Issue #4688: Add a heuristic so that tuples and dicts containing only

untrackable objects are not tracked by the garbage collector. This can
reduce the size of collections and therefore the garbage collection overhead
on long-running programs, depending on their particular use of datatypes.

(trivia: this makes the "binary_trees" benchmark from the Computer Language
Shootout 40% faster)

11 files changed, 401 insertions, 2 deletions

diff --git a/Doc/library/gc.rst b/Doc/library/gc.rst
index 9ebbf06..65f0f39 100644
--- a/Doc/library/gc.rst
+++ b/Doc/library/gc.rst
@@ -140,6 +140,31 @@ The :mod:`gc` module provides the following functions:
 
    .. versionadded:: 2.3
 
+.. function:: is_tracked(obj)
+
+   Returns True if the object is currently tracked by the garbage collector,
+   False otherwise.  As a general rule, instances of atomic types aren't
+   tracked and instances of non-atomic types (containers, user-defined
+   objects...) are.  However, some type-specific optimizations can be present
+   in order to suppress the garbage collector footprint of simple instances
+   (e.g. dicts containing only atomic keys and values)::
+
+      >>> gc.is_tracked(0)
+      False
+      >>> gc.is_tracked("a")
+      False
+      >>> gc.is_tracked([])
+      True
+      >>> gc.is_tracked({})
+      False
+      >>> gc.is_tracked({"a": 1})
+      False
+      >>> gc.is_tracked({"a": []})
+      True
+
+   .. versionadded:: 2.7
+
+
 The following variable is provided for read-only access (you can mutate its
 value but should not rebind it):
 
diff --git a/Include/dictobject.h b/Include/dictobject.h
index b83cd0e..06e0a7e 100644
--- a/Include/dictobject.h
+++ b/Include/dictobject.h
@@ -111,6 +111,7 @@ PyAPI_FUNC(PyObject *) PyDict_Copy(PyObject *mp);
 PyAPI_FUNC(int) PyDict_Contains(PyObject *mp, PyObject *key);
 PyAPI_FUNC(int) _PyDict_Contains(PyObject *mp, PyObject *key, long hash);
 PyAPI_FUNC(PyObject *) _PyDict_NewPresized(Py_ssize_t minused);
+PyAPI_FUNC(void) _PyDict_MaybeUntrack(PyObject *mp);
 
 /* PyDict_Update(mp, other) is equivalent to PyDict_Merge(mp, other, 1). */
 PyAPI_FUNC(int) PyDict_Update(PyObject *mp, PyObject *other);
diff --git a/Include/objimpl.h b/Include/objimpl.h
index ef47218..55186b1 100644
--- a/Include/objimpl.h
+++ b/Include/objimpl.h
@@ -285,6 +285,17 @@ extern PyGC_Head *_PyGC_generation0;
 	g->gc.gc_next = NULL; \
     } while (0);
 
+/* True if the object is currently tracked by the GC. */
+#define _PyObject_GC_IS_TRACKED(o) \
+	((_Py_AS_GC(o))->gc.gc_refs != _PyGC_REFS_UNTRACKED)
+ 
+/* True if the object may be tracked by the GC in the future, or already is.
+   This can be useful to implement some optimizations. */
+#define _PyObject_GC_MAY_BE_TRACKED(obj) \
+	(PyObject_IS_GC(obj) && \
+		(!PyTuple_CheckExact(obj) || _PyObject_GC_IS_TRACKED(obj)))
+
+
 PyAPI_FUNC(PyObject *) _PyObject_GC_Malloc(size_t);
 PyAPI_FUNC(PyObject *) _PyObject_GC_New(PyTypeObject *);
 PyAPI_FUNC(PyVarObject *) _PyObject_GC_NewVar(PyTypeObject *, Py_ssize_t);
diff --git a/Include/tupleobject.h b/Include/tupleobject.h
index 58479ee..a5ab733 100644
--- a/Include/tupleobject.h
+++ b/Include/tupleobject.h
@@ -44,6 +44,7 @@ PyAPI_FUNC(int) PyTuple_SetItem(PyObject *, Py_ssize_t, PyObject *);
 PyAPI_FUNC(PyObject *) PyTuple_GetSlice(PyObject *, Py_ssize_t, Py_ssize_t);
 PyAPI_FUNC(int) _PyTuple_Resize(PyObject **, Py_ssize_t);
 PyAPI_FUNC(PyObject *) PyTuple_Pack(Py_ssize_t, ...);
+PyAPI_FUNC(void) _PyTuple_MaybeUntrack(PyObject *);
 
 /* Macro, trading safety for speed */
 #define PyTuple_GET_ITEM(op, i) (((PyTupleObject *)(op))->ob_item[i])
diff --git a/Lib/test/test_dict.py b/Lib/test/test_dict.py
index 0907744..b73a53f 100644
--- a/Lib/test/test_dict.py
+++ b/Lib/test/test_dict.py
@@ -569,6 +569,104 @@ class DictTest(unittest.TestCase):
             gc.collect()
             self.assert_(ref() is None, "Cycle was not collected")
 
+    def _not_tracked(self, t):
+        # Nested containers can take several collections to untrack
+        gc.collect()
+        gc.collect()
+        self.assertFalse(gc.is_tracked(t), t)
+
+    def _tracked(self, t):
+        self.assertTrue(gc.is_tracked(t), t)
+        gc.collect()
+        gc.collect()
+        self.assertTrue(gc.is_tracked(t), t)
+
+    def test_track_literals(self):
+        # Test GC-optimization of dict literals
+        x, y, z, w = 1.5, "a", (1, None), []
+
+        self._not_tracked({})
+        self._not_tracked({x:(), y:x, z:1})
+        self._not_tracked({1: "a", "b": 2})
+        self._not_tracked({1: 2, (None, True, False, ()): int})
+        self._not_tracked({1: object()})
+
+        # Dicts with mutable elements are always tracked, even if those
+        # elements are not tracked right now.
+        self._tracked({1: []})
+        self._tracked({1: ([],)})
+        self._tracked({1: {}})
+        self._tracked({1: set()})
+
+    def test_track_dynamic(self):
+        # Test GC-optimization of dynamically-created dicts
+        class MyObject(object):
+            pass
+        x, y, z, w, o = 1.5, "a", (1, object()), [], MyObject()
+
+        d = dict()
+        self._not_tracked(d)
+        d[1] = "a"
+        self._not_tracked(d)
+        d[y] = 2
+        self._not_tracked(d)
+        d[z] = 3
+        self._not_tracked(d)
+        self._not_tracked(d.copy())
+        d[4] = w
+        self._tracked(d)
+        self._tracked(d.copy())
+        d[4] = None
+        self._not_tracked(d)
+        self._not_tracked(d.copy())
+
+        # dd isn't tracked right now, but it may mutate and therefore d
+        # which contains it must be tracked.
+        d = dict()
+        dd = dict()
+        d[1] = dd
+        self._not_tracked(dd)
+        self._tracked(d)
+        dd[1] = d
+        self._tracked(dd)
+
+        d = dict.fromkeys([x, y, z])
+        self._not_tracked(d)
+        dd = dict()
+        dd.update(d)
+        self._not_tracked(dd)
+        d = dict.fromkeys([x, y, z, o])
+        self._tracked(d)
+        dd = dict()
+        dd.update(d)
+        self._tracked(dd)
+
+        d = dict(x=x, y=y, z=z)
+        self._not_tracked(d)
+        d = dict(x=x, y=y, z=z, w=w)
+        self._tracked(d)
+        d = dict()
+        d.update(x=x, y=y, z=z)
+        self._not_tracked(d)
+        d.update(w=w)
+        self._tracked(d)
+
+        d = dict([(x, y), (z, 1)])
+        self._not_tracked(d)
+        d = dict([(x, y), (z, w)])
+        self._tracked(d)
+        d = dict()
+        d.update([(x, y), (z, 1)])
+        self._not_tracked(d)
+        d.update([(x, y), (z, w)])
+        self._tracked(d)
+
+    def test_track_subtypes(self):
+        # Dict subtypes are always tracked
+        class MyDict(dict):
+            pass
+        self._tracked(MyDict())
+
 
 from test import mapping_tests
 
diff --git a/Lib/test/test_gc.py b/Lib/test/test_gc.py
index 6e2ea41..e9af550 100644
--- a/Lib/test/test_gc.py
+++ b/Lib/test/test_gc.py
@@ -415,6 +415,37 @@ class GCTests(unittest.TestCase):
 
         self.assertEqual(gc.get_referents(1, 'a', 4j), [])
 
+    def test_is_tracked(self):
+        # Atomic built-in types are not tracked, user-defined objects and
+        # mutable containers are.
+        # NOTE: types with special optimizations (e.g. tuple) have tests
+        # in their own test files instead.
+        self.assertFalse(gc.is_tracked(None))
+        self.assertFalse(gc.is_tracked(1))
+        self.assertFalse(gc.is_tracked(1.0))
+        self.assertFalse(gc.is_tracked(1.0 + 5.0j))
+        self.assertFalse(gc.is_tracked(True))
+        self.assertFalse(gc.is_tracked(False))
+        self.assertFalse(gc.is_tracked("a"))
+        self.assertFalse(gc.is_tracked(u"a"))
+        self.assertFalse(gc.is_tracked(bytearray("a")))
+        self.assertFalse(gc.is_tracked(type))
+        self.assertFalse(gc.is_tracked(int))
+        self.assertFalse(gc.is_tracked(object))
+        self.assertFalse(gc.is_tracked(object()))
+
+        class OldStyle:
+            pass
+        class NewStyle(object):
+            pass
+        self.assertTrue(gc.is_tracked(gc))
+        self.assertTrue(gc.is_tracked(OldStyle))
+        self.assertTrue(gc.is_tracked(OldStyle()))
+        self.assertTrue(gc.is_tracked(NewStyle))
+        self.assertTrue(gc.is_tracked(NewStyle()))
+        self.assertTrue(gc.is_tracked([]))
+        self.assertTrue(gc.is_tracked(set()))
+
     def test_bug1055820b(self):
         # Corresponds to temp2b.py in the bug report.
 
diff --git a/Lib/test/test_tuple.py b/Lib/test/test_tuple.py
index 15bc295..89cbe1d 100644
--- a/Lib/test/test_tuple.py
+++ b/Lib/test/test_tuple.py
@@ -1,5 +1,7 @@
 from test import test_support, seq_tests
 
+import gc
+
 class TupleTest(seq_tests.CommonTest):
     type2test = tuple
 
@@ -82,6 +84,69 @@ class TupleTest(seq_tests.CommonTest):
         self.assertEqual(repr(a0), "()")
         self.assertEqual(repr(a2), "(0, 1, 2)")
 
+    def _not_tracked(self, t):
+        # Nested tuples can take several collections to untrack
+        gc.collect()
+        gc.collect()
+        self.assertFalse(gc.is_tracked(t), t)
+
+    def _tracked(self, t):
+        self.assertTrue(gc.is_tracked(t), t)
+        gc.collect()
+        gc.collect()
+        self.assertTrue(gc.is_tracked(t), t)
+
+    def test_track_literals(self):
+        # Test GC-optimization of tuple literals
+        x, y, z = 1.5, "a", []
+
+        self._not_tracked(())
+        self._not_tracked((1,))
+        self._not_tracked((1, 2))
+        self._not_tracked((1, 2, "a"))
+        self._not_tracked((1, 2, (None, True, False, ()), int))
+        self._not_tracked((object(),))
+        self._not_tracked(((1, x), y, (2, 3)))
+
+        # Tuples with mutable elements are always tracked, even if those
+        # elements are not tracked right now.
+        self._tracked(([],))
+        self._tracked(([1],))
+        self._tracked(({},))
+        self._tracked((set(),))
+        self._tracked((x, y, z))
+
+    def check_track_dynamic(self, tp, always_track):
+        x, y, z = 1.5, "a", []
+
+        check = self._tracked if always_track else self._not_tracked
+        check(tp())
+        check(tp([]))
+        check(tp(set()))
+        check(tp([1, x, y]))
+        check(tp(obj for obj in [1, x, y]))
+        check(tp(set([1, x, y])))
+        check(tp(tuple([obj]) for obj in [1, x, y]))
+        check(tuple(tp([obj]) for obj in [1, x, y]))
+
+        self._tracked(tp([z]))
+        self._tracked(tp([[x, y]]))
+        self._tracked(tp([{x: y}]))
+        self._tracked(tp(obj for obj in [x, y, z]))
+        self._tracked(tp(tuple([obj]) for obj in [x, y, z]))
+        self._tracked(tuple(tp([obj]) for obj in [x, y, z]))
+
+    def test_track_dynamic(self):
+        # Test GC-optimization of dynamically constructed tuples.
+        self.check_track_dynamic(tuple, False)
+
+    def test_track_subtypes(self):
+        # Tuple subtypes must always be tracked
+        class MyTuple(tuple):
+            pass
+        self.check_track_dynamic(MyTuple, True)
+
+
 def test_main():
     test_support.run_unittest(TupleTest)
 
diff --git a/Misc/NEWS b/Misc/NEWS
index e96ec17..d734536 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -12,6 +12,11 @@ What's New in Python 2.7 alpha 1
 Core and Builtins
 -----------------
 
+- Issue #4688: Add a heuristic so that tuples and dicts containing only
+  untrackable objects are not tracked by the garbage collector. This can
+  reduce the size of collections and therefore the garbage collection overhead
+  on long-running programs, depending on their particular use of datatypes.
+
 - Issue #5512: Rewrite PyLong long division algorithm (x_divrem) to
   improve its performance.  Long divisions and remainder operations
   are now between 50% and 150% faster.
diff --git a/Modules/gcmodule.c b/Modules/gcmodule.c
index 4d71591..d9bea73 100644
--- a/Modules/gcmodule.c
+++ b/Modules/gcmodule.c
@@ -432,7 +432,13 @@ move_unreachable(PyGC_Head *young, PyGC_Head *unreachable)
                         (void) traverse(op,
                                         (visitproc)visit_reachable,
                                         (void *)young);
-                        next = gc->gc.gc_next;
+			next = gc->gc.gc_next;
+			if (PyTuple_CheckExact(op)) {
+				_PyTuple_MaybeUntrack(op);
+			}
+			else if (PyDict_CheckExact(op)) {
+				_PyDict_MaybeUntrack(op);
+			}
 		}
 		else {
 			/* This *may* be unreachable.  To make progress,
@@ -1264,6 +1270,26 @@ gc_get_objects(PyObject *self, PyObject *noargs)
 	return result;
 }
 
+PyDoc_STRVAR(gc_is_tracked__doc__,
+"is_tracked(obj) -> bool\n"
+"\n"
+"Returns true if the object is tracked by the garbage collector.\n"
+"Simple atomic objects will return false.\n"
+);
+
+static PyObject *
+gc_is_tracked(PyObject *self, PyObject *obj)
+{
+	PyObject *result;
+	
+	if (PyObject_IS_GC(obj) && IS_TRACKED(obj))
+		result = Py_True;
+	else
+		result = Py_False;
+	Py_INCREF(result);
+	return result;
+}
+
 
 PyDoc_STRVAR(gc__doc__,
 "This module provides access to the garbage collector for reference cycles.\n"
@@ -1278,6 +1304,7 @@ PyDoc_STRVAR(gc__doc__,
 "set_threshold() -- Set the collection thresholds.\n"
 "get_threshold() -- Return the current the collection thresholds.\n"
 "get_objects() -- Return a list of all objects tracked by the collector.\n"
+"is_tracked() -- Returns true if a given object is tracked.\n"
 "get_referrers() -- Return the list of objects that refer to an object.\n"
 "get_referents() -- Return the list of objects that an object refers to.\n");
 
@@ -1293,6 +1320,7 @@ static PyMethodDef GcMethods[] = {
 	{"collect",	   (PyCFunction)gc_collect,
          	METH_VARARGS | METH_KEYWORDS,           gc_collect__doc__},
 	{"get_objects",    gc_get_objects,METH_NOARGS,  gc_get_objects__doc__},
+	{"is_tracked",     gc_is_tracked, METH_O,       gc_is_tracked__doc__},
 	{"get_referrers",  gc_get_referrers, METH_VARARGS,
 		gc_get_referrers__doc__},
 	{"get_referents",  gc_get_referents, METH_VARARGS,
diff --git a/Objects/dictobject.c b/Objects/dictobject.c
index f4d8683..5069c76 100644
--- a/Objects/dictobject.c
+++ b/Objects/dictobject.c
@@ -180,6 +180,24 @@ show_alloc(void)
 }
 #endif
 
+/* Debug statistic to count GC tracking of dicts */
+#ifdef SHOW_TRACK_COUNT
+static Py_ssize_t count_untracked = 0;
+static Py_ssize_t count_tracked = 0;
+
+static void
+show_track(void)
+{
+	fprintf(stderr, "Dicts created: %" PY_FORMAT_SIZE_T "d\n",
+		count_tracked + count_untracked);
+	fprintf(stderr, "Dicts tracked by the GC: %" PY_FORMAT_SIZE_T
+		"d\n", count_tracked);
+	fprintf(stderr, "%.2f%% dict tracking rate\n\n",
+		(100.0*count_tracked/(count_untracked+count_tracked)));
+}
+#endif
+
+
 /* Initialization macros.
    There are two ways to create a dict:  PyDict_New() is the main C API
    function, and the tp_new slot maps to dict_new().  In the latter case we
@@ -233,6 +251,9 @@ PyDict_New(void)
 #ifdef SHOW_ALLOC_COUNT
 		Py_AtExit(show_alloc);
 #endif
+#ifdef SHOW_TRACK_COUNT
+		Py_AtExit(show_track);
+#endif
 	}
 	if (numfree) {
 		mp = free_list[--numfree];
@@ -262,10 +283,12 @@ PyDict_New(void)
 #endif
 	}
 	mp->ma_lookup = lookdict_string;
+#ifdef SHOW_TRACK_COUNT
+	count_untracked++;
+#endif
 #ifdef SHOW_CONVERSION_COUNTS
 	++created;
 #endif
-	_PyObject_GC_TRACK(mp);
 	return (PyObject *)mp;
 }
 
@@ -433,6 +456,52 @@ lookdict_string(PyDictObject *mp, PyObject *key, register long hash)
 	return 0;
 }
 
+#ifdef SHOW_TRACK_COUNT
+#define INCREASE_TRACK_COUNT \
+	(count_tracked++, count_untracked--);
+#define DECREASE_TRACK_COUNT \
+	(count_tracked--, count_untracked++);
+#else
+#define INCREASE_TRACK_COUNT
+#define DECREASE_TRACK_COUNT
+#endif
+
+#define MAINTAIN_TRACKING(mp, key, value) \
+	do { \
+		if (!_PyObject_GC_IS_TRACKED(mp)) { \
+			if (_PyObject_GC_MAY_BE_TRACKED(key) || \
+				_PyObject_GC_MAY_BE_TRACKED(value)) { \
+				_PyObject_GC_TRACK(mp); \
+				INCREASE_TRACK_COUNT \
+			} \
+		} \
+	} while(0)
+
+void
+_PyDict_MaybeUntrack(PyObject *op)
+{
+	PyDictObject *mp;
+	PyObject *value;
+	Py_ssize_t mask, i;
+	PyDictEntry *ep;
+
+	if (!PyDict_CheckExact(op) || !_PyObject_GC_IS_TRACKED(op))
+		return;
+	
+	mp = (PyDictObject *) op;
+	ep = mp->ma_table;
+	mask = mp->ma_mask;
+	for (i = 0; i <= mask; i++) {
+		if ((value = ep[i].me_value) == NULL)
+			continue;
+		if (_PyObject_GC_MAY_BE_TRACKED(value) ||
+			_PyObject_GC_MAY_BE_TRACKED(ep[i].me_key))
+			return;
+	}
+	_PyObject_GC_UNTRACK(op);
+}
+
+
 /*
 Internal routine to insert a new item into the table.
 Used both by the internal resize routine and by the public insert routine.
@@ -453,6 +522,7 @@ insertdict(register PyDictObject *mp, PyObject *key, long hash, PyObject *value)
 		Py_DECREF(value);
 		return -1;
 	}
+	MAINTAIN_TRACKING(mp, key, value);
 	if (ep->me_value != NULL) {
 		old_value = ep->me_value;
 		ep->me_value = value;
@@ -492,6 +562,7 @@ insertdict_clean(register PyDictObject *mp, PyObject *key, long hash,
 	PyDictEntry *ep0 = mp->ma_table;
 	register PyDictEntry *ep;
 
+	MAINTAIN_TRACKING(mp, key, value);
 	i = hash & mask;
 	ep = &ep0[i];
 	for (perturb = hash; ep->me_key != NULL; perturb >>= PERTURB_SHIFT) {
@@ -2202,9 +2273,18 @@ dict_new(PyTypeObject *type, PyObject *args, PyObject *kwds)
 		assert(d->ma_table == NULL && d->ma_fill == 0 && d->ma_used == 0);
 		INIT_NONZERO_DICT_SLOTS(d);
 		d->ma_lookup = lookdict_string;
+		/* The object has been implicitely tracked by tp_alloc */
+		if (type == &PyDict_Type)
+			_PyObject_GC_UNTRACK(d);
 #ifdef SHOW_CONVERSION_COUNTS
 		++created;
 #endif
+#ifdef SHOW_TRACK_COUNT
+		if (_PyObject_GC_IS_TRACKED(d))
+			count_tracked++;
+		else
+			count_untracked++;
+#endif
 	}
 	return self;
 }
diff --git a/Objects/tupleobject.c b/Objects/tupleobject.c
index 74d392a..644d8a9 100644
--- a/Objects/tupleobject.c
+++ b/Objects/tupleobject.c
@@ -23,11 +23,36 @@ Py_ssize_t fast_tuple_allocs;
 Py_ssize_t tuple_zero_allocs;
 #endif
 
+/* Debug statistic to count GC tracking of tuples.
+   Please note that tuples are only untracked when considered by the GC, and
+   many of them will be dead before. Therefore, a tracking rate close to 100%
+   does not necessarily prove that the heuristic is inefficient.
+*/
+#ifdef SHOW_TRACK_COUNT
+static Py_ssize_t count_untracked = 0;
+static Py_ssize_t count_tracked = 0;
+
+static void
+show_track(void)
+{
+	fprintf(stderr, "Tuples created: %" PY_FORMAT_SIZE_T "d\n",
+		count_tracked + count_untracked);
+	fprintf(stderr, "Tuples tracked by the GC: %" PY_FORMAT_SIZE_T
+		"d\n", count_tracked);
+	fprintf(stderr, "%.2f%% tuple tracking rate\n\n",
+		(100.0*count_tracked/(count_untracked+count_tracked)));
+}
+#endif
+
+
 PyObject *
 PyTuple_New(register Py_ssize_t size)
 {
 	register PyTupleObject *op;
 	Py_ssize_t i;
+#ifdef SHOW_TRACK_COUNT
+	count_tracked++;
+#endif
 	if (size < 0) {
 		PyErr_BadInternalCall();
 		return NULL;
@@ -131,6 +156,32 @@ PyTuple_SetItem(register PyObject *op, register Py_ssize_t i, PyObject *newitem)
 	return 0;
 }
 
+void
+_PyTuple_MaybeUntrack(PyObject *op)
+{
+	PyTupleObject *t;
+	Py_ssize_t i, n;
+	
+	if (!PyTuple_CheckExact(op) || !_PyObject_GC_IS_TRACKED(op))
+		return;
+	t = (PyTupleObject *) op;
+	n = Py_SIZE(t);
+	for (i = 0; i < n; i++) {
+		PyObject *elt = PyTuple_GET_ITEM(t, i);
+		/* Tuple with NULL elements aren't
+		   fully constructed, don't untrack
+		   them yet. */
+		if (!elt ||
+			_PyObject_GC_MAY_BE_TRACKED(elt))
+			return;
+	}
+#ifdef SHOW_TRACK_COUNT
+	count_tracked--;
+	count_untracked++;
+#endif
+	_PyObject_GC_UNTRACK(op);
+}
+
 PyObject *
 PyTuple_Pack(Py_ssize_t n, ...)
 {
@@ -880,6 +931,9 @@ PyTuple_Fini(void)
 
 	(void)PyTuple_ClearFreeList();
 #endif
+#ifdef SHOW_TRACK_COUNT
+	show_track();
+#endif
 }
 
 /*********************** Tuple Iterator **************************/