Make binascii use byte strings everywhere (in and out).

Lib/test/test_binascii.py

@@ -7,10 +7,10 @@ import binascii
 class BinASCIITest(unittest.TestCase):
 
     # Create binary test data
-    data = "The quick brown fox jumps over the lazy dog.\r\n"
+    data = b"The quick brown fox jumps over the lazy dog.\r\n"
     # Be slow so we don't depend on other modules
-    data += "".join(map(chr, range(256)))
-    data += "\r\nHello world.\n"
+    data += bytes(range(256))
+    data += b"\r\nHello world.\n"
 
     def test_exceptions(self):
         # Check module exceptions
@@ -40,10 +40,10 @@ class BinASCIITest(unittest.TestCase):
             b = self.data[i:i+MAX_BASE64]
             a = binascii.b2a_base64(b)
             lines.append(a)
-        res = ""
+        res = bytes()
         for line in lines:
             b = binascii.a2b_base64(line)
-            res = res + b
+            res += b
         self.assertEqual(res, self.data)
 
     def test_base64invalid(self):
@@ -56,24 +56,23 @@ class BinASCIITest(unittest.TestCase):
             a = binascii.b2a_base64(b)
             lines.append(a)
 
-        fillers = ""
-        valid = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789+/"
+        fillers = bytes()
+        valid = b"abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789+/"
         for i in range(256):
-            c = chr(i)
-            if c not in valid:
-                fillers += c
+            if i not in valid:
+                fillers.append(i)
         def addnoise(line):
             noise = fillers
             ratio = len(line) // len(noise)
-            res = ""
+            res = bytes()
             while line and noise:
                 if len(line) // len(noise) > ratio:
                     c, line = line[0], line[1:]
                 else:
                     c, noise = noise[0], noise[1:]
-                res += c
+                res.append(c)
             return res + noise + line
-        res = ""
+        res = bytes()
         for line in map(addnoise, lines):
             b = binascii.a2b_base64(line)
             res += b
@@ -81,7 +80,7 @@ class BinASCIITest(unittest.TestCase):
 
         # Test base64 with just invalid characters, which should return
         # empty strings. TBD: shouldn't it raise an exception instead ?
-        self.assertEqual(binascii.a2b_base64(fillers), '')
+        self.assertEqual(binascii.a2b_base64(fillers), b'')
 
     def test_uu(self):
         MAX_UU = 45
@@ -90,23 +89,23 @@ class BinASCIITest(unittest.TestCase):
             b = self.data[i:i+MAX_UU]
             a = binascii.b2a_uu(b)
             lines.append(a)
-        res = ""
+        res = bytes()
         for line in lines:
             b = binascii.a2b_uu(line)
             res += b
         self.assertEqual(res, self.data)
 
-        self.assertEqual(binascii.a2b_uu("\x7f"), "\x00"*31)
-        self.assertEqual(binascii.a2b_uu("\x80"), "\x00"*32)
-        self.assertEqual(binascii.a2b_uu("\xff"), "\x00"*31)
-        self.assertRaises(binascii.Error, binascii.a2b_uu, "\xff\x00")
-        self.assertRaises(binascii.Error, binascii.a2b_uu, "!!!!")
+        self.assertEqual(binascii.a2b_uu(b"\x7f"), b"\x00"*31)
+        self.assertEqual(binascii.a2b_uu(b"\x80"), b"\x00"*32)
+        self.assertEqual(binascii.a2b_uu(b"\xff"), b"\x00"*31)
+        self.assertRaises(binascii.Error, binascii.a2b_uu, b"\xff\x00")
+        self.assertRaises(binascii.Error, binascii.a2b_uu, b"!!!!")
 
-        self.assertRaises(binascii.Error, binascii.b2a_uu, 46*"!")
+        self.assertRaises(binascii.Error, binascii.b2a_uu, 46*b"!")
 
     def test_crc32(self):
-        crc = binascii.crc32("Test the CRC-32 of")
-        crc = binascii.crc32(" this string.", crc)
+        crc = binascii.crc32(b"Test the CRC-32 of")
+        crc = binascii.crc32(b" this string.", crc)
         self.assertEqual(crc, 1571220330)
 
         self.assertRaises(TypeError, binascii.crc32)
@@ -115,16 +114,16 @@ class BinASCIITest(unittest.TestCase):
 
     def test_hex(self):
         # test hexlification
-        s = '{s\005\000\000\000worldi\002\000\000\000s\005\000\000\000helloi\001\000\000\0000'
+        s = b'{s\005\000\000\000worldi\002\000\000\000s\005\000\000\000helloi\001\000\000\0000'
         t = binascii.b2a_hex(s)
         u = binascii.a2b_hex(t)
         self.assertEqual(s, u)
         self.assertRaises(TypeError, binascii.a2b_hex, t[:-1])
-        self.assertRaises(TypeError, binascii.a2b_hex, t[:-1] + 'q')
+        self.assertRaises(TypeError, binascii.a2b_hex, t[:-1] + b'q')
 
         # Verify the treatment of Unicode strings
         if test_support.have_unicode:
-            self.assertEqual(binascii.hexlify(str('a', 'ascii')), '61')
+            self.assertEqual(binascii.hexlify('a'), b'61')
 
     def test_qp(self):
         # A test for SF bug 534347 (segfaults without the proper fix)
@@ -134,28 +133,29 @@ class BinASCIITest(unittest.TestCase):
             pass
         else:
             self.fail("binascii.a2b_qp(**{1:1}) didn't raise TypeError")
-        self.assertEqual(binascii.a2b_qp("= "), "= ")
-        self.assertEqual(binascii.a2b_qp("=="), "=")
-        self.assertEqual(binascii.a2b_qp("=AX"), "=AX")
+        self.assertEqual(binascii.a2b_qp(b"= "), b"= ")
+        self.assertEqual(binascii.a2b_qp(b"=="), b"=")
+        self.assertEqual(binascii.a2b_qp(b"=AX"), b"=AX")
         self.assertRaises(TypeError, binascii.b2a_qp, foo="bar")
-        self.assertEqual(binascii.a2b_qp("=00\r\n=00"), "\x00\r\n\x00")
+        self.assertEqual(binascii.a2b_qp(b"=00\r\n=00"), b"\x00\r\n\x00")
         self.assertEqual(
-            binascii.b2a_qp("\xff\r\n\xff\n\xff"),
-            "=FF\r\n=FF\r\n=FF"
+            binascii.b2a_qp(b"\xff\r\n\xff\n\xff"),
+            b"=FF\r\n=FF\r\n=FF"
         )
         self.assertEqual(
-            binascii.b2a_qp("0"*75+"\xff\r\n\xff\r\n\xff"),
-            "0"*75+"=\r\n=FF\r\n=FF\r\n=FF"
+            binascii.b2a_qp(b"0"*75+b"\xff\r\n\xff\r\n\xff"),
+            b"0"*75+b"=\r\n=FF\r\n=FF\r\n=FF"
         )
 
-        self.assertEqual(binascii.b2a_qp('\0\n'), '=00\n')
-        self.assertEqual(binascii.b2a_qp('\0\n', quotetabs=True), '=00\n')
-        self.assertEqual(binascii.b2a_qp('foo\tbar\t\n'), 'foo\tbar=09\n')
-        self.assertEqual(binascii.b2a_qp('foo\tbar\t\n', quotetabs=True), 'foo=09bar=09\n')
+        self.assertEqual(binascii.b2a_qp(b'\0\n'), b'=00\n')
+        self.assertEqual(binascii.b2a_qp(b'\0\n', quotetabs=True), b'=00\n')
+        self.assertEqual(binascii.b2a_qp(b'foo\tbar\t\n'), b'foo\tbar=09\n')
+        self.assertEqual(binascii.b2a_qp(b'foo\tbar\t\n', quotetabs=True),
+                         b'foo=09bar=09\n')
 
-        self.assertEqual(binascii.b2a_qp('.'), '=2E')
-        self.assertEqual(binascii.b2a_qp('.\n'), '=2E\n')
-        self.assertEqual(binascii.b2a_qp('a.\n'), 'a.\n')
+        self.assertEqual(binascii.b2a_qp(b'.'), b'=2E')
+        self.assertEqual(binascii.b2a_qp(b'.\n'), b'=2E\n')
+        self.assertEqual(binascii.b2a_qp(b'a.\n'), b'a.\n')
 
     def test_empty_string(self):
         # A test for SF bug #1022953.  Make sure SystemError is not raised.
@@ -164,7 +164,10 @@ class BinASCIITest(unittest.TestCase):
                   'a2b_hqx', 'a2b_base64', 'rlecode_hqx', 'b2a_uu',
                   'rledecode_hqx']:
             f = getattr(binascii, n)
-            f('')
+            try:
+                f(b'')
+            except SystemError as err:
+                self.fail("%s(b'') raises SystemError: %s" % (n, err))
         binascii.crc_hqx('', 0)
 
 def test_main():

Modules/binascii.c

@@ -138,7 +138,7 @@ static char table_a2b_base64[] = {
 #define BASE64_PAD '='
 
 /* Max binary chunk size; limited only by available memory */
-#define BASE64_MAXBIN (INT_MAX/2 - sizeof(PyStringObject) - 3)
+#define BASE64_MAXBIN ((PY_SSIZE_T_MAX - 3) / 2)
 
 static unsigned char table_b2a_base64[] =
 "ABCDEFGHIJKLMNOPQRSTUVWXYZabcdefghijklmnopqrstuvwxyz0123456789+/";
@@ -200,9 +200,9 @@ binascii_a2b_uu(PyObject *self, PyObject *args)
 	ascii_len--;
 
 	/* Allocate the buffer */
-	if ( (rv=PyString_FromStringAndSize(NULL, bin_len)) == NULL )
+	if ( (rv=PyBytes_FromStringAndSize(NULL, bin_len)) == NULL )
 		return NULL;
-	bin_data = (unsigned char *)PyString_AsString(rv);
+	bin_data = (unsigned char *)PyBytes_AS_STRING(rv);
 
 	for( ; bin_len > 0 ; ascii_len--, ascii_data++ ) {
 		/* XXX is it really best to add NULs if there's no more data */
@@ -277,9 +277,9 @@ binascii_b2a_uu(PyObject *self, PyObject *args)
 	}
 
 	/* We're lazy and allocate to much (fixed up later) */
-	if ( (rv=PyString_FromStringAndSize(NULL, bin_len*2+2)) == NULL )
+	if ( (rv=PyBytes_FromStringAndSize(NULL, bin_len*2+2)) == NULL )
 		return NULL;
-	ascii_data = (unsigned char *)PyString_AsString(rv);
+	ascii_data = (unsigned char *)PyBytes_AS_STRING(rv);
 
 	/* Store the length */
 	*ascii_data++ = ' ' + (bin_len & 077);
@@ -301,8 +301,12 @@ binascii_b2a_uu(PyObject *self, PyObject *args)
 	}
 	*ascii_data++ = '\n';	/* Append a courtesy newline */
 
-	_PyString_Resize(&rv, (ascii_data -
-			       (unsigned char *)PyString_AsString(rv)));
+	if (PyBytes_Resize(rv,
+                           (ascii_data -
+                            (unsigned char *)PyBytes_AS_STRING(rv))) < 0) {
+		Py_DECREF(rv);
+		rv = NULL;
+	}
 	return rv;
 }
 
@@ -351,9 +355,9 @@ binascii_a2b_base64(PyObject *self, PyObject *args)
 	bin_len = ((ascii_len+3)/4)*3; /* Upper bound, corrected later */
 
 	/* Allocate the buffer */
-	if ( (rv=PyString_FromStringAndSize(NULL, bin_len)) == NULL )
+	if ( (rv=PyBytes_FromStringAndSize(NULL, bin_len)) == NULL )
 		return NULL;
-	bin_data = (unsigned char *)PyString_AsString(rv);
+	bin_data = (unsigned char *)PyBytes_AS_STRING(rv);
 	bin_len = 0;
 
 	for( ; ascii_len > 0; ascii_len--, ascii_data++) {
@@ -412,13 +416,17 @@ binascii_a2b_base64(PyObject *self, PyObject *args)
 
 	/* And set string size correctly. If the result string is empty
 	** (because the input was all invalid) return the shared empty
-	** string instead; _PyString_Resize() won't do this for us.
+	** string instead; PyBytes_Resize() won't do this for us.
 	*/
-	if (bin_len > 0)
-		_PyString_Resize(&rv, bin_len);
+	if (bin_len > 0) {
+		if (PyBytes_Resize(rv, bin_len) < 0) {
+			Py_DECREF(rv);
+			rv = NULL;
+		}
+	}
 	else {
 		Py_DECREF(rv);
-		rv = PyString_FromString("");
+		rv = PyBytes_FromStringAndSize("", 0);
 	}
 	return rv;
 }
@@ -445,9 +453,9 @@ binascii_b2a_base64(PyObject *self, PyObject *args)
 	/* We're lazy and allocate too much (fixed up later).
 	   "+3" leaves room for up to two pad characters and a trailing
 	   newline.  Note that 'b' gets encoded as 'Yg==\n' (1 in, 5 out). */
-	if ( (rv=PyString_FromStringAndSize(NULL, bin_len*2 + 3)) == NULL )
+	if ( (rv=PyBytes_FromStringAndSize(NULL, bin_len*2 + 3)) == NULL )
 		return NULL;
-	ascii_data = (unsigned char *)PyString_AsString(rv);
+	ascii_data = (unsigned char *)PyBytes_AS_STRING(rv);
 
 	for( ; bin_len > 0 ; bin_len--, bin_data++ ) {
 		/* Shift the data into our buffer */
@@ -471,8 +479,12 @@ binascii_b2a_base64(PyObject *self, PyObject *args)
 	}
 	*ascii_data++ = '\n';	/* Append a courtesy newline */
 
-	_PyString_Resize(&rv, (ascii_data -
-			       (unsigned char *)PyString_AsString(rv)));
+	if (PyBytes_Resize(rv,
+			   (ascii_data -
+			    (unsigned char *)PyBytes_AS_STRING(rv))) < 0) {
+		Py_DECREF(rv);
+		rv = NULL;
+	}
 	return rv;
 }
 
@@ -495,9 +507,9 @@ binascii_a2b_hqx(PyObject *self, PyObject *args)
 	/* Allocate a string that is too big (fixed later) 
 	   Add two to the initial length to prevent interning which
 	   would preclude subsequent resizing.  */
-	if ( (rv=PyString_FromStringAndSize(NULL, len+2)) == NULL )
+	if ( (rv=PyBytes_FromStringAndSize(NULL, len+2)) == NULL )
 		return NULL;
-	bin_data = (unsigned char *)PyString_AsString(rv);
+	bin_data = (unsigned char *)PyBytes_AS_STRING(rv);
 
 	for( ; len > 0 ; len--, ascii_data++ ) {
 		/* Get the byte and look it up */
@@ -531,8 +543,12 @@ binascii_a2b_hqx(PyObject *self, PyObject *args)
 		Py_DECREF(rv);
 		return NULL;
 	}
-	_PyString_Resize(
-		&rv, (bin_data - (unsigned char *)PyString_AsString(rv)));
+	if (PyBytes_Resize(rv,
+			   (bin_data -
+			    (unsigned char *)PyBytes_AS_STRING(rv))) < 0) {
+		Py_DECREF(rv);
+		rv = NULL;
+	}
 	if (rv) {
 		PyObject *rrv = Py_BuildValue("Oi", rv, done);
 		Py_DECREF(rv);
@@ -556,9 +572,9 @@ binascii_rlecode_hqx(PyObject *self, PyObject *args)
 		return NULL;
 
 	/* Worst case: output is twice as big as input (fixed later) */
-	if ( (rv=PyString_FromStringAndSize(NULL, len*2+2)) == NULL )
+	if ( (rv=PyBytes_FromStringAndSize(NULL, len*2+2)) == NULL )
 		return NULL;
-	out_data = (unsigned char *)PyString_AsString(rv);
+	out_data = (unsigned char *)PyBytes_AS_STRING(rv);
 
 	for( in=0; in<len; in++) {
 		ch = in_data[in];
@@ -584,8 +600,12 @@ binascii_rlecode_hqx(PyObject *self, PyObject *args)
 			}
 		}
 	}
-	_PyString_Resize(&rv, (out_data -
-			       (unsigned char *)PyString_AsString(rv)));
+	if (PyBytes_Resize(rv,
+			   (out_data -
+			    (unsigned char *)PyBytes_AS_STRING(rv))) < 0) {
+		Py_DECREF(rv);
+		rv = NULL;
+	}
 	return rv;
 }
 
@@ -605,9 +625,9 @@ binascii_b2a_hqx(PyObject *self, PyObject *args)
 		return NULL;
 
 	/* Allocate a buffer that is at least large enough */
-	if ( (rv=PyString_FromStringAndSize(NULL, len*2+2)) == NULL )
+	if ( (rv=PyBytes_FromStringAndSize(NULL, len*2+2)) == NULL )
 		return NULL;
-	ascii_data = (unsigned char *)PyString_AsString(rv);
+	ascii_data = (unsigned char *)PyBytes_AS_STRING(rv);
 
 	for( ; len > 0 ; len--, bin_data++ ) {
 		/* Shift into our buffer, and output any 6bits ready */
@@ -624,8 +644,12 @@ binascii_b2a_hqx(PyObject *self, PyObject *args)
 		leftchar <<= (6-leftbits);
 		*ascii_data++ = table_b2a_hqx[leftchar & 0x3f];
 	}
-	_PyString_Resize(&rv, (ascii_data -
-			       (unsigned char *)PyString_AsString(rv)));
+	if (PyBytes_Resize(rv,
+			   (ascii_data -
+			    (unsigned char *)PyBytes_AS_STRING(rv))) < 0) {
+		Py_DECREF(rv);
+		rv = NULL;
+	}
 	return rv;
 }
 
@@ -644,14 +668,14 @@ binascii_rledecode_hqx(PyObject *self, PyObject *args)
 
 	/* Empty string is a special case */
 	if ( in_len == 0 )
-		return PyString_FromString("");
+		return PyBytes_FromStringAndSize("", 0);
 
 	/* Allocate a buffer of reasonable size. Resized when needed */
 	out_len = in_len*2;
-	if ( (rv=PyString_FromStringAndSize(NULL, out_len)) == NULL )
+	if ( (rv=PyBytes_FromStringAndSize(NULL, out_len)) == NULL )
 		return NULL;
 	out_len_left = out_len;
-	out_data = (unsigned char *)PyString_AsString(rv);
+	out_data = (unsigned char *)PyBytes_AS_STRING(rv);
 
 	/*
 	** We need two macros here to get/put bytes and handle
@@ -670,9 +694,9 @@ binascii_rledecode_hqx(PyObject *self, PyObject *args)
 #define OUTBYTE(b) \
 	do { \
 		 if ( --out_len_left < 0 ) { \
-			  _PyString_Resize(&rv, 2*out_len); \
-			  if ( rv == NULL ) return NULL; \
-			  out_data = (unsigned char *)PyString_AsString(rv) \
+			  if (PyBytes_Resize(rv, 2*out_len) < 0) \
+			    { Py_DECREF(rv); return NULL; } \
+			  out_data = (unsigned char *)PyBytes_AS_STRING(rv) \
 								 + out_len; \
 			  out_len_left = out_len-1; \
 			  out_len = out_len * 2; \
@@ -720,8 +744,12 @@ binascii_rledecode_hqx(PyObject *self, PyObject *args)
 			OUTBYTE(in_byte);
 		}
 	}
-	_PyString_Resize(&rv, (out_data -
-			       (unsigned char *)PyString_AsString(rv)));
+	if (PyBytes_Resize(rv,
+			   (out_data -
+			    (unsigned char *)PyBytes_AS_STRING(rv))) < 0) {
+		Py_DECREF(rv);
+		rv = NULL;
+	}
 	return rv;
 }
 
@@ -912,12 +940,10 @@ binascii_hexlify(PyObject *self, PyObject *args)
 	if (!PyArg_ParseTuple(args, "s#:b2a_hex", &argbuf, &arglen))
 		return NULL;
 
-	retval = PyString_FromStringAndSize(NULL, arglen*2);
+	retval = PyBytes_FromStringAndSize(NULL, arglen*2);
 	if (!retval)
 		return NULL;
-	retbuf = PyString_AsString(retval);
-	if (!retbuf)
-		goto finally;
+	retbuf = PyBytes_AS_STRING(retval);
 
 	/* make hex version of string, taken from shamodule.c */
 	for (i=j=0; i < arglen; i++) {
@@ -978,12 +1004,10 @@ binascii_unhexlify(PyObject *self, PyObject *args)
 		return NULL;
 	}
 
-	retval = PyString_FromStringAndSize(NULL, (arglen/2));
+	retval = PyBytes_FromStringAndSize(NULL, (arglen/2));
 	if (!retval)
 		return NULL;
-	retbuf = PyString_AsString(retval);
-	if (!retbuf)
-		goto finally;
+	retbuf = PyBytes_AS_STRING(retval);
 
 	for (i=j=0; i < arglen; i += 2) {
 		int top = to_int(Py_CHARMASK(argbuf[i]));
@@ -1095,7 +1119,7 @@ binascii_a2b_qp(PyObject *self, PyObject *args, PyObject *kwargs)
 			out++;
 		}
 	}
-	if ((rv = PyString_FromStringAndSize((char *)odata, out)) == NULL) {
+	if ((rv = PyBytes_FromStringAndSize((char *)odata, out)) == NULL) {
 		PyMem_Free(odata);
 		return NULL;
 	}
@@ -1295,7 +1319,7 @@ binascii_b2a_qp (PyObject *self, PyObject *args, PyObject *kwargs)
 			}
 		}
 	}
-	if ((rv = PyString_FromStringAndSize((char *)odata, out)) == NULL) {
+	if ((rv = PyBytes_FromStringAndSize((char *)odata, out)) == NULL) {
 		PyMem_Free(odata);
 		return NULL;
 	}