I didn't check anything in for that. And the problem is from sampling in continuation indents. It has heuristics that work at getting the correct indent out of the sampled indents, but it's breaking down here. I had hoped I would be able to work around it for the time being, and address it fully in the point release, but it got a little complicated.
For languages with statement parsing, we have access to the information we need to filter-out/not-visit the continuation indents, and my plan is to integrate that into adaptive formatting for the point release.