diff --git a/scripts/refcounter.py b/scripts/refcounter.py
new file mode 100755
index 00000000..adafce81
--- /dev/null
+++ b/scripts/refcounter.py
@@ -0,0 +1,111 @@
+#!/usr/bin/env python
+"""Detect reference leaks after several unit test runs.
+
+The script runs the unit test and counts the objects alive after the run. If
+the object count differs between the last two runs, a report is printed and the
+script exits with error 1.
+"""
+
+# Copyright (C) 2011 Daniele Varrazzo <daniele.varrazzo@gmail.com>
+#
+# This program is free software; you can redistribute it and/or modify
+# it under the terms of the GNU General Public License as published by
+# the Free Software Foundation; either version 2 of the License, or
+# (at your option) any later version.
+#
+# This program is distributed in the hope that it will be useful,
+# but WITHOUT ANY WARRANTY; without even the implied warranty of
+# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+# GNU General Public License for more details.
+#
+# You should have received a copy of the GNU General Public License
+# along with this program; if not, write to the Free Software
+# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
+
+import gc
+import sys
+import difflib
+import unittest
+from pprint import pprint
+from collections import defaultdict
+
+def main():
+    opt = parse_args()
+
+    import psycopg2.tests
+    test = psycopg2.tests
+    if opt.suite:
+        test = getattr(test, opt.suite)
+
+    sys.stdout.write("test suite %s\n" % test.__name__)
+
+    for i in range(1, opt.nruns + 1):
+        sys.stdout.write("test suite run %d of %d\n" % (i, opt.nruns))
+        runner = unittest.TextTestRunner()
+        runner.run(test.test_suite())
+        dump(i, opt)
+
+    f1 = open('debug-%02d.txt' % (opt.nruns - 1)).readlines()
+    f2 = open('debug-%02d.txt' % (opt.nruns)).readlines()
+    for line in difflib.unified_diff(f1, f2,
+            "run %d" % (opt.nruns - 1), "run %d" % opt.nruns):
+        sys.stdout.write(line)
+
+    rv = f1 != f2 and 1 or 0
+
+    if opt.objs:
+        f1 = open('objs-%02d.txt' % (opt.nruns - 1)).readlines()
+        f2 = open('objs-%02d.txt' % (opt.nruns)).readlines()
+        for line in difflib.unified_diff(f1, f2,
+                "run %d" % (opt.nruns - 1), "run %d" % opt.nruns):
+            sys.stdout.write(line)
+
+    return rv
+
+def parse_args():
+    import optparse
+
+    parser = optparse.OptionParser(description=__doc__)
+    parser.add_option('--nruns', type='int', metavar="N", default=3,
+        help="number of test suite runs [default: %default]")
+    parser.add_option('--suite', metavar="NAME",
+        help="the test suite to run (e.g. 'test_cursor'). [default: all]")
+    parser.add_option('--objs', metavar="TYPE",
+        help="in case of leaks, print a report of object TYPE "
+            "(support still incomplete)")
+
+    opt, args = parser.parse_args()
+    return opt
+
+
+def dump(i, opt):
+    gc.collect()
+    objs = gc.get_objects()
+
+    c = defaultdict(int)
+    for o in objs:
+        c[type(o)] += 1
+
+    pprint(
+        sorted(((v,str(k)) for k,v in c.items()), reverse=True),
+        stream=open("debug-%02d.txt" % i, "w"))
+
+    if opt.objs:
+        co = []
+        t = getattr(__builtins__, opt.objs)
+        for o in objs:
+            if type(o) is t:
+                co.append(o)
+
+        # TODO: very incomplete
+        if t is dict:
+            co.sort(key = lambda d: d.items())
+        else:
+            co.sort()
+
+        pprint(co, stream=open("objs-%02d.txt" % i, "w"))
+
+
+if __name__ == '__main__':
+    sys.exit(main())
+