merge from 3.3

#17403: urllib.parse.robotparser normalizes the urls before adding to
ruleline. This helps in handling certain types invalid urls in a conservative
manner. Patch contributed by Mher Movsisyan.

2 files changed, 13 insertions, 0 deletions

diff --git a/Lib/test/test_robotparser.py b/Lib/test/test_robotparser.py
index 8c09e7452c..d1dfd9eeec 100644
--- a/Lib/test/test_robotparser.py
+++ b/Lib/test/test_robotparser.py
@@ -234,6 +234,18 @@ bad = ['/some/path']
 
 RobotTest(15, doc, good, bad)
 
+# 16. Empty query (issue #17403). Normalizing the url first.
+doc = """
+User-agent: *
+Allow: /some/path?
+Disallow: /another/path?
+"""
+
+good = ['/some/path?']
+bad = ['/another/path?']
+
+RobotTest(16, doc, good, bad)
+
 
 class NetworkTestCase(unittest.TestCase):
 
diff --git a/Lib/urllib/robotparser.py b/Lib/urllib/robotparser.py
index 75be4af409..978ba58d84 100644
--- a/Lib/urllib/robotparser.py
+++ b/Lib/urllib/robotparser.py
@@ -157,6 +157,7 @@ class RuleLine:
         if path == '' and not allowance:
             # an empty value means allow all
             allowance = True
+        path = urllib.parse.urlunparse(urllib.parse.urlparse(path))
         self.path = urllib.parse.quote(path)
         self.allowance = allowance