URLBear: Regex for percentage-encoded URLs

Add regex for percentage-encoded URls like https://img.shields.io/badge/Maintained%3F-yes-green.svg Fixes #1290
coala · Oct 16, 2017 · 2ab24c9 · 2ab24c9
1 parent ffe13e3
commit 2ab24c9
Show file tree

Hide file tree

Showing 2 changed files with 27 additions and 6 deletions.
diff --git a/bears/general/URLBear.py b/bears/general/URLBear.py
@@ -108,12 +108,16 @@ def extract_links_from_file(file, link_ignore_regex, link_ignore_list):
             [^.:%\s_/?#[\]@\\]+         # Initial part of domain
             \.                          # A required dot `.`
             (
-                (?:[^\s()%\'"`<>|\\\[\]]+)  # Path name
-                                            # This part does not allow
-                                            # any parenthesis: balanced or
-                                            # unbalanced.
-            |                               # OR
-                \([^\s()%\'"`<>|\\\[\]]*\)  # Path name contained within ()
+                ((?:%[A-Fa-f0-9][A-Fa-f0-9])*[^\s()%\'"`<>|\\\[\]]+)
+                                        # Path name
+                                        # This part allows precentage
+                                        # encoding like %3F
+                                        # and does not allow
+                                        # any parenthesis: balanced or
+                                        # unbalanced.
+            |                           # OR
+                \((?:%[A-Fa-f0-9][A-Fa-f0-9])*[^\s()%\'"`<>|\\\[\]]*\)
+                                        # Path name contained within ()
                                         # This part allows path names that
                                         # are explicitly enclosed within one
                                         # set of parenthesis.

diff --git a/tests/general/URLBearTest.py b/tests/general/URLBearTest.py
@@ -60,6 +60,23 @@ def test_detect_url_result(self):
                              [3, 'http://www.google.com/404',
                               404, LINK_CONTEXT.no_context])
 
+    def test_precentage_encoded_url(self):
+        valid_file = """
+        # A url with a precentage-encoded character in path
+        https://img.shields.io/badge/Maintained%3F-yes-green.svg/200
+        """.splitlines()
+
+        with requests_mock.Mocker() as m:
+            m.add_matcher(custom_matcher)
+
+            result = get_results(self.uut, valid_file)
+            self.assertEqual(result[0].contents,
+                             [3,
+                              ('https://img.shields.io/badge/Maintained%3F-'
+                               'yes-green.svg/200'),
+                              200,
+                              LINK_CONTEXT.no_context])
+
 
 class URLResultTest(unittest.TestCase):