Distro-provided mechanism to clean up old kernels

[Barry Warsaw]

On Feb 16, 2012, at 03:16 PM, Scott Kitterman wrote:

>On Thursday, February 16, 2012 12:07:58 PM Steve Langasek wrote:
>> On Thu, Feb 16, 2012 at 12:04:03PM -0800, Steve Langasek wrote:
>> > Bear in mind that dbus is not running on servers by default.  So that
>> > would be a fine solution for the desktop, but there's a larger
>> > architectural decision to confront there if we think that should be the
>> > solution on servers as well.
>> 
>> Evan Broder has set me straight here that we *are* installing dbus by
>> default, and I see that it's part of the 'standard' task in precise.  So
>> nevermind. :)
>
>Right, but that doesn't make it a good thing.
>
>DBus is a desktop technology and I don't think there's a compelling use case 
>for adding this additional complexity to servers.  
>
>Myself, I'd prefer it go away.  I'm not clear why it was added.

I don't have a particularly strong opinion about whether dbus should or should
not be installed on servers, but given that c-j was originally intended (and
maybe still is <wink>) as a desktop application, I think it makes sense that
it's implemented as a dbus client/server.  It certainly improved some
usability problems in the past.

Having said that, the dbus service is a fairly thin wrapper around the backend
library, although that latter isn't really well suited to be used as a library
atm.  I think it could be with a bit of work though.

The real question as to the future of c-j is whether it's even the right
approach to cleaning up your system.  If so, then maybe a bit of engineering
to clean it up, better separate the backend from the dbus, ui, and cli
interfaces, and package it in a better way would be worth it.

Cheers,
-Barry