Thanks, Brett. So when you wrote "I've also narrowed the max expiration window to 10 days and set the default to 5 days (instead of 30 days)" that's saying that your hub will automatically delete subscriptions after 5 days (default) or 10 days (max)? If we request 0 days, does that imply the default (5 days) or the max (10)?
Or were you talking about something completely different? :-) ...doug On Wed, Sep 22, 2010 at 4:47 PM, Brett Slatkin <[email protected]> wrote: > Hey Doug, > > I totally misunderstood your first mail. My bad. > > This part is covered by the subscription refreshing part of the spec: > > http://pubsubhubbub.googlecode.com/svn/trunk/pubsubhubbub-core-0.3.html#autorefresh > > You should rely on hubs to implement this feature and always double > check with you before your subscription expires. If you're super > paranoid, refreshing your own subscription every few days seems > reasonable, but the hubs really should get this auto-refresh behavior > correct. > > -Brett > > > > On Wed, Sep 22, 2010 at 4:41 PM, Brett Slatkin <[email protected]> wrote: >> Hi Doug, >> >> On Wed, Sep 22, 2010 at 4:38 PM, Doug Kaye <[email protected]> wrote: >>> Your message points out something I've been confused about that >>> doesn't appear to be address in the protocol spec. >>> >>> What's the "right" way to subscribe to notifications long-term? We've >>> been two things that are somewhat at odds with one another: (1) >>> subscribe with expiration=0, (2) re-subscribe at every opportunity. >>> Obviously this is a case of overkill. >>> >>> For your hub, at least, what is recommended? >> >> It's not clear in the spec? >> >> http://pubsubhubbub.googlecode.com/svn/trunk/pubsubhubbub-core-0.3.html#anchor5 >> >> hub.mode=unsubscribe >> >> When it calls you back to verify, you'll see the hub.mode parameter is >> 'unsubscribe'. Return the challenge to confirm the subscription action >> along with 200. Return 404 to deny the subscription action (i.e., deny >> the unsubscription = stay subscribed). >> >> >> -Brett >> >
