jdev - 2022-10-14


  1. deimos

    is libstrophe for c considered reasonably current?

  2. cccnlv

    请问有会中文的吗?

  3. MattJ

    deimos: check what profanity is using

  4. MattJ

    I seem to recall it was using libstrophe and is actively developed

  5. pep.

    > 请问有会中文的吗? Does someone speak Chinese btw? Even though they're gone now

  6. Kev

    That was the question.

  7. Kev

    (Thanks, Google)

  8. pep.

    Yeah that's why I asked

  9. pep.

    I didn't especially want to know what they said, I was more curious if someone could answer them

  10. deimos

    > deimos: check what profanity is using ok, i don't see libstrophe there, but I might have missed it

  11. deimos

    generally, we're standardizing on xmpp/mqtt at work, and i'm soon to publish my julia lib for xmpp, which is a wrapper around libstrophe

  12. deimos

    the goal is to get the julia xmpp lib accepted into juliahub so it can be generally used by the community

  13. selurvedu

    So what do you guys think regarding the ejabberd news? https://www.process-one.net/blog/matrix-protocol-added-to-ejabberd/ (Sorry if this was discussed in another room I'm not a member of)

  14. MattJ

    selurvedu: hard to say much with no experience of it, and it's not currently open-source, so... 🙂

  15. MattJ

    Ultimately, there are already XMPP<>Matrix bridges, so it doesn't change much for the network

  16. MattJ

    Unless it's significantly better, which would be great

  17. selurvedu

    Yes, we have to wait until they release it to the public, but an announcement is intriguing still :)

  18. selurvedu

    MattJ, bridge is one thing, you need to host a separate service or use a third-party one, and bridges lack some features and aren't always reliable. Knowing how ejabberd performs in general, I presume it would be a better choice.

  19. selurvedu

    MattJ, bridge is one thing, you need to host a separate service or use a third-party one, and bridges lack some features and aren't always reliable, and a sever with multiple protocols is another. It won't let the XMPP users from other servers reach Matrix, unlike a bridge. Knowing how ejabberd performs in general, I presume it would be a better choice.

  20. nicoco_

    There is no theoretical reason for bridges to not be reliable, is there?

  21. selurvedu

    nicoco_, have you ever used bifrost at matrix.org? I guess you haven't :)

  22. selurvedu

    The aria-net.org fork is quite stable, but it is not the "officially recommended" one, and it did not exist from the beginning.

  23. nicoco_

    I’m using aria-net for some matrix rooms and it’s fine - but there have been some downtime indeed. Also it seems to fire a lot of useless rename events, but I suspect this is some of my xmpp clients not respecting the bookmarked nickname and doing their own thing.

  24. selurvedu

    nah, it's the bridge, I think

  25. selurvedu

    it doesn't handle the join with a nickname as a single event, so first you join with nick==jid localpart, then you get renamed

  26. selurvedu

    Even without the renames I got complaints from Matrix users about constant join/parts

  27. MSavoritias (fae,ve)

    selurvedu: Ejabberd is also a bridge so no difference there with aria net

  28. MSavoritias (fae,ve)

    But its ontegrated so thats a plus of course

  29. selurvedu

    They expect the users to join and remain there, like e.g. irc bouncers work, but bifrost is not a bouncer, and join-parts are just how XMPP works

  30. selurvedu

    I wonder if ejabberd would use permanent presence for matrix rooms to avoid that, or keep this behaviour

  31. selurvedu

    MSavoritias (fae,ve), well, yes, but it's not a separate program, it's still the same server (at least this is what I expect knowing it supports MQTT), and, unlike bridges, it won't let non-local jids to reach the matrix network.

  32. selurvedu

    In the rest it mostly the same, yes.