Hi Edhelas, Le mardi 9 juin 2026, 14:15:30 heure d’été d’Europe centrale Timothée Jaussoin a écrit : > [SNIP] > I'm wondering if something shorter like this is possible ? (where the > client is resolving the "space" parameter as the Pubsub node) > > Proposal 2: xmpp:spaces.server.tld?;space=x56ae32
It's a problem that we already had, and so far I think that the consensus was the query the node to check its type. However I would be happy too to have something in the URI which would avoid a round-trip. I would avoid something tied to space though, and I agree with Kev that we would need something more generic. I see two ways: - `type=urn:xmpp:spaces:0`. The URI would be: `xmpp:spaces.server.tld?;node=123;type=urn:xmpp:spaces:0` That would work and be generic, probably the cleanest option in my opinion, but I understand that you want something more user-friendly. Will those URIs be exposed easily to end-user like an HTTP URL? If not, I suspect that user- friendliness is not that important. - we could use short name, something like `s=spaces`. The URI would then be: `xmpp:spaces.server.tld?;node=123;s=spaces` It could work in most cases and be more user friendly. > * to prevent non "space-ready" XMPP clients to open it as a Pubsub > node by mistake Why would this be a problem? A pubsub node is generic, a client must expect generic things, and it may be intended to inspect the node or whatever. I'm not keen on removing the `node` query argument, but if there is a consensus on that, I suspect that using it as query type would be better. > Because this is having effect on the XMPP uris format I'm wondering if > there's some kind of rules or specific requests to follow ? I believe that is should go trough registrar: https://xmpp.org/registrar/querytypes.html Best, Goffi
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Standards mailing list -- [email protected] To unsubscribe send an email to [email protected]
