jdev - 2019-10-07


  1. Daniel

    Do servers these days inject stanza id even if MAM is disabled? Should they?

  2. jonas’

    what would be the point?

  3. guus.der.kinderen

    The Unique and Stable IDs, you mean? Openfire does (for MUC).

  4. guus.der.kinderen

    (unless explicitly disabled)

  5. guus.der.kinderen

    The rationale was that XEP-0359 might be used for other (although unidentified) purposes than MAM, and adding it didn't seem to be much of a problem.

  6. Daniel

    > what would be the point? Other deduplicational purposes. Something something

  7. guus.der.kinderen

    XEP-0359 is supposed to be a stand-alone feature, as I read the XEP. MAM depends on it, but that doesn't need to imply that there's no other use case for XEP-0359 than MAM.

  8. Daniel

    I don't really know

  9. guus.der.kinderen

    It's a lot of assumptions on my part to.

  10. lovetox_

    one simple use case could be, if a muc has disabled mam, this does not mean it will not enable it at some point

  11. lovetox_

    and on that point if i join i would like to request from my last known stanza-id

  12. lovetox_

    which implies i have to track stanza-ids even if mam is disabled

  13. jonas’

    lovetox_, or if MAM is enabled and you don’t have any stanza-ID, you request from oldest?

  14. jonas’

    that’d make more sense anyways

  15. lovetox_

    if i dont have a stanza-id, how can i request from the oldest?

  16. jonas’

    without <after/>

  17. lovetox_

    of course, but this yields duplicates

  18. lovetox_

    im not saying a client is not usable, i say it makes sense to track stanza-ids even if mam is not enabled, because it probably will be enabled at some point in the future

  19. lovetox_

    but if i think about it, if mam is enabled, the stanza-id i request will not be in the archive

  20. lovetox_

    so its useless anyway

  21. Daniel

    Yeah I don't know. I just had the vague feeling that it might be useful. Mostly triggered by a bug / very paranoid code in Conversations that absolutely requires stanza ids to be there.

  22. lovetox_

    yeah in that case i could request the whole archive and never receive duplicates

  23. Daniel

    But that's probably just Conversations being too paranoid

  24. Daniel

    I do see how they can be useful for muc though.

  25. lovetox_

    i think i only record the stanza-id if mam is available

  26. lovetox_

    how Daniel?

  27. Daniel

    To dedup even if catching up with the old history model

  28. lovetox_

    yes true

  29. lovetox_

    i wonder if it would be ok to just never use MUC history

  30. jonas’

    lovetox_, how would it yield duplicates?

  31. lovetox_

    if a server does not install MAM, does he really care about history?

  32. lovetox_

    jonas’ i already corrected that statement

  33. jonas’

    right

  34. flow

    Daniel, I would only require <stanza-id/> with by=room@muc.example.org if room@muc.example.org announces the urn:xmpp:sid:0 feature, and potentially, by implication, flag/remove those if the feature is not announced, because it is likely spoofed (although I am not sure if this could be exploited, I can't hurt doing so)

  35. allie

    anyone know if anybody is working on any modern web interfaces to MUCs that isn't ConverseJS?

  36. jonas’

    I don’t think there is

  37. allie

    hmmm... I might have to play around with writing something, if only to see if I can do it. I like conversejs, but it's a bit clunky for some things

  38. Zash

    There are other web clients, I know that much.

  39. allie

    there was that one really nice one... with a k. I forget the name. but it looks abandoned now. and candy for MUC is just bad...

  40. Zash

    Am I the only one who thought Candy was okay?

  41. allie

    Zash: maybe? lol

  42. allie

    Zash: it works, it's just not ideal UI wise. I don't know about the backend. it might be something I can fix to my own satisfaction with themeing

  43. Zash

    It doesn't look like Slack, sure. But I had never seen anything like Slack back then.

  44. Zash

    Looked like any other ~IRC-ish chat to me.

  45. allie

    there are a few sites I've used it on before and it just wasn't quite up to the task once you got a bunch of people involved.

  46. Zash

    allie, looked through the ones listed as "Browser" on https://xmpp.org/software/clients.html ?

  47. Zash

    or the raw data https://github.com/xsf/xmpp.org/blob/master/data/clients.json

  48. Zash

    allie, Kaiwa?

  49. Link Mauve

    Zash, Candy had many vulnerabilities, and nobody was maintaining it.

  50. Link Mauve

    allie, Reventlov recently started a React interface for Converse.

  51. Zash

    Link Mauve, "back then" was a decade ago.

  52. Zash

    Around when it was actively developed

  53. allie

    Zash: yeah kaiwa! it's unmaintained now tho :(

  54. Zash

    There was a fork? Or is Kaiwa the fork?

  55. Zash

    Ah, yeah, otalk.

  56. Zash

    Also unmaintained 🙁

  57. allie

    it'd be nice if rocket chat would work with an xmpp backend, but that hasn't happened so far

  58. Zash hides mod_rocketchat behind his back and whistles innocently

  59. allie

    Zash: whoa whoa wait what?!

  60. Zash

    Fun fact: Supporting the desktop client is the most trivial thing ever.

  61. Zash

    ... because it's just a web browser that loads the web client, so you can stick some JSON in the right place and serve converse.js or something 🙂

  62. allie

    Zash: but then it's converse and not rocket chat, right?

  63. Zash

    I do have some WIP that sorta works with their mobile client tho.

  64. allie

    rocket chat has a pretty nice interface. we use it at work.