Clone needs to handle regular expressions as well PYTHON-421

Author: rozza

bson/json_util.py

@@ -83,7 +83,7 @@
         json_lib = False
 
 import bson
-from bson import EPOCH_AWARE
+from bson import EPOCH_AWARE, RE_TYPE
 from bson.binary import Binary
 from bson.code import Code
 from bson.dbref import DBRef
@@ -94,9 +94,6 @@
 
 from bson.py3compat import PY3, binary_type, string_types
 
-# TODO share this with bson.py?
-_RE_TYPE = type(re.compile("foo"))
-
 
 def dumps(obj, *args, **kwargs):
     """Helper function that wraps :class:`json.dumps`.
@@ -174,7 +171,7 @@ def default(obj):
         millis = int(calendar.timegm(obj.timetuple()) * 1000 +
                      obj.microsecond / 1000)
         return {"$date": millis}
-    if isinstance(obj, _RE_TYPE):
+    if isinstance(obj, RE_TYPE):
         flags = ""
         if obj.flags & re.IGNORECASE:
             flags += "i"

pymongo/cursor.py

@@ -16,6 +16,7 @@
 import copy
 from collections import deque
 
+from bson import RE_TYPE
 from bson.code import Code
 from bson.son import SON
 from pymongo import helpers, message, read_preferences
@@ -184,7 +185,7 @@ def __clone(self, deepcopy=True):
         data = dict((k, v) for k, v in self.__dict__.iteritems()
                     if k.startswith('_Cursor__') and k[9:] in values_to_clone)
         if deepcopy:
-            data = copy.deepcopy(data)
+            data = self.__deepcopy(data)
         clone.__dict__.update(data)
         return clone
 
@@ -823,3 +824,24 @@ def __deepcopy__(self, memo):
         .. versionadded:: 2.3+
         """
         return self.__clone(deepcopy=True)
+
+    def __deepcopy(self, x, memo=None):
+        """Deepcopy helper for the data dictionary.
+
+        Regular expressions cannot be deep copied but as they are immutable we
+        don't have to copy them when cloning.
+        """
+        y = {}
+        if memo is None:
+            memo = {}
+        val_id = id(x)
+        if val_id in memo:
+            return memo.get(val_id)
+        memo[val_id] = y
+        for key, value in x.iteritems():
+            if isinstance(value, dict):
+                value = self.__deepcopy(value, memo)
+            elif not isinstance(value, RE_TYPE):
+                value = copy.deepcopy(value, memo)
+            y[copy.deepcopy(key, memo)] = value
+        return y

test/test_cursor.py

@@ -16,6 +16,7 @@
 import copy
 import itertools
 import random
+import re
 import sys
 import unittest
 sys.path[0:0] = [""]
@@ -464,7 +465,8 @@ class MyClass(dict):
         self.assertEqual(type(MyClass()), type(cursor[0]))
 
         # Just test attributes
-        cursor = self.db.test.find(skip=1,
+        cursor = self.db.test.find({"x": re.compile("^hello.*")},
+                                   skip=1,
                                    timeout=False,
                                    snapshot=True,
                                    tailable=True,
@@ -508,6 +510,19 @@ class MyClass(dict):
         cursor4._Cursor__fields['cursor4'] = False
         self.assertFalse('cursor4' in cursor._Cursor__fields)
 
+        # Test memo when deepcopying queries
+        query = {"hello": "world"}
+        query["reflexive"] = query
+        cursor = self.db.test.find(query)
+
+        cursor2 = copy.deepcopy(cursor)
+
+        self.assertNotEqual(id(cursor._Cursor__spec),
+                            id(cursor2._Cursor__spec))
+        self.assertEqual(id(cursor2._Cursor__spec['reflexive']),
+                         id(cursor2._Cursor__spec))
+        self.assertEqual(len(cursor2._Cursor__spec), 2)
+
     def test_add_remove_option(self):
         cursor = self.db.test.find()
         self.assertEqual(0, cursor._Cursor__query_options())