BIND 10 #513: b10-auth hangs in submitting statistics

[BIND 10 Development]

#513: b10-auth hangs in submitting statistics
-------------------------------------+-------------------------------------
                 Reporter:  jinmei   |                Owner:  y-aharen
                     Type:  defect   |               Status:  reviewing
                 Priority:           |            Milestone:  A-Team-
  critical                           |  Sprint-20110126
                Component:           |           Resolution:
  b10-auth                           |            Sensitive:  0
                 Keywords:           |  Add Hours to Ticket:  0
Estimated Number of Hours:  0.0      |          Total Hours:  0
                Billable?:  1        |
                Internal?:  0        |
-------------------------------------+-------------------------------------

Comment (by y-aharen):

 The branch and changelog entry are OK. It is ready to merge into trunk.

 Replying to [comment:9 jinmei]:
 > Although I agree it should be rejected at the frontend of configuration
 > parser (so I fixed it), I also believe this is primarily a problem of
 > the interface of IntervalTimer::setupTimer().  It takes an unsigned
 > integer and naively passes to boost::posix_time::seconds(), which
 assumes
 > a signed long value.  We should at least use a fully compatible type
 > (simple 'long' would be best because we cannot even assume it's 32 bit),
 > and, if we want to reject negative timer intervals as part of the
 > IntervalTimer interface, we should explicitly check and reject them
 > in its setupTimer().  But I wouldn't go into that level within this
 ticket
 > anyway.  It would have to go to a separate ticket/task.
 I intended to make sure IntervalTimer::setupTimer() only accepts positive
 value or 0 by that, although boost::posix_time::seconds() is for generic
 purpose and accepts negative value. If it is a bad idea, I will create a
 ticket to address that.

 > For that matter, while I extended the IntervalTimer class, I noticed
 > the name of "setupTimer()" was redundant because the class name already
 > clarifies that it's somehow related to the timer.  As a hindsight it
 > should have been named "setup", and so I chose "cancel()" rather than
 > "cancelTimer()", which might be better in terms of consistency.
 > So, for consistency and conciseness, I'd suggest renaming "setupTimer()"
 > to "setup()" (or at least something not redundant).  It should be
 > a topic of separate ticket, though.
 I agree with you. I will create a ticket later.

-- 
Ticket URL: <http://bind10.isc.org/ticket/513#comment:11>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development