FIX: use identifiable instead of record

The crawler has internal checks in order to decide whether a Record can be checked for existence with the server. For this, all identifying properties need to be existent and all references of those need to be resolved or also records that can be checked themselves (so that they can be checked first).

Previously, the crawler required that ALL reference Properties can be checked or are resolved in order decide that the Record at hand can be checked.

Example: RT A with properties b and c

The registered identifiable is A with b

Record A
  b = 1234

can be checked (but the old code assumed it cannot.)

Record A
  c = 2344

cannot be checked since the identifying prop is missing.

Please consider !56 (merged)

src/newcrawler/crawl.py

@@ -204,7 +204,8 @@ class Crawler(object):
 
         Returns True otherwise.
         """
-        for p in record.properties:
+        identifiable = self.identifiableAdapter.get_identifiable(record)
+        for p in identifiable.properties:
             # TODO: implement for lists?
             if (is_reference(p) and isinstance(p.value, db.Entity)
                     and p.value.id is None):
@@ -229,7 +230,8 @@ class Crawler(object):
         returns true if all references either have IDs or were checked remotely and not found (i.e.
         they exist in the local cache)
         """
-        for p in record.properties:
+        identifiable = self.identifiableAdapter.get_identifiable(record)
+        for p in identifiable.properties:
             if (is_reference(p)
                 # Entity instead of ID and not cached locally
                 # TODO: implement for lists?