Christian Clason

That's fine; I just wanted to link the context. The main points are that a) we need "message metadata" to allow different handling of messages (length is a good stand-in...

The test failures are due to the fact that you're generating the parser with ABI 15, which the bindings are not compatible with. Please use `tree-sitter g --abi=14` for this...

Ugh, CI forces generation... Can you add `generate: false` after this line? https://github.com/tree-sitter/tree-sitter-javascript/blob/6fbef40512dcd9f0a61ce03a4c9ae7597b36ab5c/.github/workflows/ci.yml#L43 (This should be a separate `ci(test): don't validate generation` commit _before_ your -- squashed! -- other commit.)

please use conventional commits, and squash and reorder your commits before this is ready to merge

Yes, hence my asking you to disable the test (which you did, which is why CI passes now). But please check that the queries are still compatbile!

As long as the old queries are still valid, that's ok. We've had PRs that change the grammar in a way that they break without thinking to adapt the reference...

https://github.com/tree-sitter/tree-sitter-javascript/tree/master/queries

I do not have write access.

Yes, but this makes it actually usable -- if even a single plugin didn't want to use it because it was not available on all platforms, this will be helpful....

We ignore `581` since it's not true when tables are involved. (And, yes, `561` is a weird one. I don't think we even enable the ~chromatic~ cyclomatic check.) Assuming that...