netmask(inet) function broken in 7.4 - Mailing list pgsql-bugs

From Joe Sunday
Subject netmask(inet) function broken in 7.4
Date
Msg-id 20031130224326.GC18154@csh.rit.edu
Whole thread Raw
Responses Re: netmask(inet) function broken in 7.4
List pgsql-bugs
============================================================================
                        POSTGRESQL BUG REPORT TEMPLATE
============================================================================


Your name        :    Joe Sunday
Your email address    :    sunday@csh.rit.edu


System Configuration
---------------------
  Architecture (example: Intel Pentium)      : UltraSparc II

  Operating System (example: Linux 2.0.26 ELF)     : Solaris 9

  PostgreSQL version (example: PostgreSQL-7.4):   PostgreSQL-7.4

  Compiler used (example:  gcc 2.95.2)        : gcc 2.95.3


Please enter a FULL description of your problem:
------------------------------------------------
After upgrading to 7.4 from 7.3.4, the behaviour of the netmask(inet) function changed,
but the documentation still specifies the old behaviour (which seems more correct than the
new version)

Old:
netmask( '192.168.1.0/23' ) returns '255.255.254.0'

New:
netmask( '192.168.1.0/23' ) returns '255.255.254.0/23'

Please describe a way to repeat the problem.   Please try to provide a
concise reproducible example, if at all possible:
----------------------------------------------------------------------

Perform a SELECT using the netmask function with the example above.

If you know how this problem might be fixed, list the solution below:
---------------------------------------------------------------------
Revert netmask to the prior behaviour, or update the documentation to match the new
behaviour. Masklen(inet) already returns just the mask length for CIDR-style addressing,
netmask(inet) should just return a mask without the redundant length.

--
Joe Sunday <sunday@csh.rit.edu>  http://www.csh.rit.edu/~sunday/
Computer Science House, Rochester Inst. Of Technology

pgsql-bugs by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: unix_socket_directory vs pg_ctl?
Next
From: Martin Edlman
Date:
Subject: Re: Repeatedly breaking indexes - SOLVED