XSF Discussion - 2018-11-09

  1. Ge0rG

    Is there a way to persistently put something into your presence for your contacts to see it without keeping a client logged in?

  2. jonas’


  3. jonas’

    use PEP for that, maybe?

  4. Ge0rG

    What if I want to limit it to individual contacts (speaking for a Moved redesign)

  5. Zash


  6. Ge0rG

    So every of my contacts(' servers) need to subscribe to my Moved node?

  7. Ge0rG

    you could send an 'unsubscribe' followed by a 'subscribe', because the server MUST store the latter!

  8. Ge0rG

    one could send a <message><moved/> as well

  9. jonas’

    Ge0rG, ahh, that’s the context

  10. jonas’

    sending a <moved/> message + the unsubscribed+unsubscribe presence is probably not bad

  11. Ge0rG

    you can send a directed/broadcast presence with <moved/>, but that will vanish as soon as you logout

  12. jonas’

    sending a <moved/> message + the unsubscribed+unsubscribe presence including the <moved/> is probably not bad

  13. jonas’

    Ge0rG, you could send type="unavailable" when you log out

  14. jonas’

    with the payload

  15. Ge0rG

    jonas’: will the server cache that?

  16. jonas’


  17. Zash


  18. Ge0rG


  19. jonas’

    I feel like it should

  20. Zash

    Prosody doesn't

  21. jonas’


  22. Ge0rG

    it should also cache an 'unsubscribe'

  23. Ge0rG

    It should, but unfortunately it only MAY.

  24. Ge0rG

    can has PR 6121?

  25. jonas’

    add erratum

  26. Ge0rG

    Where to?

  27. jonas’

    dunno, figure out ietf process

  28. jonas’

    dunno, figure out ietf process?

  29. fippo

    ge0rg: wasn't the server allowed to return your last unavailable presence?

  30. Ge0rG

    fippo: "allowed to" is not a behavior I can rely on

  31. fippo

    i think that was allowd but then not used by many servers because it means a reply to *every* probe which makes presence traffic even worse

  32. fippo

    xep-12 example 5 might do the job

  33. fippo

    but i am not sure how many servers implement that either. and sending an extra iq to each contact...

  34. jonas’

    don’t servers have to reply to every probe anyways?

  35. Ge0rG

    fippo: it doesn't even contain the whole presence, just the non-i18n status text.

  36. Ge0rG

    > Clients accepting the moved subscription SHOULD indicate to the user that that this subscription request was the result of a move operation and because of potential malicious behavior SHOULD NOT auto-accept the subscription without displaying the <moved/> JID to the user. I'm reading this as "clients MAY auto-accept the subscription if they display the <moved/> JID to the user." - am I right in that interpretation of English?

  37. Zash

    Isn't RFC 2119 language not supposed to apply to UI things?

  38. Ge0rG

    Zash: yes.

  39. MattJ

    Zash, conventionally

  40. Ge0rG

    But that doesn't answer my question.

  41. Ge0rG

    I know it's not as much fun as discussing the applicability of lower-case RFC 2119 to UI, but still.

  42. Zash

    Clients ShOuLd nOt

  43. Zash

    Ge0rG: I believe I interpret the quoted text the same way you do.

  44. Ge0rG


  45. Zash

    Implementation recommendation: Tell the user what happens.

  46. Ge0rG

    What if the user isn't at the screen.

  47. Ge0rG

    And there is still no clean way to persistently convey a <moved> when unsubscribing.

  48. Zash

    Send out the informationator bot to search for and .. inform the user.

  49. Ge0rG

    I think I'll have to <message> it.

  50. Zash

    Get MattJ to look into https://issues.prosody.im/689 :)

  51. Zash

    Blocker for all the nice things Ge0rG wants to fix :(

  52. Ge0rG

    Zash: but this is only for 'subscribe'

  53. Ge0rG

    is it still not fixed?

  54. Zash

    nice things are not to be had!

  55. Zash

    Ge0rG: Oh there's no such rule about keeping the unsubscribe(d)?

  56. Zash

    Moved doesn't help when I lost my phone along with the autogenerated password it created the account with?

  57. Ge0rG

    Zash: no.

  58. Ge0rG

    and no.

  59. Zash

    Message might be more reliable then.

  60. Ge0rG


  61. Zash

    <message> all the things

  62. Ge0rG

    so send the same payload three times: message, unsubscribe, unsubscribed

  63. Zash


  64. Ge0rG

    Zash: https://xmpp.org/rfcs/rfc6121.html#sub-cancel-inbound > Implementation Note: If the user's account has no available resources when the inbound unsubscribed notification is received, the user's server MAY keep a record of the notification (ideally the complete presence stanza) and then deliver the notification when the account next has an available resource. This behavior provides more complete signaling to the user regarding the reasons for the roster change that occurred while the user was offline.

  65. Ge0rG

    Maybe I need to rewrite PARS to use <message> then.

  66. Ge0rG

    because prosody.

  67. Ge0rG

    And then it's getting really weird. If a client auto-accepts a Moved, all other clients need to know about it as well.

  68. Ge0rG

    So they can change message archive etc.

  69. Zash


  70. Zash

    Ah, yeah

  71. Ge0rG

    The longer I think about it, the worse everything becomes.

  72. Zash

    Things are complicated, who knew?

  73. Ge0rG

    a <message> will only be there in MAM for a limited time

  74. Kev

    Will it?

  75. Ge0rG

    Kev: do you have eternal MAM?

  76. Kev

    I mean, heat death of the universe or something, sure, but possible as long as you're using your account.

  77. Kev


  78. Ge0rG

    Kev: that's great!

  79. Ge0rG

    My server had a meltdown because I was storing 14 days worth.

  80. Kev

    isode.com's running with a few years now, I think.

  81. Ge0rG

    Kev: how many indian trading bots are there on isode.com?

  82. Kev

    Possibly fewer than average.

  83. Ge0rG

    I don't know about other public servers, but there is a handful on mine.

  84. Ge0rG

    with something like 10k msg/day

  85. Ge0rG

    I suppose we don't have something like an account tombstone, right?

  86. Ge0rG

    PEP or vcard would be potential candidates to put one in.

  87. perflyst

    Hi, is there a rss feed for https://xmpp.org/blog.html available? I couldnt find one...

  88. Zash clicks RSS icon

  89. Zash gets https://xmpp.org/feeds/all.atom.xml

  90. Zash

    Browsers hate this one small trick

  91. perflyst

    I didnt see that "Feeds" on the left, thanks :)

  92. Zash

    Oh look, there too

  93. Zash

    In Firefox one can add back the RSS button that they removed because reasons.

  94. perflyst

    what do you mean? rss button in firefox browser or on the website?

  95. Zash

    in firefox

  96. perflyst

    yeah that is because they removed / are removing rss support in firefox

  97. perflyst


  98. Ge0rG

    RSS is dead. Together with NNTP and XMPP.

  99. Zash

    Can the Web be next?

  100. Ge0rG

    No, the web is way too shittily designed to go down

  101. Zash

    To shitty to fail!