Lib/test/test_urlparse.py

   1 #! /usr/bin/env python
   2
   3 from test import test_support
   4 import unittest
   5 import urlparse
   6
   7 RFC1808_BASE = "http://a/b/c/d;p?q#f"
   8 RFC2396_BASE = "http://a/b/c/d;p?q"
   9
  10 class UrlParseTestCase(unittest.TestCase):
  11     def test_frags(self):
  12         for url, parsed, split in [
  13             ('http://www.python.org',
  14              ('http', 'www.python.org', '', '', '', ''),
  15              ('http', 'www.python.org', '', '', '')),
  16             ('http://www.python.org#abc',
  17              ('http', 'www.python.org', '', '', '', 'abc'),
  18              ('http', 'www.python.org', '', '', 'abc')),
  19             ('http://www.python.org/#abc',
  20              ('http', 'www.python.org', '/', '', '', 'abc'),
  21              ('http', 'www.python.org', '/', '', 'abc')),
  22             (RFC1808_BASE,
  23              ('http', 'a', '/b/c/d', 'p', 'q', 'f'),
  24              ('http', 'a', '/b/c/d;p', 'q', 'f')),
  25             ('file:///tmp/junk.txt',
  26              ('file', '', '/tmp/junk.txt', '', '', ''),
  27              ('file', '', '/tmp/junk.txt', '', '')),
  28             ]:
  29             result = urlparse.urlparse(url)
  30             self.assertEqual(result, parsed)
  31             # put it back together and it should be the same
  32             result2 = urlparse.urlunparse(result)
  33             self.assertEqual(result2, url)
  34
  35             # check the roundtrip using urlsplit() as well
  36             result = urlparse.urlsplit(url)
  37             self.assertEqual(result, split)
  38             result2 = urlparse.urlunsplit(result)
  39             self.assertEqual(result2, url)
  40
  41     def checkJoin(self, base, relurl, expected):
  42         self.assertEqual(urlparse.urljoin(base, relurl), expected,
  43                          (base, relurl, expected))
  44
  45     def test_unparse_parse(self):
  46         for u in ['Python', './Python']:
  47             self.assertEqual(urlparse.urlunsplit(urlparse.urlsplit(u)), u)
  48             self.assertEqual(urlparse.urlunparse(urlparse.urlparse(u)), u)
  49
  50     def test_RFC1808(self):
  51         # "normal" cases from RFC 1808:
  52         self.checkJoin(RFC1808_BASE, 'g:h', 'g:h')
  53         self.checkJoin(RFC1808_BASE, 'g', 'http://a/b/c/g')
  54         self.checkJoin(RFC1808_BASE, './g', 'http://a/b/c/g')
  55         self.checkJoin(RFC1808_BASE, 'g/', 'http://a/b/c/g/')
  56         self.checkJoin(RFC1808_BASE, '/g', 'http://a/g')
  57         self.checkJoin(RFC1808_BASE, '//g', 'http://g')
  58         self.checkJoin(RFC1808_BASE, '?y', 'http://a/b/c/d;p?y')
  59         self.checkJoin(RFC1808_BASE, 'g?y', 'http://a/b/c/g?y')
  60         self.checkJoin(RFC1808_BASE, 'g?y/./x', 'http://a/b/c/g?y/./x')
  61         self.checkJoin(RFC1808_BASE, '#s', 'http://a/b/c/d;p?q#s')
  62         self.checkJoin(RFC1808_BASE, 'g#s', 'http://a/b/c/g#s')
  63         self.checkJoin(RFC1808_BASE, 'g#s/./x', 'http://a/b/c/g#s/./x')
  64         self.checkJoin(RFC1808_BASE, 'g?y#s', 'http://a/b/c/g?y#s')
  65         self.checkJoin(RFC1808_BASE, ';x', 'http://a/b/c/d;x')
  66         self.checkJoin(RFC1808_BASE, 'g;x', 'http://a/b/c/g;x')
  67         self.checkJoin(RFC1808_BASE, 'g;x?y#s', 'http://a/b/c/g;x?y#s')
  68         self.checkJoin(RFC1808_BASE, '.', 'http://a/b/c/')
  69         self.checkJoin(RFC1808_BASE, './', 'http://a/b/c/')
  70         self.checkJoin(RFC1808_BASE, '..', 'http://a/b/')
  71         self.checkJoin(RFC1808_BASE, '../', 'http://a/b/')
  72         self.checkJoin(RFC1808_BASE, '../g', 'http://a/b/g')
  73         self.checkJoin(RFC1808_BASE, '../..', 'http://a/')
  74         self.checkJoin(RFC1808_BASE, '../../', 'http://a/')
  75         self.checkJoin(RFC1808_BASE, '../../g', 'http://a/g')
  76
  77         # "abnormal" cases from RFC 1808:
  78         self.checkJoin(RFC1808_BASE, '', 'http://a/b/c/d;p?q#f')
  79         self.checkJoin(RFC1808_BASE, '../../../g', 'http://a/../g')
  80         self.checkJoin(RFC1808_BASE, '../../../../g', 'http://a/../../g')
  81         self.checkJoin(RFC1808_BASE, '/./g', 'http://a/./g')
  82         self.checkJoin(RFC1808_BASE, '/../g', 'http://a/../g')
  83         self.checkJoin(RFC1808_BASE, 'g.', 'http://a/b/c/g.')
  84         self.checkJoin(RFC1808_BASE, '.g', 'http://a/b/c/.g')
  85         self.checkJoin(RFC1808_BASE, 'g..', 'http://a/b/c/g..')
  86         self.checkJoin(RFC1808_BASE, '..g', 'http://a/b/c/..g')
  87         self.checkJoin(RFC1808_BASE, './../g', 'http://a/b/g')
  88         self.checkJoin(RFC1808_BASE, './g/.', 'http://a/b/c/g/')
  89         self.checkJoin(RFC1808_BASE, 'g/./h', 'http://a/b/c/g/h')
  90         self.checkJoin(RFC1808_BASE, 'g/../h', 'http://a/b/c/h')
  91
  92         # RFC 1808 and RFC 1630 disagree on these (according to RFC 1808),
  93         # so we'll not actually run these tests (which expect 1808 behavior).
  94         #self.checkJoin(RFC1808_BASE, 'http:g', 'http:g')
  95         #self.checkJoin(RFC1808_BASE, 'http:', 'http:')
  96
  97     def test_RFC2396(self):
  98         # cases from RFC 2396
  99
 100         ### urlparse.py as of v 1.32 fails on these two
 101         #self.checkJoin(RFC2396_BASE, '?y', 'http://a/b/c/?y')
 102         #self.checkJoin(RFC2396_BASE, ';x', 'http://a/b/c/;x')
 103
 104         self.checkJoin(RFC2396_BASE, 'g:h', 'g:h')
 105         self.checkJoin(RFC2396_BASE, 'g', 'http://a/b/c/g')
 106         self.checkJoin(RFC2396_BASE, './g', 'http://a/b/c/g')
 107         self.checkJoin(RFC2396_BASE, 'g/', 'http://a/b/c/g/')
 108         self.checkJoin(RFC2396_BASE, '/g', 'http://a/g')
 109         self.checkJoin(RFC2396_BASE, '//g', 'http://g')
 110         self.checkJoin(RFC2396_BASE, 'g?y', 'http://a/b/c/g?y')
 111         self.checkJoin(RFC2396_BASE, '#s', 'http://a/b/c/d;p?q#s')
 112         self.checkJoin(RFC2396_BASE, 'g#s', 'http://a/b/c/g#s')
 113         self.checkJoin(RFC2396_BASE, 'g?y#s', 'http://a/b/c/g?y#s')
 114         self.checkJoin(RFC2396_BASE, 'g;x', 'http://a/b/c/g;x')
 115         self.checkJoin(RFC2396_BASE, 'g;x?y#s', 'http://a/b/c/g;x?y#s')
 116         self.checkJoin(RFC2396_BASE, '.', 'http://a/b/c/')
 117         self.checkJoin(RFC2396_BASE, './', 'http://a/b/c/')
 118         self.checkJoin(RFC2396_BASE, '..', 'http://a/b/')
 119         self.checkJoin(RFC2396_BASE, '../', 'http://a/b/')
 120         self.checkJoin(RFC2396_BASE, '../g', 'http://a/b/g')
 121         self.checkJoin(RFC2396_BASE, '../..', 'http://a/')
 122         self.checkJoin(RFC2396_BASE, '../../', 'http://a/')
 123         self.checkJoin(RFC2396_BASE, '../../g', 'http://a/g')
 124         self.checkJoin(RFC2396_BASE, '', RFC2396_BASE)
 125         self.checkJoin(RFC2396_BASE, '../../../g', 'http://a/../g')
 126         self.checkJoin(RFC2396_BASE, '../../../../g', 'http://a/../../g')
 127         self.checkJoin(RFC2396_BASE, '/./g', 'http://a/./g')
 128         self.checkJoin(RFC2396_BASE, '/../g', 'http://a/../g')
 129         self.checkJoin(RFC2396_BASE, 'g.', 'http://a/b/c/g.')
 130         self.checkJoin(RFC2396_BASE, '.g', 'http://a/b/c/.g')
 131         self.checkJoin(RFC2396_BASE, 'g..', 'http://a/b/c/g..')
 132         self.checkJoin(RFC2396_BASE, '..g', 'http://a/b/c/..g')
 133         self.checkJoin(RFC2396_BASE, './../g', 'http://a/b/g')
 134         self.checkJoin(RFC2396_BASE, './g/.', 'http://a/b/c/g/')
 135         self.checkJoin(RFC2396_BASE, 'g/./h', 'http://a/b/c/g/h')
 136         self.checkJoin(RFC2396_BASE, 'g/../h', 'http://a/b/c/h')
 137         self.checkJoin(RFC2396_BASE, 'g;x=1/./y', 'http://a/b/c/g;x=1/y')
 138         self.checkJoin(RFC2396_BASE, 'g;x=1/../y', 'http://a/b/c/y')
 139         self.checkJoin(RFC2396_BASE, 'g?y/./x', 'http://a/b/c/g?y/./x')
 140         self.checkJoin(RFC2396_BASE, 'g?y/../x', 'http://a/b/c/g?y/../x')
 141         self.checkJoin(RFC2396_BASE, 'g#s/./x', 'http://a/b/c/g#s/./x')
 142         self.checkJoin(RFC2396_BASE, 'g#s/../x', 'http://a/b/c/g#s/../x')
 143
 144     def test_urldefrag(self):
 145         for url, defrag, frag in [
 146             ('http://python.org#frag', 'http://python.org', 'frag'),
 147             ('http://python.org', 'http://python.org', ''),
 148             ('http://python.org/#frag', 'http://python.org/', 'frag'),
 149             ('http://python.org/', 'http://python.org/', ''),
 150             ('http://python.org/?q#frag', 'http://python.org/?q', 'frag'),
 151             ('http://python.org/?q', 'http://python.org/?q', ''),
 152             ('http://python.org/p#frag', 'http://python.org/p', 'frag'),
 153             ('http://python.org/p?q', 'http://python.org/p?q', ''),
 154             (RFC1808_BASE, 'http://a/b/c/d;p?q', 'f'),
 155             (RFC2396_BASE, 'http://a/b/c/d;p?q', ''),
 156             ]:
 157             self.assertEqual(urlparse.urldefrag(url), (defrag, frag))
 158
 159 def test_main():
 160     test_support.run_unittest(UrlParseTestCase)
 161
 162 if __name__ == "__main__":
 163     test_main()