test(p02): document reproduced durability/replay risk, no runtime patch [skip ci]
12 KiB
12 KiB
DataStack — Phase Status & Change Log (Source of Progress Truth)
Frozen source baseline: d725841ebc828d58cbe90f517da9e600f23c4630. Generated: 2026-11-22 Asia/Dhaka.
Scope: documentation-only kickoff. Any future phase must update this log after verification, not on intention.
Completed implementation phases: 1 / 3. Remaining: 2. Documentation baseline preparation is a separate non-implementation milestone.
| Phase | Status | IDs / function | Changed source | Verified CI evidence | Commit / remote sync |
|---|---|---|---|---|---|
| P00 Notes / freeze | COMPLETED (documentation only) | 001–006 docs, source baseline and exact scope |
notes/** only |
6 Markdown files; balanced fences; staged diff --check PASS |
Documentation commit containing this file; remote SHA verification recorded in session handoff |
| P01 Tests & lockfile | CLOSED — VERIFIED | CI-01…CI-05, no runtime feature changes | 16 exact P01 code/test/CI paths, plus notes/001–006 records |
#59 Linux/Windows clean npm/SDK, 24/24 TS suites, Next standalone PASS; #60 strict TypeScript and ESLint PASS; historical #59 quality mismatch corrected | One permanent P01 main commit containing these notes; confirm remote SHA/tree parity in phase handoff |
| P02 Ingestion reliability | BLOCKED — three diagnostic cases reproduced; runtime fix requires separate approval | R-01 ACK/RAM loss; R-02 CSV duplicate replay; R-03 health liveness vs readiness | Only tests/phase02-sync.test.ts and tests/phase03-storage-buffer.test.ts, plus notes/001–006 |
Gitea #63: all three P02 subtests PASS to reproduce existing behavior; full matrix status recorded in appendix | Diagnostic branch p02-readiness-test-20261122, no production runtime changes |
| P03 CI/Docker/security | NOT STARTED | CI-06,07,08; release gates | None | Not run | — |
Existing diagnostic CI snapshot (NOT implementation phase completions)
| Run | Baseline commit | Result | Meaning |
|---|---|---|---|
| #42 | f7c1e7f |
Cancelled | Windows original lock fail logged; long combined regression stopped |
| #45 | d3a93c6 |
Failure | TS 18 clean/6 fail-or-timeout; 43 type errors; build and Windows SDK pass; Python apt / Docker CLI blockers |
| #48 | ff1333d |
Success | Python SDK 14/14 |
| #49 | 559f180 |
Success | ESLint |
| #50 | d725841 |
Failure | Compose + ingestion image pass; web image fails containerd snapshot copy; proxy unrun |
Historical baseline code stayed frozen through these CI workflow changes. Current HEAD before documentation changes = d725841ebc828d58cbe90f517da9e600f23c4630 on local/GitHub/GitLab/Gitea.
Per-phase closure record template — COPY for each P01/P02/P03
P0X — [TITLE] (OPEN / BLOCKED / CLOSED)
- Authorized baseline / parent SHA:
- Approval command / exact scope document reference:
- Findings addressed + any conditional gates:
- Exact modified paths (one path per row):
- Applied code patch summary (function + lines):
- Existing feature/UI/API/auth/SDK/data preservation audit:
- Focused CI run URL / test counts / exit codes:
- Security check / handling of secrets, data and scope exceptions:
- Verification result: PASS / FAIL / BLOCKED, not assumed:
- Closure commit SHA + GitHub/GitLab/Gitea main SHAs:
- Completed implementation phases: X/3; remaining: Y.
- Next phase: P0Y; specific approved features / allowed files:
- New issues / deferred risks / notes evidence links:
- MEMORY.md cross-IDE handoff text (copyable):
[DataStack P0X] Baseline=<SHA>; findings=<IDs>; scope=<exact paths>; CI=<run + exit>; commit=<SHA>; remote sync=<all 3>; phase=CLOSED; next=P0Y; production deployment=unchanged.
P01 attempted implementation & gate outcome (2026-11-22)
- Frozen source baseline remains
d725841ebc828d58cbe90f517da9e600f23c4630; latest source HEAD before P01 was documentation commitd8ac96e39dad4ea3b241c1d2f16f959c5215701eon all three remotes. - Disposable CI commit
90fce0629a752dd2916d8d3a2afeaf2c52ed6cbbcontains 16 P01-allowed paths:bun.lockdeletion,package-lock.jsonoptional/native metadata closure (+67 entries, no existing version or root-manifest change), 14 scoped test files, and.gitea/workflows/datastack-ci.ymlclean-install/phase-filter changes. No application runtime source/UI/SDK/Docker changes. - Gitea run: https://dev.ivib.cc/vibtools/DataStack/actions/runs/58; overall FAIL, terminal.
windows-compatibility,node-regression,quality,production-buildall failed during cleannpm ci:Missing: @tybys/wasm-util@0.10.5 from lock file. The Python and Docker jobs were intentionally skipped for P01. No current TypeScript test, strict typecheck, ESLint or build PASS may be claimed. - Root issue likely parallel incompatible transitive requirement of
@tybys/wasm-util; inspectnode_modules/@napi-rs/wasm-runtimevsnode_modules/@tailwindcss/oxide-wasm32-wasiin lock and install exact requested transitive version under correct nested path without changing existing pinned entries. Then verify npm clean-install on both platforms and rerun focused P01 acceptance once after repair; do not blindly override lock integrity. - Local original Windows device's VARUA connection reached monthly call quota; local changes were staged via safe squash on
mainbut no P01 main commit/push occurred. A separate connected desktop did not have this workspace. Do not claim local/3-remote post-P01 sync until verified. - Historical #58 status: 0/3, P01 BLOCKED before nested lockfile correction. Superseded by closure evidence below; no force push or deploy was used.
Status transition rules
- OPEN -> IN PROGRESS only with explicit approval and baseline SHA match.
- IN PROGRESS -> CLOSED only after one focused CI and exact diff audit, updated
003+005+ relevant001/004/006, Git commit, sync, and confirmed remote SHA. - Failed test -> BLOCKED with log excerpt and scope decision; never log as complete.
- Unexpected file change -> STOP and revert only own documented changes safely; never touch pre-existing work.
- At any time,
actual production deployment statusis a separate tracked dimension (do not infer deployed SHA from Git HEAD).
P01 Final Closure — CI-01…CI-05
- Status: CLOSED after one P01 acceptance CI (#59) plus one narrow quality-only correction/retry (#60); completed implementation phases 1/3, remaining 2 (P02 and P03).
- Baseline: frozen source
d725841ebc828d58cbe90f517da9e600f23c4630; pre-P01 documentation HEADd8ac96e39dad4ea3b241c1d2f16f959c5215701e. - Validated implementation code snapshot:
7d8eebd3dc867e82113343403b63ce1d4ed20584on disposable Gitea P01 branch. Permanent main commit is the single squash commit containing this closure log; remote SHAs and content-tree equality must be checked at publish, not predicted here. - Exact P01 patch: deletion
bun.lock; append-only 68 optional/native lock entries with zero original package-version changes; 13 in-scope test files with typed mocks, teardown, and shared-helper assertions; narrow.gitea/workflows/datastack-ci.ymlclean-install and focused verification inputs.package.json,tsconfig.json, all app/UI/backend/SDK/Docker/production files unchanged. - #58 failed before tests (
@tybys/wasm-util@0.10.5missing); fixed by nested, registry-hash-verified optional dependency. #59 overall FAIL due exactly one TS2339 test teardown type declaration; the same run separately proved 24/24 TS suite exits 0, Linux/Windows original committed-locknpm ci, both Node SDK builds, and Next standalone production build (22/22 pages). Fixed only the test declaration; #60 quality-only PASS fornpx tsc --noEmitandnpm run lintwith no error. - Prior independent Python SDK 14/14 PASS: #48 (SDK unmodified); not rerun during P01. P03 Docker and security/advisory findings unresolved; no Docker builds, live ingestion, network fault injection, migrations, releases or production deployment occurred.
- Diff audit: only allowlisted 16 P01 code/test/CI paths and 6
notes/*.mdfiles;git diff --checkPASS; no unrelated features, API behavior, UI or persistent data changed. 1293 original lock entries remained byte-equivalent; 68 optional entries added with integrity hashes. - Working tree caveat: original local Windows VARUA device reached its monthly quota during P01; its current local branch/working tree could not be rechecked and should be reconciled against published
mainbefore new local work. A separate authorized temporary Windows clone was used for diff and Git push verification. Never claim original PC is already updated. - Next: P02 read-only initial source audit plus targeted disposable fault-injection tests for R-01 (Redis ACK), R-02 (CSV replay), R-03 (health readiness). Runtime-source change requires reproduced issue and a separate exact method/file approval. P03 only after P02 closure.
- P02 implementation approval command:
APPROVED — DataStack P02. Implement only notes/002 P02 diagnostic/test write scope for R-01, R-02 and R-03. Keep app/UI/API/auth/SDK/runtime/data unchanged. Runtime files read-only unless a proven failure and separately approved narrow code patch. One focused disposable Gitea CI, exact diff, notes/001–006 update, commit and remote verification. No production deployment or scope expansion.
P02 Execution Log — 2026-11-22 (supersedes P02: pending until acceptance)
- Status: IN PROGRESS / diagnostic CI pending. Not CLOSED and not a production-readiness certificate.
- P01 published baseline
667a140c4e9e75382bfdcdbe1d3e9eb96f6e808a. Original PC localmainstill at prior docs HEAD with 16 user-visible staged P01 paths, preserved untouched. New isolatedp02-readiness-test-20261122worktree based on published baseline; no reset/stash/force update. - P02 disposable test commit:
bbe73d91666641743cbec6cf7203f1694094caef; only two files:tests/phase02-sync.test.ts(Redis negative-path ACK and health fallback),tests/phase03-storage-buffer.test.ts(failure between CSV append and Redis LTRIM, replay). Source contracts and auth flows unchanged. - Focused evidence: https://dev.ivib.cc/vibtools/DataStack/actions/runs/63. Record terminal suite/typecheck/build results after completion. Test assertions expose existing behavior; they do not implement durability or readiness fixes.
- Scope audited:
server/index.js,server/socket/handlers.js,server/workers/csvWriter.js,core/data/buffer.ts,app/api/health/route.tsall READ-ONLY;app/**,components/**,lib/**,sdk/**, Docker, settings, production data and deployments unchanged. - P01 remains CLOSED (1/3). If R-01 or R-02 are confirmed unsafe, P02 must remain BLOCKED for runtime remediation approval. No P03 implementation on this approval.
P02 Verified Diagnostic Gate (R-01/R-02/R-03)
- Gitea #63 (test-only commit
bbe73d91666641743cbec6cf7203f1694094caef):P02 R-01 diagnosticPASS,P02 R-03 diagnosticPASS,P02 R-02 diagnosticPASS. Each PASS validates reproduction of existing behavior, not a successful reliability remediation. tests/phase02-sync.test.tssuite exit 0;tests/phase03-storage-buffer.test.tssuite exit 0. Other CI jobs may have independent status; do not infer full job or complete workflow PASS from three cases.- BLOCKING RELEASE FINDINGS: R-01 success ACK issued for non-durable RAM fallback when Redis RPUSH fails; R-02 identical eventId can append twice on post-CSV Redis-trim failure (web-side representative test). R-03 health remains online on Redis loss; liveness-only behavior is backward-compatible, protected dependency-readiness verification remains missing.
- P02 PHASE = BLOCKED, diagnostic implementation done, release reliability acceptance NOT satisfied. Runtime remediation requires separate exact approval. 1/3 phases still CLOSED (P01 only); P02 and P03 remain unclosed. No P03 start authorization from this diagnostic completion.
- Next approval must explicitly authorize exact narrow server/core patch for ACK durability and CSV replay deduplication, plus concrete no-loss/no-duplicate tests. Any readiness endpoint change needs separate explicit approval. User's initial P02 approval is insufficient for runtime mutation.