Sebastian Pfitzner

This is a tokenizer issue. I don't think we'll ever end up fixing that in this package.

The point is that symbol resolution can only be accurate if you're pointing VS Code to the right environment (you're of course free to ignore the prompt).

Ok, so you want that button to automatically select the appropriate environment instead of selecting it yourself. Fair enough, I guess.

Sure. Take a look at the [Documenter.jl docs](https://juliadocs.github.io/Documenter.jl/stable/) and open a PR.

Our best bet here would be to use the [semantic tokens spec](https://microsoft.github.io/language-server-protocol/specifications/lsp/3.17/specification/#textDocument_semanticTokens). That allows augmenting *or* replacing the regex based grammar with additional LS provided info. IIRC we don't have...

Right, so should we rename `Tk.Text`, unexport it from `Tk` or unexport `Text` from Base?

> Personally, I do not agree that adding a new export is breaking. Instead, I personally believe that nobody should do `using Foo`. I agree, but there's plenty of code...

> bad interaction with some task magic that VS Code does Yeah, that. There's some dark and scary magic in the extension. We're basically scheduling an `InterruptException` on our user...

Yeah... this basically was a shortcut I took because I was too annoyed with figuring out corrected regexps for this. I don't think I'll look into fixing this and instead...

We had these multiline macro strings defined earlier, but there are systematic issues with that: the injected grammar is not guaranteed to "terminate" (e.g. for unclosed code blocks in markdown,...