XSF Discussion - 2021-09-17


  1. cyberduckgo

    Hi

  2. debacle

    Question about "XEP-0060: Publish-Subscribe" `pubsub#item_expire` : Is this meant to be relative to the creation time of the item or last modification? Both can make sense, depending on use case, right?

  3. flow

    true

  4. debacle

    IMHO, time of last modification makes more sense, because the standard also says: > Note: If a publisher publishes an item with an Item ID and the ItemID matches that of an existing item, the pubsub service MUST NOT fail the publication but instead MUST overwrite the existing item and generate a new event notification (i.e., *re-publication is equivalent to modification).*

  5. flow

    also true, but that makes it harder for the use case where you want to publish some information until a certain absolute deadline passes, and you edit that information after it was initially published

  6. flow

    so, two deaths, chose one

  7. debacle

    For singleton nodes and also for blog posts, I assume, that modification is better in most cases. For an absolute deadline, IMHO expiry time is not very good, because it is configured per node not by item. As long as there is no absolute expiry time per item, I would better implement that on side of the publisher, i.e. the publisher should "manually" remove the item, when time comes.

  8. debacle

    But whatever option, most important would be that the standard does not leave this question open ;-)

  9. flow

    debacle, ahh right, fair point

  10. debacle

    https://xmpp.debian.org:5281/upload/6lQnCKqHOy1J9frJ/xep-0060.patch

  11. flow

    debacle, PR it!

  12. debacle

    flow Uff, one day I need to learn all that modern stuff...

  13. Zash

    email the patch to the editor 😉

  14. flow

    debacle, well you *could* mail the patch to the editor

  15. flow

    but, think of it this way: if you PR it via github, you actually incurr cost to microsoft ;)

  16. debacle

    yes, that's how old people like myself work, inefficiently ;-)

  17. flow

    although MS may benefit at the end of the day from it anyway

  18. debacle

    I'm sure, that MS uses XMPP *somewhere* - so they benefit from everything we do to improve XMPP ;-)

  19. Maranda

    > I'm sure, that MS uses XMPP *somewhere* - so they benefit from everything we do to improve XMPP ;-) I'm sure enough you're wrong

  20. flow

    MS is a pretty big company (plus associated companies), and somewhere is a pretty low bar to jump. a single game using XMPP as basis for its lobby would be sufficient…

  21. Maranda

    There was a XMPP gateway in Skype for Business Server but that was deprecated in Skype for Business Server 2019

  22. Maranda

    (see: https://docs.microsoft.com/en-us/skypeforbusiness/migration/migrating-xmpp-federation)

  23. Maranda

    (Beside that SBS is on its way to die as well)

  24. MattJ

    https://matthewwild.co.uk/uploads/screenshot-20210917-1631871982-13372.png

  25. MattJ

    Well done Google

  26. jonas’

    hah! :)

  27. Zash

    We're doing XMPP and we're still alive 🎵️

  28. christian

    MattJ: we must be optimistic, we need better clients. Pidgin should have been improved instead to develop so many other clients to beta and then abandon them. (excuse my horrible english)

  29. MattJ

    The truth is Pidgin never was a great XMPP client

  30. MattJ

    And the same is generally true of all multi-protocol clients

  31. christian

    I used Gaim :)) and it rocked. It had voice and video on windows95 and it was working with yahoo google and aol.

  32. christian

    I forgot icq

  33. Holger

    MattJ: FWIW my impression wasn't that bad. Functionality was basic bad good enough for most users back then, and it was rock solid, which wasn't true for all clients (looking at all your tracebacks, Gajim).

  34. Holger

    MattJ: FWIW my impression wasn't that bad. Functionality was basic but good enough for most users back then, and it was rock solid, which wasn't true for all clients (looking at all your tracebacks, Gajim).

  35. christian

    https://developer.pidgin.im/wiki/GSoC2008/VoiceAndVideo

  36. flow

    hehe, although I am not sure if I want to use a client written in a non-typesafe language in 2021

  37. Holger

    I think I did A/V with Pidgin just fine, some years ago.

  38. Holger

    flow: Yeah a type safe language most probably would've avoided most of those Gajim traces 😉

  39. Holger

    flow: Yeah a type-safe language most probably would've avoided most of those Gajim traces 😉

  40. flow

    Holger, you mean statically typed languages would avoided those

  41. Holger

    Yes yes, I saw that coming.

  42. christian

    People identify xmpp with pidgin, for that reason alone we should evolve pidgin and not create a new status symbol. It is like an artifact. Believe me, I know marketing :))

  43. christian

    You can not say a religion is 2000 years old if you write a bible every 10 years.

  44. mdosch

    > People identify xmpp with pidgin It is a widely used client but I don't think this is true.

  45. meeeeson

    christian: I think the number of ppl identifying Pidgin with XMPP is declining statically (for good) :)

  46. Maranda

    (thank gods)

  47. jcbrand

    The Shiny Flakes documentary has lots of scenes where he's using XMPP, even showing addresses like xxx@jabbim.cx And guess what chat client he's using? https://www.imdb.com/title/tt15036416/

  48. wgreenhouse

    jcbrand: probably he was using tails, where it's still part of the base install

  49. jcbrand

    He ran a wordpress site on the clear net and had 300kg of drugs in his room (all while being a minor)

  50. jcbrand

    Somehow I don't think he used Tails

  51. wgreenhouse

    ope

  52. jcbrand

    Fun fact, on his website, he promised not to keep any data on his customers. In reality he had an Excel sheet with IIRC 13000 orders, with names, addresses and what they bought.

  53. jcbrand

    Many of those people were later prosecuted and he had to go as witness to those trials

  54. jcbrand

    I'm tempted to give more spoilers concerning his terrible opsec, but maybe some people want to watch the documentary. It was quite good IMO

  55. wgreenhouse

    sounds like I should watch this on a day when I need some tragicomedy

  56. christian

    This is very OT: This is all normal, you never catch the dealers until you have uncovered and arrested the whole network. The dealer is the last to be arrested, and is usually released immediately, because he uncovers the trail to the next network.

  57. mjk

    > We're doing XMPP and we're still alive 🎵️ `And when they're dead, we will be still alive`

  58. bung

    > Zash wrote: > We're doing XMPP and we're still alive 🎵️ Yes

  59. bung

    I don't think Xmpp's going to end.

  60. Zash

    All good things must come to an end.

  61. Zash

    https://en.wikipedia.org/wiki/Heat_death_of_the_universe

  62. bung

    > Zash wrote: > All good things must come to an end. I have need immortality

  63. Zash

    Turn evil you say?

  64. bung

    What?

  65. Zash

    Good things → ends Bad things → IMMORTALITY Therefore: Evil QED

  66. bung

    Why immortality is bad?

  67. Zash

    Because it does not end. Good things end. Therefore immortality ≠ good.

  68. Zash

    Therefore we must turn evil!!11!!1

  69. bung

    I think immortality is good

  70. bung

    Relaxing

  71. bung

    And why would the software that's been around for twenty-two years end?

  72. Holger

    Assuming it won't, it feels quite exciting how we've been involved with the very first decades of a spec that will live on for billions of years.

  73. mjk

    BILLIONS

  74. Holger

    Yes.

  75. Holger

    Maybe people will remember us as the early pionieers a billion years from now.

  76. Holger

    Prosody's community repo will hold billions of modules. (hg will cope, it's superior to Git.)

  77. Holger

    ejabberd will scale to billions of nodes. YY.MM release versioning scheme will become problematic tho.

  78. Holger

    Sam will be going nuts due to billions of redundant XEPs for publishing avatars.

  79. MattJ

    MIX will be nearly deployed

  80. mjk

    Eh, there's plenty of time til the end of the century to change the versioning scheme to include more Ys

  81. mjk

    Sorting versions, otoh... ugh

  82. Sam

    heh, indeed, the situation is bad enough already, I dread seeing it in a few million years

  83. mjk

    > MIX will be nearly deployed Despite the billions of "against" and "strongly against"

  84. millesimus

    Has anything changed with the newsletter feed? For some reason, my reader keeps loading all feed items since 2008, marking them as unread…

  85. MattJ

    I was worried that may happen :)

  86. MattJ

    We switched to a new site generator, and some things necessarily changed

  87. millesimus

    It happened twice or thrice already, that's why I'm asking. Were there several changes?

  88. moparisthebest

    MattJ: have to wonder if Google gives the same answer to matrix fans though...

  89. MattJ

    millesimus, that sounds plausible - the first roll-out accidentally switched to summarized articles, instead of including the full post content in the feed. That was fixed, along with some other things, so it's possible that caused the issue. I'd hope it doesn't happen again though, so feel free to shout if it does.

  90. millesimus

    Will keep an eye on it. :)

  91. lovetox

    goffi

  92. lovetox

    FYI, your email server fails for dkim and dmarc

  93. lovetox

    DKIM: 'FAIL' mit Domain goffi.org DMARC: 'FAIL'

  94. lovetox

    this means it lands in spam

  95. goffi

    lovetox: oh really? Damn I thought this was properly configured, I'll check it, thanks.

  96. lovetox

    SPF : PASS though :)

  97. lovetox

    on the topic, its sad that not more server support #item_expire

  98. lovetox

    its incredibly useful for stuff like

  99. lovetox

    posting a Status / Current Music etc to pubsub

  100. lovetox

    such information is not very long useful and accurate

  101. lovetox

    and its impossible for clients to remove it in all circumstances

  102. lovetox

    the idea is like whatsapp for example automatically expire it after some time, status for example after 24 hours

  103. Daniel

    lovetox: I think we (as a community) have a decent history of creating server side and client side implementations simultaneously at the same time

  104. Daniel

    If for example you wanted to implement WhatsApp like statuses in your fairly popular client I'm sure server developers can be convinced to implement this

  105. Holger

    > on the topic, its sad that not more server support #item_expire Funny, a customer asked me to implement this literally yesterday.

  106. Holger

    (Ah, he's joined to this room and spawned the topic, so not a coincidence, heh.)

  107. lovetox

    Daniel, i asked the prosody team for example to implement this, and i think i got an email a few days ago which indicates they might consider doing it. This was not meant as a rant. I just took the chance, while the option was discussed at the mailing list, to bring awareness how the option might be useful to users. Because its maybe not evident to server devs.

  108. lovetox

    in retrospect i probably should not have used the word "sad". its not really a good fit for the situation.

  109. Daniel

    A customer of mine recently implemented whatsapp like statuses (for an xmpp client) out of band over HTTP instead of microblogging and pubsub. Which is super sad. I would have loved to be paid to implement this with xmpp

  110. Daniel

    I actually think this would have worked out pretty nicely