BIND 10 #533: additional regression in msgq with kqueue

[BIND 10 Development]

#533: additional regression in msgq with kqueue
-------------------------------------+-------------------------------------
                 Reporter:  jinmei   |                Owner:  UnAssigned
                     Type:  defect   |               Status:  reviewing
                 Priority:           |            Milestone:  feature
  critical                           |  backlog item
                Component:  msgq     |           Resolution:
                 Keywords:           |            Sensitive:  0
Estimated Number of Hours:  0.0      |  Add Hours to Ticket:  0
                Billable?:  1        |          Total Hours:  0
                Internal?:  0        |
-------------------------------------+-------------------------------------

Comment (by vorner):

 This fails on linux:

 {{{
 Running test: msgq_test.py
 Traceback (most recent call last):
   File "/home/vorner/work/bind10/src/bin/msgq/tests/msgq_test.py", line 1,
 in <module>
     from msgq import SubscriptionManager, MsgQ
   File "/home/vorner/work/bind10/src/bin/msgq/msgq.py", line 92, in
 <module>
     class MsgQ:
   File "/home/vorner/work/bind10/src/bin/msgq/msgq.py", line 139, in MsgQ
     def add_kqueue_socket(self, socket,
 filter_type=select.KQ_FILTER_READ):
 AttributeError: 'module' object has no attribute 'KQ_FILTER_READ'
 }}}

 That is the reason why I didn't use the default parameter, but the boolean
 flag. If it is only in the code, the constant is not needed until it is
 used, but this way it is needed at the load of the module.

 Maybe I got it wrong how the events in kqueue work, but if there's an
 event with the same id (fd), isn't the older one completely replaced by
 the new one? By the patch you sent, it seems not.

 But I have no MacOS to actually review it :-(.

-- 
Ticket URL: <http://bind10.isc.org/ticket/533#comment:3>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development