test_web: remove test_GET_DIRURL_large, it failed to exercise the problem in #237 and consumes a disproportionate amount of time

src/allmydata/test/test_web.py

@@ -650,37 +650,6 @@ class Web(WebMixin, unittest.TestCase):
                                   self.public_url + "/foo?t=bogus")
         return d
 
-    def test_GET_DIRURL_large(self):
-        # Nevow has a problem showing more than about 192 children of a
-        # directory: it uses defer.success() and d.addCallback in a way that
-        # can make the stack grow very quickly. See ticket #237 for details.
-        # To work around this, I think we'll need to put a 'return
-        # defer.fireEventually' in our render_ method. This test is intended
-        # to trigger the bug (and eventually verify that our workaround
-        # actually works), but it isn't yet failing for me.
-
-        d = self.s.create_empty_dirnode()
-        COUNT = 200
-        def _created(dirnode):
-            entries = [ (str(i), self._foo_uri) for i in range(COUNT) ]
-            d2 = dirnode.set_uris(entries)
-            d2.addCallback(lambda res: dirnode)
-            return d2
-        d.addCallback(_created)
-
-        def _check(dirnode):
-            large_url = "/uri/" + dirnode.get_uri() + "/"
-            return self.GET(large_url)
-        d.addCallback(_check)
-
-        def _done(res):
-            m = r'<a href="/uri/URI%3ADIR2%3A[^"]+">' + ("%d" % (COUNT-1)) + r'</a>'
-            self.failUnless(re.search(m, res))
-            self.failIf("maximum recursion depth exceeded" in res)
-        d.addCallback(_done)
-        return d
-    test_GET_DIRURL_large.timeout= 240 # this hits 120-sec timeout on overloaded vm buildslaves
-
     def test_GET_DIRURL_json(self):
         d = self.GET(self.public_url + "/foo?t=json")
         d.addCallback(self.failUnlessIsFooJSON)