Cover the tag-merge helpers and the first enabled: line

[dmccoystephenson]

Summary

This is a Stage B (unit-test expansion) cycle. The issue backlog was empty at triage, so no issues were deferred. The change is test-only: TraceClient.java is untouched.

Four tests are added for branches that had no direct test:

• withVersion_dropsNullKeysAndValuesAndKeepsTheEventsOwnVersion: null keys and values are dropped from the copy, an event's own version is kept, and the caller's map is left unmodified.
• withVersion_addsTheVersionEvenToAnEventAlreadyAtMaxTags: a characterization test. Current behaviour is pinned: an event with MAX_TAGS tags of its own goes out with MAX_TAGS + 1. This looks like a bug and is filed as withVersion pushes an event already at MAX_TAGS over the server's tag limit #13. Production code is deliberately left unchanged, because fixing it alters what is sent on the wire and that is a maintainer decision. When withVersion pushes an event already at MAX_TAGS over the server's tag limit #13 is resolved, this test is expected to be updated.
• withInstall_addsTheIdOnlyWhenThereIsOneAndRoomAndNoneAlready: covers all four outcomes, which are added, no ID, the event's own install wins, and already at MAX_TAGS. It also checks that the caller's map is not modified.
• serverWideConfig_theFirstEnabledLineIsTheSwitch: the first enabled: line decides, and a commented-out line does not count.

No tracking issue exists for the coverage itself; the gaps were found during triage.

Test plan

• mvn -B verify run locally on JDK 21: Tests run: 66, Failures: 0, Errors: 0, Skipped: 0 (62 before, plus 4)
• CI matrix (Java 8, 17, 21) green on the PR head

This PR description was drafted during a Gardener session (https://github.com/Stephenson-Software/gardener).

🤖 Generated with Claude Code

---

drafted by Claude on behalf of Daniel Stephenson

[dmccoystephenson]

Self-review rubric (performed inline in the same session, so it is not an independent review):

• Scope: PASS. git diff --stat origin/main shows only TraceClientTest.java, +72/-0. No production code is touched.
• Tests-new: PASS. No new public members. The 4 new tests cover withVersion, withInstall and parseServerWideConfig branches that had no direct test.
• Tests-fix: N/A. This is not a bug fix. The suspected bug is filed as withVersion pushes an event already at MAX_TAGS over the server's tag limit #13 and only characterized here.
• Sibling structure / renames: PASS. The tests follow the subject_behaviour naming and Arrange/Act/Assert comments. Nothing is renamed.
• Docs: PASS. No behaviour changed, so no README or Javadoc row is affected.
• Issue resolution: N/A. There is no Closes. withVersion pushes an event already at MAX_TAGS over the server's tag limit #13 is referenced but deliberately left open.
• CI: PASS. build (8), build (17) and build (21) all pass on the PR head (run 37429266363). Locally: Tests run: 66, Failures: 0, Errors: 0, Skipped: 0.
• No dependencies / single file / Java 8: PASS. pom.xml is untouched. git ls-files src/main lists only TraceClient.java. The new tests use only Java 8 APIs (LinkedHashMap, Collections.singletonMap, Arrays.asList), and the Java 8 leg is green.
• Never throws / non-blocking / env seam / version / opt-out contract: PASS (not applicable). No production code changed.

Findings (judgment calls, not blocking):

• src/test/java/software/stephenson/trace/TraceClientTest.java (withVersion_addsTheVersionEvenToAnEventAlreadyAtMaxTags): this pins behaviour that withVersion pushes an event already at MAX_TAGS over the server's tag limit #13 calls a likely bug. Whoever fixes withVersion pushes an event already at MAX_TAGS over the server's tag limit #13 must invert this assertion. The test comment names withVersion pushes an event already at MAX_TAGS over the server's tag limit #13 so that the link is not lost.
• src/test/java/software/stephenson/trace/TraceClientTest.java (serverWideConfig_theFirstEnabledLineIsTheSwitch): this test is placed next to the withVersion/withInstall tests rather than next to the other serverWideConfig_* tests near serverWideConfig_acceptsEverySpellingOfOff. That is a minor ordering inconsistency, and the test can be moved if the grouping matters.

Merge-gate note: no do-not-auto-merge path is modified. This dispatch is not authorized to merge, so the PR is left open for human review.

This review comment was drafted during a Gardener session (https://github.com/Stephenson-Software/gardener).

---

drafted by Claude on behalf of Daniel Stephenson