XSF Discussion - 2020-04-18


  1. lovetox

    flow, role and affiliation has to be in all presences

  2. lovetox

    as these elements can only be in a muc#user extension, the answer is yes

  3. DebXWoody

    XEP-0373: OpenPGP for XMPP Is it ok to add a body stanza for XEP-0373? (e.g.: https://codeberg.org/xmpp-messenger/xmppc/src/branch/master/src/mode/openpgp.c#L74) It's not defined in the XEP.

  4. lovetox

    read XEP-0380 and follow it

  5. DebXWoody

    oh,... next XEP :-) I will check, thx

  6. DebXWoody

    I think the namespace for OMEMO is wrong in https://xmpp.org/extensions/xep-0380.html#protocols. Has been changed in XEP-0384 V 0.4.0?

  7. lovetox

    yes Link Mauve ^

  8. pep.

    it's not "wrong", it's still very much a thing

  9. pep.

    also https://github.com/xsf/xeps/pull/907

  10. flow

    jonas’, why doesn't https://xmpp.org/extensions/xep-0390.html#algorithm-input talk about sorting in the actual algorithm?

  11. jonas’

    uhh

  12. jonas’

    oversight I guess

  13. jonas’

    ah, it does

  14. jonas’

    > Join the resulting octet strings together, ordered from lesser to greater.

  15. flow

    ahh, i greped for sort

  16. jonas’

    yeah, me too :D

  17. flow

    and asumed the sorting happened before the "for each" loop

  18. Mikaela

    Hi, would you have any comments on PrivacyTools delisting XMPP due to E2EE-by-default not being required by any XEP currently? https://github.com/privacytoolsIO/privacytools.io/pull/1836 Personally I am unable to approve this PR in good conscience as I see federation more important to privacy than E2EE-by-default, but I have opened a tracking issue in case that will happen picking up promising looking XEPs from the future development of compliance suites https://github.com/privacytoolsIO/privacytools.io/issues/1838

  19. pep.

    ugh people really need to stop confusing spec and implementations

  20. jonas’

    Mikaela, there is no standards venue where requiring E2EE would be acceptable.

  21. jonas’

    XMPP is more than just instant messaging, and the XSF has to consider this.

  22. jonas’

    A specification which requires E2EE by default would be tricky at least. It could be on Informational track, it would stay controversial, and may or may not make it into IM-Core.

  23. jonas’

    there are also situations where high security demands of a group counter the E2EE interests of the individual, making E2EE impossible to use.

  24. jonas’

    so, no, it wouldn’t make it into IM-Core

  25. jonas’

    probably not IM-Advanced

  26. jonas’

    so yeah, it’s not going to happen.

  27. jonas’

    not under the XSF umbrella either way

  28. Mikaela

    I was wondering if some compliance suite would be a place for it, but that is in line with my thoughts

  29. jonas’

    (my personal judgement of the situation mostly)

  30. pep.

    I still think that basing the PrivacyTools decision on this is wrong though

  31. jonas’

    yes

  32. jonas’

    I agree

  33. pep.

    Conversations has it by default, they can list it there

  34. jonas’

    Snikket, too?

  35. pep.

    sure..

  36. jonas’

    Snikket would be a thing to place there

  37. jonas’

    more than XMPP itself anyways

  38. pep.

    snikket, and maybe things like Tigase things, Xabber things, (if they do e2ee by default, dunno)

  39. jonas’

    Xabber thinks E2EE is for drug dealers only

  40. pep.

    it's all good then, it's private :)

  41. !XSF_Martin

    But they would implement OMEMO for 2 bitcoins

  42. pep.

    Mikaela: thanks for keeping an eye on this btw :)

  43. Mikaela

    you're welcome 💜️, it's kind of my responsibility

  44. Jeybe

    I see compliance suite has added STUN & TURN discovery as informational result, is this due to the planned support of VoIP Calls in Conversations?

  45. Daniel

    Jeybe, not sure if this is the right channel for that question. but yes

  46. flow

    ahh the other compliance suites, not xep423

  47. Zash

    Did you mean the Conversations Compliance test tool?

  48. !XSF_Martin

    Zash: Yes, a STUN TURN test has been added to caas

  49. flow

    so xep128 is not really clear if there can be multiple forms. or am I wrong? xep115/390 algo however handles multiple forms. should this be clarified in xep128?

  50. Zash

    Clarification good.

  51. jonas’

    flow, disco#info uc.xmpp.zombofant.net for an example of multiple forms

  52. jonas’

    I suppose many http upload services have that because of differing versions

  53. jonas’

    (though I’m not sure if uc.xmpp.zombofant.net federates)

  54. jonas’

    here’s a textual representation: extensions: urn:xmpp:http:upload:0 var='max-file-size' values=['104857600'] urn:xmpp:http:upload var='max-file-size' values=['104857600']

  55. Zash

    > No certificate provided by uc.xmpp.zombofant.net

  56. jonas’

    right

  57. jonas’

    it’s not configured for federation for obvious reasons :)

  58. jonas’

    adding explicit wording in '128 is certainly not bad; it was quite obvious for me though that '128 allows multiple forms

  59. Jeybe

    >Did you mean the Conversations Compliance test tool? Zash‎, yes, meant the compliance checker, not the suite. pardon

  60. Zash

    Could someone remind me why XEP-0353 appears to be stuck in Proposed?

  61. Daniel

    we last called that a while ago; council rejected it

  62. Daniel

    and then editor fogot probably

  63. jonas’

    seems realistic

  64. jumplkuter

    hi

  65. Jeybe

    jumplkuter‎, hello

  66. rion

    .v

  67. rion

    wrong chat =)

  68. MattJ

    Mikaela: I feel that XMPP is unfairly demoted on privacytools.io. The Matrix section more or less assumes Riot and briefly mentions other clients may not (in my experience always are not) up to the same standard

  69. MattJ

    A fair comparison would be better against specific XMPP implementations, with a similar note

  70. MattJ

    A simple improvement (from a usability perspective also) would be to indicate which listed XMPP clients have default E2EE, instead of just bucketing them all as "inconsistent"

  71. MattJ

    That would help guide people to the more privacy-conscious implementations

  72. Zash

    If you value privacy (or really any single property) over all else with no compromises your choinces tend to be limited.

  73. MattJ

    Well I think it is good to give people informed choice