jdev - 2022-06-20


  1. lovetox

    one of the more weird things is that server put requested mam messages into offline storage

  2. lovetox

    was there not a possibility to turn that off?

  3. lovetox

    something to clear the offline cache on connect or something

  4. jcbrand

    I have a question for client developers. How seriously do you take service discovery? Do you always check for support, or do you sometimes YOLO? I'm specifically thinking of carbons. So far, I didn't check for carbons support and just always sent the IQ. I know am thinking of checking for support first, but I wonder if that might cause breakage.

  5. jcbrand

    I have a question for client developers. How seriously do you take service discovery? Do you always check for support, or do you sometimes YOLO? I'm specifically thinking of carbons. So far, I didn't check for carbons support and just always sent the IQ. I now am thinking of checking for support first, but I wonder if that might cause breakage.

  6. jonas’

    checking for support has the advantage that you don't have to rely on the service sending an understandable error back in case it doesn't support it

  7. jonas’

    for server features, I usually do discovery; for client features, I don't care (as I can't know which client I'll be talking to anyway, thinking carbons and MAM)

  8. Zash

    Can't you cache disco#info and thus checking should normally be a check without roundtrip?

  9. jonas’

    yup

  10. jcbrand

    I'm more worried about some servers not advertising support but indeed supporting carbons. I guess that's a bit silly

  11. jcbrand

    And I then don't enable carbons and then get bug reports

  12. Zash

    Also, carbons being an iq is silly, sasl2+bind2 when?

  13. jcbrand

    So really what I want to avoid is bug reports 🙂

  14. Zash

    So you want it to just work, even if everything is horribly broken, so that nobody notices and reports the bugs?

  15. jcbrand

    If simply sending an IQ without checking for support results in less bug reports, then that's what I'd prefer

  16. Zash

    I am conflicted

  17. jcbrand

    I guess I'm a bit paranoid. I can just assume server developers are sensible and will properly advertise support

  18. jcbrand

    🙂

  19. jonas’

    :-)

  20. jonas’

    and if they're not its likely an easy fix

  21. jcbrand

    Yes, but to figure out that that's the problem takes time (which is what I'm trying to avoid)

  22. jonas’

    that should be easy with proper logging and diagnostics?

  23. Zash

    warn, no carbons advertised, server is totally outdated, abort()

  24. jcbrand

    yep, I'll warn

  25. Zash

    otherwise this seems like the kind of thing that results in more annoying bug reports sent to someone else

  26. lovetox

    its a fucking pandemic

  27. lovetox

    seems the whole world has MUCs added to their roster

  28. lovetox

    im the idot who writes code that assumes a roster contact is not a muc

  29. Zash

    contagious roster-MUCs?

  30. jcbrand

    > otherwise this seems like the kind of thing that results in more annoying bug reports sent to someone else yep ok.

  31. jcbrand

    I don't think I've every encountered someone having a MUC in the roster

  32. jcbrand

    I don't think I've ever encountered someone having a MUC in the roster

  33. jcbrand

    I'm not sure what Converse would do, I'm pretty sure it also assumes roster entries are contacts

  34. qy

    > its a fucking pandemic > seems the whole world has MUCs added to their roster 😅

  35. lovetox

    its funny that Discovery is XEP number 30

  36. lovetox

    they they added 20+ other XEPs until they reached the conclusion, its probably good if we know what we are talking to