As well as fix-ups to the overall logic of DEFAULT_LEXICAL_STATE in conjunction with INCLUDE.
I wrote a new blog article here: https://parsers.org/announcements/tokens-i-do-declare/
It's a funny situation because I have been aware of how glitchy DEFAULT_LEXICAL_STATE is and also the need for something like the REQUIRE_TOKEN_DECLARATION option but I thought it would be a very quick thing to deal with. But instead, I got bogged down in a lot of detail fixing this up. Also, properly documenting it was maybe more work than (or just as much work as) fixing the code. I was amazed that the above-linked blog article turned out to be something like 2000 words long!
But I encourage everybody to read it! Any questions or feedback are welcome.