Issue #8592: PyArg_Parse*() functions raise a TypeError for "y", "u" and "Z"

formats if the string contains a null byte/character. Write unit tests for
string formats.

5 files changed, 317 insertions, 5 deletions

diff --git a/Doc/c-api/arg.rst b/Doc/c-api/arg.rst
index ea897a8..1c49c8b 100644
--- a/Doc/c-api/arg.rst
+++ b/Doc/c-api/arg.rst
@@ -125,6 +125,8 @@ Unless otherwise stated, buffers are not NUL-terminated.
    pointer variable, which will be filled with the pointer to an existing
    Unicode buffer.  Please note that the width of a :ctype:`Py_UNICODE`
    character depends on compilation options (it is either 16 or 32 bits).
+   The Python string must not contain embedded NUL characters; if it does,
+   a :exc:`TypeError` exception is raised.
 
    .. note::
       Since ``u`` doesn't give you back the length of the string, and it
diff --git a/Lib/test/test_getargs2.py b/Lib/test/test_getargs2.py
index baf70d1..7186f55 100644
--- a/Lib/test/test_getargs2.py
+++ b/Lib/test/test_getargs2.py
@@ -293,8 +293,136 @@ class Keywords_TestCase(unittest.TestCase):
         else:
             self.fail('TypeError should have been raised')
 
+class Bytes_TestCase(unittest.TestCase):
+    def test_s(self):
+        from _testcapi import getargs_s
+        self.assertEqual(getargs_s('abc\xe9'), b'abc\xc3\xa9')
+        self.assertRaises(TypeError, getargs_s, 'nul:\0')
+        self.assertRaises(TypeError, getargs_s, b'bytes')
+        self.assertRaises(TypeError, getargs_s, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_s, memoryview(b'memoryview'))
+        self.assertRaises(TypeError, getargs_s, None)
+
+    def test_s_star(self):
+        from _testcapi import getargs_s_star
+        self.assertEqual(getargs_s_star('abc\xe9'), b'abc\xc3\xa9')
+        self.assertEqual(getargs_s_star('nul:\0'), b'nul:\0')
+        self.assertEqual(getargs_s_star(b'bytes'), b'bytes')
+        self.assertEqual(getargs_s_star(bytearray(b'bytearray')), b'bytearray')
+        self.assertEqual(getargs_s_star(memoryview(b'memoryview')), b'memoryview')
+        self.assertRaises(TypeError, getargs_s_star, None)
+
+    def test_s_hash(self):
+        from _testcapi import getargs_s_hash
+        self.assertEqual(getargs_s_hash('abc\xe9'), b'abc\xc3\xa9')
+        self.assertEqual(getargs_s_hash('nul:\0'), b'nul:\0')
+        self.assertEqual(getargs_s_hash(b'bytes'), b'bytes')
+        self.assertRaises(TypeError, getargs_s_hash, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_s_hash, memoryview(b'memoryview'))
+        self.assertRaises(TypeError, getargs_s_hash, None)
+
+    def test_z(self):
+        from _testcapi import getargs_z
+        self.assertEqual(getargs_z('abc\xe9'), b'abc\xc3\xa9')
+        self.assertRaises(TypeError, getargs_z, 'nul:\0')
+        self.assertEqual(getargs_z(b'bytes'), b'bytes')
+        self.assertRaises(TypeError, getargs_z, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_z, memoryview(b'memoryview'))
+        self.assertIsNone(getargs_z(None))
+
+    def test_z_star(self):
+        from _testcapi import getargs_z_star
+        self.assertEqual(getargs_z_star('abc\xe9'), b'abc\xc3\xa9')
+        self.assertEqual(getargs_z_star('nul:\0'), b'nul:\0')
+        self.assertEqual(getargs_z_star(b'bytes'), b'bytes')
+        self.assertEqual(getargs_z_star(bytearray(b'bytearray')), b'bytearray')
+        self.assertEqual(getargs_z_star(memoryview(b'memoryview')), b'memoryview')
+        self.assertIsNone(getargs_z_star(None))
+
+    def test_z_hash(self):
+        from _testcapi import getargs_z_hash
+        self.assertEqual(getargs_z_hash('abc\xe9'), b'abc\xc3\xa9')
+        self.assertEqual(getargs_z_hash('nul:\0'), b'nul:\0')
+        self.assertEqual(getargs_z_hash(b'bytes'), b'bytes')
+        self.assertRaises(TypeError, getargs_z_hash, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_z_hash, memoryview(b'memoryview'))
+        self.assertIsNone(getargs_z_hash(None))
+
+    def test_y(self):
+        from _testcapi import getargs_y
+        self.assertRaises(TypeError, getargs_y, 'abc\xe9')
+        self.assertEqual(getargs_y(b'bytes'), b'bytes')
+        self.assertRaises(TypeError, getargs_y, b'nul:\0')
+        self.assertRaises(TypeError, getargs_y, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_y, memoryview(b'memoryview'))
+        self.assertRaises(TypeError, getargs_y, None)
+
+    def test_y_star(self):
+        from _testcapi import getargs_y_star
+        self.assertRaises(TypeError, getargs_y_star, 'abc\xe9')
+        self.assertEqual(getargs_y_star(b'bytes'), b'bytes')
+        self.assertEqual(getargs_y_star(b'nul:\0'), b'nul:\0')
+        self.assertEqual(getargs_y_star(bytearray(b'bytearray')), b'bytearray')
+        self.assertEqual(getargs_y_star(memoryview(b'memoryview')), b'memoryview')
+        self.assertRaises(TypeError, getargs_y_star, None)
+
+    def test_y_hash(self):
+        from _testcapi import getargs_y_hash
+        self.assertRaises(TypeError, getargs_y_hash, 'abc\xe9')
+        self.assertEqual(getargs_y_hash(b'bytes'), b'bytes')
+        self.assertEqual(getargs_y_hash(b'nul:\0'), b'nul:\0')
+        self.assertRaises(TypeError, getargs_y_hash, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_y_hash, memoryview(b'memoryview'))
+        self.assertRaises(TypeError, getargs_y_hash, None)
+
+
+class Unicode_TestCase(unittest.TestCase):
+    def test_u(self):
+        from _testcapi import getargs_u
+        self.assertEqual(getargs_u('abc\xe9'), 'abc\xe9')
+        self.assertRaises(TypeError, getargs_u, 'nul:\0')
+        self.assertRaises(TypeError, getargs_u, b'bytes')
+        self.assertRaises(TypeError, getargs_u, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_u, memoryview(b'memoryview'))
+        self.assertRaises(TypeError, getargs_u, None)
+
+    def test_u_hash(self):
+        from _testcapi import getargs_u_hash
+        self.assertEqual(getargs_u_hash('abc\xe9'), 'abc\xe9')
+        self.assertEqual(getargs_u_hash('nul:\0'), 'nul:\0')
+        self.assertRaises(TypeError, getargs_u_hash, b'bytes')
+        self.assertRaises(TypeError, getargs_u_hash, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_u_hash, memoryview(b'memoryview'))
+        self.assertRaises(TypeError, getargs_u_hash, None)
+
+    def test_Z(self):
+        from _testcapi import getargs_Z
+        self.assertEqual(getargs_Z('abc\xe9'), 'abc\xe9')
+        self.assertRaises(TypeError, getargs_Z, 'nul:\0')
+        self.assertRaises(TypeError, getargs_Z, b'bytes')
+        self.assertRaises(TypeError, getargs_Z, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_Z, memoryview(b'memoryview'))
+        self.assertIsNone(getargs_Z(None))
+
+    def test_Z_hash(self):
+        from _testcapi import getargs_Z_hash
+        self.assertEqual(getargs_Z_hash('abc\xe9'), 'abc\xe9')
+        self.assertEqual(getargs_Z_hash('nul:\0'), 'nul:\0')
+        self.assertRaises(TypeError, getargs_Z_hash, b'bytes')
+        self.assertRaises(TypeError, getargs_Z_hash, bytearray(b'bytearray'))
+        self.assertRaises(TypeError, getargs_Z_hash, memoryview(b'memoryview'))
+        self.assertIsNone(getargs_Z_hash(None))
+
+
 def test_main():
-    tests = [Signed_TestCase, Unsigned_TestCase, Tuple_TestCase, Keywords_TestCase]
+    tests = [
+        Signed_TestCase,
+        Unsigned_TestCase,
+        Tuple_TestCase,
+        Keywords_TestCase,
+        Bytes_TestCase,
+        Unicode_TestCase,
+    ]
     try:
         from _testcapi import getargs_L, getargs_K
     except ImportError:
diff --git a/Misc/NEWS b/Misc/NEWS
index 65ab474..6855696 100644
--- a/Misc/NEWS
+++ b/Misc/NEWS
@@ -12,6 +12,10 @@ What's New in Python 3.2 Alpha 1?
 Core and Builtins
 -----------------
 
+- Issue #8592: PyArg_Parse*() functions raise a TypeError for "y", "u" and "Z"
+  formats if the string contains a null byte/character. Write unit tests for
+  string formats.
+
 - Issue #7490: to facilitate sharing of doctests between 2.x and 3.x test
   suites, the IGNORE_EXCEPTION_DETAIL directive now also ignores the module
   location of the raised exception.
diff --git a/Modules/_testcapimodule.c b/Modules/_testcapimodule.c
index e1dc5f4..7bcc1d8 100644
--- a/Modules/_testcapimodule.c
+++ b/Modules/_testcapimodule.c
@@ -1011,6 +1011,157 @@ test_k_code(PyObject *self)
     return Py_None;
 }
 
+static PyObject *
+getargs_s(PyObject *self, PyObject *args)
+{
+    char *str;
+    if (!PyArg_ParseTuple(args, "s", &str))
+        return NULL;
+    return PyBytes_FromString(str);
+}
+
+static PyObject *
+getargs_s_star(PyObject *self, PyObject *args)
+{
+    Py_buffer buffer;
+    PyObject *bytes;
+    if (!PyArg_ParseTuple(args, "s*", &buffer))
+        return NULL;
+    bytes = PyBytes_FromStringAndSize(buffer.buf, buffer.len);
+    PyBuffer_Release(&buffer);
+    return bytes;
+}
+
+static PyObject *
+getargs_s_hash(PyObject *self, PyObject *args)
+{
+    char *str;
+    Py_ssize_t size;
+    if (!PyArg_ParseTuple(args, "s#", &str, &size))
+        return NULL;
+    return PyBytes_FromStringAndSize(str, size);
+}
+
+static PyObject *
+getargs_z(PyObject *self, PyObject *args)
+{
+    char *str;
+    if (!PyArg_ParseTuple(args, "z", &str))
+        return NULL;
+    if (str != NULL)
+        return PyBytes_FromString(str);
+    else
+        Py_RETURN_NONE;
+}
+
+static PyObject *
+getargs_z_star(PyObject *self, PyObject *args)
+{
+    Py_buffer buffer;
+    PyObject *bytes;
+    if (!PyArg_ParseTuple(args, "z*", &buffer))
+        return NULL;
+    if (buffer.buf != NULL)
+        bytes = PyBytes_FromStringAndSize(buffer.buf, buffer.len);
+    else {
+        Py_INCREF(Py_None);
+        bytes = Py_None;
+    }
+    PyBuffer_Release(&buffer);
+    return bytes;
+}
+
+static PyObject *
+getargs_z_hash(PyObject *self, PyObject *args)
+{
+    char *str;
+    Py_ssize_t size;
+    if (!PyArg_ParseTuple(args, "z#", &str, &size))
+        return NULL;
+    if (str != NULL)
+        return PyBytes_FromStringAndSize(str, size);
+    else
+        Py_RETURN_NONE;
+}
+
+static PyObject *
+getargs_y(PyObject *self, PyObject *args)
+{
+    char *str;
+    if (!PyArg_ParseTuple(args, "y", &str))
+        return NULL;
+    return PyBytes_FromString(str);
+}
+
+static PyObject *
+getargs_y_star(PyObject *self, PyObject *args)
+{
+    Py_buffer buffer;
+    PyObject *bytes;
+    if (!PyArg_ParseTuple(args, "y*", &buffer))
+        return NULL;
+    bytes = PyBytes_FromStringAndSize(buffer.buf, buffer.len);
+    PyBuffer_Release(&buffer);
+    return bytes;
+}
+
+static PyObject *
+getargs_y_hash(PyObject *self, PyObject *args)
+{
+    char *str;
+    Py_ssize_t size;
+    if (!PyArg_ParseTuple(args, "y#", &str, &size))
+        return NULL;
+    return PyBytes_FromStringAndSize(str, size);
+}
+
+static PyObject *
+getargs_u(PyObject *self, PyObject *args)
+{
+    Py_UNICODE *str;
+    Py_ssize_t size;
+    if (!PyArg_ParseTuple(args, "u", &str))
+        return NULL;
+    size = Py_UNICODE_strlen(str);
+    return PyUnicode_FromUnicode(str, size);
+}
+
+static PyObject *
+getargs_u_hash(PyObject *self, PyObject *args)
+{
+    Py_UNICODE *str;
+    Py_ssize_t size;
+    if (!PyArg_ParseTuple(args, "u#", &str, &size))
+        return NULL;
+    return PyUnicode_FromUnicode(str, size);
+}
+
+static PyObject *
+getargs_Z(PyObject *self, PyObject *args)
+{
+    Py_UNICODE *str;
+    Py_ssize_t size;
+    if (!PyArg_ParseTuple(args, "Z", &str))
+        return NULL;
+    if (str != NULL) {
+        size = Py_UNICODE_strlen(str);
+        return PyUnicode_FromUnicode(str, size);
+    } else
+        Py_RETURN_NONE;
+}
+
+static PyObject *
+getargs_Z_hash(PyObject *self, PyObject *args)
+{
+    Py_UNICODE *str;
+    Py_ssize_t size;
+    if (!PyArg_ParseTuple(args, "Z#", &str, &size))
+        return NULL;
+    if (str != NULL)
+        return PyUnicode_FromUnicode(str, size);
+    else
+        Py_RETURN_NONE;
+}
 
 /* Test the s and z codes for PyArg_ParseTuple.
 */
@@ -2062,11 +2213,24 @@ static PyMethodDef TestMethods[] = {
     {"test_long_long_and_overflow",
         (PyCFunction)test_long_long_and_overflow, METH_NOARGS},
     {"test_L_code",             (PyCFunction)test_L_code,        METH_NOARGS},
+#endif
+    {"getargs_s",               getargs_s,                       METH_VARARGS},
+    {"getargs_s_star",          getargs_s_star,                  METH_VARARGS},
+    {"getargs_s_hash",          getargs_s_hash,                  METH_VARARGS},
+    {"getargs_z",               getargs_z,                       METH_VARARGS},
+    {"getargs_z_star",          getargs_z_star,                  METH_VARARGS},
+    {"getargs_z_hash",          getargs_z_hash,                  METH_VARARGS},
+    {"getargs_y",               getargs_y,                       METH_VARARGS},
+    {"getargs_y_star",          getargs_y_star,                  METH_VARARGS},
+    {"getargs_y_hash",          getargs_y_hash,                  METH_VARARGS},
+    {"getargs_u",               getargs_u,                       METH_VARARGS},
+    {"getargs_u_hash",          getargs_u_hash,                  METH_VARARGS},
+    {"getargs_Z",               getargs_Z,                       METH_VARARGS},
+    {"getargs_Z_hash",          getargs_Z_hash,                  METH_VARARGS},
     {"codec_incrementalencoder",
      (PyCFunction)codec_incrementalencoder,      METH_VARARGS},
     {"codec_incrementaldecoder",
      (PyCFunction)codec_incrementaldecoder,      METH_VARARGS},
-#endif
     {"test_s_code",             (PyCFunction)test_s_code,        METH_NOARGS},
     {"test_u_code",             (PyCFunction)test_u_code,        METH_NOARGS},
     {"test_Z_code",             (PyCFunction)test_Z_code,        METH_NOARGS},
diff --git a/Python/getargs.c b/Python/getargs.c
index 127b147..d4d8d84 100644
--- a/Python/getargs.c
+++ b/Python/getargs.c
@@ -935,10 +935,15 @@ convertsimple(PyObject *arg, const char **p_format, va_list *p_va, int flags,
         count = convertbuffer(arg, p, &buf);
         if (count < 0)
             return converterr(buf, arg, msgbuf, bufsize);
-        else if (*format == '#') {
+        if (*format == '#') {
             FETCH_SIZE;
             STORE_SIZE(count);
             format++;
+        } else {
+            if (strlen(*p) != count)
+                return converterr(
+                    "bytes without null bytes",
+                    arg, msgbuf, bufsize);
         }
         break;
     }
@@ -1045,9 +1050,13 @@ convertsimple(PyObject *arg, const char **p_format, va_list *p_va, int flags,
 
             if (arg == Py_None)
                 *p = 0;
-            else if (PyUnicode_Check(arg))
+            else if (PyUnicode_Check(arg)) {
                 *p = PyUnicode_AS_UNICODE(arg);
-            else
+                if (Py_UNICODE_strlen(*p) != PyUnicode_GET_SIZE(arg))
+                    return converterr(
+                        "str without null character or None",
+                        arg, msgbuf, bufsize);
+            } else
                 return converterr("str or None", arg, msgbuf, bufsize);
         }
         break;
@@ -1227,6 +1236,11 @@ convertsimple(PyObject *arg, const char **p_format, va_list *p_va, int flags,
             FETCH_SIZE;
             STORE_SIZE(PyUnicode_GET_SIZE(arg));
             format++;
+        } else {
+            if (Py_UNICODE_strlen(*p) != PyUnicode_GET_SIZE(arg))
+                return converterr(
+                    "str without null character",
+                    arg, msgbuf, bufsize);
         }
         break;
     }