Re: BUG #1517: SQL interval syntax is accepted by the parser,

Roy Badami <roy(at)gnomon(dot)org(dot)uk> writes:
> ie do you accept
> interval '1 day 1 hour' day to second

I think we have to, and the reason is that this isn't different under
the hood from reading the external value '1 day 1 hour' and storing
it into a column that has the DAY TO SECOND typmod. If we reject
the above we'd be breaking existing dump files. Furthermore this
would imply that dump output from a constrained interval column
would *have to* not have any decoration; ie we could only output
'1 1' and not '1 day 1 hour'. Regardless of what the spec says,
I find the former dangerously ambiguous.

I'm happy to see our code upgraded to accept the spec's syntax.
I won't be happy to see it changed to reject input that we used
to accept, especially when the only argument for doing so is a
narrow-minded insistence that we can't accept anything beyond
what the spec says.

regards, tom lane