Christian Clason

Relevant ideas from Helix here: https://github.com/helix-editor/helix/discussions/11157; that's a bit orthogonal to how quantified nodes should be handled.

Can we -- as a first step -- _just_ add handling of quantified nodes and leave any changes to `@injection.combined` to a followup after the upstream discussion is concluded?

I think as you propose them to be handled. (Basically, quantified injections should be the "better injection.combined", as you say.) We don't need to handle _all_ situations out of the...

I think there's two different issues here: 1. **global** scope combination: _all_ injected nodes for a language get treated as _one single_ document -- this is what `@injection.combined` does. 2....

Just to be on clear side, here's what I suggest for _this_ PR: Assume you have ``` other NODETYPE other other NODETYPE other something other entirely other NODETYPE other other...

Sorry for the back and forth, but I am fairly confident that this proposal (which I believe matches yours from https://github.com/neovim/neovim/pull/29601#issuecomment-2212737235) is what matches upstream intention and will address most...

> which incidentally parses it as the same document Not incidentally; that's how it's intended to work ;)

Ah, ok, that was a misunderstanding then. CLI is correct and we are not. We need to convert the range to a region (table of ranges) then.

Yeah, I think that was an early assumption that was never revisited. I agree that this requires a bigger rewrite, unfortunately. How does the CLI (or, better, Helix) handle this?...

@leonerd