jdev - 2022-12-18


  1. soulcaramel

    I found this chat online listed as a general chat for nothing specific , is that true?

  2. soulcaramel

    Oh wow Zash and MattJ mods here, cool to see you guys here too

  3. soulcaramel

    Mb i got the description wrong: geeky discussion about XMPP software development and protocols

  4. soulcaramel

    Things make sense again

  5. soulcaramel

    In xmpp, our messages are only saved on the server hosting the chatroom we’re messaging on right? Unlike matrix which I think copies it to everyone’s server that is involved

  6. Menel

    True

  7. moparisthebest

    soulcaramel: and most people's clients

  8. soulcaramel

    moparisthebest, right most people's clients store the chat log in some sort of cache I imagine

  9. pep.

    In the case of an MSN session leaving a room, only other sessions of the same occupant should be notified? A normal leave with @jid, and nobody else? And of course the session leaving also gets a confirmation

  10. pep.

    Except if this session was the "main" session of this MSN, in which case there's a presence update broadcasted to others?

  11. pep.

    Ah hmm, stuff does get broadcasted I guess if jids are visible

  12. flow

    fwiw, I blieve the situation is not black and white. there a good reasons to not have subdomains for a service, e.g. upload.example.org was a mistake, example.org providing http upload is just fine. but on the other hand, example.org hosting MUC and users under localpart@example.org is not desirable

  13. flow

    and yes, while you can't tell from looking at localpart@foo.example.org if it is a MUC address, you can tell that foo.example.org is not a MUC address if MUC addresses are required to be exactly in localpoart@domainpart form

  14. flow

    and that appears to be a win UX wise

  15. pep.

    I maintain that it doesn't matter to users and that's up to the client to tell them

  16. pep.

    But I'm happy not to redo the same pointless chat today :)

  17. flow

    na life is to short for that

  18. flow

    that said, I still wonder which issues I had in mind

  19. flow

    assume a user types "/message foo@bar.com" how does the client decide if it should be a chat or groupchat message (simply put)

  20. flow

    bar.com would annouce MUC and User features

  21. flow

    so you probably want to disco foo@bar.com?

  22. flow

    which you can't because we want to prevent user enumeration attacks

  23. flow

    "/message foo@bar.com Hello there :)" that is…

  24. pep.

    flow: and yes I agree it's not black and white, also why I was asking for conflicting features

  25. flow

    conflicting features?

  26. pep.

    Yesterday, of rooms and components. It seems they don't answer the same thing in disco#items (but this one matters little)

  27. pep.

    (that is an entity assuming both route would have to choose which one to answer)

  28. pep.

    (that is an entity assuming both roles would have to choose which one to answer)

  29. pep.

    No clues for MSN?

  30. pep.

    https://wiki.xmpp.org/web/MUC_Extensions I added this. Feel free to edit / improve / change the format

  31. pep.

    What's up with dmuc? There's 0281: DMUC1, inbox/dmuc3, inbox/distributedmuc

  32. pep.

    Ah there's also 0282: dmuc2.

  33. MattJ

    pep.: there were discussions that this is something we wanted to do, but there are various ways of doing it. Multiple "competing" XEPs were submitted, some accepted, with the expectation that the best would "win" and advance

  34. MattJ

    I think the winner was FMUC, mainly because it's relatively straightforward (builds well upon existing MUC primitives) and got implemented by (at least) Isode

  35. pep.

    Yeah dmuc2 says "This document is one of several proposals for distributing XMPP chat rooms across multiple chat services. It is expected that the various approaches will be refined and harmonized before a final protocol is developed."

  36. MattJ

    However as a community I don't think there has actually been that much interest in this problem, and a lack of open-source implementations is probably the main reasoln it hasn't advanced

  37. MattJ

    However as a community I don't think there has actually been that much interest in this problem, and a lack of open-source implementations is probably the main reason it hasn't advanced

  38. pep.

    Sure, personally I don't feel like it's a problem worth solving for MUC. I'm just listing MUC extensions

  39. pep.

    (And there are a lot)

  40. Zash

    And then there are those who dismiss XMPP because MUCs live in singular places....

  41. pep.

    Ok it looks like I got most of the extensions..

  42. pep.

    I'm not sure if it's interesting to mark document status :/

  43. Zash

    You forgot the most important one, HATS!

  44. Zash

    You forgot the most important one, HATS 🎩️!

  45. pep.

    Oh, hats!

  46. pep.

    Also not sure about categories in this page..

  47. pep.

    Here, hats added!

  48. Zash

    👒️🎉️

  49. pep.

    Anybody implemented MucSub btw?

  50. pep.

    316

  51. pep.

    s/MucSub/MEP/ rather

  52. Zash

    Once upon a time in Prosody ... https://blog.prosody.im/multi-user-chat-gets-rich/

  53. pep.

    It was removed?

  54. Zash

    Never merged afaik, just experiments, seems to predate the XEP so might just be overlapping acronyms :)

  55. pep.

    Ok

  56. pep.

    hmm, re MSN, even in non-anon rooms users don't see all sessions of an occupant?

  57. soulcaramel

    Xmpp question: image uploads, do they give away where a person’s home server is? Or do the images get uploaded to the server hosting that MUC?

  58. Menel

    Generally they are a service of your homeserver. So generally it's the same Domain and people will know your homeserver.

  59. soulcaramel

    Hmm ok so if you wanted to be anonymous on what your exact jid was you should then change your nickname when joining a public muc

  60. Menel

    Yes

  61. pep.

    It would be nice if the MUC proposed such a service, and certainly not too hard for to support

  62. pep.

    It would be nice if the MUC proposed such a service, and certainly not too hard for clients (impls in general) to support

  63. Zash

    Nothing stops you, except all the implementations not doing it that way.