BIND 10 #1883: define tp_hash for some basic isc.dns classes

[BIND 10 Development]

#1883: define tp_hash for some basic isc.dns classes
-------------------------------------+-------------------------------------
                   Reporter:         |                 Owner:  jinmei
  jinmei                             |                Status:  reviewing
                       Type:  task   |             Milestone:
                   Priority:  high   |  Sprint-20120717
                  Component:         |            Resolution:
  libdns++                           |             Sensitive:  0
                   Keywords:         |           Sub-Project:  DNS
            Defect Severity:  N/A    |  Estimated Difficulty:  5
Feature Depending on Ticket:         |           Total Hours:  0
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------

Comment (by jinmei):

 Replying to [comment:8 jelte]:

 > My compiler (gcc 4.6) is being a PITA and fails on that code;

 Ah, I was slightly afraid about that kind of failure.  I'm not sure if
 it's a compiler "bug", but in any case we need some workaround.

 > So far the only workaround I've come up with is to replace it by a
 variable;
 [...]
 > But both these 'solutions' are quite ugly...

 I've been trying to find a cleaner solution, but couldn't come up with
 anything obviously better either.

 One possible compromise is to use the wrapper only for the case of
 `Name` (and simplifies the converter function focusing on that
 specific case), as the implicit conversion is actually safe for the
 other cases.  But I guess the not-so-clear constraint of excluding -1
 can really bite us if and when we support tp_hash for some other
 classes, so I relatively prefer providing a generic wrapper from the
 beginning even if it's less clean.

 I've committed a slightly different version of your suggested change.
 I don't seem to have easy access to the version of g++ that triggers
 this warning, so I'm not sure if it really compiles thought.  I'm also
 not sure if hash_val_bits cannot be const.  If you also agree on
 providing a generic wrapper, could you confirm these?  If you'd rather
 prefer a minimalist approach and only convert the `Name` hash for now,
 that's okay for me too.

-- 
Ticket URL: <http://bind10.isc.org/ticket/1883#comment:9>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development