There are a few cases where we prefer to spend extra lines for clarity. (Examples: In modern code we strongly prefer to avoid packing too many side-effects into an expression, we avoid defining multiple variables on a single line, and we usually prefer to avoid elaborate conditional operators.)

One such case is simultaneously defining a type and an object. We prefer to do this separately, because they happen in different "domains". (Only lambdas get to be special!)

We accumulated a few occurrences throughput the codebase, which I tracked down. I consider this convention important enough (and the transformation simple enough) to be worth updating tests/std.