Re: row_to_json bug with index only scans: empty keys!

Andrew Dunstan <andrew(at)dunslane(dot)net> writes:
> On 11/08/2014 11:24 AM, Tom Lane wrote:
>> That seems like a pretty silly move: it wouldn't actually fix anything,
>> and it would break cases that perhaps are acceptable to users today.

> What evidence do you have that it might be acceptable to today's users?
> The only evidence we have that I know of is Ross' complaint that
> indicates that it's not acceptable.

The fact that we've only gotten two bug reports in however many years the
failure has been possible might mean that few people encounter the case,
or it might mean that it doesn't break things for their usages so badly
that they need to complain. If the latter, throwing an error rather than
fixing it is not going to be an improvement.

> I assume that's what you would propose for just the stable branches, and
> that going forward we'd always use the aliases from the RTE?

That would be my druthers. But given the lack of complaints, maybe we
should just stick to the more-backwards-compatible behavior until someone
does complain. Thoughts?

regards, tom lane