Skip blob objects during Python 3 migration. (#7)

CHANGES.rst

@@ -4,6 +4,9 @@ Changes
 1.1 (unreleased)
 ----------------
 
+- Skip records for ZODB.blob when migrating database to python3 to not break
+  references to blobfiles.
+
 - When migrating databases to python3, do not fail when converting
   attributes containing None.
 

src/zodbupdate/serialize.py

@@ -25,6 +25,9 @@
 logger = logging.getLogger('zodbupdate.serialize')
 known_broken_modules = {}
 
+# types to skip when renaming/migrating databases
+SKIP_SYMBS = [('ZODB.blob', 'Blob')]
+
 
 def create_broken_module_for(symb):
     """If your pickle refer a broken class (not an instance of it, a
@@ -37,7 +40,7 @@ def create_broken_module_for(symb):
     parts = symb.__module__.split('.')
     previous = None
     for fullname, name in reversed(
-            [('.'.join(parts[0:p+1]), parts[p])
+            [('.'.join(parts[0:p + 1]), parts[p])
              for p in range(1, len(parts))]):
         if fullname not in sys.modules:
             if fullname not in known_broken_modules:
@@ -155,7 +158,10 @@ def __update_symb(self, symb_info):
         loaded and its location is checked to see if it have moved as
         well.
         """
-        if symb_info in self.__renames:
+        if symb_info in SKIP_SYMBS:
+            self.__skipped = True
+            return symb_info
+        elif symb_info in self.__renames:
             self.__changed = True
             return self.__renames[symb_info]
         else:
@@ -292,13 +298,18 @@ def rename(self, input_file):
         return None otherwise.
         """
         self.__changed = False
+        self.__skipped = False
 
         unpickler = self.__unpickler(input_file)
         class_meta = unpickler.load()
         data = unpickler.load()
 
         class_meta = self.__update_class_meta(class_meta)
 
+        if self.__skipped:
+            # do not do renames/conversions on blob records
+            return None
+
         self.__decode_data(class_meta, data)
 
         if not (self.__changed or self.__repickle_all):

src/zodbupdate/tests.py

@@ -27,6 +27,7 @@
 import six
 import zope.interface
 import zodbupdate.main
+import zodbupdate.serialize
 
 
 class TestLogHandler(object):
@@ -94,6 +95,8 @@ class OtherFactory(persistent.Persistent):
         self.conn = self.db.open()
         self.root = self.conn.root()
 
+        self._skipped_symbs = zodbupdate.serialize.SKIP_SYMBS
+
     def update(self, **args):
         self.conn.close()
         self.db.close()
@@ -132,6 +135,8 @@ def tearDown(self):
         os.unlink(self.dbfile + '.tmp')
         os.unlink(self.dbfile + '.lock')
 
+        zodbupdate.serialize.SKIP_SYMBS = self._skipped_symbs
+
     def test_no_transaction_if_no_changes(self):
         # If an update run doesn't produce any changes it won't commit the
         # transaction to avoid superfluous clutter in the DB.
@@ -175,6 +180,45 @@ def getName(self):
         renames = updater.processor.get_rules(implicit=True)
         self.assertEqual({}, renames)
 
+    def test_skipped_types_are_left_untouched(self):
+        skipped = sys.modules['module1'].Factory()
+        self.root['skipped'] = skipped
+        transaction.commit()
+
+        self.assertIn(('ZODB.blob', 'Blob'), zodbupdate.serialize.SKIP_SYMBS)
+        zodbupdate.serialize.SKIP_SYMBS += [('module1', 'Factory')]
+
+        oid = self.root['skipped']._p_oid
+        old_pickle, old_serial = self.storage.load(oid)
+
+        self.update(
+            default_renames={
+                ('module1', 'Factory'): ('module2', 'OtherFactory')})
+
+        pickle, serial = self.storage.load(oid)
+        self.assertEqual(old_pickle, pickle)
+        self.assertEqual(old_serial, serial)
+
+    def test_not_skipped_types_are_touched(self):
+        skipped = sys.modules['module1'].Factory()
+        self.root['skipped'] = skipped
+        transaction.commit()
+
+        self.assertNotIn(
+            ('module1', 'Factory'),
+            zodbupdate.serialize.SKIP_SYMBS)
+
+        oid = self.root['skipped']._p_oid
+        old_pickle, old_serial = self.storage.load(oid)
+
+        self.update(
+            default_renames={
+                ('module1', 'Factory'): ('module2', 'OtherFactory')})
+
+        pickle, serial = self.storage.load(oid)
+        self.assertNotEqual(old_pickle, pickle)
+        self.assertNotEqual(old_serial, serial)
+
 
 class Python2Tests(Tests):