BIND 10 #192: Data source hotspot cache

[BIND 10 Development]

#192: Data source hotspot cache
-------------------------+--------------------------------------------------
 Reporter:  each         |        Owner:  jinmei                                        
     Type:  enhancement  |       Status:  reviewing                                     
 Priority:  major        |    Milestone:  05. 3rd Incremental Release: Serious Secondary
Component:  b10-auth     |   Resolution:                                                
 Keywords:               |    Sensitive:  0                                             
-------------------------+--------------------------------------------------

Comment(by jinmei):

 On the !DataSrcMatch class:

 I've updated the code based on the agreement (or at least what I believe
 we agreeded on) in the review (r2298).

 A few notes, including a substantial one:

  - I guess there's a bug in the current implementation (not specific to
 this branch.  it should be the same for the current trunk version).  see
 DISABLED_initialUpdateWithNoMatch.
  - I've added many more tests and more detailed doxygen description.  I'd
 expect any new document to be as detailed as this one.  And, it helps us,
 too; I noticed the bug through this process.
  - and this is a substantial note: in the end, I'm feeling even
 !DataSrcMatchTest dropping RR type is leaky.  this interface cannot well
 explain test cases like !DataSrcMatchTest.updateWithWrongClass.  we could
 provide artificial restriction such as an exception to be thrown or simply
 ignoring RR class in update(), but it looks like counter intuitive and
 error prone.  To me, this suggests we should revisit the abstraction, and
 my updated suggestion is to simply go back to the original design of
 !NameMatch.  And we pass an RR class parameter to findClosestEnclosure()
 separately.  As we discussed, whether to include RRclass in
 Name/DataSrcMatch doesn't affect the hot spot caching itself, and, as a
 bonus, the diff will be smaller and review will be easier.

-- 
Ticket URL: <https://bind10.isc.org/ticket/192#comment:39>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development