Fix deadlock when entry is removed while receiver holds lock in lookup.

[fancycode]

Problem happened in CI tests of GrpcClients which tried to remove the entry concurrently with onLookup.

When trying to remove the entry here:

nextcloud-spreed-signaling/grpc_client.go

Line 755 in c85b31b

c.dnsMonitor.Remove(entry.entry)

The lock was acquired here:

nextcloud-spreed-signaling/grpc_client.go

Line 744 in c85b31b

c.mu.Lock()

Concurrently, the DNS monitor was notifying about the lookup here:

nextcloud-spreed-signaling/grpc_client.go

Line 511 in c85b31b

c.mu.Lock()

And previously acquired its lock here:

nextcloud-spreed-signaling/dns_monitor.go

Line 64 in c85b31b

e.mu.Lock()

Found here:
https://github.com/strukturag/nextcloud-spreed-signaling/actions/runs/7813581057/job/21313022175?pr=653

With this change, removing entries is possible while the lookup results are currently processed.

[coveralls]

Pull Request Test Coverage Report for Build 7814441439

• 0 of 34 (100.0%) changed or added relevant lines in 1 file are covered.
• No unchanged relevant lines lost coverage.
• Overall first build on dnsmonitor-deadlock at 53.531%

---

Totals
Change from base Build 7812492684:	53.5%
Covered Lines:	7702
Relevant Lines:	14388

---

💛 - Coveralls