This fixes an unsoundness around handling of UTF-8 identifiers in the parser at the expense of decreasing the performance by around 20-30%.
I may switch back to a &[u8] based stream once I figured out how to handle UTF-8 identifiers with it safely without adding a truck load of complexity, but for now correctness is more important than performance.
Fixes https://github.com/martinohmann/hcl-rs/issues/350
This fixes an unsoundness around handling of UTF-8 identifiers in the parser at the expense of decreasing the performance by around 20-30%.
I may switch back to a
&[u8]based stream once I figured out how to handle UTF-8 identifiers with it safely without adding a truck load of complexity, but for now correctness is more important than performance.