def testCronAndPublicFiles(self): '''Runs cron then requests every public file. Cron needs to be run first because the public file requests are offline. ''' if _EXPLICIT_TEST_FILES is not None: return print('Running cron...') start_time = time.time() try: response = Handler(Request.ForTest('/_cron')).Get() if response: self.assertEqual(200, response.status) self.assertEqual('Success', response.content.ToString()) else: self.fail('No response for _cron') finally: print('Took %s seconds' % (time.time() - start_time)) # TODO(kalman): Re-enable this, but it takes about an hour at the moment, # presumably because every page now has a lot of links on it from the # topnav. #print("Checking for broken links...") #start_time = time.time() #link_error_detector = LinkErrorDetector( # # TODO(kalman): Use of ChrootFileSystem here indicates a hack. Fix. # ChrootFileSystem(LocalFileSystem.Create(), EXTENSIONS), # lambda path: Handler(Request.ForTest(path)).Get(), # 'templates/public', # ('extensions/index.html', 'apps/about_apps.html')) #broken_links = link_error_detector.GetBrokenLinks() #if broken_links: # print('Found %d broken links.' % ( # len(broken_links))) # if _VERBOSE: # print(StringifyBrokenLinks(broken_links)) #broken_links_set = set(broken_links) #known_broken_links_path = os.path.join( # Server2Path('known_broken_links.json')) #try: # with open(known_broken_links_path, 'r') as f: # # The JSON file converts tuples and sets into lists, and for this # # set union/difference logic they need to be converted back. # known_broken_links = set(tuple(item) for item in json.load(f)) #except IOError: # known_broken_links = set() #newly_broken_links = broken_links_set - known_broken_links #fixed_links = known_broken_links - broken_links_set #print('Took %s seconds.' % (time.time() - start_time)) #print('Searching for orphaned pages...') #start_time = time.time() #orphaned_pages = link_error_detector.GetOrphanedPages() #if orphaned_pages: # # TODO(jshumway): Test should fail when orphaned pages are detected. # print('Found %d orphaned pages:' % len(orphaned_pages)) # for page in orphaned_pages: # print(page) #print('Took %s seconds.' % (time.time() - start_time)) public_files = _GetPublicFiles() print('Rendering %s public files...' % len(public_files.keys())) start_time = time.time() try: for path, content in public_files.iteritems(): AssertIsValid(path) if path.endswith('redirects.json'): continue # The non-example html and md files are served without their file # extensions. path_without_ext, ext = posixpath.splitext(path) if (ext in ('.html', '.md') and '/examples/' not in path and path != SITE_VERIFICATION_FILE): path = path_without_ext def check_result(response): self.assertEqual( 200, response.status, 'Got %s when rendering %s' % (response.status, path)) # This is reaaaaally rough since usually these will be tiny templates # that render large files. At least it'll catch zero-length responses. self.assertTrue( len(response.content) >= len(content), 'Rendered content length was %s vs template content length %s ' 'when rendering %s' % (len(response.content), len(content), path)) check_result(Handler(Request.ForTest(path)).Get()) if path.startswith(('apps/', 'extensions/')): # Make sure that adding the .html will temporarily redirect to # the path without the .html for APIs and articles. if '/examples/' not in path: redirect_response = Handler( Request.ForTest(path + '.html')).Get() self.assertEqual( ('/' + path, False), redirect_response.GetRedirect(), '%s.html did not (temporarily) redirect to %s (status %s)' % (path, path, redirect_response.status)) # Make sure including a channel will permanently redirect to the same # path without a channel. for channel in BranchUtility.GetAllChannelNames(): redirect_response = Handler( Request.ForTest(posixpath.join(channel, path))).Get() self.assertEqual( ('/' + path, True), redirect_response.GetRedirect(), '%s/%s did not (permanently) redirect to %s (status %s)' % (channel, path, path, redirect_response.status)) # Samples are internationalized, test some locales. if path.endswith('/samples'): for lang in ('en-US', 'es', 'ar'): check_result( Handler( Request.ForTest(path, headers={ 'Accept-Language': '%s;q=0.8' % lang })).Get()) finally: print('Took %s seconds' % (time.time() - start_time))
def testCronAndPublicFiles(self): '''Runs cron then requests every public file. Cron needs to be run first because the public file requests are offline. ''' if _EXPLICIT_TEST_FILES is not None: return print('Running cron...') start_time = time.time() try: response = Handler(Request.ForTest('/_cron/stable')).Get() self.assertEqual(200, response.status) self.assertEqual('Success', response.content.ToString()) finally: print('Took %s seconds' % (time.time() - start_time)) print("Checking for broken links...") start_time = time.time() link_error_detector = LinkErrorDetector( LocalFileSystem(os.path.join(sys.path[0], os.pardir, os.pardir)), lambda path: Handler(Request.ForTest(path)).Get(), 'templates/public', ('extensions/index.html', 'apps/about_apps.html')) broken_links = link_error_detector.GetBrokenLinks() if broken_links: # TODO(jshumway): Test should fail when broken links are detected. print('Warning: Found %d broken links:' % (len(broken_links))) print(StringifyBrokenLinks(broken_links)) print('Took %s seconds.' % (time.time() - start_time)) print('Searching for orphaned pages...') start_time = time.time() orphaned_pages = link_error_detector.GetOrphanedPages() if orphaned_pages: # TODO(jshumway): Test should fail when orphaned pages are detected. print('Warning: Found %d orphaned pages:' % len(orphaned_pages)) for page in orphaned_pages: print(page) print('Took %s seconds.' % (time.time() - start_time)) public_files = _GetPublicFiles() print('Rendering %s public files...' % len(public_files.keys())) start_time = time.time() try: for path, content in public_files.iteritems(): if path.endswith('redirects.json'): continue def check_result(response): self.assertEqual( 200, response.status, 'Got %s when rendering %s' % (response.status, path)) # This is reaaaaally rough since usually these will be tiny templates # that render large files. At least it'll catch zero-length responses. self.assertTrue( len(response.content) >= len(content), 'Content was "%s" when rendering %s' % (response.content, path)) check_result(Handler(Request.ForTest(path)).Get()) # Make sure that leaving out the .html will temporarily redirect to the # path with the .html. if path.startswith(('apps/', 'extensions/')): redirect_result = Handler( Request.ForTest(posixpath.splitext(path)[0])).Get() self.assertEqual((path, False), redirect_result.GetRedirect()) # Make sure including a channel will permanently redirect to the same # path without a channel. for channel in BranchUtility.GetAllChannelNames(): redirect_result = Handler( Request.ForTest('%s/%s' % (channel, path))).Get() self.assertEqual((path, True), redirect_result.GetRedirect()) # Samples are internationalized, test some locales. if path.endswith('/samples.html'): for lang in ['en-US', 'es', 'ar']: check_result( Handler( Request.ForTest(path, headers={ 'Accept-Language': '%s;q=0.8' % lang })).Get()) finally: print('Took %s seconds' % (time.time() - start_time))
def testCronAndPublicFiles(self): '''Runs cron then requests every public file. Cron needs to be run first because the public file requests are offline. ''' if _EXPLICIT_TEST_FILES is not None: return print('Running cron...') start_time = time.time() try: response = Handler(Request.ForTest('/_cron')).Get() self.assertEqual(200, response.status) self.assertEqual('Success', response.content.ToString()) finally: print('Took %s seconds' % (time.time() - start_time)) print("Checking for broken links...") start_time = time.time() link_error_detector = LinkErrorDetector( # TODO(kalman): Use of ChrootFileSystem here indicates a hack. Fix. ChrootFileSystem(LocalFileSystem.Create(), EXTENSIONS), lambda path: Handler(Request.ForTest(path)).Get(), 'templates/public', ('extensions/index.html', 'apps/about_apps.html')) broken_links = link_error_detector.GetBrokenLinks() if broken_links and _VERBOSE: print('The broken links are:') print(StringifyBrokenLinks(broken_links)) broken_links_set = set(broken_links) known_broken_links_path = os.path.join( sys.path[0], 'known_broken_links.json') try: with open(known_broken_links_path, 'r') as f: # The JSON file converts tuples and sets into lists, and for this # set union/difference logic they need to be converted back. known_broken_links = set(tuple(item) for item in json.load(f)) except IOError: known_broken_links = set() newly_broken_links = broken_links_set - known_broken_links fixed_links = known_broken_links - broken_links_set if _REBASE: print('Rebasing broken links with %s newly broken and %s fixed links.' % (len(newly_broken_links), len(fixed_links))) with open(known_broken_links_path, 'w') as f: json.dump(broken_links, f, indent=2, separators=(',', ': '), sort_keys=True) else: if fixed_links or newly_broken_links: print('Found %s broken links, and some have changed. ' 'If this is acceptable or expected then run %s with the --rebase ' 'option.' % (len(broken_links), os.path.split(__file__)[-1])) elif broken_links: print('Found %s broken links, but there were no changes.' % len(broken_links)) if fixed_links: print('%s broken links have been fixed:' % len(fixed_links)) print(StringifyBrokenLinks(fixed_links)) if newly_broken_links: print('There are %s new broken links:' % len(newly_broken_links)) print(StringifyBrokenLinks(newly_broken_links)) self.fail('See logging for details.') print('Took %s seconds.' % (time.time() - start_time)) print('Searching for orphaned pages...') start_time = time.time() orphaned_pages = link_error_detector.GetOrphanedPages() if orphaned_pages: # TODO(jshumway): Test should fail when orphaned pages are detected. print('Warning: Found %d orphaned pages:' % len(orphaned_pages)) for page in orphaned_pages: print(page) print('Took %s seconds.' % (time.time() - start_time)) public_files = _GetPublicFiles() print('Rendering %s public files...' % len(public_files.keys())) start_time = time.time() try: for path, content in public_files.iteritems(): if path.endswith('redirects.json'): continue def check_result(response): self.assertEqual(200, response.status, 'Got %s when rendering %s' % (response.status, path)) # This is reaaaaally rough since usually these will be tiny templates # that render large files. At least it'll catch zero-length responses. self.assertTrue(len(response.content) >= len(content), 'Content was "%s" when rendering %s' % (response.content, path)) check_result(Handler(Request.ForTest(path)).Get()) # Make sure that leaving out the .html will temporarily redirect to the # path with the .html. if path.startswith(('apps/', 'extensions/')): redirect_result = Handler( Request.ForTest(posixpath.splitext(path)[0])).Get() self.assertEqual((path, False), redirect_result.GetRedirect()) # Make sure including a channel will permanently redirect to the same # path without a channel. for channel in BranchUtility.GetAllChannelNames(): redirect_result = Handler( Request.ForTest('%s/%s' % (channel, path))).Get() self.assertEqual((path, True), redirect_result.GetRedirect()) # Samples are internationalized, test some locales. if path.endswith('/samples.html'): for lang in ['en-US', 'es', 'ar']: check_result(Handler(Request.ForTest( path, headers={'Accept-Language': '%s;q=0.8' % lang})).Get()) finally: print('Took %s seconds' % (time.time() - start_time))