before the algorithm was happy with root hints for all queries starting
at root, however they're often overloaded and result in timeouts
the updated code provides SBELT only for root NS query lookup and tries
to use cached information as much as possible

previously it was always overwritten with SBELT for root + root TA
doesn't have to be in cache (it's in trust store)

refs #33

This reverts commit 0f9e615b.

the selected address is now an array with selection, caller can then send the same query to multiple offered targets

refs #35

this could happen if the query contained a CNAME with AA=0, or missing mandatory DS in previous NS query

example:
local rr = pkt:section(kres.section.ANSWER)[1]
print(kres.rr2str(rr))

example:
local rr = pkt:section(kres.section.ANSWER)
for i = 1, #rr do
	if rr[i].type == kres.type.A then
		print(kres.dname2str(rr[i].owner))
		print(‘rdlen:’, #rr[i].rdata)
	end
end

this is useful if we want to prefetch or update data
in cache, it doesn’t affect the lookup of closest known
zone cut

sets maximum UDP/EDNS payload within <1220, 65535>
the default is max unfragmented UDP packet, but it’s
possible to set it higher if the network supports it to avoid
TCP retransmits

todo: active refresh

DS keys are injected into current set (unmanaged)
DNSKEY keys are in the managed set and their RFC5011 state is tracked

todo:
- implement timers and this AddTime/RemTime
- active refresh
- move to a separate module