Issue #16741: Fix an error reporting in int().
diff --git a/Include/longobject.h b/Include/longobject.h
index 886bc63..24cf401 100644
--- a/Include/longobject.h
+++ b/Include/longobject.h
@@ -97,6 +97,7 @@
#ifndef Py_LIMITED_API
PyAPI_FUNC(PyObject *) PyLong_FromUnicode(Py_UNICODE*, Py_ssize_t, int);
PyAPI_FUNC(PyObject *) PyLong_FromUnicodeObject(PyObject *u, int base);
+PyAPI_FUNC(PyObject *) _PyLong_FromBytes(const char *, Py_ssize_t, int);
#endif
#ifndef Py_LIMITED_API
diff --git a/Lib/test/test_int.py b/Lib/test/test_int.py
index 4921230..4e00622 100644
--- a/Lib/test/test_int.py
+++ b/Lib/test/test_int.py
@@ -73,14 +73,6 @@
x = -1-sys.maxsize
self.assertEqual(x >> 1, x//2)
- self.assertRaises(ValueError, int, '123\0')
- self.assertRaises(ValueError, int, '53', 40)
-
- # SF bug 1545497: embedded NULs were not detected with
- # explicit base
- self.assertRaises(ValueError, int, '123\0', 10)
- self.assertRaises(ValueError, int, '123\x00 245', 20)
-
x = int('1' * 600)
self.assertIsInstance(x, int)
@@ -401,14 +393,37 @@
int(TruncReturnsBadInt())
def test_error_message(self):
- testlist = ('\xbd', '123\xbd', ' 123 456 ')
- for s in testlist:
- try:
- int(s)
- except ValueError as e:
- self.assertIn(s.strip(), e.args[0])
- else:
- self.fail("Expected int(%r) to raise a ValueError", s)
+ def check(s, base=None):
+ with self.assertRaises(ValueError,
+ msg="int(%r, %r)" % (s, base)) as cm:
+ if base is None:
+ int(s)
+ else:
+ int(s, base)
+ self.assertEqual(cm.exception.args[0],
+ "invalid literal for int() with base %d: %r" %
+ (10 if base is None else base, s))
+
+ check('\xbd')
+ check('123\xbd')
+ check(' 123 456 ')
+
+ check('123\x00')
+ # SF bug 1545497: embedded NULs were not detected with explicit base
+ check('123\x00', 10)
+ check('123\x00 245', 20)
+ check('123\x00 245', 16)
+ check('123\x00245', 20)
+ check('123\x00245', 16)
+ # byte string with embedded NUL
+ check(b'123\x00')
+ check(b'123\x00', 10)
+ # non-UTF-8 byte string
+ check(b'123\xbd')
+ check(b'123\xbd', 10)
+ # lone surrogate in Unicode string
+ check('123\ud800')
+ check('123\ud800', 10)
def test_main():
support.run_unittest(IntTestCases)
diff --git a/Misc/NEWS b/Misc/NEWS
index 10c0470..61857a9 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -10,6 +10,8 @@
Core and Builtins
-----------------
+- Issue #16741: Fix an error reporting in int().
+
- Issue #17899: Fix rare file descriptor leak in os.listdir().
- Issue #9035: ismount now recognises volumes mounted below a drive root
diff --git a/Objects/abstract.c b/Objects/abstract.c
index 6c8c561..5f11c3d 100644
--- a/Objects/abstract.c
+++ b/Objects/abstract.c
@@ -1261,25 +1261,6 @@
}
-/* Add a check for embedded NULL-bytes in the argument. */
-static PyObject *
-long_from_string(const char *s, Py_ssize_t len)
-{
- char *end;
- PyObject *x;
-
- x = PyLong_FromString((char*)s, &end, 10);
- if (x == NULL)
- return NULL;
- if (end != s + len) {
- PyErr_SetString(PyExc_ValueError,
- "null byte in argument for int()");
- Py_DECREF(x);
- return NULL;
- }
- return x;
-}
-
PyObject *
PyNumber_Long(PyObject *o)
{
@@ -1327,16 +1308,16 @@
if (PyBytes_Check(o))
/* need to do extra error checking that PyLong_FromString()
- * doesn't do. In particular int('9.5') must raise an
- * exception, not truncate the float.
+ * doesn't do. In particular int('9\x005') must raise an
+ * exception, not truncate at the null.
*/
- return long_from_string(PyBytes_AS_STRING(o),
- PyBytes_GET_SIZE(o));
+ return _PyLong_FromBytes(PyBytes_AS_STRING(o),
+ PyBytes_GET_SIZE(o), 10);
if (PyUnicode_Check(o))
/* The above check is done in PyLong_FromUnicode(). */
return PyLong_FromUnicodeObject(o, 10);
if (!PyObject_AsCharBuffer(o, &buffer, &buffer_len))
- return long_from_string(buffer, buffer_len);
+ return _PyLong_FromBytes(buffer, buffer_len, 10);
return type_error("int() argument must be a string or a "
"number, not '%.200s'", o);
diff --git a/Objects/longobject.c b/Objects/longobject.c
index b47be40..e0d641a0 100644
--- a/Objects/longobject.c
+++ b/Objects/longobject.c
@@ -2000,6 +2000,14 @@
return long_normalize(z);
}
+/* Parses a long from a bytestring. Leading and trailing whitespace will be
+ * ignored.
+ *
+ * If successful, a PyLong object will be returned and 'pend' will be pointing
+ * to the first unused byte unless it's NULL.
+ *
+ * If unsuccessful, NULL will be returned.
+ */
PyObject *
PyLong_FromString(char *str, char **pend, int base)
{
@@ -2262,24 +2270,54 @@
str++;
if (*str != '\0')
goto onError;
- if (pend)
- *pend = str;
long_normalize(z);
- return (PyObject *) maybe_small_long(z);
+ z = maybe_small_long(z);
+ if (z == NULL)
+ return NULL;
+ if (pend != NULL)
+ *pend = str;
+ return (PyObject *) z;
onError:
+ if (pend != NULL)
+ *pend = str;
Py_XDECREF(z);
slen = strlen(orig_str) < 200 ? strlen(orig_str) : 200;
strobj = PyUnicode_FromStringAndSize(orig_str, slen);
if (strobj == NULL)
return NULL;
PyErr_Format(PyExc_ValueError,
- "invalid literal for int() with base %d: %R",
+ "invalid literal for int() with base %d: %.200R",
base, strobj);
Py_DECREF(strobj);
return NULL;
}
+/* Since PyLong_FromString doesn't have a length parameter,
+ * check here for possible NULs in the string.
+ *
+ * Reports an invalid literal as a bytes object.
+ */
+PyObject *
+_PyLong_FromBytes(const char *s, Py_ssize_t len, int base)
+{
+ PyObject *result, *strobj;
+ char *end = NULL;
+
+ result = PyLong_FromString((char*)s, &end, base);
+ if (end == NULL || (result != NULL && end == s + len))
+ return result;
+ Py_XDECREF(result);
+ strobj = PyBytes_FromStringAndSize(s, Py_MIN(len, 200));
+ if (strobj != NULL) {
+ PyErr_Format(PyExc_ValueError,
+ "invalid literal for int() with base %d: %.200R",
+ base, strobj);
+ Py_DECREF(strobj);
+ }
+ return NULL;
+}
+
PyObject *
PyLong_FromUnicode(Py_UNICODE *u, Py_ssize_t length, int base)
{
@@ -2294,9 +2332,8 @@
PyObject *
PyLong_FromUnicodeObject(PyObject *u, int base)
{
- PyObject *result;
- PyObject *asciidig;
- char *buffer, *end;
+ PyObject *result, *asciidig;
+ char *buffer, *end = NULL;
Py_ssize_t buflen;
asciidig = _PyUnicode_TransformDecimalAndSpaceToASCII(u);
@@ -2305,17 +2342,22 @@
buffer = PyUnicode_AsUTF8AndSize(asciidig, &buflen);
if (buffer == NULL) {
Py_DECREF(asciidig);
- return NULL;
+ if (!PyErr_ExceptionMatches(PyExc_UnicodeEncodeError))
+ return NULL;
}
- result = PyLong_FromString(buffer, &end, base);
- if (result != NULL && end != buffer + buflen) {
- PyErr_SetString(PyExc_ValueError,
- "null byte in argument for int()");
- Py_DECREF(result);
- result = NULL;
+ else {
+ result = PyLong_FromString(buffer, &end, base);
+ if (end == NULL || (result != NULL && end == buffer + buflen)) {
+ Py_DECREF(asciidig);
+ return result;
+ }
+ Py_DECREF(asciidig);
+ Py_XDECREF(result);
}
- Py_DECREF(asciidig);
- return result;
+ PyErr_Format(PyExc_ValueError,
+ "invalid literal for int() with base %d: %.200R",
+ base, u);
+ return NULL;
}
/* forward */
@@ -4319,23 +4361,12 @@
if (PyUnicode_Check(x))
return PyLong_FromUnicodeObject(x, (int)base);
else if (PyByteArray_Check(x) || PyBytes_Check(x)) {
- /* Since PyLong_FromString doesn't have a length parameter,
- * check here for possible NULs in the string. */
char *string;
- Py_ssize_t size = Py_SIZE(x);
if (PyByteArray_Check(x))
string = PyByteArray_AS_STRING(x);
else
string = PyBytes_AS_STRING(x);
- if (strlen(string) != (size_t)size || !size) {
- /* We only see this if there's a null byte in x or x is empty,
- x is a bytes or buffer, *and* a base is given. */
- PyErr_Format(PyExc_ValueError,
- "invalid literal for int() with base %d: %R",
- (int)base, x);
- return NULL;
- }
- return PyLong_FromString(string, NULL, (int)base);
+ return _PyLong_FromBytes(string, Py_SIZE(x), (int)base);
}
else {
PyErr_SetString(PyExc_TypeError,