Based on patch by Victor Stinner.
Add private C API function _PyUnicode_AsUnicode() which is similar to
PyUnicode_AsUnicode(), but checks for null characters.
PyAPI_FUNC(Py_UCS4*) PyUnicode_AsUCS4Copy(PyObject *unicode);
#endif
+#ifndef Py_LIMITED_API
/* Return a read-only pointer to the Unicode object's internal
Py_UNICODE buffer.
If the wchar_t/Py_UNICODE representation is not yet available, this
function will calculate it. */
-#ifndef Py_LIMITED_API
PyAPI_FUNC(Py_UNICODE *) PyUnicode_AsUnicode(
PyObject *unicode /* Unicode object */
) /* Py_DEPRECATED(3.3) */;
-#endif
+
+/* Similar to PyUnicode_AsUnicode(), but raises a ValueError if the string
+ contains null characters. */
+PyAPI_FUNC(const Py_UNICODE *) _PyUnicode_AsUnicode(
+ PyObject *unicode /* Unicode object */
+ );
/* Return a read-only pointer to the Unicode object's internal
Py_UNICODE buffer and save the length at size.
If the wchar_t/Py_UNICODE representation is not yet available, this
function will calculate it. */
-#ifndef Py_LIMITED_API
PyAPI_FUNC(Py_UNICODE *) PyUnicode_AsUnicodeAndSize(
PyObject *unicode, /* Unicode object */
Py_ssize_t *size /* location where to save the length */
windll["kernel32"].GetModuleHandleW
windll.LoadLibrary("kernel32").GetModuleHandleW
WinDLL("kernel32").GetModuleHandleW
+ # embedded null character
+ self.assertRaises(ValueError, windll.LoadLibrary, "kernel32\0")
@unittest.skipUnless(os.name == "nt",
'test specific to Windows')
self.assertRaises(TypeError, __import__, 1, 2, 3, 4)
self.assertRaises(ValueError, __import__, '')
self.assertRaises(TypeError, __import__, 'sys', name='sys')
+ # embedded null character
+ self.assertRaises(ModuleNotFoundError, __import__, 'string\x00')
def test_abs(self):
# int
self.assertEqual(fp.read(300), 'XXX'*100)
self.assertEqual(fp.read(1000), 'YYY'*100)
+ # embedded null bytes and characters
+ self.assertRaises(ValueError, open, 'a\x00b')
+ self.assertRaises(ValueError, open, b'a\x00b')
+
def test_open_default_encoding(self):
old_environ = dict(os.environ)
try:
win2 = curses.newwin(15,15, 5,5)
for meth in [stdscr.addch, stdscr.addstr]:
- for args in [('a'), ('a', curses.A_BOLD),
+ for args in [('a',), ('a', curses.A_BOLD),
(4,4, 'a'), (5,5, 'a', curses.A_BOLD)]:
with self.subTest(meth=meth.__qualname__, args=args):
meth(*args)
self.assertRaises(ValueError, stdscr.instr, -2)
self.assertRaises(ValueError, stdscr.instr, 2, 3, -2)
+ def test_embedded_null_chars(self):
+ # reject embedded null bytes and characters
+ stdscr = self.stdscr
+ for arg in ['a', b'a']:
+ with self.subTest(arg=arg):
+ self.assertRaises(ValueError, stdscr.addstr, 'a\0')
+ self.assertRaises(ValueError, stdscr.addnstr, 'a\0', 1)
+ self.assertRaises(ValueError, stdscr.insstr, 'a\0')
+ self.assertRaises(ValueError, stdscr.insnstr, 'a\0', 1)
def test_module_funcs(self):
"Test module-level functions"
self.assertRaises(TypeError, grp.getgrgid)
self.assertRaises(TypeError, grp.getgrnam)
self.assertRaises(TypeError, grp.getgrall, 42)
+ # embedded null character
+ self.assertRaises(ValueError, grp.getgrnam, 'a\x00b')
# try to get some errors
bynames = {}
loader.get_data(imp.__file__) # File should be closed
loader.get_data(imp.__file__) # Will need to create a newly opened file
+ def test_load_source(self):
+ with self.assertRaisesRegex(ValueError, 'embedded null'):
+ imp.load_source(__name__, __file__ + "\0")
+
class ReloadTests(unittest.TestCase):
self.assertLess(locale.strcoll('a', 'b'), 0)
self.assertEqual(locale.strcoll('a', 'a'), 0)
self.assertGreater(locale.strcoll('b', 'a'), 0)
+ # embedded null character
+ self.assertRaises(ValueError, locale.strcoll, 'a\0', 'a')
+ self.assertRaises(ValueError, locale.strcoll, 'a', 'a\0')
def test_strxfrm(self):
self.assertLess(locale.strxfrm('a'), locale.strxfrm('b'))
+ # embedded null character
+ self.assertRaises(ValueError, locale.strxfrm, 'a\0')
class TestEnUSCollation(BaseLocalizedTest, TestCollation):
except ValueError:
self.fail('conversion specifier: %r failed.' % format)
+ self.assertRaises(TypeError, time.strftime, b'%S', tt)
+ # embedded null character
+ self.assertRaises(ValueError, time.strftime, '%S\0', tt)
+
def _bounds_checking(self, func):
# Make sure that strftime() checks the bounds of the various parts
# of the time tuple (0 is valid for *all* values).
self.assertRaises(TypeError, winsound.PlaySound, "bad",
winsound.SND_MEMORY)
self.assertRaises(TypeError, winsound.PlaySound, 1, 0)
+ # embedded null character
+ self.assertRaises(ValueError, winsound.PlaySound, 'bad\0', 0)
def test_keyword_args(self):
safe_PlaySound(flags=winsound.SND_ALIAS, sound="SystemExit")
PyObject *nameobj;
PyObject *ignored;
HMODULE hMod;
- if (!PyArg_ParseTuple(args, "O|O:LoadLibrary", &nameobj, &ignored))
+
+ if (!PyArg_ParseTuple(args, "U|O:LoadLibrary", &nameobj, &ignored))
return NULL;
- name = PyUnicode_AsUnicode(nameobj);
+ name = _PyUnicode_AsUnicode(nameobj);
if (!name)
return NULL;
PyCurses_ConvertToString(PyCursesWindowObject *win, PyObject *obj,
PyObject **bytes, wchar_t **wstr)
{
+ char *str;
if (PyUnicode_Check(obj)) {
#ifdef HAVE_NCURSESW
assert (wstr != NULL);
+
*wstr = PyUnicode_AsWideCharString(obj, NULL);
if (*wstr == NULL)
return 0;
*bytes = PyUnicode_AsEncodedString(obj, win->encoding, NULL);
if (*bytes == NULL)
return 0;
+ /* check for embedded null bytes */
+ if (PyBytes_AsStringAndSize(*bytes, &str, NULL) < 0) {
+ return 0;
+ }
return 1;
#endif
}
else if (PyBytes_Check(obj)) {
Py_INCREF(obj);
*bytes = obj;
+ /* check for embedded null bytes */
+ if (PyBytes_AsStringAndSize(*bytes, &str, NULL) < 0) {
+ return 0;
+ }
return 1;
}
if (fd < 0) {
#ifdef MS_WINDOWS
- Py_ssize_t length;
if (!PyUnicode_FSDecoder(nameobj, &stringobj)) {
return -1;
}
- widename = PyUnicode_AsUnicodeAndSize(stringobj, &length);
+ widename = PyUnicode_AsUnicode(stringobj);
if (widename == NULL)
return -1;
#else
s = PyUnicode_AsWideCharString(str, &n1);
if (s == NULL)
goto exit;
+ if (wcslen(s) != (size_t)n1) {
+ PyErr_SetString(PyExc_ValueError,
+ "embedded null character");
+ goto exit;
+ }
/* assume no change in size, first */
n1 = n1 + 1;
if ((bytes = PyUnicode_EncodeFSDefault(name)) == NULL)
return NULL;
+ /* check for embedded null bytes */
if (PyBytes_AsStringAndSize(bytes, &name_chars, NULL) == -1)
goto out;
return NULL;
if ((bkey = PyUnicode_EncodeFSDefault(ukey)) == NULL)
return NULL;
+ /* check for embedded null bytes */
if (PyBytes_AsStringAndSize(bkey, &key, &keylen) == -1) {
Py_DECREF(bkey);
return NULL;
PyObject *result;
const wchar_t *path_wchar;
- path_wchar = PyUnicode_AsUnicode(path);
+ path_wchar = _PyUnicode_AsUnicode(path);
if (path_wchar == NULL)
return NULL;
))
return NULL;
- path = PyUnicode_AsUnicode(po);
+ path = _PyUnicode_AsUnicode(po);
if (path == NULL)
return NULL;
/*[clinic end generated code: output=d29a567d6b2327d2 input=ba586581c2e6105f]*/
{
const wchar_t *env;
+ Py_ssize_t size;
/* Search from index 1 because on Windows starting '=' is allowed for
defining hidden environment variables. */
if (unicode == NULL) {
return NULL;
}
- if (_MAX_ENV < PyUnicode_GET_LENGTH(unicode)) {
+
+ env = PyUnicode_AsUnicodeAndSize(unicode, &size);
+ if (env == NULL)
+ goto error;
+ if (size > _MAX_ENV) {
PyErr_Format(PyExc_ValueError,
"the environment variable is longer than %u characters",
_MAX_ENV);
goto error;
}
-
- env = PyUnicode_AsUnicode(unicode);
- if (env == NULL)
+ if (wcslen(env) != (size_t)size) {
+ PyErr_SetString(PyExc_ValueError, "embedded null character");
goto error;
+ }
+
if (_wputenv(env)) {
posix_error();
goto error;
if ((bytes = PyUnicode_EncodeFSDefault(arg)) == NULL)
return NULL;
+ /* check for embedded null bytes */
if (PyBytes_AsStringAndSize(bytes, &name, NULL) == -1)
goto out;
if ((p = getpwnam(name)) == NULL) {
if ((bytes = PyUnicode_EncodeFSDefault(arg)) == NULL)
return NULL;
+ /* check for embedded null bytes */
if (PyBytes_AsStringAndSize(bytes, &name, NULL) == -1)
goto out;
if ((p = getspnam(name)) == NULL) {
return PyUnicode_AsUnicodeAndSize(unicode, NULL);
}
+const Py_UNICODE *
+_PyUnicode_AsUnicode(PyObject *unicode)
+{
+ Py_ssize_t size;
+ const Py_UNICODE *wstr;
+
+ wstr = PyUnicode_AsUnicodeAndSize(unicode, &size);
+ if (wstr && wcslen(wstr) != (size_t)size) {
+ PyErr_SetString(PyExc_ValueError, "embedded null character");
+ return NULL;
+ }
+ return wstr;
+}
+
Py_ssize_t
PyUnicode_GetSize(PyObject *unicode)
return NULL;
if (PyUnicode_Check(data)) {
+ WCHAR *value = _PyUnicode_AsUnicode(data);
+ if (value == NULL) {
+ return NULL;
+ }
status = MsiSummaryInfoSetPropertyW(si->h, field, VT_LPSTR,
- 0, NULL, PyUnicode_AsUnicode(data));
+ 0, NULL, value);
} else if (PyLong_CheckExact(data)) {
long value = PyLong_AsLong(data);
if (value == -1 && PyErr_Occurred()) {
{
dl_funcptr p;
char funcname[258], *import_python;
- wchar_t *wpathname;
+ const wchar_t *wpathname;
#ifndef _DEBUG
_Py_CheckPython3();
#endif
- wpathname = PyUnicode_AsUnicode(pathname);
+ wpathname = _PyUnicode_AsUnicode(pathname);
if (wpathname == NULL)
return NULL;
#ifdef MS_WINDOWS
int err;
struct _stat wstatbuf;
- wchar_t *wpath;
+ const wchar_t *wpath;
- wpath = PyUnicode_AsUnicode(path);
+ wpath = _PyUnicode_AsUnicode(path);
if (wpath == NULL)
return -2;
+
err = _wstat(wpath, &wstatbuf);
if (!err)
statbuf->st_mode = wstatbuf.st_mode;
return err;
#else
int ret;
- PyObject *bytes = PyUnicode_EncodeFSDefault(path);
+ PyObject *bytes;
+ char *cpath;
+
+ bytes = PyUnicode_EncodeFSDefault(path);
if (bytes == NULL)
return -2;
- ret = stat(PyBytes_AS_STRING(bytes), statbuf);
+
+ /* check for embedded null bytes */
+ if (PyBytes_AsStringAndSize(bytes, &cpath, NULL) == -1) {
+ Py_DECREF(bytes);
+ return -2;
+ }
+
+ ret = stat(cpath, statbuf);
Py_DECREF(bytes);
return ret;
#endif
FILE *f;
int async_err = 0;
#ifdef MS_WINDOWS
- wchar_t *wpath;
+ const wchar_t *wpath;
wchar_t wmode[10];
int usize;
Py_TYPE(path));
return NULL;
}
- wpath = PyUnicode_AsUnicode(path);
+ wpath = _PyUnicode_AsUnicode(path);
if (wpath == NULL)
return NULL;