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             ('imap://mail.python.org/mbox1',
  29              ('imap', 'mail.python.org', '/mbox1', '', '', ''),
  30              ('imap', 'mail.python.org', '/mbox1', '', '')),
  31             ('mms://wms.sys.hinet.net/cts/Drama/09006251100.asf',
  32              ('mms', 'wms.sys.hinet.net', '/cts/Drama/09006251100.asf', '', '', ''),
  33              ('mms', 'wms.sys.hinet.net', '/cts/Drama/09006251100.asf', '', '')),
  34             ]:
  35             result = urlparse.urlparse(url)
  36             self.assertEqual(result, parsed)
  37             # put it back together and it should be the same
  38             result2 = urlparse.urlunparse(result)
  39             self.assertEqual(result2, url)
  40
  41             # check the roundtrip using urlsplit() as well
  42             result = urlparse.urlsplit(url)
  43             self.assertEqual(result, split)
  44             result2 = urlparse.urlunsplit(result)
  45             self.assertEqual(result2, url)
  46
  47     def checkJoin(self, base, relurl, expected):
  48         self.assertEqual(urlparse.urljoin(base, relurl), expected,
  49                          (base, relurl, expected))
  50
  51     def test_unparse_parse(self):
  52         for u in ['Python', './Python']:
  53             self.assertEqual(urlparse.urlunsplit(urlparse.urlsplit(u)), u)
  54             self.assertEqual(urlparse.urlunparse(urlparse.urlparse(u)), u)
  55
  56     def test_RFC1808(self):
  57         # "normal" cases from RFC 1808:
  58         self.checkJoin(RFC1808_BASE, 'g:h', 'g:h')
  59         self.checkJoin(RFC1808_BASE, 'g', 'http://a/b/c/g')
  60         self.checkJoin(RFC1808_BASE, './g', 'http://a/b/c/g')
  61         self.checkJoin(RFC1808_BASE, 'g/', 'http://a/b/c/g/')
  62         self.checkJoin(RFC1808_BASE, '/g', 'http://a/g')
  63         self.checkJoin(RFC1808_BASE, '//g', 'http://g')
  64         self.checkJoin(RFC1808_BASE, '?y', 'http://a/b/c/d;p?y')
  65         self.checkJoin(RFC1808_BASE, 'g?y', 'http://a/b/c/g?y')
  66         self.checkJoin(RFC1808_BASE, 'g?y/./x', 'http://a/b/c/g?y/./x')
  67         self.checkJoin(RFC1808_BASE, '#s', 'http://a/b/c/d;p?q#s')
  68         self.checkJoin(RFC1808_BASE, 'g#s', 'http://a/b/c/g#s')
  69         self.checkJoin(RFC1808_BASE, 'g#s/./x', 'http://a/b/c/g#s/./x')
  70         self.checkJoin(RFC1808_BASE, 'g?y#s', 'http://a/b/c/g?y#s')
  71         self.checkJoin(RFC1808_BASE, ';x', 'http://a/b/c/d;x')
  72         self.checkJoin(RFC1808_BASE, 'g;x', 'http://a/b/c/g;x')
  73         self.checkJoin(RFC1808_BASE, 'g;x?y#s', 'http://a/b/c/g;x?y#s')
  74         self.checkJoin(RFC1808_BASE, '.', 'http://a/b/c/')
  75         self.checkJoin(RFC1808_BASE, './', 'http://a/b/c/')
  76         self.checkJoin(RFC1808_BASE, '..', 'http://a/b/')
  77         self.checkJoin(RFC1808_BASE, '../', 'http://a/b/')
  78         self.checkJoin(RFC1808_BASE, '../g', 'http://a/b/g')
  79         self.checkJoin(RFC1808_BASE, '../..', 'http://a/')
  80         self.checkJoin(RFC1808_BASE, '../../', 'http://a/')
  81         self.checkJoin(RFC1808_BASE, '../../g', 'http://a/g')
  82
  83         # "abnormal" cases from RFC 1808:
  84         self.checkJoin(RFC1808_BASE, '', 'http://a/b/c/d;p?q#f')
  85         self.checkJoin(RFC1808_BASE, '../../../g', 'http://a/../g')
  86         self.checkJoin(RFC1808_BASE, '../../../../g', 'http://a/../../g')
  87         self.checkJoin(RFC1808_BASE, '/./g', 'http://a/./g')
  88         self.checkJoin(RFC1808_BASE, '/../g', 'http://a/../g')
  89         self.checkJoin(RFC1808_BASE, 'g.', 'http://a/b/c/g.')
  90         self.checkJoin(RFC1808_BASE, '.g', 'http://a/b/c/.g')
  91         self.checkJoin(RFC1808_BASE, 'g..', 'http://a/b/c/g..')
  92         self.checkJoin(RFC1808_BASE, '..g', 'http://a/b/c/..g')
  93         self.checkJoin(RFC1808_BASE, './../g', 'http://a/b/g')
  94         self.checkJoin(RFC1808_BASE, './g/.', 'http://a/b/c/g/')
  95         self.checkJoin(RFC1808_BASE, 'g/./h', 'http://a/b/c/g/h')
  96         self.checkJoin(RFC1808_BASE, 'g/../h', 'http://a/b/c/h')
  97
  98         # RFC 1808 and RFC 1630 disagree on these (according to RFC 1808),
  99         # so we'll not actually run these tests (which expect 1808 behavior).
 100         #self.checkJoin(RFC1808_BASE, 'http:g', 'http:g')
 101         #self.checkJoin(RFC1808_BASE, 'http:', 'http:')
 102
 103     def test_RFC2396(self):
 104         # cases from RFC 2396
 105
 106         ### urlparse.py as of v 1.32 fails on these two
 107         #self.checkJoin(RFC2396_BASE, '?y', 'http://a/b/c/?y')
 108         #self.checkJoin(RFC2396_BASE, ';x', 'http://a/b/c/;x')
 109
 110         self.checkJoin(RFC2396_BASE, 'g:h', 'g:h')
 111         self.checkJoin(RFC2396_BASE, 'g', 'http://a/b/c/g')
 112         self.checkJoin(RFC2396_BASE, './g', 'http://a/b/c/g')
 113         self.checkJoin(RFC2396_BASE, 'g/', 'http://a/b/c/g/')
 114         self.checkJoin(RFC2396_BASE, '/g', 'http://a/g')
 115         self.checkJoin(RFC2396_BASE, '//g', 'http://g')
 116         self.checkJoin(RFC2396_BASE, 'g?y', 'http://a/b/c/g?y')
 117         self.checkJoin(RFC2396_BASE, '#s', 'http://a/b/c/d;p?q#s')
 118         self.checkJoin(RFC2396_BASE, 'g#s', 'http://a/b/c/g#s')
 119         self.checkJoin(RFC2396_BASE, 'g?y#s', 'http://a/b/c/g?y#s')
 120         self.checkJoin(RFC2396_BASE, 'g;x', 'http://a/b/c/g;x')
 121         self.checkJoin(RFC2396_BASE, 'g;x?y#s', 'http://a/b/c/g;x?y#s')
 122         self.checkJoin(RFC2396_BASE, '.', 'http://a/b/c/')
 123         self.checkJoin(RFC2396_BASE, './', 'http://a/b/c/')
 124         self.checkJoin(RFC2396_BASE, '..', 'http://a/b/')
 125         self.checkJoin(RFC2396_BASE, '../', 'http://a/b/')
 126         self.checkJoin(RFC2396_BASE, '../g', 'http://a/b/g')
 127         self.checkJoin(RFC2396_BASE, '../..', 'http://a/')
 128         self.checkJoin(RFC2396_BASE, '../../', 'http://a/')
 129         self.checkJoin(RFC2396_BASE, '../../g', 'http://a/g')
 130         self.checkJoin(RFC2396_BASE, '', RFC2396_BASE)
 131         self.checkJoin(RFC2396_BASE, '../../../g', 'http://a/../g')
 132         self.checkJoin(RFC2396_BASE, '../../../../g', 'http://a/../../g')
 133         self.checkJoin(RFC2396_BASE, '/./g', 'http://a/./g')
 134         self.checkJoin(RFC2396_BASE, '/../g', 'http://a/../g')
 135         self.checkJoin(RFC2396_BASE, 'g.', 'http://a/b/c/g.')
 136         self.checkJoin(RFC2396_BASE, '.g', 'http://a/b/c/.g')
 137         self.checkJoin(RFC2396_BASE, 'g..', 'http://a/b/c/g..')
 138         self.checkJoin(RFC2396_BASE, '..g', 'http://a/b/c/..g')
 139         self.checkJoin(RFC2396_BASE, './../g', 'http://a/b/g')
 140         self.checkJoin(RFC2396_BASE, './g/.', 'http://a/b/c/g/')
 141         self.checkJoin(RFC2396_BASE, 'g/./h', 'http://a/b/c/g/h')
 142         self.checkJoin(RFC2396_BASE, 'g/../h', 'http://a/b/c/h')
 143         self.checkJoin(RFC2396_BASE, 'g;x=1/./y', 'http://a/b/c/g;x=1/y')
 144         self.checkJoin(RFC2396_BASE, 'g;x=1/../y', 'http://a/b/c/y')
 145         self.checkJoin(RFC2396_BASE, 'g?y/./x', 'http://a/b/c/g?y/./x')
 146         self.checkJoin(RFC2396_BASE, 'g?y/../x', 'http://a/b/c/g?y/../x')
 147         self.checkJoin(RFC2396_BASE, 'g#s/./x', 'http://a/b/c/g#s/./x')
 148         self.checkJoin(RFC2396_BASE, 'g#s/../x', 'http://a/b/c/g#s/../x')
 149
 150     def test_urldefrag(self):
 151         for url, defrag, frag in [
 152             ('http://python.org#frag', 'http://python.org', 'frag'),
 153             ('http://python.org', 'http://python.org', ''),
 154             ('http://python.org/#frag', 'http://python.org/', 'frag'),
 155             ('http://python.org/', 'http://python.org/', ''),
 156             ('http://python.org/?q#frag', 'http://python.org/?q', 'frag'),
 157             ('http://python.org/?q', 'http://python.org/?q', ''),
 158             ('http://python.org/p#frag', 'http://python.org/p', 'frag'),
 159             ('http://python.org/p?q', 'http://python.org/p?q', ''),
 160             (RFC1808_BASE, 'http://a/b/c/d;p?q', 'f'),
 161             (RFC2396_BASE, 'http://a/b/c/d;p?q', ''),
 162             ]:
 163             self.assertEqual(urlparse.urldefrag(url), (defrag, frag))
 164
 165 def test_main():
 166     test_support.run_unittest(UrlParseTestCase)
 167
 168 if __name__ == "__main__":
 169     test_main()