Patch 1494554: Update numeric properties to Unicode 4.1.

Doc/api/concrete.tex

@@ -965,7 +965,7 @@ These APIs can be used for fast direct character conversions:
 \end{cfuncdesc}
 
 \begin{cfuncdesc}{double}{Py_UNICODE_TONUMERIC}{Py_UNICODE ch}
-  Return the character \var{ch} converted to a (positive) double.
+  Return the character \var{ch} converted to a double.
   Return \code{-1.0} if this is not possible.  This macro does not raise
   exceptions.
 \end{cfuncdesc}

Lib/test/test_unicodedata.py

@@ -16,7 +16,7 @@ encoding = 'utf-8'
 class UnicodeMethodsTest(unittest.TestCase):
 
     # update this, if the database changes
-    expectedchecksum = 'a6555cd209d960dcfa17bfdce0c96d91cfa9a9ba'
+    expectedchecksum = 'c198ed264497f108434b3f576d4107237221cc8a'
 
     def test_method_checksum(self):
         h = hashlib.sha1()
@@ -75,7 +75,7 @@ class UnicodeDatabaseTest(unittest.TestCase):
 class UnicodeFunctionsTest(UnicodeDatabaseTest):
 
     # update this, if the database changes
-    expectedchecksum = 'b45b79f3203ee1a896d9b5655484adaff5d4964b'
+    expectedchecksum = '4e389f97e9f88b8b7ab743121fd643089116f9f2'
 
     def test_function_checksum(self):
         data = []

Misc/NEWS

@@ -12,6 +12,9 @@ What's New in Python 2.5 alpha 3?
 Core and builtins
 -----------------
 
+- Patch #1494554: Update unicodedata.numeric and unicode.isnumeric to
+  Unicode 4.1.
+
 - Patch #921466: sys.path_importer_cache is now used to cache valid and
   invalid file paths for the built-in import machinery which leads to
   fewer open calls on startup.

Modules/unicodedata.c

@@ -209,7 +209,7 @@ unicodedata_numeric(PyObject *self, PyObject *args)
         if (old->category_changed == 0) {
             /* unassigned */
             have_old = 1;
-            rc = -1;
+            rc = -1.0;
         } 
         else if (old->decimal_changed != 0xFF) {
             have_old = 1;
@@ -219,7 +219,7 @@ unicodedata_numeric(PyObject *self, PyObject *args)
 
     if (!have_old)
         rc = Py_UNICODE_TONUMERIC(*PyUnicode_AS_UNICODE(v));
-    if (rc < 0) {
+    if (rc == -1.0) {
 	if (defobj == NULL) {
 	    PyErr_SetString(PyExc_ValueError, "not a numeric character");
 	    return NULL;

Objects/unicodeobject.c

@@ -2376,6 +2376,7 @@ PyObject *_PyUnicode_DecodeUnicodeInternal(const char *s,
     end = s + size;
 
     while (s < end) {
+	*p = *(Py_UNICODE*)s;
         memcpy(p, s, sizeof(Py_UNICODE));
         /* We have to sanity check the raw data, otherwise doom looms for
            some malformed UCS-4 data. */