XSF Discussion - 2018-07-07

  1. daniel

    well it’s a good thing that we have standards https://github.com/siacs/Conversations/blob/master/src/main/java/eu/siacs/conversations/generator/IqGenerator.java#L447-L449

  2. MattJ

    Do we?

  3. MattJ

    For this, I mean

  4. daniel


  5. daniel

    Last time we couldn't agree on a name

  6. daniel

    And decided not to do it at all

  7. MattJ


  8. MattJ

    Well that's silly for something that's clearly being used widely already

  9. jonasw


  10. MattJ

    Also, just this week I added a new config option to control whether members in a members-only room are allowed to invite others (who get automatically added as members)

  11. MattJ

    That should probably also be standardized

  12. jonasw

    good thing that it’s easy to add stuff to the MUC config

  13. MattJ

    https://hg.prosody.im/trunk/rev/92f0876b9230#l1.28 FWIW

  14. daniel

    MattJ, are you also putting this in disco#info

  15. MattJ

    Ah, I meant to check that

  16. daniel

    as a client i need to know whether i should render that invite button

  17. MattJ

    My guess is probably not right now

  18. MattJ

    I'll look at that now

  19. MattJ

    Yeah, we're not, 1-line fix

  20. daniel

    also could you put the roomconfig_name in there?

  21. daniel

    while you are at it

  22. daniel

    i made this point recently on the mailinglist or here in the chat that most servers prefil the identity name with the localpart; thus preventing me from checking if the name is currently empty

  23. daniel

    looking at roomconfig_name would allow me to check that

  24. MattJ

    Ok, shall do

  25. MattJ

    daniel, oh, looks like that's not standard either?

  26. daniel

    MattJ, you can add any roomconfig_name into the disco#info

  27. MattJ

    daniel, says?

  28. daniel

    took me a second to figure that out myself

  29. MattJ

    I thought so too at first

  30. MattJ

    But look closely, and roomconfig != roominfo

  31. MattJ

    roominfo is a separate registry

  32. MattJ

    That said, roomconfig_name isn't there either

  33. daniel

    yes. but you can still add roomconfig things to the disco#info form. give me a second the find that

  34. daniel

    > he muc#roomconfig FORM_TYPE can be included in the extended service discovery fields (as shown above for the "muc#roomconfig_changesubject" field).

  35. MattJ


  36. daniel

    its roomconfig_roomname sorry

  37. daniel

    typed that from memory

  38. daniel

    i had already prepared a PR that registeres roominfo_roomname when i found that

  39. MattJ

    Ahh, ok, thanks

  40. MattJ

    So strange

  41. daniel

    there are quite a few roomconfig things that should be put in the disco#info as well. like allowpm and allowinvites

  42. daniel

    how is your allow_invites different from that by the way?

  43. MattJ

    I considered that (and may be open to persuasion)

  44. MattJ

    But I don't think that just because you can send an invite automatically indicates that you can invite non-members

  45. daniel


  46. MattJ

    In fact the XEP actually states that this should be restricted to admins

  47. MattJ

    "If the room is members-only, the service MAY also add the invitee to the member list. (Note: Invitation privileges in members-only rooms SHOULD be restricted to room admins; if a member without privileges to edit the member list attempts to invite another user, the service SHOULD return a <forbidden/> error to the occupant; for details, see the Modifying the Member List section of this document.)"

  48. daniel

    mhhh fair enough

  49. MattJ

    Prosody doesn't actually implement allowinvites

  50. MattJ

    It's pretty useless given the existence of direct invitations

  51. MattJ

    daniel, done in trunk, looks like this: https://prosody.im/pastebin/029789ae-3642-47da-be6e-957eeec0831b

  52. daniel


  53. flow

    There could be good story about the reason that ejabberd community and saas use different strings for MUC/MAM. Or is it just "don't care"?

  54. jonasw

    yeah, I’d like to hear that story too

  55. flow

    Holger, ^ maybe you can shed some light into this?

  56. Holger

    SaaS has an older code base, that's all.

  57. Holger

    I think it was changed to match the suggestion from some PR or something.

  58. daniel

    Working with ejabberd SaaS is a deja vu for a lot of the small annoying things that got fixed in ejabberd a while ago

  59. Holger

    Yes they're careful merging stuff to avoid breakage, which also avoids fixes. SaaS is a bit like Debian.