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 (
22    CONTENT_PROVIDERS, CHROME_EXTENSIONS, PUBLIC_TEMPLATES)
23from fake_fetchers import ConfigureFakeFetchers
24from special_paths import SITE_VERIFICATION_FILE
25from handler import Handler
26from link_error_detector import LinkErrorDetector, StringifyBrokenLinks
27from local_file_system import LocalFileSystem
28from local_renderer import LocalRenderer
29from path_util import AssertIsValid
30from servlet import Request
31from third_party.json_schema_compiler import json_parse
32from test_util import (
33    ChromiumPath, DisableLogging, EnableLogging, ReadFile, Server2Path)
34
35
36# Arguments set up if __main__ specifies them.
37_EXPLICIT_TEST_FILES = None
38_REBASE = False
39_VERBOSE = False
40
41
42def _ToPosixPath(os_path):
43  return os_path.replace(os.sep, '/')
44
45
46def _FilterHidden(paths):
47  '''Returns a list of the non-hidden paths from |paths|.
48  '''
49  # Hidden files start with a '.' but paths like './foo' and '../foo' are not
50  # hidden.
51  return [path for path in paths if (not path.startswith('.')) or
52                                     path.startswith('./') or
53                                     path.startswith('../')]
54
55
56def _GetPublicFiles():
57  '''Gets all public file paths mapped to their contents.
58  '''
59  def walk(path, prefix=''):
60    path = ChromiumPath(path)
61    public_files = {}
62    for root, dirs, files in os.walk(path, topdown=True):
63      relative_root = root[len(path):].lstrip(os.path.sep)
64      dirs[:] = _FilterHidden(dirs)
65      for filename in _FilterHidden(files):
66        with open(os.path.join(root, filename), 'r') as f:
67          request_path = posixpath.join(prefix, relative_root, filename)
68          public_files[request_path] = f.read()
69    return public_files
70
71  # Public file locations are defined in content_providers.json, sort of. Epic
72  # hack to pull them out; list all the files from the directories that
73  # Chromium content providers ask for.
74  public_files = {}
75  content_providers = json_parse.Parse(ReadFile(CONTENT_PROVIDERS))
76  for content_provider in content_providers.itervalues():
77    if 'chromium' in content_provider:
78      public_files.update(walk(content_provider['chromium']['dir'],
79                               prefix=content_provider['serveFrom']))
80  return public_files
81
82
83class IntegrationTest(unittest.TestCase):
84  def setUp(self):
85    ConfigureFakeFetchers()
86
87  @EnableLogging('info')
88  def testCronAndPublicFiles(self):
89    '''Runs cron then requests every public file. Cron needs to be run first
90    because the public file requests are offline.
91    '''
92    if _EXPLICIT_TEST_FILES is not None:
93      return
94
95    print('Running cron...')
96    start_time = time.time()
97    try:
98      response = Handler(Request.ForTest('/_cron')).Get()
99      if response:
100        self.assertEqual(200, response.status)
101        self.assertEqual('Success', response.content.ToString())
102      else:
103        self.fail('No response for _cron')
104    finally:
105      print('Took %s seconds' % (time.time() - start_time))
106
107    # TODO(kalman): Re-enable this, but it takes about an hour at the moment,
108    # presumably because every page now has a lot of links on it from the
109    # topnav.
110
111    #print("Checking for broken links...")
112    #start_time = time.time()
113    #link_error_detector = LinkErrorDetector(
114    #    # TODO(kalman): Use of ChrootFileSystem here indicates a hack. Fix.
115    #    ChrootFileSystem(LocalFileSystem.Create(), CHROME_EXTENSIONS),
116    #    lambda path: Handler(Request.ForTest(path)).Get(),
117    #    'templates/public',
118    #    ('extensions/index.html', 'apps/about_apps.html'))
119
120    #broken_links = link_error_detector.GetBrokenLinks()
121    #if broken_links:
122    #  print('Found %d broken links.' % (
123    #    len(broken_links)))
124    #  if _VERBOSE:
125    #    print(StringifyBrokenLinks(broken_links))
126
127    #broken_links_set = set(broken_links)
128
129    #known_broken_links_path = os.path.join(
130    #    Server2Path('known_broken_links.json'))
131    #try:
132    #  with open(known_broken_links_path, 'r') as f:
133    #    # The JSON file converts tuples and sets into lists, and for this
134    #    # set union/difference logic they need to be converted back.
135    #    known_broken_links = set(tuple(item) for item in json.load(f))
136    #except IOError:
137    #  known_broken_links = set()
138
139    #newly_broken_links = broken_links_set - known_broken_links
140    #fixed_links = known_broken_links - broken_links_set
141
142    #print('Took %s seconds.' % (time.time() - start_time))
143
144    #print('Searching for orphaned pages...')
145    #start_time = time.time()
146    #orphaned_pages = link_error_detector.GetOrphanedPages()
147    #if orphaned_pages:
148    #  # TODO(jshumway): Test should fail when orphaned pages are detected.
149    #  print('Found %d orphaned pages:' % len(orphaned_pages))
150    #  for page in orphaned_pages:
151    #    print(page)
152    #print('Took %s seconds.' % (time.time() - start_time))
153
154    public_files = _GetPublicFiles()
155
156    print('Rendering %s public files...' % len(public_files.keys()))
157    start_time = time.time()
158    try:
159      for path, content in public_files.iteritems():
160        AssertIsValid(path)
161        if path.endswith('redirects.json'):
162          continue
163
164        # The non-example html and md files are served without their file
165        # extensions.
166        path_without_ext, ext = posixpath.splitext(path)
167        if (ext in ('.html', '.md') and
168            '/examples/' not in path and
169            path != SITE_VERIFICATION_FILE):
170          path = path_without_ext
171
172        def check_result(response):
173          self.assertEqual(200, response.status,
174              'Got %s when rendering %s' % (response.status, path))
175
176          # This is reaaaaally rough since usually these will be tiny templates
177          # that render large files. At least it'll catch zero-length responses.
178          self.assertTrue(len(response.content) >= len(content),
179              'Rendered content length was %s vs template content length %s '
180              'when rendering %s' % (len(response.content), len(content), path))
181
182        check_result(Handler(Request.ForTest(path)).Get())
183
184        if path.startswith(('apps/', 'extensions/')):
185          # Make sure that adding the .html will temporarily redirect to
186          # the path without the .html for APIs and articles.
187          if '/examples/' not in path:
188            redirect_response = Handler(Request.ForTest(path + '.html')).Get()
189            self.assertEqual(
190                ('/' + path, False), redirect_response.GetRedirect(),
191                '%s.html did not (temporarily) redirect to %s (status %s)' %
192                    (path, path, redirect_response.status))
193
194          # Make sure including a channel will permanently redirect to the same
195          # path without a channel.
196          for channel in BranchUtility.GetAllChannelNames():
197            redirect_response = Handler(
198                Request.ForTest(posixpath.join(channel, path))).Get()
199            self.assertEqual(
200                ('/' + path, True),
201                redirect_response.GetRedirect(),
202                '%s/%s did not (permanently) redirect to %s (status %s)' %
203                    (channel, path, path, redirect_response.status))
204
205        # Samples are internationalized, test some locales.
206        if path.endswith('/samples'):
207          for lang in ('en-US', 'es', 'ar'):
208            check_result(Handler(Request.ForTest(
209                path,
210                headers={'Accept-Language': '%s;q=0.8' % lang})).Get())
211    finally:
212      print('Took %s seconds' % (time.time() - start_time))
213
214    #if _REBASE:
215    #  print('Rebasing broken links with %s newly broken and %s fixed links.' %
216    #        (len(newly_broken_links), len(fixed_links)))
217    #  with open(known_broken_links_path, 'w') as f:
218    #    json.dump(broken_links, f,
219    #              indent=2, separators=(',', ': '), sort_keys=True)
220    #else:
221    #  if fixed_links or newly_broken_links:
222    #    print('**********************************************\n'
223    #          'CHANGE DETECTED IN BROKEN LINKS WITHOUT REBASE\n'
224    #          '**********************************************')
225    #    print('Found %s broken links, and some have changed. '
226    #          'If this is acceptable or expected then run %s with the --rebase '
227    #          'option.' % (len(broken_links), os.path.split(__file__)[-1]))
228    #  elif broken_links:
229    #    print('%s existing broken links' % len(broken_links))
230    #  if fixed_links:
231    #    print('%s broken links have been fixed:' % len(fixed_links))
232    #    print(StringifyBrokenLinks(fixed_links))
233    #  if newly_broken_links:
234    #    print('There are %s new broken links:' % len(newly_broken_links))
235    #    print(StringifyBrokenLinks(newly_broken_links))
236    #    self.fail('See logging for details.')
237
238  # TODO(kalman): Move this test elsewhere, it's not an integration test.
239  # Perhaps like "presubmit_tests" or something.
240  def testExplicitFiles(self):
241    '''Tests just the files in _EXPLICIT_TEST_FILES.
242    '''
243    if _EXPLICIT_TEST_FILES is None:
244      return
245    for filename in _EXPLICIT_TEST_FILES:
246      print('Rendering %s...' % filename)
247      start_time = time.time()
248      try:
249        response = LocalRenderer.Render(_ToPosixPath(filename))
250        self.assertEqual(200, response.status)
251        self.assertTrue(response.content != '')
252      finally:
253        print('Took %s seconds' % (time.time() - start_time))
254
255    # TODO(jshumway): Check page for broken links (currently prohibited by the
256    # time it takes to render the pages).
257
258  @DisableLogging('warning')
259  def testFileNotFound(self):
260    response = LocalRenderer.Render('/extensions/notfound')
261    self.assertEqual(404, response.status)
262
263  def testSiteVerificationFile(self):
264    response = LocalRenderer.Render('/' + SITE_VERIFICATION_FILE)
265    self.assertEqual(200, response.status)
266
267if __name__ == '__main__':
268  parser = optparse.OptionParser()
269  parser.add_option('-a', '--all', action='store_true', default=False,
270                    help='Render all pages, not just the one specified')
271  parser.add_option('-r', '--rebase', action='store_true', default=False,
272                    help='Rewrites the known_broken_links.json file with '
273                         'the current set of broken links')
274  parser.add_option('-v', '--verbose', action='store_true', default=False,
275                    help='Show verbose output like currently broken links')
276  (opts, args) = parser.parse_args()
277  if not opts.all:
278    _EXPLICIT_TEST_FILES = args
279  _REBASE = opts.rebase
280  _VERBOSE = opts.verbose
281  # Kill sys.argv because we have our own flags.
282  sys.argv = [sys.argv[0]]
283  unittest.main()
284