Compare commits

..
Author SHA1 Message Date
SinabinaandClaude Fable 5 39f750e831 fix(migrations): scope v0.32.2 dirty-check to targeted sources; surface failed phase detail
- phaseBFenceFacts now queries legacy rows FIRST and dirty-checks only
  the source_ids it will actually write into. Zero fenceable rows (or
  rows scoped to clean sources) no longer fail on an unrelated dirty
  source. Targeted-dirty-source refusal unchanged. Fixes #927.
- apply-migrations now prints each failed phase's name + detail to
  stderr alongside 'reported status=failed', instead of burying the
  actionable message in the ledger. Fixes #921.

Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
2026-07-21 14:24:42 -07:00
8 changed files with 89 additions and 102 deletions
+1 -1
View File
@@ -2387,7 +2387,7 @@ JOBS (Minions)
jobs get <id> Job details + history
jobs cancel <id> Cancel job
jobs retry <id> Re-queue failed/dead job
jobs prune [--older-than 30d] [--status s,..] Clean old terminal jobs (0d = no age floor)
jobs prune [--older-than 30d] Clean old jobs
jobs stats Job health dashboard
jobs work [--queue Q] Start worker daemon (Postgres only)
+7
View File
@@ -438,6 +438,13 @@ export async function runApplyMigrations(args: string[]): Promise<void> {
const result = await m.orchestrator(orchestratorOptsFrom(cli));
if (result.status === 'failed') {
console.error(`Migration v${m.version} reported status=failed.`);
// Surface each failed phase's detail — the ledger records it, but
// the operator needs it on stderr to act (#921).
for (const p of result.phases) {
if (p.status === 'failed') {
console.error(` phase ${p.name}: ${p.detail ?? '(no detail)'}`);
}
}
// Record the attempt as 'partial' (not 'complete') so the cap counts
// it. Don't let a failed orchestrator look like it never ran.
try {
+5 -33
View File
@@ -106,21 +106,6 @@ export function parseMaxRssFlag(args: string[]): number | undefined {
return parsed;
}
/** Terminal statuses `jobs prune --status` accepts (PR #2282). Matches what
* queue.prune can safely delete; anything else (waiting/active/…) is live. */
export const PRUNE_STATUSES = ['completed', 'failed', 'dead', 'cancelled'] as const satisfies readonly MinionJobStatus[];
/** Parse a `--status a,b,c` value into prune statuses. Throws on any value
* outside PRUNE_STATUSES (fail-fast, mirrors parseNiceValue). */
export function parsePruneStatuses(raw: string): MinionJobStatus[] {
const requested = raw.split(',').map(s => s.trim()).filter(Boolean);
const invalid = requested.filter(s => !(PRUNE_STATUSES as readonly string[]).includes(s));
if (requested.length === 0 || invalid.length > 0) {
throw new Error(`--status accepts a comma-separated subset of [${PRUNE_STATUSES.join(', ')}]${invalid.length ? `. Invalid: ${invalid.join(', ')}` : ''}`);
}
return requested as MinionJobStatus[];
}
/** Parse `--nice N` (then `GBRAIN_NICE` env). Returns:
* - undefined if absent (no priority change — inherit)
* - the validated integer in [-20, 19] otherwise
@@ -223,9 +208,7 @@ USAGE
gbrain jobs get <id>
gbrain jobs cancel <id>
gbrain jobs retry <id>
gbrain jobs prune [--older-than 30d] [--status completed,failed,dead,cancelled]
(--older-than 0d = no age floor: deletes ALL
matching terminal jobs; pair with --status)
gbrain jobs prune [--older-than 30d]
gbrain jobs delete <id>
gbrain jobs stats
gbrain jobs smoke
@@ -617,27 +600,16 @@ HANDLER TYPES (built in)
case 'prune': {
const olderThanStr = parseFlag(args, '--older-than') ?? '30d';
const days = parseInt(olderThanStr, 10);
if (isNaN(days) || days < 0) {
console.error('Error: --older-than must be a non-negative number (days). Example: --older-than 30d; --older-than 0d removes the age floor (deletes ALL matching terminal jobs).');
if (isNaN(days) || days <= 0) {
console.error('Error: --older-than must be a positive number (days). Example: --older-than 30d');
process.exit(1);
}
const statusFlag = parseFlag(args, '--status');
let statuses: MinionJobStatus[] | undefined;
if (statusFlag !== undefined) {
try { statuses = parsePruneStatuses(statusFlag); }
catch (e) { console.error(`Error: ${e instanceof Error ? e.message : String(e)}`); process.exit(1); }
}
try { await queue.ensureSchema(); }
catch (e) { console.error(e instanceof Error ? e.message : String(e)); process.exit(1); }
const count = await queue.prune({
olderThan: new Date(Date.now() - days * 86400000),
...(statuses ? { status: statuses } : {}),
});
const statusLabel = statuses ? statuses.join('+') : 'completed+dead+cancelled';
const ageLabel = days === 0 ? 'regardless of age' : `older than ${days} days`;
console.log(`Pruned ${count} ${statusLabel} jobs ${ageLabel}.`);
const count = await queue.prune({ olderThan: new Date(Date.now() - days * 86400000) });
console.log(`Pruned ${count} jobs older than ${days} days.`);
break;
}
+15 -11
View File
@@ -186,17 +186,6 @@ async function phaseBFenceFacts(
const localPathById = new Map<string, string | null>();
for (const s of sources) localPathById.set(s.id, s.local_path);
// Dirty-tree refusal: check every source's local_path before writing.
for (const [id, localPath] of localPathById) {
if (localPath && isLocalPathDirty(localPath)) {
return {
name: 'fence_facts',
status: 'failed',
detail: `source "${id}" has uncommitted changes in ${localPath}. Commit or stash, then re-run.`,
};
}
}
// Walk legacy rows in (source_id, entity_slug) groups for per-page
// atomic writes.
const legacy = await engine.executeRaw<LegacyFactRow>(
@@ -235,6 +224,21 @@ async function phaseBFenceFacts(
groups.set(key, list);
}
// Dirty-tree refusal: check ONLY the sources we are about to write
// into. A dirty tree in an unrelated source (or zero fenceable rows
// at all) must not block a no-op or a targeted backfill (#927).
const targetSourceIds = new Set([...groups.keys()].map(k => k.split('\0')[0]));
for (const id of targetSourceIds) {
const localPath = localPathById.get(id);
if (localPath && isLocalPathDirty(localPath)) {
return {
name: 'fence_facts',
status: 'failed',
detail: `source "${id}" has uncommitted changes in ${localPath}. Commit or stash, then re-run.`,
};
}
}
for (const [key, group] of groups) {
const [sourceId, entitySlug] = key.split('\0');
const localPath = localPathById.get(sourceId)!;
+13
View File
@@ -180,3 +180,16 @@ describe('runApplyMigrations exit codes (v0.36.1.x #1062)', () => {
expect(src).toMatch(/All migrations up to date[\s\S]{0,80}process\.exit\(0\)/);
});
});
// #921: a failed orchestrator must print each failed phase's detail to
// stderr — not just "reported status=failed" — so the operator can act
// without digging through the ledger.
describe('failed migration prints phase detail (#921)', () => {
test('runner loops result.phases and console.errors failed phase details', async () => {
const { readFileSync } = await import('fs');
const src = readFileSync('src/commands/apply-migrations.ts', 'utf8');
expect(src).toMatch(
/reported status=failed[\s\S]{0,400}for \(const p of result\.phases\)[\s\S]{0,200}p\.status === 'failed'[\s\S]{0,200}console\.error\([\s\S]{0,80}p\.name[\s\S]{0,80}p\.detail/,
);
});
});
-30
View File
@@ -1,30 +0,0 @@
/**
* Unit tests for parsePruneStatuses (PR #2282) `jobs prune --status` parsing.
*/
import { describe, test, expect } from 'bun:test';
import { parsePruneStatuses, PRUNE_STATUSES } from '../src/commands/jobs.ts';
describe('parsePruneStatuses', () => {
test('parses a single status', () => {
expect(parsePruneStatuses('failed')).toEqual(['failed']);
});
test('parses a comma-separated list with whitespace', () => {
expect(parsePruneStatuses(' completed, dead ')).toEqual(['completed', 'dead']);
});
test('accepts every documented terminal status', () => {
expect(parsePruneStatuses(PRUNE_STATUSES.join(','))).toEqual([...PRUNE_STATUSES]);
});
test('throws on non-terminal statuses', () => {
expect(() => parsePruneStatuses('waiting')).toThrow(/Invalid: waiting/);
expect(() => parsePruneStatuses('completed,active')).toThrow(/Invalid: active/);
});
test('throws on empty value', () => {
expect(() => parsePruneStatuses('')).toThrow(/comma-separated subset/);
expect(() => parsePruneStatuses(',')).toThrow(/comma-separated subset/);
});
});
+48 -1
View File
@@ -10,10 +10,11 @@
* __setTestEngineOverride so we don't need a configured brain.
*/
import { describe, test, expect, beforeAll, afterAll, beforeEach } from 'bun:test';
import { describe, test, expect, beforeAll, afterAll, beforeEach, afterEach } from 'bun:test';
import { mkdtempSync, rmSync, existsSync, readFileSync, writeFileSync, mkdirSync } from 'node:fs';
import { tmpdir } from 'node:os';
import { join } from 'node:path';
import { execFileSync } from 'node:child_process';
import { PGLiteEngine } from '../src/core/pglite-engine.ts';
import { v0_32_2, __setTestEngineOverride, __testing } from '../src/commands/migrations/v0_32_2.ts';
@@ -238,6 +239,52 @@ describe('phaseBFenceFacts — happy path backfill', () => {
});
});
describe('phaseBFenceFacts — dirty-tree refusal scoping (#927)', () => {
let dirtyDir: string;
beforeEach(async () => {
// A second source whose local_path is a git repo with uncommitted changes.
dirtyDir = mkdtempSync(join(tmpdir(), 'mig-v0_32_2-dirty-'));
execFileSync('git', ['-C', dirtyDir, 'init', '-q']);
writeFileSync(join(dirtyDir, 'uncommitted.md'), 'dirty', 'utf-8');
// eslint-disable-next-line @typescript-eslint/no-explicit-any
await (engine as any).db.query(
`INSERT INTO sources (id, name, local_path) VALUES ('other', 'other', $1)`,
[dirtyDir],
);
});
afterEach(async () => {
// eslint-disable-next-line @typescript-eslint/no-explicit-any
await (engine as any).db.query(`DELETE FROM sources WHERE id = 'other'`);
rmSync(dirtyDir, { recursive: true, force: true });
});
test('no legacy facts at all → complete, dirty unrelated source ignored', async () => {
const r = await __testing.phaseBFenceFacts(engine, OPTS);
expect(r.status).toBe('complete');
expect(r.detail).toContain('scanned=0');
});
test('facts scoped to a clean source fence despite dirty unrelated source', async () => {
await seedLegacyFact({ entity_slug: 'people/alice', fact: 'Founded Acme' });
const r = await __testing.phaseBFenceFacts(engine, OPTS);
expect(r.status).toBe('complete');
expect(r.detail).toContain('fenced=1');
expect(existsSync(join(brainDir, 'people/alice.md'))).toBe(true);
});
test('still refuses when the TARGETED source is dirty', async () => {
await seedLegacyFact({ entity_slug: 'people/alice', fact: 'F1', source_id: 'other' });
const r = await __testing.phaseBFenceFacts(engine, OPTS);
expect(r.status).toBe('failed');
expect(r.detail).toContain('"other"');
expect(r.detail).toContain('uncommitted changes');
});
});
describe('phaseCVerify', () => {
test('returns complete when fence + DB row counts match', async () => {
await seedLegacyFact({ entity_slug: 'people/alice', fact: 'F1' });
-26
View File
@@ -702,32 +702,6 @@ describe('MinionQueue: Prune', () => {
const count = await queue.prune({ olderThan: new Date(Date.now() + 86400000) }); // future date = prune everything old enough
expect(count).toBe(1); // only the cancelled one
});
// PR #2282: `jobs prune --status` passes an explicit status subset through.
test('status filter prunes only the requested terminal statuses', async () => {
const cancelled = await queue.add('sync', {});
await queue.cancelJob(cancelled.id);
const dead = await queue.add('embed', {}, { max_attempts: 1 });
await queue.claim('tok1', 30000, 'default', ['embed']);
await queue.failJob(dead.id, 'tok1', 'boom', 'dead');
const count = await queue.prune({ olderThan: new Date(Date.now() + 86400000), status: ['dead'] });
expect(count).toBe(1); // only the dead one
const remaining = await queue.getJobs({ status: 'cancelled' });
expect(remaining.length).toBe(1);
});
// PR #2282: `--older-than 0d` = no age floor — olderThan of "now" deletes
// terminal jobs that finished moments ago.
test('olderThan now (0d semantics) prunes just-terminated jobs', async () => {
const job = await queue.add('sync', {});
await queue.cancelJob(job.id);
await new Promise(r => setTimeout(r, 5)); // ensure updated_at < now
const count = await queue.prune({ olderThan: new Date() });
expect(count).toBe(1);
});
});
// --- Stats (1 test) ---