Erez Shinan

Hello @swamidass and @MegaIng , I wonder if maybe the best solution is to use [mmap](https://docs.python.org/3/library/mmap.html). I imagine it's the most efficient option, and we already support `bytes` as input,...

But if unicode is a must, I don't see a better approach than what MegaIng presented, which is to constantly grow your text buffer as the lexer advances. The obvious...

Yes, we'll be happy to accept a PR that extends the existing interface to support Earley. I can't promise how soon I'll get to it, since I have other obligations,...

I see, that makes sense. I forgot it happens internally, since that function is called for LALR anyway too. Perhaps it should be refactored out. Another thing I recently noticed...

It does build the parse-tree incrementally. You can find it at `ip.parser_state.value_stack`, where `ip: InteractiveParser`. But it might be a bit tricky to use it effectively.

Like you said, tokens have a type and a value. The value is the actual bit of text being parsed, and the type is the category of that text. The...

@davidmcnabnz Yes, the Earley parser is harder to debug. I wouldn't say impossible. The "problem" with Earley (which is also why it's so powerful) is that it doesn't know what...

Congrats on the milestone! Sounds interesting, what would the API look like, more or less?

@davidmcnabnz Hard to say without seeing it first! Maybe it's safer to put it in a separate repo, and we can always copy it to the main one afterwards.

@davidmcnabnz I don't think closures are the worst thing, but I agree it was nicer if there was a more relaxed way to do it. But at the same time,...