[Rpm-maint] Commit fafd8090 (Dead code removal)

[Panu Matilainen]

On Tue, 13 Apr 2010, Michael Schroeder wrote:

> On Tue, Apr 13, 2010 at 09:51:12AM +0300, Panu Matilainen wrote:
>> OTOH: are there cases where self-obsoletes, or conflicts for that matter,
>> that actually match the package itself ever make any sense at all? At
>> least I can't think of any.
>
> The debian folks often use a self conflict to make sure just
> one flavor of something is installed.
>
> Example (made up):
>
>  sendmail:
>    Provides: smtp_daemon
>    Conflicts: smtp_daemon
>
>  postfix:
>    Provides: smtp_daemon
>    Conflicts: smtp_daemon
>
> That way you don't have to know in advance all of the flavors
> of a capability. With rpm you need to explicitely enumerate them:
>
>  sendmail:
>    Provides: smtp_daemon
>    Conflicts: postfix exim ...

Ok that's one potential use-case for self-conflicts, perhaps 
self-conflicts on self-provides should be permitted. OTOH there might be 
saner ways to express the "only allow one of these to be installed", maybe 
it's just me but this construct of specifying a "singleton" looks just 
weird:

    Provides: foo
    Conflicts: foo

 	- Panu -