mirror of https://github.com/python/cpython
728 lines
33 KiB
Python
728 lines
33 KiB
Python
import sys
|
|
sys.path = ['.'] + sys.path
|
|
|
|
from test.test_support import verbose, run_unittest, guard_warnings_filter
|
|
import re
|
|
from re import Scanner
|
|
import sys, os, traceback
|
|
from weakref import proxy
|
|
|
|
# Misc tests from Tim Peters' re.doc
|
|
|
|
# WARNING: Don't change details in these tests if you don't know
|
|
# what you're doing. Some of these tests were carefuly modeled to
|
|
# cover most of the code.
|
|
|
|
import unittest
|
|
|
|
class ReTests(unittest.TestCase):
|
|
|
|
def test_weakref(self):
|
|
s = 'QabbbcR'
|
|
x = re.compile('ab+c')
|
|
y = proxy(x)
|
|
self.assertEqual(x.findall('QabbbcR'), y.findall('QabbbcR'))
|
|
|
|
def test_search_star_plus(self):
|
|
self.assertEqual(re.search('x*', 'axx').span(0), (0, 0))
|
|
self.assertEqual(re.search('x*', 'axx').span(), (0, 0))
|
|
self.assertEqual(re.search('x+', 'axx').span(0), (1, 3))
|
|
self.assertEqual(re.search('x+', 'axx').span(), (1, 3))
|
|
self.assertEqual(re.search('x', 'aaa'), None)
|
|
self.assertEqual(re.match('a*', 'xxx').span(0), (0, 0))
|
|
self.assertEqual(re.match('a*', 'xxx').span(), (0, 0))
|
|
self.assertEqual(re.match('x*', 'xxxa').span(0), (0, 3))
|
|
self.assertEqual(re.match('x*', 'xxxa').span(), (0, 3))
|
|
self.assertEqual(re.match('a+', 'xxx'), None)
|
|
|
|
def bump_num(self, matchobj):
|
|
int_value = int(matchobj.group(0))
|
|
return str(int_value + 1)
|
|
|
|
def test_basic_re_sub(self):
|
|
self.assertEqual(re.sub("(?i)b+", "x", "bbbb BBBB"), 'x x')
|
|
self.assertEqual(re.sub(r'\d+', self.bump_num, '08.2 -2 23x99y'),
|
|
'9.3 -3 24x100y')
|
|
self.assertEqual(re.sub(r'\d+', self.bump_num, '08.2 -2 23x99y', 3),
|
|
'9.3 -3 23x99y')
|
|
|
|
self.assertEqual(re.sub('.', lambda m: r"\n", 'x'), '\\n')
|
|
self.assertEqual(re.sub('.', r"\n", 'x'), '\n')
|
|
|
|
s = r"\1\1"
|
|
self.assertEqual(re.sub('(.)', s, 'x'), 'xx')
|
|
self.assertEqual(re.sub('(.)', re.escape(s), 'x'), s)
|
|
self.assertEqual(re.sub('(.)', lambda m: s, 'x'), s)
|
|
|
|
self.assertEqual(re.sub('(?P<a>x)', '\g<a>\g<a>', 'xx'), 'xxxx')
|
|
self.assertEqual(re.sub('(?P<a>x)', '\g<a>\g<1>', 'xx'), 'xxxx')
|
|
self.assertEqual(re.sub('(?P<unk>x)', '\g<unk>\g<unk>', 'xx'), 'xxxx')
|
|
self.assertEqual(re.sub('(?P<unk>x)', '\g<1>\g<1>', 'xx'), 'xxxx')
|
|
|
|
self.assertEqual(re.sub('a',r'\t\n\v\r\f\a\b\B\Z\a\A\w\W\s\S\d\D','a'),
|
|
'\t\n\v\r\f\a\b\\B\\Z\a\\A\\w\\W\\s\\S\\d\\D')
|
|
self.assertEqual(re.sub('a', '\t\n\v\r\f\a', 'a'), '\t\n\v\r\f\a')
|
|
self.assertEqual(re.sub('a', '\t\n\v\r\f\a', 'a'),
|
|
(chr(9)+chr(10)+chr(11)+chr(13)+chr(12)+chr(7)))
|
|
|
|
self.assertEqual(re.sub('^\s*', 'X', 'test'), 'Xtest')
|
|
|
|
def test_bug_449964(self):
|
|
# fails for group followed by other escape
|
|
self.assertEqual(re.sub(r'(?P<unk>x)', '\g<1>\g<1>\\b', 'xx'),
|
|
'xx\bxx\b')
|
|
|
|
def test_bug_449000(self):
|
|
# Test for sub() on escaped characters
|
|
self.assertEqual(re.sub(r'\r\n', r'\n', 'abc\r\ndef\r\n'),
|
|
'abc\ndef\n')
|
|
self.assertEqual(re.sub('\r\n', r'\n', 'abc\r\ndef\r\n'),
|
|
'abc\ndef\n')
|
|
self.assertEqual(re.sub(r'\r\n', '\n', 'abc\r\ndef\r\n'),
|
|
'abc\ndef\n')
|
|
self.assertEqual(re.sub('\r\n', '\n', 'abc\r\ndef\r\n'),
|
|
'abc\ndef\n')
|
|
|
|
def test_sub_template_numeric_escape(self):
|
|
# bug 776311 and friends
|
|
self.assertEqual(re.sub('x', r'\0', 'x'), '\0')
|
|
self.assertEqual(re.sub('x', r'\000', 'x'), '\000')
|
|
self.assertEqual(re.sub('x', r'\001', 'x'), '\001')
|
|
self.assertEqual(re.sub('x', r'\008', 'x'), '\0' + '8')
|
|
self.assertEqual(re.sub('x', r'\009', 'x'), '\0' + '9')
|
|
self.assertEqual(re.sub('x', r'\111', 'x'), '\111')
|
|
self.assertEqual(re.sub('x', r'\117', 'x'), '\117')
|
|
|
|
self.assertEqual(re.sub('x', r'\1111', 'x'), '\1111')
|
|
self.assertEqual(re.sub('x', r'\1111', 'x'), '\111' + '1')
|
|
|
|
self.assertEqual(re.sub('x', r'\00', 'x'), '\x00')
|
|
self.assertEqual(re.sub('x', r'\07', 'x'), '\x07')
|
|
self.assertEqual(re.sub('x', r'\08', 'x'), '\0' + '8')
|
|
self.assertEqual(re.sub('x', r'\09', 'x'), '\0' + '9')
|
|
self.assertEqual(re.sub('x', r'\0a', 'x'), '\0' + 'a')
|
|
|
|
self.assertEqual(re.sub('x', r'\400', 'x'), '\0')
|
|
self.assertEqual(re.sub('x', r'\777', 'x'), '\377')
|
|
|
|
self.assertRaises(re.error, re.sub, 'x', r'\1', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\8', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\9', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\11', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\18', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\1a', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\90', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\99', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\118', 'x') # r'\11' + '8'
|
|
self.assertRaises(re.error, re.sub, 'x', r'\11a', 'x')
|
|
self.assertRaises(re.error, re.sub, 'x', r'\181', 'x') # r'\18' + '1'
|
|
self.assertRaises(re.error, re.sub, 'x', r'\800', 'x') # r'\80' + '0'
|
|
|
|
# in python2.3 (etc), these loop endlessly in sre_parser.py
|
|
self.assertEqual(re.sub('(((((((((((x)))))))))))', r'\11', 'x'), 'x')
|
|
self.assertEqual(re.sub('((((((((((y))))))))))(.)', r'\118', 'xyz'),
|
|
'xz8')
|
|
self.assertEqual(re.sub('((((((((((y))))))))))(.)', r'\11a', 'xyz'),
|
|
'xza')
|
|
|
|
def test_qualified_re_sub(self):
|
|
self.assertEqual(re.sub('a', 'b', 'aaaaa'), 'bbbbb')
|
|
self.assertEqual(re.sub('a', 'b', 'aaaaa', 1), 'baaaa')
|
|
|
|
def test_bug_114660(self):
|
|
self.assertEqual(re.sub(r'(\S)\s+(\S)', r'\1 \2', 'hello there'),
|
|
'hello there')
|
|
|
|
def test_bug_462270(self):
|
|
# Test for empty sub() behaviour, see SF bug #462270
|
|
self.assertEqual(re.sub('x*', '-', 'abxd'), '-a-b-d-')
|
|
self.assertEqual(re.sub('x+', '-', 'abxd'), 'ab-d')
|
|
|
|
def test_symbolic_refs(self):
|
|
self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<a', 'xx')
|
|
self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<', 'xx')
|
|
self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g', 'xx')
|
|
self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<a a>', 'xx')
|
|
self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<1a1>', 'xx')
|
|
self.assertRaises(IndexError, re.sub, '(?P<a>x)', '\g<ab>', 'xx')
|
|
self.assertRaises(re.error, re.sub, '(?P<a>x)|(?P<b>y)', '\g<b>', 'xx')
|
|
self.assertRaises(re.error, re.sub, '(?P<a>x)|(?P<b>y)', '\\2', 'xx')
|
|
self.assertRaises(re.error, re.sub, '(?P<a>x)', '\g<-1>', 'xx')
|
|
|
|
def test_re_subn(self):
|
|
self.assertEqual(re.subn("(?i)b+", "x", "bbbb BBBB"), ('x x', 2))
|
|
self.assertEqual(re.subn("b+", "x", "bbbb BBBB"), ('x BBBB', 1))
|
|
self.assertEqual(re.subn("b+", "x", "xyz"), ('xyz', 0))
|
|
self.assertEqual(re.subn("b*", "x", "xyz"), ('xxxyxzx', 4))
|
|
self.assertEqual(re.subn("b*", "x", "xyz", 2), ('xxxyz', 2))
|
|
|
|
def test_re_split(self):
|
|
self.assertEqual(re.split(":", ":a:b::c"), ['', 'a', 'b', '', 'c'])
|
|
self.assertEqual(re.split(":*", ":a:b::c"), ['', 'a', 'b', 'c'])
|
|
self.assertEqual(re.split("(:*)", ":a:b::c"),
|
|
['', ':', 'a', ':', 'b', '::', 'c'])
|
|
self.assertEqual(re.split("(?::*)", ":a:b::c"), ['', 'a', 'b', 'c'])
|
|
self.assertEqual(re.split("(:)*", ":a:b::c"),
|
|
['', ':', 'a', ':', 'b', ':', 'c'])
|
|
self.assertEqual(re.split("([b:]+)", ":a:b::c"),
|
|
['', ':', 'a', ':b::', 'c'])
|
|
self.assertEqual(re.split("(b)|(:+)", ":a:b::c"),
|
|
['', None, ':', 'a', None, ':', '', 'b', None, '',
|
|
None, '::', 'c'])
|
|
self.assertEqual(re.split("(?:b)|(?::+)", ":a:b::c"),
|
|
['', 'a', '', '', 'c'])
|
|
|
|
def test_qualified_re_split(self):
|
|
self.assertEqual(re.split(":", ":a:b::c", 2), ['', 'a', 'b::c'])
|
|
self.assertEqual(re.split(':', 'a:b:c:d', 2), ['a', 'b', 'c:d'])
|
|
self.assertEqual(re.split("(:)", ":a:b::c", 2),
|
|
['', ':', 'a', ':', 'b::c'])
|
|
self.assertEqual(re.split("(:*)", ":a:b::c", 2),
|
|
['', ':', 'a', ':', 'b::c'])
|
|
|
|
def test_re_findall(self):
|
|
self.assertEqual(re.findall(":+", "abc"), [])
|
|
self.assertEqual(re.findall(":+", "a:b::c:::d"), [":", "::", ":::"])
|
|
self.assertEqual(re.findall("(:+)", "a:b::c:::d"), [":", "::", ":::"])
|
|
self.assertEqual(re.findall("(:)(:*)", "a:b::c:::d"), [(":", ""),
|
|
(":", ":"),
|
|
(":", "::")])
|
|
|
|
def test_bug_117612(self):
|
|
self.assertEqual(re.findall(r"(a|(b))", "aba"),
|
|
[("a", ""),("b", "b"),("a", "")])
|
|
|
|
def test_re_match(self):
|
|
self.assertEqual(re.match('a', 'a').groups(), ())
|
|
self.assertEqual(re.match('(a)', 'a').groups(), ('a',))
|
|
self.assertEqual(re.match(r'(a)', 'a').group(0), 'a')
|
|
self.assertEqual(re.match(r'(a)', 'a').group(1), 'a')
|
|
self.assertEqual(re.match(r'(a)', 'a').group(1, 1), ('a', 'a'))
|
|
|
|
pat = re.compile('((a)|(b))(c)?')
|
|
self.assertEqual(pat.match('a').groups(), ('a', 'a', None, None))
|
|
self.assertEqual(pat.match('b').groups(), ('b', None, 'b', None))
|
|
self.assertEqual(pat.match('ac').groups(), ('a', 'a', None, 'c'))
|
|
self.assertEqual(pat.match('bc').groups(), ('b', None, 'b', 'c'))
|
|
self.assertEqual(pat.match('bc').groups(""), ('b', "", 'b', 'c'))
|
|
|
|
# A single group
|
|
m = re.match('(a)', 'a')
|
|
self.assertEqual(m.group(0), 'a')
|
|
self.assertEqual(m.group(0), 'a')
|
|
self.assertEqual(m.group(1), 'a')
|
|
self.assertEqual(m.group(1, 1), ('a', 'a'))
|
|
|
|
pat = re.compile('(?:(?P<a1>a)|(?P<b2>b))(?P<c3>c)?')
|
|
self.assertEqual(pat.match('a').group(1, 2, 3), ('a', None, None))
|
|
self.assertEqual(pat.match('b').group('a1', 'b2', 'c3'),
|
|
(None, 'b', None))
|
|
self.assertEqual(pat.match('ac').group(1, 'b2', 3), ('a', None, 'c'))
|
|
|
|
def test_re_groupref_exists(self):
|
|
self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', '(a)').groups(),
|
|
('(', 'a'))
|
|
self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', 'a').groups(),
|
|
(None, 'a'))
|
|
self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', 'a)'), None)
|
|
self.assertEqual(re.match('^(\()?([^()]+)(?(1)\))$', '(a'), None)
|
|
self.assertEqual(re.match('^(?:(a)|c)((?(1)b|d))$', 'ab').groups(),
|
|
('a', 'b'))
|
|
self.assertEqual(re.match('^(?:(a)|c)((?(1)b|d))$', 'cd').groups(),
|
|
(None, 'd'))
|
|
self.assertEqual(re.match('^(?:(a)|c)((?(1)|d))$', 'cd').groups(),
|
|
(None, 'd'))
|
|
self.assertEqual(re.match('^(?:(a)|c)((?(1)|d))$', 'a').groups(),
|
|
('a', ''))
|
|
|
|
# Tests for bug #1177831: exercise groups other than the first group
|
|
p = re.compile('(?P<g1>a)(?P<g2>b)?((?(g2)c|d))')
|
|
self.assertEqual(p.match('abc').groups(),
|
|
('a', 'b', 'c'))
|
|
self.assertEqual(p.match('ad').groups(),
|
|
('a', None, 'd'))
|
|
self.assertEqual(p.match('abd'), None)
|
|
self.assertEqual(p.match('ac'), None)
|
|
|
|
|
|
def test_re_groupref(self):
|
|
self.assertEqual(re.match(r'^(\|)?([^()]+)\1$', '|a|').groups(),
|
|
('|', 'a'))
|
|
self.assertEqual(re.match(r'^(\|)?([^()]+)\1?$', 'a').groups(),
|
|
(None, 'a'))
|
|
self.assertEqual(re.match(r'^(\|)?([^()]+)\1$', 'a|'), None)
|
|
self.assertEqual(re.match(r'^(\|)?([^()]+)\1$', '|a'), None)
|
|
self.assertEqual(re.match(r'^(?:(a)|c)(\1)$', 'aa').groups(),
|
|
('a', 'a'))
|
|
self.assertEqual(re.match(r'^(?:(a)|c)(\1)?$', 'c').groups(),
|
|
(None, None))
|
|
|
|
def test_groupdict(self):
|
|
self.assertEqual(re.match('(?P<first>first) (?P<second>second)',
|
|
'first second').groupdict(),
|
|
{'first':'first', 'second':'second'})
|
|
|
|
def test_expand(self):
|
|
self.assertEqual(re.match("(?P<first>first) (?P<second>second)",
|
|
"first second")
|
|
.expand(r"\2 \1 \g<second> \g<first>"),
|
|
"second first second first")
|
|
|
|
def test_repeat_minmax(self):
|
|
self.assertEqual(re.match("^(\w){1}$", "abc"), None)
|
|
self.assertEqual(re.match("^(\w){1}?$", "abc"), None)
|
|
self.assertEqual(re.match("^(\w){1,2}$", "abc"), None)
|
|
self.assertEqual(re.match("^(\w){1,2}?$", "abc"), None)
|
|
|
|
self.assertEqual(re.match("^(\w){3}$", "abc").group(1), "c")
|
|
self.assertEqual(re.match("^(\w){1,3}$", "abc").group(1), "c")
|
|
self.assertEqual(re.match("^(\w){1,4}$", "abc").group(1), "c")
|
|
self.assertEqual(re.match("^(\w){3,4}?$", "abc").group(1), "c")
|
|
self.assertEqual(re.match("^(\w){3}?$", "abc").group(1), "c")
|
|
self.assertEqual(re.match("^(\w){1,3}?$", "abc").group(1), "c")
|
|
self.assertEqual(re.match("^(\w){1,4}?$", "abc").group(1), "c")
|
|
self.assertEqual(re.match("^(\w){3,4}?$", "abc").group(1), "c")
|
|
|
|
self.assertEqual(re.match("^x{1}$", "xxx"), None)
|
|
self.assertEqual(re.match("^x{1}?$", "xxx"), None)
|
|
self.assertEqual(re.match("^x{1,2}$", "xxx"), None)
|
|
self.assertEqual(re.match("^x{1,2}?$", "xxx"), None)
|
|
|
|
self.assertNotEqual(re.match("^x{3}$", "xxx"), None)
|
|
self.assertNotEqual(re.match("^x{1,3}$", "xxx"), None)
|
|
self.assertNotEqual(re.match("^x{1,4}$", "xxx"), None)
|
|
self.assertNotEqual(re.match("^x{3,4}?$", "xxx"), None)
|
|
self.assertNotEqual(re.match("^x{3}?$", "xxx"), None)
|
|
self.assertNotEqual(re.match("^x{1,3}?$", "xxx"), None)
|
|
self.assertNotEqual(re.match("^x{1,4}?$", "xxx"), None)
|
|
self.assertNotEqual(re.match("^x{3,4}?$", "xxx"), None)
|
|
|
|
self.assertEqual(re.match("^x{}$", "xxx"), None)
|
|
self.assertNotEqual(re.match("^x{}$", "x{}"), None)
|
|
|
|
def test_getattr(self):
|
|
self.assertEqual(re.match("(a)", "a").pos, 0)
|
|
self.assertEqual(re.match("(a)", "a").endpos, 1)
|
|
self.assertEqual(re.match("(a)", "a").string, "a")
|
|
self.assertEqual(re.match("(a)", "a").regs, ((0, 1), (0, 1)))
|
|
self.assertNotEqual(re.match("(a)", "a").re, None)
|
|
|
|
def test_special_escapes(self):
|
|
self.assertEqual(re.search(r"\b(b.)\b",
|
|
"abcd abc bcd bx").group(1), "bx")
|
|
self.assertEqual(re.search(r"\B(b.)\B",
|
|
"abc bcd bc abxd").group(1), "bx")
|
|
self.assertEqual(re.search(r"\b(b.)\b",
|
|
"abcd abc bcd bx", re.LOCALE).group(1), "bx")
|
|
self.assertEqual(re.search(r"\B(b.)\B",
|
|
"abc bcd bc abxd", re.LOCALE).group(1), "bx")
|
|
self.assertEqual(re.search(r"\b(b.)\b",
|
|
"abcd abc bcd bx", re.UNICODE).group(1), "bx")
|
|
self.assertEqual(re.search(r"\B(b.)\B",
|
|
"abc bcd bc abxd", re.UNICODE).group(1), "bx")
|
|
self.assertEqual(re.search(r"^abc$", "\nabc\n", re.M).group(0), "abc")
|
|
self.assertEqual(re.search(r"^\Aabc\Z$", "abc", re.M).group(0), "abc")
|
|
self.assertEqual(re.search(r"^\Aabc\Z$", "\nabc\n", re.M), None)
|
|
self.assertEqual(re.search(r"\b(b.)\b",
|
|
"abcd abc bcd bx").group(1), "bx")
|
|
self.assertEqual(re.search(r"\B(b.)\B",
|
|
"abc bcd bc abxd").group(1), "bx")
|
|
self.assertEqual(re.search(r"^abc$", "\nabc\n", re.M).group(0), "abc")
|
|
self.assertEqual(re.search(r"^\Aabc\Z$", "abc", re.M).group(0), "abc")
|
|
self.assertEqual(re.search(r"^\Aabc\Z$", "\nabc\n", re.M), None)
|
|
self.assertEqual(re.search(r"\d\D\w\W\s\S",
|
|
"1aa! a").group(0), "1aa! a")
|
|
self.assertEqual(re.search(r"\d\D\w\W\s\S",
|
|
"1aa! a", re.LOCALE).group(0), "1aa! a")
|
|
self.assertEqual(re.search(r"\d\D\w\W\s\S",
|
|
"1aa! a", re.UNICODE).group(0), "1aa! a")
|
|
|
|
def test_ignore_case(self):
|
|
self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
|
|
self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
|
|
|
|
def test_bigcharset(self):
|
|
self.assertEqual(re.match("([\u2222\u2223])",
|
|
"\u2222").group(1), "\u2222")
|
|
self.assertEqual(re.match("([\u2222\u2223])",
|
|
"\u2222", re.UNICODE).group(1), "\u2222")
|
|
|
|
def test_anyall(self):
|
|
self.assertEqual(re.match("a.b", "a\nb", re.DOTALL).group(0),
|
|
"a\nb")
|
|
self.assertEqual(re.match("a.*b", "a\n\nb", re.DOTALL).group(0),
|
|
"a\n\nb")
|
|
|
|
def test_non_consuming(self):
|
|
self.assertEqual(re.match("(a(?=\s[^a]))", "a b").group(1), "a")
|
|
self.assertEqual(re.match("(a(?=\s[^a]*))", "a b").group(1), "a")
|
|
self.assertEqual(re.match("(a(?=\s[abc]))", "a b").group(1), "a")
|
|
self.assertEqual(re.match("(a(?=\s[abc]*))", "a bc").group(1), "a")
|
|
self.assertEqual(re.match(r"(a)(?=\s\1)", "a a").group(1), "a")
|
|
self.assertEqual(re.match(r"(a)(?=\s\1*)", "a aa").group(1), "a")
|
|
self.assertEqual(re.match(r"(a)(?=\s(abc|a))", "a a").group(1), "a")
|
|
|
|
self.assertEqual(re.match(r"(a(?!\s[^a]))", "a a").group(1), "a")
|
|
self.assertEqual(re.match(r"(a(?!\s[abc]))", "a d").group(1), "a")
|
|
self.assertEqual(re.match(r"(a)(?!\s\1)", "a b").group(1), "a")
|
|
self.assertEqual(re.match(r"(a)(?!\s(abc|a))", "a b").group(1), "a")
|
|
|
|
def test_ignore_case(self):
|
|
self.assertEqual(re.match(r"(a\s[^a])", "a b", re.I).group(1), "a b")
|
|
self.assertEqual(re.match(r"(a\s[^a]*)", "a bb", re.I).group(1), "a bb")
|
|
self.assertEqual(re.match(r"(a\s[abc])", "a b", re.I).group(1), "a b")
|
|
self.assertEqual(re.match(r"(a\s[abc]*)", "a bb", re.I).group(1), "a bb")
|
|
self.assertEqual(re.match(r"((a)\s\2)", "a a", re.I).group(1), "a a")
|
|
self.assertEqual(re.match(r"((a)\s\2*)", "a aa", re.I).group(1), "a aa")
|
|
self.assertEqual(re.match(r"((a)\s(abc|a))", "a a", re.I).group(1), "a a")
|
|
self.assertEqual(re.match(r"((a)\s(abc|a)*)", "a aa", re.I).group(1), "a aa")
|
|
|
|
def test_category(self):
|
|
self.assertEqual(re.match(r"(\s)", " ").group(1), " ")
|
|
|
|
def test_getlower(self):
|
|
import _sre
|
|
self.assertEqual(_sre.getlower(ord('A'), 0), ord('a'))
|
|
self.assertEqual(_sre.getlower(ord('A'), re.LOCALE), ord('a'))
|
|
self.assertEqual(_sre.getlower(ord('A'), re.UNICODE), ord('a'))
|
|
|
|
self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
|
|
self.assertEqual(re.match("abc", "ABC", re.I).group(0), "ABC")
|
|
|
|
def test_not_literal(self):
|
|
self.assertEqual(re.search("\s([^a])", " b").group(1), "b")
|
|
self.assertEqual(re.search("\s([^a]*)", " bb").group(1), "bb")
|
|
|
|
def test_search_coverage(self):
|
|
self.assertEqual(re.search("\s(b)", " b").group(1), "b")
|
|
self.assertEqual(re.search("a\s", "a ").group(0), "a ")
|
|
|
|
def test_re_escape(self):
|
|
p=""
|
|
for i in range(0, 256):
|
|
p = p + chr(i)
|
|
self.assertEqual(re.match(re.escape(chr(i)), chr(i)) is not None,
|
|
True)
|
|
self.assertEqual(re.match(re.escape(chr(i)), chr(i)).span(), (0,1))
|
|
|
|
pat=re.compile(re.escape(p))
|
|
self.assertEqual(pat.match(p) is not None, True)
|
|
self.assertEqual(pat.match(p).span(), (0,256))
|
|
|
|
def test_pickling(self):
|
|
import pickle
|
|
self.pickle_test(pickle)
|
|
# old pickles expect the _compile() reconstructor in sre module
|
|
import warnings
|
|
with guard_warnings_filter():
|
|
warnings.filterwarnings("ignore", "The sre module is deprecated",
|
|
DeprecationWarning)
|
|
from sre import _compile
|
|
|
|
def pickle_test(self, pickle):
|
|
oldpat = re.compile('a(?:b|(c|e){1,2}?|d)+?(.)')
|
|
s = pickle.dumps(oldpat)
|
|
newpat = pickle.loads(s)
|
|
self.assertEqual(oldpat, newpat)
|
|
|
|
def test_constants(self):
|
|
self.assertEqual(re.I, re.IGNORECASE)
|
|
self.assertEqual(re.L, re.LOCALE)
|
|
self.assertEqual(re.M, re.MULTILINE)
|
|
self.assertEqual(re.S, re.DOTALL)
|
|
self.assertEqual(re.X, re.VERBOSE)
|
|
|
|
def test_flags(self):
|
|
for flag in [re.I, re.M, re.X, re.S, re.L]:
|
|
self.assertNotEqual(re.compile('^pattern$', flag), None)
|
|
|
|
def test_sre_character_literals(self):
|
|
for i in [0, 8, 16, 32, 64, 127, 128, 255]:
|
|
self.assertNotEqual(re.match(r"\%03o" % i, chr(i)), None)
|
|
self.assertNotEqual(re.match(r"\%03o0" % i, chr(i)+"0"), None)
|
|
self.assertNotEqual(re.match(r"\%03o8" % i, chr(i)+"8"), None)
|
|
self.assertNotEqual(re.match(r"\x%02x" % i, chr(i)), None)
|
|
self.assertNotEqual(re.match(r"\x%02x0" % i, chr(i)+"0"), None)
|
|
self.assertNotEqual(re.match(r"\x%02xz" % i, chr(i)+"z"), None)
|
|
self.assertRaises(re.error, re.match, "\911", "")
|
|
|
|
def test_sre_character_class_literals(self):
|
|
for i in [0, 8, 16, 32, 64, 127, 128, 255]:
|
|
self.assertNotEqual(re.match(r"[\%03o]" % i, chr(i)), None)
|
|
self.assertNotEqual(re.match(r"[\%03o0]" % i, chr(i)), None)
|
|
self.assertNotEqual(re.match(r"[\%03o8]" % i, chr(i)), None)
|
|
self.assertNotEqual(re.match(r"[\x%02x]" % i, chr(i)), None)
|
|
self.assertNotEqual(re.match(r"[\x%02x0]" % i, chr(i)), None)
|
|
self.assertNotEqual(re.match(r"[\x%02xz]" % i, chr(i)), None)
|
|
self.assertRaises(re.error, re.match, "[\911]", "")
|
|
|
|
def test_bug_113254(self):
|
|
self.assertEqual(re.match(r'(a)|(b)', 'b').start(1), -1)
|
|
self.assertEqual(re.match(r'(a)|(b)', 'b').end(1), -1)
|
|
self.assertEqual(re.match(r'(a)|(b)', 'b').span(1), (-1, -1))
|
|
|
|
def test_bug_527371(self):
|
|
# bug described in patches 527371/672491
|
|
self.assertEqual(re.match(r'(a)?a','a').lastindex, None)
|
|
self.assertEqual(re.match(r'(a)(b)?b','ab').lastindex, 1)
|
|
self.assertEqual(re.match(r'(?P<a>a)(?P<b>b)?b','ab').lastgroup, 'a')
|
|
self.assertEqual(re.match("(?P<a>a(b))", "ab").lastgroup, 'a')
|
|
self.assertEqual(re.match("((a))", "a").lastindex, 1)
|
|
|
|
def test_bug_545855(self):
|
|
# bug 545855 -- This pattern failed to cause a compile error as it
|
|
# should, instead provoking a TypeError.
|
|
self.assertRaises(re.error, re.compile, 'foo[a-')
|
|
|
|
def test_bug_418626(self):
|
|
# bugs 418626 at al. -- Testing Greg Chapman's addition of op code
|
|
# SRE_OP_MIN_REPEAT_ONE for eliminating recursion on simple uses of
|
|
# pattern '*?' on a long string.
|
|
self.assertEqual(re.match('.*?c', 10000*'ab'+'cd').end(0), 20001)
|
|
self.assertEqual(re.match('.*?cd', 5000*'ab'+'c'+5000*'ab'+'cde').end(0),
|
|
20003)
|
|
self.assertEqual(re.match('.*?cd', 20000*'abc'+'de').end(0), 60001)
|
|
# non-simple '*?' still used to hit the recursion limit, before the
|
|
# non-recursive scheme was implemented.
|
|
self.assertEqual(re.search('(a|b)*?c', 10000*'ab'+'cd').end(0), 20001)
|
|
|
|
def test_bug_612074(self):
|
|
pat="["+re.escape("\u2039")+"]"
|
|
self.assertEqual(re.compile(pat) and 1, 1)
|
|
|
|
def test_stack_overflow(self):
|
|
# nasty cases that used to overflow the straightforward recursive
|
|
# implementation of repeated groups.
|
|
self.assertEqual(re.match('(x)*', 50000*'x').group(1), 'x')
|
|
self.assertEqual(re.match('(x)*y', 50000*'x'+'y').group(1), 'x')
|
|
self.assertEqual(re.match('(x)*?y', 50000*'x'+'y').group(1), 'x')
|
|
|
|
def test_scanner(self):
|
|
def s_ident(scanner, token): return token
|
|
def s_operator(scanner, token): return "op%s" % token
|
|
def s_float(scanner, token): return float(token)
|
|
def s_int(scanner, token): return int(token)
|
|
|
|
scanner = Scanner([
|
|
(r"[a-zA-Z_]\w*", s_ident),
|
|
(r"\d+\.\d*", s_float),
|
|
(r"\d+", s_int),
|
|
(r"=|\+|-|\*|/", s_operator),
|
|
(r"\s+", None),
|
|
])
|
|
|
|
self.assertNotEqual(scanner.scanner.scanner("").pattern, None)
|
|
|
|
self.assertEqual(scanner.scan("sum = 3*foo + 312.50 + bar"),
|
|
(['sum', 'op=', 3, 'op*', 'foo', 'op+', 312.5,
|
|
'op+', 'bar'], ''))
|
|
|
|
def test_bug_448951(self):
|
|
# bug 448951 (similar to 429357, but with single char match)
|
|
# (Also test greedy matches.)
|
|
for op in '','?','*':
|
|
self.assertEqual(re.match(r'((.%s):)?z'%op, 'z').groups(),
|
|
(None, None))
|
|
self.assertEqual(re.match(r'((.%s):)?z'%op, 'a:z').groups(),
|
|
('a:', 'a'))
|
|
|
|
def test_bug_725106(self):
|
|
# capturing groups in alternatives in repeats
|
|
self.assertEqual(re.match('^((a)|b)*', 'abc').groups(),
|
|
('b', 'a'))
|
|
self.assertEqual(re.match('^(([ab])|c)*', 'abc').groups(),
|
|
('c', 'b'))
|
|
self.assertEqual(re.match('^((d)|[ab])*', 'abc').groups(),
|
|
('b', None))
|
|
self.assertEqual(re.match('^((a)c|[ab])*', 'abc').groups(),
|
|
('b', None))
|
|
self.assertEqual(re.match('^((a)|b)*?c', 'abc').groups(),
|
|
('b', 'a'))
|
|
self.assertEqual(re.match('^(([ab])|c)*?d', 'abcd').groups(),
|
|
('c', 'b'))
|
|
self.assertEqual(re.match('^((d)|[ab])*?c', 'abc').groups(),
|
|
('b', None))
|
|
self.assertEqual(re.match('^((a)c|[ab])*?c', 'abc').groups(),
|
|
('b', None))
|
|
|
|
def test_bug_725149(self):
|
|
# mark_stack_base restoring before restoring marks
|
|
self.assertEqual(re.match('(a)(?:(?=(b)*)c)*', 'abb').groups(),
|
|
('a', None))
|
|
self.assertEqual(re.match('(a)((?!(b)*))*', 'abb').groups(),
|
|
('a', None, None))
|
|
|
|
def test_bug_764548(self):
|
|
# bug 764548, re.compile() barfs on str/unicode subclasses
|
|
class my_unicode(str): pass
|
|
pat = re.compile(my_unicode("abc"))
|
|
self.assertEqual(pat.match("xyz"), None)
|
|
|
|
def test_finditer(self):
|
|
iter = re.finditer(r":+", "a:b::c:::d")
|
|
self.assertEqual([item.group(0) for item in iter],
|
|
[":", "::", ":::"])
|
|
|
|
def test_bug_926075(self):
|
|
self.assert_(re.compile('bug_926075') is not
|
|
re.compile(str8('bug_926075')))
|
|
|
|
def test_bug_931848(self):
|
|
pattern = eval('"[\u002E\u3002\uFF0E\uFF61]"')
|
|
self.assertEqual(re.compile(pattern).split("a.b.c"),
|
|
['a','b','c'])
|
|
|
|
def test_bug_581080(self):
|
|
iter = re.finditer(r"\s", "a b")
|
|
self.assertEqual(next(iter).span(), (1,2))
|
|
self.assertRaises(StopIteration, next, iter)
|
|
|
|
scanner = re.compile(r"\s").scanner("a b")
|
|
self.assertEqual(scanner.search().span(), (1, 2))
|
|
self.assertEqual(scanner.search(), None)
|
|
|
|
def test_bug_817234(self):
|
|
iter = re.finditer(r".*", "asdf")
|
|
self.assertEqual(next(iter).span(), (0, 4))
|
|
self.assertEqual(next(iter).span(), (4, 4))
|
|
self.assertRaises(StopIteration, next, iter)
|
|
|
|
def test_empty_array(self):
|
|
# SF buf 1647541
|
|
import array
|
|
for typecode in 'bBuhHiIlLfd':
|
|
a = array.array(typecode)
|
|
self.assertEqual(re.compile("bla").match(a), None)
|
|
self.assertEqual(re.compile("").match(a).groups(), ())
|
|
|
|
def run_re_tests():
|
|
from test.re_tests import benchmarks, tests, SUCCEED, FAIL, SYNTAX_ERROR
|
|
if verbose:
|
|
print('Running re_tests test suite')
|
|
else:
|
|
# To save time, only run the first and last 10 tests
|
|
#tests = tests[:10] + tests[-10:]
|
|
pass
|
|
|
|
for t in tests:
|
|
sys.stdout.flush()
|
|
pattern = s = outcome = repl = expected = None
|
|
if len(t) == 5:
|
|
pattern, s, outcome, repl, expected = t
|
|
elif len(t) == 3:
|
|
pattern, s, outcome = t
|
|
else:
|
|
raise ValueError, ('Test tuples should have 3 or 5 fields', t)
|
|
|
|
try:
|
|
obj = re.compile(pattern)
|
|
except re.error:
|
|
if outcome == SYNTAX_ERROR: pass # Expected a syntax error
|
|
else:
|
|
print('=== Syntax error:', t)
|
|
except KeyboardInterrupt: raise KeyboardInterrupt
|
|
except:
|
|
print('*** Unexpected error ***', t)
|
|
if verbose:
|
|
traceback.print_exc(file=sys.stdout)
|
|
else:
|
|
try:
|
|
result = obj.search(s)
|
|
except re.error as msg:
|
|
print('=== Unexpected exception', t, repr(msg))
|
|
if outcome == SYNTAX_ERROR:
|
|
# This should have been a syntax error; forget it.
|
|
pass
|
|
elif outcome == FAIL:
|
|
if result is None: pass # No match, as expected
|
|
else: print('=== Succeeded incorrectly', t)
|
|
elif outcome == SUCCEED:
|
|
if result is not None:
|
|
# Matched, as expected, so now we compute the
|
|
# result string and compare it to our expected result.
|
|
start, end = result.span(0)
|
|
vardict={'found': result.group(0),
|
|
'groups': result.group(),
|
|
'flags': result.re.flags}
|
|
for i in range(1, 100):
|
|
try:
|
|
gi = result.group(i)
|
|
# Special hack because else the string concat fails:
|
|
if gi is None:
|
|
gi = "None"
|
|
except IndexError:
|
|
gi = "Error"
|
|
vardict['g%d' % i] = gi
|
|
for i in result.re.groupindex.keys():
|
|
try:
|
|
gi = result.group(i)
|
|
if gi is None:
|
|
gi = "None"
|
|
except IndexError:
|
|
gi = "Error"
|
|
vardict[i] = gi
|
|
repl = eval(repl, vardict)
|
|
if repl != expected:
|
|
print('=== grouping error', t, end=' ')
|
|
print(repr(repl) + ' should be ' + repr(expected))
|
|
else:
|
|
print('=== Failed incorrectly', t)
|
|
|
|
# Try the match on a unicode string, and check that it
|
|
# still succeeds.
|
|
try:
|
|
result = obj.search(str(s, "latin-1"))
|
|
if result is None:
|
|
print('=== Fails on unicode match', t)
|
|
except NameError:
|
|
continue # 1.5.2
|
|
except TypeError:
|
|
continue # unicode test case
|
|
|
|
# Try the match on a unicode pattern, and check that it
|
|
# still succeeds.
|
|
obj=re.compile(str(pattern, "latin-1"))
|
|
result = obj.search(s)
|
|
if result is None:
|
|
print('=== Fails on unicode pattern match', t)
|
|
|
|
# Try the match with the search area limited to the extent
|
|
# of the match and see if it still succeeds. \B will
|
|
# break (because it won't match at the end or start of a
|
|
# string), so we'll ignore patterns that feature it.
|
|
|
|
if pattern[:2] != '\\B' and pattern[-2:] != '\\B' \
|
|
and result is not None:
|
|
obj = re.compile(pattern)
|
|
result = obj.search(s, result.start(0), result.end(0) + 1)
|
|
if result is None:
|
|
print('=== Failed on range-limited match', t)
|
|
|
|
# Try the match with IGNORECASE enabled, and check that it
|
|
# still succeeds.
|
|
obj = re.compile(pattern, re.IGNORECASE)
|
|
result = obj.search(s)
|
|
if result is None:
|
|
print('=== Fails on case-insensitive match', t)
|
|
|
|
# Try the match with LOCALE enabled, and check that it
|
|
# still succeeds.
|
|
obj = re.compile(pattern, re.LOCALE)
|
|
result = obj.search(s)
|
|
if result is None:
|
|
print('=== Fails on locale-sensitive match', t)
|
|
|
|
# Try the match with UNICODE locale enabled, and check
|
|
# that it still succeeds.
|
|
obj = re.compile(pattern, re.UNICODE)
|
|
result = obj.search(s)
|
|
if result is None:
|
|
print('=== Fails on unicode-sensitive match', t)
|
|
|
|
def test_main():
|
|
run_unittest(ReTests)
|
|
run_re_tests()
|
|
|
|
if __name__ == "__main__":
|
|
test_main()
|