diff --git a/Lib/test/test_opcodes.py b/Lib/test/test_opcodes.py
index 8fe77ac..4403a11 100644
--- a/Lib/test/test_opcodes.py
+++ b/Lib/test/test_opcodes.py
@@ -98,6 +98,12 @@
         g = eval('lambda a=1: None')
         self.assertNotEquals(f, g)
 
+    def test_modulo_of_string_subclasses(self):
+        class MyString(str):
+            def __mod__(self, value):
+                return 42
+        self.assertEqual(MyString() % 3, 42)
+
 
 def test_main():
     run_unittest(OpcodeTest)
diff --git a/Python/ceval.c b/Python/ceval.c
index d1d6d1d..e693147 100644
--- a/Python/ceval.c
+++ b/Python/ceval.c
@@ -1446,7 +1446,10 @@
 		TARGET(BINARY_MODULO)
 			w = POP();
 			v = TOP();
-			x = PyNumber_Remainder(v, w);
+			if (PyUnicode_CheckExact(v))
+				x = PyUnicode_Format(v, w);
+			else
+				x = PyNumber_Remainder(v, w);
 			Py_DECREF(v);
 			Py_DECREF(w);
 			SET_TOP(x);
diff --git a/Tools/scripts/README b/Tools/scripts/README
index f5c9791..64fc2bd 100644
--- a/Tools/scripts/README
+++ b/Tools/scripts/README
@@ -4,6 +4,7 @@
 
 See also the Demo/scripts directory!
 
+analyze_dxp.py		Analyzes the result of sys.getdxp()
 byext.py		Print lines/words/chars stats of files by extension
 byteyears.py		Print product of a file's size and age
 checkappend.py		Search for multi-argument .append() calls
diff --git a/Tools/scripts/analyze_dxp.py b/Tools/scripts/analyze_dxp.py
new file mode 100644
index 0000000..bde931e
--- /dev/null
+++ b/Tools/scripts/analyze_dxp.py
@@ -0,0 +1,129 @@
+"""
+Some helper functions to analyze the output of sys.getdxp() (which is
+only available if Python was built with -DDYNAMIC_EXECUTION_PROFILE).
+These will tell you which opcodes have been executed most frequently
+in the current process, and, if Python was also built with -DDXPAIRS,
+will tell you which instruction _pairs_ were executed most frequently,
+which may help in choosing new instructions.
+
+If Python was built without -DDYNAMIC_EXECUTION_PROFILE, importing
+this module will raise a RuntimeError.
+
+If you're running a script you want to profile, a simple way to get
+the common pairs is:
+
+$ PYTHONPATH=$PYTHONPATH:<python_srcdir>/Tools/scripts \
+./python -i -O the_script.py --args
+...
+> from analyze_dxp import *
+> s = render_common_pairs()
+> open('/tmp/some_file', 'w').write(s)
+"""
+
+import copy
+import opcode
+import operator
+import sys
+import threading
+
+if not hasattr(sys, "getdxp"):
+    raise RuntimeError("Can't import analyze_dxp: Python built without"
+                       " -DDYNAMIC_EXECUTION_PROFILE.")
+
+
+_profile_lock = threading.RLock()
+_cumulative_profile = sys.getdxp()
+
+# If Python was built with -DDXPAIRS, sys.getdxp() returns a list of
+# lists of ints.  Otherwise it returns just a list of ints.
+def has_pairs(profile):
+    """Returns True if the Python that produced the argument profile
+    was built with -DDXPAIRS."""
+
+    return len(profile) > 0 and isinstance(profile[0], list)
+
+
+def reset_profile():
+    """Forgets any execution profile that has been gathered so far."""
+    with _profile_lock:
+        sys.getdxp()  # Resets the internal profile
+        global _cumulative_profile
+        _cumulative_profile = sys.getdxp()  # 0s out our copy.
+
+
+def merge_profile():
+    """Reads sys.getdxp() and merges it into this module's cached copy.
+
+    We need this because sys.getdxp() 0s itself every time it's called."""
+
+    with _profile_lock:
+        new_profile = sys.getdxp()
+        if has_pairs(new_profile):
+            for first_inst in range(len(_cumulative_profile)):
+                for second_inst in range(len(_cumulative_profile[first_inst])):
+                    _cumulative_profile[first_inst][second_inst] += (
+                        new_profile[first_inst][second_inst])
+        else:
+            for inst in range(len(_cumulative_profile)):
+                _cumulative_profile[inst] += new_profile[inst]
+
+
+def snapshot_profile():
+    """Returns the cumulative execution profile until this call."""
+    with _profile_lock:
+        merge_profile()
+        return copy.deepcopy(_cumulative_profile)
+
+
+def common_instructions(profile):
+    """Returns the most common opcodes in order of descending frequency.
+
+    The result is a list of tuples of the form
+      (opcode, opname, # of occurrences)
+
+    """
+    if has_pairs(profile) and profile:
+        inst_list = profile[-1]
+    else:
+        inst_list = profile
+    result = [(op, opcode.opname[op], count)
+              for op, count in enumerate(inst_list)
+              if count > 0]
+    result.sort(key=operator.itemgetter(2), reverse=True)
+    return result
+
+
+def common_pairs(profile):
+    """Returns the most common opcode pairs in order of descending frequency.
+
+    The result is a list of tuples of the form
+      ((1st opcode, 2nd opcode),
+       (1st opname, 2nd opname),
+       # of occurrences of the pair)
+
+    """
+    if not has_pairs(profile):
+        return []
+    result = [((op1, op2), (opcode.opname[op1], opcode.opname[op2]), count)
+              # Drop the row of single-op profiles with [:-1]
+              for op1, op1profile in enumerate(profile[:-1])
+              for op2, count in enumerate(op1profile)
+              if count > 0]
+    result.sort(key=operator.itemgetter(2), reverse=True)
+    return result
+
+
+def render_common_pairs(profile=None):
+    """Renders the most common opcode pairs to a string in order of
+    descending frequency.
+
+    The result is a series of lines of the form:
+      # of occurrences: ('1st opname', '2nd opname')
+
+    """
+    if profile is None:
+        profile = snapshot_profile()
+    def seq():
+        for _, ops, count in common_pairs(profile):
+            yield "%s: %s\n" % (count, ops)
+    return ''.join(seq())