1. 16 Feb, 2011 1 commit
  2. 14 Feb, 2011 1 commit
  3. 09 Feb, 2011 1 commit
  4. 04 Feb, 2011 2 commits
  5. 03 Feb, 2011 1 commit
    • Stephen Morris's avatar
      [trac496] Changes made to address review comments · ab2c79a8
      Stephen Morris authored
      In particular:
      * Use Use asiolink abstractions instead of Boost::asio directly
      * Add cross-section checks (answer/question v authority)
      * Extended description of bailiwick
      * Methods added to isc::dns::Message and asiolink::IOAddress
      ab2c79a8
  6. 02 Feb, 2011 1 commit
    • Jelte Jansen's avatar
      [trac494] addressed review comments · 430e0e89
      Jelte Jansen authored
      Removed some dead code, added two unit tests, changed sendQuery() to the
      more descriptive resolve(), and removed a superfluous call to dlog()
      430e0e89
  7. 01 Feb, 2011 1 commit
  8. 31 Jan, 2011 2 commits
  9. 28 Jan, 2011 1 commit
  10. 27 Jan, 2011 2 commits
  11. 26 Jan, 2011 1 commit
    • Jelte Jansen's avatar
      [trac494] initial abstraction for runningquery results · 8a175757
      Jelte Jansen authored
      Added a abstract placeholder for callbacks in asio (for now); which has
      two subclasses; one takes a DNSServer* and calls resume() on it when the
      runningquery finished its lookup (or fails), the other one is provided a
      callback as defined in isc::nsas::ResolverInterface.
      
      These are more objects that are newed while running and delete
      themselves, and they may not even be necessary, but for now this makes
      it work while doing the least amount of changes in other parts.
      
      RunnigQuery now takes one of these objects instead of a DNSServer*, and
      the objects decide what should be done once it is ready.
      
      Things to do:
      - remove the temporary MyCallback class from resolver (perhaps move it
        to a unittest)
      - move ResolverInterface out of nsas, as well as its Callback thingy,
        and merge it with the new callback holders
      - general cleanup and tests (although, as with current resolver tests,
        need framework to fake outside query handling)
      8a175757
  12. 25 Jan, 2011 2 commits
    • Jelte Jansen's avatar
      changed port for listenAddress test · b9b7ef39
      Jelte Jansen authored
      The port used in the test was the same as the default port for the
      bind10 server (5300), so the test failed if bind10 was already running on the
      system. Changed it to 5321.
      b9b7ef39
    • Stephen Morris's avatar
      Update comment in response_classifier.h · 98d6a12c
      Stephen Morris authored
      As suggested in the review in ticket #487, an additional non-Doxygen
      comment was added to the Category enum to indicate that for error()
      to work, non-error codes should come before error codes.
      98d6a12c
  13. 24 Jan, 2011 3 commits
  14. 22 Jan, 2011 2 commits
  15. 21 Jan, 2011 6 commits
    • JINMEI Tatuya's avatar
      more $Id$ cleanups: · b20ba8d5
      JINMEI Tatuya authored
       - did it for non .cc/.h files
       - some files had 'already substitued' $Id$'s (which were probably not
         under keyword substitution with svn; just copied from some other
         file).  they were overlooked in the previous cleanups
       - there were lower cased $Id$, which were overlooked, too.
      b20ba8d5
    • JINMEI Tatuya's avatar
      cleanup: removed $Id$, which is useless now that we've migrated to git. · e3e59cf1
      JINMEI Tatuya authored
      briefly talked about it on jabber, no one opposed to it, Jeremy explicitly
      agreed.  changes are trivial themselves, so no explicit review.
      e3e59cf1
    • Stephen Morris's avatar
      First cut at code to classify a server response · 30838487
      Stephen Morris authored
      The code analyses the response for errors as well as determining whether
      it is an answer or a referral.
      30838487
    • Jelte Jansen's avatar
      use the three types of timeout · 8f23ba4a
      Jelte Jansen authored
      Note that client_timeout at this point has the same result as
      lookup_timeout.
      
      Changed self-deletion of RunningQuery a bit, as it
      cannot delete itselfs until we are sure there are not events coming in
      anymore. If we change the clientTimeout() callback to send an answer and
      not stop, we shall need to keep track of that fact (i.e. not send back
      an answer in the end, and update stop since it won't be called by
      clientTimeout() anymore
      
      When adding the asiolink tests for these features, added a few
      convenience functions for common code in those tests.
      8f23ba4a
    • Jelte Jansen's avatar
      enable configuration of new timeouts · b3ec8dc3
      Jelte Jansen authored
      renamed existing timeout to query_timeout, added client_timeout and
      lookup_timeout
      In the config spec, these are named timeout_XX instead of XX_timeout, so
      they will automatically be shown next to each other in ordered output
      b3ec8dc3
    • JINMEI Tatuya's avatar
      b39bfcee
  16. 20 Jan, 2011 2 commits
    • Scott Mann's avatar
      First task of trac483 · ce2fead0
      Scott Mann authored
      Added support for specifying root servers in configuration.
      ce2fead0
    • Jelte Jansen's avatar
      use default timeout/retry settings · ce648a93
      Jelte Jansen authored
      resolver does not update these if these weren't set, and does not look
      at the spec's defaults. This change makes it at least use something as a
      default (using spec's default has gone to ticket #518)
      ce648a93
  17. 19 Jan, 2011 2 commits
    • Jelte Jansen's avatar
      propagate timeout/retries values · 03fddc0d
      Jelte Jansen authored
      apparently these weren't passed on to the recursivequery class
      (anymore?), resulting in the use of hardcoded defaults -1 and 0. This
      one-line change fixes that, acked by Michal on jabber.
      03fddc0d
    • Jeremy C. Reed's avatar
      Document some configurations for the resolver. · 64e656b8
      Jeremy C. Reed authored
      For b10-resolver man page:  document  configurable settings and
      shutdown command.  This is not completed yet.
      
      Also regenerate man page from docbook.
      64e656b8
  18. 18 Jan, 2011 6 commits
    • JINMEI Tatuya's avatar
      Revert "maintenance commit" · 901c9eed
      JINMEI Tatuya authored
      This reverts commit a669db43.
      
      The reverted commit was introduced due to misoperation possibly with
      git-new-workdir, and actually reverted previous changes made on master.
      By reverting it we now recover these lost changes.
      901c9eed
    • JINMEI Tatuya's avatar
      maintenance commit · a669db43
      JINMEI Tatuya authored
      a669db43
    • Jelte Jansen's avatar
      question section wasn't copied · c772a141
      Jelte Jansen authored
      c772a141
    • Jelte Jansen's avatar
      enable forward mode again · ee96cc04
      Jelte Jansen authored
      only do recursive if forward_addresses (variable upstream_ in the code)
      is empty
      ee96cc04
    • Jelte Jansen's avatar
      ends0 fix and message argument rename · 6e4c7c07
      Jelte Jansen authored
      renamed the original MessaPtr message argument in the various process()
      calls to query_message, to better reflect the content
      Also fixed EDNS0 responses; fill in our own value (currently hardcoded
      to 4096, we should make this configurable at some point)
      added a new constant DEFAULT_MAX_EDNS0_UDPSIZE to lib/dns/message.h for
      that purpose
      6e4c7c07
    • Jelte Jansen's avatar
      use answer_message for final answer · e9c8b140
      Jelte Jansen authored
      use the answer_message Message object instead of parsing it out of the
      buffer again when constructing the final answer
      e9c8b140
  19. 15 Jan, 2011 2 commits
  20. 14 Jan, 2011 1 commit