- Main change: use
str::from_utf8instead ofString::from_utf8wherever possible. This will avoid unnecessary allocations. - This also allowed direct usage of
&[u8]instead ofVec<u8>in multiple places, for additional efficiency.
Breaking Change
- Removed (undocumented)
to_utf8_lossyfunction since we only plan to support valid utf8 data for ease of use. We might introduce another serializer / deserializer that can handle arbitrary bytes later on, but this should be clear that it uses another implementation altogether.
Full Changelog: v0.1.1...v0.2.0