BIND 10 #1771: database datasource incorrectly rejects "on zonecut" glue

[BIND 10 Development]

#1771: database datasource incorrectly rejects "on zonecut" glue
-------------------------------------+-------------------------------------
                   Reporter:         |                 Owner:  jinmei
  jinmei                             |                Status:  reviewing
                       Type:         |             Milestone:
  defect                             |  Sprint-20120703
                   Priority:         |            Resolution:
  medium                             |             Sensitive:  0
                  Component:  data   |           Sub-Project:  DNS
  source                             |  Estimated Difficulty:  3
                   Keywords:         |           Total Hours:  0
            Defect Severity:  N/A    |
Feature Depending on Ticket:         |
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------

Comment (by jinmei):

 Replying to [comment:24 muks]:

 Thanks for the review.

 > This branch is fine. I agree the code is cleaner now, but should we add
 a ChangeLog entry for it because the new branch allows non-glue records
 now? Even with `GLUE_OK`, this isn't strictly glue for any of the accepted
 meanings:
 > {{{
 >      {"brokenns2.example.org.", "NS", "3600", "", "ns.example.com."},
 >      {"brokenns2.example.org.", "A", "3600", "", "192.0.2.1"},
 > }}}

 I've made the following changelog
 {{{
 451.    [bug]           muks, jinmei
         libdatasrc: the database-based data source now correctly returns
         glue records on (not under) a zone cut, such as in the case where
         the NS name of an NS record is identical to its owner name. (Note:
         libdatasrc itself doesn't judge what kind of record type can be a
         "glue"; it's the caller's responsibility.)
         (Trac #1771, git 483f1075942965f0340291e7ff7dae7806df22af)
 }}}

 I'd personally still argue the definition of glue is not fully fixed
 and that it's still open whether the above A record can be a glue or
 not, although it looks like I'm in the minority group within the BIND
 10 team and I know it's seemingly different from what's described in
 Peter Koch's draft.  In any event, the intent of GLUE_OK (which was
 borrowed from a BIND 9 API) was not to give this definition.  In
 retrospect it should have been named some like "IGNORE_ZONECUT", which
 would indicate what this option intends to be and wouldn't cause
 unnecessary confusion about the definition of "glue".

 I'll make a ticket to make this change and see whether others think it
 important enough to address.

 > Minor change: you may want to move the `is_origin` assignment into the
 `if (found.first)` block:

 > Please also repeat for `not_origin` in a previous hunk too. No need to
 get these reviewed. :)
 > You can go ahead and merge.

 Okay, changed these as suggested, merged, and closing.

 Please also delete the original trac1771 branch from the public repo.

-- 
Ticket URL: <https://bind10.isc.org/ticket/1771#comment:25>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development