jdev - 2020-03-17

  1. ankit

    hi

  2. ankit

    Kev

  3. anana

    ankit

  4. defanor

    Is a roster item's 'jid' attribute required to be a bare JID? I'd think it is so, but failing to find where it is specified.

  5. Ge0rG

    defanor: it can't be a full jid, but it may be a domain jid

  6. defanor

    Ge0rG: Oh, makes sense. But still, is it specified somewhere, or just implied?

  7. Ge0rG

    defanor: I can't find anything in RFC6121

  8. MattJ

    I thought it was one of the things clarified in the bis RFCs :/

  9. MattJ

    There has been discussion about it previously

  10. MattJ

    You can't sensibly have a subscription to a full JID though, so it doesn't make sense in many ways

  11. jonas’

    MattJ, though it would allow to Auto-GC1.0-Join MUCs

  12. Zash

    Oh no

  13. pep.

    GC1 is a myth

  14. jonas’

    rightfully so

  15. flow

    MattJ> You can't sensibly have a subscription to a full JID though, but you can have a roster item without subscription, sooo…

  16. flow

    I think it is imporant when designing xmpp to think about and specific which types of JIDs are sensible and allowed at every place an xmpp address appears ✎

  17. Zash

    Is the relevant text in the parts about subscription handling?

  18. flow

    I think it is imporant, when designing xmpp, to think about and specific which types of JIDs are sensible and allowed at every place an xmpp address appears ✏

  19. Zash

    https://xmpp.org/rfcs/rfc6121.html#sub > the server MUST stamp the outgoing presence stanza with the bare JID <localpart@domainpart> of the sending entity

  20. Zash

    https://xmpp.org/rfcs/rfc6121.html#sub-request-gen > When a user sends a presence subscription request [...], the value of the 'to' attribute MUST be a bare JID