I would like to just use Raku rules for a concise way to describe more advanced grammars; I’d then just keep my regexes to the PCRE subset that’s common between Google’s RE2 and the Rust regex crate. I doubt they’re both “regular” but both guarantee linear time matching. Part of the reason I don’t do this is portability. Not everything runs Raku, but almost every platform has a regex engine with the features I need.
