feat(creacc): translate CREACC to Java by a2chang · Pull Request #13 · augment-solutions/cbsa-java

a2chang · 2026-05-01T08:18:31Z

Source program: https://github.com/augment-solutions/cics-banking-sample-application-cbsa/blob/main/src/base/cobol_src/CREACC.cbl

Mapped REST endpoints

POST /api/v1/creacc/insert

Notable decisions

Reused InqcustService and InqacccuService to mirror the COBOL EXEC CICS LINK validation flow before persisting the account.
Replaced the COBOL named-counter ENQ/DEQ with a CockroachDB control row SELECT FOR UPDATE, wrapped in CrdbRetry, so account allocation is transactional and retry-safe.
Persisted the account, control, and proctran updates in one transaction so failed writes do not leave partially-created account state behind.
Preserved the z/OS Connect nested CreAcc JSON contract in request/response DTOs while continuing the project-wide RFC-7807 ProblemDetail error handling for failure paths.

Test coverage

Unit: CreaccServiceUnitTest
SpringBootTest: CreaccServiceIntegrationTest
MockMvc: CreaccControllerTest
Web slice: CreaccControllerWebMvcTest

a2chang · 2026-05-01T08:18:39Z

augment review

augmentcode · 2026-05-01T08:20:21Z

^{Test Coverage Guardian 🧪}

✅ Test coverage looks good. The new/changed behavior in this PR has adequate test coverage. No additional tests needed.

This PR includes comprehensive testing across multiple layers:

Unit tests: CreaccServiceUnitTest validates business logic with mocked dependencies
Integration tests: CreaccServiceIntegrationTest tests the full service layer against CockroachDB
Controller tests: CreaccControllerTest validates HTTP endpoints with SpringBootTest
Web slice tests: CreaccControllerWebMvcTest tests controller behavior in isolation

The tests cover happy paths, validation failures, edge cases (max accounts per customer), and error handling scenarios.

augment-app-staging

Review completed. 2 suggestions posted.

Comment augment review to trigger a new review at any time.

augment-app-staging · 2026-05-01T08:23:35Z

+    }
+
+    private boolean isSupportedAccountType(String accountType) {
+        return accountType.startsWith("ISA")


Using startsWith here will accept values like ISAFOO (still <= 8 chars) as “supported”, which looks like it bypasses the intended account-type validation.

Severity: medium

_{🤖 Was this useful? React with 👍 or 👎}

augment-app-staging · 2026-05-01T08:23:35Z

+    @Test
+    void createsAccountAuditTrailAndControlRow() {
+        insertCustomer(10L);
+        LocalDate today = LocalDate.now(Clock.systemUTC());


Capturing today before calling creaccService.create(...) can make this test flaky if execution crosses midnight UTC between the two calls (the service derives its own today from Instant.now(clock)).

Severity: low

Other Locations

src/test/java/com/augment/cbsa/web/creacc/CreaccControllerTest.java:51

_{🤖 Was this useful? React with 👍 or 👎}

a2chang · 2026-05-01T08:26:49Z

augment review

augment-app-staging

Review completed. 2 suggestions posted.

Comment augment review to trigger a new review at any time.

augment-app-staging · 2026-05-01T08:27:15Z

+    private static final Set<String> VALIDATION_FAIL_CODES = Set.of("A");
+    private static final Set<String> NOT_FOUND_FAIL_CODES = Set.of("1");
+    private static final Set<String> CAPACITY_FAIL_CODES = Set.of("8");
+    private static final Set<String> TRANSIENT_FAIL_CODES = Set.of("3", "5");


After replacing the COBOL ENQ/DEQ with transactional row locking, it looks like CREACC never produces failCode 3/5 anymore, so isTransientFailure() (and the controller’s 503 mapping) may be unreachable.
If counter/lock acquisition can still fail transiently, consider whether it should be surfaced as a CREACC failure code rather than bubbling up as an abend/500.

Severity: medium

Other Locations

src/main/java/com/augment/cbsa/web/creacc/CreaccController.java:67

src/main/java/com/augment/cbsa/repository/CreaccRepository.java:39

_{🤖 Was this useful? React with 👍 or 👎}

augment-app-staging · 2026-05-01T08:27:15Z

+    @Test
+    void createsAccountAuditTrailAndControlRow() {
+        insertCustomer(10L);
+        LocalDate today = LocalDate.now(Clock.systemUTC());


These assertions derive today from Clock.systemUTC() separately from the application’s injected Clock, which can make the test flaky if execution crosses midnight between the two reads.
Using the same fixed/overridden Clock for both the service and expectations would avoid intermittent failures.

Severity: low

Other Locations

src/test/java/com/augment/cbsa/web/creacc/CreaccControllerTest.java:51

_{🤖 Was this useful? React with 👍 or 👎}

augment-app-staging

Review completed. 1 suggestion posted.

Comment augment review to trigger a new review at any time.

augment-app-staging · 2026-05-01T08:32:27Z

+                    .set(ACCOUNT.ACTUAL_BALANCE, account.actualBalance())
+                    .execute();
+        } catch (DataAccessException exception) {
+            throw rollbackFailure("7", "Unable to create the account record.");


By catching DataAccessException here and converting it into a rollback failure, Cockroach serialization failures (SQLSTATE 40001) will also be treated as a business failure rather than being retried by CrdbRetry. This can make account creation intermittently fail under contention even though the operation is retryable.

Severity: medium

Other Locations

src/main/java/com/augment/cbsa/repository/CreaccRepository.java:58

src/main/java/com/augment/cbsa/repository/CreaccRepository.java:93

src/main/java/com/augment/cbsa/repository/CreaccRepository.java:126

_{🤖 Was this useful? React with 👍 or 👎}

a2chang · 2026-05-01T08:36:40Z

augment review

augment-app-staging

Review completed. 3 suggestions posted.

Comment augment review to trigger a new review at any time.

augment-app-staging · 2026-05-01T08:41:44Z

+) {
+
+    public CreaccCommareaRequestDto {
+        Objects.requireNonNull(commCustno, "commCustno must not be null");


These Objects.requireNonNull(...) checks in the request DTO constructor can throw during JSON deserialization, which will bypass the validation handlers and fall into CbsaExceptionHandler#handleUnexpected (500) instead of returning a 400 validation error. That can make missing/invalid fields surface as server errors.

Severity: medium

Other Locations

src/main/java/com/augment/cbsa/web/creacc/dto/CreaccRequestDto.java:16

src/main/java/com/augment/cbsa/web/creacc/dto/CreaccKeyDto.java:24

_{🤖 Was this useful? React with 👍 or 👎}

augment-app-staging · 2026-05-01T08:41:44Z

+    private static final String COUNTER_ABEND_CODE = "HNCS";
+    private static final String PROCTRAN_ABEND_CODE = "HWPT";
+    private static final long MAX_ACCOUNT_NUMBER = 99_999_999L;
+    private static final DateTimeFormatter COBOL_DATE_FORMATTER = DateTimeFormatter.ofPattern("ddMMyyyy");


DateTimeFormatter.ofPattern("ddMMyyyy") without an explicit locale can emit non-ASCII digits under some default locales, which can corrupt persisted PROCTRAN.DESCRIPTION (it’s meant to be COBOL-style numeric text). Consider ensuring the formatter is locale-stable for these numeric fields.

Severity: low

Other Locations

src/main/java/com/augment/cbsa/web/creacc/CreaccController.java:30

_{🤖 Was this useful? React with 👍 or 👎}

augment-app-staging · 2026-05-01T08:41:44Z

+        return new CreaccResponseDto(new CreaccCommareaResponseDto(
+                EYE_CATCHER,
+                account.customerNumber(),
+                new CreaccKeyDto(Integer.parseInt(account.sortcode()), account.accountNumber()),


Integer.parseInt(account.sortcode()) will throw if cbsa.sortcode ever contains non-digit characters or surrounding whitespace, turning a configuration issue into a 500 at runtime. It might be worth validating/normalizing sortcode before using it to build responses.

Severity: medium

_{🤖 Was this useful? React with 👍 or 👎}

a2chang · 2026-05-01T08:46:20Z

Round 3 of 3 (auto-review cap reached). Round-3 surfaced 4 distinct items; the 5th comment re-flagged the DataAccessException catch issue that was already addressed in this round on commit 5b7e924.

Captured as follow-up issues so this can merge:

CREACC: review whether transient counter-lock failures should map to a CreaccResult fail code instead of 500 #14 — transient counter-lock failure mapping (CreaccResult/Controller 503 path)
CREACC DTOs: replace Objects.requireNonNull in record constructors with bean-validation so missing fields return 400, not 500 #15 — DTO requireNonNull → bean-validation (return 400 instead of 500)
CREACC: pin Locale.ROOT on DateTimeFormatter.ofPattern("ddMMyyyy") so PROCTRAN.DESCRIPTION stays ASCII-stable #16 — pin Locale.ROOT on COBOL date formatters
CREACC: validate/normalise cbsa.sortcode before Integer.parseInt(...) in the controller #17 — validate/normalise cbsa.sortcode before parseInt

Merging per the 3-round review cap.

feat(creacc): translate CREACC to Java

5456d2e

augment-app-staging Bot reviewed May 1, 2026

View reviewed changes

fix(creacc): tighten account-type matching; share clock in tests

a780450

augment-app-staging Bot reviewed May 1, 2026

View reviewed changes

fix(creacc): rethrow CRDB serialization failures so CrdbRetry can retry

5b7e924

augment-app-staging Bot reviewed May 1, 2026

View reviewed changes

a2chang merged commit 5628a12 into main May 1, 2026
1 check passed

a2chang deleted the migrate/CREACC branch May 1, 2026 08:46

a2chang mentioned this pull request May 1, 2026

CBSA Migration Ledger #3

Closed

22 tasks

Conversation

a2chang commented May 1, 2026

Uh oh!

a2chang commented May 1, 2026

Uh oh!

augmentcode Bot commented May 1, 2026

Uh oh!

augment-app-staging Bot left a comment

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot May 1, 2026

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot May 1, 2026

Choose a reason for hiding this comment

Uh oh!

a2chang commented May 1, 2026

Uh oh!

augment-app-staging Bot left a comment

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot May 1, 2026

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot May 1, 2026

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot left a comment

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot May 1, 2026

Choose a reason for hiding this comment

Uh oh!

a2chang commented May 1, 2026

Uh oh!

augment-app-staging Bot left a comment

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot May 1, 2026

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot May 1, 2026

Choose a reason for hiding this comment

Uh oh!

augment-app-staging Bot May 1, 2026

Choose a reason for hiding this comment

Uh oh!

a2chang commented May 1, 2026

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

1 participant