Issue #19327: Fixed the working of regular expressions with too big charset.
diff --git a/Lib/sre_compile.py b/Lib/sre_compile.py
index 97c1663..bd40705 100644
--- a/Lib/sre_compile.py
+++ b/Lib/sre_compile.py
@@ -343,7 +343,7 @@
else:
code = 'I'
# Convert block indices to byte array of 256 bytes
- mapping = array.array('b', mapping).tostring()
+ mapping = array.array('B', mapping).tostring()
# Convert byte array to word array
mapping = array.array(code, mapping)
assert mapping.itemsize == _sre.CODESIZE
diff --git a/Lib/test/test_re.py b/Lib/test/test_re.py
index 8b277cf..d879bac 100644
--- a/Lib/test/test_re.py
+++ b/Lib/test/test_re.py
@@ -427,6 +427,8 @@
u"\u2222").group(1), u"\u2222")
self.assertEqual(re.match(u"([\u2222\u2223])",
u"\u2222", re.UNICODE).group(1), u"\u2222")
+ r = u'[%s]' % u''.join(map(unichr, range(256, 2**16, 255)))
+ self.assertEqual(re.match(r, u"\uff01", re.UNICODE).group(), u"\uff01")
def test_big_codesize(self):
# Issue #1160