BIND 10 #1259: framework for adding/deleting RR in datasource from xfrin

[BIND 10 Development]

#1259: framework for adding/deleting RR in datasource from xfrin
-------------------------------------+-------------------------------------
                   Reporter:         |                 Owner:  jinmei
  jinmei                             |                Status:  reviewing
                       Type:  task   |             Milestone:
                   Priority:         |  Sprint-20111011
  blocker                            |            Resolution:
                  Component:  xfrin  |             Sensitive:  0
                   Keywords:         |           Sub-Project:  DNS
            Defect Severity:  N/A    |  Estimated Difficulty:  0
Feature Depending on Ticket:         |           Total Hours:  0
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------

Comment (by jinmei):

 Replying to [comment:10 vorner]:

 > > One possibility would be making it a submodule of the datasrc package:
 > > isc.datasrc.diff (on second thought it would at least be better than a
 > > submodule of isc.dns (or pydnspp) as the diff module depends on the
 > > concept of datasrc).  But it's not a well-baked thought, just an idea.
 > > If you like it, please feel free to use it.  If you don't or simply am
 > > not sure, I'm okay with keeping the current module name.
 >
 > I don't think I like submodule of datasrc for two reasons.
 >
 > One is technical, as datasrc is C++ module and adding a python submodule
 is non-trivial (there are some rests of the original version of datasrc
 and we should get rid of them soon, the ugliness of the submoduling is one
 of the reasons).
 >
 > Another one is the fact that the C++ version of module don't have the
 submodule. That might be confusing, as we try to make the versions close
 to each other, someone might try to find the diffs there.

 Hmm, I don't think sharing .so and .py modules in the same package is
 in itself a problem (it's not uncommon for standard python libraries,
 for example).  But I missed the fact that datasrc.so is actually a
 single module containing multiple classes.  Having both a multi-class
 module (datasrc.so) and a single class one (diff.py) may be confusing.
 I also see datasrc.so might become a non-packaged module once we
 remove the old library.

 > So, can we wait until it gets shared with DDNS and then see how to name
 it?

 As a result, I'm okay with this.  (If this module is really only for
 xfrin, I'd rather move it to src/bin/xfrin, though).  But please add
 to the module pydoc (top of diff.cc) that we may revisit this
 module-name point so we don't forget.

 > > I have a couple of more comments about the revised code:
 > > - I'd add to the TTL log message something similar to BIND 9's log
 > >   "adjusting %lu -> %lu".  I'd also explain which TTL is overridden in
 > >   the detailed version of the log description.
 > > - Related to this point, I'll test the resulting TTL as part of
 > >   test_ttl (maybe for both cases of first TTL > second and first <
 > >   second to see it's just about the ordering, not the
 > >   smallest/largest)
 >
 > Added.

 I don't see an update in the branch.  Have you pushed it?

-- 
Ticket URL: <http://bind10.isc.org/ticket/1259#comment:11>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development