149 lines
13 KiB
JavaScript
149 lines
13 KiB
JavaScript
import test from 'node:test';
|
|
import assert from 'node:assert/strict';
|
|
import fs from 'node:fs';
|
|
import path from 'node:path';
|
|
import { fixture, launchWorkbench, until, compatibleFixture, call, completion, assertToolPairs } from './helpers.mjs';
|
|
import { batchFixture, assertBatchBytes } from './fixtures/change-set-project.mjs';
|
|
import { Runtime } from '../src/server/runtime.mjs';
|
|
import { Provider } from '../src/server/provider.mjs';
|
|
|
|
test('batch APIs enforce ownership and review fingerprints, apply once, and persist the completed batch across restart', { timeout: 30000 }, async t => {
|
|
const f = fixture(t), b = batchFixture(f); f.store.close();
|
|
const host = await launchWorkbench(t, f.data), route = '/sessions/' + f.session.id;
|
|
const first = await host.api(route + '/change-sets', { changeIds: b.ids }), set = first.execution.changeSets.at(-1);
|
|
assert.equal((await host.api(route + '/change-sets', { changeIds: b.ids })).execution.changeSets.at(-1).id, set.id);
|
|
assertBatchBytes(assert, f, b.before);
|
|
await assert.rejects(host.api(route + '/change-sets/' + set.id + '/apply', {}), /指纹/);
|
|
await assert.rejects(host.api(route + '/changes/' + b.ids[0] + '/apply', {}), /封存/);
|
|
await assert.rejects(host.api(route + '/messages', { content: 'different task' }), /变更集/);
|
|
const other = await host.api('/sessions', { projectId: f.project.id });
|
|
await assert.rejects(host.api('/sessions/' + other.id + '/change-sets/' + set.id), /不属于/);
|
|
await host.api(route + '/change-sets/' + set.id + '/apply', { fingerprint: set.fingerprint });
|
|
let applied;
|
|
await until(async () => { applied = await host.api(route); return applied.execution.changeSets.at(-1).status === 'completed'; });
|
|
assertBatchBytes(assert, f, b.after);
|
|
await assert.rejects(host.api(route + '/change-sets/' + set.id + '/apply', { fingerprint: set.fingerprint }), /重复应用/);
|
|
await host.close();
|
|
const restarted = await launchWorkbench(t, f.data), restored = await restarted.api(route);
|
|
assert.deepEqual(restored.execution.changeSets, applied.execution.changeSets); assertBatchBytes(assert, f, b.after);
|
|
});
|
|
|
|
test('delivery reports bind actual batch identity and reject corrupted snapshot evidence without changing project files', { timeout: 30000 }, async t => {
|
|
const f = fixture(t), b = batchFixture(f), set = await b.manager.prepare(f.session, b.ids);
|
|
await (await b.manager.start(f.session, set.id, set.fingerprint)).promise;
|
|
const snapshot = f.store.owned('artifact', set.files[1].snapshotArtifactId, f.session.id), snapshotPath = path.join(f.store.artifactStore.root, snapshot.relative);
|
|
f.store.close(); const host = await launchWorkbench(t, f.data), route = '/sessions/' + f.session.id;
|
|
const valid = await host.api(route + '/delivery-reports', {}), jsonArtifact = valid.execution.artifacts.slice().reverse().find(item => item.kind === 'delivery-report');
|
|
const preview = await host.api('/artifacts/' + jsonArtifact.id + '?session=' + f.session.id + '&limit=16000'), report = JSON.parse(preview.content);
|
|
assert.equal(report.changeSets[0].id, set.id); assert.equal(report.changeSets[0].snapshotVerified, true);
|
|
assert.ok(report.changes.every(change => change.changeSetId === set.id)); assert.equal(report.status, 'needs_checks');
|
|
fs.writeFileSync(snapshotPath, 'Corrupt owned QA snapshot');
|
|
const invalid = await host.api(route + '/delivery-reports', {}), artifact = invalid.execution.artifacts.slice().reverse().find(item => item.kind === 'delivery-report');
|
|
const invalidReport = JSON.parse((await host.api('/artifacts/' + artifact.id + '?session=' + f.session.id + '&limit=16000')).content);
|
|
assert.equal(invalidReport.status, 'needs_verification'); assert.equal(invalidReport.changeSets[0].snapshotVerified, false); assertBatchBytes(assert, f, b.after);
|
|
});
|
|
|
|
test('a compatible model can seal proposals through the real HTTP loop while application stays behind user review', { timeout: 60000 }, async t => {
|
|
const f = fixture(t); f.store.close(); let count = 0;
|
|
const provider = await compatibleFixture(t, (url, input) => {
|
|
if (!input?.messages) return { data: [{ id: 'fixture-model' }] };
|
|
count++;
|
|
if (count === 1) return completion([
|
|
call('first-file', 'propose_file', { path: 'src/new.mjs', content: 'export const newValue = 1;\n', reason: '新增模块' }),
|
|
call('second-file', 'propose_file', { path: 'docs/new.md', content: '# New\n', reason: '补充说明' }),
|
|
], '准备两个文件提案。');
|
|
if (count === 2) {
|
|
const ids = input.messages.filter(message => message.role === 'tool').map(message => JSON.parse(message.content).change_id).filter(Boolean);
|
|
return completion([call('seal-batch', 'prepare_change_set', { change_ids: ids })], '封存整批审查范围。');
|
|
}
|
|
return completion([], '文件批次已封存,等待用户审查应用。');
|
|
});
|
|
const host = await launchWorkbench(t, f.data);
|
|
await host.api('/config', { mode: 'compatible', baseUrl: provider.baseUrl, model: 'fixture-model', apiKey: 'SYNTHETIC-QA-ONLY' });
|
|
await host.api('/sessions/' + f.session.id + '/messages', { content: '创建两个文件并组合审查' });
|
|
let session;
|
|
await until(async () => { session = await host.api('/sessions/' + f.session.id); return session.status === 'needs_review'; }, 30000);
|
|
assert.equal(session.execution.changeSets.length, 1); assert.equal(session.execution.changeSets[0].status, 'prepared');
|
|
assert.equal(session.execution.changeSets[0].files.length, 2); assert.equal(fs.existsSync(path.join(f.projectRoot, 'src')), false);
|
|
assert.equal(fs.existsSync(path.join(f.projectRoot, 'docs')), false);
|
|
assert.equal(session.events.find(event => event.tool === 'prepare_change_set').phase, 'completed');
|
|
assert.equal(count, 2); assert.equal(session.execution.run.stopReason, 'change_set_review');
|
|
const tools = provider.requests.find(request => request.input?.tools)?.input.tools.map(tool => tool.function.name);
|
|
assert.ok(tools.includes('prepare_change_set')); assert.ok(!tools.includes('apply_change_set')); assert.ok(!tools.includes('rollback_change_set'));
|
|
});
|
|
|
|
test('a sealed batch at the step limit hands off and continues the same Task into actual checks without replaying saved tools', { timeout: 60000 }, async t => {
|
|
const f = fixture(t);
|
|
fs.mkdirSync(path.join(f.projectRoot, 'test'));
|
|
const acceptance = "import test from 'node:test'; import assert from 'node:assert/strict'; import { left } from '../src/left.mjs'; import { right } from '../src/right.mjs';\ntest('left value', () => assert.equal(left, 1)); test('right value', () => assert.equal(right, 2));\n";
|
|
fs.writeFileSync(path.join(f.projectRoot, 'test/values.test.mjs'), acceptance);
|
|
fs.writeFileSync(path.join(f.projectRoot, 'package.json'), JSON.stringify({ type: 'module', scripts: { test: 'node --test test/*.test.mjs' } }));
|
|
f.store.close(); let count = 0;
|
|
const model = await compatibleFixture(t, (url, input) => {
|
|
if (!input?.messages) return { data: [{ id: 'batch-continuation-fixture' }] };
|
|
count++;
|
|
if (count === 1) return completion([
|
|
call('paused-left', 'propose_file', { path: 'src/left.mjs', content: 'export const left = 1;\n', reason: '第一份受控模块' }),
|
|
call('paused-right', 'propose_file', { path: 'src/right.mjs', content: 'export const right = 2;\n', reason: '第二份受控模块' }),
|
|
]);
|
|
const results = input.messages.filter(message => message.role === 'tool').map(message => JSON.parse(message.content));
|
|
if (count === 2) return completion([call('paused-seal', 'prepare_change_set', { change_ids: results.map(result => result.change_id).filter(Boolean) })]);
|
|
if (count === 3) return completion([call('continued-check-plan', 'plan_checks', { names: ['test'] })]);
|
|
if (count === 4) return completion([call('continued-actual-check', 'run_checks', { plan_id: results.findLast(result => result.checks && result.fingerprint).id })]);
|
|
return completion([], '原任务的两份文件已通过真实工程检查。');
|
|
});
|
|
const host = await launchWorkbench(t, f.data), route = '/sessions/' + f.session.id;
|
|
await host.api('/config', { mode: 'compatible', baseUrl: model.baseUrl, model: 'batch-continuation-fixture', maxSteps: 2 });
|
|
await host.api(route + '/messages', { content: '完成原始双模块工程目标并运行测试' });
|
|
let paused;
|
|
await until(async () => { paused = await host.api(route); return paused.status === 'needs_review'; }, 30000);
|
|
assert.equal(paused.execution.run.stopReason, 'change_set_review'); assert.equal(count, 2);
|
|
const set = paused.execution.changeSets.at(-1), firstRun = structuredClone(paused.execution.run);
|
|
assert.equal(set.status, 'prepared'); assert.equal(fs.existsSync(path.join(f.projectRoot, 'src')), false);
|
|
const savedMessages = (await host.api(route + '/export')).session.providerMessages;
|
|
const savedResults = savedMessages.filter(message => message.role === 'tool'); assert.equal(savedResults.length, 3);
|
|
await assert.rejects(host.api(route + '/resume', {}), { status: 409, message: /请先审查待处理改动/ });
|
|
await assert.rejects(host.api(route + '/continue-reviewed', {}), { status: 409 });
|
|
await host.api(route + '/change-sets/' + set.id + '/apply', { fingerprint: set.fingerprint });
|
|
await until(async () => (await host.api(route)).execution.changeSets.at(-1).status === 'completed', 30000);
|
|
await host.api('/config', { ...(await host.api('/bootstrap')).config, maxSteps: 5 });
|
|
await host.api(route + '/continue-reviewed', {});
|
|
let waiting;
|
|
await until(async () => { waiting = await host.api(route); return waiting.approvals.some(approval => approval.status === 'pending'); }, 30000);
|
|
await host.api(route + '/approvals/' + waiting.approvals.find(approval => approval.status === 'pending').id, { allow: true });
|
|
let completed;
|
|
await until(async () => { completed = await host.api(route); return completed.status === 'completed'; }, 30000);
|
|
assert.equal(completed.execution.task.id, paused.execution.task.id);
|
|
assert.equal(completed.execution.task.goal, '完成原始双模块工程目标并运行测试');
|
|
assert.equal(completed.execution.run.resumeOf, firstRun.id); assert.equal(completed.execution.task.verification.status, 'passed');
|
|
assert.equal(completed.execution.checks.length, 1); assert.equal(completed.execution.checks[0].results[0].tests.passed, 2);
|
|
assert.equal(completed.execution.runs.find(run => run.id === firstRun.id).status, 'needs_review');
|
|
assert.deepEqual(completed.execution.runs.find(run => run.id === firstRun.id).usage, firstRun.usage);
|
|
assert.equal(completed.execution.changeSets[0].runId, firstRun.id); assert.equal(count, 5);
|
|
const exported = (await host.api(route + '/export')).session.providerMessages;
|
|
assert.deepEqual(exported.filter(message => message.role === 'tool').slice(0, 3), savedResults); assertToolPairs(exported);
|
|
assert.equal(completed.events.filter(event => event.kind === 'tool' && ['propose_file', 'prepare_change_set'].includes(event.tool)).length, 3);
|
|
assert.equal(fs.readFileSync(path.join(f.projectRoot, 'test/values.test.mjs'), 'utf8'), acceptance);
|
|
await assert.rejects(host.api(route + '/continue-reviewed', {}), { status: 409 });
|
|
});
|
|
|
|
test('applied batches cannot turn cancellation, other budgets, restart or unknown commands into a review continuation', async t => {
|
|
const f = fixture(t), b = batchFixture(f), set = await b.manager.prepare(f.session, b.ids);
|
|
await (await b.manager.start(f.session, set.id, set.fingerprint)).promise;
|
|
const runtime = new Runtime(f.store, new Provider()), run = f.store.core.run.get(f.session.runId);
|
|
for (const reason of ['cancelled', 'time_limit', 'token_limit', 'task_token_limit', 'restart']) {
|
|
run.status = 'paused'; run.stopReason = reason;
|
|
const before = JSON.stringify(f.store.publicSession(f.session));
|
|
assert.throws(() => runtime.continueReviewed(f.session.id), { status: 409 });
|
|
assert.equal(JSON.stringify(f.store.publicSession(f.session)), before); assertBatchBytes(assert, f, b.after);
|
|
}
|
|
const invocation = call('unknown-command', 'run_command', { command: 'echo owned-fixture', reason: 'Unknown result must be reviewed' });
|
|
const event = f.store.event(f.session, { kind: 'tool', tool: 'run_command', toolCallId: invocation.id, phase: 'running' });
|
|
f.store.beginInvocation(f.session, invocation, run, event); f.store.invocationState(f.session, event, 'running');
|
|
f.store.interruptPending(f.session, run); f.session.status = 'paused'; f.store.finishRun(f.session, run, 'step_limit');
|
|
const before = JSON.stringify(f.store.publicSession(f.session));
|
|
assert.equal(f.store.publicSession(f.session).execution.progress.requiresVerification.length, 1);
|
|
assert.throws(() => runtime.continueReviewed(f.session.id), { status: 409 });
|
|
assert.equal(JSON.stringify(f.store.publicSession(f.session)), before); assertBatchBytes(assert, f, b.after);
|
|
});
|