XMPP Council - 2024-09-10


  1. daniel

    Sorry I didn't find the time for editor work the past week and I also didn't send an agenda. I did however just read the two new proto xeps just now and I'm +1 on both

  2. daniel

    Note that the vote expires today so if you haven't read them yet now might be a good time 😊

  3. dan.caseley

    > Client Access XEP - I've got a nit about using service-unavailable rather than not-allowed/not-acceptable, but that's not really enough to reject on. It also has no XML schema. What really gives me pause is the opaque client id - there's no implementation notes or suggestions, just that it's 'opaque'. Is there prior art that makes my concern unusual or unreasonable? Anyone got a take on my thoughts on the Client Access XEP? Is it overly harsh?

  4. daniel

    dan.caseley: not having a schema is pretty common at this stage

  5. Zash

    > In order for a Standards Track XEP to advance from Proposed to Stable, it must: > 6. be formally defined by an XML schema per XEP-0001, so shouldn't be a hard requirement until then

  6. dan.caseley

    The opaque client ID bothers me most.

  7. daniel

    And id being opaque. I thinks that's like an archive id for example? I don't understand why opaque should be an issue

  8. daniel

    It's not User defined or anything is it?

  9. daniel

    Or iq ids...

  10. singpolyma

    I think all ids in XMPP are opaque

  11. daniel

    I mean we can make it a UUID like in sasl2/bind2 but I find that pretty nit picky for an experimental xep

  12. singpolyma

    Which is usually what you want

  13. singpolyma

    Even in sasl2 doesn't it just suggest uuid? Obviously that's a sane choice usually but there's no point to require id

  14. singpolyma

    Even in sasl2 doesn't it just suggest uuid? Obviously that's a sane choice usually but there's no point to require it

  15. daniel

    I think it suggest uuidv4 specifically and some server implementations check

  16. daniel

    But I'm fine with it simply being opaque

  17. moparisthebest

    I will be +1 on both Though I highly dislike adding the misfeature of user-agent to XMPP, don't most modern things instead ask users to add a nickname for clients? That's just feedback for MattJ I guess, nothing blocking

  18. dan.caseley

    Okay. If it's established in history as being normal, I'm happy being +0

  19. MattJ

    moparisthebest, I can't recall something asking me to name a login session

  20. moparisthebest

    Also thanks for the great security considerations on both that's a pet peeve of mine (:

  21. moparisthebest

    MattJ: every "grant access to this" ux I've seen, uh tokens in gitlab and GitHub and Gmail etc

  22. singpolyma

    We have lots of user agent equivalents already just not yours part of auth

  23. MattJ

    moparisthebest, really? Tokens, perhaps, but I don't recall ever seeing that as part of an OAuth flow

  24. daniel

    It's time

  25. moparisthebest

    I'm not sure I've ever seen an oauth flow in real life, so token management is the only example of granting/revoking account access I'm familiar with

  26. daniel

    1) roll call

  27. moparisthebest

    Hello!

  28. dan.caseley

    Howdy!

  29. daniel

    I assume singpolyma dan.caseley are here

  30. daniel

    Is larma here

  31. larma

    yes

  32. daniel

    2) agenda bashing

  33. daniel

    No agenda so nothing to bash

  34. daniel

    3) editors update

  35. daniel

    I didn't find the time to do editor tasks this week. But my availability will improve towards the end of the day

  36. daniel

    4) items for voting

  37. daniel

    None

  38. daniel

    5) pending votes

  39. daniel

    None I think. Since people voted just now

  40. daniel

    6) date of next

  41. daniel

    +1w wfm

  42. moparisthebest

    +1w wfm

  43. singpolyma

    +1w wfm

  44. dan.caseley

    +1w wfm

  45. daniel

    7) AOB

  46. moparisthebest

    Nothing here

  47. daniel

    Assuming none

  48. daniel

    8) close

  49. daniel

    Thank you all. See you next week

  50. dan.caseley

    Thanks daniel!

  51. moparisthebest

    Thanks daniel and all