Hi goffi, Thanks for brining this up.
On Dienstag, 27. Juli 2021 17:57:29 CEST goffi wrote: > it is said that: > > If a service persists published items, a node owner may want to purge > > the node of all published items (thus removing all items from the > > persistent store, **with the exception of the last published item, > > which MAY be cached**). > > (emphasis is from me). > > That means, apparently, that a service may or may not keep the last > item. This uncertainty make the behavior unpredictable thus the caching > impossible. I have provided a workaround in the currently discussed > protoXEP "Pubsub Caching Hints" (cf. > https://xmpp.org/extensions/inbox/pubsub-caching-hints.html#purge), but > it would be better to get totally rid of this formulation, and make sure > that a pubsub node purge remove **all** items. > > I've discussed about that on xsf@ muc, and at least on author of Pubsub > XEP (Ralph Meijer) tends to agree. This is indeed strange text. The notion of the present council members in the last meeting was broadly "this is weird, we should get rid of it", but the question came up what such services might be which misbehave. @All server developers: Please check your code and see how you implemented the purge. Does it keep any item? The current idea is to put a note in the text like this into the document to help clients with servers which actually do the MAY thing: <p class="box">Note: Some service may preserve the last item despite the purge request. In such a case and if an entity needs to be certain that all items have been deleted, a query for the remaining items after a purge combined with a specific retraction request for the remaining items should be used.</p> Any opinions? kind regards, Jonas
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Standards mailing list Info: https://mail.jabber.org/mailman/listinfo/standards Unsubscribe: [email protected] _______________________________________________
