1. 28 Oct, 2015 1 commit
  2. 13 Oct, 2015 1 commit
  3. 28 Jul, 2015 1 commit
    • Dave Cridland's avatar
      Use Deque instead of Map, and add delay stamping · a6ea2b4b
      Dave Cridland authored
      After discussion with Jonny, switched the existing Map to a Deque, and in
      response to Tom's suggestion, added XEP-0203 (and legacy) delay stamping.
      
      Tested using Gajim and Swift; testing the delay stamping was done by disabling
      the acknowledgement handling to force retransmission.
      a6ea2b4b
  4. 27 Jul, 2015 1 commit
  5. 24 Jul, 2015 1 commit
    • JonnyHeavey's avatar
      Initial xep-0198 message ack implementation · ea56ccdd
      JonnyHeavey authored
      This patch adds a basic, non-resumable, XEP-0198 acking solution. It will
      redirect "lost" messages to offline storage, but does not attempt to suspend
      the session, so will not allow any resumptions.
      ea56ccdd
  6. 17 Apr, 2015 1 commit
  7. 09 Apr, 2015 1 commit
    • Dave Cridland's avatar
      Change OF-888 fix to be based on error · 5115f9b8
      Dave Cridland authored
      OF-888 is presumed to be a recursion due to an attempt to bounce a bounce, in
      other words it's failing to handle a double-bounce.
      
      It appears to be caused when routing a stanza from a MUC fails, and for some
      reason I don't yet understand, the routing of the bounce to the originating
      MUC source also fails.
      
      The test for the double bounce is only present for the IQ case, but this fails
      in any case because it tests only for the IQ case, and uses a test which checks
      the IQ's symbolic type; this is not actually set by Packet.setError(), so does
      not trigger.
      
      Tom's fix inserted a sentinel into the original failing stanza, however a new
      stanza is created for the bounce, which will not contain the sentinel; therefore
      the fix will not protect from a recursion.
      
      Therefore this patch:
      
      1) Removes Tom's fix for OF-888.
      
      2) Tests for the message and presence error types.
      
      3) Tests for the presence of a stanza error (though this is a warning).
      
      4) Explicitly sets the stanza error type.
      5115f9b8
  8. 06 Mar, 2015 2 commits
  9. 11 Jan, 2015 2 commits
  10. 23 Dec, 2014 1 commit
  11. 05 Aug, 2014 1 commit
    • Dave Cridland's avatar
      More S2S fixes · dc21027b
      Dave Cridland authored
      Kim 'Zash' Alvefur commented that an empty authzid in EXTERNAL wasn't working.
      
      This patch adds this handling, and also changes authorization checks from a
      domain.contains() to a domain.equals().
      dc21027b
  12. 04 Aug, 2014 1 commit
  13. 17 Jun, 2014 3 commits
    • Dave Cridland's avatar
      Add TLS/Authentication diagnostics · 5eb60080
      Dave Cridland authored
      This adds TLS information and Authentication choices to the server session
      details page.
      
      In doing so, it factors out a ServerSession interface, and LocalServerSession
      class.
      5eb60080
    • Dave Cridland's avatar
      Tidy up outgoing sessions · 1aadb51f
      Dave Cridland authored
      A few changes here:
      
      1) Don't recurse up the DNS tree. That's just wrong.
      
      2) Also, don't assume that a subdomain is handled by a parent domain's server.
      Still wrong.
      
      3) Check certificates post-connect using our new logic, and drop the session
      if they don't match and we're not meant to be doing dialback.
      
      4) Do use EXTERNAL if offered, even if we're using a self-signed certificate.
      There's no value in not doing so, it's a bizarre behaviour.
      
      5) Disable S2S Compression; it's currently not working. XPP reset seems to fail,
      so doing replacement of the input stream instead.
      
      6) Protect against a null features after TLS. Seems unlikely to happen, but
      still.
      1aadb51f
    • Dave Cridland's avatar
      Support dialback errors · 026c3f2f
      Dave Cridland authored
      See XEP-0220, Dialback Errors.
      
      This reduces disconnect in the case of piggybacking errors, and provides better
      diagnostics.
      026c3f2f
  14. 07 Jun, 2014 1 commit
  15. 05 Jun, 2014 1 commit
  16. 02 Jun, 2014 1 commit
  17. 07 May, 2014 1 commit
  18. 06 May, 2014 2 commits
  19. 05 May, 2014 1 commit
  20. 26 Apr, 2014 2 commits
  21. 25 Apr, 2014 1 commit
  22. 09 Apr, 2014 1 commit
  23. 05 Apr, 2014 1 commit
  24. 24 Feb, 2014 1 commit
  25. 05 Feb, 2014 1 commit
  26. 30 Jan, 2014 1 commit
  27. 20 Sep, 2013 1 commit
  28. 21 May, 2013 1 commit
  29. 04 Nov, 2012 1 commit
  30. 29 Sep, 2012 1 commit
  31. 01 Jan, 2012 1 commit
  32. 24 Sep, 2011 1 commit
    • guus's avatar
      OF-443: No longer use the dialback namespace definition as a trigger to start... · 55d8c08f
      guus authored
      OF-443: No longer use the dialback namespace definition as a trigger to start dialback (but use the db:verify / db:result elements instead). When the reported version of the stream is absent (prevents compatibility issues with openfire before version 3.7.1) or smaller than one, nothing should be sent on the stream before a db:verify / db:result stanza is sent (no stream features, for example). This mimics old behaviour. This way, dialback appears to be triggered by the namespace definition in the stream (even though in actuality it's triggered by a child element) which makes things backwards compatible.
      
      git-svn-id: http://svn.igniterealtime.org/svn/repos/openfire/trunk@12837 b35dd754-fafc-0310-a699-88a17e54d16e
      55d8c08f
  33. 10 Apr, 2011 1 commit
  34. 09 Apr, 2011 1 commit