Close LIST_ALL_QUERY (null-prefix) and S3-handler empty-prefix coverage gaps

[zichengl]

Summary

Adds two integration tests that close coverage gaps remaining after #3265:

1. MySqlNamedBlobDbListOperationIntegrationTest — the existing parametrized matrix never invoked namedBlobDb.list(...) with a null prefix, so the LIST_ALL_QUERY path (and the new option-4 LIST_ALL_SQL window-function variant introduced in Fix option-4 empty-prefix LIST regression; add window-function LIST_ALL #3265) had no parametrized integration coverage. The TMC failure that motivated Fix option-4 empty-prefix LIST regression; add window-function LIST_ALL #3265 traversed exactly this path (prefix="" → parseS3 collapses to null → list() with null → LIST_ALL_QUERY under option 4). This PR adds two test methods exercising it across options 2/3/4 × hard-delete on/off.

2. S3IntegrationTest — no end-to-end test sent the request shape ?prefix= (parameter present, value empty) that the AWS S3 SDK emits when ListObjectsRequest is built with an unset/empty prefix. Added s3ListEmptyPrefixTest covering HTTP → Netty → S3ListHandler → NamedBlobPath.parseS3 → NamedBlobListHandler → NamedBlobDb#list.

What this catches

Failure class	Caught by
Option-2/3/4 LIST_ALL_QUERY returns wrong rows on null prefix	testListNamedBlobsWithNullPrefix (across all 6 parameter combinations)
Option-4 LIST_ALL_SQL semantic bug — deleted_ts-placement footgun on the null-prefix path	testListAllNullPrefixHidesDeletedLatestUnderOption4 (option 4 only via Assume, mirrors the with-prefix invariant test in #3260)
Regression in NamedBlobPath.parseS3 empty-prefix-to-null collapse	s3ListEmptyPrefixTest (end-to-end through HTTP)
Regression in S3ListHandler empty-prefix routing	s3ListEmptyPrefixTest

What's left for a follow-up

Full stitched coverage of S3 handler → real MySQL with option 4 on a realistic container is not in this PR. s3ListEmptyPrefixTest runs against InMemNamedBlobDbFactory (the default backing in S3IntegrationTest), so it catches handler-layer regressions but not SQL-side regressions on the stitched path. Adding a MySQL-backed sibling of S3IntegrationTest would require refactoring buildFrontendVProps to accept a configurable NamedBlobDbFactory plus a port-collision-aware sibling class. Reserving that for a follow-up to keep this PR's diff small and review-focused.

The SQL-side coverage gap is fully closed by (1) — MySqlNamedBlobDbListOperationIntegrationTest runs against a real MySQL container (matching the OSS CI setup) on every PR.

Diff size

ambry-frontend/src/integration-test/.../S3IntegrationTest.java                                  | +59
ambry-named-mysql/src/integration-test/.../MySqlNamedBlobDbListOperationIntegrationTest.java   | +96
2 files changed, +155 / -0

Testing Done

• ./gradlew :ambry-named-mysql:compileIntTestJava → BUILD SUCCESSFUL.
• ./gradlew :ambry-frontend:compileIntTestJava → BUILD SUCCESSFUL.
• ./gradlew :ambry-named-mysql:intTest --tests '*MySqlNamedBlobDbListOperationIntegrationTest.testListNamedBlobsWithNullPrefix*' --tests '*MySqlNamedBlobDbListOperationIntegrationTest.testListAllNullPrefixHidesDeletedLatestUnderOption4*' → 12 runs, 8 passed, 0 failed, 4 skipped. The 4 skipped are the option-4-only invariant test on options 2/3 (via Assume.assumeTrue).
• ./gradlew :ambry-frontend:intTest --tests '*S3IntegrationTest.s3ListEmptyPrefixTest' → 1 passed.
• Full :ambry-named-mysql:intTest matrix (MySqlNamedBlobDbListOperationIntegrationTest*) against a local MySQL 8.0 container → 42 total, 38 passed, 4 skipped, 0 failed (30 baseline + 12 new, with the 4 Assume-skips for the option-4-only invariant).

Durability risk

Test-only addition. No production code path touched, no schema, no config defaults. Zero durability risk.

AI Usage

Used Claude Code (Opus 4.7) to identify the coverage gap (both MySqlNamedBlobDbListOperationIntegrationTest's 9 list() invocations passing non-null prefix, and S3IntegrationTest's reliance on the helper method that drops empty prefix values), draft the test methods, run them against a local Docker MySQL 8.0 container, and author this PR description. Author reviewed.

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 50.94%. Comparing base (52ba813) to head (2572752).
⚠️ Report is 396 commits behind head on master.

Additional details and impacted files

@@              Coverage Diff              @@
##             master    #3270       +/-   ##
=============================================
- Coverage     64.24%   50.94%   -13.30%     
+ Complexity    10398     8699     -1699     
=============================================
  Files           840      937       +97     
  Lines         71755    80199     +8444     
  Branches       8611     9633     +1022     
=============================================
- Hits          46099    40859     -5240     
- Misses        23004    35956    +12952     
- Partials       2652     3384      +732     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.