[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Bug#681783: Are Recommends really important (especially for metapackages)?

On Lu, 16 iul 12, 15:50:06, Ian Jackson wrote:
> Andrei POPESCU writes ("Bug#681783: Are Recommends really important (especially for metapackages)?"):
> > Package: tech-ctte
> These questions seems rather abstract to me.  I would rather deal with
> them iff they turn out to be relevant to an actual specific package or
> package(s).  Are you referring to the discussion surrounding
> gnome-core and network-manager ?
> If you would like the TC to make a ruling on that specific question
> then I'm sure we'd be happy to do so.
I am on purpose trying to avoid any specific issue, as I believe at this 
time there are too many in the archive and it would be unproductive for 
all parties involved.
> > Following the big sub-thread "Recommends for metapackages" on 
> > debian-devel[1] I would like to submit following questions to you as per 
> > Debian Constitution 6.1, item 5 ("Offer advice"). Examples below are not 
> > meant to request specific decision(s) from your side or even to point 
> > fingers at any one package Maintainer (DD or not), but to better explain 
> > what I think is a more general problem that you could address.
> I don't think it would be a good idea for the TC to formally offer
> such abstract advice, at least in this case.  The result would seem to
> be likely to be arguments over the interpretation of our advice.

If by "this case" you mean gnome-core and network-manager, then by all 
means, delay any ruling until the spirits have settled (this is why I 
mentioned after the release) :)

> But I'm happy to give my personal opinion:
> > 1. Is running a system with Recommends turned off a supported 
> > configuration?
> I would say that "yes it is supported" but that violating a Recommends
> may result in unexpected or less favourable behaviour.
> Ie, I think a maintainer is at liberty to close bug reports resulting
> from violated Recommends with "don't do that then", providing that the
> failure mode is reasonable.

I completely agree. However, maintainers may not want to expose 
themselves to a flood of bugs triggered by some missing un-essential 
functionality[1]. It might help them more if for example all Debian 
documentation was amended first to discourage disabling Recommends. 
Before starting any work on this I'd rather have a "go" from you.
[1] I have argued this myself on debian-user on the occasion of Xorg -> 
hal controversy.

Kind regards,
Offtopic discussions among Debian users and developers:

Attachment: signature.asc
Description: Digital signature

Reply to: