Tests-027 GatewayMetrics exposes its internal Meter; the
StreamEvents_WhenEventIsWritten_RecordsSendDuration listener
now filters by ReferenceEquals(instrument.Meter, metrics.Meter)
instead of Meter.Name, so parallel tests with their own
GatewayMetrics no longer cross-contaminate the families list.
Tests-028 FakeWorkerClient.Kill now captures LastKillReason;
SessionManager.KillWorkerAsync tests pin the reason
propagation end-to-end and cover the blank/null guard. The
DashboardSessionAdminService kill test pins the literal
dashboard-admin-kill reason.
Tests-029 Added CloseSessionAsync_BlankSessionId_ReturnsFailure to mirror
the existing KillWorkerAsync blank-id coverage.
Tests-030 DeleteAsync_WhenStoreRefuses_ReportsFriendlyError renamed and
extended to assert the dashboard-delete-key audit row with
Details = not-found-or-active. Added
DeleteAsync_BlankKeyId_ReturnsFailure.
Tests-031 DashboardSnapshotPublisher reconnect test now measures the
gap from the first throw inside the fake (firstThrowAt) to
secondSubscribeAt, isolating Task.Delay from StartAsync /
scheduling overhead.
All resolved at 2026-05-24; 512/512 gateway tests pass.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Append 5 new findings (Tests-027..031) covering the
StreamEvents_WhenEventIsWritten_RecordsSendDuration flake root cause
(shared MeterListener by meter name), missing kill-path coverage
(reason propagation + concurrent-kill double-count), asymmetric guard
coverage between Close and Kill, missing audit-failure-path coverage
for ApiKey Delete, and the DashboardSnapshotPublisher reconnect-window
timer sensitivity.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Tests-025 (Conventions): Extracted the previously-duplicated
NullDashboardEventBroadcaster into TestSupport/NullDashboardEventBroadcaster.cs
(singleton Instance, private ctor). The two nested copies in
EventStreamServiceTests and GatewayEndToEndFakeWorkerSmokeTests were
removed; both files now use the shared type via
'using ZB.MOM.WW.MxGateway.Tests.TestSupport;'. The Server-041 regression
test's ThrowingDashboardEventBroadcaster is intentionally left nested —
single-file usage doesn't warrant promotion to TestSupport. The third
copy in IntegrationTests/WorkerLiveMxAccessSmokeTests was handled by
IntegrationTests-024 in its own commit.
Tests-026 (Testing coverage): Added a new RecordingDashboardEventBroadcaster
test double in TestSupport — a thread-safe (ConcurrentQueue<DashboardEventCapture>)
recorder. New fixture
StreamEventsAsync_PublishesEachEventToDashboardBroadcaster in
EventStreamServiceTests pushes two events through the fake session and
asserts the broadcaster received both with the correct sessionId and
WorkerSequence. TDD red→green confirmed: the deliberately-wrong
"Expected 3, Actual 2" red phase proved the recording fake was actually
invoked by the production code path.
Verification: 486/486 server tests passing (485 previous + 1 new).
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Restores the `code-reviews/` tree (was unwritten on this working copy)
and re-reviews every module per `REVIEW-PROCESS.md` against HEAD
`d692232`. The diff in scope is the five commits since the last sweep:
`dc9c0c9` (ZB.MOM.WW gateway-side rename + slnx migrate),
`397d3c5` (client SDK rename + the missing alarm-RPC proto types and
the .NET DiscoverHierarchyOptions POCO), `27ed651` (role-based LDAP
auth + HubToken bearer, drop PathBase), `6594359` (sidebar layout +
three SignalR push hubs), and `d692232` (EventsHub publisher + doc
refresh).
Module status
| Module | Open | Total | Delta this pass |
|---|---|---|---|
| Server | 8 | 43 | +6 |
| Contracts | 2 | 17 | +2 |
| Tests | 2 | 26 | +2 |
| IntegrationTests | 3 | 24 | +3 |
| Client.Java | 5 | 31 | +5 |
| Client.Rust | 1 | 21 | +1 |
| Worker | 0 | 25 | 0 (rename-only diff, clean) |
| Worker.Tests | 0 | 30 | 0 (rename-only diff, clean) |
| Client.Dotnet | 0 | 17 | 0 (rename + alarm-fix diff, clean) |
| Client.Python | 0 | 21 | 0 (rename + alarm-fix diff, clean) |
| Client.Go | 0 | 21 | 0 (rename + alarm-fix diff, clean) |
Total new findings: 19. Severity breakdown: 1 Medium-security
(Server-038), 4 Medium-documentation/coverage, 14 Low.
New findings
* Server-038 (Medium / Security) — EventsHub.SubscribeSession accepts
any session id from any Viewer; no per-session ACL guards the
EventsHub group fan-out.
* Server-039 (Low / Error handling) — HubTokenService.Validate
accepts a payload with null Name/NameIdentifier.
* Server-040 (Low / Conventions) — MapGroupsToRoles undocumented
full-vs-RDN lookup precedence.
* Server-041 (Low / Design adherence) — EventStreamService calls
IDashboardEventBroadcaster.Publish without a try/catch — fragile
seam relying on the never-throw contract.
* Server-042 (Low / Performance) — DashboardSnapshotPublisher tight
retry loop with no backoff (vs AlarmsHubPublisher 5s delay).
* Server-043 (Low / Documentation) — HubTokenService singleton
sharing across login + hub-token validation undocumented.
* Contracts-016 (Low / Conventions) — QueryActiveAlarmsRequest.session_id
reserved-for-future-use ambiguity.
* Contracts-017 (Low / Documentation) — rpc QueryActiveAlarms doc
omits the alarm_filter_prefix filter description.
* Tests-025 (Low / Conventions) — duplicate NullDashboardEventBroadcaster
fakes in EventStreamServiceTests and GatewayEndToEndFakeWorkerSmokeTests.
* Tests-026 (Medium / Testing coverage) — no test proves
EventStreamService actually calls IDashboardEventBroadcaster.Publish.
* IntegrationTests-022 (Low / Conventions) — ResolveRepositoryRoot
silent fallback to Directory.GetCurrentDirectory().
* IntegrationTests-023 (Low / Testing coverage) — DashboardLdapLiveTests
success-path asserts ldap_group but not the Role claim.
* IntegrationTests-024 (Low / Conventions) — inline
NullDashboardEventBroadcaster fake duplicates Tests-side copies.
* Client.Java-027 (Medium / Documentation) — README + JavaClientDesign
Gradle task names still use the old short project names.
* Client.Java-028 (Medium / Design adherence) — JavaClientDesign
build-layout shows the old `com/dohertylan/mxgateway/` package paths.
* Client.Java-029 (Low / Documentation) — README installDist path
cites the wrong directory.
* Client.Java-030 (Low / Testing coverage) — no Java test exercises
the regenerated QueryActiveAlarmsRequest RPC.
* Client.Java-031 (Low / Conventions) — README prose uses old short
project names instead of canonical prefixed ones.
* Client.Rust-021 (Low / Design adherence) — RustClientDesign.md
"Crate layout" shows an aspirational nested `crates/zb-mom-ww-mxgateway-client/`
that does not exist; actual layout is the flat top-level crate.
Two pre-existing pending findings (Server-031 lock-contention,
Server-032 bounded event channel) remain unchanged — neither was
touched by this wave of commits.
Process notes
- The `code-reviews/` tree was not in this working copy's git
history (the local extract pre-dates the divergent branch that
carried the reviews). Restored from `dd7ca16` via
`git checkout dd7ca16 -- code-reviews/` before the re-review.
- Some "Resolved" entries in the restored findings.md reference
fixes that landed on the divergent branch (the same one that
carried the reviews) and are not present on the current main
lineage. The re-review treats those statuses as historical;
the new pass only files findings against HEAD's actual state.
- `python code-reviews/regen-readme.py --check` is green.
Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
Joseph Doherty