this comes with some extra CPU cost, however:
- parsed packets are prealloc'd based on header information, so no
  reallocs
- constructed packets are alloc'd by steps
- packet construction / copy is  cheaper now
- doesn't matter when the mm backend is mempool anyway

the rationale for this change is the use in resolver and in smaller
installations

Static records substitution and DNS proxy modules.
(merge request #317, merged by hand)

 - valgrind slowed the previous version too much

Journal patches

See merge request !321

 - When reloading a server, there's a possibility of two readers reading
   the journal, but that should be fine. This is because the zone
structure is recreated during reloads, and so is the lock. Only events
can write to journal though, so there's no possibility of concurrent
read and write.

IXFR special cases handling (2)

Fixes #307, see the issue description and the discussion below for details. I went for option A for case (2) as no other people voted. ;-). Case (3) has been modified as proposed by @jvcelak.

See merge request !320

- Only for the first packet.
- Also restructured the code a bit.

- If the IXFR response contains no records or the first record is
  not SOA, consider it malformed and fail the transfer.
- AXFR-style IXFR both master side and slave side are checked in
  one test, removed the other.

this is to guarantee that there is at least one vacant entry between the
qtail and the qhead