Exact SDK revision validation and webhook backlog-age correction. All 14 targeted regression tests passed after Docker recovered.
Previous review · Findings · Scores · Validation dataLatest: Round 6 recovery improvements and scoring
Previous review: 96/100
Previous review: 95/100
Round 5 targeted follow-up: reviewed release-source validation and webhook recovery aging, with unchanged SDK runtime and wallet code. Round 4 broad validation remains historical evidence, not a new full-suite run. All 14 targeted database, migration and webhook tests passed after Docker recovered. Scores remain 96/95 because the remaining category deductions are still open.
Scores are engineering judgments against the same 30/25/20/15/10 rubric. They are not percentages of bug-free code or a security certification. Historical scores are preserved.
| Category | Maximum | TrustWeave | SaaS |
|---|---|---|---|
| Security | 30 | 29 | 29 |
| Correctness | 25 | 24 | 24 |
| Architecture | 20 | 19 | 19 |
| Testing | 15 | 15 | 14 |
| Experience | 10 | 9 | 9 |
| Area | Result | Scope and limits |
|---|---|---|
| Release source guard | 2 Python tests passed; actual SDK revision and fingerprint match | Local committed SDK. No publication, remote CI or deployment. |
| Backend compilation | Production and test Kotlin compilation passed | Production and test Kotlin compile; V25 migration and retry behavior also passed their database regressions. |
| Webhook unit tests | 5 passed in final targeted run | Two recovery monitor tests and three Stripe controller tests; all passed. |
| PostgreSQL integration and migration | 9 passed; 14 total targeted tests with zero failures or skips | Five Flyway and four recovery/access tests. BUILD SUCCESSFUL in 3m 14s. The targeted run excludes the whole-server coverage threshold; no full-suite/coverage rerun is claimed. |
| SDK and wallet | Unchanged runtime; Round 4 evidence retained | 3,729 JVM tests with 15 skips, 28 wallet unit tests and 9 browser tests were validated in Round 4, not rerun here. |
| SaaS broad suite | Round 4 evidence retained: 526 tests, 2 skips; coverage passed then | Historical broad-suite baseline. The new webhook database change is now additionally covered by successful Round 5 targeted validation. |
Results distinguish local database/HTTP fixtures from hosted-service validation. Optional tests that did not run remain reported as skips. Timing excludes fixture setup.
Updated the SaaS pin to the committed round-4 SDK and added exact HEAD validation. CI passes its explicitly selected immutable revision into the same guard.
Evidence: Two Python regressions pass, including rejection of branch names, abbreviated SHAs and a mismatched full SHA. The actual SDK checkout passes revision and fingerprint validation.
Limit: The commit exists locally; publication and remote CI remain unverified.
Preserve first_failed_at with a PostgreSQL atomic upsert, while updating latest failure details. The oldest-failure metric uses the original failure time. Flyway V25 backfills existing rows and permits inserts from older application versions.
Evidence: All 14 targeted tests passed: five Flyway migration tests, four PostgreSQL webhook recovery/access tests, three Stripe controller tests and two monitor tests. This includes preserved first-failure age after retry and V25 backfill/older-version insert compatibility.
Limit: Pre-migration first-failure history cannot be reconstructed. Hosted alert delivery remains untested.
The SDK is now committed and pinned locally. Publication, exact-pair remote Linux CI and designated staging validation remain pending.
Evidence: No publication or staging execution is claimed.
Limit: A matching local fingerprint cannot substitute for a reviewed, published and tested immutable pair.
Most plugin capabilities remain unassessed. The reference wallet has no hardware/user-presence signing or recoverable holder-key service; supported issuer and disclosure profiles remain limited.
Evidence: Full local JVM suite passed with 15 optional/template tests skipped; live provider environments and hardware-backed custody remain unvalidated.
Limit: A complete product/profile definition and supported provider matrix are needed before these deductions can be closed.
Two Accountly live-contract tests require a separately configured Accountly/PSP stand-in. Legacy Stripe invoice handlers remain logging-only for settlement/notification purposes.
Evidence: Local signed-webhook recovery and aggregate monitoring are tested; hosted alert delivery and live Accountly flow are not.
Limit: The supported Accountly cutover/legacy Stripe workflow needs to be defined and exercised before claiming completeness.
After Docker became available again, its engine responded with version 29.4.1 and the blocked test selection was rerun successfully.
Evidence: Gradle BUILD SUCCESSFUL in 3m 14s; 14 tests, zero failures, errors or skips. Per-suite results are recorded in backend-validation.json.
Limit: This was a targeted regression run, not a new full-server coverage run or hosted staging validation.
docs/round-5-operations.md covers the repaired release pin, Flyway V25, preserved failure age and deployment limits. Staging environment and release-branch details were requested during this pass and remain prerequisites for external release checks.
Implementation references: PostgreSQL advisory locks, Next.js CSP, Azure listing semantics.