Franco, I agree the case is a problem, but I don't think it can be fixed at the spec level. The spec indicates that the topic URL must be exactly the self url... In all honesty, I think Feedburner is being an jerk for not handling these things correctly.
How can the hub check that it *will* get a ping? What we do at Superfeedr (and what I think the Google Hub should probably do) is that we check when somebody subscribes that the topic url matches the self link. If it does not, then we refuse the subscription. If it does, it does not mean that the publisher will ping us, but I don't think there is a way to check that. The protocol cannot say what should happen when any involved party does not respect it. It can only be 'positive' and specify what happens when certain conditions are matched. To take an extreme example, the protocol cannot state what should happen when you send a PUT request to a hub, because that's outside the spec. Using that rule, the protocol cannot specify what the hub should do when the subscriber does not behave by the protocol. That being said, I'm convinced that a 'sensible' implementation should be as helpful as possible by providing context on obvious errors: http://en.wikipedia.org/wiki/Robustness_principle I hope this makes sense. -- *Got a blog? Make following it simple: https://www.subtome.com/ <https://www.subtome.com/> * Julien Genestoux, http://twitter.com/julien51 +1 (415) 830 6574 +33 (0)9 70 44 76 29 On Fri, Nov 22, 2013 at 10:25 PM, Franco <[email protected]> wrote: > Julien, > > Thanks, that did the trick! > > Would it make sense for the pshb spec to specify that hub's treat topic > url's as case-insensitive? Actually just looked at the URI spec (RFC 3986) > and it seems that resources after the domain are technically > case-insensitive but really left up to the server to decide. So technically > those two url's (TechCrunch vs Techcrunch) could serve up different > resources, though I wonder how common that actually is in practice, > particularly for a feed xml. > > Alternatively, would it make sense for the hub to check if there is > actually a publisher registered send pings for that topic url and respond > with failure to a client subscription attempt for such a topic url that's > not registered? It looks like currently the hub successfully verifies a > subscription to any random url http://www.google.com/blah/blah/blah for > example. Is there a reason this is allowed by the protocol currently? > > Thanks! > Franco > > > On Thursday, November 21, 2013 2:09:17 AM UTC-8, Franco wrote: >> >> I am not receiving notifications for http://feeds.feedburner. >> com/TechCrunch. My client has successfully subscribed and I verified my >> callback address is subscribed for this topic at https://pubsubhubbub. >> appspot.com/subscribe >> >> I am successfully receiving notifications for several other subscriptions >> including feedburner feeds. >> > -- > > --- > You received this message because you are subscribed to the Google Groups > "Pubsubhubbub" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to [email protected]. > For more options, visit https://groups.google.com/groups/opt_out. > -- --- You received this message because you are subscribed to the Google Groups "Pubsubhubbub" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
