Commit ea134da9 authored by Serhiy Storchaka's avatar Serhiy Storchaka

Issue #16840: Tkinter now supports 64-bit integers added in Tcl 8.4 and

arbitrary precision integers added in Tcl 8.5.
parent 8e44aa5a
......@@ -133,9 +133,22 @@ class TclTest(unittest.TestCase):
tcl = self.interp
self.assertRaises(TclError,tcl.unsetvar,'a')
def get_integers(self):
integers = (0, 1, -1, 2**31-1, -2**31)
if tcl_version >= (8, 4): # wideInt was added in Tcl 8.4
integers += (2**31, -2**31-1, 2**63-1, -2**63)
if tcl_version >= (8, 5): # bignum was added in Tcl 8.5
integers += (2**63, -2**63-1, 2**1000, -2**1000)
return integers
def test_getint(self):
tcl = self.interp.tk
self.assertEqual(tcl.getint(' 42 '), 42)
for i in self.get_integers():
self.assertEqual(tcl.getint(' %d ' % i), i)
self.assertEqual(tcl.getint(' %#o ' % i), i)
self.assertEqual(tcl.getint(' %#x ' % i), i)
if tcl_version < (8, 5): # bignum was added in Tcl 8.5
self.assertRaises(TclError, tcl.getint, str(2**1000))
self.assertEqual(tcl.getint(42), 42)
self.assertRaises(TypeError, tcl.getint)
self.assertRaises(TypeError, tcl.getint, '42', '10')
......@@ -270,7 +283,7 @@ class TclTest(unittest.TestCase):
check('"a\xbd\u20ac"', 'a\xbd\u20ac')
check(r'"a\xbd\u20ac"', 'a\xbd\u20ac')
check(r'"a\0b"', 'a\x00b')
if tcl_version >= (8, 5):
if tcl_version >= (8, 5): # bignum was added in Tcl 8.5
check('2**64', str(2**64))
def test_exprdouble(self):
......@@ -302,7 +315,7 @@ class TclTest(unittest.TestCase):
check('[string length "a\xbd\u20ac"]', 3.0)
check(r'[string length "a\xbd\u20ac"]', 3.0)
self.assertRaises(TclError, tcl.exprdouble, '"abc"')
if tcl_version >= (8, 5):
if tcl_version >= (8, 5): # bignum was added in Tcl 8.5
check('2**64', float(2**64))
def test_exprlong(self):
......@@ -334,7 +347,7 @@ class TclTest(unittest.TestCase):
check('[string length "a\xbd\u20ac"]', 3)
check(r'[string length "a\xbd\u20ac"]', 3)
self.assertRaises(TclError, tcl.exprlong, '"abc"')
if tcl_version >= (8, 5):
if tcl_version >= (8, 5): # bignum was added in Tcl 8.5
self.assertRaises(TclError, tcl.exprlong, '2**64')
def test_exprboolean(self):
......@@ -375,7 +388,7 @@ class TclTest(unittest.TestCase):
check('[string length "a\xbd\u20ac"]', True)
check(r'[string length "a\xbd\u20ac"]', True)
self.assertRaises(TclError, tcl.exprboolean, '"abc"')
if tcl_version >= (8, 5):
if tcl_version >= (8, 5): # bignum was added in Tcl 8.5
check('2**64', True)
def test_booleans(self):
......@@ -397,6 +410,21 @@ class TclTest(unittest.TestCase):
check('1 < 2', True)
check('1 > 2', False)
def test_expr_bignum(self):
tcl = self.interp
for i in self.get_integers():
result = tcl.call('expr', str(i))
if self.wantobjects:
self.assertEqual(result, i)
self.assertIsInstance(result, int)
else:
self.assertEqual(result, str(i))
self.assertIsInstance(result, str)
if tcl_version < (8, 5): # bignum was added in Tcl 8.5
result = tcl.call('expr', str(2**1000))
self.assertEqual(result, str(2**1000))
self.assertIsInstance(result, str)
def test_passing_values(self):
def passValue(value):
return self.interp.call('set', '_', value)
......@@ -414,8 +442,10 @@ class TclTest(unittest.TestCase):
b'str\xc0\x80ing' if self.wantobjects else 'str\xc0\x80ing')
self.assertEqual(passValue(b'str\xbding'),
b'str\xbding' if self.wantobjects else 'str\xbding')
for i in (0, 1, -1, 2**31-1, -2**31):
for i in self.get_integers():
self.assertEqual(passValue(i), i if self.wantobjects else str(i))
if tcl_version < (8, 5): # bignum was added in Tcl 8.5
self.assertEqual(passValue(2**1000), str(2**1000))
for f in (0.0, 1.0, -1.0, 1/3,
sys.float_info.min, sys.float_info.max,
-sys.float_info.min, -sys.float_info.max):
......@@ -473,8 +503,10 @@ class TclTest(unittest.TestCase):
check(b'str\x00ing', 'str\x00ing')
check(b'str\xc0\x80ing', 'str\xc0\x80ing')
check(b'str\xc0\x80ing\xe2\x82\xac', 'str\xc0\x80ing\xe2\x82\xac')
for i in (0, 1, -1, 2**31-1, -2**31):
for i in self.get_integers():
check(i, str(i))
if tcl_version < (8, 5): # bignum was added in Tcl 8.5
check(2**1000, str(2**1000))
for f in (0.0, 1.0, -1.0):
check(f, repr(f))
for f in (1/3.0, sys.float_info.min, sys.float_info.max,
......
......@@ -24,6 +24,9 @@ Core and Builtins
Library
-------
- Issue #16840: Tkinter now supports 64-bit integers added in Tcl 8.4 and
arbitrary precision integers added in Tcl 8.5.
- Issue #23834: Fix socket.sendto(), use the C Py_ssize_t type to store the
result of sendto() instead of the C int type.
......
......@@ -69,6 +69,11 @@ Copyright (C) 1994 Steen Lumholt.
#error "Tk older than 8.3.1 not supported"
#endif
#if TK_VERSION_HEX >= 0x08050000
#define HAVE_LIBTOMMAMTH
#include <tclTomMath.h>
#endif
#if !(defined(MS_WINDOWS) || defined(__CYGWIN__))
#define HAVE_CREATEFILEHANDLER
#endif
......@@ -247,6 +252,8 @@ typedef struct {
const Tcl_ObjType *ByteArrayType;
const Tcl_ObjType *DoubleType;
const Tcl_ObjType *IntType;
const Tcl_ObjType *WideIntType;
const Tcl_ObjType *BignumType;
const Tcl_ObjType *ListType;
const Tcl_ObjType *ProcBodyType;
const Tcl_ObjType *StringType;
......@@ -583,6 +590,8 @@ Tkapp_New(char *screenName, char *className,
v->ByteArrayType = Tcl_GetObjType("bytearray");
v->DoubleType = Tcl_GetObjType("double");
v->IntType = Tcl_GetObjType("int");
v->WideIntType = Tcl_GetObjType("wideInt");
v->BignumType = Tcl_GetObjType("bignum");
v->ListType = Tcl_GetObjType("list");
v->ProcBodyType = Tcl_GetObjType("procbody");
v->StringType = Tcl_GetObjType("string");
......@@ -875,28 +884,92 @@ static PyType_Spec PyTclObject_Type_spec = {
#define CHECK_STRING_LENGTH(s)
#endif
#ifdef HAVE_LIBTOMMAMTH
static Tcl_Obj*
asBignumObj(PyObject *value)
{
Tcl_Obj *result;
int neg;
PyObject *hexstr;
char *hexchars;
mp_int bigValue;
neg = Py_SIZE(value) < 0;
hexstr = _PyLong_Format(value, 16);
if (hexstr == NULL)
return NULL;
hexchars = PyUnicode_AsUTF8(hexstr);
if (hexchars == NULL) {
Py_DECREF(hexstr);
return NULL;
}
hexchars += neg + 2; /* skip sign and "0x" */
mp_init(&bigValue);
if (mp_read_radix(&bigValue, hexchars, 16) != MP_OKAY) {
mp_clear(&bigValue);
Py_DECREF(hexstr);
PyErr_NoMemory();
return NULL;
}
Py_DECREF(hexstr);
bigValue.sign = neg ? MP_NEG : MP_ZPOS;
result = Tcl_NewBignumObj(&bigValue);
mp_clear(&bigValue);
if (result == NULL) {
PyErr_NoMemory();
return NULL;
}
return result;
}
#endif
static Tcl_Obj*
AsObj(PyObject *value)
{
Tcl_Obj *result;
long longVal;
int overflow;
if (PyBytes_Check(value))
return Tcl_NewByteArrayObj((unsigned char *)PyBytes_AS_STRING(value),
PyBytes_GET_SIZE(value));
else if (PyBool_Check(value))
if (PyBool_Check(value))
return Tcl_NewBooleanObj(PyObject_IsTrue(value));
else if (PyLong_CheckExact(value) &&
((longVal = PyLong_AsLongAndOverflow(value, &overflow)),
!overflow)) {
if (PyLong_CheckExact(value)) {
int overflow;
long longValue;
#ifdef TCL_WIDE_INT_TYPE
Tcl_WideInt wideValue;
#endif
longValue = PyLong_AsLongAndOverflow(value, &overflow);
if (!overflow) {
return Tcl_NewLongObj(longValue);
}
/* If there is an overflow in the long conversion,
fall through to wideInt handling. */
#ifdef TCL_WIDE_INT_TYPE
if (_PyLong_AsByteArray((PyLongObject *)value,
(unsigned char *)(void *)&wideValue,
sizeof(wideValue),
PY_LITTLE_ENDIAN,
/* signed */ 1) == 0) {
return Tcl_NewWideIntObj(wideValue);
}
PyErr_Clear();
#endif
/* If there is an overflow in the wideInt conversion,
fall through to bignum handling. */
#ifdef HAVE_LIBTOMMAMTH
return asBignumObj(value);
#endif
/* If there is no wideInt or bignum support,
fall through to default object handling. */
return Tcl_NewLongObj(longVal);
}
else if (PyFloat_Check(value))
if (PyFloat_Check(value))
return Tcl_NewDoubleObj(PyFloat_AS_DOUBLE(value));
else if (PyTuple_Check(value)) {
if (PyTuple_Check(value)) {
Tcl_Obj **argv;
Py_ssize_t size, i;
......@@ -916,7 +989,8 @@ AsObj(PyObject *value)
ckfree(FREECAST argv);
return result;
}
else if (PyUnicode_Check(value)) {
if (PyUnicode_Check(value)) {
void *inbuf;
Py_ssize_t size;
int kind;
......@@ -966,12 +1040,14 @@ AsObj(PyObject *value)
ckfree(FREECAST outbuf);
return result;
}
else if(PyTclObject_Check(value)) {
if (PyTclObject_Check(value)) {
Tcl_Obj *v = ((PyTclObject*)value)->value;
Tcl_IncrRefCount(v);
return v;
}
else {
{
PyObject *v = PyObject_Str(value);
if (!v)
return 0;
......@@ -990,6 +1066,62 @@ fromBoolean(PyObject* tkapp, Tcl_Obj *value)
return PyBool_FromLong(boolValue);
}
#ifdef TCL_WIDE_INT_TYPE
static PyObject*
fromWideIntObj(PyObject* tkapp, Tcl_Obj *value)
{
Tcl_WideInt wideValue;
if (Tcl_GetWideIntFromObj(Tkapp_Interp(tkapp), value, &wideValue) == TCL_OK) {
#ifdef HAVE_LONG_LONG
if (sizeof(wideValue) <= SIZEOF_LONG_LONG)
return PyLong_FromLongLong(wideValue);
#endif
return _PyLong_FromByteArray((unsigned char *)(void *)&wideValue,
sizeof(wideValue),
PY_LITTLE_ENDIAN,
/* signed */ 1);
}
return NULL;
}
#endif
#ifdef HAVE_LIBTOMMAMTH
static PyObject*
fromBignumObj(PyObject* tkapp, Tcl_Obj *value)
{
mp_int bigValue;
unsigned long numBytes;
unsigned char *bytes;
PyObject *res;
if (Tcl_GetBignumFromObj(Tkapp_Interp(tkapp), value, &bigValue) != TCL_OK)
return Tkinter_Error(tkapp);
numBytes = mp_unsigned_bin_size(&bigValue);
bytes = PyMem_Malloc(numBytes);
if (bytes == NULL) {
mp_clear(&bigValue);
return PyErr_NoMemory();
}
if (mp_to_unsigned_bin_n(&bigValue, bytes,
&numBytes) != MP_OKAY) {
mp_clear(&bigValue);
PyMem_Free(bytes);
return PyErr_NoMemory();
}
res = _PyLong_FromByteArray(bytes, numBytes,
/* big-endian */ 0,
/* unsigned */ 0);
PyMem_Free(bytes);
if (res != NULL && bigValue.sign == MP_NEG) {
PyObject *res2 = PyNumber_Negative(res);
Py_DECREF(res);
res = res2;
}
mp_clear(&bigValue);
return res;
}
#endif
static PyObject*
FromObj(PyObject* tkapp, Tcl_Obj *value)
{
......@@ -1017,8 +1149,32 @@ FromObj(PyObject* tkapp, Tcl_Obj *value)
}
if (value->typePtr == app->IntType) {
return PyLong_FromLong(value->internalRep.longValue);
long longValue;
if (Tcl_GetLongFromObj(interp, value, &longValue) == TCL_OK)
return PyLong_FromLong(longValue);
/* If there is an error in the long conversion,
fall through to wideInt handling. */
}
#ifdef TCL_WIDE_INT_TYPE
if (value->typePtr == app->IntType ||
value->typePtr == app->WideIntType) {
result = fromWideIntObj(tkapp, value);
if (result != NULL || PyErr_Occurred())
return result;
Tcl_ResetResult(interp);
/* If there is an error in the wideInt conversion,
fall through to bignum handling. */
}
#endif
#ifdef HAVE_LIBTOMMAMTH
if (value->typePtr == app->IntType ||
value->typePtr == app->WideIntType ||
value->typePtr == app->BignumType) {
return fromBignumObj(tkapp, value);
}
#endif
if (value->typePtr == app->ListType) {
int size;
......@@ -1067,6 +1223,15 @@ FromObj(PyObject* tkapp, Tcl_Obj *value)
}
#endif
#ifdef HAVE_LIBTOMMAMTH
if (app->BignumType == NULL &&
strcmp(value->typePtr->name, "bignum") == 0) {
/* bignum type is not registered in Tcl */
app->BignumType = value->typePtr;
return fromBignumObj(tkapp, value);
}
#endif
return newPyTclObject(value);
}
......@@ -1700,7 +1865,12 @@ static PyObject *
Tkapp_GetInt(PyObject *self, PyObject *args)
{
char *s;
int v;
#if defined(TCL_WIDE_INT_TYPE) || defined(HAVE_LIBTOMMAMTH)
Tcl_Obj *value;
PyObject *result;
#else
int intValue;
#endif
if (PyTuple_Size(args) == 1) {
PyObject* o = PyTuple_GetItem(args, 0);
......@@ -1712,9 +1882,29 @@ Tkapp_GetInt(PyObject *self, PyObject *args)
if (!PyArg_ParseTuple(args, "s:getint", &s))
return NULL;
CHECK_STRING_LENGTH(s);
if (Tcl_GetInt(Tkapp_Interp(self), s, &v) == TCL_ERROR)
#if defined(TCL_WIDE_INT_TYPE) || defined(HAVE_LIBTOMMAMTH)
value = Tcl_NewStringObj(s, -1);
if (value == NULL)
return Tkinter_Error(self);
return Py_BuildValue("i", v);
/* Don't use Tcl_GetInt() because it returns ambiguous result for value
in ranges -2**32..-2**31-1 and 2**31..2**32-1 (on 32-bit platform).
Prefer bignum because Tcl_GetWideIntFromObj returns ambiguous result for
value in ranges -2**64..-2**63-1 and 2**63..2**64-1 (on 32-bit platform).
*/
#ifdef HAVE_LIBTOMMAMTH
result = fromBignumObj(self, value);
#else
result = fromWideIntObj(self, value);
#endif
Tcl_DecrRefCount(value);
if (result != NULL || PyErr_Occurred())
return result;
#else
if (Tcl_GetInt(Tkapp_Interp(self), s, &intValue) == TCL_OK)
return PyLong_FromLong(intValue);
#endif
return Tkinter_Error(self);
}
static PyObject *
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment