From owner-cvs-gnu  Tue May  9 18:36:01 1995
Return-Path: cvs-gnu-owner
Received: (from majordom@localhost)
          by freefall.cdrom.com (8.6.10/8.6.6) id SAA03242
          for cvs-gnu-outgoing; Tue, 9 May 1995 18:36:01 -0700
Received: from forgery.CS.Berkeley.EDU (forgery.CS.Berkeley.EDU [128.32.33.75])
          by freefall.cdrom.com (8.6.10/8.6.6) with ESMTP id SAA03236
          ; Tue, 9 May 1995 18:35:57 -0700
Received: (from asami@localhost) by forgery.CS.Berkeley.EDU (8.6.11/8.6.9) id SAA16944; Tue, 9 May 1995 18:35:54 -0700
Date: Tue, 9 May 1995 18:35:54 -0700
Message-Id: <199505100135.SAA16944@forgery.CS.Berkeley.EDU>
To: wollman@halloran-eldar.lcs.mit.edu
CC: paul@isl.cf.ac.uk, bde@zeta.org.au, ache@freefall.cdrom.com,
        CVS-commiters@freefall.cdrom.com, cvs-gnu@freefall.cdrom.com
In-reply-to: <9505091517.AA09817@halloran-eldar.lcs.mit.edu> (message from Garrett Wollman on Tue, 9 May 1995 11:17:29 -0400)
Subject: Re: cvs commit: src/gnu/lib/libreadline Makefile
From: asami@cs.berkeley.edu (Satoshi Asami | =?ISO-2022-JP?B?GyRCQHUbKEI=?= =?ISO-2022-JP?B?GyRCOCsbKEIgGyRCOC0bKEI=?=)
Sender: cvs-gnu-owner@freebsd.org
Precedence: bulk

 * > I thought we agreed that it would get bunped *once* per release.
 * 
 * Indeed we did so agree.
 * 
 * (Well, closer to the mark, I recommended establishing such a policy, a
 * few people stated their agreement, and nobody objected at the time.)

Ok...then let me reiterate the proposal.

========<from here>========

The version number gets bumped the first time someone changes
something in a library (in a visible way from the caller) after a
release.  It stays there until the next release (except, of course
when the first change requires a minor number change and the second
requires a major number change), so people modifying the library
should check the commit logs before changing the version numbers.

Note:

The following events cause major version numbers to be bumped

(1) deleting a function
(2) interface change (different argument order, etc.)

and the following events will require the minor number to be bumped

(3) adding a function

In general, a x.y library can be substituted for x.z if y > z, and
this is the only possible substitution.

=========<to here>=========

If nobody objects, then this will become the official policy. :)

Satoshi