외부 리뷰("현재 상태를 유지하기 위한 검증이 너무 많고, 그 검증 자체를
다시 검증하는 구조까지 생겼다")를 설계 문서로 정리하고 코드로 반영한다.
설계·판단 근거는 docs/superpowers/specs/2026-09-16-verification-surface-reduction-design.md.
삭제
- .github/ci-gate-matrix.yml(1,025줄) + verify-gate-matrix.sh(568줄):
Gradle task graph와 workflow graph에 이미 있는 정보의 3중 복제
- verify-gradle-wrapper.sh(799줄): workflow 바이트 해시 잠금.
wrapper 검증은 gradle/actions/wrapper-validation(full SHA 핀)에 위임
- DeveloperExperienceContractTest 등의 CI YAML mutation 테스트:
애플리케이션 test suite가 GitHub Actions YAML 파서를 검증하던 계층 역전
- 문서 drift 파서: verifyReadmeCommands, verifyRunbookReferences,
verifyDocumentedLeafCount, verifyTestSourceSetRegistry
- 빈 레지스트리를 지키던 커스텀 YAML 파서: verifyTrivyignore,
verifyQuarantineSunset, flaky-quarantine.yaml
- verifyConfigurationPropertiesProcessor, verifyOneTypePerFile:
각각 ca.spring-config convention과 Checkstyle OneTopLevelClass가 대체
- 정상 입력으로도 성공할 수 없던 messaging always-fail task
- ModuleRegistry의 JSON 필드 집합 정확 일치, sample-portfolio negative guard
이동
- java/quality/spring 공통 설정을 configure(subprojects) 블록에서
ca.java-conventions / ca.quality-conventions / ca.java-library /
ca.spring-library convention plugin으로
- 아키텍처 검증을 ca.architecture로, JPA·messaging qualification을
gradle/qualification/ 아래로, verifyEnvKeys를 :app-bootstrap 소유로
완화
- Git revision은 releaseCheck·아카이브 생성에서만 요구. 일반 빌드는 SNAPSHOT
- SpotBugs/FindSecBugs는 로컬 check에서 빼고 qualityCheck 레인으로
task 계층
- leaf check는 그 leaf만. architectureCheck / qualityCheck /
configContractCheck / integrationCheck / ci / releaseCheck로 이름 분리
CI
- _reusable-gradle.yml 신규. checkout + wrapper validation + JDK/캐시 공통화
- fileserver-release.yml -> fileserver-certification.yml (CD가 아니라 certification)
- GitHub Actions = CI + artifact, Argo CD = CD 경계를 docs/ci-cd/boundary.md로 고정
순증감 +3,274 / -7,483.
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
63 lines
3.7 KiB
Groovy
63 lines
3.7 KiB
Groovy
apply plugin: 'ca.platform-module'
|
|
apply plugin: 'ca.spring-config'
|
|
|
|
// Scopes, not a flat list of `api`.
|
|
//
|
|
// Sixteen internal leaves were exposed as `api`, so every adopter of the starter compiled against
|
|
// the Kafka adapter, the Rabbit adapter, both JDBC reliability leaves and the admin plane whether
|
|
// they used them or not — and both brokers' `@ConditionalOnClass` guards were true at once because
|
|
// the starter itself dragged in both clients.
|
|
//
|
|
// `api` is now the types that genuinely appear in this module's public signatures; everything else
|
|
// is `implementation`. Splitting the starter into per-capability artifacts (core / kafka / rabbit /
|
|
// reliability / admin) is the structural fix the review asks for and is a registry change of its
|
|
// own; this is the part that stops an adopter's compile classpath from containing two broker SDKs.
|
|
dependencies {
|
|
// Types an adopter writes against: envelopes, destinations, publish results, profiles, ports.
|
|
api project(':messaging:messaging-core-api')
|
|
api project(':messaging:messaging-schema-api')
|
|
api project(':messaging:messaging-policy')
|
|
api project(':messaging:messaging-transport-spi')
|
|
api project(':messaging:messaging-observability')
|
|
api project(':messaging:messaging-security')
|
|
api project(':messaging:messaging-reliability-api')
|
|
|
|
// Wired by the auto-configurations and never named in an adopter's source.
|
|
implementation project(':messaging:messaging-schema-json')
|
|
implementation project(':messaging:messaging-cloudevents')
|
|
// The publisher and its collaborators live here. The starter did not depend on it at all,
|
|
// which is why MessagingCoreAutoConfiguration consumed a MessagePublisher that nothing in
|
|
// reach could produce (MSG-INT-003).
|
|
implementation project(':messaging:messaging-runtime-core')
|
|
implementation project(':messaging:messaging-kafka')
|
|
implementation project(':messaging:messaging-rabbit')
|
|
implementation project(':messaging:messaging-outbox-jdbc-postgresql')
|
|
implementation project(':messaging:messaging-inbox-jdbc-postgresql')
|
|
implementation project(':messaging:messaging-claim-check')
|
|
implementation project(':messaging:messaging-admin-api')
|
|
implementation project(':messaging:messaging-admin-runtime')
|
|
|
|
implementation 'org.springframework.boot:spring-boot-autoconfigure'
|
|
implementation 'org.springframework.boot:spring-boot-actuator'
|
|
|
|
// The Reactor facade lives here, not in core-api: the core contract stays CompletionStage so
|
|
// that a service which does not use Reactor never inherits it. api, because
|
|
// ReactiveMessagePublisher returns Mono — a facade whose return type an adopter cannot name
|
|
// is a facade they cannot call.
|
|
api 'io.projectreactor:reactor-core'
|
|
testImplementation 'io.projectreactor:reactor-test'
|
|
testImplementation 'org.springframework.boot:spring-boot-test'
|
|
testImplementation 'org.springframework:spring-test'
|
|
// Test-only, and specifically so the configuration reference can be executed rather than read:
|
|
// MessagingConfigurationBindingTest loads the YAML blocks out of the shipped document and boots
|
|
// a context on them. Spring Boot 4 treats snakeyaml as optional, so a document-driven test has
|
|
// to ask for it; production code does not, and adding it here keeps that asymmetry visible.
|
|
testImplementation 'org.yaml:snakeyaml'
|
|
// The live round trip. MSG-INT-003's membership rule is that the starter earns
|
|
// app-bootstrap membership in the same change that proves a real broker round trip,
|
|
// so the proof has to run somewhere — and the starter is where the whole platform
|
|
// assembles.
|
|
testImplementation 'org.testcontainers:testcontainers-junit-jupiter'
|
|
testImplementation 'org.testcontainers:testcontainers-kafka'
|
|
}
|