BIND 10 #1186: libdhcp implementation - DHCPv6 part
BIND 10 Development
do-not-reply at isc.org
Thu Oct 13 14:34:46 UTC 2011
#1186: libdhcp implementation - DHCPv6 part
-------------------------------------+-------------------------------------
Reporter: tomek | Owner: stephen
Type: | Status: reviewing
enhancement | Milestone: Sprint-
Priority: major | DHCP-20111026
Component: dhcp | Resolution:
Keywords: | Sensitive: 0
Defect Severity: N/A | Sub-Project: DHCP
Feature Depending on Ticket: 878 | Estimated Difficulty: 0
Add Hours to Ticket: 0 | Total Hours: 0
Internal?: 0 |
-------------------------------------+-------------------------------------
Comment (by stephen):
Replying to [comment:11 tomek]:
>> '''src/lib/dhcp/tests/option6_addrlst_unittest.cc'''
>> All tests. Is there a need to declare opt1/2/3 as a pointer to
Option6AddrLst and delete them afterwards? As they are being initialized
at creation, why not declare them as objects of type Option6AddrLst? It
eliminates the need to delete them at the end of the test method.
(Alternatively, encapsulate them in a std::auto_ptr to ensure that they
are deleted even if the method returns prematurely.)
> In tests, I prefer to simplest approach possible. There is nothing that
could go wrong with new and delete (except running out of memory, but that
would indicate severe memory leak in code).
Not sure I follow the logic here. I would have thought that declaring
them as a local variable (as opposed to a pointer) would be easier.
However, your reply later on in the comment about using new/delete to test
the full lifecycle makes sense. I would suggest that we add
EXPECT_NO_THROW round the construction/deletion to be complete.
BTW, we still have declarations of the form:
{{{
Option6AddrLst * opt
}}}
The space before the "*" should be removed.
--
Ticket URL: <http://bind10.isc.org/ticket/1186#comment:17>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development
More information about the bind10-tickets
mailing list