refresh SRV records and send out ServiceRemoved for expired SRV

[keepsimple1]

This patch is to fix issue raised in #176. There are two parts:

• [x] Add checks of refresh for SRV record so that they will be updated if the responder is alive.
• [x] Also triggers ServiceRemoved when SRV record expires as the responder is no longer active.

Added a new setter _set_host_ttl() for DnsRecord to support testing.

Also addressed issue #203 : only apply cache-flush bit if the record is new, i.e. rdata changed. Otherwise, it is only a refresh of TTL. This makes sense to me as it can avoid unnecessary new ServiceResolved events.

Also removed -Wclippy::all from CI as it refers the default warnings, hence is redundant with -D warnings. (https://doc.rust-lang.org/clippy/)

[keepsimple1]

Rebased to the latest main branch and update the PR description.