Be far, far noisier about 'generated file; do not edit' to keep the important message from disappearing between copyright notices (say, in rdatastruct.h)

edns1 I-D has expired, no one knows what edns level 1 will mean, and it may end up not
implying support for GLOBAL16.  We are still liberal in what we accept; GLOBAL16
is accepted in packets we receive.

dns_[de]compress_setmethods instead of separately in every caller

macro

so users can easily get (for example) SOA list from all authorative servers
with 'dig +short +nssearch isc.org soa' rather than stringing dozens of
options together.  (Rather these options remain or are removed in the
release version is still subject to discussion.)

don't access freed ftcx when validation canceled

so that we can cope with parent/child NXT pairs.  DNSSEC sucks.

Changes to conform to standard coding style.
Internally reorder batch line options to properly handle @ options in
batch files.

Split dns_log_init() into dns_log_init() which sets up module and category names, and dns_log_setcontext() which sets the logging context.  Call isc_log_setcontext(), dns_log_init(), and dns_log_setcontext().

Add isc_log_setcontext() to set the isc_lctx global (ala dns_lctx) and add module names for isc/socket

Fix bug in batch processing in host.
Add fixed origin code.

zone to be parsed and then ignored). Currently must
come after the 'type' clause.

support optional ports in masters and also-notify
statements:

	masters [ port xxx ] { y.y.y.y [ port zzz ] ; }

removed some unused arguments.

stop issuing some warnings about some features that
were not implemented, but now are.

Add support for ndots (from resolv.conf and command line) to dig and host
Add usage information
Add support for retrying UDP sends, with command line options to set
how many sends are allowed to dig and host
Add multiple option t dig, like -C on host
Add options in dig to enable and disable display of comment lines and
individual sections

files as root

on the real stack.

headers and adding it to source files that need it.

headers and adding it to source files that need it.

                        than continuing to itemize every header which changed,
                        this changelog entry just notes that if a header file
                        did not need another header file that it was including
                        in order to provide its advertized functionality, the
                        inclusion of the other header file was removed.  See
                        util/check-includes for how this was tested.

Brian should look at this section of code, since I believe it will leak
a reference to 'keyset'

                        ISC_LANG_ENDDECLS to header files that had function
                        prototypes, and removed it from those that did not.