Fix URL checking for Windows OS

  - Add tests for URL helper function and remove unused code

tests/test_resources.py

@@ -28,6 +28,7 @@
     etree_element, py_etree_element
 from xmlschema.namespaces import XSD_NAMESPACE
 from xmlschema.helpers import is_etree_element
+from xmlschema.resources import is_url, is_local_url, is_remote_url, url_path_is_file
 from xmlschema.documents import get_context
 from xmlschema.testing import SKIP_REMOTE_TESTS
 
@@ -94,6 +95,29 @@ def check_url(self, url, expected):
             expected_path = PurePath(expected_parts.path)
         self.assertEqual(path, expected_path, "%r: Paths differ." % url)
 
+    def test_url_helper_functions(self):
+        self.assertTrue(is_url(self.col_xsd_file))
+        self.assertFalse(is_url(' \t<root/>'))
+        self.assertFalse(is_url('line1\nline2'))
+        self.assertFalse(is_url(None))
+
+        self.assertTrue(is_local_url(self.col_xsd_file))
+        self.assertTrue(is_local_url('/home/user/'))
+        self.assertTrue(is_local_url('/home/user/schema.xsd'))
+        self.assertTrue(is_local_url('  /home/user/schema.xsd  '))
+        self.assertTrue(is_local_url('C:\\Users\\foo\\schema.xsd'))
+        self.assertTrue(is_local_url(' file:///home/user/schema.xsd'))
+        self.assertFalse(is_local_url('http://example.com/schema.xsd'))
+
+        self.assertFalse(is_remote_url(self.col_xsd_file))
+        self.assertFalse(is_remote_url('/home/user/'))
+        self.assertFalse(is_remote_url('/home/user/schema.xsd'))
+        self.assertFalse(is_remote_url(' file:///home/user/schema.xsd'))
+        self.assertTrue(is_remote_url('  http://example.com/schema.xsd'))
+
+        self.assertTrue(url_path_is_file(self.col_xml_file))
+        self.assertFalse(url_path_is_file(self.col_dir))
+
     def test_normalize_url_posix(self):
         url1 = "https://example.com/xsd/other_schema.xsd"
         self.check_url(normalize_url(url1, base_url="/path_my_schema/schema.xsd"), url1)

xmlschema/resources.py

@@ -42,65 +42,7 @@ class XmlResourceXPathParser(XPath1Parser):
 
 
 ###
-# Internal helper functions
-
-def is_url(obj):
-    """
-    Checks if and object can be an URL, restricting to strings that cannot be XML data.
-    """
-    if not isinstance(obj, (str, bytes)):
-        return False
-    elif '\n' in obj or obj.lstrip().startswith('<'):
-        return False
-
-    try:
-        urlsplit(obj)
-    except ValueError:
-        return False
-    else:
-        return True
-
-
-def is_remote_url(url):
-    return is_url(url) and urlsplit(url).scheme not in ('', 'file')
-
-
-def is_local_url(url):
-    return is_url(url) and urlsplit(url).scheme in ('', 'file')
-
-
-def url_path_is_directory(url):
-    return is_local_url(url) and os.path.isdir(urlsplit(url).path)
-
-
-def url_path_is_file(url):
-    return is_local_url(url) and os.path.isfile(urlsplit(url).path)
-
-
-def update_prefix(namespaces, prefix, uri):
-    """Update namespace registration without overwrite an existing one."""
-    if not prefix:
-        if '' not in namespaces:
-            namespaces[prefix] = uri
-            return
-        elif namespaces[''] == uri:
-            return
-        prefix = 'default'
-
-    while prefix in namespaces:
-        if namespaces[prefix] == uri:
-            return
-        match = re.search(r'(\d+)$', prefix)
-        if match:
-            index = int(match.group()) + 1
-            prefix = prefix[:match.span()[0]] + str(index)
-        else:
-            prefix += '0'
-    namespaces[prefix] = uri
-
-
-###
-# API for XML resources
+# URL normalization (that fixes many headaches :)
 
 def normalize_url(url, base_url=None, keep_relative=False):
     """
@@ -191,6 +133,65 @@ def filter_url(x):
     return filter_url(normalized_url)
 
 
+###
+# Internal helper functions
+
+def is_url(obj):
+    """
+    Checks if and object can be an URL, restricting to strings that cannot be XML data.
+    """
+    if not isinstance(obj, (str, bytes)):
+        return False
+    elif '\n' in obj or obj.lstrip().startswith('<'):
+        return False
+
+    try:
+        urlsplit(obj.strip())
+    except ValueError:
+        return False
+    else:
+        return True
+
+
+def is_remote_url(url):
+    return is_url(url) and urlsplit(normalize_url(url)).scheme not in ('', 'file')
+
+
+def is_local_url(url):
+    return is_url(url) and urlsplit(normalize_url(url)).scheme in ('', 'file')
+
+
+def url_path_is_file(url):
+    if not is_local_url(url):
+        return False
+    return os.path.isfile(url) or os.path.isfile(urlsplit(normalize_url(url)).path)
+
+
+def update_prefix(namespaces, prefix, uri):
+    """Update namespace registration without overwrite an existing one."""
+    if not prefix:
+        if '' not in namespaces:
+            namespaces[prefix] = uri
+            return
+        elif namespaces[''] == uri:
+            return
+        prefix = 'default'
+
+    while prefix in namespaces:
+        if namespaces[prefix] == uri:
+            return
+        match = re.search(r'(\d+)$', prefix)
+        if match:
+            index = int(match.group()) + 1
+            prefix = prefix[:match.span()[0]] + str(index)
+        else:
+            prefix += '0'
+    namespaces[prefix] = uri
+
+
+###
+# API for XML resources
+
 def normalize_locations(locations, base_url=None, keep_relative=False):
     """
     Returns a list of normalized locations. The locations are normalized using

xmlschema/validators/schema.py

@@ -302,7 +302,8 @@ def __init__(self, source, namespace=None, validation='strict', global_maps=None
             logger.setLevel(logging.WARNING)
 
         if allow == 'sandbox' and base_url is None and is_local_url(source):
-            base_url = os.path.dirname(source)
+            # Allow sandbox mode without a base_url using the initial schema URL as base
+            base_url = os.path.dirname(normalize_url(source))
 
         self.source = XMLResource(source, base_url, allow, defuse, timeout, lazy=False)
         logger.debug("Read schema from %r", self.source)