jdev - 2022-09-11


  1. pep.

    I'm confusing which JIDs should go where.. what kind of terminology do you use for your MUC-component impls? - roomjid, barejid of the room - occupant? Many resources (fulljid) from the same account to one participant fulljid - Session? A real(full)jid. - Anything I'm missing?

  2. pep.

    - Participant -- which I named in the occupant line --, fulljid for the occupant in the room

  3. MattJ

    Don't look to Prosody for inspiration here... most of the variables called 'nick' contain an occupant's room JID

  4. MattJ

    That "most" is the best part

  5. pep.

    ^^'

  6. MattJ

    Gradually cleaning that up

  7. MattJ

    Occupant JID is generally clear, I think

  8. MattJ

    And real JID is always clear

  9. pep.

    Occupant JID being.. what I called participant jid?

  10. pep.

    In my list above, an occupant isn't actually visible from the outside

  11. pep.

    Maybe I should just call that occupant jid yeah

  12. pep.

    Well it's visible as a participant

  13. pep.

    But I guess I just took that from roles and it's not actually needed

  14. pep.

    But I guess I just took that word from roles and it's not actually needed

  15. pep.

    Yeah realjid is always clear to me too

  16. pep.

    Even though.. it can be full or bare

  17. pep.

    That's why I also like session

  18. MattJ

    Apart from the multi-session nick stuff, MUC is pretty much based on full JIDs only

  19. pep.

    I start a new MUC implementation, I'm going to include MSN :p

  20. MattJ

    Though given trends, it may be something I'd change in a new implementation

  21. pep.

    But yeah I realized it was mostly fulljids. I was a bit surprised coming mostly from the client side

  22. pep.

    When sending presences at join, I send occupant presences right, not sessions. Unless in non-anon rooms, or.. is it that I just include multiple realjids in the presence?

  23. nicoco_

    when you master all this, will you help me with slidge's MUC support pep, please ? 🙂

  24. nicoco_

    when you master all this, will you help me with slidge's MUC support pep, please? 🙂

  25. pep.

    heh

  26. pep.

    Or I just include one of the session jid, and which one is implementation defined?

  27. pep.

    Because MSN wasn't planned from the start so it's just not there

  28. MattJ

    pep.: Prosody sends multiple real JIDs in presence, though it's not in the spec (or wasn't, can't remember if we updated it)

  29. pep.

    “ If allowed by the service, a user can associate more than one full JID with the same occupant JID (e.g., the user juliet@capulet.lit is allowed to log in simultaneously as the nick "JuliC" in the characters@chat.shakespeare.lit chatroom from both juliet@capulet.lit/balcony and juliet@capulet.lit/chamber). Multi-session nicks are not currently defined in this document.”

  30. pep.

    “If allowed by the service, a user can associate more than one full JID with the same occupant JID (e.g., the user juliet@capulet.lit is allowed to log in simultaneously as the nick "JuliC" in the characters@chat.shakespeare.lit chatroom from both juliet@capulet.lit/balcony and juliet@capulet.lit/chamber). Multi-session nicks are not currently defined in this document.”

  31. pep.

    What does this look like in XML? multiple <item/>?

  32. MattJ

    Yes

  33. pep.

    ta

  34. pep.

    https://xmpp.org/extensions/xep-0045.html#bizrules-presence > A room MUST silently ignore unavailable presence received from a user who has a role of "none". It seems like prosody is sending cancel/service-unavailable instead? That's when a realjid has no existing session and sends an unavailable presence (with no payload). Prosody also sends the same when an there's an existing session but the unavailable presence comes from another resource. But I guess here that may be ok?

  35. MattJ

    I'm not sure why the MUST NOT, I don't see what harm it causes

  36. MattJ

    But it's probably just oversight on our part... we return that error for any stanza not "handled", by default

  37. MattJ

    So MUC probably does ignore it, and then Prosody sees no code did anything useful and lets the sender know with an error

  38. pep.

    Yeah I'm also not sure, both make sense I guess

  39. lovetox

    role = none? means im kicked?

  40. lovetox

    can a joined user have role = none

  41. lovetox

    does this not simply mean, dont broadcast presences for not joined users?

  42. lovetox

    or it goes into this direction to not confuse clients

  43. lovetox

    with presence broadcasts

  44. MattJ

    Uh, well, we might all be talking about different things

  45. MattJ

    I was talking about the response to the sender

  46. MattJ

    It obviously shouldn't be broadcast

  47. MattJ

    Now I'm not sure what pep. was talking about

  48. pep.

    The response to the sender yeah

  49. MattJ

    My guess is that the XEP is thinking about broadcast at this point

  50. MattJ

    Which makes the MUST seem more sensible

  51. MattJ

    More than sensible 🙂

  52. lovetox

    That's what I meant, the xep probably wants to say "don't follow the normal muc roules for this presence"

  53. lovetox

    That's what I meant, the xep probably wants to say "don't follow the normal muc rules for this presence"

  54. lovetox

    But silently ignore is probably to much

  55. Sam

    Random picture of Ralph half way through this article, plus a secretly planted XMPP mention in the middle (even though the article isn't about that): https://www.forbes.com/sites/michaeldelcastillo/2022/09/11/jack-dorseys-former-boss-is-building-a-decentralized-twitter/?sh=1b053b5457c3

  56. pep.

    403. Another website that doesn't like Tor.

  57. moparisthebest

    Or https://www.forbes.com/sites/michaeldelcastillo/2022/09/11/jack-dorseys-former-boss-is-building-a-decentralized-twitter/ with the creepy tracking link removed

  58. pep.

    I guess Mastodon wasn't decentralized enough for Jack.

  59. pep.

    TIL about Twetch. "Decentralized social network -- where everything lives on the blockchain". Are they being serious?

  60. techmetx11

    blockchain is some magical place

  61. techmetx11

    atleast... that's what they say