python-rapidjson · hkxs · Dec 30, 2024 · Jan 2, 2025 · Jan 2, 2025 · Jan 2, 2025
diff --git a/README.rst b/README.rst
@@ -32,7 +32,7 @@ capabilities.
 
 Latest version documentation is automatically rendered by `Read the Docs`__.
 
-__ http://json-schema.org/documentation.html
+__ http://json-schema.org/docs.html
 __ https://python-rapidjson.readthedocs.io/en/latest/
 
 
@@ -123,3 +123,27 @@ some aspects. See `this section`__ in the documentation for further details.
 __ https://python-rapidjson.readthedocs.io/en/latest/quickstart.html#incompatibilities
 
 .. _RapidJSON: http://rapidjson.org/
+
+Regex Engine
+------------
+
+By default RapidJson uses a simple NFA regular expression engine for it's schema
+validation, see the section `RapidJson Regular Expression`__, it is possible
+to use ``std::regex`` instead of the original implementation by setting the
+environmental variable ``RAPIDJSON_SCHEMA_USE_STDREGEX=1``.
+
+.. code-block:: bash
+
+    $ export RAPIDJSON_SCHEMA_USE_STDREGEX=1
+    $ pip install python-rapidjson
+
+After installation, you can verify that std::regex is being used by checking the
+constant in the module:
+
+.. code-block:: python
+
+    >>> import rapidjson
+    >>> rapidjson.RAPIDJSON_SCHEMA_USE_STDREGEX
+    1
+
+__ http://rapidjson.org/md_doc_schema.html#Regex
diff --git a/docs/validator.rst b/docs/validator.rst
@@ -75,3 +75,27 @@
          Traceback (most recent call last):
            File "<stdin>", line 1, in <module>
          rapidjson.JSONDecodeError: Invalid JSON
+
+============
+Regex Engine
+============
+By default RapidJson uses a simple NFA regular expression engine for it's schema
+validation, see the section `RapidJson Regular Expression`__, it is possible
+to use ``std::regex`` instead of the original implementation by setting the
+environmental variable ``RAPIDJSON_SCHEMA_USE_STDREGEX=1``.
+
+.. code-block:: bash
+
+    $ export RAPIDJSON_SCHEMA_USE_STDREGEX=1
+    $ pip install python-rapidjson
+
+After installation, you can verify that std::regex is being used by checking the
+constant in the module:
+
+.. code-block:: python
+
+    >>> import rapidjson
+    >>> rapidjson.RAPIDJSON_SCHEMA_USE_STDREGEX
+    1
+
+__ http://rapidjson.org/md_doc_schema.html#Regex
diff --git a/rapidjson.cpp b/rapidjson.cpp
@@ -4046,6 +4046,7 @@ module_exec(PyObject* m)
                                    MM_COERCE_KEYS_TO_STRINGS)
         || PyModule_AddIntConstant(m, "MM_SKIP_NON_STRING_KEYS", MM_SKIP_NON_STRING_KEYS)
         || PyModule_AddIntConstant(m, "MM_SORT_KEYS", MM_SORT_KEYS)
+        || PyModule_AddIntConstant(m, "RAPIDJSON_SCHEMA_USE_STDREGEX", RAPIDJSON_SCHEMA_USE_STDREGEX)
 
         || PyModule_AddStringConstant(m, "__version__",
                                       STRINGIFY(PYTHON_RAPIDJSON_VERSION))

diff --git a/setup.py b/setup.py
@@ -8,6 +8,7 @@
 
 import os.path
 import sys
+from os import environ
 
 try:
     from setuptools import setup, Extension
@@ -52,13 +53,25 @@
 with open('CHANGES.rst', encoding='utf-8') as f:
     CHANGES = f.read()
 
+if environ.get('RAPIDJSON_SCHEMA_USE_STDREGEX', None):
+    RAPIDJSON_SCHEMA_USE_STDREGEX = 1
+    RAPIDJSON_SCHEMA_USE_INTERNALREGEX = 0
+else:
+    RAPIDJSON_SCHEMA_USE_STDREGEX = 0
+    RAPIDJSON_SCHEMA_USE_INTERNALREGEX = 1
+
 extension_options = {
     'sources': ['./rapidjson.cpp'],
     'include_dirs': [rj_include_dir],
-    'define_macros': [('PYTHON_RAPIDJSON_VERSION', VERSION)],
+    'define_macros': [
+        ('PYTHON_RAPIDJSON_VERSION', VERSION),
+        ('RAPIDJSON_SCHEMA_USE_INTERNALREGEX', RAPIDJSON_SCHEMA_USE_INTERNALREGEX),
+        ('RAPIDJSON_SCHEMA_USE_STDREGEX', RAPIDJSON_SCHEMA_USE_STDREGEX),
+    ],
     'depends': ['./rapidjson_exact_version.txt'],
 }
 
+
 if os.path.exists('rapidjson_exact_version.txt'):
     with open('rapidjson_exact_version.txt', encoding='utf-8') as f:
         extension_options['define_macros'].append(

diff --git a/tests/test_validator.py b/tests/test_validator.py
@@ -65,3 +65,22 @@ def test_invalid(schema, json, details):
 def test_additional_and_pattern_properties_valid(schema, json):
     validate = rj.Validator(schema)
     validate(json)
+
+
+def test_std_regex_used():
+    schema = rj.dumps({
+        "type": "object",
+        "patternProperties": {
+            "^(?!a).+": {
+                "type": "string"
+            }
+        }
+    })
+    validate = rj.Validator(schema)
+    validate('{"b": "A string"}')
+    if rj.RAPIDJSON_SCHEMA_USE_STDREGEX:
+        with pytest.raises(ValueError) as error:
+            validate('{"b": 1}')
+        assert error.value.args == ('patternProperties', '#', '#/b')
+    else:
+        validate('{"b": 1}')
diff --git a/typings/rapidjson/__init__.pyi b/typings/rapidjson/__init__.pyi
@@ -12,6 +12,8 @@ import typing as t
 __rapidjson_exact_version__: str
 __rapidjson_version__: str
 
+RAPIDJSON_SCHEMA_USE_STDREGEX: bool
+
 
 _JSONType = t.Union[
     str,