jdev - 2023-02-10


  1. ralphm

    Zash: the work started earlier. The first draft submitted to the IETF was https://datatracker.ietf.org/doc/html/draft-miller-jabber-00 in February 2002. The start of the `ietf-discuss` mailing list coincides with the first mention of work on that draft on February 1: https://mail.jabber.org/pipermail/standards/2002-February/000207.html. I'm curious if the archives of that old mailing list are still saved somewhere.

  2. ralphm

    The old messages in the standards-jig mailing list archive are a treasure trove, by the way, but on the topic of 5223, see https://mail.jabber.org/pipermail/standards/2002-March/000566.html

  3. ralphm

    And of course https://xmpp.org/extensions/xep-0035.html (SSL/TLS Integration)

  4. Zash

    What! TIL

  5. Zash

    What! TIL (that there was a XEP about it)

  6. ralphm

    This was some preparatory work that eventually went into XMPP Core

  7. moparisthebest

    > 5270 for server-to-server communications TIL on that one, I don't think I've ever seen that in the wild or in server configs

  8. Zash

    I've seen it in the wild, first in ejabberd examples

  9. Zash

    Didn't know it dated back to 2002 tho

  10. ralphm

    Probably earlier even, but I don't have access to other archives right now. Initial development happened mostly on IRC

  11. Wojtek

    quick question - is error message stanza with multiple child erorrs legal? Looking at the specs I don't see anywhere an upper limit in case of stanza error akin to stream errors.

  12. singpolyma

    For message I'm sure it's fine, IIRC iq says only one child and then has a special case for 2 if one is an error

  13. MattJ

    There is only one <error> element

  14. MattJ

    The text of the RFC says that error stanzas contain "an <error/>" not "one or more <error/>" elements

  15. MattJ

    The schema confirms that there may be 0 or 1 <error/> element

  16. MattJ

    If you want to convey additional information about an error, the correct place to do this is inside the <error/> element as an "application-defined error condition", the rules for which are covered in the RFC

  17. MattJ

    There was a discussion recently (here?) about how many of *those* are allowed, and there isn't a clear answer

  18. MattJ

    I would only emit one to be on the safe side, since this is what the RFC would suggest

  19. Wojtek

    Actually I don't want to include multiple errors, I was wondering how to handle incomming message with multiple errors :-) The specification says "contains and error element" and you are right that it's not "one or more" but https://xmpp.org/rfcs/rfc6120.html#schemas-client seems to indicate only lower limit `<xs:element ref='error' minOccurs='0'/>` ?

  20. Peter Waher

    In a sequence, minOccurs and maxOccurs have default value of 1

  21. MattJ

    Yes

  22. Peter Waher

    If writong only minOccurs=”0”, means element is optional, and maxOccurs=”1” implicitly

  23. Peter Waher

    If writing only minOccurs=”0”, means element is optional, and maxOccurs=”1” implicitly

  24. Wojtek

    got it