All notable changes to this project will be documented in this file. This project adheres to Semantic Versioning.
Nothing yet.
0-21 - 2019-11-25
Added:
- array.array now has support in to_array(), but no type checking yet
0-20 - 2019-07-07
Changed:
- Updated to work with Python3.7
- Now typing module is a hard dependency as the fake_typing module was too hard to support
0-17 - 2019-05-03
Fixed:
- more slight performance tweaks
0-16 - 2019-05-02
Fixed:
- some slight performance issues
- to_array now support unicode
Changed:
- Anno() doesn't take typ anymore in the constructor, use set_typ() instead
0-15 - 2019-03-21
Fixed:
- serialize_object now works on Jython with scisoftpy
0-14 - 2019-03-21
Fixed:
- serialize_object now works on platforms without enum package
0-13 - 2019-03-20
Added:
- Sphinx call_types extension to get correct Serializable documentation
- json_encode and json_decode functions that operate on Serializable
0-12 - 2019-03-12
Fixed:
- Fix serialization of nested objects in Arrays
0-11 - 2019-03-04
Added:
- typeid based Serializable, serialize_object and deserialize_object
0-10 - 2018-12-13
Fixed:
- Fix Array.__ne__ behaviour, and fix Array.__eq__ for numpy arrays
0-9-1 - 2018-10-29
Fixed:
- Fix "make install" step in Makefile used at Diamond Light Source
0-9 - 2018-07-03
Fixed:
- Make Array.__eq__ behave like list.__eq__ even with numpy arrays
0-8 - 2018-04-25
Fixed:
- self and cls as first arguments are now optionally given types
- make_annotations with optional globals_d now works with dotted names
0-7 - 2018-04-24
Fixed:
- make_annotations now always returns a dict
0-6 - 2018-04-23
Added:
- Add make_annotations to public API, and allow globals_d to be optional
0-5 - 2018-02-16
Added:
- Support for *args and **kwargs annotations
- Support for bare Any as return type
- Support for setting Anno.default in constructor
- Support for Array.__eq__
Changed:
- Annotation creation now only takes into account globals, not locals
Fixed:
- WithCallTypes subclasses may now also subclass Generic
- Array handling of numpy ndarrays better
- WithCallTypes subclasses with no __init__ now works
- Repr on instance where not all call_types are stored attributes now works
0-4 - 2018-01-10
Fixed:
- Example name that may cause Pycharm to fail
- Allow to_array to take sequence=None
0-3 - 2018-01-04
Added:
- Support for Any
0-2 - 2018-01-04
Added:
- Support for Mapping
0-1-1 - 2018-01-02
Fixed:
- Fixed PyPI packaging
Initial release