On Mon, 29 Mar 2010 14:38:35 +0200
Markus Armbruster <arm...@redhat.com> wrote:
> Luiz Capitulino <lcapitul...@redhat.com> writes:
>
> > On Thu, 25 Mar 2010 20:30:33 +0100
> > Markus Armbruster <arm...@redhat.com> wrote:
> >
> >> Luiz Capitulino <lcapitul...@redhat.com> writes:
> >>
> >> > On Thu, 25 Mar 2010 18:37:25 +0100
> >> > Markus Armbruster <arm...@redhat.com> wrote:
> >> >
> >> > [...]
> >> >
> >> >> >> @@ -86,12 +86,13 @@ int do_migrate(Monitor *mon, const QDict *qdict,
> >> >> >> QObject **ret_data)
> >> >> >> (int)qdict_get_int(qdict,
> >> >> >> "inc"));
> >> >> >> #endif
> >> >> >> } else {
> >> >> >> - monitor_printf(mon, "unknown migration protocol: %s\n",
> >> >> >> uri);
> >> >> >> + qerror_report(QERR_INVALID_PARAMETER, "uri");
> >> >> >> return -1;
> >> >> >> }
> >> >> >>
> >> >> >> if (s == NULL) {
> >> >> >> - monitor_printf(mon, "migration failed\n");
> >> >> >> + /* TODO push error reporting into the
> >> >> >> FOO_start_outgoing_migration() */
> >> >> >> + qerror_report(QERR_MIGRATION_FAILED);
> >> >> >> return -1;
> >> >> >> }
> >> >> >
> >> >> > I think this one is no better than the automatic UndefinedError
> >> >> > which is going to be triggered. I would only touch this when/if
> >> >> > we get the migration functions converted.
> >> >>
> >> >> I feel it is a bit better, because:
> >> >>
> >> >> * It doesn't dilute the nice "this is a bug, and I should report it"
> >> >> property of UndefinedError.
> >> >>
> >> >> * It avoids the "returned failure but did not pass an error" message.
> >> >> Minor, because it's under CONFIG_DEBUG_MONITOR.
> >> >
> >> > But this is exactly what we want because having QERR_MIGRATION_FAILED
> >> > there doesn't fix the problems those warnings are making us aware of.
> >>
> >> Except we are already aware of the problem, and additional warnings are
> >> just noise.
> >
> > Our memory is not the best place for it, not only because we can forget,
> > but also because it's limited on you and me.
> >
> > Anyone who enables QMP's debugging support should be able to know what's
> > there to be done.
>
> That's what TODO comments are for.
It's case by case and in this very case I'd only add QERR_MIGRATION_FAILED
if it makes difference for clients, because besides silencing the warning
w/o the proper fix, this error is probably going to be dropped when the
*_outgoing_migration() functions are converted.
