mirror of
https://github.com/garrytan/gstack.git
synced 2026-10-31 22:14:25 +01:00
* v1.91.23.0 feat(cso): private two-architecture release scoring and self-checking runtime release gates
* Keep the Linux-only native-evidence and requalification tests out of the Windows lane
* Skip the two Linux-release-only checks on Windows in-file instead of growing the shard list
* fix(cso): opaque producer inputs and receipts; requalification triggers independent of cwd
Producer inputs are {schemaVersion: 2, cellRef, skill, source, execution} and
receipts bind cellRef + inputHash; collection rebuilds the cell from the
matrix, so a cell VM never learns its case or variant. requalificationTriggers
resolves Bun's cwd-relative metafile inputs against the repository root.
* test(evals): bound the review judge's thinking; guarded browser cleanup is not a probe
review/SKILL.md workflow: the ~64k-token bundle let default-effort thinking
spend all 8192 output tokens without an answer in CI. Medium effort with a
64k streaming cap kept 24/24 measured samples under 5k tokens and 63 s with
the same gating scores (as ship already does).
qa-only-no-fix: the fixture prompt routes browser cleanup through the deadline
guard, but the checkpoint detector treated every guarded run as a probe that
needs a checkpoint. A child that only closes the tab with the supplied browse
binary, after the last probe, is cleanup; anything else stays a probe. The
captured CI transcript (run 37252022510) now passes and new scenarios pin
the negative controls.
* test(evals): keep the review judge at default effort; only lift its output cap
Truncation was the defect: thinking and the answer share max_tokens, and the
~64k-token review bundle once spent all 8192. The registration keeps default
effort and response format and gets ship's 64k streaming cap; the guard test
pins exactly that.
69 lines
4.3 KiB
TypeScript
69 lines
4.3 KiB
TypeScript
#!/usr/bin/env bun
|
|
/**
|
|
* Requalification triggers: the inputs a runtime qualification was measured
|
|
* under. A promoted catalog records them; a catalog whose recorded triggers
|
|
* differ from the current source is older than its inputs and must be
|
|
* requalified (or its runtimes withdrawn) before it ships.
|
|
*/
|
|
import * as fs from 'node:fs';
|
|
import * as path from 'node:path';
|
|
import { canonical, sha256 } from '../lib/cso/contracts';
|
|
import { ISOLATION_POLICY_HASH } from '../lib/cso/docker';
|
|
import { CSO_HELPER_ABI, isEvaluationRuntimeCatalog, type RequalificationTriggers, type RuntimeCatalog } from '../lib/cso/runtime-catalog';
|
|
|
|
const ROOT = path.resolve(import.meta.dir, '..');
|
|
/** Compiled into every runtime image as gstack-cso-verifier and gstack-cso-preparation. */
|
|
export const IMAGE_HELPER_ENTRYPOINTS = ['lib/cso/verifier.ts', 'lib/cso/preparation-container.ts'] as const;
|
|
export const IMAGE_CONTEXT = 'lib/cso/images';
|
|
/** Image-context files that describe qualification rather than enter an image; build-inputs.json is its own trigger. */
|
|
const IMAGE_CONTEXT_EXCLUDED = new Set(['README.md', 'qualification.json', 'build-inputs.json']);
|
|
|
|
/** Every repository file whose bytes enter a runtime image: the helpers' import closure plus the image build context. */
|
|
export async function imageSourceFiles(root = ROOT): Promise<string[]> {
|
|
const build = await Bun.build({ entrypoints: IMAGE_HELPER_ENTRYPOINTS.map(entry => path.join(root, entry)), target: 'bun', metafile: true });
|
|
if (!build.success || !build.metafile) throw new Error('REQUALIFICATION_SOURCE_CLOSURE_FAILED');
|
|
// Bun reports metafile inputs relative to the process working directory, not to the entrypoints.
|
|
const closure = Object.keys(build.metafile.inputs).map(input => path.relative(root, path.resolve(process.cwd(), input)).split(path.sep).join('/'));
|
|
if (closure.some(file => file.startsWith('..') || file.includes('node_modules/'))) throw new Error('REQUALIFICATION_SOURCE_OUTSIDE_REPOSITORY');
|
|
const context = fs.readdirSync(path.join(root, IMAGE_CONTEXT), { withFileTypes: true })
|
|
.filter(entry => entry.isFile() && !IMAGE_CONTEXT_EXCLUDED.has(entry.name))
|
|
.map(entry => `${IMAGE_CONTEXT}/${entry.name}`);
|
|
return [...new Set([...closure, ...context])].sort();
|
|
}
|
|
|
|
export async function requalificationTriggers(root = ROOT): Promise<RequalificationTriggers> {
|
|
const files = await imageSourceFiles(root);
|
|
const buildInputs = JSON.parse(fs.readFileSync(path.join(root, IMAGE_CONTEXT, 'build-inputs.json'), 'utf8'));
|
|
if (typeof buildInputs?.revision !== 'string' || !buildInputs.revision) throw new Error('INVALID_BUILD_INPUTS_REVISION');
|
|
return {
|
|
helperAbi: CSO_HELPER_ABI,
|
|
isolationPolicyHash: ISOLATION_POLICY_HASH,
|
|
preparationSha256: sha256(canonical(files.map(file => [file, sha256(fs.readFileSync(path.join(root, file)))]))),
|
|
buildInputsRevision: buildInputs.revision,
|
|
};
|
|
}
|
|
|
|
/** Names of the triggers that changed since the catalog's runtimes were qualified; empty when nothing is promoted. */
|
|
export function staleRequalificationTriggers(catalog: RuntimeCatalog, current: RequalificationTriggers): string[] {
|
|
if (isEvaluationRuntimeCatalog(catalog) || catalog.runtimes.length === 0) return [];
|
|
const recorded = catalog.promotion?.requalification;
|
|
if (!recorded) return Object.keys(current).sort();
|
|
return (Object.keys(current) as Array<keyof RequalificationTriggers>).filter(key => recorded[key] !== current[key]).sort();
|
|
}
|
|
|
|
if (import.meta.main) {
|
|
try {
|
|
const [command, file, ...rest] = process.argv.slice(2);
|
|
if (command === 'triggers' && file === undefined) {
|
|
process.stdout.write(JSON.stringify(await requalificationTriggers(), null, 2) + '\n');
|
|
} else if (command === 'check' && file && rest.length === 0) {
|
|
const stale = staleRequalificationTriggers(JSON.parse(fs.readFileSync(path.resolve(file), 'utf8')), await requalificationTriggers());
|
|
if (stale.length) throw new Error(`STALE_RUNTIME_QUALIFICATION: ${stale.join(', ')} changed since qualification; requalify the runtimes or withdraw them from the catalog`);
|
|
process.stdout.write('REQUALIFICATION CURRENT\n');
|
|
} else throw new Error('Usage: cso-requalification.ts triggers | check <runtime-catalog.json>');
|
|
} catch (error) {
|
|
process.stderr.write((error instanceof Error ? error.message : 'REQUALIFICATION_CHECK_FAILED') + '\n');
|
|
process.exitCode = 1;
|
|
}
|
|
}
|