First, atom-linter is awesome infrastructure.
However I want to run clang-tidy, which takes >10 seconds to finish and I don’t really want to queue up invocations of it so I would rather run it on demand but still display the results in atom-linter when they arrive. Any idea as to how that can be achieved? Can I activate the package for one run and then force-invoke atom-linter for example? Is there a better way?