zephyr: Fix broken postgres regular expression logic and add tests.

Like many rare-case code with new tests, it turns out that the logic
for handling null characters in our Zephyr postgres query escaping
never worked, in multiple ways.  First, it always changed the second
character in s, not the current one being inspected, and second, the
value it replaced it with was no the correct postgres escape of the
null byte.  We fix this and add tests.

This completes the effort to get zerver/views/messages.py to 100%
test coverage.

Fixes #1006.

tools/test-backend

@@ -96,7 +96,6 @@ not_yet_fully_covered = {
     'zerver/tests/tests.py',
     # Getting views file coverage to 100% is a major project goal
     'zerver/views/auth.py',
-    'zerver/views/messages.py',
     'zerver/views/home.py',
     'zerver/views/registration.py',
     'zerver/views/events_register.py',

zerver/tests/test_narrow.py

@@ -85,6 +85,18 @@ class NarrowBuilderTest(ZulipTestCase):
         term = dict(operator='stream', operand='Scotland')
         self._do_add_term_test(term, 'WHERE recipient_id = :recipient_id_1')
 
+    def test_add_term_using_stream_operator_zephyr_hack_null(self):
+        # type: () -> None
+        """This tests the _pg_re_escape code path with a null byte"""
+        null_stream_name = u'fun\u0000test'
+        s = Stream.objects.create(realm=get_realm("mit"), name=null_stream_name)
+        Recipient.objects.create(type=Recipient.STREAM, type_id=s.id)
+        user_profile = get_user_profile_by_email("sipbtest@mit.edu")
+        builder = NarrowBuilder(user_profile, column('id'))
+        term = dict(operator='stream', operand=null_stream_name)
+        result = str(builder.add_term(self.raw_query, term))
+        self.assertTrue("WHERE recipient_id IN (:recipient_id_1)" in result)
+
     def test_add_term_using_stream_operator_and_negated(self):  # NEGATED
         # type: () -> None
         term = dict(operator='stream', operand='Scotland', negated=True)
@@ -541,6 +553,22 @@ class GetOldMessagesTest(ZulipTestCase):
             stream_id = stream_messages[i].recipient.id
             self.assertEqual(message["recipient_id"], stream_id)
 
+    def test_get_old_messages_with_narrow_stream_mit_unicode_null_regex(self):
+        # type: () -> None
+        self.login("starnine@mit.edu")
+        null_stream_name = u"foo\000-stream"
+        self.subscribe_to_stream("starnine@mit.edu", null_stream_name)
+        self.send_message("starnine@mit.edu", null_stream_name, Recipient.STREAM)
+        narrow = [dict(operator='stream', operand=null_stream_name)]
+        result = self.get_and_check_messages(dict(num_before=2, num_after=2,
+                                                  narrow=ujson.dumps(narrow)))
+
+        messages = get_user_messages(get_user_profile_by_email("starnine@mit.edu"))
+        stream_messages = [msg for msg in messages if msg.recipient.type == Recipient.STREAM]
+        self.assertEqual(len(result["messages"]), 1)
+        self.assertEqual(result["messages"][0]["type"], "stream")
+        self.assertEqual(result["messages"][0]["recipient_id"], stream_messages[0].recipient.id)
+
     def test_get_old_messages_with_narrow_topic_mit_unicode_regex(self):
         # type: () -> None
         """

zerver/views/messages.py

@@ -154,7 +154,7 @@ class NarrowBuilder(object):
         for i, c in enumerate(s):
             if c not in self._alphanum:
                 if c == '\000':
-                    s[1] = '\\000'
+                    s[i] = '\0'
                 elif ord(c) >= 128:
                     # convert the character to hex postgres regex will take
                     # \uXXXX