BIND 10 #441: MemoryZone refactor

BIND 10 Development do-not-reply at isc.org
Tue Dec 21 16:04:21 UTC 2010


#441: MemoryZone refactor
------------------------------+---------------------------------------------
      Reporter:  vorner       |        Owner:  UnAssigned
          Type:  task         |       Status:  reviewing 
      Priority:  major        |    Milestone:            
     Component:  data source  |   Resolution:            
      Keywords:               |    Sensitive:  0         
Estimatedhours:  0.0          |        Hours:  0         
      Billable:  1            |   Totalhours:  0         
      Internal:  0            |  
------------------------------+---------------------------------------------

Comment(by jinmei):

 Replying to [comment:1 vorner]:

 > The find does handle delegation, because it was like 5 more lines of
 code and it seemed like comparable amount of work to adding TODO. It,
 however, returns SOA, not NS, because I think it is cleaner (anyway, both
 will be needed when the delegation is used), but it is agains the
 description in Zone::find. So I'd like to ask, is there any reason why we
 want to return NS instead of SOA (eg. which should be changed - MemoryZone
 or Zone's documentation)?
 >
 Although not looking into the code deeply, I'm not sure why we wanted to
 return SOA in the case of "delegation".  In this context "delegation" was
 intended to mean the following case:
  - we serve "example.com" zone
  - this zone has
 {{{
 child.example.com. NS ns.child.example.com.
 (with glue A/AAAA, but it's irrelevant to this discussion)
 }}}
  - we have a query for www.child.example.com.

 In this case the server is expected to return the NS (with glue).  That's
 why the find() is supposed to return the NS.  Actually, this is the way
 BIND 9 currently works and it simply follows that implementation (we don't
 necessarily have to do the same if there's a reason, but it seems
 reasonable to me).

 Maybe you have a difference scenario with the word "delegation" in your
 mind?

 > As this is part of greater work on the in-memory datasource, does it
 need a changelog?
 >
 I'd say it's up to you.  There's no fixed rule about whether to add a
 ChangeLog entry.  (In my understanding) it's left to developer's
 discretion, and as far as I can see we all have behaved reasonably.

-- 
Ticket URL: <http://bind10.isc.org/ticket/441#comment:3>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development


More information about the bind10-tickets mailing list