And remove the now renamed/removed internal only zone_new_data_ready
command.

Regenerate nroff file too (which includes

No new changelog entry. This goes with changelog entry 402.

Somewhat discussed on jabber. Didn't get consensus on one sentence,
but that can always be clarified later.

Also removed the garbage '?' mark.  It appears I accidentally committed/pushed
the proposed entry to master...so this is just for cleanup.

using debug level of 40 for now.

This is for ticket #1102.

Also clarify the log description.

No changelog entry.

Discussed some on jabber.

this will fix a crash in asiolink tests on OpenBSD due to its deviant
behavior that could result in different configuration on the use of
boost threads in different .cc files.

This is because random() always returns 0 on OpenBSD 5.0, when
the RNG is initialized with srandom(0).

see http://bind10.isc.org/ticket/1784#comment:5

Unrelated cleanup.

They are equivalent, but this looks more consistent.

It is very similar to the findPreviousName, just that it wraps around
when no hash is found.

Naturally, the tests fail, the method is not implemented.

The one for providing the "normal" lookups is just slightly modified to
accept another SELECT statement. The SELECT statement has a small hack -
since the table does not contain a sigtype column (because it makes
little sense in the nsec3 namespace), it is replaced with column where
everything equals to 1, to position the rdata column to the proper
position.

Conflicts:
	ChangeLog

also make sure the internal (reusable) MessageRenderer is always used.

thanks to the introduction of DNSServiceBase we don't need it any more.

also introduce mock DNSService and use it to test the behavior, and to test
the resolver behavior doesn't change.

the essential part should be quite simple; most of the changes are
straightforward interface adjustments.
it also includes some editorial cleanups such as adding header-file guards
or style guideline fixes.

Only some part of portconfig need to be adjusted; others will still use
the derived class directly.

For the findPreviousNSEC3Hash.

Also, renamed the nsec_namespace_ variable to more correct
nsec3_namespace_.

This is for the newly introduced interface. It makes stuff compile, it
complained about pure virtual methods.