Re: Bugs in copyfuncs/equalfuncs support for JSON node types

In reviewing Peter's patch to auto-generate the backend/nodes
support files, I compared what the patch's script produces to
what is in the code now.  I found several discrepancies in the
recently-added parse node types for JSON functions, and as far
as I can see every one of those discrepancies is an error in
the existing code.  Some of them are relatively harmless
(e.g. COPY_LOCATION_FIELD isn't really different from
COPY_SCALAR_FIELD), but some of them definitely are live bugs.
I propose the attached patch.

                        regards, tom lane