This file is indexed.

/usr/lib/python2.7/test/test_pep263.py is in libpython2.7-testsuite 2.7.6-8ubuntu0.5.

This file is owned by root:root, with mode 0o644.

The actual contents of the file can be viewed below.

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
# -*- coding: koi8-r -*-

import unittest
from test import test_support

class PEP263Test(unittest.TestCase):

    def test_pep263(self):
        self.assertEqual(
            u"ðÉÔÏÎ".encode("utf-8"),
            '\xd0\x9f\xd0\xb8\xd1\x82\xd0\xbe\xd0\xbd'
        )
        self.assertEqual(
            u"\ð".encode("utf-8"),
            '\\\xd0\x9f'
        )

    def test_compilestring(self):
        # see #1882
        c = compile("\n# coding: utf-8\nu = u'\xc3\xb3'\n", "dummy", "exec")
        d = {}
        exec c in d
        self.assertEqual(d['u'], u'\xf3')


    def test_issue3297(self):
        c = compile("a, b = '\U0001010F', '\\U0001010F'", "dummy", "exec")
        d = {}
        exec(c, d)
        self.assertEqual(d['a'], d['b'])
        self.assertEqual(len(d['a']), len(d['b']))

    def test_issue7820(self):
        # Ensure that check_bom() restores all bytes in the right order if
        # check_bom() fails in pydebug mode: a buffer starts with the first
        # byte of a valid BOM, but next bytes are different

        # one byte in common with the UTF-16-LE BOM
        self.assertRaises(SyntaxError, eval, '\xff\x20')

        # two bytes in common with the UTF-8 BOM
        self.assertRaises(SyntaxError, eval, '\xef\xbb\x20')

    def test_error_message(self):
        compile('# -*- coding: iso-8859-15 -*-\n', 'dummy', 'exec')
        compile('\xef\xbb\xbf\n', 'dummy', 'exec')
        compile('\xef\xbb\xbf# -*- coding: utf-8 -*-\n', 'dummy', 'exec')
        with self.assertRaisesRegexp(SyntaxError, 'fake'):
            compile('# -*- coding: fake -*-\n', 'dummy', 'exec')
        with self.assertRaisesRegexp(SyntaxError, 'iso-8859-15'):
            compile('\xef\xbb\xbf# -*- coding: iso-8859-15 -*-\n',
                    'dummy', 'exec')
        with self.assertRaisesRegexp(SyntaxError, 'BOM'):
            compile('\xef\xbb\xbf# -*- coding: iso-8859-15 -*-\n',
                    'dummy', 'exec')
        with self.assertRaisesRegexp(SyntaxError, 'fake'):
            compile('\xef\xbb\xbf# -*- coding: fake -*-\n', 'dummy', 'exec')
        with self.assertRaisesRegexp(SyntaxError, 'BOM'):
            compile('\xef\xbb\xbf# -*- coding: fake -*-\n', 'dummy', 'exec')

    def test_non_unicode_codec(self):
        with self.assertRaisesRegexp(SyntaxError,
                                     'codec did not return a unicode'):
            from test import bad_coding3


def test_main():
    test_support.run_unittest(PEP263Test)

if __name__=="__main__":
    test_main()