With `Layout`s in Compose, is there anything intrinsically wrong with a measure policy simply neglecting to measure and place one of the items of content it is passed? (Besides the fact that all of the passed measurables will have been composed already only for layout and drawing to be skipped on one). The docs don’t really say anything either way, it’s a neat solution (in this case using the multi-content overload of
Layout
to select between one of two different input composables. I assume the only caveat is the slight performance cost of composing something that won’t get used. I also assume that the semantics tree is only constructed after layout since it relies on the relative position of nodes.