Horton

Re: BUG #16419: wrong parsing BC year in to_date() function

Tom Lane <tgl@sss.pgh.pa.us>

View thread View on postgr.es
From: Tom Lane <tgl@sss.pgh.pa.us>
To: Bruce Momjian <bruce@momjian.us>
Cc: "David G. Johnston" <david.g.johnston@gmail.com>, Laurenz Albe <laurenz.albe@cybertec.at>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2020-09-29T18:18:58Z
Lists: pgsql-bugs, pgsql-hackers

Attachments

I wrote:
> I think this is nuts.  The current behavior is obviously broken;
> we should just treat it as a bug and fix it, including back-patching.
> I do not think there is a compatibility problem of any significance.
> Who out there is going to have an application that is relying on the
> ability to insert BC dates in this way?

Concretely, I propose the attached.  This adjusts Dar Alathar-Yemen's
patch (it didn't do the right thing IMO for the combination of bc
and year < 0) and adds test cases and docs.

Oracle would have us throw an error for year zero, but our historical
behavior has been to read it as 1 BC.  That's not so obviously wrong
that I'd want to change it in the back branches.  Maybe it could be
done as a follow-up change in HEAD.

			regards, tom lane

Commits

  1. Fix handling of BC years in to_date/to_timestamp.

  2. Fix make_timestamp[tz] to accept negative years as meaning BC.

  3. doc: PG 13 relnotes, update TOAST item to mention decompression