Better handle stubbing of trial types in ce

The following discussion from !204747 (merged) should be addressed:

• @dstull started a discussion: (+1 comment)

  thought (non-blocking):

  I'm wondering how we should longer term deal with situations like this where:

  1. FOSS breaks because the helper is only defined in ee
  2. We only technically need this stub in ee, but due to code execution with full gitlab ce + ee, we need a stub here.

  I wonder if we do something like:

  1. in ce create a helper with needed empty methods
  2. in ee the helper will have actual methods.

  I think the above matches our non spec code override patterns the best.

  Any thoughts?