fix conflict in changelog

Some slight grammar changes.
Mostly in comments, but some for api docs.
Changes a isc_throw output too.

Reviewed via jabber.

because it is a message ID, add a changelog entry too.
(reviewed via jabber)

and use the holder in the unit tests as well

Half-related change, if available, call mysql_library_end() at the end of the unit test run

Clean up docs/design/cc-protocol.txt

Conflicts:
	doc/design/cc-protocol.txt

it was a leftover from an intermediate attempt.

assuming the compiler optimizes the former (more understandable but
possibly less efficient) to the latter (less understandable but
efficient).

reviewed by jelte via jabber.

Changes some log message descriptions and some
function arguments and some output.

adjusted one unit test case based on the message change.

we previously relied on boost rejecting the padding chracter and throwing
an exception; since Boost 1.53 paddding characters are considered valid
and conerted to all-0 bits, so we now have to do it ourselves.  The detection
logic doesn't rely on the new behavior of Boost, so it should be backward
compatible.

transform_width in Boost 1.53 requires this version, so build will fail
without it.

Also don't catch specific exception errnos. Just handle them all
equally.

This is so that we can avoid merging RdataSets every single time which
can be inefficient.

We still only queue (without flushing) upto the point where a different
owner name is encountered. This is to limit the size of NodeRRsets used.