Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 23 Mar 2009 19:17:47 -0700
From:      Doug Barton <dougb@FreeBSD.org>
To:        obrien@freebsd.org
Cc:        cvs-ports@FreeBSD.org, Pav Lucistnik <pav@FreeBSD.org>, cvs-all@FreeBSD.org, ports-committers@FreeBSD.org
Subject:   Re: cvs commit: ports/shells/bash Makefile pkg-plist
Message-ID:  <49C842CB.6070900@FreeBSD.org>
In-Reply-To: <20090323231412.GA94221@hub.freebsd.org>
References:  <200903120954.n2C9s2ev063133@repoman.freebsd.org> <20090313023956.GA49511@dragon.NUXI.org> <49BA52D2.8090209@FreeBSD.org> <20090323231412.GA94221@hub.freebsd.org>

next in thread | previous in thread | raw e-mail | index | archive | help
David O'Brien wrote:
> On Fri, Mar 13, 2009 at 01:34:26PM +0100, Pav Lucistnik wrote:
>> David O'Brien wrote:
>>> On Thu, Mar 12, 2009 at 09:54:01AM +0000, Sergey A. Osokin wrote:
>>>> osa         2009-03-12 09:54:01 UTC
>>>>  FreeBSD ports repository
>>>>  Modified files:
>>>>    shells/bash          Makefile pkg-plist 
>>>>  Log:
>>>>  Fix pkg-plist.
>>>>  Bump PORTREVISION.
>>> Thanks for fixing the PLIST (much appreciated), but why did you bump
>>> PORTREVISION for a PLIST change?
>>>
>>> There is zero reason to force a reinstall for a PLIST change.  Either the
>>> port is already installed (and the user can wait for some other reason to
>>> update), or the port isn't installed and bumping PORTREVISION does
>>> nothing.

Actually if you bump PORTREVISION without changing anything else the
users who have the thing installed will get the corrected plist. This
can matter if they deinstall the port later rather than upgrading it.
It can also matter if the upgrade changes the plist again.

>> It's needed for package cluster, otherwise it does not know to rebuild 
>> and will serve incomplete package forever.
> 
> Is there ever a change then that doesn't require a bump in either
> PORTREVISION or PORTVERSION?

Yes. If the package isn't going to change, no PORTREVISION bump should
be done. Someone else already provided examples, but they are numerous.

> If not, maybe we should do away with PORTREVISION and use something like:
> 
> ${PORTNAME}-${PORTVERSION}_${VCS_ID}

I actually have in mind a different scheme that replaces both
PORTREVISION and PORTEPOCH with a date string like 200903231 that
would be appended to each PKGNAME. It would completely remove the
ambiguity (and the kludgy mess that both of the current variables have
created), and has the extra added bonus that you could set it
differently depending on which options the user has set without fear
of creating ambiguity. But I digress ...


Doug

-- 

    This .signature sanitized for your protection




Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?49C842CB.6070900>