mirror of
https://github.com/paperclipai/paperclip.git
synced 2026-11-09 07:10:34 +01:00
Verify default skill sync through two real cold sleep cycles (#19191)
## Thinking Path > - Paperclip is the open source app people use to manage AI agents for work. > - Automatic skill sources check upstream every five minutes. > - The domain deadline contract can hand future work to a durable wake alarm. > - A saved deadline or a worker unit test does not prove that a stopped tenant resumes the real producer. > - This pull request checks the default cadence through the actual app and a matching service checkout. > - It checks two cold starts and three distinct persisted skill revisions with real clocks. > - The benefit is repeatable local evidence without changing production timing or authority. ## Linked Issues or Issue Description Refs: #19001, #17008, #17923. This is a test and documentation follow-up to #19001. #19001 is now merged; this PR targets master. Feature code remains in that merged foundation. #17008 covers the earlier source worker change. #17923 covers separate skill-cache retention; this PR does not change that behavior. A scoped search of public PRs and issues found no duplicate default-cadence cold-cycle acceptance test. No matching public issue was found. The enhancement fields follow. **Pre-submission checklist** This improves checks for an existing feature. It adds no product endpoint or execution authority. The public related-work search is complete. **What existing behavior does this improve?** Verification of automatic skill-source updates across a durable idle handoff and cold restart. **Subsystem affected** `server/` test fixtures and development documentation. **Current behavior** Domain tests cover claims, access and future deadlines. They do not prove the full default-300-second path from ordinary source attachment through signed tenant reports, durable alarms, stopped compute and a changed persisted skill. **Proposed behavior** An opt-in launcher checks bootstrap first, then two full real-clock cycles. Ordinary CI can run the transport controls without the host fixture. The launcher-specific worker assertion runs only when prewire is requested, and fails closed in that lane. **Reason and benefit** The test can detect an early refresh, a lost deadline, a stale proof, an ACK after physical stop, a failed cold restart or a producer that does not publish changed contents. **Breaking changes** None to product APIs, schema, cadence, authorization, claims or retry policy. Existing maintenance-helper defaults stay unchanged. **Additional context** The opt-in host flow needs an authorized matching service checkout, its normal compiled acceptance bridge, Node 22 and two owned PostgreSQL test databases. The standalone controls do not need that service or PostgreSQL. Roadmap review places this under existing Skills Manager and Cloud deployment verification, not a new core feature. ## What Changed - Add an opt-in default-cadence acceptance test and checked-in launcher. - Reuse real server startup, managed tenant authority and the ordinary project attachment route. Control only the remote Git snapshot boundary; keep actual authorization, scanning, claims and publication. - Use actual signed tenant routes and the matching service's gate, sleep controller and PostgreSQL wake-alarm store. Stop and restart real app children through a controlled local provider. - Preload the Node transport guard before coordinator, worker and child module work. Check worker-local and child PID, thread and source-hash receipts. Keep caught unexpected transport attempts fatal. - Require both persisted refresh times and actual upstream handler timestamps to be at or after each original deadline. - Join physical callbacks during bridge close, stop a late wake-created child, and preserve the original cleanup budget. Add a real IPC/HTTP-child counterfactual. - Accept valid owned HTTP options paths without allowing Unix sockets or unknown origins. Remove the unused App helper export. - Add finite transport and launcher controls. Defer only the launcher-specific parent assertion in vanilla CI. Keep all subprocess controls default-runnable. - Add optional child entry, preload and bounded pool settings to the existing test helper. Preserve its defaults and the real observer test body. - Remove added explanatory comments under the contribution policy. Preserve executable syntax, directives and literals. - Document pinned setup, fresh owned databases, bootstrap and full two-cycle commands in `doc/DEVELOPING.md`. The diff against the intended base contains nine test/helper/launcher files and this one document. It adds no production feature code or Cloud source to the App repository. ## Verification The latest two-file CI repair preserves both original failed jobs. A newer master workflow introduced two required step inputs absent from its extracted-step test environment; the faithful false/false inputs now pass all five existing target-authorization cases, after an actual missing-input counterfactual failed. A controlled socket case also reproduces legal reuse of the released runner port; the fixture now holds that reservation until the preview binds, releases it in finally, and asserts distinct ports. All original retry/result/accounting assertions and timeouts stay unchanged. The original Linux listener owner remains unknown; the control proves this fixture collision mechanism, not that original attribution. Fresh current-head full CI and complete reviews are pending. Current merge-integration candidate: `f6f686103b7bfd0a4ab83085634c622af3896ecb`, a normal merge after #19001 landed, including actual master `1c9767d323cf872a266962b9e6b4bd5cbcb57f1c`. All nine acceptance code files are byte-identical to148a82e5; both parents' documentation additions survive. Every other path equals actual master. The only conflict was the old foundation's startup block; current master was retained in full. Fresh exact-head CI and complete reviews are pending. The local measurements below belong to their explicitly named revisions; no fresh full614-second run is claimed. Current candidate: `148a82e54c502759785407357001c6b50c1439dd`, normally carrying the current base `8c912e1ba4fc66397a7341ade59373af8075068f`. All nine acceptance code blobs remain identical to qualified repair `e092746f4da181c8414c63be71474619a84b97e7`; documentation retains both parents. The matching built service for the fresh bootstrap is `389878364b42dd129835d263f91882c5774d5123`, using Node 22.23.4. The App uses Node 24.22.2 and pnpm 9.15.4. Use the exact controlled setup and commands in [doc/DEVELOPING.md](doc/DEVELOPING.md#controlled-default-cadence-sleep-acceptance). Fresh owned databases and the matching compiled service are prerequisites; the launcher does not create a cluster or build dependencies. - **Current affected semantic check passed**, using actual native TypeScript 7 and the current generated CI options/references (0.640 seconds). Normal declaration references passed in 11.726 seconds, with source-matched SDK/shared/kernel/Runner prerequisites and the frozen current lockfile. - **Current guarded controls: 44 passed**, plus the same single explicit real-PG observer deferral, in 4.790 seconds. The executing worker checked its own global PID/thread/preload source hash and matching receipt. Guard hash: `7daf9ae7d2f26e8fcd389b1d560fac1d4e218a8304a2c58e49f646af108165bc`. No unexpected attempts or PG sockets occurred. Vanilla controls on the identical repair files passed 22 cases with one explicit launcher-parent assertion deferral (2.044 seconds). - **Real cleanup counterfactual:** unchanged old helper failed the actual endpoint assertion in 1.568 seconds: its late HTTP child remained reachable after close returned. The same fixed-helper case passed in 1.485 seconds: close waited, the child exited, the port was unreachable and no IPC error occurred. The isolated driver records only the known legacy closed-IPC error long enough to inspect cleanup; current helper requires zero such errors. All recorded processes and ports were independently checked empty. - **Fresh current bootstrap passed once**, 16.159 seconds wall / 14.12 seconds Vitest / 12.237 seconds body. The ordinary enabled producer persisted useful revision 1, cleared its lease and retained exactly 300,000 ms between success and original next deadline. A signed current-generation report produced the matching generation/operation-lock SQL ACK before physical child stop (ACK 06:54:02.020 UTC; stop signal 06:54:02.025 UTC). The alarm retained the original D and D-120 eligibility. Both fixture databases ended with zero sessions; the coordinator independently audited and normally removed them. This bootstrap does not execute a due wake or the new v2/v3 timing assertions. - **Historical full two-cycle result remains 1/1 passed in 614.213 seconds wall**, at measured App `74f7c012ca17ad0d7302d76403f5033affb1430b` and service `2cb54f904031dcf69a32902cd853b43f74d6f301`. Two new processes published distinct immutable v2/v3 at original D+115 ms and D+98 ms. Exact signed SQL ACKs preceded both physical stops; real SQL D-120 claims caused both cold starts. Features, claims and original business deadlines stayed unchanged. - Both historical 10.002-second quiet windows kept app SQL, upstream and outgoing HTTP counts unchanged. Two awake PG sockets remained. Quiet counts do not mean all connections closed. - The stronger timing checks are additional assertions, not retroactively executed results. Offline inspection of the original receipts confirms persisted successes at D+115/+98 ms and earliest upstream handler entries at D+44/+30 ms; all five requests in each refresh group started after D. The exact new pre-sleep request counter and dispatch-before-handler time were not independently recorded historically. - Source review found 19 critical automatic Git producer/runtime files identical to the measured App. The current active cache helper change does not run on automatic Git publication and adds no pool. The matching service carries unchanged measured helper/sleep/controller/alarm seams through its normal master merge. The current bootstrap and focused controls qualify the changed cleanup/guard and parent integration. No fresh current-head 614-second pass or whole-tree identity is claimed. - Runtime is serialized: worker 1, per-pool limits at most 4, configured App potential 10 + shared service pool 2 + one observer against server max 40. Current bootstrap point samples observed at most 3 owned clients plus the observer; those samples are not a continuous aggregate peak. App/service guard peaks were 2 each, with zero remaining sockets at cleanup. Failures remain accounted for. The first historical bootstrap cleanup masked an unknown primary error. Later attempts exposed missing fixture-owned domain metadata and lost origin snapshots; test-only ownership/cleanup corrections preserved the original assertions and timeouts. The new causal driver first lacked a final observation, then encountered sandbox bind EPERM, then exposed the old closed-IPC error; only the final real endpoint failure above is labeled the cleanup counterfactual. An offline dependency miss and a generator-cleared ignored compiler config were retained and corrected without weakening compiler options. The former inherited policy failure (397 findings) is resolved by the genuine current-master base/scanner, with zero findings and seven scanner tests; no unrelated UI site was edited and the gate was not relaxed. Historical real TCP observer coverage passed 1/1 in 3.688 seconds. Comment-only and opt-in-selection carries remain documented in prior evidence; self-hash changes are metadata exceptions with fresh executing-worker checks. The full workspace typecheck, test, build and browser commands were not run locally for this test-only slice. The current base passes its full exact-head CI and Greptile 5/5; this leaf now has all 56 exact-head check runs terminal (54 success, one neutral, one skip), the additional legacy security status passing, complete current Greptile 5/5 and Ponytail 5/5 reviews, and all four review threads resolved across all pages. No live provider or deployed host/database suspension was exercised. ## Risks - The host test requires a matching service checkout and compiled bridge. An unrelated service revision is not a valid substitute. - Controlled localhost Git, S3 and physical process operations do not qualify deployed Railway or Neon suspension, external permissions or provider uncertainty. - The Node guard does not prove native subprocess transport isolation. Only finite local Git metadata and Node version commands were needed. - Two short SQL/HTTP quiet windows do not establish strict 90-second host acceptance. Separate plugin-host LISTEN, MCP wake and other domain gaps remain open. - Explicit test URLs are destructive fixture inputs. Use fresh, owned `_test` databases. The launcher does not start a cluster or choose a database implicitly. - Passing CI and review do not qualify deployed suspension; the current proof boundaries above still apply. > For core feature work, check `ROADMAP.md` first and discuss it in `#dev` before opening the PR. This PR adds verification to the existing feature in its base. ## Model Used OpenAI `gpt-6.1-sol` assisted with implementation, repository review and local code execution. OpenAI Codex, based on GPT-6, performed managing review and integration. The context-window sizes and exact served backend snapshots were not exposed. ## Checklist - [x] I have included a thinking path that traces from project context to this change - [x] I have specified the model used (with version and capability details) - [x] I have checked ROADMAP.md and confirmed this PR does not duplicate planned core work - [x] I have searched GitHub for duplicate or related PRs and linked them above - [x] I have either (a) linked existing issues with `Fixes: #` / `Closes #` / `Refs #` OR (b) described the issue in-PR following the relevant issue template - [x] I have not referenced internal/instance-local Paperclip issues or links (only public GitHub `#NNN` / `github.com/paperclipai/paperclip` URLs) - [x] My branch name describes the change and contains no internal Paperclip ticket id - [x] I have run tests locally and they pass - [x] I have added or updated tests where applicable - [x] I have updated relevant documentation to reflect my changes - [x] I have considered and documented any risks above - [ ] All Paperclip CI gates are green — fresh merge-head gates pending - [ ] Greptile is 5/5 with no open P2s, recommendations, or follow-ups — fresh merge-head review pending - [x] I will address all Greptile and reviewer comments before requesting merge ### Current master reconciliation (2026-11-06) Merged actual master `e28158b353fd274f0322be76d62cd223e0df599f` normally. All nine qualified acceptance code files and the held-port repair remain byte-identical. Upstream independently landed the workflow input fix in #19602; the fixture now uses that exact upstream version once, avoiding duplicate object keys in a squash merge. All other master changes are preserved. The earlier two-cycle local evidence remains source-specific; no fresh 614-second run or deployed suspension is claimed. This new head requires fresh CI and bot reviews. ### Actual combined-tree typecheck failure and repair CI job `128156887548` failed with TS1117 at workflow-target.test.ts:22: two independently landed master fixes declared the same `IMAGE_ONLY` and `TESTOPS_REQUESTED` keys. The complete original log is retained. A further normal master merge now removes the duplicate declarations, retaining exactly one false/false pair and every original assertion. All nine qualified acceptance code paths and the held-port repair remain byte-identical; unrelated master changes are preserved. This is a concrete compiler failure and source repair, with fresh exact-head CI/reviews required. --------- Co-authored-by: Paperclip <noreply@paperclip.ing>
This commit is contained in:
DottaandPaperclip authored and GitHub committed
1 parent
e482f0f632
commit
44c60c7221
11 files changed
+792
-8
No files matched your search
@@ -4744,6 +4744,102 @@ unavailable. Use a separate, disposable database whose name ends in `_test` for
|
||||
`PAPERCLIP_SKILL_SOURCES_TEST_DATABASE_URL`. The suites apply migrations and create
|
||||
test data; the synchronization suite truncates company data between tests.
|
||||
|
||||
#### Controlled default-cadence sleep acceptance
|
||||
|
||||
The opt-in `domain_deadlines_v1` negotiation adds owner-bound skill deadline
|
||||
handoff coverage. Legacy callers retain the conservative blockers described above.
|
||||
The checked-in `scripts/test-skill-default-cadence.mjs` launcher checks this local
|
||||
flow with the actual app, signed tenant routes, Cloud controller and PostgreSQL
|
||||
alarm store. Git contents, S3 and physical child stop/start use controlled local
|
||||
providers. A matching service checkout with
|
||||
`src/sleep/fixtures/default-cadence-bridge.ts` and its normal compiled output is
|
||||
required for the host flow. The PG0 controls do not require that checkout or a DB.
|
||||
The launcher does not start a cluster, build artifacts, or contact live providers.
|
||||
|
||||
Use Node **24.22.2**, pnpm **9.15.4**, and Node **22.23.4** for the matching service.
|
||||
Prepare the server dependency closure with a frozen lockfile. Build only missing
|
||||
TypeScript prerequisites explicitly; the launcher does not install or build them.
|
||||
From the app repository root:
|
||||
|
||||
```sh
|
||||
export NODE24=/absolute/path/to/node-v24.22.2/bin/node
|
||||
export PNPM9=/absolute/path/to/pnpm-9.15.4/bin/pnpm
|
||||
export PATH="$(dirname "$NODE24"):$(dirname "$PNPM9"):$PATH"
|
||||
"$PNPM9" --filter @paperclipai/server... install --frozen-lockfile --ignore-scripts
|
||||
for package in packages/shared packages/plugins/sdk packages/paperclip-eval-kernel packages/paperclip-runner; do
|
||||
(cd "$package" && "$NODE24" node_modules/typescript/bin/tsc -p tsconfig.json)
|
||||
done
|
||||
(cd packages/paperclip-runner && "$NODE24" node_modules/typescript/bin/tsc -p tsconfig.surfaces.json)
|
||||
export SKILL_SLEEP_RESULTS="$(mktemp -d "${TMPDIR:-/tmp}/paperclip-skill-sleep.XXXXXX")"
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG="$SKILL_SLEEP_RESULTS/controls-guard.jsonl"
|
||||
"$NODE24" scripts/test-skill-default-cadence.mjs --controls
|
||||
```
|
||||
|
||||
This PG0 lane runs 44 controls. It explicitly defers the single real-PostgreSQL
|
||||
observer case before its support probe. It requires the actual worker's preloaded
|
||||
guard marker and matching receipt. Ordinary Vitest runs keep all subprocess
|
||||
controls and defer only that launcher-specific parent assertion. An explicit
|
||||
prewire mode or guard-log setting makes the unchanged assertion mandatory; a
|
||||
missing or invalid marker fails. The launcher never installs the guard late.
|
||||
|
||||
For host acceptance, separately create empty, owned databases on an existing local
|
||||
PostgreSQL 17 server. Use distinct names ending in `_test`. The test applies normal
|
||||
migrations and writes fixtures; it never starts an embedded cluster. Use a fresh
|
||||
pair for each attempt, including the full run after bootstrap. Prepare the matching
|
||||
service dependencies separately and compile its normal `tsconfig.build.json` with
|
||||
Node 22. The recorded local runtime used service source
|
||||
`2cb54f904031dcf69a32902cd853b43f74d6f301`. Current service publication
|
||||
`389878364b42dd129835d263f91882c5774d5123` carries that measured helper source through
|
||||
its normal upstream merge; the runtime was not repeated on that service merge.
|
||||
An arbitrary service revision is not an interchangeable prerequisite.
|
||||
|
||||
```sh
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_CHECKOUT=/absolute/path/to/matching-service
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_NODE=/absolute/path/to/node-v22.23.4/bin/node
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_SHA=389878364b42dd129835d263f91882c5774d5123
|
||||
test "$(git -C "$PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_CHECKOUT" rev-parse HEAD)" = "$PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_SHA"
|
||||
(cd "$PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_CHECKOUT" && "$PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_NODE" node_modules/typescript/bin/tsc -p tsconfig.build.json)
|
||||
export PAPERCLIP_TEST_POSTGRES_ADMIN_URL=postgres://paperclip@127.0.0.1:55448/postgres
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_TEST_DATABASE_URL=postgres://paperclip@127.0.0.1:55448/skill_sleep_bootstrap_app_test
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_TEST_DATABASE_URL=postgres://paperclip@127.0.0.1:55448/skill_sleep_bootstrap_cloud_test
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_EVIDENCE="$SKILL_SLEEP_RESULTS/bootstrap"
|
||||
mkdir -p "$PAPERCLIP_SKILL_DEFAULT_CADENCE_EVIDENCE"
|
||||
unset PAPERCLIP_MAINTENANCE_PG0_CONTROLS
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_STAGE=bootstrap
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG="$SKILL_SLEEP_RESULTS/bootstrap-guard.jsonl"
|
||||
"$NODE24" scripts/test-skill-default-cadence.mjs
|
||||
```
|
||||
|
||||
Bootstrap checks revision 1, the legacy refusal, the signed bounded handoff and the
|
||||
exact durable ACK before child stop. It does not check a due wake. For the complete
|
||||
run, first create a different empty pair, then run:
|
||||
|
||||
```sh
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_TEST_DATABASE_URL=postgres://paperclip@127.0.0.1:55448/skill_sleep_cycles_app_test
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_TEST_DATABASE_URL=postgres://paperclip@127.0.0.1:55448/skill_sleep_cycles_cloud_test
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_EVIDENCE="$SKILL_SLEEP_RESULTS/two-cycle"
|
||||
mkdir -p "$PAPERCLIP_SKILL_DEFAULT_CADENCE_EVIDENCE"
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_STAGE=two-cycle
|
||||
export PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG="$SKILL_SLEEP_RESULTS/two-cycle-guard.jsonl"
|
||||
"$NODE24" scripts/test-skill-default-cadence.mjs
|
||||
```
|
||||
|
||||
Use a new guard-log path for each attempt. The full mode takes about 10-14 minutes.
|
||||
It retains the original 300-second source cadence, real app/SQL clocks, D-120 alarm
|
||||
eligibility and the existing 850-second test timeout. Do not shorten the cadence,
|
||||
write deadlines, or substitute a fake clock. The launcher uses one worker and no
|
||||
file parallelism. Budget runtime pools for the app's potential 10 connections plus
|
||||
Cloud's shared 2 and one transient observer; each pool is at most 4. The recorded
|
||||
server limit was 40. Run this fixture without other heavy tests or builds.
|
||||
|
||||
Receipts retain both original deadlines, durable alarm generations, distinct cold
|
||||
processes and persisted revisions 2 and 3. Each cold start also checks a measured
|
||||
10-second app SQL/upstream/outgoing-HTTP quiet window. Awake guards can still show
|
||||
2 PostgreSQL sockets: quiet SQL does not mean all connections are closed. This is
|
||||
controlled local acceptance, not strict 90-second host acceptance, live-provider
|
||||
qualification or proof of deployed Railway/Neon suspension. Keep failed attempts
|
||||
and audit owned sessions/processes before database cleanup.
|
||||
|
||||
### Public MCP and Dot event delivery
|
||||
|
||||
Public MCP task events and Dot mailbox events use separate topics in the shared
|
||||
|
||||
@@ -0,0 +1,97 @@
|
||||
|
||||
const fs = require('node:fs'), path = require('node:path'), os = require('node:os');
|
||||
const net = require('node:net'), tls = require('node:tls'), http = require('node:http'), https = require('node:https'), dns = require('node:dns');
|
||||
const { createHash } = require('node:crypto');
|
||||
const { threadId, isMainThread } = require('node:worker_threads');
|
||||
const key = Symbol.for('paperclip.skill.default-cadence.prewire');
|
||||
const log = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG;
|
||||
if (!log) throw new Error('Default cadence preload requires an explicit guard log');
|
||||
if (globalThis[key]) { if (globalThis[key].log !== log) throw new Error('Default cadence guard log changed'); return; }
|
||||
const loopback = new Set(['127.0.0.1', 'localhost', '::1', '[::1]']);
|
||||
const listeners = new Set(), endpoints = new Set(), pgEndpoints = new Set(), attempts = [], active = new Set();
|
||||
const tracked = new WeakSet(); let peak = 0;
|
||||
const identity = 'skill-default-cadence-guard-v1';
|
||||
const sourceHash = createHash('sha256').update(fs.readFileSync(__filename)).digest('hex');
|
||||
const record = (event, value = {}) => fs.appendFileSync(log, JSON.stringify({ event, identity, sourceHash, pid: process.pid, ppid: process.ppid, threadId, isMainThread, uid: process.getuid?.(), at: Date.now(), ...value }) + '\n');
|
||||
const deny = reason => { attempts.push(reason); process.exitCode = 98; record('blocked', { reason }); throw new Error(`Default cadence guard blocked ${reason} before transport`); };
|
||||
const local = host => { if (!loopback.has(String(host).toLowerCase())) deny(`offhost ${host}`); };
|
||||
function allowOrigin(value) {
|
||||
const url = new URL(value); local(url.hostname);
|
||||
if (!['http:', 'https:'].includes(url.protocol) || url.username || url.password || url.search || url.hash || url.pathname !== '/') deny('invalid owned HTTP origin');
|
||||
endpoints.add(`${url.hostname}:${url.port || (url.protocol === 'https:' ? 443 : 80)}`);
|
||||
}
|
||||
function allowDatabase(value) {
|
||||
const url = new URL(value); local(url.hostname);
|
||||
if (!['postgres:', 'postgresql:'].includes(url.protocol) || !/^[a-zA-Z0-9_]+_test$/.test(decodeURIComponent(url.pathname.slice(1)))
|
||||
|| ['options', 'database', 'dbname'].some(k => url.searchParams.has(k))) deny('invalid dedicated database');
|
||||
const endpoint = `${url.hostname}:${url.port || 5432}`; endpoints.add(endpoint); pgEndpoints.add(endpoint);
|
||||
}
|
||||
for (const name of ['PAPERCLIP_SKILL_DEFAULT_CADENCE_TEST_DATABASE_URL', 'PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_TEST_DATABASE_URL', 'DATABASE_URL']) if (process.env[name]) allowDatabase(process.env[name]);
|
||||
for (const value of JSON.parse(process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_ORIGINS || '[]')) allowOrigin(value);
|
||||
function check(host = 'localhost', port, protocol) {
|
||||
local(host);
|
||||
if (protocol && !['http:', 'https:'].includes(protocol)) deny(`protocol ${protocol}`);
|
||||
if (!port || !endpoints.has(`${host}:${port}`) && !listeners.has(Number(port))) deny(`unowned loopback ${host}:${port}`);
|
||||
}
|
||||
function checkArgs(args, httpProtocol) {
|
||||
if (Array.isArray(args[0])) return checkArgs(args[0], httpProtocol);
|
||||
const first = args[0];
|
||||
if (first instanceof URL || typeof first === 'string' && /^https?:/.test(first)) {
|
||||
const url = first instanceof URL ? first : new URL(first); check(url.hostname, url.port || (url.protocol === 'https:' ? 443 : 80), url.protocol);
|
||||
if (args[1] && typeof args[1] === 'object') {
|
||||
const override = args[1]; if (override.hostname || override.host) check(override.hostname || override.host, override.port || url.port || (url.protocol === 'https:' ? 443 : 80), override.protocol || url.protocol);
|
||||
}
|
||||
} else if (first && typeof first === 'object') {
|
||||
if (first.socketPath || !httpProtocol && first.path && !first.port) deny('unowned Unix socket');
|
||||
const port = first.port || (httpProtocol === 'https:' ? 443 : httpProtocol === 'http:' ? 80 : undefined);
|
||||
check(first.hostname || first.host, port, first.protocol || httpProtocol);
|
||||
} else if (typeof first === 'number') check(typeof args[1] === 'string' ? args[1] : undefined, first);
|
||||
else deny('unknown connection arguments');
|
||||
}
|
||||
const listen = net.Server.prototype.listen;
|
||||
net.Server.prototype.listen = function(...args) {
|
||||
this.once('listening', () => { const address = this.address(); if (address && typeof address === 'object' && loopback.has(address.address)) {
|
||||
listeners.add(address.port); this.once('close', () => listeners.delete(address.port));
|
||||
} });
|
||||
return Reflect.apply(listen, this, args);
|
||||
};
|
||||
for (const module of [http, https]) for (const name of ['request', 'get']) {
|
||||
const original = module[name]; module[name] = function(...args) { checkArgs(args, module === https ? 'https:' : 'http:'); return Reflect.apply(original, this, args); };
|
||||
}
|
||||
for (const module of [dns, dns.promises]) for (const name of ['lookup', 'resolve', 'resolve4', 'resolve6', 'resolveAny', 'resolveCname', 'resolveMx', 'resolveNaptr', 'resolveNs', 'resolvePtr', 'resolveSoa', 'resolveSrv', 'resolveTxt', 'reverse']) if (typeof module[name] === 'function') {
|
||||
const original = module[name]; module[name] = function(...args) { local(args[0]); if (name !== 'lookup') deny(`unowned DNS ${name}`); return Reflect.apply(original, this, args); };
|
||||
}
|
||||
const connect = net.Socket.prototype.connect;
|
||||
net.Socket.prototype.connect = function(...args) {
|
||||
const normalized = Array.isArray(args[0]) ? args[0] : args;
|
||||
const first = normalized[0]; const socketPath = typeof first === 'string' ? first : first?.port ? undefined : first?.path;
|
||||
if (socketPath) {
|
||||
const expected = path.join(os.tmpdir(), `tsx-${process.getuid?.()}`, `${process.ppid}.pipe`);
|
||||
let live = false; try { process.kill(process.ppid, 0); live = true; } catch {}
|
||||
if (!live || path.resolve(socketPath) !== path.resolve(expected) || fs.existsSync(expected)) deny('unowned or existing Unix socket');
|
||||
record('current-parent-pipe-probe', { path: expected, parent: process.ppid });
|
||||
this.once('connect', () => { attempts.push('unexpected parent pipe connection'); process.exitCode = 98; record('blocked', { reason: 'unexpected parent pipe connection' }); this.destroy(); });
|
||||
this.once('error', error => {
|
||||
if (error.code !== 'ENOENT') { attempts.push(`parent pipe ${error.code}`); process.exitCode = 98; record('blocked', { reason: `parent pipe ${error.code}` }); }
|
||||
else record('current-parent-pipe-ENOENT', { path: expected, parent: process.ppid });
|
||||
});
|
||||
return Reflect.apply(connect, this, args);
|
||||
}
|
||||
checkArgs(normalized);
|
||||
const host = first && typeof first === 'object' ? first.host || first.hostname || 'localhost' : typeof normalized[1] === 'string' ? normalized[1] : 'localhost';
|
||||
const port = first && typeof first === 'object' ? first.port : first;
|
||||
if (pgEndpoints.has(`${host}:${port}`) && !tracked.has(this)) {
|
||||
tracked.add(this); this.once('connect', () => { active.add(this); peak = Math.max(peak, active.size); }); this.once('close', () => active.delete(this));
|
||||
}
|
||||
return Reflect.apply(connect, this, args);
|
||||
};
|
||||
const tlsConnect = tls.connect; tls.connect = function(...args) { checkArgs(args); return Reflect.apply(tlsConnect, this, args); };
|
||||
const fetch = globalThis.fetch; if (fetch) globalThis.fetch = function(input, init) { checkArgs([input instanceof Request ? input.url : input]); return fetch(input, init); };
|
||||
const exit = process.exit;
|
||||
process.exit = function(code) { return exit.call(process, attempts.length ? 98 : code); };
|
||||
const stats = () => ({ loaded: true, identity, sourceHash, pid: process.pid, ppid: process.ppid, threadId, isMainThread, uid: process.getuid?.(), preload: __filename,
|
||||
peakPgSockets: peak, remainingPgSockets: active.size, unexpectedAttempts: attempts.slice() });
|
||||
const state = Object.freeze({ log, stats, allowOrigin, allowDatabase });
|
||||
globalThis[key] = state; process[key] = state;
|
||||
record('preloaded', { node: process.version, execPath: process.execPath, preload: __filename });
|
||||
process.on('exit', () => record('exit', { ...stats(), exitCode: process.exitCode || 0 }));
|
||||
@@ -0,0 +1,58 @@
|
||||
import { createRequire } from 'node:module';
|
||||
import { spawn } from 'node:child_process';
|
||||
import { existsSync, readFileSync, statSync } from 'node:fs';
|
||||
import path from 'node:path';
|
||||
import { fileURLToPath } from 'node:url';
|
||||
|
||||
const root = fileURLToPath(new URL('../', import.meta.url));
|
||||
const controls = process.argv.slice(2).join(' ') === '--controls';
|
||||
if (!controls && process.argv.length > 2) throw new Error('Use no arguments for the witness or --controls for finite PG0 controls');
|
||||
const names = ['PAPERCLIP_SKILL_DEFAULT_CADENCE_TEST_DATABASE_URL', 'PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_TEST_DATABASE_URL'];
|
||||
if (!controls && process.env.PAPERCLIP_MAINTENANCE_PG0_CONTROLS) throw new Error('PG0 controls flag is forbidden for the witness');
|
||||
if (!controls) {
|
||||
const urls = names.map(name => {
|
||||
if (!process.env[name]) throw new Error(`${name} requires an explicit dedicated test database`);
|
||||
let url; try { url = new URL(process.env[name]); } catch { throw new Error(`${name} requires a loopback dedicated test database`); }
|
||||
if (!['postgres:', 'postgresql:'].includes(url.protocol) || !['127.0.0.1', 'localhost', '::1', '[::1]'].includes(url.hostname)
|
||||
|| !/^[a-zA-Z0-9_]+_test$/.test(decodeURIComponent(url.pathname.slice(1)))
|
||||
|| ['options', 'dbname', 'database'].some(key => url.searchParams.has(key))) throw new Error(`${name} requires a loopback dedicated test database`);
|
||||
return url;
|
||||
});
|
||||
if (urls[0].pathname === urls[1].pathname) throw new Error('App and Cloud require distinct test databases');
|
||||
const cloud = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_CHECKOUT;
|
||||
if (!cloud || !path.isAbsolute(cloud) || !existsSync(path.join(cloud, 'dist/sleep/fixtures/default-cadence-bridge.js'))) throw new Error('Set the exact compiled Cloud checkout');
|
||||
if (!path.isAbsolute(process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_NODE || '')) throw new Error('Set the actual Node22 executable');
|
||||
if (!/^[a-f0-9]{40}$/.test(process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_SHA || '')) throw new Error('Set exact Cloud source SHA');
|
||||
}
|
||||
const log = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG;
|
||||
if (!log || !path.isAbsolute(log) || !existsSync(path.dirname(log))) throw new Error('Set an absolute guard log in the existing evidence directory');
|
||||
if (existsSync(log) && statSync(log).size !== 0) throw new Error('Use a fresh per-attempt guard log');
|
||||
if (!process.version.startsWith('v24.')) throw new Error('Launch acceptance with the actual Node24 executable');
|
||||
const require = createRequire(new URL('../server/package.json', import.meta.url));
|
||||
const guard = path.join(root, 'scripts/fixtures/skill-default-cadence-network-guard.cjs');
|
||||
|
||||
|
||||
require(guard);
|
||||
const vitest = path.join(path.dirname(require.resolve('vitest/package.json')), 'vitest.mjs');
|
||||
const env = { PATH: process.env.PATH, HOME: process.env.HOME, LANG: process.env.LANG, TMPDIR: process.env.TMPDIR,
|
||||
NODE_ENV: 'test', NODE_OPTIONS: `--require=${guard}`, PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG: log };
|
||||
if (controls) env.PAPERCLIP_MAINTENANCE_PG0_CONTROLS = '1';
|
||||
if (!controls) {
|
||||
for (const name of [...names, 'PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_CHECKOUT', 'PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_NODE',
|
||||
'PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_SHA', 'PAPERCLIP_SKILL_DEFAULT_CADENCE_EVIDENCE', 'PAPERCLIP_SKILL_DEFAULT_CADENCE_STAGE',
|
||||
'PAPERCLIP_TEST_POSTGRES_ADMIN_URL']) if (process.env[name]) env[name] = process.env[name];
|
||||
env.PAPERCLIP_SKILL_DEFAULT_CADENCE_ACCEPTANCE = '1'; env.PAPERCLIP_MAINTENANCE_KEEP_ARTIFACTS = '1';
|
||||
}
|
||||
const files = controls ? ['server/src/__tests__/skill-default-cadence-prewire.test.ts', 'server/src/__tests__/maintenance-harness-resources.test.ts',
|
||||
'server/src/__tests__/maintenance-postgres-proxy.test.ts', 'server/src/services/skill-source-idle-schedule.test.ts']
|
||||
: ['server/src/__tests__/skill-source-default-cadence-acceptance.test.ts'];
|
||||
const child = spawn(process.execPath, ['--require', guard, vitest, 'run', ...files,
|
||||
'--maxWorkers=1', '--no-file-parallelism', ...(controls ? ['--testNamePattern=^(?!.*maintenance observer against PostgreSQL)'] : []), `--execArgv=--require=${guard}`], { cwd: root, env, stdio: 'inherit' });
|
||||
for (const signal of ['SIGINT', 'SIGTERM']) process.on(signal, () => child.kill(signal));
|
||||
child.once('error', error => { console.error(error); process.exitCode = 1; });
|
||||
child.once('close', (code, signal) => {
|
||||
const events = readFileSync(log, 'utf8').trim().split('\n').filter(Boolean).map(line => JSON.parse(line));
|
||||
const violation = events.some(event => event.event === 'blocked' || event.unexpectedAttempts?.length);
|
||||
if (violation) console.error('Default cadence guard recorded an unexpected attempt; caught exceptions do not qualify');
|
||||
process.exitCode = violation ? 98 : signal ? 1 : code ?? 1;
|
||||
});
|
||||
@@ -0,0 +1,89 @@
|
||||
import { fileURLToPath } from 'node:url';
|
||||
import { defaultCadenceNetwork } from './default-cadence-network.js';
|
||||
import { fork } from 'node:child_process';
|
||||
import path from 'node:path';
|
||||
import { setTimeout as delay } from 'node:timers/promises';
|
||||
|
||||
|
||||
export async function startDefaultCadenceBridge(options: {
|
||||
checkout: string; node: string;
|
||||
physicalStop(): Promise<void>; physicalWake(): Promise<{ origin: string }>;
|
||||
}) {
|
||||
const child = fork(path.join(options.checkout, 'dist/sleep/fixtures/default-cadence-bridge.js'), [], {
|
||||
cwd: options.checkout, execPath: options.node, execArgv: ["--require", fileURLToPath(new URL("../../../../../scripts/fixtures/skill-default-cadence-network-guard.cjs", import.meta.url))],
|
||||
env: { PATH: process.env.PATH, TMPDIR: process.env.TMPDIR, NODE_ENV: 'test', PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG: process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG }, stdio: ['ignore', 'pipe', 'pipe', 'ipc'],
|
||||
});
|
||||
let output = '', id = 0;
|
||||
child.stdout?.on('data', bytes => { output += String(bytes); });
|
||||
child.stderr?.on('data', bytes => { output += String(bytes); });
|
||||
const pending = new Map<number, { resolve(value: any): void; reject(error: Error): void }>();
|
||||
const physicalWork = new Set<Promise<void>>();
|
||||
const physicalCleanupFailures: unknown[] = [];
|
||||
let closing = false;
|
||||
let readyResolve!: () => void, readyReject!: (error: Error) => void;
|
||||
const ready = new Promise<void>((resolve, reject) => { readyResolve = resolve; readyReject = reject; });
|
||||
child.on('message', (message: any) => {
|
||||
if (message.ready) {
|
||||
const guard = message.defaultCadenceGuard;
|
||||
if (!guard?.loaded || guard.pid !== child.pid || guard.ppid !== process.pid || guard.sourceHash !== defaultCadenceNetwork().sourceHash) {
|
||||
readyReject(new Error('Actual Cloud child did not report the expected preloaded guard identity')); return;
|
||||
}
|
||||
readyResolve(); return;
|
||||
}
|
||||
if (message.callbackId) {
|
||||
if (closing) {
|
||||
if (child.connected) child.send({ callbackId: message.callbackId, error: 'Default cadence bridge is closing' });
|
||||
return;
|
||||
}
|
||||
const operation = message.physical === 'stop' ? options.physicalStop() : options.physicalWake();
|
||||
const handled = operation.then(async result => {
|
||||
if (closing && message.physical === 'wake') await options.physicalStop();
|
||||
if (child.connected) child.send({ callbackId: message.callbackId, result });
|
||||
}, error => {
|
||||
if (child.connected) child.send({ callbackId: message.callbackId, error: error instanceof Error ? error.stack : String(error) });
|
||||
});
|
||||
physicalWork.add(handled);
|
||||
void handled.then(() => physicalWork.delete(handled), error => {
|
||||
physicalWork.delete(handled); physicalCleanupFailures.push(error);
|
||||
});
|
||||
return;
|
||||
}
|
||||
const request = pending.get(message.id); pending.delete(message.id);
|
||||
if (message.error) {
|
||||
output += `\nCloud RPC failure: ${message.error}`;
|
||||
console.error(`[default-cadence-bridge] ${message.error}`);
|
||||
request?.reject(new Error(message.error));
|
||||
} else request?.resolve(message.result);
|
||||
});
|
||||
child.once('error', readyReject);
|
||||
const exited = new Promise<void>(resolve => child.once('exit', () => {
|
||||
const failure = new Error(`Default cadence bridge exited: ${output.slice(-8000)}`);
|
||||
readyReject(failure); for (const request of pending.values()) request.reject(failure); resolve();
|
||||
}));
|
||||
async function bounded<T>(work: Promise<T>, ms: number) {
|
||||
const abort = new AbortController();
|
||||
try { return await Promise.race([work, delay(ms, undefined, { signal: abort.signal }).then(() => { throw new Error(`Default cadence bridge deadline: ${output.slice(-8000)}`); })]); }
|
||||
finally { abort.abort(); }
|
||||
}
|
||||
try { await bounded(ready, 15_000); } catch (error) { child.kill("SIGKILL"); await exited; throw error; }
|
||||
async function rpc<T = any>(method: string, input?: unknown, budgetMs = 210_000): Promise<T> {
|
||||
const requestId = ++id;
|
||||
const response = new Promise<T>((resolve, reject) => pending.set(requestId, { resolve, reject }));
|
||||
child.send({ id: requestId, method, input });
|
||||
try { return await bounded(response, budgetMs); } finally { pending.delete(requestId); }
|
||||
}
|
||||
return { rpc: <T = any>(method: string, input?: unknown) => rpc<T>(method, input), pid: child.pid, get output() { return output; }, async close() {
|
||||
closing = true;
|
||||
const deadline = performance.now() + 210_000;
|
||||
const remaining = () => Math.max(0, deadline - performance.now());
|
||||
try {
|
||||
await bounded(Promise.all([...physicalWork]), remaining());
|
||||
if (physicalCleanupFailures.length) throw new AggregateError(physicalCleanupFailures, 'Physical callback cleanup failed');
|
||||
if (child.exitCode === null && child.connected) { await rpc('close', undefined, remaining()); console.info(`[default-cadence-bridge] close acknowledged pid=${child.pid}`); }
|
||||
} finally {
|
||||
console.info(`[default-cadence-bridge] disconnect pid=${child.pid} exit=${child.exitCode} signal=${child.signalCode}`);
|
||||
if (child.connected) child.disconnect();
|
||||
await bounded(exited, 10_000).catch(async error => { child.kill('SIGKILL'); await exited; throw error; });
|
||||
}
|
||||
} };
|
||||
}
|
||||
@@ -0,0 +1,35 @@
|
||||
import { installDefaultCadenceNetwork, defaultCadenceNetwork } from './default-cadence-network.js';
|
||||
import assert from 'node:assert/strict';
|
||||
import { mock } from 'node:test';
|
||||
import { githubFixture } from '../github-skills.js';
|
||||
import type { openGitSkillSnapshot } from '../../../services/skill-source-git-snapshot.js';
|
||||
|
||||
|
||||
|
||||
|
||||
installDefaultCadenceNetwork();
|
||||
const send = process.send;
|
||||
if (send) process.send = function(message: unknown, ...args: unknown[]) {
|
||||
const value = message && typeof message === 'object' ? { ...message, defaultCadenceGuard: defaultCadenceNetwork() } : message;
|
||||
return Reflect.apply(send, process, [value, ...args]);
|
||||
} as NonNullable<typeof process.send>;
|
||||
const origin = process.env.PAPERCLIP_MAINTENANCE_GITHUB_ORIGIN;
|
||||
assert(origin && new URL(origin).hostname === '127.0.0.1');
|
||||
const actual = await import('../../../services/skill-source-git-snapshot.js');
|
||||
const open: typeof openGitSkillSnapshot = async (input, options = {}) => {
|
||||
assert.equal(input.repositoryUrl, 'https://github.com/acme/skills');
|
||||
assert(['HEAD', 'main'].includes(input.ref) || /^[a-f0-9]{40}$/i.test(input.ref));
|
||||
options.signal?.throwIfAborted();
|
||||
options.beforeDownload?.();
|
||||
const response = await fetch(new URL('/__skill_snapshot', origin), { signal: options.signal });
|
||||
assert(response.ok);
|
||||
const snapshot = await response.json() as { commit: string; files: Record<string, string> };
|
||||
assert.match(snapshot.commit, /^[a-f0-9]{40}$/);
|
||||
if (input.commitSha) assert.equal(input.commitSha, snapshot.commit);
|
||||
return githubFixture(snapshot.files, {}, snapshot.commit).openSnapshot!(input, options);
|
||||
};
|
||||
mock.module(new URL('../../../services/skill-source-git-snapshot.ts', import.meta.url).href,
|
||||
{ namedExports: { ...actual, openGitSkillSnapshot: open } });
|
||||
assert.equal((await import('../../../services/skill-source-git-snapshot.js')).openGitSkillSnapshot, open,
|
||||
'Module loader bypassed the named external Git boundary');
|
||||
await import('./server-child.js');
|
||||
@@ -0,0 +1,12 @@
|
||||
type GuardStats = { loaded: true; identity: string; sourceHash: string; preload: string; pid: number; ppid: number; unexpectedAttempts: string[]; peakPgSockets: number; remainingPgSockets: number };
|
||||
type GuardState = { stats(): GuardStats; allowOrigin(origin: string): void; allowDatabase(url: string): void };
|
||||
function current(): GuardState {
|
||||
const state = (globalThis as unknown as Record<symbol, GuardState | undefined>)[Symbol.for('paperclip.skill.default-cadence.prewire')];
|
||||
if (!state || state.stats().pid !== process.pid || state.stats().identity !== 'skill-default-cadence-guard-v1') {
|
||||
throw new Error('Default cadence requires the checked-in preload before module operations');
|
||||
}
|
||||
return state;
|
||||
}
|
||||
export function installDefaultCadenceNetwork() { current(); }
|
||||
export function defaultCadenceNetwork() { const stats = current().stats(); return { ...stats, offhostAttempts: stats.unexpectedAttempts }; }
|
||||
export function allowDefaultCadenceOrigin(origin: string) { current().allowOrigin(origin); }
|
||||
@@ -61,6 +61,9 @@ export interface MaintenanceConfiguration {
|
||||
storageProvider?: "local_disk" | "s3";
|
||||
}
|
||||
export interface MaintenanceServerOptions {
|
||||
childEntry?: URL;
|
||||
childPreload?: URL;
|
||||
databasePoolMax?: number;
|
||||
artifactRoot?: string; enabled?: boolean; clock?: boolean; suppressed?: boolean; githubOrigin?: string;
|
||||
connector?: { origin: string; signPrivateKey: string; sealPrivateKey: string; instanceId?: string };
|
||||
cloud?: { jwks: string; s3Origin: string; standby: boolean };
|
||||
@@ -71,6 +74,9 @@ export interface MaintenanceServerOptions {
|
||||
}
|
||||
export async function startMaintenanceServer(databaseUrl: string, options: MaintenanceServerOptions = {}) {
|
||||
const configuration = options.configuration;
|
||||
if (options.databasePoolMax !== undefined && (!Number.isInteger(options.databasePoolMax) || options.databasePoolMax < 1 || options.databasePoolMax > 4)) {
|
||||
throw new Error('Maintenance database pool must be an integer from 1 through 4');
|
||||
}
|
||||
if (options.captureIdleWork && options.clock) throw new Error("Passive setup diagnostics require the real clock");
|
||||
if (options.tenantActor && (!options.cloud || !configuration?.managedInstance || configuration.deploymentMode !== 'authenticated' || configuration.storageProvider !== 's3')) {
|
||||
throw new Error('Managed tenant fixture requires the supported authenticated S3 profile');
|
||||
@@ -132,18 +138,22 @@ export async function startMaintenanceServer(databaseUrl: string, options: Maint
|
||||
const captureStartedAt = options.captureIdleWork ? Date.now() : undefined;
|
||||
const captureSources = options.captureIdleWork ? new Map<number, string[]>() : undefined;
|
||||
let settledDiagnosticId = 0;
|
||||
const child = fork(fileURLToPath(new URL("./server-child.ts", import.meta.url)), [], {
|
||||
const child = fork(fileURLToPath(options.childEntry ?? new URL("./server-child.ts", import.meta.url)), [], {
|
||||
cwd: fileURLToPath(new URL("../../../../..", import.meta.url)),
|
||||
execArgv: [...((configuration?.announcements || options.captureIdleWork) ? ['--experimental-test-module-mocks'] : []), "--import", fileURLToPath(new URL("../../../../node_modules/tsx/dist/loader.mjs", import.meta.url))],
|
||||
execArgv: [...(options.childPreload ? ["--require", fileURLToPath(options.childPreload)] : []), ...((configuration?.announcements || options.captureIdleWork || options.childEntry) ? ['--experimental-test-module-mocks'] : []), "--import", fileURLToPath(new URL("../../../../node_modules/tsx/dist/loader.mjs", import.meta.url))],
|
||||
env: {
|
||||
...(captureStartedAt === undefined ? {} : {
|
||||
PAPERCLIP_MAINTENANCE_IDLE_WORK_CAPTURE: '1',
|
||||
PAPERCLIP_DB_RESPONSE_CAPTURE_WINDOW: `${new Date(captureStartedAt).toISOString()}/${new Date(captureStartedAt + 60_000).toISOString()}`,
|
||||
}),
|
||||
...(options.childPreload ? {
|
||||
PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG: process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG,
|
||||
PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_ORIGINS: JSON.stringify([options.githubOrigin, options.cloud?.s3Origin].filter(Boolean)),
|
||||
} : {}),
|
||||
PATH: process.env.PATH, TMPDIR: process.env.TMPDIR, NODE_ENV: "production",
|
||||
PAPERCLIP_HOME: directory, PAPERCLIP_CONFIG: configPath, PAPERCLIP_INSTANCE_ID: "maintenance-test",
|
||||
PAPERCLIP_RUNNER_STATE_DIR: path.join(directory, "durable-sessions"),
|
||||
DATABASE_URL: proxy.url, DATABASE_POOL_MAX: configuration ? "4" : "6", DATABASE_APPLICATION_NAME: "maintenance-integration-test",
|
||||
DATABASE_URL: proxy.url, DATABASE_POOL_MAX: options.databasePoolMax === undefined ? (configuration ? "4" : "6") : String(options.databasePoolMax), DATABASE_APPLICATION_NAME: "maintenance-integration-test",
|
||||
PAPERCLIP_AGENT_JWT_SECRET: randomUUID(), PAPERCLIP_DECISION_SIGNING_SECRET: randomUUID(),
|
||||
PAPERCLIP_TOOL_ACTION_SIGNING_SECRET: randomUUID(), BETTER_AUTH_SECRET: randomUUID(),
|
||||
CODEX_HOME: path.join(directory, "codex"), PAPERCLIP_ANNOUNCEMENTS_ENABLED: String(configuration?.announcements ?? false),
|
||||
@@ -180,6 +190,7 @@ export async function startMaintenanceServer(databaseUrl: string, options: Maint
|
||||
const log = createWriteStream(path.join(directory, "server.log"));
|
||||
stopLog = () => new Promise<void>(resolve => log.end(resolve));
|
||||
let serverReady = false;
|
||||
let childGuard: unknown;
|
||||
child.stdout?.on("data", chunk => { output += String(chunk); log.write(chunk); });
|
||||
child.stderr?.on("data", chunk => { output += String(chunk); log.write(chunk); });
|
||||
const requests = new Map<number, { resolve: (value: any) => void; reject: (error: Error) => void }>();
|
||||
@@ -187,7 +198,7 @@ export async function startMaintenanceServer(databaseUrl: string, options: Maint
|
||||
const ready = new Promise<number>((resolve, reject) => { readyResolve = resolve; readyReject = reject; });
|
||||
const exited = new Promise<void>(resolve => child.once("exit", () => resolve()));
|
||||
child.on("message", (message: any) => {
|
||||
if (message.kind === "ready") { serverReady = true; readyResolve(message.port); }
|
||||
if (message.kind === "ready") { childGuard = message.defaultCadenceGuard; serverReady = true; readyResolve(message.port); }
|
||||
else if (message.id) {
|
||||
const request = requests.get(message.id); requests.delete(message.id);
|
||||
if (message.kind === "error") request?.reject(new Error(message.error)); else request?.resolve(message);
|
||||
@@ -227,7 +238,7 @@ export async function startMaintenanceServer(databaseUrl: string, options: Maint
|
||||
};
|
||||
const port = await withDeadline(ready, 90_000, () => `Maintenance boot timed out: ${output.slice(-12_000)}`);
|
||||
return {
|
||||
proxy, directory, retainedDirectory, port, stop,
|
||||
proxy, directory, retainedDirectory, port, stop, pid: child.pid, get childGuard() { return childGuard; },
|
||||
async api<T>(method: string, route: string, body?: unknown, headers: Record<string, string> = {}): Promise<T> {
|
||||
if (options.clock) await request("realtime");
|
||||
try {
|
||||
|
||||
@@ -58,7 +58,9 @@ describe("passive maintenance PostgreSQL observer", () => {
|
||||
|
||||
let databaseUrl = resolveExternalTestDatabaseUrl("PAPERCLIP_MAINTENANCE_TEST_DATABASE_URL");
|
||||
let temporary: Awaited<ReturnType<typeof startEmbeddedPostgresTestDatabase>> | undefined;
|
||||
const databaseSupport = await maintenanceDatabaseSupport(databaseUrl);
|
||||
const databaseSupport = process.env.PAPERCLIP_MAINTENANCE_PG0_CONTROLS === '1'
|
||||
? { supported: false, reason: 'Explicit PG0 controls defer the complete real-PostgreSQL observer suite without probing native support.' }
|
||||
: await maintenanceDatabaseSupport(databaseUrl);
|
||||
const describeDatabase = databaseSupport.supported ? describe : describe.skip;
|
||||
if (!databaseSupport.supported) console.warn(`Skipping maintenance PostgreSQL tests: ${databaseSupport.reason ?? "unsupported environment"}`);
|
||||
describeDatabase("maintenance observer against PostgreSQL", () => {
|
||||
|
||||
@@ -0,0 +1,196 @@
|
||||
import { threadId, isMainThread } from 'node:worker_threads';
|
||||
import { spawnSync } from 'node:child_process';
|
||||
import { mkdtempSync, readFileSync, rmSync } from 'node:fs';
|
||||
import { tmpdir } from 'node:os';
|
||||
import path from 'node:path';
|
||||
import { fileURLToPath } from 'node:url';
|
||||
import { expect, it } from 'vitest';
|
||||
|
||||
const guard = fileURLToPath(new URL('../../../scripts/fixtures/skill-default-cadence-network-guard.cjs', import.meta.url));
|
||||
const launcher = fileURLToPath(new URL('../../../scripts/test-skill-default-cadence.mjs', import.meta.url));
|
||||
const prewireRequested = Object.hasOwn(process.env, 'PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG')
|
||||
|| process.env.PAPERCLIP_MAINTENANCE_PG0_CONTROLS === '1'
|
||||
|| process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_ACCEPTANCE === '1';
|
||||
it.skipIf(!prewireRequested)('the actual Vitest worker is preloaded with its own PID before fixture execution', () => {
|
||||
const guardState = (globalThis as any)[Symbol.for('paperclip.skill.default-cadence.prewire')];
|
||||
expect(guardState).toBeDefined();
|
||||
const observed = guardState.stats();
|
||||
expect(observed.pid).toBe(process.pid); expect(observed.preload).toBe(guard);
|
||||
expect(observed.threadId).toBe(threadId); expect(observed.isMainThread).toBe(isMainThread);
|
||||
const receipts = readFileSync(process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG!, 'utf8').trim().split('\n').map(line => JSON.parse(line));
|
||||
expect(receipts).toContainEqual(expect.objectContaining({ event: 'preloaded', pid: process.pid, threadId, isMainThread, sourceHash: observed.sourceHash }));
|
||||
expect(observed.unexpectedAttempts).toEqual([]);
|
||||
console.info(JSON.stringify({ actualVitestWorkerGuard: observed }));
|
||||
});
|
||||
function child(code: string, extra: Record<string, string> = {}, nodeArgs: string[] = []) {
|
||||
const root = mkdtempSync(path.join(tmpdir(), 'skill-prewire-'));
|
||||
const log = path.join(root, 'guard.jsonl');
|
||||
try {
|
||||
const result = spawnSync(process.execPath, ['--require', guard, ...nodeArgs, '-e', code], { timeout: 5_000, encoding: 'utf8',
|
||||
env: { PATH: process.env.PATH, HOME: process.env.HOME, TMPDIR: root, NODE_OPTIONS: '', PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG: log, ...extra } });
|
||||
const events = readFileSync(log, 'utf8').trim().split('\n').map(line => JSON.parse(line));
|
||||
return { result, events };
|
||||
} finally { rmSync(root, { recursive: true, force: true }); }
|
||||
}
|
||||
const stats = "globalThis[Symbol.for('paperclip.skill.default-cadence.prewire')].stats()";
|
||||
it.each([
|
||||
"require('node:http').get('http://example.invalid/')",
|
||||
"require('node:https').get('https://example.invalid/')",
|
||||
"require('node:net').connect(443,'example.invalid')",
|
||||
"require('node:tls').connect(443,'example.invalid')",
|
||||
"fetch('https://example.invalid/')",
|
||||
"require('node:dns').lookup('example.invalid',()=>{})",
|
||||
"require('node:dns').promises.resolve4('example.invalid')",
|
||||
"new(require('node:net').Socket)().connect([{host:'example.invalid',port:443},()=>{}])",
|
||||
"require('node:net').connect(55448,'127.0.0.1')",
|
||||
"require('node:net').connect(require('node:path').join(require('node:os').tmpdir(),'tsx-'+process.getuid(),(process.ppid+12345)+'.pipe'))",
|
||||
])('preload rejects and records a caught unowned transport: %s', operation => {
|
||||
const { result, events } = child(`try { ${operation}; } catch {} console.log(JSON.stringify(${stats}));`);
|
||||
expect(result.error).toBeUndefined(); expect(result.status).toBe(98);
|
||||
const observed = JSON.parse(result.stdout.trim());
|
||||
expect(observed.pid).toBe(result.pid); expect(observed.unexpectedAttempts).toHaveLength(1);
|
||||
expect(observed.peakPgSockets).toBe(0);
|
||||
expect(events[0]).toMatchObject({ event: 'preloaded', pid: result.pid });
|
||||
expect(events.filter(event => event.event === 'blocked')).toHaveLength(1);
|
||||
expect(events.at(-1)).toMatchObject({ event: 'exit', exitCode: 98 });
|
||||
});
|
||||
it('checks an overriding HTTP hostname even when the URL names a real owned listener', () => {
|
||||
const { result, events } = child(`const http=require('node:http');let dnsCalls=0;
|
||||
const server=http.createServer();server.listen(0,'127.0.0.1',()=>{
|
||||
require('node:dns').lookup=()=>{dnsCalls++;throw Error('DNS must not run');};
|
||||
try{http.get(new URL('http://127.0.0.1:'+server.address().port),{hostname:'example.invalid'});}catch{}
|
||||
server.close(()=>console.log(JSON.stringify({...${stats},dnsCalls})));
|
||||
});`);
|
||||
expect(result.error).toBeUndefined(); expect(result.status).toBe(98);
|
||||
expect(JSON.parse(result.stdout.trim())).toMatchObject({ dnsCalls: 0, unexpectedAttempts: ['offhost example.invalid'] });
|
||||
expect(events.filter(event => event.event === 'blocked')).toHaveLength(1);
|
||||
});
|
||||
it('never permits an existing Unix path even for the exact current parent', () => {
|
||||
const { result, events } = child(`const fs=require('node:fs'),path=require('node:path'),os=require('node:os');
|
||||
const dir=path.join(os.tmpdir(),'tsx-'+process.getuid());fs.mkdirSync(dir,{recursive:true});
|
||||
const pipe=path.join(dir,process.ppid+'.pipe');fs.writeFileSync(pipe,'not a socket');
|
||||
try{require('node:net').connect(pipe);}catch{}console.log(JSON.stringify(${stats}));`);
|
||||
expect(result.error).toBeUndefined(); expect(result.status).toBe(98);
|
||||
expect(events.filter(event => event.event === 'blocked')).toHaveLength(1);
|
||||
});
|
||||
it('permits actual owned loopback HTTP only after the real listener binds', () => {
|
||||
const { result, events } = child(`const http=require('node:http'); const server=http.createServer((req,res)=>res.end('useful response'));
|
||||
server.listen(0,'127.0.0.1',()=>http.get('http://127.0.0.1:'+server.address().port,res=>{
|
||||
let body='';res.on('data',chunk=>body+=chunk);res.on('end',()=>server.close(()=>{
|
||||
if(body!=='useful response')throw Error('missing useful body');console.log(JSON.stringify(${stats}));
|
||||
}));
|
||||
}));`);
|
||||
expect(result.error).toBeUndefined(); expect(result.status).toBe(0);
|
||||
expect(JSON.parse(result.stdout.trim())).toMatchObject({ pid: result.pid, loaded: true, unexpectedAttempts: [] });
|
||||
expect(events.at(-1)).toMatchObject({ event: 'exit', exitCode: 0 });
|
||||
});
|
||||
it('permits only an absent current live-parent TSX probe and records the actual ENOENT', () => {
|
||||
const { result, events } = child(`const path=require('node:path'),os=require('node:os');
|
||||
const socket=require('node:net').connect(path.join(os.tmpdir(),'tsx-'+process.getuid(),process.ppid+'.pipe'));
|
||||
socket.on('error',error=>{if(error.code!=='ENOENT')throw error;console.log(JSON.stringify(${stats}));});`);
|
||||
expect(result.error).toBeUndefined(); expect(result.status).toBe(0);
|
||||
expect(JSON.parse(result.stdout.trim()).unexpectedAttempts).toEqual([]);
|
||||
expect(events.filter(event => event.event === 'current-parent-pipe-ENOENT')).toHaveLength(1);
|
||||
});
|
||||
it.each([{},
|
||||
{ PAPERCLIP_SKILL_DEFAULT_CADENCE_TEST_DATABASE_URL: 'postgres://u@127.0.0.1/live', PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_TEST_DATABASE_URL: 'postgres://u@127.0.0.1/cloud_test' },
|
||||
{ PAPERCLIP_SKILL_DEFAULT_CADENCE_TEST_DATABASE_URL: 'postgres://u@127.0.0.1/same_test', PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_TEST_DATABASE_URL: 'postgres://u@127.0.0.1/same_test' },
|
||||
{ PAPERCLIP_SKILL_DEFAULT_CADENCE_TEST_DATABASE_URL: 'postgres://u@example.invalid/app_test', PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_TEST_DATABASE_URL: 'postgres://u@127.0.0.1/cloud_test' },
|
||||
])('launcher refuses absent/unsafe settings before loading Vitest', settings => {
|
||||
const result = spawnSync(process.execPath, [launcher], { timeout: 5_000, encoding: 'utf8', env: { PATH: process.env.PATH, NODE_OPTIONS: '', ...settings } });
|
||||
expect(result.error).toBeUndefined(); expect(result.status).toBe(1);
|
||||
expect(result.stderr).toMatch(/test database|loopback|distinct/); expect(result.stdout).not.toContain('RUN');
|
||||
});
|
||||
|
||||
it('permits approved HTTP options with a request path and default port through an owned agent', () => {
|
||||
const { result } = child(`const http=require('node:http'),net=require('node:net');
|
||||
const state=globalThis[Symbol.for('paperclip.skill.default-cadence.prewire')];let agentCalls=0;
|
||||
const server=http.createServer((req,res)=>res.end('approved options response'));
|
||||
server.listen(0,'127.0.0.1',()=>{
|
||||
state.allowOrigin('http://127.0.0.1');
|
||||
const agent=new http.Agent();agent.createConnection=()=>{agentCalls++;return net.connect(server.address().port,'127.0.0.1');};
|
||||
http.get({host:'127.0.0.1',path:'/useful',agent},res=>{let body='';res.on('data',chunk=>body+=chunk);
|
||||
res.on('end',()=>{agent.destroy();server.close(()=>console.log(JSON.stringify({...state.stats(),body,agentCalls})));});});
|
||||
});`);
|
||||
expect(result.error).toBeUndefined(); expect(result.status).toBe(0);
|
||||
expect(JSON.parse(result.stdout.trim())).toMatchObject({ body: 'approved options response', agentCalls: 1, unexpectedAttempts: [] });
|
||||
});
|
||||
it.each([
|
||||
['unknown.example.invalid', 'offhost unknown.example.invalid'],
|
||||
['127.0.0.1', 'unowned loopback 127.0.0.1:80'],
|
||||
])('rejects an unknown HTTP options origin before agent work: %s', (host, reason) => {
|
||||
const { result } = child(`const http=require('node:http');let agentCalls=0;const agent=new http.Agent();
|
||||
agent.createConnection=()=>{agentCalls++;throw Error('Agent must not run');};
|
||||
try{http.get({host:${JSON.stringify(host)},path:'/',agent});}catch{}
|
||||
console.log(JSON.stringify({...${stats},agentCalls}));`);
|
||||
expect(result.error).toBeUndefined(); expect(result.status).toBe(98);
|
||||
expect(JSON.parse(result.stdout.trim())).toMatchObject({ agentCalls: 0, unexpectedAttempts: [reason] });
|
||||
});
|
||||
it('joins an admitted deferred wake and closes its real child listener before bridge close returns', () => {
|
||||
const bridgeSource = process.env.PAPERCLIP_SKILL_BRIDGE_CONTROL_SOURCE
|
||||
?? fileURLToPath(new URL('./helpers/maintenance/default-cadence-bridge.ts', import.meta.url));
|
||||
const networkSource = fileURLToPath(new URL('./helpers/maintenance/default-cadence-network.ts', import.meta.url));
|
||||
const loader = fileURLToPath(new URL('../../node_modules/tsx/dist/loader.mjs', import.meta.url));
|
||||
const { result, events } = child(String.raw`(async()=>{
|
||||
const assert=require('node:assert/strict'),fs=require('node:fs'),path=require('node:path'),os=require('node:os');
|
||||
const {fork}=require('node:child_process'),{pathToFileURL}=require('node:url');
|
||||
const delay=ms=>new Promise(resolve=>setTimeout(resolve,ms));
|
||||
const ipcErrors=[];process.on('uncaughtException',error=>{if(error.code!=='ERR_IPC_CHANNEL_CLOSED')throw error;ipcErrors.push({code:error.code,message:error.message});});
|
||||
const root=fs.mkdtempSync(path.join(os.tmpdir(),'bridge-close-')),app=path.join(root,'app');
|
||||
const helpers=path.join(app,'server/src/__tests__/helpers/maintenance');fs.mkdirSync(helpers,{recursive:true});
|
||||
fs.writeFileSync(path.join(app,'package.json'),JSON.stringify({type:'module'}));
|
||||
const source=path.join(helpers,'default-cadence-bridge.ts');fs.copyFileSync(process.env.BRIDGE_SOURCE,source);
|
||||
fs.copyFileSync(process.env.NETWORK_SOURCE,path.join(helpers,'default-cadence-network.ts'));
|
||||
const preload=path.join(app,'scripts/fixtures/skill-default-cadence-network-guard.cjs');fs.mkdirSync(path.dirname(preload),{recursive:true});fs.copyFileSync(process.env.GUARD_SOURCE,preload);
|
||||
const checkout=path.join(root,'cloud'),entry=path.join(checkout,'dist/sleep/fixtures/default-cadence-bridge.js');fs.mkdirSync(path.dirname(entry),{recursive:true});
|
||||
fs.writeFileSync(entry,[
|
||||
"process.on('message',message=>{if(message.method==='trigger'){process.send({callbackId:'wake',physical:'wake'},()=>process.exit(0));}else if(message.method==='close'){process.send({id:message.id,result:true});}});",
|
||||
"process.send({ready:true,pid:process.pid,defaultCadenceGuard:globalThis[Symbol.for('paperclip.skill.default-cadence.prewire')].stats()});"
|
||||
].join('\n'));
|
||||
const physicalFile=path.join(root,'physical-app.cjs');fs.writeFileSync(physicalFile,[
|
||||
"const http=require('node:http');const server=http.createServer((req,res)=>res.end('late real child'));",
|
||||
"server.listen(0,'127.0.0.1',()=>process.send({origin:'http://127.0.0.1:'+server.address().port}));",
|
||||
"process.on('message',message=>{if(message.stop)server.close(()=>process.disconnect());});"
|
||||
].join('\n'));
|
||||
let release,start,finish,fail,physical,physicalExit,origin,body,physicalOutput='',stops=0;
|
||||
const released=new Promise(resolve=>{release=resolve;}),started=new Promise(resolve=>{start=resolve;}),finished=new Promise((resolve,reject)=>{finish=resolve;fail=reject;});
|
||||
const state=globalThis[Symbol.for('paperclip.skill.default-cadence.prewire')];
|
||||
const stop=async()=>{stops++;if(physical&&physical.exitCode===null){
|
||||
if(physical.connected)physical.send({stop:true});else physical.kill('SIGTERM');
|
||||
await Promise.race([physicalExit,delay(1000).then(async()=>{if(physical.exitCode===null)physical.kill('SIGKILL');await physicalExit;})]);
|
||||
}};
|
||||
const {startDefaultCadenceBridge}=await import(pathToFileURL(source).href);let bridge;
|
||||
try{
|
||||
bridge=await startDefaultCadenceBridge({checkout,node:process.execPath,physicalStop:stop,physicalWake:async()=>{
|
||||
start();await released;
|
||||
try{
|
||||
physical=fork(physicalFile,[],{execArgv:['--require',preload],env:process.env,stdio:['ignore','ignore','pipe','ipc']});
|
||||
physical.stderr.on('data',bytes=>{physicalOutput+=String(bytes);});
|
||||
physicalExit=new Promise(resolve=>physical.once('exit',resolve));
|
||||
origin=await new Promise((resolve,reject)=>{physical.once('message',message=>resolve(message.origin));physical.once('error',reject);physical.once('exit',code=>reject(Error('Physical child exited before origin: '+code+' '+physicalOutput)));});
|
||||
state.allowOrigin(origin);body=await(await fetch(origin)).text();finish();return{origin};
|
||||
}catch(error){fail(error);throw error;}
|
||||
}});
|
||||
const trigger=bridge.rpc('trigger').then(()=>null,error=>String(error));await started;await trigger;
|
||||
const closing=bridge.close().then(()=>({ok:true}),error=>({ok:false,error:String(error)}));
|
||||
const closedEarly=await Promise.race([closing.then(()=>true),delay(40).then(()=>false)]);
|
||||
release();await finished;const closed=await closing;
|
||||
let reachable=false;try{reachable=(await fetch(origin)).ok;}catch{}
|
||||
console.log(JSON.stringify({...state.stats(),body,closedEarly,closed,reachable,stops,origin,ipcErrors,physicalPid:physical.pid,physicalExitCode:physical.exitCode}));
|
||||
assert.equal(body,'late real child');assert.equal(reachable,false,'Late child listener remained reachable after bridge.close');
|
||||
assert.equal(physical.exitCode,0);assert.equal(stops,1);assert.equal(closedEarly,false);assert.equal(closed.ok,true);assert.deepEqual(ipcErrors,[]);
|
||||
}finally{
|
||||
release();if(physical&&physical.exitCode===null)await stop();fs.rmSync(root,{recursive:true,force:true});
|
||||
}
|
||||
})().catch(error=>{console.error(error.stack);process.exitCode=1;});`, {
|
||||
BRIDGE_SOURCE: bridgeSource, NETWORK_SOURCE: networkSource, GUARD_SOURCE: guard,
|
||||
}, ['--import', loader]);
|
||||
const diagnostics = JSON.stringify({ stdout: result.stdout, stderr: result.stderr, events });
|
||||
expect(result.error, diagnostics).toBeUndefined(); expect(result.status, diagnostics).toBe(0);
|
||||
const observation = result.stdout.trim().split('\n').at(-1)!;
|
||||
expect(observation, diagnostics).toMatch(/^\{/);
|
||||
console.info(JSON.stringify({ cleanupObservation: JSON.parse(observation), guardEvents: events }));
|
||||
expect(JSON.parse(observation)).toMatchObject({
|
||||
body: 'late real child', closedEarly: false, closed: { ok: true }, reachable: false, stops: 1, physicalExitCode: 0, ipcErrors: [], unexpectedAttempts: [],
|
||||
});
|
||||
});
|
||||
@@ -0,0 +1,186 @@
|
||||
import { defaultCadenceNetwork, installDefaultCadenceNetwork, allowDefaultCadenceOrigin } from './helpers/maintenance/default-cadence-network.js';
|
||||
import assert from 'node:assert/strict';
|
||||
import { generateKeyPairSync } from 'node:crypto';
|
||||
import { execFileSync } from 'node:child_process';
|
||||
import { createServer } from 'node:http';
|
||||
import { readFile, writeFile } from 'node:fs/promises';
|
||||
import path from 'node:path';
|
||||
import { setTimeout as delay } from 'node:timers/promises';
|
||||
import { describe, expect, it } from 'vitest';
|
||||
import { applyPendingMigrations, createDb } from '@paperclipai/db';
|
||||
import { resolveExternalTestDatabaseUrl } from './helpers/external-test-database.js';
|
||||
import { githubFixture } from './helpers/github-skills.js';
|
||||
import { startMaintenanceServer, withMaintenanceProvider } from './helpers/maintenance/server-harness.js';
|
||||
import { createMaintenanceS3Boundary } from './helpers/maintenance/s3-boundary.js';
|
||||
import { provisionManagedSleepCompany, managedTenantActor } from './helpers/maintenance/managed-profile.js';
|
||||
import { startDefaultCadenceBridge } from './helpers/maintenance/default-cadence-bridge.js';
|
||||
|
||||
const enabled = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_ACCEPTANCE === '1';
|
||||
const stage = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_STAGE ?? 'two-cycle';
|
||||
if (enabled && !['bootstrap', 'two-cycle'].includes(stage)) throw new Error('Default cadence stage must be bootstrap or two-cycle');
|
||||
if (enabled) installDefaultCadenceNetwork();
|
||||
const databaseUrl = resolveExternalTestDatabaseUrl('PAPERCLIP_SKILL_DEFAULT_CADENCE_TEST_DATABASE_URL');
|
||||
const cloudDatabaseUrl = resolveExternalTestDatabaseUrl('PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_TEST_DATABASE_URL');
|
||||
const cloudCheckout = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_CHECKOUT;
|
||||
const cloudSha = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_SHA;
|
||||
const cloudNode = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_CLOUD_NODE;
|
||||
const evidence = process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_EVIDENCE;
|
||||
if (enabled && (!databaseUrl || !cloudDatabaseUrl || !cloudCheckout || !cloudNode || !evidence || !/^[a-f0-9]{40}$/.test(cloudSha ?? '')
|
||||
|| process.env.PAPERCLIP_MAINTENANCE_KEEP_ARTIFACTS !== '1')) throw new Error('Default cadence witness requires two explicit owned DBs, pinned Cloud checkout/Node22, and retained evidence');
|
||||
|
||||
|
||||
|
||||
describe.skipIf(!enabled)('default skill cadence actual signed host handoff', () => {
|
||||
it(stage === 'bootstrap' ? 'qualifies actual producer, signed legacy refusal and original-D durable ACK only'
|
||||
: 'sleeps and cold-reactivates twice under each original default300s deadline', async () => {
|
||||
assert.equal(execFileSync('git', ['-C', cloudCheckout!, 'rev-parse', 'HEAD'], { encoding: 'utf8' }).trim(), cloudSha);
|
||||
assert.match(execFileSync(cloudNode!, ['--version'], { encoding: 'utf8' }), /^v22\./);
|
||||
await applyPendingMigrations(databaseUrl!);
|
||||
|
||||
let upstream = 1;
|
||||
const files = () => ({ 'workflow/SKILL.md': `---\nname: default-cadence\ndescription: Useful default cadence workflow\n---\nPersisted useful revision ${upstream}.\n` });
|
||||
const commit = () => String(upstream).repeat(40);
|
||||
const requests: Array<{ at: number; path: string }> = [];
|
||||
const upstreamServer = createServer(async (req, res) => {
|
||||
requests.push({ at: Date.now(), path: req.url! });
|
||||
try {
|
||||
const body = req.url === '/__skill_snapshot' ? { files: files(), commit: commit() }
|
||||
: await githubFixture(files(), {}, commit())(req.url!);
|
||||
res.setHeader('Content-Type', 'application/json'); res.end(JSON.stringify(body));
|
||||
} catch (error) { res.statusCode = 500; res.end(JSON.stringify({ error: String(error) })); }
|
||||
});
|
||||
const s3 = createMaintenanceS3Boundary();
|
||||
const pair = generateKeyPairSync('ed25519');
|
||||
const jwk = { ...pair.publicKey.export({ format: 'jwk' }), kid: 'default-cadence', alg: 'EdDSA', use: 'sig' };
|
||||
const privateJwk = JSON.stringify({ ...pair.privateKey.export({ format: 'jwk' }), kid: jwk.kid, alg: 'EdDSA', use: 'sig' });
|
||||
const jwks = JSON.stringify({ keys: [jwk] });
|
||||
const cycles: unknown[] = [];
|
||||
const childPids: number[] = [];
|
||||
let server: Awaited<ReturnType<typeof startMaintenanceServer>> | undefined;
|
||||
let bridge: Awaited<ReturnType<typeof startDefaultCadenceBridge>> | undefined;
|
||||
try {
|
||||
await withMaintenanceProvider(s3.server, s3Origin => withMaintenanceProvider(upstreamServer, async githubOrigin => {
|
||||
try {
|
||||
const boot = async () => {
|
||||
const started = await startMaintenanceServer(databaseUrl!, {
|
||||
childEntry: new URL('./helpers/maintenance/default-cadence-child.ts', import.meta.url),
|
||||
childPreload: new URL('../../../scripts/fixtures/skill-default-cadence-network-guard.cjs', import.meta.url),
|
||||
databasePoolMax: 2, artifactRoot: evidence, githubOrigin, clock: false, controlJwks: jwks,
|
||||
cloud: { jwks, s3Origin, standby: false }, tenantActor: managedTenantActor,
|
||||
configuration: { managedInstance: true, deploymentMode: 'authenticated', storageProvider: 's3', announcements: false },
|
||||
});
|
||||
assert(started.pid); childPids.push(started.pid);
|
||||
const guard = started.childGuard as ReturnType<typeof defaultCadenceNetwork>;
|
||||
expect(guard.loaded).toBe(true); expect(guard.pid).toBe(started.pid); expect(guard.ppid).toBe(process.pid);
|
||||
expect(guard.sourceHash).toBe(defaultCadenceNetwork().sourceHash);
|
||||
allowDefaultCadenceOrigin(`http://127.0.0.1:${started.port}`);
|
||||
return started;
|
||||
};
|
||||
server = await boot();
|
||||
|
||||
|
||||
const setupDb = createDb(databaseUrl!, { maxConnections: 1 });
|
||||
const company = await provisionManagedSleepCompany(server, setupDb).finally(() => setupDb.$client.end({ timeout: 1 }));
|
||||
await server.settled();
|
||||
bridge = await startDefaultCadenceBridge({ checkout: cloudCheckout!, node: cloudNode!,
|
||||
physicalStop: async () => { assert(server); await server.stop(); server = undefined; upstream++; },
|
||||
physicalWake: async () => { assert(!server); server = await boot(); return { origin: `http://127.0.0.1:${server.port}` }; } });
|
||||
assert(bridge.pid); childPids.push(bridge.pid);
|
||||
await bridge.rpc('initialize', { databaseUrl: cloudDatabaseUrl, migrationsPath: path.join(cloudCheckout!, 'migrations'),
|
||||
appOrigin: `http://127.0.0.1:${server.port}`, privateJwk, jwks });
|
||||
await server.settled();
|
||||
|
||||
await server.api('POST', `/companies/${company.id}/projects`, { name: 'Default cadence source',
|
||||
workspace: { name: 'Repository', repoUrl: 'https://github.com/acme/skills', defaultRef: 'HEAD' } });
|
||||
const persisted = async () => {
|
||||
const observerUrl = new URL(databaseUrl!); observerUrl.searchParams.set('application_name', 'skill-default-readonly-observer');
|
||||
const observer = createDb(observerUrl.href, { maxConnections: 1 });
|
||||
try {
|
||||
const rows = await observer.$client.begin('read only', tx => tx`select s.id,s.current_version_id,s.markdown,c.next_sync_at,c.last_success_at,c.last_scan_commit,c.lease_token,c.automatic_updates,c.enabled
|
||||
from company_skills s join company_skill_source_entries e on e.skill_id=s.id and e.company_id=s.company_id
|
||||
join company_skill_sources c on c.id=e.source_id and c.company_id=s.company_id where s.company_id=${company.id}`);
|
||||
assert.equal(rows.length, 1); return rows[0]!;
|
||||
} finally { await observer.$client.end({ timeout: 1 }); }
|
||||
};
|
||||
await expect.poll(async () => {
|
||||
try { return (await persisted()).markdown; } catch { return ''; }
|
||||
}, { timeout: 30_000 }).toContain('Persisted useful revision 1.');
|
||||
const initial = await persisted();
|
||||
let priorVersion = initial.current_version_id;
|
||||
const legacy = await bridge.rpc('legacy');
|
||||
expect(legacy.idleSleepSafe).not.toBe(true);
|
||||
for (let cycle = 1; cycle <= 2; cycle++) {
|
||||
assert(server);
|
||||
await server.settled();
|
||||
const source = await persisted();
|
||||
const originalD = new Date(source.next_sync_at).getTime();
|
||||
expect(originalD - new Date(source.last_success_at).getTime()).toBeGreaterThanOrEqual(300_000);
|
||||
expect(originalD - new Date(source.last_success_at).getTime()).toBeLessThan(300_050);
|
||||
expect(source.automatic_updates && source.enabled).toBe(true);
|
||||
expect(source.lease_token).toBeNull();
|
||||
const refreshRequestStart = requests.length;
|
||||
const sleep = await bridge.rpc('sleep');
|
||||
expect(server).toBeUndefined();
|
||||
expect(new Date(sleep.alarm.next_wake_at).getTime()).toBe(originalD);
|
||||
const wakeAt = new Date(sleep.alarm.next_attempt_at).getTime();
|
||||
expect(wakeAt).toBe(originalD - 120_000);
|
||||
await writeFile(path.join(evidence!, `cycle-${cycle}-durable-ack.json`), JSON.stringify({ originalD, sleep, requests }, null, 2));
|
||||
if (stage === 'bootstrap') {
|
||||
await writeFile(path.join(evidence!, 'bootstrap-only-receipt.json'), JSON.stringify({ stage, originalD, sleep,
|
||||
initial, requests, boundary: 'Actual initial useful producer/signed gate/durable ACK/physical child stop only. No due wake or second cycle executed.' }, null, 2));
|
||||
return;
|
||||
}
|
||||
|
||||
await delay(Math.max(0, wakeAt - Date.now()) + 25);
|
||||
const wake = await bridge.rpc('wake');
|
||||
assert(server, 'Actual controller provider callback must cold-start a new child');
|
||||
await server.settled();
|
||||
|
||||
const before = await persisted();
|
||||
expect(before.current_version_id).toBe(priorVersion);
|
||||
const snapBefore = await server.snapshot(), queryCount = server.proxy.queries.length, providerCount = requests.length;
|
||||
await delay(10_000);
|
||||
const snapAfter = await server.snapshot();
|
||||
expect(server.proxy.queries.length).toBe(queryCount);
|
||||
expect(requests.length).toBe(providerCount);
|
||||
expect(snapAfter.outgoingHttpRequests).toBe(snapBefore.outgoingHttpRequests);
|
||||
await delay(Math.max(0, originalD - Date.now()) + 25);
|
||||
await expect.poll(async () => (await persisted()).markdown, { timeout: 30_000 }).toContain(`Persisted useful revision ${cycle + 1}.`);
|
||||
const changed = await persisted();
|
||||
expect(changed.current_version_id).not.toBe(priorVersion);
|
||||
expect(new Date(changed.last_success_at).getTime()).toBeGreaterThanOrEqual(originalD);
|
||||
const refreshRequests = requests.slice(refreshRequestStart);
|
||||
expect(refreshRequests.length).toBeGreaterThan(0);
|
||||
expect(Math.min(...refreshRequests.map(request => request.at))).toBeGreaterThanOrEqual(originalD);
|
||||
priorVersion = changed.current_version_id;
|
||||
await server.settled();
|
||||
cycles.push({ cycle, originalD, wakeAt, sleep, wake, changed, refreshRequestStart, refreshRequests, quiet: { snapBefore, snapAfter, queryCount, providerCount } });
|
||||
await writeFile(path.join(evidence!, 'two-cycle-receipt.json'), JSON.stringify({ cycles, requests, s3: s3.snapshot(),
|
||||
boundary: 'Actual app index/routes/source owner and signed Cloud gate/controller/Postgres alarms; controlled localhost Git/S3/process stop only; no Railway/Neon claim' }, null, 2));
|
||||
}
|
||||
} catch (error) {
|
||||
await writeFile(path.join(evidence!, 'primary-failure-before-cleanup.json'), JSON.stringify({
|
||||
error: error instanceof Error ? error.stack : String(error), bridgeOutput: bridge?.output, serverOutput: server?.output,
|
||||
}, null, 2));
|
||||
throw error;
|
||||
} finally {
|
||||
try { const current = bridge; bridge = undefined; await current?.close(); }
|
||||
finally { const current = server; server = undefined; await current?.stop(); }
|
||||
}
|
||||
}));
|
||||
} catch (error) {
|
||||
await writeFile(path.join(evidence!, 'attempt-failure.json'), JSON.stringify({ at: new Date().toISOString(),
|
||||
error: error instanceof Error ? error.stack : String(error), cycles, requests, bridgeOutput: bridge?.output,
|
||||
serverOutput: server?.output, boundary: 'Actual first failure; no automatic retry or assertion/time policy change' }, null, 2));
|
||||
throw error;
|
||||
} finally {
|
||||
try { await bridge?.close(); } finally { await server?.stop(); }
|
||||
expect(defaultCadenceNetwork().offhostAttempts).toEqual([]);
|
||||
const events = (await readFile(process.env.PAPERCLIP_SKILL_DEFAULT_CADENCE_GUARD_LOG!, 'utf8')).trim().split('\n').filter(Boolean).map(line => JSON.parse(line));
|
||||
expect(events.filter(event => event.event === 'blocked')).toEqual([]);
|
||||
for (const pid of [process.pid, process.ppid, ...childPids]) {
|
||||
expect(events).toContainEqual(expect.objectContaining({ event: 'preloaded', pid, sourceHash: defaultCadenceNetwork().sourceHash }));
|
||||
}
|
||||
}
|
||||
}, 850_000);
|
||||
});
|
||||
+4
-2
@@ -128,7 +128,6 @@ describeEmbeddedPostgres("native ingress bootstrap retry through the real Runner
|
||||
const address = portProbe.address();
|
||||
if (!address || typeof address === "string") throw new Error("Missing fixture port");
|
||||
const runnerPort = address.port;
|
||||
await new Promise<void>(done => portProbe.close(() => done()));
|
||||
let healthy = false, acquisitions = 0, previewUrl = "";
|
||||
const preview = createHttpsServer({ key: await readFile(key), cert: await readFile(certificate) }, (request, response) => {
|
||||
if (request.url !== "/preview") { response.writeHead(404).end(); return; }
|
||||
@@ -159,9 +158,12 @@ describeEmbeddedPostgres("native ingress bootstrap retry through the real Runner
|
||||
});
|
||||
});
|
||||
preview.on("tlsClientError", error => { diagnostics.push(`fixture TLS: ${error.message}`); });
|
||||
await new Promise<void>(done => preview.listen(0, "127.0.0.1", done));
|
||||
try {
|
||||
await new Promise<void>(done => preview.listen(0, "127.0.0.1", done));
|
||||
} finally { await new Promise<void>(done => portProbe.close(() => done())); }
|
||||
const previewAddress = preview.address();
|
||||
if (!previewAddress || typeof previewAddress === "string") throw new Error("Missing preview address");
|
||||
expect(previewAddress.port).not.toBe(runnerPort);
|
||||
previewUrl = `wss://127.0.0.1:${previewAddress.port}/api/runner/v1/connect/${runId}`;
|
||||
const bundles: ReturnType<typeof createRunnerdCodexTransport>[] = [];
|
||||
const handles: RunnerProcessHandle[] = [];
|
||||
|
||||
Reference in new issue
Block a user