Epoch 0 (was: rpms/lft/devel lft.spec,1.6,1.7)

[Dag Wieers]

On Thu, 3 Mar 2005, Michael Schwendt wrote:

> On Thu, 3 Mar 2005 13:17:44 +0100 (CET), Dag Wieers wrote:
> 
> > > > > Frankly I could not give a damn about the fact that we break Freshen - these
> > > > > things are fed through yum from a yum repository, so I really
> > > > > don't care that rpm -F might fail in some weird conditions.
> > > > > 
> > > > > Epoch: 0 is stupid, useless and moronic.
> > > > 
> > > > Where were you 2 years ago ? We needed you back then :)
> > > 
> > > Two years ago, explicit Epoch 0 made sense.
> > > 
> > > Nowadays, the supported versions of RPM no longer require it.
> > 
> > Explicit Epoch 0 never made sense, we never required it for anything, Red 
> > Hat never required it and JBJ admits he was misquoted on the subject.
> > 
> > But please do include your line of thought.
> 
> My line of thought is the following:
> 
> I was not involved in making explicit Epoch 0 a policy. I understood the
> problems with missing Epoch and RPM < 4.1.1. I didn't see the explicit
> Epoch 0 rule as a burden. A burden so high that I would complain forever
> and rather choose to boycott a community project than get productive.

It's funny how your interpretation differs from mine. The Epoch 0 problem 
is one of the several indications of how decisions were taken against the 
majority of the community and lack of arguments in favor, and those were 
enough for me to consider not being part of that community anylonger.
(Communication largely took place on IRC and bugzilla)

(And I'm only repeating this because you imply things about me that are 
incorrect or shortsighted.)

Decisions are seldom important in itself, but the consequences of 
decisions are important. So yes, I object to this and a few other 
decisions, not because of the decisions in itself, but how they came to 
exist. And what they resulted in.

I also fail to see how I have boycotted the aformentioned community 
project. Was it because I didn't join or because I voiced my opinion ? I 
hope I can still voice my opinion, even when it's not shared by you or the 
decision makers. You probably do not agree, but I consider that 
contributing too.

Ignoring things would have been much easier than arguing (believe me), 
still it affects me enough to try to prevent some of the decisions.


> There was at least one case (modplug is the one I remember) where the
> explicit Epoch 0 rule avoided upgrade problems. Since I knew we would be
> save, I didn't track any other corner-cases, which would be affected by
> the Epoch problem. The good effect of the explicit Epoch 0 rule was, we
> didn't need to care about related upgrade problems.

A missing epoch is considered to be 0 (implicit epoch 0), so an explicit 
epoch 0 is plain silly. We have iterated over this many times, there's no 
way to ratify this. Maybe there was a bug somewhere that should have been 
fixed, but we're not basing our policy on bugs ! And the aim was to be 
compatible with Red Hat all along, and they were not doing it...

--   dag wieers,  dag at wieers.com,  http://dag.wieers.com/   --
[all I want is a warm bed and a kind word and unlimited power]