Index: django/test/testcases.py =================================================================== --- django/test/testcases.py (revision 6352) +++ django/test/testcases.py (working copy) @@ -186,3 +186,141 @@ self.failIf(template_name in template_names, (u"Template '%s' was used unexpectedly in rendering the" u" response") % template_name) + + def assertNoBrokenLinks(self, response, internal_only=True): + """ + Asserts that all the links within the response, when followed, return + a valid page (a 200) or a redirect (302). + + Blank links are also identified (such as ) as this is helpful + to identify when the url tag in + fails. + + Internal page links (such as Skip to content are + also checked to ensure they are not broken (ie. that an element with the + id exists on the page). + + Current issues/thoughts: + * Should we follow 302's to verify the page redirects to a 200 result? + """ + non_broken_status_codes = (200, 301, 302, 304) + + # Create the parser to grab the internal and external links + import HTMLParser + + class AnchorParser(HTMLParser.HTMLParser): + external_href_re = re.compile(r'^https?://', re.IGNORECASE) + ignore_href_re = re.compile(r'^(mailto|ftp):', re.IGNORECASE) + internal_id_href_re = re.compile(r'^#(.*)') + + def __init__(self): + self.hrefs_internal = [] + self.hrefs_external = [] + self.interal_page_link_ids = [] + self.element_ids = [] + self.reset() + + def handle_starttag(self, tag, attrs): + if tag == "a": + for k, v in attrs: + if k == "href": + # For each href that we're not ignoring, save the + # value and position + if self.ignore_href_re.match(v): + pass + elif self.external_href_re.match(v): + self.hrefs_external.append((v, self.getpos())) + elif self.internal_id_href_re.match(v): + # If this is of the form href="#content" then + # remember the actual id "content". + self.interal_page_link_ids.append( + ( + self.internal_id_href_re.match(v).groups()[0], + self.getpos() + ) + ) + else: + self.hrefs_internal.append((v, self.getpos())) + elif k == "id": + # An anchor link can have an id and be linked to + # via an internal page link too. + self.element_ids.append(v) + else: + # Go through the attributes of all the other tags so we know all + # the element id's within the page for internal page links. + for k, v in attrs: + if k == "id": + self.element_ids.append(v) + + p = AnchorParser() + p.feed(response.content) + p.close() + + # Check the internal links first: + for link, (lineno, offset) in p.hrefs_internal: + self.failIf( + ''==link, + (u"The page contains a link with an empty href on line %(lineno)d.") % { + 'page': 'pagename', + 'lineno': lineno, + 'response': response + } + ) + + link_response = response.client.get(link) + self.failUnless( + link_response.status_code in non_broken_status_codes, + (u"The link '%(link)s' on line %(lineno)d appears to be broken (status is %(status)s)") % { + 'link': link, + 'lineno': lineno, + 'status': link_response.status_code + } + ) + + # Next, check the internal page links: + for id, (lineno, offset) in p.interal_page_link_ids: + # If the id wasn't blank (ie. then make sure that there + # was an element with the same id on the page somewhere. + if id: + self.failUnless( + id in p.element_ids, + ( + u"The internal link to #%(id)s on line %(lineno)d does" + u" not link to a corresponding element with an " + u"id=\"%(id)s\"." % { + 'id': id, + 'lineno': lineno + } + ) + ) + + # Then check the external links + if not internal_only: + import urllib2 + from django.conf import settings + + headers = { + "Accept": "text/xml,application/xml,application/xhtml+xml,text/html;q=0.9,text/plain;q=0.8,image/png,*/*;q=0.5", + "Accept-Language": "en-us,en;q=0.5", + "Accept-Charset": "ISO-8859-1,utf-8;q=0.7,*;q=0.7", + "Connection": "close", + "User-Agent": settings.URL_VALIDATOR_USER_AGENT, + } + + for link, (lineno, offset) in p.hrefs_external: + try: + req = urllib2.Request(link, None, headers) + u = urllib2.urlopen(req) + except ValueError: + self.fail( + u"The link '%(link)s' on line %(lineno)d appears to be invalid." % { + 'link': link, + 'lineno': lineno + } + ) + except: # urllib2.URLError, httplib.InvalidURL, etc. + self.fail(u"The link '%(link)s' on line %(lineno)d appears to be broken." % { + 'link': link, + 'lineno': lineno + } + ) Index: tests/regressiontests/test_client_regress/views.py =================================================================== --- tests/regressiontests/test_client_regress/views.py (revision 6352) +++ tests/regressiontests/test_client_regress/views.py (working copy) @@ -1,6 +1,6 @@ from django.contrib.auth.decorators import login_required from django.core.mail import EmailMessage, SMTPConnection -from django.http import HttpResponse, HttpResponseRedirect, HttpResponseServerError +from django.http import HttpResponse, HttpResponseRedirect, HttpResponseServerError, HttpResponseNotFound from django.shortcuts import render_to_response def no_template_view(request): @@ -27,4 +27,111 @@ def login_protected_redirect_view(request): "A view that redirects all requests to the GET view" return HttpResponseRedirect('/test_client_regress/get_view/') -login_protected_redirect_view = login_required(login_protected_redirect_view) \ No newline at end of file +login_protected_redirect_view = login_required(login_protected_redirect_view) + +def no_broken_links_view(request): + return HttpResponse( + """ + Lots of html stuff, including a few local links: + here + This should redirect to add the / + + a redirect view. Doesn't matter if closing tag missing. + + Some external links: + Django blog + + Some anchor links should be ignored, such as: + Send me an email, or + Download from here + + Including a Missing slash + Whole bunch of other stuff before the page ends. + """ + ) + +def broken_external_link_view(request): + return HttpResponse( + """ + Lots of html stuff, including a few local links: + here + This should redirect to add the / + + a redirect view. Doesn't matter if closing tag missing. + + Some external links: + Django blog + But this one's a + Broken link + Whole bunch of other stuff before the page ends. + """ + ) + +def broken_internal_link_view(request): + return HttpResponse( + """ + Lots of html stuff, including a few local links: + + A broken view + + """ + ) + +def bad_internal_link_view(request): + return HttpResponse( + """ + Lots of html stuff, including a few local links: + + A bad view + + """ + ) + +def invalid_external_link_view(request): + return HttpResponse( + """ + Lots of html stuff, including a few local links: + + An invalid link + + """ + ) + +def blank_link_view(request): + return HttpResponse( + """ + If a link uses the url template tag to create the link for the href like + this: + href="{% url my-url-name arg %}" + and fails, it will end up with blank href="", this would be useful to + catch! + + A blank link. + + """ + ) + +def internal_page_link_view(request): + return HttpResponse( + """ + A link which is just internal to the page, href="#content" needs to + have a matching element with an id="content" on the page. + + This one should be fine + + This one should be ignored (lots of JS uses) + + But this one isn't valid as there's no corresponding + element with the id="footer" + +
+ Here's the content +
+ """ + ) + +def broken_view(request): + return HttpResponseServerError() + +def bad_view(request): + return HttpResponseNotFound() Index: tests/regressiontests/test_client_regress/models.py =================================================================== --- tests/regressiontests/test_client_regress/models.py (revision 6352) +++ tests/regressiontests/test_client_regress/models.py (working copy) @@ -233,6 +233,132 @@ except AssertionError, e: self.assertEqual(str(e), "The form 'form' in context 0 does not contain the non-field error 'Some error.' (actual errors: )") +class AssertNoBrokenLinksTests(TestCase): + def test_no_broken_links(self): + "Tests that assertion confirms internal and external non-broken links." + + response = self.client.get('/test_client_regress/no_broken_links_view/') + self.assertEqual(response.status_code, 200) + + self.assertNoBrokenLinks(response, internal_only=False) + + def test_broken_external_link(self): + "Tests that assertion finds broken external links" + + response = self.client.get('/test_client_regress/broken_external_link_view/') + self.assertEqual(response.status_code, 200) + + # No internal broken links: + self.assertNoBrokenLinks(response) + + # But there is an external broken link: + assertion_raised=False + try: + self.assertNoBrokenLinks(response, internal_only=False) + except AssertionError, e: + assertion_raised = True # Should always get here + self.assertEqual( + str(e), + "The link 'http://djangoproject.com/badlink.html' on line 11" + " appears to be broken." + ) + + self.assertTrue(assertion_raised) + + def test_invalid_external_link(self): + "Tests that assertion finds invalid external links" + + response = self.client.get('/test_client_regress/invalid_external_link_view/') + self.assertEqual(response.status_code, 200) + + assertion_raised=False + try: + self.assertNoBrokenLinks(response, internal_only=False) + except AssertionError, e: + assertion_raised = True # Should always get here + self.assertEqual( + str(e), + "The link 'http://djangoproject&.com' on line 4 appears" + " to be broken." + ) + + self.assertTrue(assertion_raised) + + def test_broken_internal_link(self): + "Tests that assertion finds broken internal links" + + response = self.client.get('/test_client_regress/broken_internal_link_view/') + self.assertEqual(response.status_code, 200) + + assertion_raised=False + try: + self.assertNoBrokenLinks(response) + except AssertionError, e: + assertion_raised = True # Should always get here + self.assertEqual( + str(e), + "The link '/test_client_regress/broken_view/' on line 4" + " appears to be broken (status is 500)" + ) + + self.assertTrue(assertion_raised) + + def test_bad_internal_link(self): + "Tests that assertion finds bad internal links" + + response = self.client.get('/test_client_regress/bad_internal_link_view/') + self.assertEqual(response.status_code, 200) + + assertion_raised=False + try: + self.assertNoBrokenLinks(response) + except AssertionError, e: + assertion_raised = True # Should always get here + self.assertEqual( + str(e), + "The link '/test_client_regress/bad_view/' on line 4 appears to" + " be broken (status is 404)" + ) + + self.assertTrue(assertion_raised) + + def test_blank_link(self): + "Tests that links with blank hrefs are identified appropriately" + + response = self.client.get('/test_client_regress/blank_link_view/') + self.assertEqual(response.status_code, 200) + + assertion_raised=False + try: + self.assertNoBrokenLinks(response) + except AssertionError, e: + assertion_raised = True # Should always get here + self.assertEqual( + str(e), + "The page contains a link with an empty href on line 8." + ) + + self.assertTrue(assertion_raised) + + def test_internal_page_link(self): + "Tests that internal page links are valid" + + response = self.client.get('/test_client_regress/internal_page_link_view/') + self.assertEqual(response.status_code, 200) + + assertion_raised=False + try: + self.assertNoBrokenLinks(response) + except AssertionError, e: + assertion_raised = True # Should always get here + self.assertEqual( + str(e), + "The internal link to #footer on line 9 does not link to a" + " corresponding element with an id=\"footer\"." + ) + + self.assertTrue(assertion_raised) + class FileUploadTests(TestCase): def test_simple_upload(self): fd = open(os.path.join(os.path.dirname(__file__), "views.py")) Index: tests/regressiontests/test_client_regress/urls.py =================================================================== --- tests/regressiontests/test_client_regress/urls.py (revision 6352) +++ tests/regressiontests/test_client_regress/urls.py (working copy) @@ -5,5 +5,15 @@ (r'^no_template_view/$', views.no_template_view), (r'^file_upload/$', views.file_upload_view), (r'^get_view/$', views.get_view), - (r'^login_protected_redirect_view/$', views.login_protected_redirect_view) + (r'^login_protected_redirect_view/$', views.login_protected_redirect_view), + # All the following urls are for the assertNoBrokenLinks feature: + (r'^no_broken_links_view/$', views.no_broken_links_view), + (r'^broken_external_link_view/$', views.broken_external_link_view), + (r'^broken_internal_link_view/$', views.broken_internal_link_view), + (r'^bad_internal_link_view/$', views.bad_internal_link_view), + (r'^invalid_external_link_view/$', views.invalid_external_link_view), + (r'^blank_link_view/$', views.blank_link_view), + (r'^internal_page_link_view/$', views.internal_page_link_view), + (r'^broken_view/$', views.broken_view), + (r'^bad_view/$', views.bad_view) ) Index: AUTHORS =================================================================== --- AUTHORS (revision 6352) +++ AUTHORS (working copy) @@ -210,6 +210,7 @@ Jason McBrayer mccutchen@gmail.com michael.mcewan@gmail.com + Michael Nelson mikko@sorl.net Slawek Mikula mitakummaa@gmail.com Index: docs/testing.txt =================================================================== --- docs/testing.txt (revision 6352) +++ docs/testing.txt (working copy) @@ -846,6 +846,22 @@ The name is a string such as ``'admin/index.html'``. +``assertNoBrokenLinks(response, internal_only=True)`` + Asserts that all the anchor links within the response are not broken (ie. + result in a status of 200 or a redirect). By default only links internal to + the site will be checked (ie. those not beginning with http:// or https://). + + Internal page links such as are checked to ensure + that they are not broken (ie. that an element with the id="content" exists + on the page). + + Blank links, such as are also identified, which is helpful + to check when the url tag in fails. + + Note: As this assertion effectively clicks on all the links within the + response, care needs to be taken if any link has a side effect (such as + modifying your database). + E-mail services ---------------