jdev - 2023-06-12


  1. lovetox

    in the reactions XEP it says

  2. lovetox

    > For other message types the sender should use the 'id' from a Unique and Stable Stanza IDs (XEP-0359) [6] <origin-id> if present, or the value of the 'id' attribute on the <message> otherwise.

  3. lovetox

    this is the use case for direct conversations ( *not* MUC)

  4. lovetox

    so why origin id OR message id?

  5. lovetox

    what is the assumption here, that there exist clients who go through the work and attach a unique stable origin-id, but delibaretly leave the message-id non stable?

  6. lovetox

    as i see it, either a client does not support origin-id -> means i need to use the message-id

  7. lovetox

    or it supports origin-id, then it is expected that the client mirrors this to the message-id -> means i can also use the message-id

  8. lovetox

    on that note, origin-id should just be a tag <origin-id> which indicates that the message-id is stable

  9. lovetox

    but i guess this would need some agreement that message-ids are not allowed to be rewritten

  10. singpolyma

    > origin-id should just be a tag <origin-id> which indicates that the message-id is stable

  11. singpolyma

    that is a great idea

  12. lovetox

    problem with that is, originally origin-id was invennted because some MUCs replaced the user set message-id

  13. lovetox

    and there is nothing in the spec that really forbids that

  14. lovetox

    anyway in MUCs we have now stanza-id which is fine

  15. lovetox

    and in direct conversations this mostly does not matter

  16. lovetox

    like if you really have a contact which uses some broken client, yeah stuff will look broken

  17. lovetox

    but its like contained to this one contact

  18. lovetox

    if in 2023 really clients are active in the wild, that dont use stable message-ids

  19. lovetox

    i will tell the user to drop these clients

  20. lovetox

    hm about Replies XEp

  21. lovetox

    we want to implement that, but i fear many people will not want to use this really

  22. lovetox

    prime use case for quoting is, that you exactly dont want to quote the whole message

  23. lovetox

    this seems not be considered in the XEP

  24. lovetox

    but seems a easy fix, simply include the some "start" and "end" markers inside the <reply> element

  25. singpolyma

    I think reply and quoting as not the same thing

  26. singpolyma

    I have been surprised how many people in XMPP are using email style quoting ever since I rolled out reply features

  27. lovetox

    what is the difference? i fail to see it, it must be very subtle

  28. lovetox

    excerp from the XEP

  29. lovetox

    > When commenting on a previous message, it is sometimes desired to clarify which message the comment refers to. Currently, Message Styling (XEP-0393) [1] quotes are commonly used for that purpose. However, this format does not allow receiving clients to properly identify the referenced message or to adjust the user interface accordingly.

  30. lovetox

    This XEP tries to add more metadata to the commonly used XEP-0393 Quotes

  31. lovetox

    so i dont think its meant as something "different"

  32. lovetox

    its ironic, that in trying to add these metadata, it removes the ability to quote

  33. lovetox

    hm or does it? i could simply not add the <fallback body> tag

  34. lovetox

    then no client can strip the quotes from the body

  35. singpolyma

    Some people use quotes of parts of multiple messages, sent by different people. People also use quotes for stuff that was never originally a message

  36. singpolyma

    I think quoting is a legitimate kind of formatting which just happens to partly overlap with replies

  37. moparisthebest

    > quoting is a legitimate kind of formatting which just happens to partly overlap with replies This