*EXPERIMENTAL* speedup of slot_sq_item. This sped up the following
test dramatically:
class T(tuple): __dynamic__ = 1
t = T(range(1000))
for i in range(1000): tt = tuple(t)
The speedup was about 5x compared to the previous state of CVS (1.7
vs. 8.8, in arbitrary time units). But it's still more than twice as
slow as as the same test with __dynamic__ = 0 (0.8).
I'm not sure that I really want to go through the trouble of this kind
of speedup for every slot. Even doing it just for the most popular
slots will be a major effort (the new slot_sq_item is 40+ lines, while
the old one was one line with a powerful macro -- unfortunately the
speedup comes from expanding the macro and doing things in a way
specific to the slot signature).
An alternative that I'm currently considering is sketched in PLAN.txt:
trap setattr on type objects. But this will require keeping track of
all derived types using weak references.
diff --git a/Objects/typeobject.c b/Objects/typeobject.c
index a681d33..d976945 100644
--- a/Objects/typeobject.c
+++ b/Objects/typeobject.c
@@ -2113,12 +2113,16 @@
PyObject *arg;
int i;
- if (!PyArg_ParseTuple(args, "O", &arg))
- return NULL;
- i = getindex(self, arg);
- if (i == -1 && PyErr_Occurred())
- return NULL;
- return (*func)(self, i);
+ if (PyTuple_GET_SIZE(args) == 1) {
+ arg = PyTuple_GET_ITEM(args, 0);
+ i = getindex(self, arg);
+ if (i == -1 && PyErr_Occurred())
+ return NULL;
+ return (*func)(self, i);
+ }
+ PyArg_ParseTuple(args, "O", &arg);
+ assert(PyErr_Occurred());
+ return NULL;
}
static struct wrapperbase tab_getitem_int[] = {
@@ -2825,7 +2829,57 @@
SLOT1(slot_sq_concat, "__add__", PyObject *, "O")
SLOT1(slot_sq_repeat, "__mul__", int, "i")
-SLOT1(slot_sq_item, "__getitem__", int, "i")
+
+/* Super-optimized version of slot_sq_item.
+ Other slots could do the same... */
+static PyObject *
+slot_sq_item(PyObject *self, int i)
+{
+ static PyObject *getitem_str;
+ PyObject *func, *args = NULL, *ival = NULL, *retval = NULL;
+ descrgetfunc f;
+
+ if (getitem_str == NULL) {
+ getitem_str = PyString_InternFromString("__getitem__");
+ if (getitem_str == NULL)
+ return NULL;
+ }
+ func = _PyType_Lookup(self->ob_type, getitem_str);
+ if (func != NULL) {
+ if (func->ob_type == &PyWrapperDescr_Type) {
+ PyWrapperDescrObject *wrapper =
+ (PyWrapperDescrObject *)func;
+ if (wrapper->d_base->wrapper == wrap_sq_item) {
+ intargfunc f;
+ f = (intargfunc)(wrapper->d_wrapped);
+ return f(self, i);
+ }
+ }
+ if ((f = func->ob_type->tp_descr_get) == NULL)
+ Py_INCREF(func);
+ else
+ func = f(func, self, (PyObject *)(self->ob_type));
+ ival = PyInt_FromLong(i);
+ if (ival != NULL) {
+ args = PyTuple_New(1);
+ if (args != NULL) {
+ PyTuple_SET_ITEM(args, 0, ival);
+ retval = PyObject_Call(func, args, NULL);
+ Py_XDECREF(args);
+ Py_XDECREF(func);
+ return retval;
+ }
+ }
+ }
+ else {
+ PyErr_SetObject(PyExc_AttributeError, getitem_str);
+ }
+ Py_XDECREF(args);
+ Py_XDECREF(ival);
+ Py_XDECREF(func);
+ return NULL;
+}
+
SLOT2(slot_sq_slice, "__getslice__", int, int, "ii")
static int