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
  12     def checkRoundtrips(self, url, parsed, split):
  13         result = urlparse.urlparse(url)
  14         self.assertEqual(result, parsed)
  15         # put it back together and it should be the same
  16         result2 = urlparse.urlunparse(result)
  17         self.assertEqual(result2, url)
  18
  19         # check the roundtrip using urlsplit() as well
  20         result = urlparse.urlsplit(url)
  21         self.assertEqual(result, split)
  22         result2 = urlparse.urlunsplit(result)
  23         self.assertEqual(result2, url)
  24
  25     def test_roundtrips(self):
  26         testcases = [
  27             ('file:///tmp/junk.txt',
  28              ('file', '', '/tmp/junk.txt', '', '', ''),
  29              ('file', '', '/tmp/junk.txt', '', '')),
  30             ('imap://mail.python.org/mbox1',
  31              ('imap', 'mail.python.org', '/mbox1', '', '', ''),
  32              ('imap', 'mail.python.org', '/mbox1', '', '')),
  33             ('mms://wms.sys.hinet.net/cts/Drama/09006251100.asf',
  34              ('mms', 'wms.sys.hinet.net', '/cts/Drama/09006251100.asf',
  35               '', '', ''),
  36              ('mms', 'wms.sys.hinet.net', '/cts/Drama/09006251100.asf',
  37               '', '')),
  38             ('svn+ssh://svn.zope.org/repos/main/ZConfig/trunk/',
  39              ('svn+ssh', 'svn.zope.org', '/repos/main/ZConfig/trunk/',
  40               '', '', ''),
  41              ('svn+ssh', 'svn.zope.org', '/repos/main/ZConfig/trunk/',
  42               '', ''))
  43             ]
  44         for url, parsed, split in testcases:
  45             self.checkRoundtrips(url, parsed, split)
  46
  47     def test_http_roundtrips(self):
  48         # urlparse.urlsplit treats 'http:' as an optimized special case,
  49         # so we test both 'http:' and 'https:' in all the following.
  50         # Three cheers for white box knowledge!
  51         testcases = [
  52             ('://www.python.org',
  53              ('www.python.org', '', '', '', ''),
  54              ('www.python.org', '', '', '')),
  55             ('://www.python.org#abc',
  56              ('www.python.org', '', '', '', 'abc'),
  57              ('www.python.org', '', '', 'abc')),
  58             ('://www.python.org?q=abc',
  59              ('www.python.org', '', '', 'q=abc', ''),
  60              ('www.python.org', '', 'q=abc', '')),
  61             ('://www.python.org/#abc',
  62              ('www.python.org', '/', '', '', 'abc'),
  63              ('www.python.org', '/', '', 'abc')),
  64             ('://a/b/c/d;p?q#f',
  65              ('a', '/b/c/d', 'p', 'q', 'f'),
  66              ('a', '/b/c/d;p', 'q', 'f')),
  67             ]
  68         for scheme in ('http', 'https'):
  69             for url, parsed, split in testcases:
  70                 url = scheme + url
  71                 parsed = (scheme,) + parsed
  72                 split = (scheme,) + split
  73                 self.checkRoundtrips(url, parsed, split)
  74
  75     def checkJoin(self, base, relurl, expected):
  76         self.assertEqual(urlparse.urljoin(base, relurl), expected,
  77                          (base, relurl, expected))
  78
  79     def test_unparse_parse(self):
  80         for u in ['Python', './Python']:
  81             self.assertEqual(urlparse.urlunsplit(urlparse.urlsplit(u)), u)
  82             self.assertEqual(urlparse.urlunparse(urlparse.urlparse(u)), u)
  83
  84     def test_RFC1808(self):
  85         # "normal" cases from RFC 1808:
  86         self.checkJoin(RFC1808_BASE, 'g:h', 'g:h')
  87         self.checkJoin(RFC1808_BASE, 'g', 'http://a/b/c/g')
  88         self.checkJoin(RFC1808_BASE, './g', 'http://a/b/c/g')
  89         self.checkJoin(RFC1808_BASE, 'g/', 'http://a/b/c/g/')
  90         self.checkJoin(RFC1808_BASE, '/g', 'http://a/g')
  91         self.checkJoin(RFC1808_BASE, '//g', 'http://g')
  92         self.checkJoin(RFC1808_BASE, 'g?y', 'http://a/b/c/g?y')
  93         self.checkJoin(RFC1808_BASE, 'g?y/./x', 'http://a/b/c/g?y/./x')
  94         self.checkJoin(RFC1808_BASE, '#s', 'http://a/b/c/d;p?q#s')
  95         self.checkJoin(RFC1808_BASE, 'g#s', 'http://a/b/c/g#s')
  96         self.checkJoin(RFC1808_BASE, 'g#s/./x', 'http://a/b/c/g#s/./x')
  97         self.checkJoin(RFC1808_BASE, 'g?y#s', 'http://a/b/c/g?y#s')
  98         self.checkJoin(RFC1808_BASE, 'g;x', 'http://a/b/c/g;x')
  99         self.checkJoin(RFC1808_BASE, 'g;x?y#s', 'http://a/b/c/g;x?y#s')
 100         self.checkJoin(RFC1808_BASE, '.', 'http://a/b/c/')
 101         self.checkJoin(RFC1808_BASE, './', 'http://a/b/c/')
 102         self.checkJoin(RFC1808_BASE, '..', 'http://a/b/')
 103         self.checkJoin(RFC1808_BASE, '../', 'http://a/b/')
 104         self.checkJoin(RFC1808_BASE, '../g', 'http://a/b/g')
 105         self.checkJoin(RFC1808_BASE, '../..', 'http://a/')
 106         self.checkJoin(RFC1808_BASE, '../../', 'http://a/')
 107         self.checkJoin(RFC1808_BASE, '../../g', 'http://a/g')
 108
 109         # "abnormal" cases from RFC 1808:
 110         self.checkJoin(RFC1808_BASE, '', 'http://a/b/c/d;p?q#f')
 111         self.checkJoin(RFC1808_BASE, '../../../g', 'http://a/../g')
 112         self.checkJoin(RFC1808_BASE, '../../../../g', 'http://a/../../g')
 113         self.checkJoin(RFC1808_BASE, '/./g', 'http://a/./g')
 114         self.checkJoin(RFC1808_BASE, '/../g', 'http://a/../g')
 115         self.checkJoin(RFC1808_BASE, 'g.', 'http://a/b/c/g.')
 116         self.checkJoin(RFC1808_BASE, '.g', 'http://a/b/c/.g')
 117         self.checkJoin(RFC1808_BASE, 'g..', 'http://a/b/c/g..')
 118         self.checkJoin(RFC1808_BASE, '..g', 'http://a/b/c/..g')
 119         self.checkJoin(RFC1808_BASE, './../g', 'http://a/b/g')
 120         self.checkJoin(RFC1808_BASE, './g/.', 'http://a/b/c/g/')
 121         self.checkJoin(RFC1808_BASE, 'g/./h', 'http://a/b/c/g/h')
 122         self.checkJoin(RFC1808_BASE, 'g/../h', 'http://a/b/c/h')
 123
 124         # RFC 1808 and RFC 1630 disagree on these (according to RFC 1808),
 125         # so we'll not actually run these tests (which expect 1808 behavior).
 126         #self.checkJoin(RFC1808_BASE, 'http:g', 'http:g')
 127         #self.checkJoin(RFC1808_BASE, 'http:', 'http:')
 128
 129     def test_RFC2396(self):
 130         # cases from RFC 2396
 131
 132         self.checkJoin(RFC2396_BASE, '?y', 'http://a/b/c/?y')
 133         self.checkJoin(RFC2396_BASE, ';x', 'http://a/b/c/;x')
 134
 135         self.checkJoin(RFC2396_BASE, 'g:h', 'g:h')
 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/g')
 140         self.checkJoin(RFC2396_BASE, '//g', 'http://g')
 141         self.checkJoin(RFC2396_BASE, 'g?y', 'http://a/b/c/g?y')
 142         self.checkJoin(RFC2396_BASE, '#s', 'http://a/b/c/d;p?q#s')
 143         self.checkJoin(RFC2396_BASE, 'g#s', 'http://a/b/c/g#s')
 144         self.checkJoin(RFC2396_BASE, 'g?y#s', 'http://a/b/c/g?y#s')
 145         self.checkJoin(RFC2396_BASE, 'g;x', 'http://a/b/c/g;x')
 146         self.checkJoin(RFC2396_BASE, 'g;x?y#s', 'http://a/b/c/g;x?y#s')
 147         self.checkJoin(RFC2396_BASE, '.', 'http://a/b/c/')
 148         self.checkJoin(RFC2396_BASE, './', 'http://a/b/c/')
 149         self.checkJoin(RFC2396_BASE, '..', 'http://a/b/')
 150         self.checkJoin(RFC2396_BASE, '../', 'http://a/b/')
 151         self.checkJoin(RFC2396_BASE, '../g', 'http://a/b/g')
 152         self.checkJoin(RFC2396_BASE, '../..', 'http://a/')
 153         self.checkJoin(RFC2396_BASE, '../../', 'http://a/')
 154         self.checkJoin(RFC2396_BASE, '../../g', 'http://a/g')
 155         self.checkJoin(RFC2396_BASE, '', RFC2396_BASE)
 156         self.checkJoin(RFC2396_BASE, '../../../g', 'http://a/../g')
 157         self.checkJoin(RFC2396_BASE, '../../../../g', 'http://a/../../g')
 158         self.checkJoin(RFC2396_BASE, '/./g', 'http://a/./g')
 159         self.checkJoin(RFC2396_BASE, '/../g', 'http://a/../g')
 160         self.checkJoin(RFC2396_BASE, 'g.', 'http://a/b/c/g.')
 161         self.checkJoin(RFC2396_BASE, '.g', 'http://a/b/c/.g')
 162         self.checkJoin(RFC2396_BASE, 'g..', 'http://a/b/c/g..')
 163         self.checkJoin(RFC2396_BASE, '..g', 'http://a/b/c/..g')
 164         self.checkJoin(RFC2396_BASE, './../g', 'http://a/b/g')
 165         self.checkJoin(RFC2396_BASE, './g/.', 'http://a/b/c/g/')
 166         self.checkJoin(RFC2396_BASE, 'g/./h', 'http://a/b/c/g/h')
 167         self.checkJoin(RFC2396_BASE, 'g/../h', 'http://a/b/c/h')
 168         self.checkJoin(RFC2396_BASE, 'g;x=1/./y', 'http://a/b/c/g;x=1/y')
 169         self.checkJoin(RFC2396_BASE, 'g;x=1/../y', 'http://a/b/c/y')
 170         self.checkJoin(RFC2396_BASE, 'g?y/./x', 'http://a/b/c/g?y/./x')
 171         self.checkJoin(RFC2396_BASE, 'g?y/../x', 'http://a/b/c/g?y/../x')
 172         self.checkJoin(RFC2396_BASE, 'g#s/./x', 'http://a/b/c/g#s/./x')
 173         self.checkJoin(RFC2396_BASE, 'g#s/../x', 'http://a/b/c/g#s/../x')
 174
 175     def test_urldefrag(self):
 176         for url, defrag, frag in [
 177             ('http://python.org#frag', 'http://python.org', 'frag'),
 178             ('http://python.org', 'http://python.org', ''),
 179             ('http://python.org/#frag', 'http://python.org/', 'frag'),
 180             ('http://python.org/', 'http://python.org/', ''),
 181             ('http://python.org/?q#frag', 'http://python.org/?q', 'frag'),
 182             ('http://python.org/?q', 'http://python.org/?q', ''),
 183             ('http://python.org/p#frag', 'http://python.org/p', 'frag'),
 184             ('http://python.org/p?q', 'http://python.org/p?q', ''),
 185             (RFC1808_BASE, 'http://a/b/c/d;p?q', 'f'),
 186             (RFC2396_BASE, 'http://a/b/c/d;p?q', ''),
 187             ]:
 188             self.assertEqual(urlparse.urldefrag(url), (defrag, frag))
 189
 190 def test_main():
 191     test_support.run_unittest(UrlParseTestCase)
 192
 193 if __name__ == "__main__":
 194     test_main()