Fix issue #4730: cPickle corrupts high-unicode strings.

alexandre.vassalotti · alexandre.vassalotti · commit 292f72dfe439 · 2008-12-27T07:08:47.000Z
Update outdated copy of PyUnicode_EncodeRawUnicodeEscape. Add a test case. git-svn-id: http://svn.python.org/projects/python/trunk@67934 6015fed2-1504-0410-9fe1-9d1591cc4771
diff --git a/Lib/test/pickletester.py b/Lib/test/pickletester.py
@@ -480,14 +480,21 @@ def test_insecure_strings(self):
 
     if have_unicode:
         def test_unicode(self):
-            endcases = [unicode(''), unicode('<\\u>'), unicode('<\\\u1234>'),
-                        unicode('<\n>'),  unicode('<\\>')]
+            endcases = [u'', u'<\\u>', u'<\\\\u1234>', u'<\n>',
+                        u'<\\>', u'<\\\\U00012345>']
             for proto in protocols:
                 for u in endcases:
                     p = self.dumps(u, proto)
                     u2 = self.loads(p)
                     self.assertEqual(u2, u)
 
+        def test_unicode_high_plane(self):
+            t = u'\U00012345'
+            for proto in protocols:
+                p = self.dumps(t, proto)
+                t2 = self.loads(p)
+                self.assertEqual(t2, t)
+
     def test_ints(self):
         import sys
         for proto in protocols:
diff --git a/Modules/cPickle.c b/Modules/cPickle.c
@@ -1255,41 +1255,90 @@ save_string(Picklerobject *self, PyObject *args, int doput)
 /* A copy of PyUnicode_EncodeRawUnicodeEscape() that also translates
    backslash and newline characters to \uXXXX escapes. */
 static PyObject *
-modified_EncodeRawUnicodeEscape(const Py_UNICODE *s, int size)
+modified_EncodeRawUnicodeEscape(const Py_UNICODE *s, Py_ssize_t size)
 {
-	PyObject *repr;
-	char *p;
-	char *q;
+    PyObject *repr;
+    char *p;
+    char *q;
 
-	static const char *hexdigit = "0123456789ABCDEF";
+    static const char *hexdigit = "0123456789abcdef";
+#ifdef Py_UNICODE_WIDE
+    const Py_ssize_t expandsize = 10;
+#else
+    const Py_ssize_t expandsize = 6;
+#endif
 
-	repr = PyString_FromStringAndSize(NULL, 6 * size);
-	if (repr == NULL)
-		return NULL;
-	if (size == 0)
-		return repr;
-
-	p = q = PyString_AS_STRING(repr);
-	while (size-- > 0) {
-		Py_UNICODE ch = *s++;
-		/* Map 16-bit characters to '\uxxxx' */
-		if (ch >= 256 || ch == '\\' || ch == '\n') {
-			*p++ = '\\';
-			*p++ = 'u';
-			*p++ = hexdigit[(ch >> 12) & 0xf];
-			*p++ = hexdigit[(ch >> 8) & 0xf];
-			*p++ = hexdigit[(ch >> 4) & 0xf];
-			*p++ = hexdigit[ch & 15];
-		}
-		/* Copy everything else as-is */
-		else
-			*p++ = (char) ch;
-	}
-	*p = '\0';
-	_PyString_Resize(&repr, p - q);
+    if (size > PY_SSIZE_T_MAX / expandsize)
+        return PyErr_NoMemory();
+
+    repr = PyString_FromStringAndSize(NULL, expandsize * size);
+    if (repr == NULL)
+        return NULL;
+    if (size == 0)
 	return repr;
-}
 
+    p = q = PyString_AS_STRING(repr);
+    while (size-- > 0) {
+        Py_UNICODE ch = *s++;
+#ifdef Py_UNICODE_WIDE
+	/* Map 32-bit characters to '\Uxxxxxxxx' */
+	if (ch >= 0x10000) {
+            *p++ = '\\';
+            *p++ = 'U';
+            *p++ = hexdigit[(ch >> 28) & 0xf];
+            *p++ = hexdigit[(ch >> 24) & 0xf];
+            *p++ = hexdigit[(ch >> 20) & 0xf];
+            *p++ = hexdigit[(ch >> 16) & 0xf];
+            *p++ = hexdigit[(ch >> 12) & 0xf];
+            *p++ = hexdigit[(ch >> 8) & 0xf];
+            *p++ = hexdigit[(ch >> 4) & 0xf];
+            *p++ = hexdigit[ch & 15];
+        }
+        else
+#else
+	/* Map UTF-16 surrogate pairs to '\U00xxxxxx' */
+	if (ch >= 0xD800 && ch < 0xDC00) {
+	    Py_UNICODE ch2;
+	    Py_UCS4 ucs;
+
+	    ch2 = *s++;
+	    size--;
+	    if (ch2 >= 0xDC00 && ch2 <= 0xDFFF) {
+		ucs = (((ch & 0x03FF) << 10) | (ch2 & 0x03FF)) + 0x00010000;
+		*p++ = '\\';
+		*p++ = 'U';
+		*p++ = hexdigit[(ucs >> 28) & 0xf];
+		*p++ = hexdigit[(ucs >> 24) & 0xf];
+		*p++ = hexdigit[(ucs >> 20) & 0xf];
+		*p++ = hexdigit[(ucs >> 16) & 0xf];
+		*p++ = hexdigit[(ucs >> 12) & 0xf];
+		*p++ = hexdigit[(ucs >> 8) & 0xf];
+		*p++ = hexdigit[(ucs >> 4) & 0xf];
+		*p++ = hexdigit[ucs & 0xf];
+		continue;
+	    }
+	    /* Fall through: isolated surrogates are copied as-is */
+	    s--;
+	    size++;
+	}
+#endif
+	/* Map 16-bit characters to '\uxxxx' */
+	if (ch >= 256 || ch == '\\' || ch == '\n') {
+            *p++ = '\\';
+            *p++ = 'u';
+            *p++ = hexdigit[(ch >> 12) & 0xf];
+            *p++ = hexdigit[(ch >> 8) & 0xf];
+            *p++ = hexdigit[(ch >> 4) & 0xf];
+            *p++ = hexdigit[ch & 15];
+        }
+	/* Copy everything else as-is */
+	else
+            *p++ = (char) ch;
+    }
+    *p = '\0';
+    _PyString_Resize(&repr, p - q);
+    return repr;
+}
 
 static int
 save_unicode(Picklerobject *self, PyObject *args, int doput)
