Horton

Re: [PATCH] seg: preserve the upper boundary's certainty indicator in seg_out()

Andrew Dunstan <andrew@dunslane.net>

View thread View on postgr.es
From: Andrew Dunstan <andrew@dunslane.net>
To: Heikki Linnakangas <hlinnaka@iki.fi>, Ewan Young <kdbase.hack@gmail.com>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2026-06-12T18:58:52Z
Lists: pgsql-hackers 
On 2026-06-11 Th 5:45 AM, Heikki Linnakangas wrote:
> On 11/06/2026 10:03, Ewan Young wrote:
>> Certainty indicators are documented to be preserved on output (they are
>> ignored by the operators, but kept as a comment), so this breaks the
>> input/output round-trip for the affected values.
>
> As a side note, while ignoring the boundaries makes sense for 
> comparison operators, seg_union() and seg_intersect() need to do with 
> them. That's not documented anywhere, and their current behavior seems 
> pretty arbitrary. We haven't actually documented those functions at 
> all, I think they were added just for the GiST support and calling 
> them directly from SQL was an afterthought.
>
>> The attached patch fixes the one-character typo and adds regression
>> tests that place each indicator on both boundaries of an interval, so
>> the upper-boundary case is now covered.  make installcheck passes with
>> the fix and fails without it.
>
> Applied to master and all stable branches, thanks!


This is upsetting cross version upgrade tests, I assume since we didn't 
backport it to branches older than 14.

Not sure what the best solution is. Drop the table, or at least the 
offending row?


cheers


andrew


--
Andrew Dunstan
EDB: https://www.enterprisedb.com

Commits

  1. Adjust cross-version upgrade tests for seg_out() fix

  2. seg: Fix seg_out() to preserve the upper boundary's certainty indicator