1#! /usr/bin/env python 2 3from test import test_support 4import unittest 5import urlparse 6 7RFC1808_BASE = "http://a/b/c/d;p?q#f" 8RFC2396_BASE = "http://a/b/c/d;p?q" 9RFC3986_BASE = 'http://a/b/c/d;p?q' 10SIMPLE_BASE = 'http://a/b/c/d' 11 12# A list of test cases. Each test case is a a two-tuple that contains 13# a string with the query and a dictionary with the expected result. 14 15parse_qsl_test_cases = [ 16 ("", []), 17 ("&", []), 18 ("&&", []), 19 ("=", [('', '')]), 20 ("=a", [('', 'a')]), 21 ("a", [('a', '')]), 22 ("a=", [('a', '')]), 23 ("a=", [('a', '')]), 24 ("&a=b", [('a', 'b')]), 25 ("a=a+b&b=b+c", [('a', 'a b'), ('b', 'b c')]), 26 ("a=1&a=2", [('a', '1'), ('a', '2')]), 27] 28 29class UrlParseTestCase(unittest.TestCase): 30 31 def checkRoundtrips(self, url, parsed, split): 32 result = urlparse.urlparse(url) 33 self.assertEqual(result, parsed) 34 t = (result.scheme, result.netloc, result.path, 35 result.params, result.query, result.fragment) 36 self.assertEqual(t, parsed) 37 # put it back together and it should be the same 38 result2 = urlparse.urlunparse(result) 39 self.assertEqual(result2, url) 40 self.assertEqual(result2, result.geturl()) 41 42 # the result of geturl() is a fixpoint; we can always parse it 43 # again to get the same result: 44 result3 = urlparse.urlparse(result.geturl()) 45 self.assertEqual(result3.geturl(), result.geturl()) 46 self.assertEqual(result3, result) 47 self.assertEqual(result3.scheme, result.scheme) 48 self.assertEqual(result3.netloc, result.netloc) 49 self.assertEqual(result3.path, result.path) 50 self.assertEqual(result3.params, result.params) 51 self.assertEqual(result3.query, result.query) 52 self.assertEqual(result3.fragment, result.fragment) 53 self.assertEqual(result3.username, result.username) 54 self.assertEqual(result3.password, result.password) 55 self.assertEqual(result3.hostname, result.hostname) 56 self.assertEqual(result3.port, result.port) 57 58 # check the roundtrip using urlsplit() as well 59 result = urlparse.urlsplit(url) 60 self.assertEqual(result, split) 61 t = (result.scheme, result.netloc, result.path, 62 result.query, result.fragment) 63 self.assertEqual(t, split) 64 result2 = urlparse.urlunsplit(result) 65 self.assertEqual(result2, url) 66 self.assertEqual(result2, result.geturl()) 67 68 # check the fixpoint property of re-parsing the result of geturl() 69 result3 = urlparse.urlsplit(result.geturl()) 70 self.assertEqual(result3.geturl(), result.geturl()) 71 self.assertEqual(result3, result) 72 self.assertEqual(result3.scheme, result.scheme) 73 self.assertEqual(result3.netloc, result.netloc) 74 self.assertEqual(result3.path, result.path) 75 self.assertEqual(result3.query, result.query) 76 self.assertEqual(result3.fragment, result.fragment) 77 self.assertEqual(result3.username, result.username) 78 self.assertEqual(result3.password, result.password) 79 self.assertEqual(result3.hostname, result.hostname) 80 self.assertEqual(result3.port, result.port) 81 82 def test_qsl(self): 83 for orig, expect in parse_qsl_test_cases: 84 result = urlparse.parse_qsl(orig, keep_blank_values=True) 85 self.assertEqual(result, expect, "Error parsing %s" % repr(orig)) 86 87 def test_roundtrips(self): 88 testcases = [ 89 ('file:///tmp/junk.txt', 90 ('file', '', '/tmp/junk.txt', '', '', ''), 91 ('file', '', '/tmp/junk.txt', '', '')), 92 ('imap://mail.python.org/mbox1', 93 ('imap', 'mail.python.org', '/mbox1', '', '', ''), 94 ('imap', 'mail.python.org', '/mbox1', '', '')), 95 ('mms://wms.sys.hinet.net/cts/Drama/09006251100.asf', 96 ('mms', 'wms.sys.hinet.net', '/cts/Drama/09006251100.asf', 97 '', '', ''), 98 ('mms', 'wms.sys.hinet.net', '/cts/Drama/09006251100.asf', 99 '', '')), 100 ('nfs://server/path/to/file.txt', 101 ('nfs', 'server', '/path/to/file.txt', '', '', ''), 102 ('nfs', 'server', '/path/to/file.txt', '', '')), 103 ('svn+ssh://svn.zope.org/repos/main/ZConfig/trunk/', 104 ('svn+ssh', 'svn.zope.org', '/repos/main/ZConfig/trunk/', 105 '', '', ''), 106 ('svn+ssh', 'svn.zope.org', '/repos/main/ZConfig/trunk/', 107 '', '')), 108 ('git+ssh://git@github.com/user/project.git', 109 ('git+ssh', 'git@github.com','/user/project.git', 110 '','',''), 111 ('git+ssh', 'git@github.com','/user/project.git', 112 '', '')) 113 ] 114 for url, parsed, split in testcases: 115 self.checkRoundtrips(url, parsed, split) 116 117 def test_http_roundtrips(self): 118 # urlparse.urlsplit treats 'http:' as an optimized special case, 119 # so we test both 'http:' and 'https:' in all the following. 120 # Three cheers for white box knowledge! 121 testcases = [ 122 ('://www.python.org', 123 ('www.python.org', '', '', '', ''), 124 ('www.python.org', '', '', '')), 125 ('://www.python.org#abc', 126 ('www.python.org', '', '', '', 'abc'), 127 ('www.python.org', '', '', 'abc')), 128 ('://www.python.org?q=abc', 129 ('www.python.org', '', '', 'q=abc', ''), 130 ('www.python.org', '', 'q=abc', '')), 131 ('://www.python.org/#abc', 132 ('www.python.org', '/', '', '', 'abc'), 133 ('www.python.org', '/', '', 'abc')), 134 ('://a/b/c/d;p?q#f', 135 ('a', '/b/c/d', 'p', 'q', 'f'), 136 ('a', '/b/c/d;p', 'q', 'f')), 137 ] 138 for scheme in ('http', 'https'): 139 for url, parsed, split in testcases: 140 url = scheme + url 141 parsed = (scheme,) + parsed 142 split = (scheme,) + split 143 self.checkRoundtrips(url, parsed, split) 144 145 def checkJoin(self, base, relurl, expected): 146 self.assertEqual(urlparse.urljoin(base, relurl), expected, 147 (base, relurl, expected)) 148 149 def test_unparse_parse(self): 150 for u in ['Python', './Python','x-newscheme://foo.com/stuff','x://y','x:/y','x:/','/',]: 151 self.assertEqual(urlparse.urlunsplit(urlparse.urlsplit(u)), u) 152 self.assertEqual(urlparse.urlunparse(urlparse.urlparse(u)), u) 153 154 def test_RFC1808(self): 155 # "normal" cases from RFC 1808: 156 self.checkJoin(RFC1808_BASE, 'g:h', 'g:h') 157 self.checkJoin(RFC1808_BASE, 'g', 'http://a/b/c/g') 158 self.checkJoin(RFC1808_BASE, './g', 'http://a/b/c/g') 159 self.checkJoin(RFC1808_BASE, 'g/', 'http://a/b/c/g/') 160 self.checkJoin(RFC1808_BASE, '/g', 'http://a/g') 161 self.checkJoin(RFC1808_BASE, '//g', 'http://g') 162 self.checkJoin(RFC1808_BASE, 'g?y', 'http://a/b/c/g?y') 163 self.checkJoin(RFC1808_BASE, 'g?y/./x', 'http://a/b/c/g?y/./x') 164 self.checkJoin(RFC1808_BASE, '#s', 'http://a/b/c/d;p?q#s') 165 self.checkJoin(RFC1808_BASE, 'g#s', 'http://a/b/c/g#s') 166 self.checkJoin(RFC1808_BASE, 'g#s/./x', 'http://a/b/c/g#s/./x') 167 self.checkJoin(RFC1808_BASE, 'g?y#s', 'http://a/b/c/g?y#s') 168 self.checkJoin(RFC1808_BASE, 'g;x', 'http://a/b/c/g;x') 169 self.checkJoin(RFC1808_BASE, 'g;x?y#s', 'http://a/b/c/g;x?y#s') 170 self.checkJoin(RFC1808_BASE, '.', 'http://a/b/c/') 171 self.checkJoin(RFC1808_BASE, './', 'http://a/b/c/') 172 self.checkJoin(RFC1808_BASE, '..', 'http://a/b/') 173 self.checkJoin(RFC1808_BASE, '../', 'http://a/b/') 174 self.checkJoin(RFC1808_BASE, '../g', 'http://a/b/g') 175 self.checkJoin(RFC1808_BASE, '../..', 'http://a/') 176 self.checkJoin(RFC1808_BASE, '../../', 'http://a/') 177 self.checkJoin(RFC1808_BASE, '../../g', 'http://a/g') 178 179 # "abnormal" cases from RFC 1808: 180 self.checkJoin(RFC1808_BASE, '', 'http://a/b/c/d;p?q#f') 181 self.checkJoin(RFC1808_BASE, '../../../g', 'http://a/../g') 182 self.checkJoin(RFC1808_BASE, '../../../../g', 'http://a/../../g') 183 self.checkJoin(RFC1808_BASE, '/./g', 'http://a/./g') 184 self.checkJoin(RFC1808_BASE, '/../g', 'http://a/../g') 185 self.checkJoin(RFC1808_BASE, 'g.', 'http://a/b/c/g.') 186 self.checkJoin(RFC1808_BASE, '.g', 'http://a/b/c/.g') 187 self.checkJoin(RFC1808_BASE, 'g..', 'http://a/b/c/g..') 188 self.checkJoin(RFC1808_BASE, '..g', 'http://a/b/c/..g') 189 self.checkJoin(RFC1808_BASE, './../g', 'http://a/b/g') 190 self.checkJoin(RFC1808_BASE, './g/.', 'http://a/b/c/g/') 191 self.checkJoin(RFC1808_BASE, 'g/./h', 'http://a/b/c/g/h') 192 self.checkJoin(RFC1808_BASE, 'g/../h', 'http://a/b/c/h') 193 194 # RFC 1808 and RFC 1630 disagree on these (according to RFC 1808), 195 # so we'll not actually run these tests (which expect 1808 behavior). 196 #self.checkJoin(RFC1808_BASE, 'http:g', 'http:g') 197 #self.checkJoin(RFC1808_BASE, 'http:', 'http:') 198 199 def test_RFC2368(self): 200 # Issue 11467: path that starts with a number is not parsed correctly 201 self.assertEqual(urlparse.urlparse('mailto:1337@example.org'), 202 ('mailto', '', '1337@example.org', '', '', '')) 203 204 def test_RFC2396(self): 205 # cases from RFC 2396 206 self.checkJoin(RFC2396_BASE, 'g:h', 'g:h') 207 self.checkJoin(RFC2396_BASE, 'g', 'http://a/b/c/g') 208 self.checkJoin(RFC2396_BASE, './g', 'http://a/b/c/g') 209 self.checkJoin(RFC2396_BASE, 'g/', 'http://a/b/c/g/') 210 self.checkJoin(RFC2396_BASE, '/g', 'http://a/g') 211 self.checkJoin(RFC2396_BASE, '//g', 'http://g') 212 self.checkJoin(RFC2396_BASE, 'g?y', 'http://a/b/c/g?y') 213 self.checkJoin(RFC2396_BASE, '#s', 'http://a/b/c/d;p?q#s') 214 self.checkJoin(RFC2396_BASE, 'g#s', 'http://a/b/c/g#s') 215 self.checkJoin(RFC2396_BASE, 'g?y#s', 'http://a/b/c/g?y#s') 216 self.checkJoin(RFC2396_BASE, 'g;x', 'http://a/b/c/g;x') 217 self.checkJoin(RFC2396_BASE, 'g;x?y#s', 'http://a/b/c/g;x?y#s') 218 self.checkJoin(RFC2396_BASE, '.', 'http://a/b/c/') 219 self.checkJoin(RFC2396_BASE, './', 'http://a/b/c/') 220 self.checkJoin(RFC2396_BASE, '..', 'http://a/b/') 221 self.checkJoin(RFC2396_BASE, '../', 'http://a/b/') 222 self.checkJoin(RFC2396_BASE, '../g', 'http://a/b/g') 223 self.checkJoin(RFC2396_BASE, '../..', 'http://a/') 224 self.checkJoin(RFC2396_BASE, '../../', 'http://a/') 225 self.checkJoin(RFC2396_BASE, '../../g', 'http://a/g') 226 self.checkJoin(RFC2396_BASE, '', RFC2396_BASE) 227 self.checkJoin(RFC2396_BASE, '../../../g', 'http://a/../g') 228 self.checkJoin(RFC2396_BASE, '../../../../g', 'http://a/../../g') 229 self.checkJoin(RFC2396_BASE, '/./g', 'http://a/./g') 230 self.checkJoin(RFC2396_BASE, '/../g', 'http://a/../g') 231 self.checkJoin(RFC2396_BASE, 'g.', 'http://a/b/c/g.') 232 self.checkJoin(RFC2396_BASE, '.g', 'http://a/b/c/.g') 233 self.checkJoin(RFC2396_BASE, 'g..', 'http://a/b/c/g..') 234 self.checkJoin(RFC2396_BASE, '..g', 'http://a/b/c/..g') 235 self.checkJoin(RFC2396_BASE, './../g', 'http://a/b/g') 236 self.checkJoin(RFC2396_BASE, './g/.', 'http://a/b/c/g/') 237 self.checkJoin(RFC2396_BASE, 'g/./h', 'http://a/b/c/g/h') 238 self.checkJoin(RFC2396_BASE, 'g/../h', 'http://a/b/c/h') 239 self.checkJoin(RFC2396_BASE, 'g;x=1/./y', 'http://a/b/c/g;x=1/y') 240 self.checkJoin(RFC2396_BASE, 'g;x=1/../y', 'http://a/b/c/y') 241 self.checkJoin(RFC2396_BASE, 'g?y/./x', 'http://a/b/c/g?y/./x') 242 self.checkJoin(RFC2396_BASE, 'g?y/../x', 'http://a/b/c/g?y/../x') 243 self.checkJoin(RFC2396_BASE, 'g#s/./x', 'http://a/b/c/g#s/./x') 244 self.checkJoin(RFC2396_BASE, 'g#s/../x', 'http://a/b/c/g#s/../x') 245 246 def test_RFC3986(self): 247 # Test cases from RFC3986 248 self.checkJoin(RFC3986_BASE, '?y','http://a/b/c/d;p?y') 249 self.checkJoin(RFC2396_BASE, ';x', 'http://a/b/c/;x') 250 self.checkJoin(RFC3986_BASE, 'g:h','g:h') 251 self.checkJoin(RFC3986_BASE, 'g','http://a/b/c/g') 252 self.checkJoin(RFC3986_BASE, './g','http://a/b/c/g') 253 self.checkJoin(RFC3986_BASE, 'g/','http://a/b/c/g/') 254 self.checkJoin(RFC3986_BASE, '/g','http://a/g') 255 self.checkJoin(RFC3986_BASE, '//g','http://g') 256 self.checkJoin(RFC3986_BASE, '?y','http://a/b/c/d;p?y') 257 self.checkJoin(RFC3986_BASE, 'g?y','http://a/b/c/g?y') 258 self.checkJoin(RFC3986_BASE, '#s','http://a/b/c/d;p?q#s') 259 self.checkJoin(RFC3986_BASE, 'g#s','http://a/b/c/g#s') 260 self.checkJoin(RFC3986_BASE, 'g?y#s','http://a/b/c/g?y#s') 261 self.checkJoin(RFC3986_BASE, ';x','http://a/b/c/;x') 262 self.checkJoin(RFC3986_BASE, 'g;x','http://a/b/c/g;x') 263 self.checkJoin(RFC3986_BASE, 'g;x?y#s','http://a/b/c/g;x?y#s') 264 self.checkJoin(RFC3986_BASE, '','http://a/b/c/d;p?q') 265 self.checkJoin(RFC3986_BASE, '.','http://a/b/c/') 266 self.checkJoin(RFC3986_BASE, './','http://a/b/c/') 267 self.checkJoin(RFC3986_BASE, '..','http://a/b/') 268 self.checkJoin(RFC3986_BASE, '../','http://a/b/') 269 self.checkJoin(RFC3986_BASE, '../g','http://a/b/g') 270 self.checkJoin(RFC3986_BASE, '../..','http://a/') 271 self.checkJoin(RFC3986_BASE, '../../','http://a/') 272 self.checkJoin(RFC3986_BASE, '../../g','http://a/g') 273 274 #Abnormal Examples 275 276 # The 'abnormal scenarios' are incompatible with RFC2986 parsing 277 # Tests are here for reference. 278 279 #self.checkJoin(RFC3986_BASE, '../../../g','http://a/g') 280 #self.checkJoin(RFC3986_BASE, '../../../../g','http://a/g') 281 #self.checkJoin(RFC3986_BASE, '/./g','http://a/g') 282 #self.checkJoin(RFC3986_BASE, '/../g','http://a/g') 283 284 self.checkJoin(RFC3986_BASE, 'g.','http://a/b/c/g.') 285 self.checkJoin(RFC3986_BASE, '.g','http://a/b/c/.g') 286 self.checkJoin(RFC3986_BASE, 'g..','http://a/b/c/g..') 287 self.checkJoin(RFC3986_BASE, '..g','http://a/b/c/..g') 288 self.checkJoin(RFC3986_BASE, './../g','http://a/b/g') 289 self.checkJoin(RFC3986_BASE, './g/.','http://a/b/c/g/') 290 self.checkJoin(RFC3986_BASE, 'g/./h','http://a/b/c/g/h') 291 self.checkJoin(RFC3986_BASE, 'g/../h','http://a/b/c/h') 292 self.checkJoin(RFC3986_BASE, 'g;x=1/./y','http://a/b/c/g;x=1/y') 293 self.checkJoin(RFC3986_BASE, 'g;x=1/../y','http://a/b/c/y') 294 self.checkJoin(RFC3986_BASE, 'g?y/./x','http://a/b/c/g?y/./x') 295 self.checkJoin(RFC3986_BASE, 'g?y/../x','http://a/b/c/g?y/../x') 296 self.checkJoin(RFC3986_BASE, 'g#s/./x','http://a/b/c/g#s/./x') 297 self.checkJoin(RFC3986_BASE, 'g#s/../x','http://a/b/c/g#s/../x') 298 #self.checkJoin(RFC3986_BASE, 'http:g','http:g') # strict parser 299 self.checkJoin(RFC3986_BASE, 'http:g','http://a/b/c/g') # relaxed parser 300 301 # Test for issue9721 302 self.checkJoin('http://a/b/c/de', ';x','http://a/b/c/;x') 303 304 def test_urljoins(self): 305 self.checkJoin(SIMPLE_BASE, 'g:h','g:h') 306 self.checkJoin(SIMPLE_BASE, 'http:g','http://a/b/c/g') 307 self.checkJoin(SIMPLE_BASE, 'http:','http://a/b/c/d') 308 self.checkJoin(SIMPLE_BASE, 'g','http://a/b/c/g') 309 self.checkJoin(SIMPLE_BASE, './g','http://a/b/c/g') 310 self.checkJoin(SIMPLE_BASE, 'g/','http://a/b/c/g/') 311 self.checkJoin(SIMPLE_BASE, '/g','http://a/g') 312 self.checkJoin(SIMPLE_BASE, '//g','http://g') 313 self.checkJoin(SIMPLE_BASE, '?y','http://a/b/c/d?y') 314 self.checkJoin(SIMPLE_BASE, 'g?y','http://a/b/c/g?y') 315 self.checkJoin(SIMPLE_BASE, 'g?y/./x','http://a/b/c/g?y/./x') 316 self.checkJoin(SIMPLE_BASE, '.','http://a/b/c/') 317 self.checkJoin(SIMPLE_BASE, './','http://a/b/c/') 318 self.checkJoin(SIMPLE_BASE, '..','http://a/b/') 319 self.checkJoin(SIMPLE_BASE, '../','http://a/b/') 320 self.checkJoin(SIMPLE_BASE, '../g','http://a/b/g') 321 self.checkJoin(SIMPLE_BASE, '../..','http://a/') 322 self.checkJoin(SIMPLE_BASE, '../../g','http://a/g') 323 self.checkJoin(SIMPLE_BASE, '../../../g','http://a/../g') 324 self.checkJoin(SIMPLE_BASE, './../g','http://a/b/g') 325 self.checkJoin(SIMPLE_BASE, './g/.','http://a/b/c/g/') 326 self.checkJoin(SIMPLE_BASE, '/./g','http://a/./g') 327 self.checkJoin(SIMPLE_BASE, 'g/./h','http://a/b/c/g/h') 328 self.checkJoin(SIMPLE_BASE, 'g/../h','http://a/b/c/h') 329 self.checkJoin(SIMPLE_BASE, 'http:g','http://a/b/c/g') 330 self.checkJoin(SIMPLE_BASE, 'http:','http://a/b/c/d') 331 self.checkJoin(SIMPLE_BASE, 'http:?y','http://a/b/c/d?y') 332 self.checkJoin(SIMPLE_BASE, 'http:g?y','http://a/b/c/g?y') 333 self.checkJoin(SIMPLE_BASE, 'http:g?y/./x','http://a/b/c/g?y/./x') 334 335 def test_RFC2732(self): 336 for url, hostname, port in [ 337 ('http://Test.python.org:5432/foo/', 'test.python.org', 5432), 338 ('http://12.34.56.78:5432/foo/', '12.34.56.78', 5432), 339 ('http://[::1]:5432/foo/', '::1', 5432), 340 ('http://[dead:beef::1]:5432/foo/', 'dead:beef::1', 5432), 341 ('http://[dead:beef::]:5432/foo/', 'dead:beef::', 5432), 342 ('http://[dead:beef:cafe:5417:affe:8FA3:deaf:feed]:5432/foo/', 343 'dead:beef:cafe:5417:affe:8fa3:deaf:feed', 5432), 344 ('http://[::12.34.56.78]:5432/foo/', '::12.34.56.78', 5432), 345 ('http://[::ffff:12.34.56.78]:5432/foo/', 346 '::ffff:12.34.56.78', 5432), 347 ('http://Test.python.org/foo/', 'test.python.org', None), 348 ('http://12.34.56.78/foo/', '12.34.56.78', None), 349 ('http://[::1]/foo/', '::1', None), 350 ('http://[dead:beef::1]/foo/', 'dead:beef::1', None), 351 ('http://[dead:beef::]/foo/', 'dead:beef::', None), 352 ('http://[dead:beef:cafe:5417:affe:8FA3:deaf:feed]/foo/', 353 'dead:beef:cafe:5417:affe:8fa3:deaf:feed', None), 354 ('http://[::12.34.56.78]/foo/', '::12.34.56.78', None), 355 ('http://[::ffff:12.34.56.78]/foo/', 356 '::ffff:12.34.56.78', None), 357 ]: 358 urlparsed = urlparse.urlparse(url) 359 self.assertEqual((urlparsed.hostname, urlparsed.port) , (hostname, port)) 360 361 for invalid_url in [ 362 'http://::12.34.56.78]/', 363 'http://[::1/foo/', 364 'ftp://[::1/foo/bad]/bad', 365 'http://[::1/foo/bad]/bad', 366 'http://[::ffff:12.34.56.78']: 367 self.assertRaises(ValueError, urlparse.urlparse, invalid_url) 368 369 def test_urldefrag(self): 370 for url, defrag, frag in [ 371 ('http://python.org#frag', 'http://python.org', 'frag'), 372 ('http://python.org', 'http://python.org', ''), 373 ('http://python.org/#frag', 'http://python.org/', 'frag'), 374 ('http://python.org/', 'http://python.org/', ''), 375 ('http://python.org/?q#frag', 'http://python.org/?q', 'frag'), 376 ('http://python.org/?q', 'http://python.org/?q', ''), 377 ('http://python.org/p#frag', 'http://python.org/p', 'frag'), 378 ('http://python.org/p?q', 'http://python.org/p?q', ''), 379 (RFC1808_BASE, 'http://a/b/c/d;p?q', 'f'), 380 (RFC2396_BASE, 'http://a/b/c/d;p?q', ''), 381 ]: 382 self.assertEqual(urlparse.urldefrag(url), (defrag, frag)) 383 384 def test_urlsplit_attributes(self): 385 url = "HTTP://WWW.PYTHON.ORG/doc/#frag" 386 p = urlparse.urlsplit(url) 387 self.assertEqual(p.scheme, "http") 388 self.assertEqual(p.netloc, "WWW.PYTHON.ORG") 389 self.assertEqual(p.path, "/doc/") 390 self.assertEqual(p.query, "") 391 self.assertEqual(p.fragment, "frag") 392 self.assertEqual(p.username, None) 393 self.assertEqual(p.password, None) 394 self.assertEqual(p.hostname, "www.python.org") 395 self.assertEqual(p.port, None) 396 # geturl() won't return exactly the original URL in this case 397 # since the scheme is always case-normalized 398 #self.assertEqual(p.geturl(), url) 399 400 url = "http://User:Pass@www.python.org:080/doc/?query=yes#frag" 401 p = urlparse.urlsplit(url) 402 self.assertEqual(p.scheme, "http") 403 self.assertEqual(p.netloc, "User:Pass@www.python.org:080") 404 self.assertEqual(p.path, "/doc/") 405 self.assertEqual(p.query, "query=yes") 406 self.assertEqual(p.fragment, "frag") 407 self.assertEqual(p.username, "User") 408 self.assertEqual(p.password, "Pass") 409 self.assertEqual(p.hostname, "www.python.org") 410 self.assertEqual(p.port, 80) 411 self.assertEqual(p.geturl(), url) 412 413 # Addressing issue1698, which suggests Username can contain 414 # "@" characters. Though not RFC compliant, many ftp sites allow 415 # and request email addresses as usernames. 416 417 url = "http://User@example.com:Pass@www.python.org:080/doc/?query=yes#frag" 418 p = urlparse.urlsplit(url) 419 self.assertEqual(p.scheme, "http") 420 self.assertEqual(p.netloc, "User@example.com:Pass@www.python.org:080") 421 self.assertEqual(p.path, "/doc/") 422 self.assertEqual(p.query, "query=yes") 423 self.assertEqual(p.fragment, "frag") 424 self.assertEqual(p.username, "User@example.com") 425 self.assertEqual(p.password, "Pass") 426 self.assertEqual(p.hostname, "www.python.org") 427 self.assertEqual(p.port, 80) 428 self.assertEqual(p.geturl(), url) 429 430 431 def test_attributes_bad_port(self): 432 """Check handling of non-integer ports.""" 433 p = urlparse.urlsplit("http://www.example.net:foo") 434 self.assertEqual(p.netloc, "www.example.net:foo") 435 self.assertRaises(ValueError, lambda: p.port) 436 437 p = urlparse.urlparse("http://www.example.net:foo") 438 self.assertEqual(p.netloc, "www.example.net:foo") 439 self.assertRaises(ValueError, lambda: p.port) 440 441 def test_attributes_without_netloc(self): 442 # This example is straight from RFC 3261. It looks like it 443 # should allow the username, hostname, and port to be filled 444 # in, but doesn't. Since it's a URI and doesn't use the 445 # scheme://netloc syntax, the netloc and related attributes 446 # should be left empty. 447 uri = "sip:alice@atlanta.com;maddr=239.255.255.1;ttl=15" 448 p = urlparse.urlsplit(uri) 449 self.assertEqual(p.netloc, "") 450 self.assertEqual(p.username, None) 451 self.assertEqual(p.password, None) 452 self.assertEqual(p.hostname, None) 453 self.assertEqual(p.port, None) 454 self.assertEqual(p.geturl(), uri) 455 456 p = urlparse.urlparse(uri) 457 self.assertEqual(p.netloc, "") 458 self.assertEqual(p.username, None) 459 self.assertEqual(p.password, None) 460 self.assertEqual(p.hostname, None) 461 self.assertEqual(p.port, None) 462 self.assertEqual(p.geturl(), uri) 463 464 def test_caching(self): 465 # Test case for bug #1313119 466 uri = "http://example.com/doc/" 467 unicode_uri = unicode(uri) 468 469 urlparse.urlparse(unicode_uri) 470 p = urlparse.urlparse(uri) 471 self.assertEqual(type(p.scheme), type(uri)) 472 self.assertEqual(type(p.hostname), type(uri)) 473 self.assertEqual(type(p.path), type(uri)) 474 475 def test_noslash(self): 476 # Issue 1637: http://foo.com?query is legal 477 self.assertEqual(urlparse.urlparse("http://example.com?blahblah=/foo"), 478 ('http', 'example.com', '', '', 'blahblah=/foo', '')) 479 480 def test_anyscheme(self): 481 # Issue 7904: s3://foo.com/stuff has netloc "foo.com". 482 self.assertEqual(urlparse.urlparse("s3://foo.com/stuff"), 483 ('s3','foo.com','/stuff','','','')) 484 self.assertEqual(urlparse.urlparse("x-newscheme://foo.com/stuff"), 485 ('x-newscheme','foo.com','/stuff','','','')) 486 487 def test_withoutscheme(self): 488 # Test urlparse without scheme 489 # Issue 754016: urlparse goes wrong with IP:port without scheme 490 # RFC 1808 specifies that netloc should start with //, urlparse expects 491 # the same, otherwise it classifies the portion of url as path. 492 self.assertEqual(urlparse.urlparse("path"), 493 ('','','path','','','')) 494 self.assertEqual(urlparse.urlparse("//www.python.org:80"), 495 ('','www.python.org:80','','','','')) 496 self.assertEqual(urlparse.urlparse("http://www.python.org:80"), 497 ('http','www.python.org:80','','','','')) 498 499 def test_portseparator(self): 500 # Issue 754016 makes changes for port separator ':' from scheme separator 501 self.assertEqual(urlparse.urlparse("path:80"), 502 ('','','path:80','','','')) 503 self.assertEqual(urlparse.urlparse("http:"),('http','','','','','')) 504 self.assertEqual(urlparse.urlparse("https:"),('https','','','','','')) 505 self.assertEqual(urlparse.urlparse("http://www.python.org:80"), 506 ('http','www.python.org:80','','','','')) 507 508 509def test_main(): 510 test_support.run_unittest(UrlParseTestCase) 511 512if __name__ == "__main__": 513 test_main() 514