constify, typo, removing redundant space, etc.

so that failed constructor does not leak

that makes a number of tests unnecessary (as well as impossible)

since they can now be done through getRecords()

(to replace searchForRecords->getNextRecord combo)

Conflicts:
	src/lib/datasrc/database.h

to be consistent

The content changes were done in the source files previously.

When we iterate and find out that there are RRs in an RRset with
different TTL, we go for the lowest of them instead of throwing and give
a warning.

Conflicts:
	src/lib/dns/Makefile.am
	src/lib/util/python/gen_wiredata.py.in

While here also sort the configurations (no content change for that).

Regenerate the nroff file.

The query_acl now has some beginning docs here, but needs more.

the RRsets and class were reversed. This was the only
example in all .mes files that had the numbers out of order.

Okayed via jabber.

This fixes:

2011-08-15 14:37:22.991 DEBUG [b10-resolver.cache] CACHE_RRSET_INIT initializing RRset cache for IN RRsets of class 10000

to:

2011-08-16 11:01:50.899 DEBUG [b10-resolver.cache] CACHE_RRSET_INIT initializing RRset cache for 10000 RRsets of class IN

One thing to note:  If this was a production server, we would need
to consider deprecating this message ID and creating a new message ID
as use of the logged results would be broken, because problem is in
the logged brief explanation and not the full description.
I will send an email about this.

Conflicts:
	src/lib/datasrc/database.cc
	src/lib/datasrc/tests/database_unittest.cc

(missing boost::)

Reviewed on jabber.