(dom, True)
-        obj_json = self.walk_element(dom, True)
-        if links_json:
-            obj_json['links'] = links_json['links']
-        return obj_json
+        return self.walk_element(dom, True)
 
-    def _deserialize_links(self, links, links_json):
-        for link in links:
-            links_json['links'][link.attrib['rel']] = link.attrib['href']
-
-    def _find_and_remove_links_from_root(self, dom, namespace):
-        """Special-case links element
-
-        If "links" is in the elements, convert it and remove it from root
-        element. "links" will be placed back into the root of the converted
-        JSON object.
-
-        """
-        for element in dom:
-            decoded_tag = XmlDeserializer._tag_name(element.tag, namespace)
-            if decoded_tag == 'links':
-                links_json = {'links': {}}
-                self._deserialize_links(element, links_json)
-                dom.remove(element)
-                # TODO(gyee): are 'next' and 'previous' mandatory? If so,
-                # setting them to None if they don't exist?
-                links_json['links'].setdefault('previous')
-                links_json['links'].setdefault('next')
-                return links_json
+    def _deserialize_links(self, links):
+        return dict((x.attrib['rel'], x.attrib['href']) for x in links)
 
     @staticmethod
     def _tag_name(tag, namespace):
@@ -161,23 +136,31 @@ class XmlDeserializer(object):
             else:
                 list_item_tag = decoded_tag[:-1]
 
-        # links is a special dict
         if decoded_tag == 'links':
-            links_json = {'links': {}}
-            self._deserialize_links(element, links_json)
-            return links_json
+            return {'links': self._deserialize_links(element)}
 
+        links = None
         for child in [self.walk_element(x) for x in element
                       if not isinstance(x, ENTITY_TYPE)]:
             if list_item_tag:
-                # FIXME(gyee): special-case lists for now unti we
+                # FIXME(gyee): special-case lists for now until we
                 # figure out how to properly handle them.
                 # If any key ends with an 's', we are assuming it is a list.
-                values.append(child[list_item_tag])
+                if list_item_tag in child:
+                    values.append(child[list_item_tag])
+                else:
+                    links = child['links']
             else:
                 values = dict(values.items() + child.items())
 
-        return {XmlDeserializer._tag_name(element.tag, namespace): values}
+        d = {XmlDeserializer._tag_name(element.tag, namespace): values}
+
+        if links:
+            d['links'] = links
+            d['links'].setdefault('next')
+            d['links'].setdefault('previous')
+
+        return d
 
 
 class XmlSerializer(object):