1
# Copyright (C) 2005 by Canonical Ltd
3
# This program is free software; you can redistribute it and/or modify
4
# it under the terms of the GNU General Public License as published by
5
# the Free Software Foundation; either version 2 of the License, or
6
# (at your option) any later version.
8
# This program is distributed in the hope that it will be useful,
9
# but WITHOUT ANY WARRANTY; without even the implied warranty of
10
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11
# GNU General Public License for more details.
13
# You should have received a copy of the GNU General Public License
14
# along with this program; if not, write to the Free Software
15
# Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
17
"""Tests for the urlutils wrapper."""
23
from bzrlib.errors import InvalidURL, InvalidURLJoin
24
import bzrlib.urlutils as urlutils
25
from bzrlib.tests import TestCaseInTempDir, TestCase, TestSkipped
28
class TestUrlToPath(TestCase):
30
def test_basename(self):
31
# bzrlib.urlutils.basename
32
# Test bzrlib.urlutils.split()
33
basename = urlutils.basename
34
if sys.platform == 'win32':
35
self.assertRaises(InvalidURL, basename, 'file:///path/to/foo')
36
self.assertEqual('foo', basename('file:///C|/foo'))
37
self.assertEqual('', basename('file:///C|/'))
39
self.assertEqual('foo', basename('file:///foo'))
40
self.assertEqual('', basename('file:///'))
42
self.assertEqual('foo', basename('http://host/path/to/foo'))
43
self.assertEqual('foo', basename('http://host/path/to/foo/'))
45
basename('http://host/path/to/foo/', exclude_trailing_slash=False))
46
self.assertEqual('path', basename('http://host/path'))
47
self.assertEqual('', basename('http://host/'))
48
self.assertEqual('', basename('http://host'))
49
self.assertEqual('path', basename('http:///nohost/path'))
51
self.assertEqual('path', basename('random+scheme://user:pass@ahost:port/path'))
52
self.assertEqual('path', basename('random+scheme://user:pass@ahost:port/path/'))
53
self.assertEqual('', basename('random+scheme://user:pass@ahost:port/'))
56
self.assertEqual('foo', basename('path/to/foo'))
57
self.assertEqual('foo', basename('path/to/foo/'))
58
self.assertEqual('', basename('path/to/foo/',
59
exclude_trailing_slash=False))
60
self.assertEqual('foo', basename('path/../foo'))
61
self.assertEqual('foo', basename('../path/foo'))
63
def test_normalize_url_files(self):
64
# Test that local paths are properly normalized
65
normalize_url = urlutils.normalize_url
67
def norm_file(expected, path):
68
url = normalize_url(path)
69
self.assertStartsWith(url, 'file:///')
70
if sys.platform == 'win32':
71
url = url[len('file:///C:'):]
73
url = url[len('file://'):]
75
self.assertEndsWith(url, expected)
77
norm_file('path/to/foo', 'path/to/foo')
78
norm_file('/path/to/foo', '/path/to/foo')
79
norm_file('path/to/foo', '../path/to/foo')
81
# Local paths are assumed to *not* be escaped at all
83
u'uni/\xb5'.encode(bzrlib.user_encoding)
85
# locale cannot handle unicode
88
norm_file('uni/%C2%B5', u'uni/\xb5')
90
norm_file('uni/%25C2%25B5', u'uni/%C2%B5')
91
norm_file('uni/%20b', u'uni/ b')
92
# All the crazy characters get escaped in local paths => file:/// urls
93
norm_file('%27%3B/%3F%3A%40%26%3D%2B%24%2C%23%20', "';/?:@&=+$,# ")
95
def test_normalize_url_hybrid(self):
96
# Anything with a scheme:// should be treated as a hybrid url
97
# which changes what characters get escaped.
98
normalize_url = urlutils.normalize_url
100
eq = self.assertEqual
101
eq('file:///foo/', normalize_url(u'file:///foo/'))
102
eq('file:///foo/%20', normalize_url(u'file:///foo/ '))
103
eq('file:///foo/%20', normalize_url(u'file:///foo/%20'))
104
# Don't escape reserved characters
105
eq('file:///ab_c.d-e/%f:?g&h=i+j;k,L#M$',
106
normalize_url('file:///ab_c.d-e/%f:?g&h=i+j;k,L#M$'))
107
eq('http://ab_c.d-e/%f:?g&h=i+j;k,L#M$',
108
normalize_url('http://ab_c.d-e/%f:?g&h=i+j;k,L#M$'))
110
# Escape unicode characters, but not already escaped chars
111
eq('http://host/ab/%C2%B5/%C2%B5',
112
normalize_url(u'http://host/ab/%C2%B5/\xb5'))
114
# Normalize verifies URLs when they are not unicode
115
# (indicating they did not come from the user)
116
self.assertRaises(InvalidURL, normalize_url, 'http://host/\xb5')
117
self.assertRaises(InvalidURL, normalize_url, 'http://host/ ')
119
def test_url_scheme_re(self):
120
# Test paths that may be URLs
121
def test_one(url, scheme_and_path):
122
"""Assert that _url_scheme_re correctly matches
124
:param scheme_and_path: The (scheme, path) that should be matched
125
can be None, to indicate it should not match
127
m = urlutils._url_scheme_re.match(url)
128
if scheme_and_path is None:
129
self.assertEqual(None, m)
131
self.assertEqual(scheme_and_path[0], m.group('scheme'))
132
self.assertEqual(scheme_and_path[1], m.group('path'))
135
test_one('/path', None)
136
test_one('C:/path', None)
137
test_one('../path/to/foo', None)
138
test_one(u'../path/to/fo\xe5', None)
141
test_one('http://host/path/', ('http', 'host/path/'))
142
test_one('sftp://host/path/to/foo', ('sftp', 'host/path/to/foo'))
143
test_one('file:///usr/bin', ('file', '/usr/bin'))
144
test_one('file:///C:/Windows', ('file', '/C:/Windows'))
145
test_one('file:///C|/Windows', ('file', '/C|/Windows'))
146
test_one(u'readonly+sftp://host/path/\xe5', ('readonly+sftp', u'host/path/\xe5'))
149
# Can't have slashes or colons in the scheme
150
test_one('/path/to/://foo', None)
151
test_one('path:path://foo', None)
152
# Must have more than one character for scheme
153
test_one('C://foo', None)
154
test_one('ab://foo', ('ab', 'foo'))
156
def test_dirname(self):
157
# Test bzrlib.urlutils.dirname()
158
dirname = urlutils.dirname
159
if sys.platform == 'win32':
160
self.assertRaises(InvalidURL, dirname, 'file:///path/to/foo')
161
self.assertEqual('file:///C|/', dirname('file:///C|/foo'))
162
self.assertEqual('file:///C|/', dirname('file:///C|/'))
164
self.assertEqual('file:///', dirname('file:///foo'))
165
self.assertEqual('file:///', dirname('file:///'))
167
self.assertEqual('http://host/path/to', dirname('http://host/path/to/foo'))
168
self.assertEqual('http://host/path/to', dirname('http://host/path/to/foo/'))
169
self.assertEqual('http://host/path/to/foo',
170
dirname('http://host/path/to/foo/', exclude_trailing_slash=False))
171
self.assertEqual('http://host/', dirname('http://host/path'))
172
self.assertEqual('http://host/', dirname('http://host/'))
173
self.assertEqual('http://host', dirname('http://host'))
174
self.assertEqual('http:///nohost', dirname('http:///nohost/path'))
176
self.assertEqual('random+scheme://user:pass@ahost:port/',
177
dirname('random+scheme://user:pass@ahost:port/path'))
178
self.assertEqual('random+scheme://user:pass@ahost:port/',
179
dirname('random+scheme://user:pass@ahost:port/path/'))
180
self.assertEqual('random+scheme://user:pass@ahost:port/',
181
dirname('random+scheme://user:pass@ahost:port/'))
184
self.assertEqual('path/to', dirname('path/to/foo'))
185
self.assertEqual('path/to', dirname('path/to/foo/'))
186
self.assertEqual('path/to/foo',
187
dirname('path/to/foo/', exclude_trailing_slash=False))
188
self.assertEqual('path/..', dirname('path/../foo'))
189
self.assertEqual('../path', dirname('../path/foo'))
192
def test(expected, *args):
193
joined = urlutils.join(*args)
194
self.assertEqual(expected, joined)
196
# Test a single element
199
# Test relative path joining
200
test('foo/bar', 'foo', 'bar')
201
test('http://foo/bar', 'http://foo', 'bar')
202
test('http://foo/bar', 'http://foo', '.', 'bar')
203
test('http://foo/baz', 'http://foo', 'bar', '../baz')
204
test('http://foo/bar/baz', 'http://foo', 'bar/baz')
205
test('http://foo/baz', 'http://foo', 'bar/../baz')
208
test('http://bar', 'http://foo', 'http://bar')
209
test('sftp://bzr/foo', 'http://foo', 'bar', 'sftp://bzr/foo')
210
test('file:///bar', 'foo', 'file:///bar')
213
# Cannot go above root
214
self.assertRaises(InvalidURLJoin, urlutils.join,
215
'http://foo', '../baz')
217
def test_function_type(self):
218
if sys.platform == 'win32':
219
self.assertEqual(urlutils._win32_local_path_to_url, urlutils.local_path_to_url)
220
self.assertEqual(urlutils._win32_local_path_from_url, urlutils.local_path_from_url)
222
self.assertEqual(urlutils._posix_local_path_to_url, urlutils.local_path_to_url)
223
self.assertEqual(urlutils._posix_local_path_from_url, urlutils.local_path_from_url)
225
def test_posix_local_path_to_url(self):
226
to_url = urlutils._posix_local_path_to_url
227
self.assertEqual('file:///path/to/foo',
228
to_url('/path/to/foo'))
231
result = to_url(u'/path/to/r\xe4ksm\xf6rg\xe5s')
233
raise TestSkipped("local encoding cannot handle unicode")
235
self.assertEqual('file:///path/to/r%C3%A4ksm%C3%B6rg%C3%A5s', result)
237
def test_posix_local_path_from_url(self):
238
from_url = urlutils._posix_local_path_from_url
239
self.assertEqual('/path/to/foo',
240
from_url('file:///path/to/foo'))
241
self.assertEqual(u'/path/to/r\xe4ksm\xf6rg\xe5s',
242
from_url('file:///path/to/r%C3%A4ksm%C3%B6rg%C3%A5s'))
243
self.assertEqual(u'/path/to/r\xe4ksm\xf6rg\xe5s',
244
from_url('file:///path/to/r%c3%a4ksm%c3%b6rg%c3%a5s'))
246
self.assertRaises(InvalidURL, from_url, '/path/to/foo')
248
def test_win32_local_path_to_url(self):
249
to_url = urlutils._win32_local_path_to_url
250
self.assertEqual('file:///C|/path/to/foo',
251
to_url('C:/path/to/foo'))
254
result = to_url(u'd:/path/to/r\xe4ksm\xf6rg\xe5s')
256
raise TestSkipped("local encoding cannot handle unicode")
258
self.assertEqual('file:///D|/path/to/r%C3%A4ksm%C3%B6rg%C3%A5s', result)
260
def test_win32_local_path_from_url(self):
261
from_url = urlutils._win32_local_path_from_url
262
self.assertEqual('C:/path/to/foo',
263
from_url('file:///C|/path/to/foo'))
264
self.assertEqual(u'D:/path/to/r\xe4ksm\xf6rg\xe5s',
265
from_url('file:///d|/path/to/r%C3%A4ksm%C3%B6rg%C3%A5s'))
266
self.assertEqual(u'D:/path/to/r\xe4ksm\xf6rg\xe5s',
267
from_url('file:///d|/path/to/r%c3%a4ksm%c3%b6rg%c3%a5s'))
269
self.assertRaises(InvalidURL, from_url, '/path/to/foo')
270
# Not a valid _win32 url, no drive letter
271
self.assertRaises(InvalidURL, from_url, 'file:///path/to/foo')
273
def test_split(self):
274
# Test bzrlib.urlutils.split()
275
split = urlutils.split
276
if sys.platform == 'win32':
277
self.assertRaises(InvalidURL, split, 'file:///path/to/foo')
278
self.assertEqual(('file:///C|/', 'foo'), split('file:///C|/foo'))
279
self.assertEqual(('file:///C|/', ''), split('file:///C|/'))
281
self.assertEqual(('file:///', 'foo'), split('file:///foo'))
282
self.assertEqual(('file:///', ''), split('file:///'))
284
self.assertEqual(('http://host/path/to', 'foo'), split('http://host/path/to/foo'))
285
self.assertEqual(('http://host/path/to', 'foo'), split('http://host/path/to/foo/'))
286
self.assertEqual(('http://host/path/to/foo', ''),
287
split('http://host/path/to/foo/', exclude_trailing_slash=False))
288
self.assertEqual(('http://host/', 'path'), split('http://host/path'))
289
self.assertEqual(('http://host/', ''), split('http://host/'))
290
self.assertEqual(('http://host', ''), split('http://host'))
291
self.assertEqual(('http:///nohost', 'path'), split('http:///nohost/path'))
293
self.assertEqual(('random+scheme://user:pass@ahost:port/', 'path'),
294
split('random+scheme://user:pass@ahost:port/path'))
295
self.assertEqual(('random+scheme://user:pass@ahost:port/', 'path'),
296
split('random+scheme://user:pass@ahost:port/path/'))
297
self.assertEqual(('random+scheme://user:pass@ahost:port/', ''),
298
split('random+scheme://user:pass@ahost:port/'))
301
self.assertEqual(('path/to', 'foo'), split('path/to/foo'))
302
self.assertEqual(('path/to', 'foo'), split('path/to/foo/'))
303
self.assertEqual(('path/to/foo', ''),
304
split('path/to/foo/', exclude_trailing_slash=False))
305
self.assertEqual(('path/..', 'foo'), split('path/../foo'))
306
self.assertEqual(('../path', 'foo'), split('../path/foo'))
308
def test_strip_trailing_slash(self):
309
sts = urlutils.strip_trailing_slash
310
if sys.platform == 'win32':
311
self.assertEqual('file:///C|/', sts('file:///C|/'))
312
self.assertEqual('file:///C|/foo', sts('file:///C|/foo'))
313
self.assertEqual('file:///C|/foo', sts('file:///C|/foo/'))
315
self.assertEqual('file:///', sts('file:///'))
316
self.assertEqual('file:///foo', sts('file:///foo'))
317
self.assertEqual('file:///foo', sts('file:///foo/'))
319
self.assertEqual('http://host/', sts('http://host/'))
320
self.assertEqual('http://host/foo', sts('http://host/foo'))
321
self.assertEqual('http://host/foo', sts('http://host/foo/'))
323
# No need to fail just because the slash is missing
324
self.assertEqual('http://host', sts('http://host'))
325
# TODO: jam 20060502 Should this raise InvalidURL?
326
self.assertEqual('file://', sts('file://'))
328
self.assertEqual('random+scheme://user:pass@ahost:port/path',
329
sts('random+scheme://user:pass@ahost:port/path'))
330
self.assertEqual('random+scheme://user:pass@ahost:port/path',
331
sts('random+scheme://user:pass@ahost:port/path/'))
332
self.assertEqual('random+scheme://user:pass@ahost:port/',
333
sts('random+scheme://user:pass@ahost:port/'))
335
# Make sure relative paths work too
336
self.assertEqual('path/to/foo', sts('path/to/foo'))
337
self.assertEqual('path/to/foo', sts('path/to/foo/'))
338
self.assertEqual('../to/foo', sts('../to/foo/'))
339
self.assertEqual('path/../foo', sts('path/../foo/'))
341
def test_unescape_for_display_utf8(self):
342
# Test that URLs are converted to nice unicode strings for display
343
def test(expected, url, encoding='utf-8'):
344
disp_url = urlutils.unescape_for_display(url, encoding=encoding)
345
self.assertIsInstance(disp_url, unicode)
346
self.assertEqual(expected, disp_url)
347
test('http://foo', 'http://foo')
348
if sys.platform == 'win32':
349
test('C:/foo/path', 'file:///C|foo/path')
351
test('/foo/path', 'file:///foo/path')
353
test('http://foo/%2Fbaz', 'http://foo/%2Fbaz')
354
test(u'http://host/r\xe4ksm\xf6rg\xe5s',
355
'http://host/r%C3%A4ksm%C3%B6rg%C3%A5s')
357
# Make sure special escaped characters stay escaped
358
test(u'http://host/%3B%2F%3F%3A%40%26%3D%2B%24%2C%23',
359
'http://host/%3B%2F%3F%3A%40%26%3D%2B%24%2C%23')
361
# Can we handle sections that don't have utf-8 encoding?
362
test(u'http://host/%EE%EE%EE/r\xe4ksm\xf6rg\xe5s',
363
'http://host/%EE%EE%EE/r%C3%A4ksm%C3%B6rg%C3%A5s')
365
# Test encoding into output that can handle some characters
366
test(u'http://host/%EE%EE%EE/r\xe4ksm\xf6rg\xe5s',
367
'http://host/%EE%EE%EE/r%C3%A4ksm%C3%B6rg%C3%A5s',
368
encoding='iso-8859-1')
370
# This one can be encoded into utf8
371
test(u'http://host/\u062c\u0648\u062c\u0648',
372
'http://host/%d8%ac%d9%88%d8%ac%d9%88',
375
# This can't be put into 8859-1 and so stays as escapes
376
test(u'http://host/%d8%ac%d9%88%d8%ac%d9%88',
377
'http://host/%d8%ac%d9%88%d8%ac%d9%88',
378
encoding='iso-8859-1')
380
def test_escape(self):
381
self.assertEqual('%25', urlutils.escape('%'))
382
self.assertEqual('%C3%A5', urlutils.escape(u'\xe5'))
384
def test_unescape(self):
385
self.assertEqual('%', urlutils.unescape('%25'))
386
self.assertEqual(u'\xe5', urlutils.unescape('%C3%A5'))
388
self.assertRaises(InvalidURL, urlutils.unescape, u'\xe5')
389
self.assertRaises(InvalidURL, urlutils.unescape, '\xe5')
390
self.assertRaises(InvalidURL, urlutils.unescape, '%E5')
392
def test_escape_unescape(self):
393
self.assertEqual(u'\xe5', urlutils.unescape(urlutils.escape(u'\xe5')))
394
self.assertEqual('%', urlutils.unescape(urlutils.escape('%')))
396
def test_relative_url(self):
397
def test(expected, base, other):
398
result = urlutils.relative_url(base, other)
399
self.assertEqual(expected, result)
401
test('a', 'http://host/', 'http://host/a')
402
test('http://entirely/different', 'sftp://host/branch',
403
'http://entirely/different')
404
test('../person/feature', 'http://host/branch/mainline',
405
'http://host/branch/person/feature')
406
test('..', 'http://host/branch', 'http://host/')
407
test('http://host2/branch', 'http://host1/branch', 'http://host2/branch')
408
test('.', 'http://host1/branch', 'http://host1/branch')
409
test('../../../branch/2b', 'file:///home/jelmer/foo/bar/2b',
410
'file:///home/jelmer/branch/2b')
411
test('../../branch/2b', 'sftp://host/home/jelmer/bar/2b',
412
'sftp://host/home/jelmer/branch/2b')
413
test('../../branch/feature/2b', 'http://host/home/jelmer/bar/2b',
414
'http://host/home/jelmer/branch/feature/2b')
415
test('../../branch/feature/2b', 'http://host/home/jelmer/bar/2b/',
416
'http://host/home/jelmer/branch/feature/2b')
417
# relative_url should preserve a trailing slash
418
test('../../branch/feature/2b/', 'http://host/home/jelmer/bar/2b/',
419
'http://host/home/jelmer/branch/feature/2b/')
420
test('../../branch/feature/2b/', 'http://host/home/jelmer/bar/2b',
421
'http://host/home/jelmer/branch/feature/2b/')
423
# TODO: treat http://host as http://host/
424
# relative_url is typically called from a branch.base or
425
# transport.base which always ends with a /
426
#test('a', 'http://host', 'http://host/a')
427
test('http://host/a', 'http://host', 'http://host/a')
428
#test('.', 'http://host', 'http://host/')
429
test('http://host/', 'http://host', 'http://host/')
430
#test('.', 'http://host/', 'http://host')
431
test('http://host', 'http://host/', 'http://host')
added
removed
bzrlib/tests/test_urlutils.py
