tests: verify release input boundaries and forced full qualification

docs/ops/li-bridge-review-comments.md

@@ -17,7 +17,7 @@ limitations under the License.
 
 # Review comment dispositions
 
-All 62 original threads have replies with published source decisions. A fresh GraphQL readback across 37 PRs verified every expected reply, found no mismatches and found no new review threads. Resolved status alone was not accepted as proof. Re-fetch after the final publication and check the actual source/test coverage before closing the review.
+All 62 original threads have replies with published source decisions. A fresh GraphQL readback across 41 PRs verified every expected reply, found no mismatches and found no new review threads. Resolved status alone was not accepted as proof. Re-fetch after the final publication and check the actual source/test coverage before closing the review.
 
 ## Later qualification findings
 
@@ -27,7 +27,7 @@ All 62 original threads have replies with published source decisions. A fresh Gr
 | [F19: recreated topic already assigned to returning replica](https://github.com/linkedin/kafka/pull/584#issuecomment-5631079569) | Paired identity recovery in 585/586; missing/zero IDs, errors, retries, current/future copies and mixed-batch tests. [Qualification update](https://github.com/linkedin/kafka/pull/586#issuecomment-5638005151). | All four revision-4 record checks passed, but full final-source and wrapper qualification remain open. |
 | [F20: rotated protocol logs omitted](https://github.com/linkedin/kafka/pull/584#issuecomment-5638004839) | PR 584 retains and scans hourly rotations; 586 is restacked on it. New positive and negative tests fail before the fix and pass afterward. | The previous failed CI job stays failed. Re-run the corrected collector. |
 | [F21: churn exits during controller movement](https://github.com/linkedin/kafka/pull/586#issuecomment-5638182322) | PR 584 fixes the workload retry policy without changing the upstream broker response. [Test and code update](https://github.com/linkedin/kafka/pull/584#issuecomment-5638482746). | The complete revision-4 process run and audit pass; final F22/wrapper qualification remains required. |
-| F22: bridge-state MBeans register by default | Paired [588](https://github.com/linkedin/kafka/pull/588)/[589](https://github.com/linkedin/kafka/pull/589) add a default-off diagnostics flag. Tests cover disabled registration, enabled readings, restart scope, KRaft and cleanup. | All 72 Python tests pass. The wrapper mapping still needs matching-jar qualification. |
+| F22: bridge-state MBeans register by default | Paired [588](https://github.com/linkedin/kafka/pull/588)/[589](https://github.com/linkedin/kafka/pull/589) add a default-off diagnostics flag. Tests cover disabled registration, enabled readings, restart scope, KRaft and cleanup. | All 72 Python tests pass. The matching-jar wrapper suite now passes 133 tests; the complete final-source verifier remains open. |
 
 ## PR 541
 
@@ -101,7 +101,7 @@ All 62 original threads have replies with published source decisions. A fresh Gr
 | Comment | Decision | Code/test evidence | Reply |
 |---|---|---|---|
 | [1](https://github.com/linkedin/kafka/pull/551#discussion_r3925877328) | Capture the post-truncation offset and size while holding the log lock. | UnifiedLog.truncateTo; truncation test and mixed stale-leader process scenario. | [reply](https://github.com/linkedin/kafka/pull/551#discussion_r3984229174) |
-| [2](https://github.com/linkedin/kafka/pull/551#discussion_r3925877401) | Convert malformed numbers to ConfigException. Keep empty lists as an intentional way to disable a bucket dimension. | KafkaConfigTest.testInvalidRequestMetricBuckets and testEmptyRequestMetricBuckets; empty is not rejected because it is supported. | [reply](https://github.com/linkedin/kafka/pull/551#discussion_r3984229326) |
+| [2](https://github.com/linkedin/kafka/pull/551#discussion_r3925877401) | Reject empty bucket lists with ConfigException, as requested. The previous reply and ledger incorrectly claimed support for empty lists. | KafkaConfigTest.testEmptyRequestMetricBuckets asserts rejection; malformed, negative and unordered cases have separate assertions. | [correction](https://github.com/linkedin/kafka/pull/551#discussion_r3992669600) |
 | [3](https://github.com/linkedin/kafka/pull/551#discussion_r3925877442) | Close cumulative counters when their associated ingress metric is removed. | BrokerTopicMetricsTest.testCloseMetricClosesCumulativeIngressCounters; moved to the storage-metrics split. | [reply](https://github.com/linkedin/kafka/pull/551#discussion_r3984229507) |
 | [4](https://github.com/linkedin/kafka/pull/551#discussion_r3926413418) | Reject negative and non-increasing boundaries. | KafkaConfigTest.testRequestMetricBucketsMustBeOrderedAndNonNegative. | [reply](https://github.com/linkedin/kafka/pull/551#discussion_r3984229700) |
 | [5](https://github.com/linkedin/kafka/pull/551#discussion_r3926413465) | Assert on the specific metric names, not global registry size. | BrokerTopicMetricsTest uses per-metric presence checks; storage split tests pass. | [reply](https://github.com/linkedin/kafka/pull/551#discussion_r3984229872) |