XSF Discussion - 2022-02-04


  1. msavoritias

    Cant clients have new groups as MIX but still keep MUC as fallback? From what i got muc is "good enough". That's why people didnt move to Mix quickly. Plus slow development.

  2. moparisthebest

    Sure you could use both

  3. Menel

    Or the server has an automatic bridge module, so you always have a legacy access šŸ˜„

  4. moparisthebest

    Sure, but if your server is going to function as a muc bouncer for you... Then you've already just fixed all the problems with muc, congrats! :)

  5. moparisthebest

    There's even a prosody module that does this already

  6. Menel

    I meant a muc MUX bridge with auto creation of the muc if a mix is created

  7. msavoritias

    I dont know about this approach. Atsome point there has to be some cut off. Otherwise you are stuck in an endless migration

  8. Menel

    And thats why we stay with muc for a while longer.. There are always more important problems. And thats ok IMHO

  9. guus.der.kinderen

    jonasā€™: there's some interest in helping with XMPP.net 2.0 being expressed on the operators mailing list. Shall I forward them to your repo?

  10. jonasā€™

    guus.der.kinderen, replying out of band

  11. Guus

    XEP-0045 hints at sending IQ Pings to detect ghost users. This proves to be less straightforward as expected. What strategies around failure to respond (in a timely manner) would be appropriate?

  12. jonasā€™

    from a server side?

  13. Guus

    yeah.

  14. flow

    related: I was wondering if presence probes wouldn't be better than IQ Pings, as ideally you get an updated presence in return. But I haven't looked at the detailed working of presence probes in a while, so they may not even be an option

  15. Guus

    Stream Management is annoying, in this respect. Can a remote domain figure out what the maximum allowable delay in responding to a IQ request is for users on the local domain?

  16. jonasā€™

    timeouts are always wrong

  17. Guus

    Ghost detecting kind of depends on them, I think?

  18. flow

    Guus, I doubt that the information of the stream management timeout would be helpful

  19. jonasā€™

    you'll find the ghosts when you get type="error" in reply to presence/message broadcasts

  20. flow

    for one, jonas is somehow right that timeouts are always wrong

  21. flow

    and furthermore, there are more timeouts than the stream management timeout, for example the tcp state machince's timeout

  22. flow

    I always assumed that MUC ghost detection is mostly about detecting ghosts of users that are long gone (say > 60min), and not about users who have a "short" ghost period where they are disconnected (< 60min)

  23. flow

    So I'd probably trigger the ghost detection after 45 minutes of inactivity with a timeout of 15 minutes or so

  24. Guus

    That might be the pragmatic thing to do to fix this Flow, thanks.

  25. Link Mauve

    ā€œ18:56:34 Zash> I do wonder if maybe it's time to have a dedicated roster-of-sorts for MUC bookmarks.ā€, yay, fourth^Wfifth way of doing bookmarks!

  26. Link Mauve

    ā€œ15:35:45 Zash> Save us Link Mauve, you're our only hope!ā€, Iā€™ve stalled that for way too long, and Iā€™m sorry about that. v_v

  27. Link Mauve

    Now that I got my new Mesa driver out, Iā€™ll try to context switch back to XMPP a bit, itā€™ll be nice for me too.

  28. Link Mauve

    Please hold me off from doing any Nintendo hacking if you see me doing so!

  29. Link Mauve

    emus, ā€œIf you know a university or good place where to promote please go ahead: https://xmpp.org/community/gsoc-2022/ā€, the logo at the top looks like it was made for a light background, but I donā€™t know CSS well enough to know how to replace it with a lighter logo (especially for the ā€œXMPPā€ text).

  30. Andrzej

    Menel wrote: > I meant a muc MUX bridge with auto creation of the muc if a mix is created That is what our implemenetation does. It provides MUC room as a compatibility fallback on the same jid as the MIX channel has.

  31. emus

    Link Mauve: yes. its just translaprent. Im not good at working with svg etc. so its just a png lets not worry too much about it

  32. Link Mauve

    Right.

  33. Link Mauve

    Speaking of SVG and transparency, https://github.com/xsf/xmpp.org/pull/1067

  34. Link Mauve

    If you ever want me to look at a SVG, or even do it for you, feel free to ping me.

  35. Link Mauve

    I might not always be available for it, but itā€™s something I enjoy doing.

  36. emus

    Link Mauve: Many thanks. May I ask first to take a look at the clients page PR? That would have more impact I think

  37. Link Mauve

    Sure, I was on my way to do that, but found a yak with very long hair mid-way.

  38. emus

    Great - I dont understand the idiom but I assume you run into issues?

  39. Link Mauve

    Ah no, itā€™s just threading the metaphorical yak shaving. ^^

  40. jonasā€™

    emus, https://en.wiktionary.org/wiki/yak_shaving

  41. jonasā€™

    > Any apparently useless activity which, by allowing you to overcome intermediate difficulties, allows you to solve a larger problem.

  42. jonasā€™

    > A less useful activity done consciously or subconsciously to procrastinate about a larger but more useful task.

  43. jonasā€™

    also: http://www.catb.org/jargon/html/Y/yak-shaving.html

  44. emus

    Hah, ok. Well, its alright Link Mauve. Highly appreciated ā˜ŗ

  45. mjk

    Link Mauve: > Use a better logo for Linux It's happening!!11 > I traced simple geometry behind the existing logo and ran svgcleaner on it, in order to not increase its size too much. The "official" logo isn't exactly heavy either, it terms of bytes. Is it too shiny though? :) There's one with flat speculars too! http://www.home.unix-ag.org/simon/penguin/ Looks even better, imo

  46. Link Mauve

    Compare the size of any of these with my version. :p

  47. Link Mauve

    The ruby logo is much larger than the other ones on our page, maybe we could find an alternative one.

  48. Sam

    *troll* Nah, that sounds about right for anything involving Ruby

  49. Sam ducks

  50. Link Mauve

    :D

  51. Link Mauve

    Before optimising the logos though, compressing the CSS and JS would give a much higher gain of transfer size, and enabling caching on most of the resources would do even better for subsequent requests.

  52. mjk

    Link Mauve: oops, yeah, the non-flat ones are bloat. I went by the 90k one on wikipedia. Anyway, the 19k of the flat one seems perfectly reasonable (maybe optimizable too)

  53. Link Mauve

    mjk, my version is 3.7Ā KiB, or 1.6Ā KiB after brotli compression.

  54. Link Mauve

    And given the size we render it at, I think thatā€™s good enough.

  55. mjk

    but it loses important detail T^T

  56. mjk

    *Yak shaving intensifies*

  57. Sam

    actually, jokes aside, I haven't touched Ruby probably in a decade. Is there good xmpp support? What is the library most ruby folks choose?

  58. Zash

    Heh, those glorious days of RoR hype, when writing a blog engine was fun

  59. Link Mauve

    Last time I tried to update omnipresence, xmpp4r was completely dead and it was easier to RIIP than to revive it.

  60. Link Mauve

    (Omnipresence was a component for showing your presence info, avatar, PEP mood and such, on the web.)

  61. Zash

    And that Ruby XMPP server seems to have died after the author got hired by Github

  62. mjk should go back to suffering windows scripts for Dino portable

  63. emus

    > mjk escribiĆ³: > *Yak shaving intensifies* šŸ™ˆ the users want clients !!1! No is alright, take your time and most important is you keep up the fun!

  64. emus

    ahh its you mjk. got confusdd

  65. Sam

    Yah, looking through RubyGems I don't see anything that's been updated since 2018 at the latest and most XMPP things seem to have been abandoned around 2014. Oh well.

  66. emus

    RubyGems?

  67. Sam

    Their package repo

  68. jonasā€™

    is that retracing of the penguin even ok, trademark-law wise?

  69. Sam

    "Permission to use and/or modify this image is granted provided you acknowledge me lewing@isc.tamu.edu and The GIMP if someone asks."

  70. Sam

    As far as I know the status has never changed since Larry Ewing drew Tux the first time

  71. mjk

    > šŸ™ˆ the users want clients !!1! > No is alright, take your time and most important is you keep up the fun! > ahh its you mjk. got confusdd Yea, I'm not Georg, far from it

  72. jonasā€™

    Sam: that sounds like copyright stuff, not trademark stuff

  73. Sam

    It doesn't matter in this case. They said anyone could use it'

  74. emus

    http://opensource.googleblog.com/2022/02/Announcing Season of Docs 2022.html are actually experiences with this in XMPP?

  75. moparisthebest

    needs more url encoding...

  76. moparisthebest

    https://opensource.googleblog.com/2022/02/Announcing%20Season%20of%20Docs%202022.html

  77. Zash

    I do wonder if XEPs are within scope of that.

  78. emus

    moparisthebest: thanks

  79. Zash

    But also, all documentation of all open source projects are perpetually in dire need of help, so might be good to at least look into it more

  80. emus

    Zash: yes me too

  81. emus

    would there be volunteers to check this out. Even if I would have time I wonder if im the right person

  82. emus

    We were just discussing to tweet significant publications of the XEP work. wurstsalat things its a very good idea also showing XEP development is not dead. Proposal welcome. Maybe we can start with MUJI

  83. emus

    edhelas are you alright I use one image from here? https://mov.im/?node/pubsub.movim.eu/Movim/end-to-end-encryption-in-movim-omemo-is-finally-there-yudZPP in the newsletter?

  84. edhelas

    Yup

  85. emus

    Great

  86. edhelas

    You can buy the NFT if you want šŸ˜Œ

  87. emus

    šŸ˜ƒļø where should I transfer da money?

  88. emus

    I can trust you of course, right?.... right?

  89. Zash

    `<payment-required/>`

  90. edhelas

    On the XEPChain

  91. emus

    I'm buying XEPs only