XSF Discussion - 2017-02-27


  1. jonasw

    will there be anything interesting going on XSF-wise on the Chemitzer Linux Tage?

  2. Ge0rG

    Oh, it's that time again. I don't think there are any plans, but I'm going there

  3. jonasw

    I’ll be there too, I think. Not sure which day(s) yet, gotta check the schedule of the event and my own :)

  4. Flow

    Any BOSH experts in here who can/want comment on https://github.com/igniterealtime/jbosh/pull/4 ?

  5. Zash

    jonasw: What where when?

  6. Flow

    zash: CLT in Karl-Marx-Stadt, err, chemnitz: https://chemnitzer.linux-tage.de/2017/de

  7. Flow

    Ge0rG, daniel and I where there last year

  8. mathieui

    §W 5

  9. MattJ

    Flow, the logic failure is that the client doesn't need to wait for an acknowledgement before it can send more requests (putting aside the normal BOSH rules about multiple open requests)

  10. dwd

    Flow, What PR#4 is saying doesn't immediately seem wrong - that one could use HTTP responses as acks - but the code doesn't appear to do this, and also what MattJ says.

  11. Flow

    Thanks MattJ, dwd. I think OPs main problem is using BOSH with only one (processing) thread. That will always cause a delay in one of the directions (if I'm not mistaken).

  12. Zash

    Blocking HTTP requests?

  13. daniel

    Guus: re that openfire connection issue. Is there a chance that the the upgrade somehow caused sasl mechanisms to vanish

  14. daniel

    Let's say sha1 disappearing for example

  15. Zash

    http://download.igniterealtime.org/openfire/docs/latest/changelog.html

  16. Flow

    Zash: blocking http requests?

  17. Zash

    daniel: some SCRAM changes there, maybe

  18. Zash

    Flow: Send HTTP request, wait for response before continuing with processing, have a bad time.

  19. Flow

    well that's what I think is happening in case only one thread is used with jbosh

  20. Zash

    Oh it's a client side library?

  21. Zash

    The coffee, it does nothing :|

  22. Zash

    Hey, Guus or dwd, does Openfire still do DIGEST-MD5?

  23. Zash

    Given the recent rush to hate on SHA-1, I'm impressed that nobody cares that DIGEST-MD5 is still around.

  24. daniel

    Zash: well the 24 hours news cycle and the general alarmism applies to it security news as well

  25. Holger

    My university's server offers only PLAIN so I'm on the safe side.

  26. Link Mauve

    <3

  27. Flow

    PSA: Google announces today the accepted GSOC orgs

  28. Flow

    In 2 h 45 minutes

  29. dwd

    Zash, DIGEST-MD5's security state hasn't really changed; the biggest weakness remains that you can churn through a lot of MD5's each second, and DIGEST-MD5 only uses three per cycle.

  30. dwd

    Zash, The fact it uses MD5 is *almost* irrelevant.

  31. jonasw

    isn’t it with DIGEST-MD5 that, like with PLAIN, it is enough to listen in on the connection to be able to authenticate as that user later?

  32. jonasw

    (I haven’t looked into it; it has been deprecated so I didn’t bother implementing it)

  33. Zash glares at daniel

  34. Tobias

    not to forget the interop issues with digest-md5

  35. Zash

    dwd: I'm sure we'll receive bug reports about SCRAM-SHA-1 being terrible because SHA-1 is broken soon.

  36. dwd

    Tobias, Oh, there are lots of problems. But using MD5 isn't really one of them.

  37. dwd

    jonasw, No, it's not subject to replay.

  38. Tobias

    dwd, reply with "Patches welcome!" :)

  39. Zash

    What about active MITM?

  40. dwd

    Zash, No channel binding, so yeah, an active MITM can work.

  41. dwd

    Zash, But couldn't replay, still.

  42. Zash

    dwd: I remember there being issues with SCRAM if you could get a client to try to auth with you.

  43. dwd

    Zash, Only in as much as you can potentially brute-force the SHA-1 and extract the plaintext equiv in a reasonable timeframe these days.

  44. Flow

    narf, google doesn't mention 17:00 UTC any more

  45. daniel

    Flow: the time line still says 1600Z

  46. Flow

    daniel: here → https://summerofcode.withgoogle.com/how-it-works/ ?

  47. jonasw

    How it works: 1. we freeze your browser because you’re not using chromium (jk)

  48. daniel

    Flow, https://developers.google.com/open-source/gsoc/timeline

  49. Link Mauve

    jonasw, loaded quite fast here on Firefox nightly.

  50. jonasw

    Link Mauve: it behaves oddly when XHR is forbidden :)

  51. jonasw

    (it blocks instead of reacting on the error O_o)

  52. Link Mauve

    Weird.

  53. jonasw

    yes.

  54. Kev

    https://summerofcode.withgoogle.com/organizations/6327289865306112/

  55. intosi

    \o/

  56. dwd

    \o/

  57. daniel

    Awesome

  58. dwd

    Tobias, Want to tweet something from @xmpp?

  59. dwd

    (Assuming Tobias is an Approved Tweeter)

  60. Kev

    I believe he is, yes. I need to get those credentials from Bear at some point.

  61. Tobias

    sure

  62. Tobias

    although i don't have any cerdentials

  63. intosi

    We won't hold it against you.

  64. arc

    Kev: did you get the email yet?

  65. Kev

    arc: Nope.

  66. Kev

    I just went straight to the source :)

  67. arc

    and?

  68. Kev

    Kev 17:01 https://summerofcode.withgoogle.com/organizations/6327289865306112/

  69. arc

    nice

  70. arc

    copyleft games is in too

  71. arc

    i'll cross-link xsf on our ideas page for related organizations

  72. dwd

    I got the email, came here, and Kev had already posted.

  73. arc

    Kev: you should link xmpp related orgs on the ideas page. it helps steer students in the right direction while they're looking ;-)

  74. kaboom

    are there any restrictions which project/persons can become a gsoc mentor for xsf?

  75. Kev

    arc: If there's stuff you think I should do, can yo umail please?

  76. Kev

    I'm in the office until Wed night, so my mind is highly lossy at the moment.

  77. Kev

    And now to dinner...

  78. Ge0rG

    Wow, the white house forbids staff to use Signal. http://www.politico.com/story/2017/02/sean-spicer-targets-own-staff-in-leak-crackdown-235413

  79. moparisthebest

    Ge0rG, as usual the media gets it wrong, the headline anyway, isn't it they were forbidden from leaking private info, meh

  80. moparisthebest

    that's how I read it anyhow, either way no mention of xmpp/conversations/omemo in there and I don't know whether to be happy or sad about it lol

  81. mimi89999

    I heart "fake news" more often in the last several months/year than during my entire life until the election crisis.

  82. mathieui

    because "fake news" were just called "lies" before then

  83. Link Mauve

    Or conspiracy theories.

  84. Link Mauve

    Or “conspiracy theories”.