Re: removing useless dependencies

On 9 July 2015 at 20:08, Joerg Sonnenberger <joerg%britannica.bec.de@localhost> wrote:

On Thu, Jul 09, 2015 at 05:26:50AM +0000, David Holland wrote:
> On Tue, Jul 07, 2015 at 07:56:51PM +0200, Joerg Sonnenberger wrote:
> > > > They create a lot of noise that hides real issues for the question of
> > > > pkgdir not matching best match. As such, they are harmful.
> > >
> > > That's a pbulk bug :-)
> >
> > Which part? Reporting potential problems? Implementing the same logic
> > pkg_add would trigger? Things are done for a reason...
>
> The fact that perfectly good alternative patterns cause it to issue
> warnings.

Have you read what I said? The patterns don't work most people believe
them to work. As such, they are hardly perfectly good...

I have read what you said, but I'm still not totally clear :)

I think what you are saying is that the problem is not so much with the implementation of alternatives, but is with the potential and sometimes actual inconsistency of behaviour between installations, some of which may be using one alternative and some using the other. So, regardless of whether alternatives do a useful job [which I think they do at least in theory], the cost of supporting them is not worth the benefit?

I can't say I've seen any fallout from inconsistent behaviour, but I take it you have. I suspect there's little practical concern about removing perl alternatives - realistically, rebuilding p5-* modules is way down the cheap end of the scale of inconveniences.