Objects/object.c

/* Generic object operations; and implementation of None (NoObject) */

#include "Python.h"

#ifdef __cplusplus
extern "C" {
#endif

#ifdef Py_REF_DEBUG
Py_ssize_t _Py_RefTotal;

Py_ssize_t
_Py_GetRefTotal(void)
{
	PyObject *o;
	Py_ssize_t total = _Py_RefTotal;
        /* ignore the references to the dummy object of the dicts and sets
           because they are not reliable and not useful (now that the
           hash table code is well-tested) */
	o = _PyDict_Dummy();
	if (o != NULL)
		total -= o->ob_refcnt;
	o = _PySet_Dummy();
	if (o != NULL)
		total -= o->ob_refcnt;
	return total;
}
#endif /* Py_REF_DEBUG */

int Py_DivisionWarningFlag;

/* Object allocation routines used by NEWOBJ and NEWVAROBJ macros.
   These are used by the individual routines for object creation.
   Do not call them otherwise, they do not initialize the object! */

#ifdef Py_TRACE_REFS
/* Head of circular doubly-linked list of all objects.  These are linked
 * together via the _ob_prev and _ob_next members of a PyObject, which
 * exist only in a Py_TRACE_REFS build.
 */
static PyObject refchain = {&refchain, &refchain};

/* Insert op at the front of the list of all objects.  If force is true,
 * op is added even if _ob_prev and _ob_next are non-NULL already.  If
 * force is false amd _ob_prev or _ob_next are non-NULL, do nothing.
 * force should be true if and only if op points to freshly allocated,
 * uninitialized memory, or you've unlinked op from the list and are
 * relinking it into the front.
 * Note that objects are normally added to the list via _Py_NewReference,
 * which is called by PyObject_Init.  Not all objects are initialized that
 * way, though; exceptions include statically allocated type objects, and
 * statically allocated singletons (like Py_True and Py_None).
 */
void
_Py_AddToAllObjects(PyObject *op, int force)
{
#ifdef  Py_DEBUG
	if (!force) {
		/* If it's initialized memory, op must be in or out of
		 * the list unambiguously.
		 */
		assert((op->_ob_prev == NULL) == (op->_ob_next == NULL));
	}
#endif
	if (force || op->_ob_prev == NULL) {
		op->_ob_next = refchain._ob_next;
		op->_ob_prev = &refchain;
		refchain._ob_next->_ob_prev = op;
		refchain._ob_next = op;
	}
}
#endif	/* Py_TRACE_REFS */

#ifdef COUNT_ALLOCS
static PyTypeObject *type_list;
/* All types are added to type_list, at least when
   they get one object created. That makes them
   immortal, which unfortunately contributes to
   garbage itself. If unlist_types_without_objects
   is set, they will be removed from the type_list
   once the last object is deallocated. */
int unlist_types_without_objects;
extern int tuple_zero_allocs, fast_tuple_allocs;
extern int quick_int_allocs, quick_neg_int_allocs;
extern int null_strings, one_strings;
void
dump_counts(FILE* f)
{
	PyTypeObject *tp;

	for (tp = type_list; tp; tp = tp->tp_next)
		fprintf(f, "%s alloc'd: %d, freed: %d, max in use: %d\n",
			tp->tp_name, tp->tp_allocs, tp->tp_frees,
			tp->tp_maxalloc);
	fprintf(f, "fast tuple allocs: %d, empty: %d\n",
		fast_tuple_allocs, tuple_zero_allocs);
	fprintf(f, "fast int allocs: pos: %d, neg: %d\n",
		quick_int_allocs, quick_neg_int_allocs);
	fprintf(f, "null strings: %d, 1-strings: %d\n",
		null_strings, one_strings);
}

PyObject *
get_counts(void)
{
	PyTypeObject *tp;
	PyObject *result;
	PyObject *v;

	result = PyList_New(0);
	if (result == NULL)
		return NULL;
	for (tp = type_list; tp; tp = tp->tp_next) {
		v = Py_BuildValue("(snnn)", tp->tp_name, tp->tp_allocs,
				  tp->tp_frees, tp->tp_maxalloc);
		if (v == NULL) {
			Py_DECREF(result);
			return NULL;
		}
		if (PyList_Append(result, v) < 0) {
			Py_DECREF(v);
			Py_DECREF(result);
			return NULL;
		}
		Py_DECREF(v);
	}
	return result;
}

void
inc_count(PyTypeObject *tp)
{
	if (tp->tp_next == NULL && tp->tp_prev == NULL) {
		/* first time; insert in linked list */
		if (tp->tp_next != NULL) /* sanity check */
			Py_FatalError("XXX inc_count sanity check");
		if (type_list)
			type_list->tp_prev = tp;
		tp->tp_next = type_list;
		/* Note that as of Python 2.2, heap-allocated type objects
		 * can go away, but this code requires that they stay alive
		 * until program exit.  That's why we're careful with
		 * refcounts here.  type_list gets a new reference to tp,
		 * while ownership of the reference type_list used to hold
		 * (if any) was transferred to tp->tp_next in the line above.
		 * tp is thus effectively immortal after this.
		 */
		Py_INCREF(tp);
		type_list = tp;
#ifdef Py_TRACE_REFS
		/* Also insert in the doubly-linked list of all objects,
		 * if not already there.
		 */
		_Py_AddToAllObjects((PyObject *)tp, 0);
#endif
	}
	tp->tp_allocs++;
	if (tp->tp_allocs - tp->tp_frees > tp->tp_maxalloc)
		tp->tp_maxalloc = tp->tp_allocs - tp->tp_frees;
}

void dec_count(PyTypeObject *tp)
{
	tp->tp_frees++;
	if (unlist_types_without_objects &&
	    tp->tp_allocs == tp->tp_frees) {
		/* unlink the type from type_list */
		if (tp->tp_prev)
			tp->tp_prev->tp_next = tp->tp_next;
		else
			type_list = tp->tp_next;
		if (tp->tp_next)
			tp->tp_next->tp_prev = tp->tp_prev;
		tp->tp_next = tp->tp_prev = NULL;
		Py_DECREF(tp);
	}
}

#endif

#ifdef Py_REF_DEBUG
/* Log a fatal error; doesn't return. */
void
_Py_NegativeRefcount(const char *fname, int lineno, PyObject *op)
{
	char buf[300];

	PyOS_snprintf(buf, sizeof(buf),
		      "%s:%i object at %p has negative ref count "
		      "%" PY_FORMAT_SIZE_T "d",
		      fname, lineno, op, op->ob_refcnt);
	Py_FatalError(buf);
}

#endif /* Py_REF_DEBUG */

void
Py_IncRef(PyObject *o)
{
    Py_XINCREF(o);
}

void
Py_DecRef(PyObject *o)
{
    Py_XDECREF(o);
}

PyObject *
PyObject_Init(PyObject *op, PyTypeObject *tp)
{
	if (op == NULL)
		return PyErr_NoMemory();
	/* Any changes should be reflected in PyObject_INIT (objimpl.h) */
	op->ob_type = tp;
	_Py_NewReference(op);
	return op;
}

PyVarObject *
PyObject_InitVar(PyVarObject *op, PyTypeObject *tp, Py_ssize_t size)
{
	if (op == NULL)
		return (PyVarObject *) PyErr_NoMemory();
	/* Any changes should be reflected in PyObject_INIT_VAR */
	op->ob_size = size;
	op->ob_type = tp;
	_Py_NewReference((PyObject *)op);
	return op;
}

PyObject *
_PyObject_New(PyTypeObject *tp)
{
	PyObject *op;
	op = (PyObject *) PyObject_MALLOC(_PyObject_SIZE(tp));
	if (op == NULL)
		return PyErr_NoMemory();
	return PyObject_INIT(op, tp);
}

PyVarObject *
_PyObject_NewVar(PyTypeObject *tp, Py_ssize_t nitems)
{
	PyVarObject *op;
	const size_t size = _PyObject_VAR_SIZE(tp, nitems);
	op = (PyVarObject *) PyObject_MALLOC(size);
	if (op == NULL)
		return (PyVarObject *)PyErr_NoMemory();
	return PyObject_INIT_VAR(op, tp, nitems);
}

/* for binary compatibility with 2.2 */
#undef _PyObject_Del
void
_PyObject_Del(PyObject *op)
{
	PyObject_FREE(op);
}

/* Implementation of PyObject_Print with recursion checking */
static int
internal_print(PyObject *op, FILE *fp, int flags, int nesting)
{
	int ret = 0;
	if (nesting > 10) {
		PyErr_SetString(PyExc_RuntimeError, "print recursion");
		return -1;
	}
	if (PyErr_CheckSignals())
		return -1;
#ifdef USE_STACKCHECK
	if (PyOS_CheckStack()) {
		PyErr_SetString(PyExc_MemoryError, "stack overflow");
		return -1;
	}
#endif
	clearerr(fp); /* Clear any previous error condition */
	if (op == NULL) {
		fprintf(fp, "<nil>");
	}
	else {
		if (op->ob_refcnt <= 0)
			/* XXX(twouters) cast refcount to long until %zd is
			   universally available */
			fprintf(fp, "<refcnt %ld at %p>",
				(long)op->ob_refcnt, op);
		else if (op->ob_type->tp_print == NULL) {
			PyObject *s;
			if (flags & Py_PRINT_RAW)
				s = PyObject_Str(op);
			else
				s = PyObject_Repr(op);
			if (s == NULL)
				ret = -1;
			else {
				ret = internal_print(s, fp, Py_PRINT_RAW,
						     nesting+1);
			}
			Py_XDECREF(s);
		}
		else
			ret = (*op->ob_type->tp_print)(op, fp, flags);
	}
	if (ret == 0) {
		if (ferror(fp)) {
			PyErr_SetFromErrno(PyExc_IOError);
			clearerr(fp);
			ret = -1;
		}
	}
	return ret;
}

int
PyObject_Print(PyObject *op, FILE *fp, int flags)
{
	return internal_print(op, fp, flags, 0);
}


/* For debugging convenience.  See Misc/gdbinit for some useful gdb hooks */
void _PyObject_Dump(PyObject* op)
{
	if (op == NULL)
		fprintf(stderr, "NULL\n");
	else {
		fprintf(stderr, "object  : ");
		(void)PyObject_Print(op, stderr, 0);
		/* XXX(twouters) cast refcount to long until %zd is
		   universally available */
		fprintf(stderr, "\n"
			"type    : %s\n"
			"refcount: %ld\n"
			"address : %p\n",
			op->ob_type==NULL ? "NULL" : op->ob_type->tp_name,
			(long)op->ob_refcnt,
			op);
	}
}

PyObject *
PyObject_Repr(PyObject *v)
{
	if (PyErr_CheckSignals())
		return NULL;
#ifdef USE_STACKCHECK
	if (PyOS_CheckStack()) {
		PyErr_SetString(PyExc_MemoryError, "stack overflow");
		return NULL;
	}
#endif
	if (v == NULL)
		return PyString_FromString("<NULL>");
	else if (v->ob_type->tp_repr == NULL)
		return PyString_FromFormat("<%s object at %p>",
					   v->ob_type->tp_name, v);
	else {
		PyObject *res;
		res = (*v->ob_type->tp_repr)(v);
		if (res == NULL)
			return NULL;
#ifdef Py_USING_UNICODE
		if (PyUnicode_Check(res)) {
			PyObject* str;
			str = PyUnicode_AsEncodedString(res, NULL, NULL);
			Py_DECREF(res);
			if (str)
				res = str;
			else
				return NULL;
		}
#endif
		if (!PyString_Check(res)) {
			PyErr_Format(PyExc_TypeError,
				     "__repr__ returned non-string (type %.200s)",
				     res->ob_type->tp_name);
			Py_DECREF(res);
			return NULL;
		}
		return res;
	}
}

PyObject *
_PyObject_Str(PyObject *v)
{
	PyObject *res;
	int type_ok;
	if (v == NULL)
		return PyString_FromString("<NULL>");
	if (PyString_CheckExact(v)) {
		Py_INCREF(v);
		return v;
	}
#ifdef Py_USING_UNICODE
	if (PyUnicode_CheckExact(v)) {
		Py_INCREF(v);
		return v;
	}
#endif
	if (v->ob_type->tp_str == NULL)
		return PyObject_Repr(v);

	res = (*v->ob_type->tp_str)(v);
	if (res == NULL)
		return NULL;
	type_ok = PyString_Check(res);
#ifdef Py_USING_UNICODE
	type_ok = type_ok || PyUnicode_Check(res);
#endif
	if (!type_ok) {
		PyErr_Format(PyExc_TypeError,
			     "__str__ returned non-string (type %.200s)",
			     res->ob_type->tp_name);
		Py_DECREF(res);
		return NULL;
	}
	return res;
}

PyObject *
PyObject_Str(PyObject *v)
{
	PyObject *res = _PyObject_Str(v);
	if (res == NULL)
		return NULL;
#ifdef Py_USING_UNICODE
	if (PyUnicode_Check(res)) {
		PyObject* str;
		str = PyUnicode_AsEncodedString(res, NULL, NULL);
		Py_DECREF(res);
		if (str)
			res = str;
		else
		    	return NULL;
	}
#endif
	assert(PyString_Check(res));
	return res;
}

#ifdef Py_USING_UNICODE
PyObject *
PyObject_Unicode(PyObject *v)
{
	PyObject *res;
	PyObject *func;
	PyObject *str;
	static PyObject *unicodestr;

	if (v == NULL) {
		res = PyString_FromString("<NULL>");
		if (res == NULL)
			return NULL;
		str = PyUnicode_FromEncodedObject(res, NULL, "strict");
		Py_DECREF(res);
		return str;
	} else if (PyUnicode_CheckExact(v)) {
		Py_INCREF(v);
		return v;
	}
	/* XXX As soon as we have a tp_unicode slot, we should
	   check this before trying the __unicode__
	   method. */
	if (unicodestr == NULL) {
		unicodestr= PyString_InternFromString("__unicode__");
		if (unicodestr == NULL)
			return NULL;
	}
	func = PyObject_GetAttr(v, unicodestr);
	if (func != NULL) {
		res = PyEval_CallObject(func, (PyObject *)NULL);
		Py_DECREF(func);
	}
	else {
		PyErr_Clear();
		if (PyUnicode_Check(v)) {
			/* For a Unicode subtype that's didn't overwrite __unicode__,
			   return a true Unicode object with the same data. */
			return PyUnicode_FromUnicode(PyUnicode_AS_UNICODE(v),
			                             PyUnicode_GET_SIZE(v));
		}
		if (PyString_CheckExact(v)) {
			Py_INCREF(v);
			res = v;
		}
		else {
			if (v->ob_type->tp_str != NULL)
				res = (*v->ob_type->tp_str)(v);
			else
				res = PyObject_Repr(v);
		}
	}
	if (res == NULL)
		return NULL;
	if (!PyUnicode_Check(res)) {
		str = PyUnicode_FromEncodedObject(res, NULL, "strict");
		Py_DECREF(res);
		res = str;
	}
	return res;
}
#endif


/* Helper to warn about deprecated tp_compare return values.  Return:
   -2 for an exception;
   -1 if v <  w;
    0 if v == w;
    1 if v  > w.
   (This function cannot return 2.)
*/
static int
adjust_tp_compare(int c)
{
	if (PyErr_Occurred()) {
		if (c != -1 && c != -2) {
			PyObject *t, *v, *tb;
			PyErr_Fetch(&t, &v, &tb);
			if (PyErr_Warn(PyExc_RuntimeWarning,
				       "tp_compare didn't return -1 or -2 "
				       "for exception") < 0) {
				Py_XDECREF(t);
				Py_XDECREF(v);
				Py_XDECREF(tb);
			}
			else
				PyErr_Restore(t, v, tb);
		}
		return -2;
	}
	else if (c < -1 || c > 1) {
		if (PyErr_Warn(PyExc_RuntimeWarning,
			       "tp_compare didn't return -1, 0 or 1") < 0)
			return -2;
		else
			return c < -1 ? -1 : 1;
	}
	else {
		assert(c >= -1 && c <= 1);
		return c;
	}
}


/* Macro to get the tp_richcompare field of a type if defined */
#define RICHCOMPARE(t) ((t)->tp_richcompare)

/* Map rich comparison operators to their swapped version, e.g. LT --> GT */
int _Py_SwappedOp[] = {Py_GT, Py_GE, Py_EQ, Py_NE, Py_LT, Py_LE};

/* Try a genuine rich comparison, returning an object.  Return:
   NULL for exception;
   NotImplemented if this particular rich comparison is not implemented or
     undefined;
   some object not equal to NotImplemented if it is implemented
     (this latter object may not be a Boolean).
*/
static PyObject *
try_rich_compare(PyObject *v, PyObject *w, int op)
{
	richcmpfunc f;
	PyObject *res;

	if (v->ob_type != w->ob_type &&
	    PyType_IsSubtype(w->ob_type, v->ob_type) &&
	    (f = RICHCOMPARE(w->ob_type)) != NULL) {
		res = (*f)(w, v, _Py_SwappedOp[op]);
		if (res != Py_NotImplemented)
			return res;
		Py_DECREF(res);
	}
	if ((f = RICHCOMPARE(v->ob_type)) != NULL) {
		res = (*f)(v, w, op);
		if (res != Py_NotImplemented)
			return res;
		Py_DECREF(res);
	}
	if ((f = RICHCOMPARE(w->ob_type)) != NULL) {
		return (*f)(w, v, _Py_SwappedOp[op]);
	}
	res = Py_NotImplemented;
	Py_INCREF(res);
	return res;
}

/* Try a genuine rich comparison, returning an int.  Return:
   -1 for exception (including the case where try_rich_compare() returns an
      object that's not a Boolean);
    0 if the outcome is false;
    1 if the outcome is true;
    2 if this particular rich comparison is not implemented or undefined.
*/
static int
try_rich_compare_bool(PyObject *v, PyObject *w, int op)
{
	PyObject *res;
	int ok;

	if (RICHCOMPARE(v->ob_type) == NULL && RICHCOMPARE(w->ob_type) == NULL)
		return 2; /* Shortcut, avoid INCREF+DECREF */
	res = try_rich_compare(v, w, op);
	if (res == NULL)
		return -1;
	if (res == Py_NotImplemented) {
		Py_DECREF(res);
		return 2;
	}
	ok = PyObject_IsTrue(res);
	Py_DECREF(res);
	return ok;
}

/* Try rich comparisons to determine a 3-way comparison.  Return:
   -2 for an exception;
   -1 if v  < w;
    0 if v == w;
    1 if v  > w;
    2 if this particular rich comparison is not implemented or undefined.
*/
static int
try_rich_to_3way_compare(PyObject *v, PyObject *w)
{
	static struct { int op; int outcome; } tries[3] = {
		/* Try this operator, and if it is true, use this outcome: */
		{Py_EQ, 0},
		{Py_LT, -1},
		{Py_GT, 1},
	};
	int i;

	if (RICHCOMPARE(v->ob_type) == NULL && RICHCOMPARE(w->ob_type) == NULL)
		return 2; /* Shortcut */

	for (i = 0; i < 3; i++) {
		switch (try_rich_compare_bool(v, w, tries[i].op)) {
		case -1:
			return -2;
		case 1:
			return tries[i].outcome;
		}
	}

	return 2;
}

/* Try a 3-way comparison, returning an int.  Return:
   -2 for an exception;
   -1 if v <  w;
    0 if v == w;
    1 if v  > w;
    2 if this particular 3-way comparison is not implemented or undefined.
*/
static int
try_3way_compare(PyObject *v, PyObject *w)
{
	int c;
	cmpfunc f;

	/* Comparisons involving instances are given to instance_compare,
	   which has the same return conventions as this function. */

	f = v->ob_type->tp_compare;
	if (PyInstance_Check(v))
		return (*f)(v, w);
	if (PyInstance_Check(w))
		return (*w->ob_type->tp_compare)(v, w);

	/* If both have the same (non-NULL) tp_compare, use it. */
	if (f != NULL && f == w->ob_type->tp_compare) {
		c = (*f)(v, w);
		return adjust_tp_compare(c);
	}

	/* If either tp_compare is _PyObject_SlotCompare, that's safe. */
	if (f == _PyObject_SlotCompare ||
	    w->ob_type->tp_compare == _PyObject_SlotCompare)
		return _PyObject_SlotCompare(v, w);

	/* If we're here, v and w,
	    a) are not instances;
	    b) have different types or a type without tp_compare; and
	    c) don't have a user-defined tp_compare.
	   tp_compare implementations in C assume that both arguments
	   have their type, so we give up if the coercion fails or if
	   it yields types which are still incompatible (which can
	   happen with a user-defined nb_coerce).
	*/
	c = PyNumber_CoerceEx(&v, &w);
	if (c < 0)
		return -2;
	if (c > 0)
		return 2;
	f = v->ob_type->tp_compare;
	if (f != NULL && f == w->ob_type->tp_compare) {
		c = (*f)(v, w);
		Py_DECREF(v);
		Py_DECREF(w);
		return adjust_tp_compare(c);
	}

	/* No comparison defined */
	Py_DECREF(v);
	Py_DECREF(w);
	return 2;
}

/* Final fallback 3-way comparison, returning an int.  Return:
   -2 if an error occurred;
   -1 if v <  w;
    0 if v == w;
    1 if v >  w.
*/
static int
default_3way_compare(PyObject *v, PyObject *w)
{
	int c;
	const char *vname, *wname;

	if (v->ob_type == w->ob_type) {
		/* When comparing these pointers, they must be cast to
		 * integer types (i.e. Py_uintptr_t, our spelling of C9X's
		 * uintptr_t).  ANSI specifies that pointer compares other
		 * than == and != to non-related structures are undefined.
		 */
		Py_uintptr_t vv = (Py_uintptr_t)v;
		Py_uintptr_t ww = (Py_uintptr_t)w;
		return (vv < ww) ? -1 : (vv > ww) ? 1 : 0;
	}

#ifdef Py_USING_UNICODE
	/* Special case for Unicode */
	if (PyUnicode_Check(v) || PyUnicode_Check(w)) {
		c = PyUnicode_Compare(v, w);
		if (!PyErr_Occurred())
			return c;
		/* TypeErrors are ignored: if Unicode coercion fails due
		   to one of the arguments not having the right type, we
		   continue as defined by the coercion protocol (see
		   above).  Luckily, decoding errors are reported as
		   ValueErrors and are not masked by this technique. */
		if (!PyErr_ExceptionMatches(PyExc_TypeError))
			return -2;
		PyErr_Clear();
	}
#endif

	/* None is smaller than anything */
	if (v == Py_None)
		return -1;
	if (w == Py_None)
		return 1;

	/* different type: compare type names; numbers are smaller */
	if (PyNumber_Check(v))
		vname = "";
	else
		vname = v->ob_type->tp_name;
	if (PyNumber_Check(w))
		wname = "";
	else
		wname = w->ob_type->tp_name;
	c = strcmp(vname, wname);
	if (c < 0)
		return -1;
	if (c > 0)
		return 1;
	/* Same type name, or (more likely) incomparable numeric types */
	return ((Py_uintptr_t)(v->ob_type) < (
		Py_uintptr_t)(w->ob_type)) ? -1 : 1;
}

/* Do a 3-way comparison, by hook or by crook.  Return:
   -2 for an exception (but see below);
   -1 if v <  w;
    0 if v == w;
    1 if v >  w;
   BUT: if the object implements a tp_compare function, it returns
   whatever this function returns (whether with an exception or not).
*/
static int
do_cmp(PyObject *v, PyObject *w)
{
	int c;
	cmpfunc f;

	if (v->ob_type == w->ob_type
	    && (f = v->ob_type->tp_compare) != NULL) {
		c = (*f)(v, w);
		if (PyInstance_Check(v)) {
			/* Instance tp_compare has a different signature.
			   But if it returns undefined we fall through. */
			if (c != 2)
				return c;
			/* Else fall through to try_rich_to_3way_compare() */
		}
		else
			return adjust_tp_compare(c);
	}
	/* We only get here if one of the following is true:
	   a) v and w have different types
	   b) v and w have the same type, which doesn't have tp_compare
	   c) v and w are instances, and either __cmp__ is not defined or
	      __cmp__ returns NotImplemented
	*/
	c = try_rich_to_3way_compare(v, w);
	if (c < 2)
		return c;
	c = try_3way_compare(v, w);
	if (c < 2)
		return c;
	return default_3way_compare(v, w);
}

/* Compare v to w.  Return
   -1 if v <  w or exception (PyErr_Occurred() true in latter case).
    0 if v == w.
    1 if v > w.
   XXX The docs (C API manual) say the return value is undefined in case
   XXX of error.
*/
int
PyObject_Compare(PyObject *v, PyObject *w)
{
	int result;

	if (v == NULL || w == NULL) {
		PyErr_BadInternalCall();
		return -1;
	}
	if (v == w)
		return 0;
	if (Py_EnterRecursiveCall(" in cmp"))
		return -1;
	result = do_cmp(v, w);
	Py_LeaveRecursiveCall();
	return result < 0 ? -1 : result;
}

/* Return (new reference to) Py_True or Py_False. */
static PyObject *
convert_3way_to_object(int op, int c)
{
	PyObject *result;
	switch (op) {
	case Py_LT: c = c <  0; break;
	case Py_LE: c = c <= 0; break;
	case Py_EQ: c = c == 0; break;
	case Py_NE: c = c != 0; break;
	case Py_GT: c = c >  0; break;
	case Py_GE: c = c >= 0; break;
	}
	result = c ? Py_True : Py_False;
	Py_INCREF(result);
	return result;
}

/* We want a rich comparison but don't have one.  Try a 3-way cmp instead.
   Return
   NULL      if error
   Py_True   if v op w
   Py_False  if not (v op w)
*/
static PyObject *
try_3way_to_rich_compare(PyObject *v, PyObject *w, int op)
{
	int c;

	c = try_3way_compare(v, w);
	if (c >= 2)
		c = default_3way_compare(v, w);
	if (c <= -2)
		return NULL;
	return convert_3way_to_object(op, c);
}

/* Do rich comparison on v and w.  Return
   NULL      if error
   Else a new reference to an object other than Py_NotImplemented, usually(?):
   Py_True   if v op w
   Py_False  if not (v op w)
*/
static PyObject *
do_richcmp(PyObject *v, PyObject *w, int op)
{
	PyObject *res;

	res = try_rich_compare(v, w, op);
	if (res != Py_NotImplemented)
		return res;
	Py_DECREF(res);

	return try_3way_to_rich_compare(v, w, op);
}

/* Return:
   NULL for exception;
   some object not equal to NotImplemented if it is implemented
     (this latter object may not be a Boolean).
*/
PyObject *
PyObject_RichCompare(PyObject *v, PyObject *w, int op)
{
	PyObject *res;

	assert(Py_LT <= op && op <= Py_GE);
	if (Py_EnterRecursiveCall(" in cmp"))
		return NULL;

	/* If the types are equal, and not old-style instances, try to
	   get out cheap (don't bother with coercions etc.). */
	if (v->ob_type == w->ob_type && !PyInstance_Check(v)) {
		cmpfunc fcmp;
		richcmpfunc frich = RICHCOMPARE(v->ob_type);
		/* If the type has richcmp, try it first.  try_rich_compare
		   tries it two-sided, which is not needed since we've a
		   single type only. */
		if (frich != NULL) {
			res = (*frich)(v, w, op);
			if (res != Py_NotImplemented)
				goto Done;
			Py_DECREF(res);
		}
		/* No richcmp, or this particular richmp not implemented.
		   Try 3-way cmp. */
		fcmp = v->ob_type->tp_compare;
		if (fcmp != NULL) {
			int c = (*fcmp)(v, w);
			c = adjust_tp_compare(c);
			if (c == -2) {
				res = NULL;
				goto Done;
			}
			res = convert_3way_to_object(op, c);
			goto Done;
		}
	}

	/* Fast path not taken, or couldn't deliver a useful result. */
	res = do_richcmp(v, w, op);
Done:
	Py_LeaveRecursiveCall();
	return res;
}

/* Return -1 if error; 1 if v op w; 0 if not (v op w). */
int
PyObject_RichCompareBool(PyObject *v, PyObject *w, int op)
{
	PyObject *res;
	int ok;

	/* Quick result when objects are the same.
	   Guarantees that identity implies equality. */
	if (v == w) {
		if (op == Py_EQ)
			return 1;
		else if (op == Py_NE)
			return 0;
	}

	res = PyObject_RichCompare(v, w, op);
	if (res == NULL)
		return -1;
	if (PyBool_Check(res))
		ok = (res == Py_True);
	else
		ok = PyObject_IsTrue(res);
	Py_DECREF(res);
	return ok;
}

/* Set of hash utility functions to help maintaining the invariant that
	if a==b then hash(a)==hash(b)

   All the utility functions (_Py_Hash*()) return "-1" to signify an error.
*/

long
_Py_HashDouble(double v)
{
	double intpart, fractpart;
	int expo;
	long hipart;
	long x;		/* the final hash value */
	/* This is designed so that Python numbers of different types
	 * that compare equal hash to the same value; otherwise comparisons
	 * of mapping keys will turn out weird.
	 */

	fractpart = modf(v, &intpart);
	if (fractpart == 0.0) {
		/* This must return the same hash as an equal int or long. */
		if (intpart > LONG_MAX || -intpart > LONG_MAX) {
			/* Convert to long and use its hash. */
			PyObject *plong;	/* converted to Python long */
			if (Py_IS_INFINITY(intpart))
				/* can't convert to long int -- arbitrary */
				v = v < 0 ? -271828.0 : 314159.0;
			plong = PyLong_FromDouble(v);
			if (plong == NULL)
				return -1;
			x = PyObject_Hash(plong);
			Py_DECREF(plong);
			return x;
		}
		/* Fits in a C long == a Python int, so is its own hash. */
		x = (long)intpart;
		if (x == -1)
			x = -2;
		return x;
	}
	/* The fractional part is non-zero, so we don't have to worry about
	 * making this match the hash of some other type.
	 * Use frexp to get at the bits in the double.
	 * Since the VAX D double format has 56 mantissa bits, which is the
	 * most of any double format in use, each of these parts may have as
	 * many as (but no more than) 56 significant bits.
	 * So, assuming sizeof(long) >= 4, each part can be broken into two
	 * longs; frexp and multiplication are used to do that.
	 * Also, since the Cray double format has 15 exponent bits, which is
	 * the most of any double format in use, shifting the exponent field
	 * left by 15 won't overflow a long (again assuming sizeof(long) >= 4).
	 */
	v = frexp(v, &expo);
	v *= 2147483648.0;	/* 2**31 */
	hipart = (long)v;	/* take the top 32 bits */
	v = (v - (double)hipart) * 2147483648.0; /* get the next 32 bits */
	x = hipart + (long)v + (expo << 15);
	if (x == -1)
		x = -2;
	return x;
}

long
_Py_HashPointer(void *p)
{
#if SIZEOF_LONG >= SIZEOF_VOID_P
	return (long)p;
#else
	/* convert to a Python long and hash that */
	PyObject* longobj;
	long x;

	if ((longobj = PyLong_FromVoidPtr(p)) == NULL) {
		x = -1;
		goto finally;
	}
	x = PyObject_Hash(longobj);

finally:
	Py_XDECREF(longobj);
	return x;
#endif
}


long
PyObject_Hash(PyObject *v)
{
	PyTypeObject *tp = v->ob_type;
	if (tp->tp_hash != NULL)
		return (*tp->tp_hash)(v);
	if (tp->tp_compare == NULL && RICHCOMPARE(tp) == NULL) {
		return _Py_HashPointer(v); /* Use address as hash value */
	}
	/* If there's a cmp but no hash defined, the object can't be hashed */
	PyErr_Format(PyExc_TypeError, "unhashable type: '%.200s'",
		     v->ob_type->tp_name);
	return -1;
}

PyObject *
PyObject_GetAttrString(PyObject *v, const char *name)
{
	PyObject *w, *res;

	if (v->ob_type->tp_getattr != NULL)
		return (*v->ob_type->tp_getattr)(v, (char*)name);
	w = PyString_InternFromString(name);
	if (w == NULL)
		return NULL;
	res = PyObject_GetAttr(v, w);
	Py_XDECREF(w);
	return res;
}

int
PyObject_HasAttrString(PyObject *v, const char *name)
{
	PyObject *res = PyObject_GetAttrString(v, name);
	if (res != NULL) {
		Py_DECREF(res);
		return 1;
	}
	PyErr_Clear();
	return 0;
}

int
PyObject_SetAttrString(PyObject *v, const char *name, PyObject *w)
{
	PyObject *s;
	int res;

	if (v->ob_type->tp_setattr != NULL)
		return (*v->ob_type->tp_setattr)(v, (char*)name, w);
	s = PyString_InternFromString(name);
	if (s == NULL)
		return -1;
	res = PyObject_SetAttr(v, s, w);
	Py_XDECREF(s);
	return res;
}

PyObject *
PyObject_GetAttr(PyObject *v, PyObject *name)
{
	PyTypeObject *tp = v->ob_type;

	if (!PyString_Check(name)) {
#ifdef Py_USING_UNICODE
		/* The Unicode to string conversion is done here because the
		   existing tp_getattro slots expect a string object as name
		   and we wouldn't want to break those. */
		if (PyUnicode_Check(name)) {
			name = _PyUnicode_AsDefaultEncodedString(name, NULL);
			if (name == NULL)
				return NULL;
		}
		else
#endif
		{
			PyErr_Format(PyExc_TypeError,
				     "attribute name must be string, not '%.200s'",
				     name->ob_type->tp_name);
			return NULL;
		}
	}
	if (tp->tp_getattro != NULL)
		return (*tp->tp_getattro)(v, name);
	if (tp->tp_getattr != NULL)
		return (*tp->tp_getattr)(v, PyString_AS_STRING(name));
	PyErr_Format(PyExc_AttributeError,
		     "'%.50s' object has no attribute '%.400s'",
		     tp->tp_name, PyString_AS_STRING(name));
	return NULL;
}

int
PyObject_HasAttr(PyObject *v, PyObject *name)
{
	PyObject *res = PyObject_GetAttr(v, name);
	if (res != NULL) {
		Py_DECREF(res);
		return 1;
	}
	PyErr_Clear();
	return 0;
}

int
PyObject_SetAttr(PyObject *v, PyObject *name, PyObject *value)
{
	PyTypeObject *tp = v->ob_type;
	int err;

	if (!PyString_Check(name)){
#ifdef Py_USING_UNICODE
		/* The Unicode to string conversion is done here because the
		   existing tp_setattro slots expect a string object as name
		   and we wouldn't want to break those. */
		if (PyUnicode_Check(name)) {
			name = PyUnicode_AsEncodedString(name, NULL, NULL);
			if (name == NULL)
				return -1;
		}
		else
#endif
		{
			PyErr_Format(PyExc_TypeError,
				     "attribute name must be string, not '%.200s'",
				     name->ob_type->tp_name);
			return -1;
		}
	}
	else
		Py_INCREF(name);

	PyString_InternInPlace(&name);
	if (tp->tp_setattro != NULL) {
		err = (*tp->tp_setattro)(v, name, value);
		Py_DECREF(name);
		return err;
	}
	if (tp->tp_setattr != NULL) {
		err = (*tp->tp_setattr)(v, PyString_AS_STRING(name), value);
		Py_DECREF(name);
		return err;
	}
	Py_DECREF(name);
	if (tp->tp_getattr == NULL && tp->tp_getattro == NULL)
		PyErr_Format(PyExc_TypeError,
			     "'%.100s' object has no attributes "
			     "(%s .%.100s)",
			     tp->tp_name,
			     value==NULL ? "del" : "assign to",
			     PyString_AS_STRING(name));
	else
		PyErr_Format(PyExc_TypeError,
			     "'%.100s' object has only read-only attributes "
			     "(%s .%.100s)",
			     tp->tp_name,
			     value==NULL ? "del" : "assign to",
			     PyString_AS_STRING(name));
	return -1;
}

/* Helper to get a pointer to an object's __dict__ slot, if any */

PyObject **
_PyObject_GetDictPtr(PyObject *obj)
{
	Py_ssize_t dictoffset;
	PyTypeObject *tp = obj->ob_type;

	dictoffset = tp->tp_dictoffset;
	if (dictoffset == 0)
		return NULL;
	if (dictoffset < 0) {
		Py_ssize_t tsize;
		size_t size;

		tsize = ((PyVarObject *)obj)->ob_size;
		if (tsize < 0)
			tsize = -tsize;
		size = _PyObject_VAR_SIZE(tp, tsize);

		dictoffset += (long)size;
		assert(dictoffset > 0);
		assert(dictoffset % SIZEOF_VOID_P == 0);
	}
	return (PyObject **) ((char *)obj + dictoffset);
}

PyObject *
PyObject_SelfIter(PyObject *obj)
{
	Py_INCREF(obj);
	return obj;
}

/* Generic GetAttr functions - put these in your tp_[gs]etattro slot */

PyObject *
PyObject_GenericGetAttr(PyObject *obj, PyObject *name)
{
	PyTypeObject *tp = obj->ob_type;
	PyObject *descr = NULL;
	PyObject *res = NULL;
	descrgetfunc f;
	Py_ssize_t dictoffset;
	PyObject **dictptr;

	if (!PyString_Check(name)){
#ifdef Py_USING_UNICODE
		/* The Unicode to string conversion is done here because the
		   existing tp_setattro slots expect a string object as name
		   and we wouldn't want to break those. */
		if (PyUnicode_Check(name)) {
			name = PyUnicode_AsEncodedString(name, NULL, NULL);
			if (name == NULL)
				return NULL;
		}
		else
#endif
		{
			PyErr_Format(PyExc_TypeError,
				     "attribute name must be string, not '%.200s'",
				     name->ob_type->tp_name);
			return NULL;
		}
	}
	else
		Py_INCREF(name);

	if (tp->tp_dict == NULL) {
		if (PyType_Ready(tp) < 0)
			goto done;
	}

	/* Inline _PyType_Lookup */
	{
		Py_ssize_t i, n;
		PyObject *mro, *base, *dict;

		/* Look in tp_dict of types in MRO */
		mro = tp->tp_mro;
		assert(mro != NULL);
		assert(PyTuple_Check(mro));
		n = PyTuple_GET_SIZE(mro);
		for (i = 0; i < n; i++) {
			base = PyTuple_GET_ITEM(mro, i);
			if (PyClass_Check(base))
				dict = ((PyClassObject *)base)->cl_dict;
			else {
				assert(PyType_Check(base));
				dict = ((PyTypeObject *)base)->tp_dict;
			}
			assert(dict && PyDict_Check(dict));
			descr = PyDict_GetItem(dict, name);
			if (descr != NULL)
				break;
		}
	}

	Py_XINCREF(descr);

	f = NULL;
	if (descr != NULL) {
		f = descr->ob_type->tp_descr_get;
		if (f != NULL && PyDescr_IsData(descr)) {
			res = f(descr, obj, (PyObject *)obj->ob_type);
			Py_DECREF(descr);
			goto done;
		}
	}

	/* Inline _PyObject_GetDictPtr */
	dictoffset = tp->tp_dictoffset;
	if (dictoffset != 0) {
		PyObject *dict;
		if (dictoffset < 0) {
			Py_ssize_t tsize;
			size_t size;

			tsize = ((PyVarObject *)obj)->ob_size;
			if (tsize < 0)
				tsize = -tsize;
			size = _PyObject_VAR_SIZE(tp, tsize);

			dictoffset += (long)size;
			assert(dictoffset > 0);
			assert(dictoffset % SIZEOF_VOID_P == 0);
		}
		dictptr = (PyObject **) ((char *)obj + dictoffset);
		dict = *dictptr;
		if (dict != NULL) {
			res = PyDict_GetItem(dict, name);
			if (res != NULL) {
				Py_INCREF(res);
				Py_XDECREF(descr);
				goto done;
			}
		}
	}

	if (f != NULL) {
		res = f(descr, obj, (PyObject *)obj->ob_type);
		Py_DECREF(descr);
		goto done;
	}

	if (descr != NULL) {
		res = descr;
		/* descr was already increfed above */
		goto done;
	}

	PyErr_Format(PyExc_AttributeError,
		     "'%.50s' object has no attribute '%.400s'",
		     tp->tp_name, PyString_AS_STRING(name));
  done:
	Py_DECREF(name);
	return res;
}

int
PyObject_GenericSetAttr(PyObject *obj, PyObject *name, PyObject *value)
{
	PyTypeObject *tp = obj->ob_type;
	PyObject *descr;
	descrsetfunc f;
	PyObject **dictptr;
	int res = -1;

	if (!PyString_Check(name)){
#ifdef Py_USING_UNICODE
		/* The Unicode to string conversion is done here because the
		   existing tp_setattro slots expect a string object as name
		   and we wouldn't want to break those. */
		if (PyUnicode_Check(name)) {
			name = PyUnicode_AsEncodedString(name, NULL, NULL);
			if (name == NULL)
				return -1;
		}
		else
#endif
		{
			PyErr_Format(PyExc_TypeError,
				     "attribute name must be string, not '%.200s'",
				     name->ob_type->tp_name);
			return -1;
		}
	}
	else
		Py_INCREF(name);

	if (tp->tp_dict == NULL) {
		if (PyType_Ready(tp) < 0)
			goto done;
	}

	descr = _PyType_Lookup(tp, name);
	f = NULL;
	if (descr != NULL) {
		f = descr->ob_type->tp_descr_set;
		if (f != NULL && PyDescr_IsData(descr)) {
			res = f(descr, obj, value);
			goto done;
		}
	}

	dictptr = _PyObject_GetDictPtr(obj);
	if (dictptr != NULL) {
		PyObject *dict = *dictptr;
		if (dict == NULL && value != NULL) {
			dict = PyDict_New();
			if (dict == NULL)
				goto done;
			*dictptr = dict;
		}
		if (dict != NULL) {
			if (value == NULL)
				res = PyDict_DelItem(dict, name);
			else
				res = PyDict_SetItem(dict, name, value);
			if (res < 0 && PyErr_ExceptionMatches(PyExc_KeyError))
				PyErr_SetObject(PyExc_AttributeError, name);
			goto done;
		}
	}

	if (f != NULL) {
		res = f(descr, obj, value);
		goto done;
	}

	if (descr == NULL) {
		PyErr_Format(PyExc_AttributeError,
			     "'%.100s' object has no attribute '%.200s'",
			     tp->tp_name, PyString_AS_STRING(name));
		goto done;
	}

	PyErr_Format(PyExc_AttributeError,
		     "'%.50s' object attribute '%.400s' is read-only",
		     tp->tp_name, PyString_AS_STRING(name));
  done:
	Py_DECREF(name);
	return res;
}

/* Test a value used as condition, e.g., in a for or if statement.
   Return -1 if an error occurred */

int
PyObject_IsTrue(PyObject *v)
{
	Py_ssize_t res;
	if (v == Py_True)
		return 1;
	if (v == Py_False)
		return 0;
	if (v == Py_None)
		return 0;
	else if (v->ob_type->tp_as_number != NULL &&
		 v->ob_type->tp_as_number->nb_nonzero != NULL)
		res = (*v->ob_type->tp_as_number->nb_nonzero)(v);
	else if (v->ob_type->tp_as_mapping != NULL &&
		 v->ob_type->tp_as_mapping->mp_length != NULL)
		res = (*v->ob_type->tp_as_mapping->mp_length)(v);
	else if (v->ob_type->tp_as_sequence != NULL &&
		 v->ob_type->tp_as_sequence->sq_length != NULL)
		res = (*v->ob_type->tp_as_sequence->sq_length)(v);
	else
		return 1;
	/* if it is negative, it should be either -1 or -2 */
	return (res > 0) ? 1 : Py_SAFE_DOWNCAST(res, Py_ssize_t, int);
}

/* equivalent of 'not v'
   Return -1 if an error occurred */

int
PyObject_Not(PyObject *v)
{
	int res;
	res = PyObject_IsTrue(v);
	if (res < 0)
		return res;
	return res == 0;
}

/* Coerce two numeric types to the "larger" one.
   Increment the reference count on each argument.
   Return value:
   -1 if an error occurred;
   0 if the coercion succeeded (and then the reference counts are increased);
   1 if no coercion is possible (and no error is raised).
*/
int
PyNumber_CoerceEx(PyObject **pv, PyObject **pw)
{
	register PyObject *v = *pv;
	register PyObject *w = *pw;
	int res;

	if (v->ob_type->tp_as_number && v->ob_type->tp_as_number->nb_coerce) {
		res = (*v->ob_type->tp_as_number->nb_coerce)(pv, pw);
		if (res <= 0)
			return res;
	}
	if (w->ob_type->tp_as_number && w->ob_type->tp_as_number->nb_coerce) {
		res = (*w->ob_type->tp_as_number->nb_coerce)(pw, pv);
		if (res <= 0)
			return res;
	}
	return 1;
}

/* Coerce two numeric types to the "larger" one.
   Increment the reference count on each argument.
   Return -1 and raise an exception if no coercion is possible
   (and then no reference count is incremented).
*/
int
PyNumber_Coerce(PyObject **pv, PyObject **pw)
{
	int err = PyNumber_CoerceEx(pv, pw);
	if (err <= 0)
		return err;
	PyErr_SetString(PyExc_TypeError, "number coercion failed");
	return -1;
}


/* Test whether an object can be called */

int
PyCallable_Check(PyObject *x)
{
	if (x == NULL)
		return 0;
	if (PyInstance_Check(x)) {
		PyObject *call = PyObject_GetAttrString(x, "__call__");
		if (call == NULL) {
			PyErr_Clear();
			return 0;
		}
		/* Could test recursively but don't, for fear of endless
		   recursion if some joker sets self.__call__ = self */
		Py_DECREF(call);
		return 1;
	}
	else {
		return x->ob_type->tp_call != NULL;
	}
}

/* Helper for PyObject_Dir.
   Merge the __dict__ of aclass into dict, and recursively also all
   the __dict__s of aclass's base classes.  The order of merging isn't
   defined, as it's expected that only the final set of dict keys is
   interesting.
   Return 0 on success, -1 on error.
*/

static int
merge_class_dict(PyObject* dict, PyObject* aclass)
{
	PyObject *classdict;
	PyObject *bases;

	assert(PyDict_Check(dict));
	assert(aclass);

	/* Merge in the type's dict (if any). */
	classdict = PyObject_GetAttrString(aclass, "__dict__");
	if (classdict == NULL)
		PyErr_Clear();
	else {
		int status = PyDict_Update(dict, classdict);
		Py_DECREF(classdict);
		if (status < 0)
			return -1;
	}

	/* Recursively merge in the base types' (if any) dicts. */
	bases = PyObject_GetAttrString(aclass, "__bases__");
	if (bases == NULL)
		PyErr_Clear();
	else {
		/* We have no guarantee that bases is a real tuple */
		Py_ssize_t i, n;
		n = PySequence_Size(bases); /* This better be right */
		if (n < 0)
			PyErr_Clear();
		else {
			for (i = 0; i < n; i++) {
				int status;
				PyObject *base = PySequence_GetItem(bases, i);
				if (base == NULL) {
					Py_DECREF(bases);
					return -1;
				}
				status = merge_class_dict(dict, base);
				Py_DECREF(base);
				if (status < 0) {
					Py_DECREF(bases);
					return -1;
				}
			}
		}
		Py_DECREF(bases);
	}
	return 0;
}

/* Helper for PyObject_Dir.
   If obj has an attr named attrname that's a list, merge its string
   elements into keys of dict.
   Return 0 on success, -1 on error.  Errors due to not finding the attr,
   or the attr not being a list, are suppressed.
*/

static int
merge_list_attr(PyObject* dict, PyObject* obj, const char *attrname)
{
	PyObject *list;
	int result = 0;

	assert(PyDict_Check(dict));
	assert(obj);
	assert(attrname);

	list = PyObject_GetAttrString(obj, attrname);
	if (list == NULL)
		PyErr_Clear();

	else if (PyList_Check(list)) {
		int i;
		for (i = 0; i < PyList_GET_SIZE(list); ++i) {
			PyObject *item = PyList_GET_ITEM(list, i);
			if (PyString_Check(item)) {
				result = PyDict_SetItem(dict, item, Py_None);
				if (result < 0)
					break;
			}
		}
	}

	Py_XDECREF(list);
	return result;
}

/* Like __builtin__.dir(arg).  See bltinmodule.c's builtin_dir for the
   docstring, which should be kept in synch with this implementation. */

PyObject *
PyObject_Dir(PyObject *arg)
{
	/* Set exactly one of these non-NULL before the end. */
	PyObject *result = NULL;	/* result list */
	PyObject *masterdict = NULL;	/* result is masterdict.keys() */

	/* If NULL arg, return the locals. */
	if (arg == NULL) {
		PyObject *locals = PyEval_GetLocals();
		if (locals == NULL)
			goto error;
		result = PyMapping_Keys(locals);
		if (result == NULL)
			goto error;
	}

	/* Elif this is some form of module, we only want its dict. */
	else if (PyModule_Check(arg)) {
		masterdict = PyObject_GetAttrString(arg, "__dict__");
		if (masterdict == NULL)
			goto error;
		if (!PyDict_Check(masterdict)) {
			PyErr_SetString(PyExc_TypeError,
					"module.__dict__ is not a dictionary");
			goto error;
		}
	}

	/* Elif some form of type or class, grab its dict and its bases.
	   We deliberately don't suck up its __class__, as methods belonging
	   to the metaclass would probably be more confusing than helpful. */
	else if (PyType_Check(arg) || PyClass_Check(arg)) {
		masterdict = PyDict_New();
		if (masterdict == NULL)
			goto error;
		if (merge_class_dict(masterdict, arg) < 0)
			goto error;
	}

	/* Else look at its dict, and the attrs reachable from its class. */
	else {
		PyObject *itsclass;
		/* Create a dict to start with.  CAUTION:  Not everything
		   responding to __dict__ returns a dict! */
		masterdict = PyObject_GetAttrString(arg, "__dict__");
		if (masterdict == NULL) {
			PyErr_Clear();
			masterdict = PyDict_New();
		}
		else if (!PyDict_Check(masterdict)) {
			Py_DECREF(masterdict);
			masterdict = PyDict_New();
		}
		else {
			/* The object may have returned a reference to its
			   dict, so copy it to avoid mutating it. */
			PyObject *temp = PyDict_Copy(masterdict);
			Py_DECREF(masterdict);
			masterdict = temp;
		}
		if (masterdict == NULL)
			goto error;

		/* Merge in __members__ and __methods__ (if any).
		   XXX Would like this to go away someday; for now, it's
		   XXX needed to get at im_self etc of method objects. */
		if (merge_list_attr(masterdict, arg, "__members__") < 0)
			goto error;
		if (merge_list_attr(masterdict, arg, "__methods__") < 0)
			goto error;

		/* Merge in attrs reachable from its class.
		   CAUTION:  Not all objects have a __class__ attr. */
		itsclass = PyObject_GetAttrString(arg, "__class__");
		if (itsclass == NULL)
			PyErr_Clear();
		else {
			int status = merge_class_dict(masterdict, itsclass);
			Py_DECREF(itsclass);
			if (status < 0)
				goto error;
		}
	}

	assert((result == NULL) ^ (masterdict == NULL));
	if (masterdict != NULL) {
		/* The result comes from its keys. */
		assert(result == NULL);
		result = PyDict_Keys(masterdict);
		if (result == NULL)
			goto error;
	}

	assert(result);
	if (!PyList_Check(result)) {
		PyErr_Format(PyExc_TypeError,
			"Expected keys() to be a list, not '%.200s'",
			result->ob_type->tp_name);
		goto error;
	}
	if (PyList_Sort(result) != 0)
		goto error;
	else
		goto normal_return;

  error:
	Py_XDECREF(result);
	result = NULL;
	/* fall through */
  normal_return:
  	Py_XDECREF(masterdict);
	return result;
}

/*
NoObject is usable as a non-NULL undefined value, used by the macro None.
There is (and should be!) no way to create other objects of this type,
so there is exactly one (which is indestructible, by the way).
(XXX This type and the type of NotImplemented below should be unified.)
*/

/* ARGSUSED */
static PyObject *
none_repr(PyObject *op)
{
	return PyString_FromString("None");
}

/* ARGUSED */
static void
none_dealloc(PyObject* ignore)
{
	/* This should never get called, but we also don't want to SEGV if
	 * we accidently decref None out of existance.
	 */
	Py_FatalError("deallocating None");
}


static PyTypeObject PyNone_Type = {
	PyObject_HEAD_INIT(&PyType_Type)
	0,
	"NoneType",
	0,
	0,
	none_dealloc,	/*tp_dealloc*/ /*never called*/
	0,		/*tp_print*/
	0,		/*tp_getattr*/
	0,		/*tp_setattr*/
	0,		/*tp_compare*/
	none_repr,	/*tp_repr*/
	0,		/*tp_as_number*/
	0,		/*tp_as_sequence*/
	0,		/*tp_as_mapping*/
	0,		/*tp_hash */
};

PyObject _Py_NoneStruct = {
	PyObject_HEAD_INIT(&PyNone_Type)
};

/* NotImplemented is an object that can be used to signal that an
   operation is not implemented for the given type combination. */

static PyObject *
NotImplemented_repr(PyObject *op)
{
	return PyString_FromString("NotImplemented");
}

static PyTypeObject PyNotImplemented_Type = {
	PyObject_HEAD_INIT(&PyType_Type)
	0,
	"NotImplementedType",
	0,
	0,
	none_dealloc,	/*tp_dealloc*/ /*never called*/
	0,		/*tp_print*/
	0,		/*tp_getattr*/
	0,		/*tp_setattr*/
	0,		/*tp_compare*/
	NotImplemented_repr, /*tp_repr*/
	0,		/*tp_as_number*/
	0,		/*tp_as_sequence*/
	0,		/*tp_as_mapping*/
	0,		/*tp_hash */
};

PyObject _Py_NotImplementedStruct = {
	PyObject_HEAD_INIT(&PyNotImplemented_Type)
};

void
_Py_ReadyTypes(void)
{
	if (PyType_Ready(&PyType_Type) < 0)
		Py_FatalError("Can't initialize 'type'");

	if (PyType_Ready(&_PyWeakref_RefType) < 0)
		Py_FatalError("Can't initialize 'weakref'");

	if (PyType_Ready(&PyBool_Type) < 0)
		Py_FatalError("Can't initialize 'bool'");

	if (PyType_Ready(&PyBytes_Type) < 0)
		Py_FatalError("Can't initialize 'bytes'");

	if (PyType_Ready(&PyString_Type) < 0)
		Py_FatalError("Can't initialize 'str'");

	if (PyType_Ready(&PyList_Type) < 0)
		Py_FatalError("Can't initialize 'list'");

	if (PyType_Ready(&PyNone_Type) < 0)
		Py_FatalError("Can't initialize type(None)");

	if (PyType_Ready(&PyNotImplemented_Type) < 0)
		Py_FatalError("Can't initialize type(NotImplemented)");
}


#ifdef Py_TRACE_REFS

void
_Py_NewReference(PyObject *op)
{
	_Py_INC_REFTOTAL;
	op->ob_refcnt = 1;
	_Py_AddToAllObjects(op, 1);
	_Py_INC_TPALLOCS(op);
}

void
_Py_ForgetReference(register PyObject *op)
{
#ifdef SLOW_UNREF_CHECK
        register PyObject *p;
#endif
	if (op->ob_refcnt < 0)
		Py_FatalError("UNREF negative refcnt");
	if (op == &refchain ||
	    op->_ob_prev->_ob_next != op || op->_ob_next->_ob_prev != op)
		Py_FatalError("UNREF invalid object");
#ifdef SLOW_UNREF_CHECK
	for (p = refchain._ob_next; p != &refchain; p = p->_ob_next) {
		if (p == op)
			break;
	}
	if (p == &refchain) /* Not found */
		Py_FatalError("UNREF unknown object");
#endif
	op->_ob_next->_ob_prev = op->_ob_prev;
	op->_ob_prev->_ob_next = op->_ob_next;
	op->_ob_next = op->_ob_prev = NULL;
	_Py_INC_TPFREES(op);
}

void
_Py_Dealloc(PyObject *op)
{
	destructor dealloc = op->ob_type->tp_dealloc;
	_Py_ForgetReference(op);
	(*dealloc)(op);
}

/* Print all live objects.  Because PyObject_Print is called, the
 * interpreter must be in a healthy state.
 */
void
_Py_PrintReferences(FILE *fp)
{
	PyObject *op;
	fprintf(fp, "Remaining objects:\n");
	for (op = refchain._ob_next; op != &refchain; op = op->_ob_next) {
		fprintf(fp, "%p [%" PY_FORMAT_SIZE_T "d] ", op, op->ob_refcnt);
		if (PyObject_Print(op, fp, 0) != 0)
			PyErr_Clear();
		putc('\n', fp);
	}
}

/* Print the addresses of all live objects.  Unlike _Py_PrintReferences, this
 * doesn't make any calls to the Python C API, so is always safe to call.
 */
void
_Py_PrintReferenceAddresses(FILE *fp)
{
	PyObject *op;
	fprintf(fp, "Remaining object addresses:\n");
	for (op = refchain._ob_next; op != &refchain; op = op->_ob_next)
		fprintf(fp, "%p [%" PY_FORMAT_SIZE_T "d] %s\n", op,
			op->ob_refcnt, op->ob_type->tp_name);
}

PyObject *
_Py_GetObjects(PyObject *self, PyObject *args)
{
	int i, n;
	PyObject *t = NULL;
	PyObject *res, *op;

	if (!PyArg_ParseTuple(args, "i|O", &n, &t))
		return NULL;
	op = refchain._ob_next;
	res = PyList_New(0);
	if (res == NULL)
		return NULL;
	for (i = 0; (n == 0 || i < n) && op != &refchain; i++) {
		while (op == self || op == args || op == res || op == t ||
		       (t != NULL && op->ob_type != (PyTypeObject *) t)) {
			op = op->_ob_next;
			if (op == &refchain)
				return res;
		}
		if (PyList_Append(res, op) < 0) {
			Py_DECREF(res);
			return NULL;
		}
		op = op->_ob_next;
	}
	return res;
}

#endif


/* Hack to force loading of cobject.o */
PyTypeObject *_Py_cobject_hack = &PyCObject_Type;


/* Hack to force loading of abstract.o */
Py_ssize_t (*_Py_abstract_hack)(PyObject *) = PyObject_Size;


/* Python's malloc wrappers (see pymem.h) */

void *
PyMem_Malloc(size_t nbytes)
{
	return PyMem_MALLOC(nbytes);
}

void *
PyMem_Realloc(void *p, size_t nbytes)
{
	return PyMem_REALLOC(p, nbytes);
}

void
PyMem_Free(void *p)
{
	PyMem_FREE(p);
}


/* These methods are used to control infinite recursion in repr, str, print,
   etc.  Container objects that may recursively contain themselves,
   e.g. builtin dictionaries and lists, should used Py_ReprEnter() and
   Py_ReprLeave() to avoid infinite recursion.

   Py_ReprEnter() returns 0 the first time it is called for a particular
   object and 1 every time thereafter.  It returns -1 if an exception
   occurred.  Py_ReprLeave() has no return value.

   See dictobject.c and listobject.c for examples of use.
*/

#define KEY "Py_Repr"

int
Py_ReprEnter(PyObject *obj)
{
	PyObject *dict;
	PyObject *list;
	Py_ssize_t i;

	dict = PyThreadState_GetDict();
	if (dict == NULL)
		return 0;
	list = PyDict_GetItemString(dict, KEY);
	if (list == NULL) {
		list = PyList_New(0);
		if (list == NULL)
			return -1;
		if (PyDict_SetItemString(dict, KEY, list) < 0)
			return -1;
		Py_DECREF(list);
	}
	i = PyList_GET_SIZE(list);
	while (--i >= 0) {
		if (PyList_GET_ITEM(list, i) == obj)
			return 1;
	}
	PyList_Append(list, obj);
	return 0;
}

void
Py_ReprLeave(PyObject *obj)
{
	PyObject *dict;
	PyObject *list;
	Py_ssize_t i;

	dict = PyThreadState_GetDict();
	if (dict == NULL)
		return;
	list = PyDict_GetItemString(dict, KEY);
	if (list == NULL || !PyList_Check(list))
		return;
	i = PyList_GET_SIZE(list);
	/* Count backwards because we always expect obj to be list[-1] */
	while (--i >= 0) {
		if (PyList_GET_ITEM(list, i) == obj) {
			PyList_SetSlice(list, i, i + 1, NULL);
			break;
		}
	}
}

/* Trashcan support. */

/* Current call-stack depth of tp_dealloc calls. */
int _PyTrash_delete_nesting = 0;

/* List of objects that still need to be cleaned up, singly linked via their
 * gc headers' gc_prev pointers.
 */
PyObject *_PyTrash_delete_later = NULL;

/* Add op to the _PyTrash_delete_later list.  Called when the current
 * call-stack depth gets large.  op must be a currently untracked gc'ed
 * object, with refcount 0.  Py_DECREF must already have been called on it.
 */
void
_PyTrash_deposit_object(PyObject *op)
{
	assert(PyObject_IS_GC(op));
	assert(_Py_AS_GC(op)->gc.gc_refs == _PyGC_REFS_UNTRACKED);
	assert(op->ob_refcnt == 0);
	_Py_AS_GC(op)->gc.gc_prev = (PyGC_Head *)_PyTrash_delete_later;
	_PyTrash_delete_later = op;
}

/* Dealloccate all the objects in the _PyTrash_delete_later list.  Called when
 * the call-stack unwinds again.
 */
void
_PyTrash_destroy_chain(void)
{
	while (_PyTrash_delete_later) {
		PyObject *op = _PyTrash_delete_later;
		destructor dealloc = op->ob_type->tp_dealloc;

		_PyTrash_delete_later =
			(PyObject*) _Py_AS_GC(op)->gc.gc_prev;

		/* Call the deallocator directly.  This used to try to
		 * fool Py_DECREF into calling it indirectly, but
		 * Py_DECREF was already called on this object, and in
		 * assorted non-release builds calling Py_DECREF again ends
		 * up distorting allocation statistics.
		 */
		assert(op->ob_refcnt == 0);
		++_PyTrash_delete_nesting;
		(*dealloc)(op);
		--_PyTrash_delete_nesting;
	}
}

#ifdef __cplusplus
}
#endif