This should make it slightly more re-usable in other controllers. In
particular this allows the `instance` to be `nil`, which might be the
case if the template was not created.

Signed-off-by: Justin Kulikauskas <jkulikau@redhat.com>
(cherry picked from commit c6dadad)

This might be slightly more performant, and other things can use this
clientset.

Signed-off-by: Justin Kulikauskas <jkulikau@redhat.com>
(cherry picked from commit c01cf0a)

Previously, the controller-runtime event recorder was used for these
events. Other policy controllers have moved away from that, for various
reasons. In this case, if a policy went from pending to noncompliant and
back to pending, the "old" pending event would be re-used by the event
recorder, and only the `lastTimestamp` would be updated. In this case,
if a policy controller emitted a compliance event within the same second
as the Pending event, the status-sync would see it as a tie, and use the
hex-encoded nanoseconds in the event name. But the event name was not
updated from the original instance when the policy was pending, so the
events would be ordered incorrectly.

Most error cases from this synchronous sending can be ignored because
they are already error cases that would be requeued.

Refs:
 - https://issues.redhat.com/browse/ACM-4699

Signed-off-by: Justin Kulikauskas <jkulikau@redhat.com>
(cherry picked from commit f0e2c60)

Signed-off-by: Justin Kulikauskas <jkulikau@redhat.com>
(cherry picked from commit d6cb733)

The KinD tests action will now run the gosec-scan, and that target will
fail if any vulnerabilities are found. The target was also configured to
ignore the test code.

Signed-off-by: Justin Kulikauskas <jkulikau@redhat.com>
(cherry picked from commit 8c251da)

Information about the gatekeeper pods might help if those tests fail.

Signed-off-by: Justin Kulikauskas <jkulikau@redhat.com>
(cherry picked from commit 0034b03)

The test is meant to ensure that the gatekeeper-sync is not emitting the
same event multiple times in a row. But the assertion was failing
sometimes because of duplicate events from template-errors. Those will
sometimes occur during normal (correct) operation of the template-sync.

Signed-off-by: Justin Kulikauskas <jkulikau@redhat.com>
(cherry picked from commit 906dcff)