BIND 10 #2922: enhance group subscriber management in msgq

[BIND 10 Development]

#2922: enhance group subscriber management in msgq
-------------------------------------+-------------------------------------
            Reporter:  jinmei        |                        Owner:
                Type:  task          |  pselkirk
            Priority:  high          |                       Status:
           Component:  msgq          |  reviewing
            Keywords:                |                    Milestone:
           Sensitive:  0             |  Sprint-20130625
         Sub-Project:  Core          |                   Resolution:
Estimated Difficulty:  4             |                 CVSS Scoring:
         Total Hours:  0             |              Defect Severity:  N/A
                                     |  Feature Depending on Ticket:
                                     |  shared memory data source
                                     |          Add Hours to Ticket:  0
                                     |                    Internal?:  0
-------------------------------------+-------------------------------------
Changes (by vorner):

 * owner:  vorner => pselkirk


Comment:

 Hello

 Replying to [comment:14 pselkirk]:
 > That is confusing, and unfortunately it's throughout the whole test
 module. The simplest thing at this point is probably to rename the local
 variable in the one test case to 'b10_init', which seems to be what other
 test cases do when they need to refer to the init module.

 ACK, updated.

 > I'm well aware of how perl uses single vs double quotes, and AIUI python
 does the same. And where the string is a literal that will be read the
 same with either kind of quotes, it comes down to matters of taste and
 convention. But when I see what appears to be inconsistency, I tend to
 wonder what the intent was. In any case, it's fine.

 Just for information, I just tested, python treats both kinds of quotes
 the same:
 {{{
 >>> print("Hello\nWorld")
 Hello
 World
 >>> print('Hello\nWorld')
 Hello
 World
 }}}

 But that probably doesn't matter, as I already updated them.

 > So the result of this is that there will be no notifications from this
 'unsubscribe' command, because no state actually changed. That's a
 legitimate approach to take, and probably the safest thing to do when
 other users are listening to these notifications. (It's also been argued
 in similar contexts that the end result is that user is unsubscribed from
 this group, so it should return success.) In any case, I'll leave it to
 you. My main reason for bringing this up was that we should know what
 happens in this case.

 Well, the unsubscribe still succeeds. It's just that there's no
 notification. It didn't sound logical to broadcast a notification for
 action with no effect. Anyway, I believe it wouldn't matter much, if the
 receiver is written correctly (eg. without race conditions), it would have
 to be able to cope with such situations anyway.

-- 
Ticket URL: <http://bind10.isc.org/ticket/2922#comment:15>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development