[m-rev.] for review: Fix abstract equivalence types.

[Peter Wang]

On Tue, 22 Aug 2017 09:16:13 +1000 (AEST), "Zoltan Somogyi" <zoltan.somogyi at runbox.com> wrote:
> 
> Other than that minor quibble, the diff is fine.

Thanks for the review.

Peter