Phase 23 closed as PARTIAL: fake coverage was gone, but the E2E layer itself
was still empty — no conftest.py anywhere, ~15 fixtures that never existed,
6 files crashing at import on os.environ["E2E_PASSWORD"], and CI never ran E2E.
STATUS's "10 E2E files / ~80+ cases ✅" described files that had never executed.
Infrastructure delivered:
- tests/e2e/conftest.py: ApiClient (trust_env=False — a host http_proxy turns
localhost calls into 502), operator_token(username) / new_user / admin_token
fixtures, and a session probe that calls pytest.exit(4) when the stack is
unreachable. NOT skip: skip would launder "did not run" into "passed".
- pytest.ini, tests/e2e/requirements.txt, and tests/e2e/sql/
seed_e2e_operators.sql — migrations contain NO operator seed at all, so a
fresh database has zero logins; the seed creates 6 e2e accounts with role
bindings, idempotent, password injected via `psql -v pwd`.
- CI job `e2e`: compose up -> poll /health/live (timeout = fail) -> seed ->
pytest -> always teardown. Gate D-23-D is now closed.
- 8 E2E files migrated off import-time env reads; tautological assertions
(`x == t or not x`), a counselor token standing in for content_editor, and a
bare `status_code == 200` claiming to prove INV-020 were replaced with real
permission/behavior assertions.
Six real production defects found only by running against real PostgreSQL
(H2 has zero coverage for all of them):
- D-23B-A (Critical): stale V042__add_appointment_operator_register_permission.sql
left in target/classes; src had replaced it with V042__add_service_location_
announcements.sql. Packing without clean put two V042 in the jar -> Flyway
"Found more than one migration with version 042" -> app would not start.
Guard added: testBuildOutputHasNoDuplicateMigrationVersions scans build output
(src is clean, so only this catches it).
- D-23B-B (High): non-UUID path variable reached the DB -> 500 + leaked
`invalid input syntax for type uuid`. Now rejected at the controller.
- D-23B-C (High): 6 E2E files raised KeyError at import -> never collected.
- D-23B-E (Critical): permissions = 48, role_permissions for super_admin = 18,
matchmaker = 0. V015's "grant all to super_admin" was a point-in-time snapshot;
the 30 permissions added later were never granted. On a fresh stack an admin
cannot create a service station and matchmakers can do nothing. V044 restores
super_admin to its documented definition (not an expansion); the matchmaker
permission set needs a product decision and is left as a residual.
- D-23B-I (Medium): IllegalArgumentException (e.g. invalid activity type) was
unmapped -> 500. Now BAD_REQUEST.
- D-23B-J (High): `operator does not exist: uuid = character varying` — Java
ids are String, columns are uuid, PostgreSQL does not implicitly cast in
comparisons. Fixed at the source with JDBC `stringtype=unspecified` (docker
profile); production datasource comes from deploy env and needs the same.
One judgement retracted: an initial read suggested service_appointment.register
was missing. It is not — the endpoint uses service.assisted_registration, seeded
in V029, and the code deliberately avoids a duplicate permission. Nothing was
changed on that basis.
Verification: 33 E2E passed against a live stack; unreachable stack and a
missing operator password both exit 4 (never skip/pass); backend regression
656 / 0 failures / 0 errors / 3 skipped (baseline 655, +1 guard, negatively
validated by injecting a duplicate V042 -> exactly one test turned red).
Phase 23 "可信系统测试重建 (REFACTOR)" re-audit. STATUS claimed "10 E2E
files / ~80+ cases" and "INV-011~020 ✅"; measurement showed the opposite.
D-23-A (Critical): V11InvariantsTest's 14 cases asserted only Constants
strings. INV-016 "Banner target types" actually asserted ARTICLE_STATUS_*;
INV-017 "home 8 entries" declared local arrays then asserted
firstRow.length == 4 (tautology, never called the home service);
INV-018/019/020 used appointment/article/account status constants as
stand-ins for permission and data-scope checks. Rewritten to 12 cases:
4 real service-behavior guards (INV-013 via real CounselingService.getDetail
— Matchmaker sees null userRequestSummary, with an assigned-Counselor
counter-proof so global sanitization cannot fake a pass; INV-016 reads the
production ALLOWED_TARGET_TYPES whitelist and asserts no external/http/url/
web entry; INV-017 calls real getHome and pins all 8 codes + order;
INV-019 asserts content_editor reads null), plus assertBehaviorTestExists
turning "行为验证见 XxxTest" pointers into executable Class.forName contracts.
D-23-B (High): 44 `pass` shells across 6 files in tests/e2e (4 files were
100% shells). Gate forbids skip/pass shells. Rewritten in place as honest
placeholders with the full scenario lists preserved.
D-23-C (High, residual): the pytest E2E suite is not executable at all —
no conftest.py anywhere, ~15 fixtures (api_client/verified_user/
banner_factory/...) do not exist. Only run_v11_tests.py really runs
(16 cases, needs a live stack + E2E_PASSWORD).
D-23-D (High, residual, NOT fixed): CI does not chain E2E as the plan
requires. Needs an infrastructure decision on how CI provisions a live
stack; not implemented rather than written unverified.
D-23-E (Medium): the CI migration job claimed to "验证 migration SQL 语法"
but only checked file non-emptiness, while the real syntax/baseline checks
live in MigrationValidationTest (already run by the backend job). Replaced
with a real sequence check (naming / non-empty / unique versions / no
gaps); verified by extracting the exact inline script and running it.
New gate: scripts/check-no-empty-tests.py (AST scan, blocks pass-only and
unconditionally skipped tests) + a CI test-shell-guard job. Verified both
ways: 0 findings on the real tree, 2 precise findings on injected shells.
Negative validation: 3 injected production defects (external_url in the
banner whitelist, swapped home entry order, isMatchmaker returning the full
record) turned exactly inv016/inv017/inv013 red, 9 others green, zero
unrelated failures; md5sum -c confirmed byte-identical restore.
Full regression: 655 tests / 652 passed / 0 failures / 0 errors / 3 skipped
(baseline 657; -2 because 14 vacuous cases were merged into 12). The 3
skips carry explicit reasons (2x @EnabledIfEnvironmentVariable(CI=true),
1x @Disabled "需要完整数据库连接") — none is a shell.
Phase 23 is PARTIALLY closed: fake coverage is gone, but the UAT gate
requires the end-to-end chain to actually execute, so "失败为零可进入 UAT"
is NOT claimed.
WorkBuddy
admin