[packagekit] Ignoring updates
Robin Norwood
rnorwood at redhat.com
Tue Apr 15 08:06:25 PDT 2008
On Tue, 15 Apr 2008 09:24:34 -0400
Jesse Keating <jkeating at redhat.com> wrote:
> On Tue, 2008-04-15 at 09:17 -0400, Robin Norwood wrote:
> > Implementing this gets tricky really fast. If I ignore hal
> > updates, but something pulls in a hal update as a dep...?
> >
> > This will be hard for the backends to get right, unless they
> > already support blacklists. Yum has the exclude option, though,
> > which should make it fairly easy for the yum backend. We'd need to
> > decide if we want this to apply to the global yum configuration
> > (editing yum.conf) or just apply it for PK transactions. We
> > already edit the repo config globally, so for consistency's sake we
> > should probably do the same here, but not trample on the existing
> > blacklist config when doing so.
>
> Honestly I think any such ignorings should be done on a
> per-transaction basis. Trying to guess what a user meant when they
> ignored a certain update is rather doomed to fail. Even at the yum
> level does the user mean "exclude it now, and for ever", "exclude
> just this one, but future ones are OK", "exclude it just from this
> repo, but from that other repo it's ok", permutations of the above?
>
> Is it unfair to have PK frontends do per-transaction exclusions, and
> anything more permanent should be written to the backend's native
> configs by the user/admin?
That's really the same as unchecking the update from the list, which we
(soon) support.
If we want to implement this on a more permanent basis, we really need
to decide if PK works with the backends existing configuration, or if
it always does it's own thing - as a design decision for PK as a
whole. I lean towards using the backend's existing configuration,
including changing it when you change something in PK.
-RN
--
Robin Norwood
Red Hat, Inc.
"The Sage does nothing, yet nothing remains undone."
-Lao Tzu, Te Tao Ching
More information about the PackageKit
mailing list