Tests with #[should_panic] macro are ignored

[morgsmccauley]

Adding the #[should_panic] macro, below the #[test] macro, to an existing test causes it to be ignored. Replacing the order, i.e. #[should_panic] followed by #[test] on the next line, seems to work.

This seems to be related to the treesitter query which looks for the #[test] macro on a function. Perhaps it just picks up the first macro defined?

[rouge8]

It sounds like you have a workaround, but I'd accept a PR making the query accept #[test] followed by #[should_panic].