Investigate AI-agent analysis report on the server-discovery-and-monitoring specification. This ticket tracks findings identified during a systematic review of the spec: missing test coverage, ambiguous normative language, and spec/test inconsistencies.

Missing Tests

• topologyVersion comparison when processId is explicitly null (vs. absent) — spec: "assume response is more
  recent if either topologyVersion is unset"
• ServerDescription.isCryptd field: mongocryptd server detection and exclusion from normal topology
• maxElectionId/maxSetVersion comparison for MongoDB [ 6.0 (setVersion-first comparison) — version-dependent
  comparison logic not integration-tested (DRIVERS-2412 implementing:
  "SDAM should prioritize electionId over setVersion only on ]=6.0 servers")
• RSGhost server incorrectly reporting a non-null hosts list — spec: "client MUST NOT attempt to use its hosts
  list"
• Load-balanced topology error handling: serviceId-based pool clearing in all error paths
• Wire protocol compatibility error message format: driver name and version substitution in error string
• Hostname with Unicode/IDNA characters — only ASCII hostnames tested; spec requires normalization to lowercase

Ambiguities

• replacement MUST happen even if new ServerDescription compares equal: Does this apply to LoadBalanced topology
  where comparison semantics differ?
• "SHOULD emit a warning if constructed with no seeds": Warning format/delivery mechanism not specified.
• MongoDB 4.0 pool clearing: Spec says clear pool for stale primary on v4.0 or earlier — 4.0 is EOL; should drivers
  still support this path?
• compareTopologyVersion uses >= instead of >: Rationale for >= (equal versions considered stale) is
  counterintuitive and not explained.

Inconsistencies

• Error handling pseudocode vs. TopologyType table: "not writable primary" error handling differs between the
  pseudo-code and the table — table shows checkIfHasPrimary action, pseudocode has conditional logic.
• Network error boundary: Handling differs for errors "before handshake completes" vs. "after handshake completes,"
  but the exact boundary is undefined in practice.
• v4.0/v4.2 pool clearing boundary: "Mark old primary Unknown and clear its pool for v4.0 or earlier" but v4.2+
  should NOT clear pool — the 4.0/4.2 boundary behavior described in two contradictory-sounding places.

Notes

• ~450 test files covering all topologies. Error handling scenarios are moderately covered but edge cases around stale
  topologies and concurrent monitor updates lack depth.
• topologyVersion handling is tested but processId equality edge cases need coverage.
• Single-threaded monitoring scanning order is under-tested relative to multi-threaded cases.
• DRIVERS-3275 (Backlog): centralize SDAM tests.
• DRIVERS-1670 (Implementing): add log messages to SDAM spec.