Re: [RFC] Changing APT to pre-depend on ${shlibs:Depends}

[Ian Jackson <ijackson@chiark.greenend.org.uk>]

Julian Andres Klode writes ("Re: [RFC] Changing APT to pre-depend on ${shlibs:Depends}"):
> On Thu, 2011-05-19 at 13:28 +0100, Ian Jackson wrote:
> > The purpose is to make sure that we have considered all the up- and
> > down-sides of the proposal, and specifically to make sure that if
> > there are things that are going to go wrong we discover them sooner
> > rather than later.
>
> Basically yes, but I like to have an overview of the opinions as well
> while collecting reasons.

Right, good.

> > In the first instance the maintainers are the persons who will decide
> > whether the consensus favours the proposal.  So I suggest that people
> > who think the proposal is a bad idea concentrate on producing good
> > reasons which will persuade the maintainers.
>
> We basically know the reasons on both sides. Most objections are of
> social nature (fear of "APT has a Pre-Depends, let's add one to"), and
> Eugene thinks it is unfair if APT were to pre-depend on things while
> Cupt would not, ignoring the differences in numbers of users and
> priority (important vs optional).

I think an argument could be made for cupt too.  cupt having
comparable predependencies would only affect users (and testers) of
cupt, after all.

Ian.