SLLS-546 Don't use system clocks in tests

[tomshafir-sonarsource]

No description provided.

[hashicorp-vault-sonar-prod]

SLLS-546

[sonar-review-alpha]

Summary

This PR replaces system clock calls with fixed timestamps across test files to eliminate non-determinism and improve test reliability.

Changes:

• Instant.now() → Instant.EPOCH (Java epoch: 1970-01-01)
• System.currentTimeMillis() → Instant.parse("2026-01-01T00:00:00Z").toEpochMilli() in ConnectedModeMediumTests (using a shared constant)
• Affected files: CommandManagerTests, SonarLintVSCodeClientTests, TaintIssueTest, ConnectedModeMediumTests, DelegatingCellIssueTest, VersionedOpenNotebookTests

These changes ensure test behavior is consistent regardless of when tests are executed, making them more stable and faster (no ordering or timing dependencies).

What reviewers should know

Where to focus:

• ConnectedModeMediumTests: The main file with the most changes. Note the new CURRENT_TIME constant at line 109 using a fixed 2026-01-01 date.
• Other test files: All use Instant.EPOCH for timestamp fields, which is appropriate for mock/test data where the actual timestamp value is irrelevant.

Key design decisions:

• Instant.EPOCH is used where timestamps are incidental to the test logic (just filling required fields)
• ConnectedModeMediumTests uses a future-dated constant (2026-01-01) instead of epoch—verify this choice aligns with any time-sensitive assertions in that test class

What to verify:

• Confirm no tests have assertions that depend on relative timestamps (e.g., comparing timestamps to "now")
• Check whether the 2026-01-01 date in ConnectedModeMediumTests was chosen for a specific reason or could also use EPOCH

---

• Generate Walkthrough
• Generate Diagram

🗣️ Give feedback

[sonar-review-alpha]

LGTM! ✅

Clean, well-scoped PR. All Instant.now() and System.currentTimeMillis() usages in test files have been replaced — no remaining non-deterministic clock calls found.

One thing worth knowing for future work: ConnectedModeMediumTests uses a named constant (2026-01-01) instead of Instant.EPOCH. This is intentional — the same timestamp is threaded through both the initial pull response's queryTimestamp field and the changedSince=<value> query parameter in the incremental sync mock URL. The value needs to be self-consistent within each test, and a semantically meaningful named constant (CURRENT_TIME) makes the intent clearer than Instant.EPOCH. The codebase already uses Clock injection in LanguageClientLogger as the pattern for production-side time control; this PR correctly limits itself to the test side only.

🗣️ Give feedback

[sonarqube-next]

Quality Gate passed

Issues
0 New issues
20 Fixed issues
0 Accepted issues

Measures
0 Security Hotspots
0 Dependency risks
No data about Coverage
0.0% Duplication on New Code

See analysis details on SonarQube