Skip to content

test: pin the srt-macos probe PATH and scale the BYOK serial timeout budget #397

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
hetaoBackend merged 2 commits into from
Sep 29, 2026
Merged

test: pin the srt-macos probe PATH and scale the BYOK serial timeout budget #397

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
b6b8282
test(sandbox): pin the probe PATH instead of inheriting the ambient one
hetaoBackend Sep 29, 2026
1d7ca05
test(byok): scale the serial BYOK budget by platform
hetaoBackend Sep 29, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
17 changes: 14 additions & 3 deletions packages/local-runtime-v2/src/service/sandbox/backend/srt-macos.test.ts
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -1487,6 +1487,17 @@ async function expectMandatoryDenySurface(
}
}

/**
* Probes must measure the sandbox, not the caller's shell. Every probe spawns a
* real process and resolves its command through `PATH`, so inheriting the
* ambient one lets a wrapper ahead of `/bin` decide what `rm` means — for
* example the recoverable-delete shim MCode installs into agent shells, whose
* trash move the sandbox then denies, turning a correct allow-probe into a
* reported sandbox regression. Pin the probes to the system binaries they
* actually use (`cat`, `rm`, `rmdir`, `find`, `mv`, `printf`, `git`, `true`).
*/
const PROBE_PATH = "/usr/bin:/bin";

async function expectProbeResult(
command: string,
filesystem: SandboxInvocationFilesystemPolicy,
Expand Down Expand Up @@ -1520,7 +1531,7 @@ async function runProbe(
cwd: fixture.workspace,
baseEnv: {
HOME: process.env.HOME ?? "/var/empty",
PATH: process.env.PATH ?? "/usr/bin:/bin",
PATH: PROBE_PATH,
},
sandboxTempDir: fixture.sessionTemp,
commandId,
Expand Down Expand Up @@ -1593,7 +1604,7 @@ function createGitProbeFixture(): GitProbeFixture {
function gitProbeEnv(fixture: GitProbeFixture): Record<string, string> {
return {
HOME: fixture.home,
PATH: process.env.PATH ?? "/usr/bin:/bin",
PATH: PROBE_PATH,
GIT_CONFIG_NOSYSTEM: "1",
GIT_AUTHOR_NAME: "probe",
GIT_AUTHOR_EMAIL: "probe@example.invalid",
Expand Down Expand Up @@ -1712,7 +1723,7 @@ async function wrapProfile(
await profileBackend.wrap({
command: "true",
cwd: profileFixture,
baseEnv: { PATH: process.env.PATH },
baseEnv: { PATH: PROBE_PATH },
sandboxTempDir: profileFixture,
commandId: opaqueId,
commandText: opaqueId,
Expand Down
10 changes: 9 additions & 1 deletion test/byok.test.mjs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -22,10 +22,18 @@ import { parse as parseYaml, stringify as stringifyYaml } from "yaml";
import { withoutProxyEnvironment } from "./offline-environment.mjs";

const cli = process.env.MCODE_TEST_CLI ?? fileURLToPath(new URL("../dist/cli.js", import.meta.url));
// This case is 30+ serial CLI spawns, each booting the whole runtime, so its
// wall-clock cost tracks machine speed rather than the transport it asserts.
// A flat budget sized for an idle laptop turns any busy runner (parallel build,
// shared CI host) into a `testTimeoutFailure` that reports a product failure
// while every assertion would have passed. Budget it the way smoke.test.mjs
// budgets runtime startup: a base allowance plus a Windows multiplier for
// slower process spawn, rather than a number that only holds on one machine.
const byokSerialTimeoutMs = process.platform === "win32" ? 360000 : 180000;
// This fixture validates BYOK transport and real Runtime persistence, not model quality.
test(
"BYOK runs without managed login and resumes its saved conversation",
{ timeout: 90000 },
{ timeout: byokSerialTimeoutMs },
async (t) => {
const fixtureDir = mkdtempSync(path.join(tmpdir(), "minimax-code-byok-"));
const dataDir = path.join(fixtureDir, "data");
Expand Down
Loading