v0.11.3

• Fixed configuration file unmarshalling of JSON floating-point values (#253). The unmarshalling of primitive types has been improved to ensure accurate conversion and prevent potential errors or data corruption. The updated functionality now correctly handles the conversion of JSON floating-point values to integers by refusing to truncate precision and instead raising a SerdeError when necessary. Additionally, string-to-boolean conversions are now strictly validated to only accept true or false (case-insensitive). Furthermore, configuration file unmarshalling has been enhanced with additional type checks to verify that loaded values match the expected types, such as strings, integers, and floats, thereby preventing incorrect type conversions and ensuring that the loaded data retains its original precision and type.
• Fixed unmarshalling of forward references on Python ≥ 3.12.4 (#252). The library's unmarshalling functionality has been updated to support forward references on Python versions 3.12.4 and later, which introduced changes to the _evaluate method of ForwardRef. A new internal utility method has been added to handle these changes, ensuring compatibility with different Python versions by conditionally passing the recursive_guard parameter as a keyword argument and including additional type information. Additionally, a new test function has been introduced to verify the correct handling of forward references in class fields, simulating future annotations and testing the save and load process of an instance with various field types, including strings, integers, and JSON values, to ensure that forward references are correctly resolved.
• Support detecting file encoding from XML declaration (#254). The library's file encoding detection capabilities have been enhanced to support XML files, allowing for the extraction of encoding information from the XML declaration at the start of the file. A new detection method has been introduced, which reads the initial bytes of the file to determine the potential encoding, attempts to decode the XML declaration, and returns the specified encoding if successful. This functionality is accessible through the updated decode_with_bom and read_text functions, which now accept an optional detect_xml parameter to enable XML declaration-based encoding detection. If no encoding is detected via the byte order mark (BOM) or XML declaration, the library defaults to the locale's preferred encoding. Additionally, new test cases have been added to verify the correct detection of XML file encodings, including scenarios with encoding declarations, byte order marks, and default UTF-8 encoding.

Contributors: @asnare, @sundarshankar89