Issue #23728: binascii.crc_hqx() could return an integer outside of the range

0-0xffff for empty data.

Lib/test/test_binascii.py

@@ -135,6 +135,18 @@ class BinASCIITest(unittest.TestCase):
         # Issue #7701 (crash on a pydebug build)
         self.assertEqual(binascii.b2a_uu(b'x'), b'!>   \n')
 
+    def test_crc_hqx(self):
+        crc = binascii.crc_hqx(self.type2test(b"Test the CRC-32 of"), 0)
+        crc = binascii.crc_hqx(self.type2test(b" this string."), crc)
+        self.assertEqual(crc, 14290)
+
+        self.assertRaises(TypeError, binascii.crc_hqx)
+        self.assertRaises(TypeError, binascii.crc_hqx, self.type2test(b''))
+
+        for crc in 0, 1, 0x1234, 0x12345, 0x12345678, -1:
+            self.assertEqual(binascii.crc_hqx(self.type2test(b''), crc),
+                             crc & 0xffff)
+
     def test_crc32(self):
         crc = binascii.crc32(self.type2test(b"Test the CRC-32 of"))
         crc = binascii.crc32(self.type2test(b" this string."), crc)

Misc/NEWS

@@ -37,6 +37,9 @@ Core and Builtins
 Library
 -------
 
+- Issue #23728: binascii.crc_hqx() could return an integer outside of the range
+  0-0xffff for empty data.
+
 - Issue #16914: new debuglevel 2 in smtplib adds timestamps to debug output.
 
 - Issue #7159: urllib.request now supports sending auth credentials

Modules/binascii.c

@@ -908,31 +908,31 @@ binascii_rledecode_hqx_impl(PyModuleDef *module, Py_buffer *data)
 
 
 /*[clinic input]
-binascii.crc_hqx -> int
+binascii.crc_hqx -> unsigned_int
 
     data: Py_buffer
-    crc: int
+    crc: unsigned_int(bitwise=True)
     /
 
 Compute hqx CRC incrementally.
 [clinic start generated code]*/
 
-static int
+static unsigned int
-binascii_crc_hqx_impl(PyModuleDef *module, Py_buffer *data, int crc)
+binascii_crc_hqx_impl(PyModuleDef *module, Py_buffer *data, unsigned int crc)
-/*[clinic end generated code: output=634dac18dfa863d7 input=68060931b2f51c8a]*/
+/*[clinic end generated code: output=167c2dac62625717 input=add8c53712ccceda]*/
 {
     unsigned char *bin_data;
-    unsigned int ucrc = (unsigned int)crc;
     Py_ssize_t len;
 
+    crc &= 0xffff;
     bin_data = data->buf;
     len = data->len;
 
     while(len-- > 0) {
-        ucrc=((ucrc<<8)&0xff00)^crctab_hqx[((ucrc>>8)&0xff)^*bin_data++];
+        crc = ((crc<<8)&0xff00) ^ crctab_hqx[(crc>>8)^*bin_data++];
     }
 
-    return (int)ucrc;
+    return crc;
 }
 
 #ifndef USE_ZLIB_CRC32

Modules/clinic/binascii.c.h

@@ -267,25 +267,25 @@ PyDoc_STRVAR(binascii_crc_hqx__doc__,
 #define BINASCII_CRC_HQX_METHODDEF    \
     {"crc_hqx", (PyCFunction)binascii_crc_hqx, METH_VARARGS, binascii_crc_hqx__doc__},
 
-static int
+static unsigned int
-binascii_crc_hqx_impl(PyModuleDef *module, Py_buffer *data, int crc);
+binascii_crc_hqx_impl(PyModuleDef *module, Py_buffer *data, unsigned int crc);
 
 static PyObject *
 binascii_crc_hqx(PyModuleDef *module, PyObject *args)
 {
     PyObject *return_value = NULL;
     Py_buffer data = {NULL, NULL};
-    int crc;
+    unsigned int crc;
-    int _return_value;
+    unsigned int _return_value;
 
     if (!PyArg_ParseTuple(args,
-        "y*i:crc_hqx",
+        "y*I:crc_hqx",
         &data, &crc))
         goto exit;
     _return_value = binascii_crc_hqx_impl(module, &data, crc);
-    if ((_return_value == -1) && PyErr_Occurred())
+    if ((_return_value == (unsigned int)-1) && PyErr_Occurred())
         goto exit;
-    return_value = PyLong_FromLong((long)_return_value);
+    return_value = PyLong_FromUnsignedLong((unsigned long)_return_value);
 
 exit:
     /* Cleanup for data */
@@ -544,4 +544,4 @@ exit:
 
     return return_value;
 }
-/*[clinic end generated code: output=175025a8a94fbdd1 input=a9049054013a1b77]*/
+/*[clinic end generated code: output=5f8d3578618b3432 input=a9049054013a1b77]*/