Skip site navigation (1)Skip section navigation (2)
Date:      Fri, 22 Aug 2014 03:21:57 +1000
From:      Kubilay Kocak <koobs@FreeBSD.org>
To:        Bryan Drewery <bdrewery@FreeBSD.org>, Matthias Andree <mandree@FreeBSD.org>, ports-committers@freebsd.org, svn-ports-all@freebsd.org, svn-ports-head@freebsd.org
Subject:   Re: svn commit: r365566 - head/Tools/scripts
Message-ID:  <53F62AB5.6060802@FreeBSD.org>
In-Reply-To: <53F62802.3030006@FreeBSD.org>
References:  <201408211556.s7LFuE0p041046@svn.freebsd.org> <53F61E42.4050104@FreeBSD.org> <53F6267E.9020909@FreeBSD.org> <53F62802.3030006@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 22/08/2014 3:10 AM, Bryan Drewery wrote:
> On 8/21/2014 12:03 PM, Matthias Andree wrote:
>> Am 21.08.2014 um 18:28 schrieb Bryan Drewery:
>>> On 8/21/2014 10:56 AM, Matthias Andree wrote:
>>>> Author: mandree
>>>> Date: Thu Aug 21 15:56:14 2014
>>>> New Revision: 365566
>>>> URL: http://svnweb.freebsd.org/changeset/ports/365566
>>>> QAT: https://qat.redports.org/buildarchive/r365566/
>>>>
>>>> Log:
>>>>   Add a BerkeleyDB upgrade helper script in preparation of 4...4.7 removal.
>>>
>>> Thanks for making things simpler.
>>>
>>> <joke> We now have a script to run another script that was made to make
>>> using ports simpler.
>>
>> The point is to abstract away the chaotic (miniscule input difference,
>> major difference in result) differences along the various dimensions of
>> port upgrading tools and local database management tools.
>>
>> I don't think a textual description for UPDATING would have been much
>> shorter and the whole migration will look scary enough to users with the
>> around-ten-step upgrade...
>>
>> And people WILL have multiple db4* versions on their typical system
>> unless they manipulated /etc/make.conf.
>>
> 
> No no, I agree with it. It's nice for the users. We have other such
> scripts too such as the perl-after-upgrade was IIRC.
> 
> Actually I find it appalling we need to use a tool to use ports at all
> and wish it was all self-contained with 1 official interface.
> 

We probably could have used one of these in python@ for cleaning up
setuptools (old) and distribute remnants that caused many issues for
users as we chased upstream, even with good instructions in UPDATING
(since we couldnt cover all cases)

I wonder too, to what extent a relatively generic interface for these
things might be possible.

--
koobs



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53F62AB5.6060802>