fixes potential cache key oob write

previously a CNAME RR could be merged as a queried type RR, leading to failed cache lookups as ‘expired’

the cache_peek() api was reworked to return an error code instead, from this the caller can tell whether the record is present (but expired) or missing. this save a secondary CNAME lookup in case the original record was just expired

prevents pointless caching of TTL={0,1} records

previously the NS address list was wiped out, this prevents them from being added in the first place

some servers break qname m12n by sending REFUSED, this accepts such answer and requeries with m12n turned off 

the only condition of caching an answer is that it resolves a query and is not already cached
(ok, it must be negative at this point…)

if an authoritative answer comes and the server responds correctly, but appends out-of-bailiwick NS records, ignore them but resolve the query

this fixes a bug when storing NODATA without SOA,NS or any other ‘authoritative answer’ hint

since the zone cut is looked up after layers now, it
is pointless to search for minimized names from cache,
since this is the same operation as searching for a zone cut

since the next CNAME target query was added on top, it replaced
the original ‘current query’, this caused recaching of an answer
if it came from cache (as the query flags have changed)

this prepares cache/txn structures to hold API as well, so we can get
rid of the global api

this fixes a regression when a final name has a dependency with a short
TTL in the search path, but has longer TTL itself
in another words - if there's a CDN in the middle and the name doesn't
exist, answer it from the cache without requerying the CDN

This reverts commit 3d1ee641.

This reverts commit ee2cce4d.

previously only queried names were resolved from cache, this meant that if the target name was present in cache but a server on the search path dropped from it, it would refetch it - this is a problem when a loadbalancer with very short TTL was in search path

the resolution has to start from the root for zonecut resolution, otherwise it might stall resolution if the only good NS drops out of cache and only the bad remain