Re: Crash in BRIN minmax-multi indexes

Tomas Vondra <tomas.vondra@enterprisedb.com>

From: Tomas Vondra <tomas.vondra@enterprisedb.com>
To: Zhihong Yu <zyu@yugabyte.com>
Cc: Jaime Casanova <jcasanov@systemguards.com.ec>, PostgreSQL Hackers <pgsql-hackers@lists.postgresql.org>
Date: 2021-04-01T01:39:30Z
Lists: pgsql-hackers
On 4/1/21 3:22 AM, Zhihong Yu wrote:
> Hi,
> -       delta += (float8) addrb[i] - (float8) addra[i];
> -       delta /= 256;
> ...
> +       delta /= 255;
> 
> May I know why the divisor was changed ?
> 

Yeah, that's a mistake, it should remain 256. Consider two subtractions

1.1.2.255 - 1.1.1.0 = [0, 0, 1, 255]

1.1.2.255 - 1.1.0.255 = [0, 0, 2, 0]

With the divisor being 255 those would be the same (2 * 256), but we
want the first one to be a bit smaller. It's also consistent with how
inet does subtractions:

test=# select '1.1.2.255'::inet - '1.1.0.255'::inet;
 ?column?
----------
      512
(1 row)

test=# select '1.1.2.255'::inet - '1.1.1.0'::inet;
 ?column?
----------
      511
(1 row)

So I'll keep the 256.


regards

-- 
Tomas Vondra
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company



Commits

  1. Fix assert in BRIN build_distances

  2. Fix bug in brin_minmax_multi_union

  3. Fix order of parameters in BRIN minmax-multi calls

  4. Fix BRIN minmax-multi distance for inet type

  5. Fix BRIN minmax-multi distance for timetz type

  6. Fix BRIN minmax-multi distance for interval type