Isue #5084: unpickling now interns the attribute names of pickled objects,

saving memory and avoiding growth in size of subsequent pickles. Proposal
and original patch by Jake McGuire.

diff --git a/Lib/pickle.py b/Lib/pickle.py
index abed1ca4fab0ab4497212fbd7e199a1e79db9cfb..8c5d51fc15be2691aa4ecea6286cf176cb4acd8a 100644 (file)
--- a/Lib/pickle.py
+++ b/Lib/pickle.py
@@ -1221,7 +1221,15 @@ class Unpickler:
             state, slotstate = state
         if state:
             try:
-                inst.__dict__.update(state)
+                d = inst.__dict__
+                try:
+                    for k, v in state.iteritems():
+                        d[intern(k)] = v
+                # keys in state don't have to be strings
+                # don't blow up, but don't go out of our way
+                except TypeError:
+                    d.update(state)
+
             except RuntimeError:
                 # XXX In restricted execution, the instance's __dict__
                 # is not accessible.  Use the old way of unpickling

diff --git a/Lib/test/pickletester.py b/Lib/test/pickletester.py
index 4ffa7028400a2f08bf2849add229d783de84bb87..bc0be1f845c50a3c00d18d8f09ed0dde3adfbb9e 100644 (file)
--- a/Lib/test/pickletester.py
+++ b/Lib/test/pickletester.py
@@ -938,6 +938,20 @@ class AbstractPickleTests(unittest.TestCase):
                              "Failed protocol %d: %r != %r"
                              % (proto, obj, loaded))
 
+    def test_attribute_name_interning(self):
+        # Test that attribute names of pickled objects are interned when
+        # unpickling.
+        for proto in protocols:
+            x = C()
+            x.foo = 42
+            x.bar = "hello"
+            s = self.dumps(x, proto)
+            y = self.loads(s)
+            x_keys = sorted(x.__dict__)
+            y_keys = sorted(y.__dict__)
+            for x_key, y_key in zip(x_keys, y_keys):
+                self.assertIs(x_key, y_key)
+
 
 # Test classes for reduce_ex
 

diff --git a/Misc/NEWS b/Misc/NEWS
index 1ccb739edc7f4920b0e9ae3e159b5cea4f7a6fda..e7516bb01164f3148e5fabfcfd39253042f5b1f4 100644 (file)
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -261,6 +261,10 @@ Core and Builtins
 Library
 -------
 
+- Issue #5084: unpickling now interns the attribute names of pickled objects,
+  saving memory and avoiding growth in size of subsequent pickles. Proposal
+  and original patch by Jake McGuire.
+
 - Issue #3002: ``shutil.copyfile()`` and ``shutil.copytree()`` now raise an
   error when a named pipe is encountered, rather than blocking infinitely.
 

diff --git a/Modules/cPickle.c b/Modules/cPickle.c
index 6c7ed9987e1f3f07752e84af715d2aa1cbb10b8c..a0e443ed940e9aa208bd723958d77b7ece610972 100644 (file)
--- a/Modules/cPickle.c
+++ b/Modules/cPickle.c
@@ -4473,8 +4473,16 @@ load_build(Unpicklerobject *self)
 
                i = 0;
                while (PyDict_Next(state, &i, &d_key, &d_value)) {
-                       if (PyObject_SetItem(dict, d_key, d_value) < 0)
+                       /* normally the keys for instance attributes are
+                          interned.  we should try to do that here. */
+                       Py_INCREF(d_key);
+                       if (PyString_CheckExact(d_key))
+                               PyString_InternInPlace(&d_key);
+                       if (PyObject_SetItem(dict, d_key, d_value) < 0) {
+                               Py_DECREF(d_key);
                                goto finally;
+                       }
+                       Py_DECREF(d_key);
                }
                Py_DECREF(dict);
        }