[packagekit] Things that need fixing

Robin Norwood rnorwood at redhat.com
Fri Feb 29 06:59:41 PST 2008


On Fri, 29 Feb 2008 09:33:38 +0100
Tim Lauridsen <tim.lauridsen at googlemail.com> wrote:

> Richard Hughes wrote:
> > On Fri, 2008-02-29 at 08:59 +0100, Tim Lauridsen wrote:
> >> Yum >= 3.2,10 has a new improved '--skip-broken' broken feature
> >> builtin to skip packages with broken dependencies.
> >> it is controlled by the self.yumbase.conf.skip_broken flag.
> > 
> > YAY!!! In which case can we use this by default for UpdateSystem or
> > is much slower? If it's much slower, we could try the transaction
> > without this set, then try it with set, else erroring out.
> 
> I is not slower if there is no problem with transaction, only if
> there is problem it will be slower.

Let's just make it the default, then, at least for the update system
call.  For installs, or more targetted updates, having things atomic
makes more sense to me.

> >> I would be nice if there was some way to control this from pk, but
> >> we have to extend the signals some how to signal packages being
> >> skipped from the transaction.
> > 
> > Just don't emit the Package() for the skipped package. We'll always
> > get it again when we requeue the GetUpdates.
> > 
>   I think the users will find it strange, if they update the system,
> and there is some updates left after the update, without some notice.
> is there a signal to send a post transaction message ?

I agree, there should still be some sort of error message to the user.

-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