Preserve empty JSON Pointer keys and unique validator names

CHANGELOG.txt

@@ -1,5 +1,7 @@
 === UNRELEASED
 
+* Fixed references through empty root keys and colliding generated validator names
+
 * Fixed property names with `{` or `}` being evaluated as format string in error messages
 * Fixed numeric keywords (minimum, maximum, multipleOf, exclusiveMinimum, exclusiveMaximum) being applied to booleans
 * Fixed `propertyNames: false` crashing on non-object values

fastjsonschema/ref_resolver.py

@@ -35,8 +35,10 @@ def resolve_path(schema: Any, fragment: str) -> Any:
 
     Path is unescaped according https://tools.ietf.org/html/rfc6901
     """
-    fragment = fragment.lstrip('/')
-    parts = unquote(fragment).split('/') if fragment else []
+    if not fragment:
+        return schema
+    fragment = fragment[1:] if fragment.startswith('/') else fragment
+    parts = unquote(fragment).split('/')
     for part in parts:
         part = part.replace('~1', '/').replace('~0', '~')
         if isinstance(schema, list):
@@ -76,7 +78,7 @@ def resolve_remote(uri: str, handlers: Mapping[str, Callable[[str], Any]]) -> An
     return result
 
 
-class RefResolver:
+class RefResolver:  # pylint: disable=too-many-instance-attributes
     """
     Resolve JSON References.
     """
@@ -105,6 +107,8 @@ def __init__(
         self.store = store
         self.cache = cache
         self.handlers = handlers
+        self._scope_names: dict[str, str] = {}
+        self._used_scope_names: set[str] = set()
         self._walked_uris: set[str] = set()
         self.walk(schema)
         self._walked_uris.add(normalize(base_uri) if base_uri else '')
@@ -185,11 +189,21 @@ def get_uri(self) -> str:
 
     def get_scope_name(self) -> str:
         """
-        Get current scope and return it as a valid function name.
+        Get a stable, unique function name for the current scope.
         """
+        uri = self.get_uri()
+        if uri in self._scope_names:
+            return self._scope_names[uri]
         name = 'validate_' + unquote(self.resolution_scope).replace('~1', '_').replace('~0', '_').replace('"', '')
         name = re.sub(r'($[^a-zA-Z]|[^a-zA-Z0-9])', '_', name)
         name = name.lower().rstrip('_')
+        # Distinct URIs, such as the root and #/, can sanitize to the same name.
+        base_name, suffix = name, 1
+        while name in self._used_scope_names:
+            name = '{}_{}'.format(base_name, suffix)
+            suffix += 1
+        self._scope_names[uri] = name
+        self._used_scope_names.add(name)
         return name
 
     def walk(self, node: dict[str, Any] | bool, depth: int = 0, rewrite_refs: bool = True) -> None:

tests/test_ref_resolver.py

@@ -0,0 +1,82 @@
+import pytest
+
+from fastjsonschema import JsonSchemaValueException, compile, compile_to_code
+from fastjsonschema.ref_resolver import RefResolver, resolve_path
+
+
+@pytest.mark.parametrize('fragment, expected', [
+    ('/', {'child': 1, '': 2, 'a/b': 3, '~': 4, 'café': 5}),
+    ('//child', 1),
+    ('//', 2),
+    ('//a~1b', 3),
+    ('//~0', 4),
+    ('//caf%C3%A9', 5),
+])
+def test_empty_leading_pointer_tokens(fragment, expected):
+    schema = {'': {'child': 1, '': 2, 'a/b': 3, '~': 4, 'café': 5}}
+    assert resolve_path(schema, fragment) == expected
+
+
+def test_root_and_ordinary_pointer_tokens():
+    schema = {'items': [{'type': 'integer'}]}
+    assert resolve_path(schema, '') is schema
+    assert resolve_path(schema, '/items/0') == {'type': 'integer'}
+    assert resolve_path(schema, 'items/0') == {'type': 'integer'}
+
+
+@pytest.mark.parametrize('generated', [False, True])
+@pytest.mark.parametrize('ref, subschema', [
+    ('#/', {'type': 'integer'}),
+    ('#//child', {'child': {'type': 'integer'}}),
+    ('#//', {'': {'type': 'integer'}}),
+])
+def test_compile_references_through_empty_root_key(generated, ref, subschema):
+    schema = {
+        '': subschema,
+        'type': 'object',
+        'properties': {'value': {'$ref': ref}},
+    }
+    if generated:
+        namespace = {}
+        exec(compile_to_code(schema), namespace)
+        validator = namespace['validate']
+    else:
+        validator = compile(schema)
+
+    assert validator({'value': 42}) == {'value': 42}
+    with pytest.raises(JsonSchemaValueException):
+        validator({'value': 'wrong type'})
+
+
+@pytest.mark.parametrize('generated', [False, True])
+def test_empty_key_in_handler_loaded_schema(generated):
+    remote = {
+        '': {'type': 'integer'},
+        'type': 'object',
+        'properties': {'value': {'$ref': '#/'}},
+    }
+    schema = {'$ref': 'https://schemas.example/schema.json'}
+    handlers = {'https': lambda _uri: remote}
+    if generated:
+        namespace = {}
+        exec(compile_to_code(schema, handlers=handlers), namespace)
+        validator = namespace['validate']
+    else:
+        validator = compile(schema, handlers=handlers)
+
+    assert validator({'value': 42}) == {'value': 42}
+    with pytest.raises(JsonSchemaValueException):
+        validator({'value': 'wrong type'})
+
+
+def test_empty_key_scope_names_stay_unique_and_stable():
+    resolver = RefResolver.from_schema({})
+    root_name = resolver.get_scope_name()
+    with resolver.in_scope('#/'):
+        child_name = resolver.get_scope_name()
+        assert resolver.get_scope_name() == child_name
+    with resolver.in_scope('#//'):
+        grandchild_name = resolver.get_scope_name()
+    with resolver.in_scope('#'):
+        assert resolver.get_scope_name() == root_name
+    assert len({root_name, child_name, grandchild_name}) == 3