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

Re: stupid dependencies on update-inetd



On Sun, Jul 29, 2007 at 03:59:13AM +0200, Marco d'Itri wrote:
> On Jul 29, Russ Allbery <rra@debian.org> wrote:
> > So is anything ever valid other than openbsd-inetd | inet-superserver as a
> > dependency?  I keep getting confused on the rules around using virtual
> > packages.  Would rlinetd | inet-superserver be okay?  Would
> Formally yes, but I do not think there is a reason for a package to
> choose a different default.
> > inet-superserver all by itself be okay?
> No, but this would be taken care of by
> virtual-package-depends-without-real-package-depends (which explains
> the part you are missing).

Isn't openbsd-inetd priority:standard? That's enough to make the
real-package unnecessary, afaik (and that lets the default inetd be
changed simply by changing the priorities of the packages, rather than
the dependencies of lots of packages).

I would've thought the Build-Dependency stuff isn't relevant for inetds.

I think policy has had confusing advice on this for a long time now --
ttbomk the only cases where an explicit dependency on a real package is
either useful or needed is when all the alternatives can be installed
concurrently and are at priority:optional so that dselect and apt can
select one easily when there's no hints to be taken from the priority
levels; and when consistency is useful for the buildds, for much the
same reason.

Cheers,
aj

Attachment: signature.asc
Description: Digital signature


Reply to: