BIND 10 #3112: boost::optional failure on Debian i686

[BIND 10 Development]

#3112: boost::optional failure on Debian i686
-------------------------------------+-------------------------------------
            Reporter:  muks          |                        Owner:
                Type:  defect        |  UnAssigned
            Priority:  high          |                       Status:
           Component:  libdns++      |  reviewing
            Keywords:                |                    Milestone:
           Sensitive:  0             |  Sprint-20130917
         Sub-Project:  DNS           |                   Resolution:
Estimated Difficulty:  3             |                 CVSS Scoring:
         Total Hours:  0             |              Defect Severity:  N/A
                                     |  Feature Depending on Ticket:
                                     |          Add Hours to Ticket:  0
                                     |                    Internal?:  0
-------------------------------------+-------------------------------------

Comment (by vorner):

 Replying to [comment:7 kean]:
 > Just FYI this only causes a build failure because of -Werror. No sane
 compiler would make aliasing issues an error and gcc/g++ certainly
 doesn't. It's hackinh as hell but if there is no other way of avoiding the
 warning there is always the distasteful fallback option of using GCC's
 pragmas's to stifle the warning. Or to turn off -Werror for this file.
 Last option though, it is obviously better to eliminate the aliased
 pointer in the first place.

 Not directly related to the topic of the ticket, but just for the record,
 I'm strongly opposed to silencing this warning. While warning about
 assignment inside conditional or unused variable is harmless, this one
 should really read as „I may (and probably will) generate bogus code“. The
 compiler may assume that pointers that are not allowed to alias are
 distinct and base some optimisations on it (and it does). If this is not
 true, the resulting code can break quite badly.

 There's option that turns off not only the warning, but also this kind of
 optimisations.

-- 
Ticket URL: <http://bind10.isc.org/ticket/3112#comment:11>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development