inet increment with int

From: Patrick Welche <prlw1(at)newn(dot)cam(dot)ac(dot)uk>
To: pgsql-hackers(at)postgresql(dot)org
Subject: inet increment with int
Date: 2005-09-05 18:25:01
Message-ID: 20050905182501.GS8469@quartz.itdept.newn.cam.ac.uk
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

Ilya Kovalenko posted some code at in a thread starting at

http://archives.postgresql.org/pgsql-hackers/2005-04/msg00417.php

which lead to the TODO item:

* Allow INET + INT4 to increment the host part of the address, or
throw an error on overflow

I think that the naively coded function attached does what is needed, e.g.,

CREATE OR REPLACE FUNCTION inet_inc(inet, int4)
RETURNS inet
AS '/tmp/inet.so','inet_inc'
LANGUAGE C STRICT;

CREATE OPERATOR + (
leftarg = inet,
rightarg = int4,
procedure = inet_inc
);

test=# select '192.168.0.1/24'::inet + 300;
ERROR: Increment (300) too big for network (/24)
test=# select '192.168.0.1/24'::inet + 254;
?column?
------------------
192.168.0.255/24
(1 row)

test=# select '192.168.0.1/24'::inet + 255;
ERROR: Increment (255) takes address (192.168.0.1) out of its network (/24)
test=# select '192.168.0.1/24'::inet + -2;
ERROR: Increment (-2) takes address (192.168.0.1) out of its network (/24)
test=# select '255.255.255.254/0'::inet + 2;
ERROR: Increment (2) takes address (255.255.255.254) out of its network (/0)

and just for fun:

create table list (
host inet
);

insert into list values ('192.168.0.1/24');
insert into list values ('192.168.0.2/24');
insert into list values ('192.168.0.4/24');
insert into list values ('192.168.0.5/24');
insert into list values ('192.168.0.6/24');
insert into list values ('192.168.0.8/24');
insert into list values ('192.168.0.9/24');
insert into list values ('192.168.0.10/24');
insert into list values ('192.168.1.1/24');
insert into list values ('192.168.1.3/24');

select host+1 from list
where host+1 <<= '192.168.1.0/24'
and not exists
( select 1
from list
where host=host+1
and host << '192.168.1.0/24' )
limit 1;

If you agree that this is the right thing, I can code it less
naively, (Ilya rightly uses ntohl/htonl), create the operator's
commutator, provide a patch which makes it a built-in, and some
obvious documentation.

Cheers,

Patrick

Attachment Content-Type Size
inet.c text/plain 1.9 KB

Responses

Browse pgsql-hackers by date

  From Date Subject
Next Message Patrick Welche 2005-09-05 18:42:16 Re: Proof of concept COLLATE support with patch
Previous Message Peter Eisentraut 2005-09-05 15:43:16 Re: Call for 7.5 feature completion