dohertj2/lmxopcua
1
0
Fork 0
Code Issues 2 Pull Requests 1 Actions Packages Projects Releases Wiki Activity

review(Driver.FOCAS.Cli): FlushLogging() in finally + fix misleading detach comment

Browse Source 
Re-review at 7286d320. -006 (Low): FlushLogging() in all command finally blocks + tests.
-007: rewrite the inaccurate handler-detach comment (cleanup is via await using disposal).
This commit is contained in:
Joseph Doherty
2026-06-19 12:08:45 -04:00
parent b50fd6c34a
commit 754c5a3684
6 changed files with 191 additions and 25 deletions
Download Patch File Download Diff File Expand all files Collapse all files
code-reviews/Driver.FOCAS.Cli/findings.md
+82 -2
View File
@@ -4,8 +4,8 @@
|---|---|
| Module | `src/Drivers/Cli/ZB.MOM.WW.OtOpcUa.Driver.FOCAS.Cli` |
| Reviewer | Claude Code |
| Review date | 2026-05-22 |
| Commit reviewed | `76d35d1` |
| Review date | 2026-06-19 |
| Commit reviewed | `7286d320` |
| Status | Reviewed |
| Open findings | 0 |
@@ -229,3 +229,83 @@ promises, restoring parity between the docs and the shipped behaviour. Regressio
shortlist again; the existing well-known shortlist `[Theory]` was extended with
the same five entries to enforce the exact `0x... (Name)` rendering. Suite now
47 green (was 42).
## Re-review 2026-06-19 (commit 7286d320)
All five prior findings (001–005) remain Resolved. The test suite now has 52 tests
(48 pre-existing + 4 added in this re-review). Two new findings were identified.
#### Checklist coverage (re-review)
| # | Category | Result |
|---|---|---|
| 1 | Correctness & logic bugs | No new issues found |
| 2 | OtOpcUa conventions | No new issues found |
| 3 | Concurrency & thread safety | No new issues found |
| 4 | Error handling & resilience | No new issues found |
| 5 | Security | No new issues found |
| 6 | Performance & resource management | No new issues found |
| 7 | Design-document adherence | No new issues found |
| 8 | Code organization & conventions | Driver.FOCAS.Cli-006, Driver.FOCAS.Cli-007 |
| 9 | Testing coverage | No new issues found |
| 10 | Documentation & comments | Driver.FOCAS.Cli-007 |
### Driver.FOCAS.Cli-006
| Field | Value |
|---|---|
| Severity | Low |
| Category | Code organization & conventions |
| Location | `Commands/ProbeCommand.cs`, `Commands/ReadCommand.cs`, `Commands/WriteCommand.cs`, `Commands/SubscribeCommand.cs` |
| Status | Resolved |
**Description:** `DriverCommandBase.ConfigureLogging` was extended between review
commits to document that callers must call `FlushLogging()` in a `finally` block
to ensure buffered Serilog output is flushed before process exit. Every sibling
driver CLI (Modbus, AbCip, AbLegacy, TwinCAT) honours this contract, but all four
FOCAS CLI commands were missing the call. With the Console sink (Serilog 6.0.0)
there is no actual data loss today because `System.Console.Out` auto-flushes;
however the gap is a maintenance hazard: if a file sink or an async sink is ever
added to the logger configuration, the omission becomes data-losing. The
inconsistency with sibling CLIs also makes the codebase misleading for reviewers.
**Recommendation:** Add `try/finally` wrappers in ProbeCommand, ReadCommand, and
WriteCommand (which had no finally blocks since the -004 cleanup) and extend
SubscribeCommand's existing finally block, calling `FlushLogging()` in each.
Add a convention test (parallel to `CommandDisposalConventionsTests`) that
scans source files and asserts the call is present, preventing silent regressions.
**Resolution:** Resolved 2026-06-19 — added `try/finally` wrappers with
`FlushLogging()` to `ProbeCommand`, `ReadCommand`, and `WriteCommand`; extended
SubscribeCommand's existing `finally` block with the same call. Added
`FlushLoggingConventionsTests` with a `[Theory]` over all four command files so
the call cannot be silently dropped in future refactors. Suite now 52 green (was 48).
### Driver.FOCAS.Cli-007
| Field | Value |
|---|---|
| Severity | Low |
| Category | Documentation & comments |
| Location | `Commands/SubscribeCommand.cs:113-120` |
| Status | Resolved |
**Description:** The comment in the `SubscribeCommand.finally` block (added by
the -002 resolution) said "detach the OnDataChange handler before unsubscribe +
disposal" — but no `-=` operator call exists in the block. The comment's claim
was factually incorrect: handler cleanup is achieved by driver disposal (via
`await using`), which tears down the `PollGroupEngine` and prevents further
`OnDataChange` events; not by explicit unsubscription from the event. A
parenthetical note at the end of the original comment partially explained the
real mechanism, but the leading sentence was still misleading for future readers.
**Recommendation:** Reword the comment to accurately describe what the code does:
`UnsubscribeAsync` stops the poll-group ticker (a subscription lifecycle step);
the anonymous handler is never explicitly removed via `-=`; driver disposal via
`await using` is the real cleanup.
**Resolution:** Resolved 2026-06-19 — rewrote the comment to correctly describe
the two-step teardown: `UnsubscribeAsync` halts the poll ticker (subscription
lifecycle), then `await using` disposes the driver and tears down
`PollGroupEngine` (real cleanup). Removed the incorrect claim about `-=`
detachment. No behavioral change; documentation only.