Feature caching and custom transformers

It would be nice if there was some sort of toggle option when running with feature caching as to whether a custom transformer produced a single output cache (like a collapsed bookmark) or if every transformer within it had a cache (current behaviour).

I can see the latter being useful when actual developing the custom transformer, but in most cases I don't care about the interim portions of the custom transformer, when developing a workspace that uses one.

I could put a bookmark around each custom transformer and collapse it, but that becomes tedious when there are several (dozens) of custom transformers.