0.7.2 (2025.10.06)
* Fix broken decoding of multiple chunks of UTF-16 in some cases
* Slightly optimize decode iterator

0.7.1 (2025.09.06)
* Fix decoding of 4-byte GB-18030 characters
* Reduce map sizes
* Unbreak compilation on Nim 1.6.10

0.7.0 (2024.12.07)
* Update GB-18030 encoder and decoder with the spec changes from October
This is a breaking change - although it does not change the API, it
changes its behavior.

0.6.0 (2024.10.05)
* Take openArray[char] instead of string in encoder/decoder functions

0.5.6 (2024.09.06)
* Fix a broken decodeAll overload

0.5.5 (2024.09.01)
* Fix streaming EUC-JP decoder

0.5.4 (2024.08.01)
* Reduce map sizes
* Promise no exceptions in public APIs

0.5.3 (2024.07.28)
* Fix skipDirs

0.5.2 (2024.06.14)
* Misc. bugfixes and improvements

0.5.0, 0.5.1 (2024.06.13)
* Re-designed API:
	- Unified validator/validatorcore with decoder/decodercore
	- Consumers no longer need to buffer the input queue for UTF-8
	  validation
This is a breaking change.

0.4.2 (2024.02.26)
* Fixed compilation on Nim 1.6.14

0.4.1 (2024.02.23)
* Fixed horribly broken validatorcore API

0.4.0 (2024.02.21)

* Renamed project
* Re-designed interface: it's now push-based
* We now encode/decode from/to UTF-8
	- actually, encoding accepts WTF-8 too
* Added multibyte tests from https://github.com/hsivonen/encoding_rs
* Fixed lots of bugs in multi-byte encoding:
	- EUC-KR encoding was completely broken (it was using the wrong map)
	- GB18030 encoding was completely broken (it didn't handle ASCII)
	- all 16-bit maps were somewhat broken, as they had no de-duplication
* Various other correctness bugfixes
* Reduced binary size somewhat
