New features
- First release with go module support.
- Import directly from
github.com/src-d/enry/v2 - Optional Flex-based tokenizer, same as Linguist uses.
Hidden behind-tags flex, improves content classifier accuracy.
Fixes
- Optional oniguruma-based tokenizer now is based on new Oniguruma v6.x and
produces consistent with RE2 results on all the samples from Linguist (including non-utf8).
Full list of issues tracked under v2.0.0 milestone.