- ChangeLog updated
 - Current reconfiguration limitations described in User's Guide
 - Dhcp{4,6}ParserTest.reconfigureRemoveSubnet added
 - Subnet::getNextID -> generateNextID
 - Various comments improved

The obsolete usage of AC_OUTPUT was actually causing config.status
to be executed multiple times, which is a big waste of time. Changed
to use the correct AC_CONFIG_FILES and AC_CONFIG_COMMANDS macros
instead.

Conflicts:
	src/bin/d2/d2_messages.mes

Fixed two missed minor cleanups from the initial review.

Changes are largely clean up and commentary.

Some files were using #include <full/path/to/header> syntax which is
not desirable as the directory may move about in the source tree, and
all of the files it was including are local to perfdhcp so they should
be using #include "filename.h".

Recent changes to the perfdhcp tool introduced a new breakage when
configuring with srcdir != objdir. These were introuced by the code
using #include <tests/tools/perfdhcp/SOME_FILE.h> but there were no
-I flags set to allow for including files from the top level.

Conflicts:
	ChangeLog
	doc/guide/bind10-guide.xml

 - flag => configuration parameter
 - echoClientId test uses ASSERTs rather than EXPECTs

Conflicts:
	ChangeLog

 - tests/methods descriptions improved
 - CfgMgrTest.echoClientId improved
 - Redundant checks removed from Dhcpv4SrvTest.discoverEchoClientId
 - BIND10 Guide clarified
 - ChangeLog added

Conflicts:
	ChangeLog

d2/nc_add.cc contained some inaccurate RFC references.

Added methods for constructing all three types of DNS update requests
required by d2::NameRemoveTransaction to complete the implementation of its
state machine.  Also refactored some unit test code into nc_test_utils.h
and .cc. Renamed request verification functions in nc_test_utils to match
the build request function names.

 - another patch contributed by dclink (thanks!)
 - BIND10 Guide updated
 - BSD/Sun interface detection cleaned up
 - MAC address checking improved
 - Developer's guide updated
 - Several whitespace cleanups

Both macros should be accepted, but CMSG_LEN appears not to work on
OpenBSD.

Rather than using terminal-specific hard-coded escape sequences that are
not guaranteed to work on all terminals, use the tput utility. Also use
the terminal defined standout mode rather than trying to set a specific
colour. The old way enabled bright yellow text which, while very visible
if your background is black, is all but invisible if your terminal
background is white.