diff --git a/Lib/sre_compile.py b/Lib/sre_compile.py index d6dc60c73ed..e08ec661aaa 100644 --- a/Lib/sre_compile.py +++ b/Lib/sre_compile.py @@ -339,7 +339,7 @@ def _optimize_unicode(charset, fixup): else: code = 'I' # Convert block indices to byte array of 256 bytes - mapping = array.array('b', mapping).tobytes() + mapping = array.array('B', mapping).tobytes() # 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 841d3a38c53..5e685850653 100644 --- a/Lib/test/test_re.py +++ b/Lib/test/test_re.py @@ -482,6 +482,9 @@ class ReTests(unittest.TestCase): "\u2222").group(1), "\u2222") self.assertEqual(re.match("([\u2222\u2223])", "\u2222", re.UNICODE).group(1), "\u2222") + r = '[%s]' % ''.join(map(chr, range(256, 2**16, 255))) + self.assertEqual(re.match(r, + "\uff01", re.UNICODE).group(), "\uff01") def test_big_codesize(self): # Issue #1160 diff --git a/Misc/NEWS b/Misc/NEWS index 9a3e08a856f..26e418b2a8d 100644 --- a/Misc/NEWS +++ b/Misc/NEWS @@ -19,6 +19,8 @@ Core and Builtins Library ------- +- Issue #19327: Fixed the working of regular expressions with too big charset. + - Issue #17400: New 'is_global' attribute for ipaddress to tell if an address is allocated by IANA for global or private networks. diff --git a/Modules/_sre.c b/Modules/_sre.c index 84330ef7152..dd9d6645d70 100644 --- a/Modules/_sre.c +++ b/Modules/_sre.c @@ -447,7 +447,7 @@ SRE_CHARSET(SRE_CODE* set, SRE_CODE ch) count = *(set++); if (sizeof(SRE_CODE) == 2) { - block = ((char*)set)[ch >> 8]; + block = ((unsigned char*)set)[ch >> 8]; set += 128; if (set[block*16 + ((ch & 255)>>4)] & (1 << (ch & 15))) return ok; @@ -457,7 +457,7 @@ SRE_CHARSET(SRE_CODE* set, SRE_CODE ch) /* !(c & ~N) == (c < N+1) for any unsigned c, this avoids * warnings when c's type supports only numbers < N+1 */ if (!(ch & ~65535)) - block = ((char*)set)[ch >> 8]; + block = ((unsigned char*)set)[ch >> 8]; else block = -1; set += 64;