potential fix for #189

caldav/lib/url.py

@@ -4,10 +4,11 @@
 from six import PY3
 from caldav.lib.python_utilities import to_unicode, to_normal_str
 if PY3:
-    from urllib.parse import ParseResult, SplitResult, urlparse
+    from urllib.parse import ParseResult, SplitResult, urlparse, unquote, quote, urlunparse
 else:
     from urlparse import ParseResult, SplitResult
-    from urlparse import urlparse
+    from urlparse import urlparse, urlunparse
+    from urllib import unquote, quote
 
 
 def uc2utf8(input):
@@ -133,21 +134,32 @@ def canonical(self):
         """
         a canonical URL ... remove authentication details, make sure there
         are no double slashes, and to make sure the URL is always the same,
-        run it through the urlparser
+        run it through the urlparser, and make sure path is properly quoted
         """
         url = self.unauth()
 
         # this is actually already done in the unauth method ...
         if '//' in url.path:
             raise NotImplementedError("remove the double slashes")
 
-        # This looks like a noop - but it may have the side effect
-        # that urlparser be run (actually not - unauth ensures we
-        # have an urlParseResult object)
-        url.scheme
+        arr = list(self.url_parsed)
+        ## quoting path
+        arr[2] = quote(unquote(url.path))
+        ## sensible defaults
+        if not arr[0]:
+            arr[0] = 'https'
+        if arr[1] and not ':' in arr[1]:
+            if arr[0] == 'https':
+                portpart = ':443'
+            elif arr[0] == 'http':
+                portpart = ':80'
+            else:
+                portpart = ''
+            arr[1] += portpart
 
         # make sure to delete the string version
-        url.url_raw = None
+        url.url_raw = urlunparse(arr)
+        url.url_parsed = None
 
         return url
 

caldav/objects.py

@@ -117,8 +117,8 @@ def children(self, type=None):
                 # And why is the strip_trailing_slash-method needed?
                 # The collection URL should always end with a slash according
                 # to RFC 2518, section 5.2.
-                if (self.url.strip_trailing_slash() !=
-                        self.url.join(path).strip_trailing_slash()):
+                if (self.url.canonical().strip_trailing_slash() !=
+                        self.url.join(path).canonical().strip_trailing_slash()):
                     c.append((self.url.join(path), resource_types,
                               resource_name))
 

tests/test_caldav_unit.py

@@ -769,6 +769,9 @@ def testURL(self):
         assert_equal(URL('http://www.example.com:8080/bar/').strip_trailing_slash(), URL('http://www.example.com:8080/bar'))
         assert_equal(URL('http://www.example.com:8080/bar/').strip_trailing_slash(), URL('http://www.example.com:8080/bar').strip_trailing_slash())
 
+        # 9) canonical
+        assert_equal(URL('https://www.example.com:443/b%61r/').canonical(), URL('//www.example.com/bar/').canonical())
+
     def testFilters(self):
         filter = \
             cdav.Filter().append(