jdev - 2022-12-15

  1. Beherit

    XSF Announcement We are interested for volunteers to support automation of our tooling to manage the standard documents and allow a more efficient processing of editorial work! Overview: https://github.com/xsf/xeps/issues?q=is%3Aopen+is%3Aissue+label%3A%22Editor+Tooling%22 Contact: editor@muc.xmpp.org Reach out if you are interested or have questions! This is based on: https://wiki.xmpp.org/web/Board-Meeting-2022-12-14 Spread the word: https://fosstodon.org/@xmpp/109519363570575768 https://twitter.com/xmpp/status/1603473596105560079

  2. nicoco

    hey all! is that a valid JID? -\20̗̀\20cucu\20𓆩♡𓆪\20̖́-#2332@discord.localhost/slidge

  3. Zash

    Prosody votes no

  4. nicoco

    prosody the software or prosody the humans behind it? ^^

  5. nicoco

    cause it sure seems to me like the software doesn't mind the horrorshow that this JID is

  6. Zash

    wait, sorry, wrong string syntax

  7. Zash

    Prosody votes unknown

  8. deuill

    Apparently here's what RFC7564 says:

  9. nicoco

    prosody's OK, clients are more or less happy with it. movim and conversation don't particularly mind. gajim is in denial, just wont even mention anything in its logs, understandably. probably in shock.

  10. deuill

    o Code points traditionally used as letters and numbers in writing > systems, i.e., the LetterDigits ("A") category first defined in > [RFC5892] and listed here under Section 9.1. > > o Code points in the range U+0021 through U+007E, i.e., the > (printable) ASCII7 ("K") category defined under Section 9.11. > These code points are "grandfathered" into PRECIS and thus are > valid even if they would otherwise be disallowed according to the > property-based rules specified in the next section.

  11. deuill

    Messed up the quoting but anyways...

  12. Zash

    Check what Unicode version U+131A9 and U+131AA is in

  13. Zash

    Those are rejected by Prosody in strict mode

  14. nicoco

    mmm some day I should try to understand some of these words

  15. Zash

    Prosody lives in the magical past where Unicode 3.2 is the latest one

  16. nicoco

    ok, apart from the "validity" of the JID, it's a bad idea to use these, isn't it?

  17. deuill

    RFC7622 refers to RFC7613, which in turn refers to RFC7654, but essentially the local part is any UTF-8 character in the Letter class?

  18. deuill

    Which doesn't include hearts with wings

  19. deuill


  20. nicoco

    the heart with wings is actually not the worse part. unescaped with rules frop XEP-0106, this is what it looks like: - ̗̀ cucu 𓆩♡𓆪 ̖́-

  21. deuill


  22. Zash

    Based on https://en.wikipedia.org/wiki/Unicode#Versions it looks like Unicode 5.2

  23. nicoco

    I actually find that it's some sort of piece of art. anyway these discord usernames are also case sensitive so it's probably looking for trouble to try and map them to jids directly. nicoco#1234@discord.example.com looks nicer than 1457984984684681681684@discord.example.com though

  24. Zash

    Yeeeeeeeeeah, good luck with that

  25. Zash

    Stuff the fancy thing in https://xmpp.org/extensions/xep-0172.html and do Something™ for JIDs

  26. Zash

    XEP-0106 doesn't do anything for case sensitiveness right? I find nothing

  27. nicoco

    that's already what I'm doing, but for a while I dreamt that "puppets" could have a nice looking JID. when PEP username is used, most clients just hide the JID though, so it's not that bad.

  28. deuill

    Usernames can change on Discord, right? Not user IDs though

  29. Zash

    JIDs can _never_ change

  30. deuill

    So it's likely better to not have the username in the JID

  31. nicoco

    case sensitiveness is not adressed by xep-0106, which is in itself a good reason to forget the nice-looking JIDs

  32. Zash

    So '172 the fancy nicknames (it's existence is motivated by mutable fancy nicknames after all) and pray that clients mostly hide ugly JIDs from users :)

  33. nicoco

    I had worked around it by keeping a map of lowercase to CaseSensitive usernames , and it worked for me though :) It only took a second person to try my thing and boom they broke it.

  34. Zash

    A XEP-0106 variant for arbitrary unicode might be useful in some cases tho...

  35. deuill

    For gateways specifically, I'd vote for keeping the JID typeable since it's much easier to map between XMPP and the legacy protocol if discovery fails on the XMPP side for whatever reason

  36. Zash

    Like those case sensitive Matrix IDs

  37. nicoco

    what do you mean typeable?

  38. Zash

    overlap with keyboards?

  39. nicoco

    oh right. was thinking types like in "type system", and didn't get the point

  40. deuill

    Nah just something you can either write or copy-paste

  41. deuill

    Networks like Discord don't care so much because discovery works well in most cases and the focus isn't 1:1 chats anyways

  42. deuill

    (Which don't exist in Discord AFAIK, 1:1 chats are modeled as group chats with 2 participants)

  43. nicoco

    discord integers ID are not trivial to get from the official discord clients though, unlike username#1234

  44. nicoco

    but I guess we should use jabber:iq:gateway for that. such a pleasant surprise to see it implemented in cheogram

  45. deuill

    I thought you could hover over a person's avatar and get it?

  46. Zash

    deuill, not the same as being easy for "bots" or whatever to get

  47. deuill

    Oh right yeah, that's where discovery comes in, or ad-hoc commands for searching etc.

  48. pep.

    "I'd vote for keeping the JID typeable" < on which layout with which IME?

  49. nicoco

    I just tried and no, the ID is never exposed just the "username#4865" thing - at least I couldn't find it

  50. deuill

    ASCII 4 life

  51. pep.

    Westerners :)

  52. deuill

    Land of the lowest common denominator! Obviously having sane JIDs wouldn't matter if discovery is guaranteed to work, but it doesn't always work and sometimes you have to resort to copypasta

  53. deuill

    Though I'm Greek and am still deathly afraid of IDN

  54. Zash

    Don't make me write a witty retort in runes on a 90kg stone and have it delivered to you via trebuchet

  55. pulkomandy

    ascii is not "common" at all. If you want the common denominator you get… whitespace? Maybe?

  56. Zash

    or maybe https://en.wikipedia.org/wiki/List_of_writing_systems#List_of_writing_systems_by_adoption

  57. moparisthebest

    Nice, so XMPP can go back to ASCII only and the Cyrillic users can just use telegram, problem solved