Files
document-haness/docs/clean-architecture-backend-template/final/evidence/raw/323-experimental-adapters-shared-findings.txt
T
DongHyeonkaandClaude Opus 5 b2963105a8 docs(keycloak-session-store): import the session-storage lab as a new project
The keycloak project ended with four open questions that design could not
settle. A two-VM lab was built to answer them by measurement, and this is
that material: 26 experiments, 125 raw command outputs, 22 browser captures.

Follows the import procedure in README.md.

  source/     the originating repository verbatim — 78 documents, 28 SVGs,
              8 manifests, plus .source-revision recording the commit
  final/      the SSOT
    document.md   729 lines written from the 29 experiment documents, not
                  concatenated: what was predicted, what was measured, and
                  where the measurement itself was wrong
    evidence/raw    125 outputs, flattened to <experiment>__<file> because
                    the originals collided (01-baseline.txt appeared three
                    times) and the audit only globs the top level
    evidence/meta   one per raw file; command and exitCode are null and the
                    README says why rather than inventing them
    evidence/browser  22 captures
    assets/       three diagrams through techviz
    .techviz/     their VizSpecs

A separate project rather than an addition to keycloak: the B-layer answers
that project's four questions, but the A, C and D layers are about cluster
failure, SSO and operations, and one document.md should hold one subject.
The four question records there can point here through 관계.

Recorded rather than papered over: only three of the 28 diagrams were
remade. The repository forbids hand-drawn SVG and forbids titles inside the
canvas; all 28 originals carry both, so converting them is redrawing, not
reformatting. They stay in source/ and the gap is written into the document.

verify-pipeline.py passes. audit-records.py reports no issues.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
2026-09-04 22:51:59 +09:00

67 lines
4.4 KiB
Plaintext

# 주제: messaging-nats-experimental / messaging-pulsar-experimental 공통 측정
# revision: 21234e38cdb9a926cbc92bb97a2aee2e4a7d2916
# ---- (1) PreSendRejection 이 같은 클래스 두 벌이다 ----
# NatsPreSendRejection.java 65줄
# PulsarPreSendRejection.java 65줄
# command: 패키지·클래스명·도메인어 3개(publish/send operation, connection/producer closed,
# subject/topic)를 치환한 뒤 diff
# exit: 1 — 차이는 javadoc 한 문단의 **줄바꿈 위치 4줄뿐**이다. 문장 내용은 동일하다.
# 15,18c15,18
# < * <p>The OP throws this only when it knows the send never left the client: an
# < * argument the SDK refused, a CLOSED, a UNIT the client could not
# ---
# > * <p>The OP throws this only when it knows the send never left the client: an argument
# > * the SDK refused, a CLOSED, a UNIT the client could not resolve.
# => 필드·생성자 2개·requireCode·code() 가 전부 동일. 같은 클래스가 두 패키지에 있다.
#
# 두 javadoc 이 같은 과거 결함을 인용한다:
# "Classification used to work by reading the exception's class simple name and matching
# substrings — "Timeout" meant ambiguous, anything else meant rejected. … A class name is not
# part of any SDK's contract … And defaulting the unknown case to REJECTED tells the caller
# nothing was transmitted, so a caller that republishes under a new identity duplicates a
# message the server may well have stored."
#
# 같은 질문("이 쓰기가 서버에 닿았는가")에 대한 접근이 플랫폼에 네 가지 있다:
# messaging-kafka KafkaPublishFailureClassifier 허용목록 7종 + 모호 기본값
# messaging-rabbit RabbitPublishFailureClassifier confirm/return 상관 + 분류기
# messaging-nats NatsPreSendRejection 명시적 pre-send 예외만 REJECTED
# messaging-pulsar PulsarPreSendRejection 동
# 앞의 둘은 SDK 예외를 분류하고, 뒤의 둘은 "분류하지 않는다" 를 선택했다.
# 뒤의 둘의 판단이 더 보수적이며 그 근거가 명시적이다 — 다만 그 판단이 두 번 구현되어 있다.
# ---- (2) 선언 의존 7개 중 4개가 import 0건 (양쪽 동일) ----
# command: git grep -h "^import dev.caskeleton.messaging.<pkg>" -- <leaf>/src/main | wc -l
# nats pulsar
# messaging-core-api 17 18 O
# messaging-transport-spi 5 5 O
# messaging-policy 1 1 O
# messaging-schema-api 0 0 X
# messaging-observability 0 0 X
# messaging-security 0 0 X
# messaging-admin-api 0 0 X
# => 각 리프에서 7개 중 4개 미사용. 지금까지 본 messaging 리프 중 최대 비율이다.
# ---- (3) *AdapterContractTest 가 공유 계약을 상속하지 않는다 ----
# command: grep -n "class .*ContractTest|extends|@Nested" <각 테스트>
# NatsAdapterContractTest.java:49 class NatsAdapterContractTest { (extends 없음)
# PulsarAdapterContractTest.java:46 class PulsarAdapterContractTest { (extends 없음)
# 대조: KafkaProducerContractTest.java:17 class Contract extends MessagingAdapterContract {
# RabbitAdapterContractTest.java:17 class Contract extends MessagingAdapterContract {
# => 이름이 "AdapterContractTest" 이지만 messaging-testkit 의 공유 계약을 돌리지 않는다.
# 각자 자기 단언을 쓴다(NATS 17건 / Pulsar 14건).
# CompatibilityMatrix 는 두 리프 모두 runsSharedContract=false 로 둔다 — **정확하다.**
# 다만 클래스 이름이 반대 인상을 준다.
# ---- (4) 두 리프 모두 runtime_memberships:[] 이고 외부 import 0건 ----
# command: git grep -ln "import dev.caskeleton.messaging.nats" -- src | grep -v messaging-nats-experimental/
# exit: 1 (출력 없음). pulsar 도 동일.
# => 인큐베이팅 리프다. messaging-schema-avro / -protobuf / -kafka-share-experimental /
# -spring-cloud-stream-bridge 와 같은 상태이며, 이 둘은 그 상태가 등급으로도 선언되어 있다
# (CompatibilityMatrix Tier.EXPERIMENTAL, enabledByDefault=false).
# ---- (5) 두 전송 모두 3-arg/4-arg 생성자 쌍을 갖고 소비 불가를 타입으로 표현한다 ----
# NatsJetStreamTransport.java:85-89 NATS_CONSUMER_NOT_CONFIGURED
# PulsarMessagingTransport.java(동형) PULSAR_CONSUMER_NOT_CONFIGURED
# => Kafka·Rabbit 과 같은 패턴. 네 어댑터가 동일한 관용구를 쓴다.