Files
picpeak/backend/__tests__/integration/restoreService.pgBranch.test.js
T
Paul Nothaft d868aac703 fix(security): close 5 Trivy alerts — postcss/tar bumps + drop npm from the runtime image (stable) (#879)
* fix(security): close the 5 open Trivy alerts — dep bumps + drop npm from the runtime image

Backend deps:
- postcss 8.5.10 -> 8.5.18 (CVE-2026-45623, GHSA-r28c-9q8g-f849; the pin
  exists to force sanitize-html's transitive copy onto a fixed version)
- tar pin/override >=7.5.16 -> >=7.5.21, resolves 7.5.22
  (GHSA-r292-9mhp-454m)

Runtime image:
- Remove the npm CLI from the final stage instead of upgrading it: npm's
  bundled node_modules ship tar 7.5.19 and brace-expansion 5.0.7 (no npm
  release bundles the fixed versions — checked 11.18.0 and 12.0.1), and
  npm never runs in production. wait-for-db.sh now invokes the migration
  runners via node directly. This ends the recurring npm-bundled-CVE
  alert class; the previous 'npm install -g npm@11' line was itself a
  patch for the last batch. (stable)

* fix(restore): run post-restore migrations via node — the image ships no npm

restoreService still shelled out to 'npm run migrate:safe' after a
restore; with npm removed from the runtime image that would ENOENT into
the non-fatal catch, silently leaving a restored older backup on a
schema behind the running code until the next container restart. Invoke
migrations/run-migrations-safe.js through node directly, matching
wait-for-db.sh. The PR #596 source-contract test now pins the new
invocation. (stable)
2026-07-27 09:54:43 +02:00

262 lines
11 KiB
JavaScript

/**
* Pins the fix for the PR #596 review blocker.
*
* **The bug**
*
* `preservedMeta` was declared with `let` INSIDE the PostgreSQL
* `else` branch of `performDatabaseRestore`, then read AFTER the
* `else` block closed at the shared replay site (~L1030). On every
* real PG restore:
*
* ReferenceError: preservedMeta is not defined
*
* would fire — psql had already completed the data restore, but
* the operator-meta replay never ran, the trigger file was left
* in place by `_installFromBackupBoot.js` because the restore
* "failed", and `combined.log` got a loud FAILED line even though
* the data was back. Caught on PR #596 review by the maintainer.
*
* **Why CI missed it**
*
* The integration tests around `performFullRestore` only exercise
* the SQLite branch via `this.dbType === 'sqlite'`. The PG branch
* (~L827-984) requires a real PG connection + real `psql` binary,
* neither of which are in the test environment. So the scope leak
* sat untested until the maintainer ran a real DR cycle.
*
* **What this test does**
*
* Reads the source of `restoreService.js` and asserts the scope
* contract: the `preservedMeta` declaration sits ABOVE the
* SQLite/PG branch split, so the replay block at the bottom of the
* try{} can read it on either branch.
*
* Source-inspection is uglier than a runtime test but it has two
* advantages here: (a) it doesn't require a real PG cluster + psql
* binary in CI, (b) it pins the EXACT contract — "the declaration
* must be visible to the replay block" — which is the property
* that broke, more directly than a runtime test would.
*
* The follow-up "real-PG integration test in CI" (separate task)
* would replace this with an end-to-end exercise, at which point
* this can be deleted.
*/
const fs = require('fs');
const path = require('path');
describe('restoreService — PG branch scope contract (PR #596 review)', () => {
let src;
let lines;
beforeAll(() => {
src = fs.readFileSync(
path.join(__dirname, '..', '..', 'src', 'services', 'restoreService.js'),
'utf8',
);
lines = src.split(/\r?\n/);
});
/** Return the 1-based line number of the FIRST line matching `re`. */
function findFirst(re) {
const idx = lines.findIndex((l) => re.test(l));
return idx >= 0 ? idx + 1 : -1;
}
/** Return the 1-based line number of the LAST line matching `re`. */
function findLast(re) {
let last = -1;
lines.forEach((l, i) => { if (re.test(l)) last = i + 1; });
return last;
}
it('preservedMetaSnapshot lives on `this` and is initialised in the constructor', () => {
// PR #596 round 3 moved the snapshot from a block-scoped local to
// an instance variable so the replay can happen in `restore()`
// AFTER post-restore verification — preventing the replay row
// from inflating the row-count check.
//
// Contract:
// 1. The constructor initialises `this.preservedMetaSnapshot = []`
// 2. The `restore()` entry point resets it per call (no leak
// across consecutive runs in the singleton service instance)
// 3. `performDatabaseRestore` assigns to `this.preservedMetaSnapshot`
// inside the PG branch (must run before DROP)
// 4. The replay reads `this.preservedMetaSnapshot` — NOT a bare
// `preservedMeta` local — so a future refactor can't
// accidentally drop the snapshot half on the floor again.
const constructorInit = lines.some((l) =>
/this\.preservedMetaSnapshot\s*=\s*\[\s*\]/.test(l)
);
expect(constructorInit).toBe(true);
const assignmentSites = lines.filter((l) =>
/this\.preservedMetaSnapshot\s*=\s*(\[\s*\]|await\s+db)/.test(l)
);
// Constructor init + restore() per-run reset + the PG-branch
// assignment from db query. Three writes.
expect(assignmentSites.length).toBeGreaterThanOrEqual(3);
// No stray bare `preservedMeta` local-scoped declaration in
// performDatabaseRestore — would indicate someone re-introduced
// the round-1 footgun.
const dangerousLocalDecl = lines.filter((l) =>
/^\s*(let|const)\s+preservedMeta\s*=/.test(l)
);
expect(dangerousLocalDecl).toEqual([]);
});
it('every .count() result is coerced to Number before comparison', () => {
// PR #596 review caught a second PG-only landmine: pg-driver
// returns COUNT(*) as a string ("16" not 16) to preserve bigint
// precision. The original code compared `result.count !==
// expected.rowCount` and every match flagged as a mismatch on PG.
//
// The fix coerces with `Number(...)` at every comparison +
// interpolation site. This test catches a future regression where
// a refactor uses `.count` directly in a `===` / `!==` / `>` /
// `<` comparison without coercing.
//
// Heuristic: find every `.count` access in the file and make sure
// the line either:
// (a) wraps it in `Number(...)`, or
// (b) is purely an interpolation that already coerced upstream
// (e.g. `validation.warnings.push(`... ${eventCountN} ...`)`
// where eventCountN is the coerced local), or
// (c) is the docstring/comment line (filtered separately).
//
// We approximate this by listing every `.count` reference site
// and asserting that lines doing comparisons (`===`/`!==`/`>`/
// `<`/`>=`/`<=`) on a raw `.count` access without `Number(...)`
// around it are zero.
const dangerousLines = lines
.map((l, i) => ({ line: i + 1, text: l }))
// Filter to lines that compare a .count result
.filter(({ text }) => {
// Skip comments
if (/^\s*(\/\/|\*)/.test(text)) return false;
// Detect a `.count` (followed by `)` for `?.count` or by space/operator)
// being directly compared via ===/!==/>/<.
// Match the BAD pattern: `<something>.count <op> <something>`
// where <op> is === / !== / > / < / >= / <=
const bareCountInComparison = /\w+\??\.count\s*(?:!==|===|>=?|<=?)\s+/;
// ALLOW if the .count is preceded by `Number(` in the same line
const wrappedInNumber = /Number\(\s*\w+\??\.count/;
return bareCountInComparison.test(text) && !wrappedInNumber.test(text);
});
expect(dangerousLines).toEqual([]);
});
it('the completed-restore update sets was_successful=true', () => {
// Without this, every successful restore ends up with
// status='completed', was_successful=false — the dashboard's
// "last successful restore" widget then filters out the row +
// any future audit query gating on was_successful misses it.
// Caught locally + maintainer PR #596 review.
//
// Contract: the update payload that writes status='completed' on
// the SUCCESS branch ALSO includes was_successful: true. We pin
// it by source inspection so any future refactor of the success
// payload keeps both fields together.
// The success-branch update lives AFTER performPostRestoreVerification.
// There's also a `status: 'completed'` in the dry-run / early-return
// path (failure handling has its own block too) — we want the
// SUCCESS-branch one specifically.
const verifyLine = findFirst(/performPostRestoreVerification\s*\(/);
expect(verifyLine).toBeGreaterThan(0);
const completedStatusLineIdx = lines
.map((l, i) => ({ line: i + 1, text: l }))
.find(({ line, text }) =>
line > verifyLine && /status:\s*['"]completed['"]/.test(text)
);
expect(completedStatusLineIdx).toBeDefined();
// Look in the next ~10 lines for was_successful: true. The actual
// payload is small (no nested objects between status and the
// closing })), so a fixed-window search is reliable.
const window = lines.slice(
completedStatusLineIdx.line - 1,
completedStatusLineIdx.line + 10,
).join('\n');
expect(window).toMatch(/was_successful:\s*true/);
});
it('the safe migration runner is invoked after the replay in restore()', () => {
// Contract from PR #596 round 4: backups taken on older picpeak
// versions must restore COMPLETELY on a newer image — even if new
// migrations have been added since the backup was taken. The
// restore() flow shells out to the safe migration runner AFTER the
// operator-meta replay so the schema catches up to the running
// code WITHIN the restore boundary (not on the next container
// restart). Invoked as `node migrations/run-migrations-safe.js` —
// the runtime image ships no npm, so the former `npm run
// migrate:safe` would ENOENT into the non-fatal catch.
//
// Contract:
// 1. A run-migrations-safe shell-out exists somewhere in restoreService
// 2. It sits AFTER the replay drain — verification → replay →
// migrations is the documented order
// 3. It does NOT sit inside performDatabaseRestore (must run
// against the reinit'd pool from the parent restore())
const migrateLine = findFirst(/run-migrations-safe\.js/);
expect(migrateLine).toBeGreaterThan(0);
const replayLine = findLast(/this\.preservedMetaSnapshot\.length\s*>\s*0/);
expect(replayLine).toBeGreaterThan(0);
expect(migrateLine).toBeGreaterThan(replayLine);
// Must NOT live inside performDatabaseRestore (same scope as the
// replay check above).
const dbRestoreStart = findFirst(/async\s+performDatabaseRestore\s*\(/);
let dbRestoreEnd = -1;
for (let i = dbRestoreStart; i < lines.length; i++) {
if (/^ \}\s*$/.test(lines[i])) {
dbRestoreEnd = i + 1;
break;
}
}
expect(migrateLine < dbRestoreStart || migrateLine > dbRestoreEnd).toBe(true);
});
it('the replay site lives in restore() AFTER performPostRestoreVerification', () => {
// PR #596 round 3 moved the replay out of performDatabaseRestore
// and into the parent restore() method, sequenced AFTER the
// post-restore verification. Otherwise the replay's upserted row
// count was being flagged as a verification mismatch (e.g.
// "expected 190, got 191" because the fresh-install seeded
// `restore_allow_force_auto_upgraded` that wasn't in the backup).
//
// Contract: the line that drains `this.preservedMetaSnapshot`
// must come AFTER `performPostRestoreVerification` AND must NOT
// sit inside `performDatabaseRestore`.
const verificationLine = findFirst(/performPostRestoreVerification\s*\(/);
expect(verificationLine).toBeGreaterThan(0);
const replayLine = findLast(/this\.preservedMetaSnapshot\.length\s*>\s*0/);
expect(replayLine).toBeGreaterThan(0);
expect(replayLine).toBeGreaterThan(verificationLine);
// `performDatabaseRestore` must not contain the replay drain.
// Find the function bounds + assert no drain line falls inside.
const dbRestoreStart = findFirst(/async\s+performDatabaseRestore\s*\(/);
expect(dbRestoreStart).toBeGreaterThan(0);
// Find the closing brace of performDatabaseRestore. Lazy heuristic:
// the first `^ \}\s*$` (two-space indent + }) after the function
// start. Brittle to indent changes but unambiguous in this codebase.
let dbRestoreEnd = -1;
for (let i = dbRestoreStart; i < lines.length; i++) {
if (/^ \}\s*$/.test(lines[i])) {
dbRestoreEnd = i + 1;
break;
}
}
expect(dbRestoreEnd).toBeGreaterThan(dbRestoreStart);
// The replay drain line must be OUTSIDE [dbRestoreStart, dbRestoreEnd].
expect(replayLine < dbRestoreStart || replayLine > dbRestoreEnd).toBe(true);
});
});