From 4862a7d7fea87e0861ff5429673b8ab131f6a1e9 Mon Sep 17 00:00:00 2001
From: alexanderlukanin13 <alexander.lukanin.13@gmail.com>
Date: Sat, 11 Jan 2014 17:29:44 +0600
Subject: [PATCH] RobotsTxtMiddlewareTest

---
 .../test_downloadermiddleware_robotstxt.py    | 35 ++++++++++++++-----
 1 file changed, 27 insertions(+), 8 deletions(-)

diff --git a/scrapy/tests/test_downloadermiddleware_robotstxt.py b/scrapy/tests/test_downloadermiddleware_robotstxt.py
index fbd4b2339..a98522510 100644
--- a/scrapy/tests/test_downloadermiddleware_robotstxt.py
+++ b/scrapy/tests/test_downloadermiddleware_robotstxt.py
@@ -1,9 +1,10 @@
+import re
 import mock
 from twisted.internet import reactor
 from twisted.internet.defer import Deferred
 from twisted.trial import unittest
 from scrapy.contrib.downloadermiddleware.robotstxt import RobotsTxtMiddleware
-from scrapy.exceptions import IgnoreRequest
+from scrapy.exceptions import IgnoreRequest, NotConfigured
 from scrapy.http import Request, Response
 from scrapy.settings import CrawlerSettings
 
@@ -14,14 +15,32 @@ class RobotsTxtMiddlewareTest(unittest.TestCase):
         crawler = mock.MagicMock()
         crawler.settings = CrawlerSettings()
         crawler.settings.overrides['USER_AGENT'] = 'CustomAgent'
+        self.assertRaises(NotConfigured, RobotsTxtMiddleware, crawler)
         crawler.settings.overrides['ROBOTSTXT_OBEY'] = True
         crawler.engine.download = mock.MagicMock()
-        deferred = Deferred()
-        crawler.engine.download.return_value = deferred
-        ROBOTS = ''
+        ROBOTS = re.sub(r'^\s+(?m)', '', '''
+        User-Agent: *
+        Disallow: /admin/
+        Disallow: /static/
+        ''')
         response = Response('http://site.local/robots.txt', body=ROBOTS)
-        reactor.callLater(0, deferred.callback, response)
+        def return_response(request, spider):
+            deferred = Deferred()
+            reactor.callFromThread(deferred.callback, response)
+            return deferred
+        crawler.engine.download.side_effect = return_response
         middleware = RobotsTxtMiddleware(crawler)
-        spider = None # Not actually used
-        self.assertIsNone(middleware.process_request(Request('http://site.local/dummy'), spider))
-        self.assertRaises(IgnoreRequest, middleware.process_request, Request('http://site.local/forbidden'), spider)
\ No newline at end of file
+        spider = None  # not actually used
+        # There is a bit of neglect in robotstxt.py: robots.txt is fetched asynchronously,
+        # and it is actually fetched only *after* first process_request completes.
+        # So, first process_request will always succeed.
+        # We defer test() because otherwise robots.txt download mock will be called after assertRaises failure.
+        self.assertIsNone(middleware.process_request(Request('http://site.local'), spider))  # not affected by robots.txt
+        def test(r):
+            self.assertIsNone(middleware.process_request(Request('http://site.local/allowed'), spider))
+            self.assertRaises(IgnoreRequest, middleware.process_request, Request('http://site.local/admin/main'), spider)
+            self.assertRaises(IgnoreRequest, middleware.process_request, Request('http://site.local/static/'), spider)
+        deferred = Deferred()
+        deferred.addCallback(test)
+        reactor.callFromThread(deferred.callback, None)
+        return deferred