BIND 10 #905: TSIG: complete python library update

[BIND 10 Development]

#905: TSIG: complete python library update
-------------------------------------+-------------------------------------
                   Reporter:         |                 Owner:  jinmei
  jinmei                             |                Status:  reviewing
                       Type:         |             Milestone:
  enhancement                        |  Sprint-20110517
                   Priority:  major  |            Resolution:
                  Component:         |             Sensitive:  0
  DNSPacket API                      |           Sub-Project:  DNS
                   Keywords:         |  Estimated Difficulty:  2.0
            Defect Severity:  N/A    |           Total Hours:  0
Feature Depending on Ticket:         |
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------

Comment (by jinmei):

 Replying to [comment:11 stephen]:
 > > (We don't have/need a python binding for the copy constructor, but
 anyway) you mean documentation for the python lib embedded in the C++
 code, right? If so, fair enough...
 > Actually I meant commenting the calls to PyArg_ParseTuple() in
 TSIGError_init() - what constructor call do they correspond to?

 Ah...okay.  Updated the comment.

 > > and I've decided to use this opportunity to address a long standing
 issue: reducing the overhead of providing the python doc.
 > ... but I'm pleased to see that it has triggered a documentation
 improvement elsewhere.

 I had a feeling that I may have misunderstood you, but due to the time
 difference I couldn't check that timely...anyway, this (half)
 automation has been needed, and it's a sooner-is-better thing, so this
 is probably a good misunderstanding.

 > > I suspect it's not always that simple due to possible circular
 dependency.
 > Nothing ever is simple :-)  But if that is the case, you would probably
 have a problem with inclusion order anyway.

 In fact, that's one major motivation of my proposed change.  As we
 extend libdns++ and pydnspp, it will be more probable that we actually
 have this trouble.

 > > But, to be clear, if that's your strong preference, I don't
 necessarily object and am okay with moving forward with it as long as we
 separate header files and avoid 'using namespace' before including header
 files.
 > I don't have a preference, they were just some suggestions for tackling
 the problem; use whatever you thing best.  But having said that, I do
 think the suggestion of using scope-limited "using namespace" declarations
 is the most viable.

 Okay, I'm not arguing separating .cc files are *the best*, but for now
 I'll keep the current style.  If someone finds a stronger need for
 re-combining them with limited scope using directives, that's their
 decision.

 > Changes are OK, please merge.

 Okay, thanks, merged.  Closing ticket.

-- 
Ticket URL: <https://bind10.isc.org/ticket/905#comment:12>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development