Dmitry

Doh, you are right. Sorry to comment based only on recollections which were incorrect :) And without even looking carefully at which issue the thread is about. So scratch my...

Interesting. Yes, it does seem useful to include surrounding parens for generator expressions, to match Python 38 behavior. But it's a behavior change for older versions.

For attribute lookup, the `Attribute` node's `last_token` will always be the name of the attribute itself. It's a single token. So I think that's a simple case, and not sure...

So you mean a helper like `get_span(attrNode, “attr”)` or `get_span(tupleNode, “elts”)`? That sounds good to me, and seems helpful to include into the asttokens library (rather than a separate one...

This sounds fine to me. I did a little thinking to clarify the goals for this. Such rules (e.g. for getting parts of Attribute or List nodes) only make sense...

I assigned to you. You make a good point about `iter_fields` (in the generic example) -- it makes the string constants less magic. Thanks for the analysis of issues and...

You are right, the built-in `ast` module doesn't seem to annotate the "alias" node with any useful information, so it doesn't have any location info. One reasonable thing we could...

It would be nice setting a list of allowed origins on a per-document basis. It would allow things like adding a record to Grist (including hosted Grist) directly from a...

I think it does make sense to allow for separate configuration of what Reference cells show vs. what autocomplete options are offered in the dropdown. Here are some related requests:...

I am glad to see this request, and I am extra glad for your offer to work on it. Definitely worth it. Let me share a couple of pointers to...