Conflicts:
	ChangeLog
	src/bin/dhcp4/config_parser.cc
	src/bin/dhcp6/config_parser.cc
	src/lib/dhcpsrv/lease_mgr.cc
	src/lib/dhcpsrv/lease_mgr.h
	src/lib/dhcpsrv/memfile_lease_mgr.cc
	src/lib/dhcpsrv/subnet.cc
	src/lib/dhcpsrv/tests/cfgmgr_unittest.cc

This uses the workaround suggested by Jinmei in:
https://lists.isc.org/pipermail/bind10-dev/2013-January/004192.html

It is less optimal than the normal code, but we have no choice if we
have to support Debian 6 stable which currently ships with a broken
Boost.

Rather, it does a getValue on all top-level elements and puts them in the map to return
This changes behaviour slightly, but it should reflect the actual use better than the old ones (even in existing cases)

Conflicts:
	src/lib/dhcpsrv/cfgmgr.cc
	src/lib/dhcpsrv/cfgmgr.h

totally irrelevant to the subject of the task, but since there was
no test for this method before, I'm trying to make it a little better
at this opportunity.

they were accidentally (incorrectly) accepted due to the lenience of
string version of Name constructor but should have been rejected (and
fixed).

this will pass on completion of this task.

The updated description specifies that the change was in the DHCP code.
Also, the updated entry points out that new code is used internally by
the DHCP server. Okayed on jabber.

Conflicts:
	src/bin/dhcp4/config_parser.cc
	src/bin/dhcp6/config_parser.cc

Conflicts:
	src/lib/dhcpsrv/Makefile.am

it's probably more intuitive than False.