Hi Mark,
On Fri, Feb 12, 2021, at 20:56, Mark Rofail wrote:
>Indeed you are right, to support the correct behaviour we have to use @>>(anycompatiblearray, anycompatiblenonarry) and >this throws a sanity error in opr_santiy since the left operand doesn't equal the gin opclass which is anyarray. I am thinking >to solve this we need to add a new opclass under gin "compatible_array_ops" beside the already existing "array_ops", >what do you think?
I'm afraid I have no idea. I don't really understand how these "anycompatible"-types work, I only knew of "anyarray" and "anyelement" until recently. I will study these in detail to get a better understanding. But perhaps you could just explain a quick question first:
Why couldn't/shouldn't @>> and <<@ be operating on anyarray and anyelement?
This would seem more natural to me since the Array Operators versions of @> and <@ operate on anyarray.
/Joel
