GitLab

Conflicts:
	src/lib/datasrc/memory_datasrc.cc

addRRsig does not change its arguments, so there is no reason why
they cannont be declared "const Xxxx&", thereby saving the overhead
of a copy when the method is called.

... instead of throwing an exception.

notice by outside person

CNAME + non-NSEC is about to coexist.

cases.

NXRRSET (including empty) for any-query was already supported and tested.

to add in the main code for them, just confirming it also works for these
cases.

case.

to test expected list of RRsets.  This should be generally better as
explained for the same update to findTest().

this makes the test code shorter, and perfomrs more flexible than pointer
comparison.  this is not necessarily crucial for this branch, but will
eventually be necessary when we update the internal representation of in-memory
data.  at that point returned RRset by find() will be different from the one
installed by add() wrt pointer comprison.

in contextCheck.  passing reference to a shared pointer doesn't make sense
(and IMO could be rather confusing for a quite little benefit) when we
often dereference it in the function anyway.  in this case we can simply
use a normal reference as the ownership isn't transferred.

g++ somehow seems to require it to compile with toupper.  updated comments
about it.