Provide own version of urlsplit and urldefrag that work with unknown schemes, fix #104

jsonschema/_utils.py

@@ -3,7 +3,7 @@ import json
 import re
 import os
 
-from jsonschema.compat import str_types, urlparse, MutableMapping
+from jsonschema.compat import str_types, MutableMapping, urlsplit
 
 
 class URIDict(MutableMapping):
@@ -13,7 +13,7 @@ class URIDict(MutableMapping):
     """
 
     def normalize(self, uri):
-        return urlparse.urlsplit(uri).geturl()
+        return urlsplit(uri).geturl()
 
     def __init__(self, *args, **kwargs):
         self.store = dict()

jsonschema/compat.py

@@ -11,15 +11,18 @@ PY3 = sys.version_info[0] >= 3
 
 if PY3:
     zip = zip
-    from urllib import parse as urlparse
-    from urllib.parse import unquote
+    from urllib.parse import (
+        unquote, urljoin, urlunsplit, SplitResult, urlsplit as _urlsplit
+    )
     from urllib.request import urlopen
     str_types = str,
     int_types = int,
     iteritems = operator.methodcaller("items")
 else:
     from itertools import izip as zip  # noqa
-    import urlparse  # noqa
+    from urlparse import (
+        urljoin, urlunsplit, SplitResult, urlsplit as _urlsplit # noqa
+    )
     from urllib import unquote  # noqa
     from urllib2 import urlopen  # noqa
     str_types = basestring
@@ -27,4 +30,22 @@ else:
     iteritems = operator.methodcaller("iteritems")
 
 
+# On python < 3.3 fragments are not handled properly with unknown schemes
+def urlsplit(url):
+    scheme, netloc, path, query, fragment = _urlsplit(url)
+    if "#" in path:
+        path, fragment = path.split("#", 1)
+    return SplitResult(scheme, netloc, path, query, fragment)
+
+
+def urldefrag(url):
+    if "#" in url:
+        s, n, p, q, frag = urlsplit(url)
+        defrag = urlunsplit((s, n, p, q, ''))
+    else:
+        defrag = url
+        frag = ''
+    return defrag, frag
+
+
 # flake8: noqa

jsonschema/validators.py

@@ -15,7 +15,7 @@ except ImportError:
 
 from jsonschema import _utils
 from jsonschema.compat import (
-    PY3, Sequence, urlparse, unquote, urlopen, str_types, int_types, iteritems,
+    PY3, Sequence, urljoin, urlsplit, urldefrag, unquote, urlopen, str_types, int_types, iteritems,
 )
 from jsonschema._format import FormatError
 
@@ -629,7 +629,7 @@ class RefResolver(object):
     @contextlib.contextmanager
     def in_scope(self, scope):
         old_scope = self.resolution_scope
-        self.resolution_scope = urlparse.urljoin(old_scope, scope)
+        self.resolution_scope = urljoin(old_scope, scope)
         try:
             yield
         finally:
@@ -645,8 +645,8 @@ class RefResolver(object):
 
         """
 
-        full_uri = urlparse.urljoin(self.resolution_scope, ref)
-        uri, fragment = urlparse.urldefrag(full_uri)
+        full_uri = urljoin(self.resolution_scope, ref)
+        uri, fragment = urldefrag(full_uri)
         if not uri:
             uri = self.base_uri
 
@@ -718,7 +718,7 @@ class RefResolver(object):
 
         """
 
-        scheme = urlparse.urlsplit(uri).scheme
+        scheme = urlsplit(uri).scheme
 
         if scheme in self.handlers:
             result = self.handlers[scheme](uri)

test_jsonschema.py

@@ -861,11 +861,11 @@ class TestRefResolver(unittest.TestCase):
             self.assertEqual(resolved, self.referrer["properties"]["foo"])
 
     def test_it_resolves_local_refs_with_id(self):
-        schema = {"id": "/bar/schema#", "a": {"foo": "bar"}}
+        schema = {"id": "foo://bar/schema#", "a": {"foo": "bar"}}
         resolver = RefResolver.from_schema(schema)
         with resolver.resolving("#/a") as resolved:
             self.assertEqual(resolved, schema["a"])
-        with resolver.resolving("/bar/schema#/a") as resolved:
+        with resolver.resolving("foo://bar/schema#/a") as resolved:
             self.assertEqual(resolved, schema["a"])
 
     def test_it_retrieves_stored_refs(self):