From 496431b4f0d9c7c430f6f2330434e51bfa080ef4 Mon Sep 17 00:00:00 2001
From: Xavier Morel <xmo@openerp.com>
Date: Tue, 5 Nov 2013 15:50:47 +0100
Subject: [PATCH] [ADD] CMS crawling case, with a bunch of hardcoded values to
 fix

bzr revid: xmo@openerp.com-20131105145047-4ccrchj79lohry5q
---
 addons/website/tests/__init__.py      |   2 +-
 addons/website/tests/cases.py         |  32 ++++++++
 addons/website/tests/test_requests.py | 111 ++++++++++++++++++++++++++
 3 files changed, 144 insertions(+), 1 deletion(-)
 create mode 100644 addons/website/tests/cases.py
 create mode 100644 addons/website/tests/test_requests.py

diff --git a/addons/website/tests/__init__.py b/addons/website/tests/__init__.py
index b00c5984aa0..24de4d7ddad 100644
--- a/addons/website/tests/__init__.py
+++ b/addons/website/tests/__init__.py
@@ -1,4 +1,4 @@
 # -*- coding: utf-8 -*-
-import test_views, test_converter
+import test_views, test_converter, test_requests
 
 checks = [ test_views, test_converter, ]
diff --git a/addons/website/tests/cases.py b/addons/website/tests/cases.py
new file mode 100644
index 00000000000..246fe6f2320
--- /dev/null
+++ b/addons/website/tests/cases.py
@@ -0,0 +1,32 @@
+# -*- coding: utf-8 -*-
+import unittest2
+
+class URLCase(unittest2.TestCase):
+    """
+    URLCase moved out of test_requests, otherwise discovery attempts to
+    instantiate and run it
+    """
+    def __init__(self, user, url, result):
+        super(URLCase, self).__init__()
+        self.user = user
+        self.url = url
+        self.result = result
+
+    @property
+    def username(self):
+        return self.user or "Anonymous Coward"
+
+    def __str__(self):
+        return "%s (as %s)" % (self.url, self.username)
+
+    __repr__ = __str__
+
+    def shortDescription(self):
+        return ""
+
+    def runTest(self):
+        code = self.result.getcode()
+        self.assertIn(
+            code, xrange(200, 300),
+            "Fetching %s as %s returned an error response (%d)" % (
+                self.url, self.username, code))
diff --git a/addons/website/tests/test_requests.py b/addons/website/tests/test_requests.py
new file mode 100644
index 00000000000..8d1e7baf469
--- /dev/null
+++ b/addons/website/tests/test_requests.py
@@ -0,0 +1,111 @@
+# -*- coding: utf-8 -*-
+import collections
+import urllib
+import urlparse
+import unittest2
+import urllib2
+
+import lxml.html
+
+from openerp import tools
+
+from . import cases
+
+__all__ = ['load_tests', 'CrawlSuite']
+
+class RedirectHandler(urllib2.HTTPRedirectHandler):
+    """
+    HTTPRedirectHandler is predicated upon HTTPErrorProcessor being used and
+    works by intercepting 3xy "errors".
+
+    Inherit from it to handle 3xy non-error responses instead, as we're not
+    using the error processor
+    """
+
+    def http_response(self, request, response):
+        code, msg, hdrs = response.code, response.msg, response.info()
+
+        if 300 <= code < 400:
+            return self.parent.error(
+                'http', request, response, code, msg, hdrs)
+
+        return response
+
+    https_response = http_response
+
+class CrawlSuite(unittest2.TestSuite):
+    """ Test suite crawling an openerp CMS instance and checking that all
+    internal links lead to a 200 response.
+
+    If a username and a password are provided, authenticates the user before
+    starting the crawl
+    """
+
+    def __init__(self, user=None, password=None):
+        super(CrawlSuite, self).__init__()
+
+        self.opener = urllib2.OpenerDirector()
+        self.opener.add_handler(urllib2.UnknownHandler())
+        self.opener.add_handler(urllib2.HTTPHandler())
+        self.opener.add_handler(urllib2.HTTPSHandler())
+        self.opener.add_handler(RedirectHandler())
+        self.opener.add_handler(urllib2.HTTPCookieProcessor())
+
+        self._authenticate(user, password)
+        self.user = user
+
+    def _request(self, path):
+        # FIXME: host, db?
+        return self.opener.open(urlparse.urlunsplit([
+            'http', 'localhost:%s' % tools.config['xmlrpc_port'],
+            path, '', ''
+        ]))
+
+    def _authenticate(self, user, password):
+        if user is None: return
+
+        # FIXME: host, port, db?
+        auth = self.opener.open(
+            'http://localhost:8069/login?' + urllib.urlencode(
+                {'db': 'website', 'login': user, 'key': password, }))
+        assert auth.getcode() < 400, "Auth failure %d" % auth.getcode()
+
+    def _wrapped_run(self, result, debug=False):
+        paths = collections.deque(['/'])
+        seen = set(paths)
+
+        while paths:
+            url = paths.popleft()
+            r = self._request(url)
+            cases.URLCase(self.user, url, r).run(result)
+
+            if r.info().gettype() != 'text/html':
+                continue
+
+            doc = lxml.html.fromstring(r.read())
+            for link in doc.xpath('//a[@href]'):
+                href = link.get('href')
+
+                # avoid repeats, even for links we won't crawl no need to
+                # bother splitting them if we've already ignored them
+                # previously
+                if href in seen: continue
+                seen.add(href)
+
+                parts = urlparse.urlsplit(href)
+
+                if parts.netloc or \
+                    not parts.path.startswith('/') or \
+                    parts.path == '/web' or\
+                    parts.path.startswith('/web/') or \
+                    (parts.scheme and parts.scheme not in ('http', 'https')):
+                    continue
+
+                paths.append(href)
+
+def load_tests(loader, base, _):
+    base.addTest(CrawlSuite())
+    # blog duplicate (&al?) are on links
+    #base.addTest(CrawlSuite('admin', 'admin'))
+    #base.addTest(CrawlSuite('demo', 'demo'))
+    return base