| 1 | n/a | import unittest |
|---|
| 2 | n/a | from test import support |
|---|
| 3 | n/a | |
|---|
| 4 | n/a | import contextlib |
|---|
| 5 | n/a | import socket |
|---|
| 6 | n/a | import urllib.request |
|---|
| 7 | n/a | import os |
|---|
| 8 | n/a | import email.message |
|---|
| 9 | n/a | import time |
|---|
| 10 | n/a | |
|---|
| 11 | n/a | |
|---|
| 12 | n/a | support.requires('network') |
|---|
| 13 | n/a | |
|---|
| 14 | n/a | class URLTimeoutTest(unittest.TestCase): |
|---|
| 15 | n/a | # XXX this test doesn't seem to test anything useful. |
|---|
| 16 | n/a | |
|---|
| 17 | n/a | TIMEOUT = 30.0 |
|---|
| 18 | n/a | |
|---|
| 19 | n/a | def setUp(self): |
|---|
| 20 | n/a | socket.setdefaulttimeout(self.TIMEOUT) |
|---|
| 21 | n/a | |
|---|
| 22 | n/a | def tearDown(self): |
|---|
| 23 | n/a | socket.setdefaulttimeout(None) |
|---|
| 24 | n/a | |
|---|
| 25 | n/a | def testURLread(self): |
|---|
| 26 | n/a | with support.transient_internet("www.example.com"): |
|---|
| 27 | n/a | f = urllib.request.urlopen("http://www.example.com/") |
|---|
| 28 | n/a | x = f.read() |
|---|
| 29 | n/a | |
|---|
| 30 | n/a | |
|---|
| 31 | n/a | class urlopenNetworkTests(unittest.TestCase): |
|---|
| 32 | n/a | """Tests urllib.reqest.urlopen using the network. |
|---|
| 33 | n/a | |
|---|
| 34 | n/a | These tests are not exhaustive. Assuming that testing using files does a |
|---|
| 35 | n/a | good job overall of some of the basic interface features. There are no |
|---|
| 36 | n/a | tests exercising the optional 'data' and 'proxies' arguments. No tests |
|---|
| 37 | n/a | for transparent redirection have been written. |
|---|
| 38 | n/a | |
|---|
| 39 | n/a | setUp is not used for always constructing a connection to |
|---|
| 40 | n/a | http://www.pythontest.net/ since there a few tests that don't use that address |
|---|
| 41 | n/a | and making a connection is expensive enough to warrant minimizing unneeded |
|---|
| 42 | n/a | connections. |
|---|
| 43 | n/a | |
|---|
| 44 | n/a | """ |
|---|
| 45 | n/a | |
|---|
| 46 | n/a | url = 'http://www.pythontest.net/' |
|---|
| 47 | n/a | |
|---|
| 48 | n/a | @contextlib.contextmanager |
|---|
| 49 | n/a | def urlopen(self, *args, **kwargs): |
|---|
| 50 | n/a | resource = args[0] |
|---|
| 51 | n/a | with support.transient_internet(resource): |
|---|
| 52 | n/a | r = urllib.request.urlopen(*args, **kwargs) |
|---|
| 53 | n/a | try: |
|---|
| 54 | n/a | yield r |
|---|
| 55 | n/a | finally: |
|---|
| 56 | n/a | r.close() |
|---|
| 57 | n/a | |
|---|
| 58 | n/a | def test_basic(self): |
|---|
| 59 | n/a | # Simple test expected to pass. |
|---|
| 60 | n/a | with self.urlopen(self.url) as open_url: |
|---|
| 61 | n/a | for attr in ("read", "readline", "readlines", "fileno", "close", |
|---|
| 62 | n/a | "info", "geturl"): |
|---|
| 63 | n/a | self.assertTrue(hasattr(open_url, attr), "object returned from " |
|---|
| 64 | n/a | "urlopen lacks the %s attribute" % attr) |
|---|
| 65 | n/a | self.assertTrue(open_url.read(), "calling 'read' failed") |
|---|
| 66 | n/a | |
|---|
| 67 | n/a | def test_readlines(self): |
|---|
| 68 | n/a | # Test both readline and readlines. |
|---|
| 69 | n/a | with self.urlopen(self.url) as open_url: |
|---|
| 70 | n/a | self.assertIsInstance(open_url.readline(), bytes, |
|---|
| 71 | n/a | "readline did not return a string") |
|---|
| 72 | n/a | self.assertIsInstance(open_url.readlines(), list, |
|---|
| 73 | n/a | "readlines did not return a list") |
|---|
| 74 | n/a | |
|---|
| 75 | n/a | def test_info(self): |
|---|
| 76 | n/a | # Test 'info'. |
|---|
| 77 | n/a | with self.urlopen(self.url) as open_url: |
|---|
| 78 | n/a | info_obj = open_url.info() |
|---|
| 79 | n/a | self.assertIsInstance(info_obj, email.message.Message, |
|---|
| 80 | n/a | "object returned by 'info' is not an " |
|---|
| 81 | n/a | "instance of email.message.Message") |
|---|
| 82 | n/a | self.assertEqual(info_obj.get_content_subtype(), "html") |
|---|
| 83 | n/a | |
|---|
| 84 | n/a | def test_geturl(self): |
|---|
| 85 | n/a | # Make sure same URL as opened is returned by geturl. |
|---|
| 86 | n/a | with self.urlopen(self.url) as open_url: |
|---|
| 87 | n/a | gotten_url = open_url.geturl() |
|---|
| 88 | n/a | self.assertEqual(gotten_url, self.url) |
|---|
| 89 | n/a | |
|---|
| 90 | n/a | def test_getcode(self): |
|---|
| 91 | n/a | # test getcode() with the fancy opener to get 404 error codes |
|---|
| 92 | n/a | URL = self.url + "XXXinvalidXXX" |
|---|
| 93 | n/a | with support.transient_internet(URL): |
|---|
| 94 | n/a | with self.assertWarns(DeprecationWarning): |
|---|
| 95 | n/a | open_url = urllib.request.FancyURLopener().open(URL) |
|---|
| 96 | n/a | try: |
|---|
| 97 | n/a | code = open_url.getcode() |
|---|
| 98 | n/a | finally: |
|---|
| 99 | n/a | open_url.close() |
|---|
| 100 | n/a | self.assertEqual(code, 404) |
|---|
| 101 | n/a | |
|---|
| 102 | n/a | def test_bad_address(self): |
|---|
| 103 | n/a | # Make sure proper exception is raised when connecting to a bogus |
|---|
| 104 | n/a | # address. |
|---|
| 105 | n/a | |
|---|
| 106 | n/a | # Given that both VeriSign and various ISPs have in |
|---|
| 107 | n/a | # the past or are presently hijacking various invalid |
|---|
| 108 | n/a | # domain name requests in an attempt to boost traffic |
|---|
| 109 | n/a | # to their own sites, finding a domain name to use |
|---|
| 110 | n/a | # for this test is difficult. RFC2606 leads one to |
|---|
| 111 | n/a | # believe that '.invalid' should work, but experience |
|---|
| 112 | n/a | # seemed to indicate otherwise. Single character |
|---|
| 113 | n/a | # TLDs are likely to remain invalid, so this seems to |
|---|
| 114 | n/a | # be the best choice. The trailing '.' prevents a |
|---|
| 115 | n/a | # related problem: The normal DNS resolver appends |
|---|
| 116 | n/a | # the domain names from the search path if there is |
|---|
| 117 | n/a | # no '.' the end and, and if one of those domains |
|---|
| 118 | n/a | # implements a '*' rule a result is returned. |
|---|
| 119 | n/a | # However, none of this will prevent the test from |
|---|
| 120 | n/a | # failing if the ISP hijacks all invalid domain |
|---|
| 121 | n/a | # requests. The real solution would be to be able to |
|---|
| 122 | n/a | # parameterize the framework with a mock resolver. |
|---|
| 123 | n/a | bogus_domain = "sadflkjsasf.i.nvali.d." |
|---|
| 124 | n/a | try: |
|---|
| 125 | n/a | socket.gethostbyname(bogus_domain) |
|---|
| 126 | n/a | except OSError: |
|---|
| 127 | n/a | # socket.gaierror is too narrow, since getaddrinfo() may also |
|---|
| 128 | n/a | # fail with EAI_SYSTEM and ETIMEDOUT (seen on Ubuntu 13.04), |
|---|
| 129 | n/a | # i.e. Python's TimeoutError. |
|---|
| 130 | n/a | pass |
|---|
| 131 | n/a | else: |
|---|
| 132 | n/a | # This happens with some overzealous DNS providers such as OpenDNS |
|---|
| 133 | n/a | self.skipTest("%r should not resolve for test to work" % bogus_domain) |
|---|
| 134 | n/a | failure_explanation = ('opening an invalid URL did not raise OSError; ' |
|---|
| 135 | n/a | 'can be caused by a broken DNS server ' |
|---|
| 136 | n/a | '(e.g. returns 404 or hijacks page)') |
|---|
| 137 | n/a | with self.assertRaises(OSError, msg=failure_explanation): |
|---|
| 138 | n/a | urllib.request.urlopen("http://{}/".format(bogus_domain)) |
|---|
| 139 | n/a | |
|---|
| 140 | n/a | |
|---|
| 141 | n/a | class urlretrieveNetworkTests(unittest.TestCase): |
|---|
| 142 | n/a | """Tests urllib.request.urlretrieve using the network.""" |
|---|
| 143 | n/a | |
|---|
| 144 | n/a | @contextlib.contextmanager |
|---|
| 145 | n/a | def urlretrieve(self, *args, **kwargs): |
|---|
| 146 | n/a | resource = args[0] |
|---|
| 147 | n/a | with support.transient_internet(resource): |
|---|
| 148 | n/a | file_location, info = urllib.request.urlretrieve(*args, **kwargs) |
|---|
| 149 | n/a | try: |
|---|
| 150 | n/a | yield file_location, info |
|---|
| 151 | n/a | finally: |
|---|
| 152 | n/a | support.unlink(file_location) |
|---|
| 153 | n/a | |
|---|
| 154 | n/a | def test_basic(self): |
|---|
| 155 | n/a | # Test basic functionality. |
|---|
| 156 | n/a | with self.urlretrieve(self.logo) as (file_location, info): |
|---|
| 157 | n/a | self.assertTrue(os.path.exists(file_location), "file location returned by" |
|---|
| 158 | n/a | " urlretrieve is not a valid path") |
|---|
| 159 | n/a | with open(file_location, 'rb') as f: |
|---|
| 160 | n/a | self.assertTrue(f.read(), "reading from the file location returned" |
|---|
| 161 | n/a | " by urlretrieve failed") |
|---|
| 162 | n/a | |
|---|
| 163 | n/a | def test_specified_path(self): |
|---|
| 164 | n/a | # Make sure that specifying the location of the file to write to works. |
|---|
| 165 | n/a | with self.urlretrieve(self.logo, |
|---|
| 166 | n/a | support.TESTFN) as (file_location, info): |
|---|
| 167 | n/a | self.assertEqual(file_location, support.TESTFN) |
|---|
| 168 | n/a | self.assertTrue(os.path.exists(file_location)) |
|---|
| 169 | n/a | with open(file_location, 'rb') as f: |
|---|
| 170 | n/a | self.assertTrue(f.read(), "reading from temporary file failed") |
|---|
| 171 | n/a | |
|---|
| 172 | n/a | def test_header(self): |
|---|
| 173 | n/a | # Make sure header returned as 2nd value from urlretrieve is good. |
|---|
| 174 | n/a | with self.urlretrieve(self.logo) as (file_location, info): |
|---|
| 175 | n/a | self.assertIsInstance(info, email.message.Message, |
|---|
| 176 | n/a | "info is not an instance of email.message.Message") |
|---|
| 177 | n/a | |
|---|
| 178 | n/a | logo = "http://www.pythontest.net/" |
|---|
| 179 | n/a | |
|---|
| 180 | n/a | def test_data_header(self): |
|---|
| 181 | n/a | with self.urlretrieve(self.logo) as (file_location, fileheaders): |
|---|
| 182 | n/a | datevalue = fileheaders.get('Date') |
|---|
| 183 | n/a | dateformat = '%a, %d %b %Y %H:%M:%S GMT' |
|---|
| 184 | n/a | try: |
|---|
| 185 | n/a | time.strptime(datevalue, dateformat) |
|---|
| 186 | n/a | except ValueError: |
|---|
| 187 | n/a | self.fail('Date value not in %r format' % dateformat) |
|---|
| 188 | n/a | |
|---|
| 189 | n/a | def test_reporthook(self): |
|---|
| 190 | n/a | records = [] |
|---|
| 191 | n/a | def recording_reporthook(blocks, block_size, total_size): |
|---|
| 192 | n/a | records.append((blocks, block_size, total_size)) |
|---|
| 193 | n/a | |
|---|
| 194 | n/a | with self.urlretrieve(self.logo, reporthook=recording_reporthook) as ( |
|---|
| 195 | n/a | file_location, fileheaders): |
|---|
| 196 | n/a | expected_size = int(fileheaders['Content-Length']) |
|---|
| 197 | n/a | |
|---|
| 198 | n/a | records_repr = repr(records) # For use in error messages. |
|---|
| 199 | n/a | self.assertGreater(len(records), 1, msg="There should always be two " |
|---|
| 200 | n/a | "calls; the first one before the transfer starts.") |
|---|
| 201 | n/a | self.assertEqual(records[0][0], 0) |
|---|
| 202 | n/a | self.assertGreater(records[0][1], 0, |
|---|
| 203 | n/a | msg="block size can't be 0 in %s" % records_repr) |
|---|
| 204 | n/a | self.assertEqual(records[0][2], expected_size) |
|---|
| 205 | n/a | self.assertEqual(records[-1][2], expected_size) |
|---|
| 206 | n/a | |
|---|
| 207 | n/a | block_sizes = {block_size for _, block_size, _ in records} |
|---|
| 208 | n/a | self.assertEqual({records[0][1]}, block_sizes, |
|---|
| 209 | n/a | msg="block sizes in %s must be equal" % records_repr) |
|---|
| 210 | n/a | self.assertGreaterEqual(records[-1][0]*records[0][1], expected_size, |
|---|
| 211 | n/a | msg="number of blocks * block size must be" |
|---|
| 212 | n/a | " >= total size in %s" % records_repr) |
|---|
| 213 | n/a | |
|---|
| 214 | n/a | |
|---|
| 215 | n/a | if __name__ == "__main__": |
|---|
| 216 | n/a | unittest.main() |
|---|