* plugin/forward: expose TLSConfig and error messages to public

* Add IsTLS() instead of TLSConfig()

* request.Match check Response bit as well

We should check this bit and reject them as invalid.

* Fix test

This allocates memory because of the append, just unroll the loop.
Also add benchmark.

Before:
goos: linux
goarch: amd64
pkg: github.com/coredns/coredns/plugin/cache
BenchmarkCacheResponse-4   	  100000	     11910 ns/op
BenchmarkMinMsgTTL-4       	 1000000	      1494 ns/op
PASS

After:
goos: linux
goarch: amd64
pkg: github.com/coredns/coredns/plugin/cache
BenchmarkCacheResponse-4   	  100000	     12016 ns/op
BenchmarkMinMsgTTL-4       	 2000000	       668 ns/op
PASS

These are not stored with newItem so we don't have to check them later.

Another thing we can test automatically, we sorta settled on using an
uppercase letter in in t.Log and t.Fatal calls.

Let's just check for this.

This reverts commit 6416925a.

Count and export number of panics we see.

Fixes #1294

* make pb

* Makefile for pb

Move the Makefile into the pb directory to remove some clutter from the
main Makefile.

* plugin/forward: expose few methods and attributes to public

* Update comments

* Remove Compress by default

Set Compress = true in Scrub only when the message doesn not fit the
advertized buffer. Doing compression is expensive, so try to avoid it.

Master vs this branch
pkg: github.com/coredns/coredns/plugin/cache
BenchmarkCacheResponse-2   	   50000	     24774 ns/op

pkg: github.com/coredns/coredns/plugin/cache
BenchmarkCacheResponse-2   	  100000	     21960 ns/op

* and make it compile

* plugin/etcd: small refactor

I think this function can be smaller.

* and make it compile

* plugin/cache: per server metrics

Use per server metrics in the cache plugin as well. This required
some plumbing changes. Also use request.Request more.

* fix cherry-pick

* plugin/dnssec: add per server metrics

final plugin.

Fixes #1696 #1492 #1189

* Move cache cap into handler so we can access the server label

* Remove cache-capacity from it entirely

Add server label to the metrics and update the README.

Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

Create separate function for zone check
Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

Add tests for zone A records
Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

Remove pointer from checkZoneForRecord func signature, Add documentation
Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

Change apex to zone, Update readme information, Add additional tests
Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

Change zone to apex
Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

Change readme to reflect apex change
Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

Correct code comment
Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

Correct string join for apex.dns
Signed-off-by: Mario Kleinsasser <mario.kleinsasser@gmail.com>

some other small bits as well.

* handle continue in response rewrite

* add test

Fixes #1720

Do ' go vet' after we have fetched the repos

Rework the TestProxyClose - close the proxy in the *same* goroutine
as where we started it. Close channels as long as we don't get dataraces
(this may need another fix).

Move the Dial goroutine out of the connManager - this simplifies things
*and* makes another goroutine go away and removes the need for connErr
channels - can now just be dns.Conn.

Also:

Revert "plugin/forward: gracefull stop (#1701)"
This reverts commit 135377bf.

Revert "rework TestProxyClose (#1735)"
This reverts commit 9e8893a0.

* plugin/health: update README

Make more clear in the readme that health is limited to 1 server.

Fixes #1722

* rephrase and remove ~~~ corefile because it will fail

* tweak parse error messaging, add tests

* looser err msg checking

Implement per server metrics in autopath; also don't export the metric.

Updated readme as well.

* core: remove HostAddresses()

config.HostAddresses() is a weird function that gathers
some data from the server and returns a string.

It is *only* used the trace plugin, to figure out what
server starts the trace.

Looks to be better to fit in the with metrics.WithServer label
on the trace itself to show which server handled the trace.

Remove HostAddresses() and cleanup trace a small bit.:w

* lint

https_google and reverse will be removed in 2 releases. Add warnings.

We will move reverse to coredns/reverse and list it as external after
this switch.

See #1713

Spin this out the metrics package so we can use it in the health
one of well to fix some reload bugs.

* plugin/forward: fix TLS setup

Way smaller PR than #1679. Fixes same thing.

* remove println

* put overwritten test back

* context

* update tests

Close a bunch of channels, also change the test to just use a for loop
with a counter.

Remove these 2 channels, as they were only used in testing and
complicate things if we want to do a clean stop()

Less is more.