Jeff Davis <pgsql(at)j-davis(dot)com> writes:
> So, should we proceed assuming an index AM API change, or try to avoid
> it? If we should change the AM API, is Dean's API change acceptable?
There is no reason at all to avoid an index AM API change if one is
useful. If you look at the history, we tend to change that API every
release or two anyway.
Whether this particular design is good is a different question, and
I don't have time right now to think about that. But please don't
bend the system out of shape just to avoid an API change.
regards, tom lane