Peter Law

> I don't know if anyone relies on them NOT being included, but since it's a change, it should perhaps get a major version bump. Not quite rely, but I'm...

It looks like the answer at https://stackoverflow.com/a/42217872/67873 (note: _not_ the accepted answer) has a mechanism for this.

The challenge I see here is what the Python API would look like. On the one hand, I'd really quite like it to be something a little bit like Django's...

> It's worth keeping in mind that README.md isn't installed - so changes there are going to help people landing here on github, but not help people once it's installed....

@bbmokhtari just wondering if there's anything I can help with to get this merged?

Hi @vincentnavetat, thanks for the contribution! Just wanted to let you know that we've seen this PR, though I can't promise we'll get to reviewing it soon unfortunately :(. I'll...

> How does this solve the last problem mentioned in [#94 (comment)](https://github.com/gristlabs/asttokens/issues/94#issuecomment-1252844046)? I need to be able to avoid using tokens when possible, but for `ASTTokens` to automatically start using...

Having _some_ of the functionality of the class have lazy-evaluation while others error about the missing data feels like a recipe for a very confusing API and a lot of...

> Using a token-specific method after specifying `init_tokens=False` is generally a mistake that should be complained about. I'm not really sure I follow the logic here given that `get_text` depends...

I appreciate that `ipython` and others are important consumers, I'm not at all saying that the lazy evaluation approach shouldn't happen. Faster `asttokens` is great for everyone, `ipython` included :)...