Max Brunsfeld

This has something to do with the word starting with a `$`. I can't reproduce it otherwise.

Oh, this has the same underlying cause as https://github.com/atom/autocomplete-plus/issues/922. The subsequence provider does not use the `editor.nonWordCharacters` setting for the current file (it uses the *global* value of the setting)....

I think it may be because the wasm binding uses the UTF16 encoding, due to javascript’s string semantics. Do you still see a mismatch if you transcode to UTF16 in...

That sounds good. I can't really remember the story behind the `default` modifier. If removing it allows us to parse code successfully that we couldn't parse before, I'm 👍 .

It looks like both types of attributes may be allowed. And in fact, it seems that `attribute_item` can appear before any *element* of the array. First, I tested that the...

I can't reproduce either of these problems on the latest nightly. Is anyone still seeing this? It may have been fixed by some upgrade of either `tree-sitter` or `tree-sitter-ruby`.

Overall, I really like the way you did this, and I'd like to merge it, since I know this feature is ~stabilizing~ stable. My only worry is that with this...

Yeah, I like the `let_chain` idea. I think ideally, you'd still create a `binary_expression` node wherever possible. So for this code: ```rust if a && b let Some(c) = d...

> Feasibility aside, I don't think I agree with this parse tree. It implies that e && f is a sub-expression, but that's not actually the case since && is...

> I think the only consistent parse tree would be... I don't think we should use the first parse tree, because it makes it much harder to find all of...