integration_test.py revision 5d1f7b1de12d16ceb2c938c56701a3e8bfa558f7 
1#!/usr/bin/env python
2# Copyright 2013 The Chromium Authors. All rights reserved.
3# Use of this source code is governed by a BSD-style license that can be
4# found in the LICENSE file.
5
6# Run build_server so that files needed by tests are copied to the local
7# third_party directory.
8import build_server
9build_server.main()
10
11import json
12import optparse
13import os
14import posixpath
15import sys
16import time
17import unittest
18
19from branch_utility import BranchUtility
20from chroot_file_system import ChrootFileSystem
21from extensions_paths import CONTENT_PROVIDERS, EXTENSIONS, PUBLIC_TEMPLATES
22from fake_fetchers import ConfigureFakeFetchers
23from special_paths import SITE_VERIFICATION_FILE
24from handler import Handler
25from link_error_detector import LinkErrorDetector, StringifyBrokenLinks
26from local_file_system import LocalFileSystem
27from local_renderer import LocalRenderer
28from path_util import AssertIsValid
29from servlet import Request
30from third_party.json_schema_compiler import json_parse
31from test_util import (
32    ChromiumPath, DisableLogging, EnableLogging, ReadFile, Server2Path)
33
34
35# Arguments set up if __main__ specifies them.
36_EXPLICIT_TEST_FILES = None
37_REBASE = False
38_VERBOSE = False
39
40
41def _ToPosixPath(os_path):
42  return os_path.replace(os.sep, '/')
43
44
45def _FilterHidden(paths):
46  '''Returns a list of the non-hidden paths from |paths|.
47  '''
48  # Hidden files start with a '.' but paths like './foo' and '../foo' are not
49  # hidden.
50  return [path for path in paths if (not path.startswith('.')) or
51                                     path.startswith('./') or
52                                     path.startswith('../')]
53
54
55def _GetPublicFiles():
56  '''Gets all public file paths mapped to their contents.
57  '''
58  def walk(path, prefix=''):
59    path = ChromiumPath(path)
60    public_files = {}
61    for root, dirs, files in os.walk(path, topdown=True):
62      relative_root = root[len(path):].lstrip(os.path.sep)
63      dirs[:] = _FilterHidden(dirs)
64      for filename in _FilterHidden(files):
65        with open(os.path.join(root, filename), 'r') as f:
66          request_path = posixpath.join(prefix, relative_root, filename)
67          public_files[request_path] = f.read()
68    return public_files
69
70  # Public file locations are defined in content_providers.json, sort of. Epic
71  # hack to pull them out; list all the files from the directories that
72  # Chromium content providers ask for.
73  public_files = {}
74  content_providers = json_parse.Parse(ReadFile(CONTENT_PROVIDERS))
75  for content_provider in content_providers.itervalues():
76    if 'chromium' in content_provider:
77      public_files.update(walk(content_provider['chromium']['dir'],
78                               prefix=content_provider['serveFrom']))
79  return public_files
80
81
82class IntegrationTest(unittest.TestCase):
83  def setUp(self):
84    ConfigureFakeFetchers()
85
86  @EnableLogging('info')
87  def testCronAndPublicFiles(self):
88    '''Runs cron then requests every public file. Cron needs to be run first
89    because the public file requests are offline.
90    '''
91    if _EXPLICIT_TEST_FILES is not None:
92      return
93
94    print('Running cron...')
95    start_time = time.time()
96    try:
97      response = Handler(Request.ForTest('/_cron')).Get()
98      if response:
99        self.assertEqual(200, response.status)
100        self.assertEqual('Success', response.content.ToString())
101      else:
102        self.fail('No response for _cron')
103    finally:
104      print('Took %s seconds' % (time.time() - start_time))
105
106    # TODO(kalman): Re-enable this, but it takes about an hour at the moment,
107    # presumably because every page now has a lot of links on it from the
108    # topnav.
109
110    #print("Checking for broken links...")
111    #start_time = time.time()
112    #link_error_detector = LinkErrorDetector(
113    #    # TODO(kalman): Use of ChrootFileSystem here indicates a hack. Fix.
114    #    ChrootFileSystem(LocalFileSystem.Create(), EXTENSIONS),
115    #    lambda path: Handler(Request.ForTest(path)).Get(),
116    #    'templates/public',
117    #    ('extensions/index.html', 'apps/about_apps.html'))
118
119    #broken_links = link_error_detector.GetBrokenLinks()
120    #if broken_links:
121    #  print('Found %d broken links.' % (
122    #    len(broken_links)))
123    #  if _VERBOSE:
124    #    print(StringifyBrokenLinks(broken_links))
125
126    #broken_links_set = set(broken_links)
127
128    #known_broken_links_path = os.path.join(
129    #    Server2Path('known_broken_links.json'))
130    #try:
131    #  with open(known_broken_links_path, 'r') as f:
132    #    # The JSON file converts tuples and sets into lists, and for this
133    #    # set union/difference logic they need to be converted back.
134    #    known_broken_links = set(tuple(item) for item in json.load(f))
135    #except IOError:
136    #  known_broken_links = set()
137
138    #newly_broken_links = broken_links_set - known_broken_links
139    #fixed_links = known_broken_links - broken_links_set
140
141    #print('Took %s seconds.' % (time.time() - start_time))
142
143    #print('Searching for orphaned pages...')
144    #start_time = time.time()
145    #orphaned_pages = link_error_detector.GetOrphanedPages()
146    #if orphaned_pages:
147    #  # TODO(jshumway): Test should fail when orphaned pages are detected.
148    #  print('Found %d orphaned pages:' % len(orphaned_pages))
149    #  for page in orphaned_pages:
150    #    print(page)
151    #print('Took %s seconds.' % (time.time() - start_time))
152
153    public_files = _GetPublicFiles()
154
155    print('Rendering %s public files...' % len(public_files.keys()))
156    start_time = time.time()
157    try:
158      for path, content in public_files.iteritems():
159        AssertIsValid(path)
160        if path.endswith('redirects.json'):
161          continue
162
163        # The non-example html and md files are served without their file
164        # extensions.
165        path_without_ext, ext = posixpath.splitext(path)
166        if (ext in ('.html', '.md') and
167            '/examples/' not in path and
168            path != SITE_VERIFICATION_FILE):
169          path = path_without_ext
170
171        def check_result(response):
172          self.assertEqual(200, response.status,
173              'Got %s when rendering %s' % (response.status, path))
174
175          # This is reaaaaally rough since usually these will be tiny templates
176          # that render large files. At least it'll catch zero-length responses.
177          self.assertTrue(len(response.content) >= len(content),
178              'Rendered content length was %s vs template content length %s '
179              'when rendering %s' % (len(response.content), len(content), path))
180
181        check_result(Handler(Request.ForTest(path)).Get())
182
183        if path.startswith(('apps/', 'extensions/')):
184          # Make sure that adding the .html will temporarily redirect to
185          # the path without the .html for APIs and articles.
186          if '/examples/' not in path:
187            redirect_response = Handler(Request.ForTest(path + '.html')).Get()
188            self.assertEqual(
189                ('/' + path, False), redirect_response.GetRedirect(),
190                '%s.html did not (temporarily) redirect to %s (status %s)' %
191                    (path, path, redirect_response.status))
192
193          # Make sure including a channel will permanently redirect to the same
194          # path without a channel.
195          for channel in BranchUtility.GetAllChannelNames():
196            redirect_response = Handler(
197                Request.ForTest(posixpath.join(channel, path))).Get()
198            self.assertEqual(
199                ('/' + path, True),
200                redirect_response.GetRedirect(),
201                '%s/%s did not (permanently) redirect to %s (status %s)' %
202                    (channel, path, path, redirect_response.status))
203
204        # Samples are internationalized, test some locales.
205        if path.endswith('/samples'):
206          for lang in ('en-US', 'es', 'ar'):
207            check_result(Handler(Request.ForTest(
208                path,
209                headers={'Accept-Language': '%s;q=0.8' % lang})).Get())
210    finally:
211      print('Took %s seconds' % (time.time() - start_time))
212
213    #if _REBASE:
214    #  print('Rebasing broken links with %s newly broken and %s fixed links.' %
215    #        (len(newly_broken_links), len(fixed_links)))
216    #  with open(known_broken_links_path, 'w') as f:
217    #    json.dump(broken_links, f,
218    #              indent=2, separators=(',', ': '), sort_keys=True)
219    #else:
220    #  if fixed_links or newly_broken_links:
221    #    print('**********************************************\n'
222    #          'CHANGE DETECTED IN BROKEN LINKS WITHOUT REBASE\n'
223    #          '**********************************************')
224    #    print('Found %s broken links, and some have changed. '
225    #          'If this is acceptable or expected then run %s with the --rebase '
226    #          'option.' % (len(broken_links), os.path.split(__file__)[-1]))
227    #  elif broken_links:
228    #    print('%s existing broken links' % len(broken_links))
229    #  if fixed_links:
230    #    print('%s broken links have been fixed:' % len(fixed_links))
231    #    print(StringifyBrokenLinks(fixed_links))
232    #  if newly_broken_links:
233    #    print('There are %s new broken links:' % len(newly_broken_links))
234    #    print(StringifyBrokenLinks(newly_broken_links))
235    #    self.fail('See logging for details.')
236
237  # TODO(kalman): Move this test elsewhere, it's not an integration test.
238  # Perhaps like "presubmit_tests" or something.
239  def testExplicitFiles(self):
240    '''Tests just the files in _EXPLICIT_TEST_FILES.
241    '''
242    if _EXPLICIT_TEST_FILES is None:
243      return
244    for filename in _EXPLICIT_TEST_FILES:
245      print('Rendering %s...' % filename)
246      start_time = time.time()
247      try:
248        response = LocalRenderer.Render(_ToPosixPath(filename))
249        self.assertEqual(200, response.status)
250        self.assertTrue(response.content != '')
251      finally:
252        print('Took %s seconds' % (time.time() - start_time))
253
254    # TODO(jshumway): Check page for broken links (currently prohibited by the
255    # time it takes to render the pages).
256
257  @DisableLogging('warning')
258  def testFileNotFound(self):
259    response = LocalRenderer.Render('/extensions/notfound')
260    self.assertEqual(404, response.status)
261
262  def testSiteVerificationFile(self):
263    response = LocalRenderer.Render('/' + SITE_VERIFICATION_FILE)
264    self.assertEqual(200, response.status)
265
266if __name__ == '__main__':
267  parser = optparse.OptionParser()
268  parser.add_option('-a', '--all', action='store_true', default=False,
269                    help='Render all pages, not just the one specified')
270  parser.add_option('-r', '--rebase', action='store_true', default=False,
271                    help='Rewrites the known_broken_links.json file with '
272                         'the current set of broken links')
273  parser.add_option('-v', '--verbose', action='store_true', default=False,
274                    help='Show verbose output like currently broken links')
275  (opts, args) = parser.parse_args()
276  if not opts.all:
277    _EXPLICIT_TEST_FILES = args
278  _REBASE = opts.rebase
279  _VERBOSE = opts.verbose
280  # Kill sys.argv because we have our own flags.
281  sys.argv = [sys.argv[0]]
282  unittest.main()
283