Robert Haas <robertmhaas@gmail.com> writes:
> Err, actually, now that I think about it, that might be a problem:
> what happens if we're trying to test two characters for equality and
> the encoding conversion fails?
This is surely all entirely doable given the encoding infrastructure
we already have. We might need some minor refactoring, eg to have
a way of not throwing an error, but it's not going to be that hard
to achieve if somebody wants to do it. So I still see little reason
for making the JSON type behave visibly differently in non-UTF8 database
encodings.
regards, tom lane