BIND 10 #1470: address trivial fixes from francis' mail

[BIND 10 Development]

#1470: address trivial fixes from francis' mail
-------------------------------------+-------------------------------------
                   Reporter:  jelte  |                 Owner:  jelte
                       Type:         |                Status:  reviewing
  defect                             |             Milestone:
                   Priority:  major  |  Sprint-20111220
                  Component:         |            Resolution:
  Unclassified                       |             Sensitive:  0
                   Keywords:         |           Sub-Project:  DNS
            Defect Severity:  N/A    |  Estimated Difficulty:  5
Feature Depending on Ticket:  none   |           Total Hours:  0
        Add Hours to Ticket:  0      |
                  Internal?:  0      |
-------------------------------------+-------------------------------------
Changes (by stephen):

 * owner:  stephen => jelte


Comment:

 > Do you know if this had a ticket too?
 On looking, it did - ticket #866.  I will close that this when this one is
 closed.

 > (btw, should we add a comment to the added test that this is what it
 tests?)
 No, none of the other tests are commented.  Also, it's not really to check
 that - I had assumed that it was a Windows only bug had been relying on
 Francis to test it.  (However, when I discovered #866, I rebuilt the code
 - both with and without the change - with _GLIBCXX_DEBUG defined: the bug
 appears with the old version of data.cc and goes away with the new
 version.)

 The test itself is more to check that the order of the data in the maps
 does not matter - all other tests check removing "b, c" from "a, b, c"
 gives "a".  This check tests that removing "c, b" from "a, b, c" still
 gives "a".

 BTW, on checking after your comment I realised that I had only updated the
 test of the non-const version of removeIdentical().  I've updated the
 "const" version as well.

 > Are there tickets for this yet?
 I don't know.  There are TODOs all over the place in BIND 10 (on the
 latest version of master, I counted 154 in .cc files, 62 in .h files and
 49 in .py files) so I'm not certain why this one should be special.

 > I'd personally prefer exceptions that would result in a LOG_ERROR rather
 than asserts that result in abort()...
 I had a similar conversation with Jinmei on another ticket.  I think he
 wanted to raise this at the F2F.

 One more change to check please - no !ChangeLog entry is planned as these
 are fixes invisible to the user.

-- 
Ticket URL: <http://bind10.isc.org/ticket/1470#comment:8>
BIND 10 Development <http://bind10.isc.org>
BIND 10 Development