Skip site navigation (1)Skip section navigation (2)
Date:      Wed, 16 Jul 2014 11:39:51 +0100
From:      Vsevolod Stakhov <vsevolod@FreeBSD.org>
To:        Alexey Dokuchaev <danfe@FreeBSD.org>
Cc:        svn-ports-head <svn-ports-head@freebsd.org>, "Timur I. Bakeyev" <timur@FreeBSD.org>, Steve Wills <swills@freebsd.org>, svn-ports-all <svn-ports-all@freebsd.org>, marino@freebsd.org, "ports-committers@freebsd.org" <ports-committers@freebsd.org>
Subject:   Re: svn commit: r361646 - in head/net/samba36: . files
Message-ID:  <53C65677.8060603@FreeBSD.org>
In-Reply-To: <20140716094453.GA53961@FreeBSD.org>
References:  <201407122229.s6CMTN42057554@svn.freebsd.org> <CALdFvJGQq=PNpAqBBEZXi-q0GT=Ro-fRSjGcHFR1R01=FBtQOQ@mail.gmail.com> <53C322A7.2090705@marino.st> <20140714003112.GA54756@mouf.net> <CALdFvJEvf4-RSJNUVxX08T8K-tq9PoKge-XxmhDafAn_QxjEcg@mail.gmail.com> <53C451FA.2020304@marino.st> <20140715170501.GA73101@FreeBSD.org> <53C5618F.2020104@FreeBSD.org> <20140716094453.GA53961@FreeBSD.org>

next in thread | previous in thread | raw e-mail | index | archive | help
On 16/07/14 10:44, Alexey Dokuchaev wrote:
> On Tue, Jul 15, 2014 at 06:14:55PM +0100, Vsevolod Stakhov wrote:
>> Let me explain the situation with pkg. Pkg needs to find so called
>> ``upgrade chains'' that are used to upgrade packages. To find out
>> packages that are suitable for upgrade we use origins in pkg 1.2 and
>> name~origin in pkg 1.3.
>>
>> However, each package is identified by a special field called
>> `manifestdigest'. In pkg 1.2, this field is just sha256(manifest).
>> Unfortunately, this means that if *any* field of a package is changed a
>> version bump is required. By fields I mean files and directories as well
>> which leads thus to a policy where we need to bump a revision even if we
>> have meaningless changes in the files a package provides (that happens
>> after this particular change).
>>
>> With pkg 1.3 this behaviour has been changed to recognize the following
>> fields only:
>>
>> *  name
>> *  origin
>> *  version
>> *  arch
>> *  maintainer
>> *  www
>> *  message
>> *  comment
>> *  options
>>
>> Hence, I think that with the release 1.3 of pkg we should define
>> revision bump policy to reflect this change.
> 
> I do not find this design good enough.  I don't use pkg, I like to build
> things.  I'm happy with pkgng for keeping track of what I have installed,
> but still prefer to build stuff from ports the old way.

Why do you think that pkg is designed for binary packages merely? I'd
like to mention that my comment was not about binary packages but about
packages generated from the ports. To distinguish one package from
another pkg uses (from 1.3) the following fields. Therefore, I would
suggest to bump revision to reflect that the package turns to be
changed. This is *not* for pkg, it is for users as they want to do `pkg
upgrade` from time to time that requires versions to be increased for
different packages.

> Now perhaps my gear is not as fast as yours; it takes about two hours to
> build heavy ports like gcc47 or Boost on PowerPC 7447A (1250.35 MHz), and
> even on much faster Pentium M 780 (2.26GHz) building e.g. Clang takes an
> hour or so.  So while I like to build things from source code, I certainly
> do not like to *rebuild* them over and over again for gratuitous reasons.

I'd like to ask a single question only: do you really think that the
whole FreeBSD project should just fulfil your needs? That's nonsense: we
should take care of *users* first. And believe me, they won't be happy
if their upgrade is broken just because you use archaic hardware to
build ports.

> Tell me, why on earth shall i bump revision for a typo fix in COMMENT or
> pkg-message, www, maintainer change?  Why do I have to waste time and CPU
> cycles for rebuilding my otherwise perfectly fine packages?

Fell free to suggest your set of fields. Unless we have released pkg 1.3
this set is a subject for discussions.

> There should be a saner way to figure out those upgrade chains.  If even
> a trivial change requires revbump to allow pkg to work correctly, then it
> must be doing it wrong, sorry.

Again, I've announced the suggested set of ``significant'' fields in the
pkg@ mailing list but I got no objections. I really want to discuss it
and it *can* be changed a.t.m. if we conclude that it is inadequate from
the ports perspective, for instance.

-- 
Vsevolod Stakhov



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?53C65677.8060603>