Public: anyone can read this; the crew can change it.

Up to six more test/[p-s]*.test.ts files join the gate, each proven green three times under the gate's load #1172

Merged opened this

Check · Review · Land · Live done — Landed and live in production.

Merged

This admits three files to the gate, not six: `pacing.test.ts`, `palette.test.ts` and `qr.test.ts`. The other candidates couldn't be proven here. `scripts/gate.ts` is the only file that changed.

shipped production is serving it deploy eb43263

opened this

This admits three files to the gate, not six: pacing.test.ts, palette.test.ts and qr.test.ts. The other candidates couldn't be proven here. scripts/gate.ts is the only file that changed.

Results. Each file was run three times with FORGE_CHECK_SHARDS=1 bun scripts/test-shards.ts --timeout 60000 <file>. I ran each file alone, not alongside the full gate.

FileRun 1Run 2Run 3
pacing.test.tsgreengreengreen
palette.test.tsgreengreengreen
qr.test.tsgreengreengreen
  • Together: after the move, the three files ran as one slice three more times, all green (37 tests, about 0.5 s).
  • check-list.test.ts: it is green.
  • Wall time: I did not time the full gate, so I have no before and after figures. The three files add well under a second of work, and I didn't record the gate's total. The forge's check on this push will report the real number.

Not admitted. Sixteen other files from the range failed here, and I did not try the rest. They stay in EXCLUDED as "not yet admitted" for #261, and I did not edit any test.

  • Sandbox socket failures: the sandbox here refuses Bun.serve on a unix socket (EPERM … listen …sock, src/http/server.ts:1719). That broke policy, prompts, public-url, parity, probe, probes, secret-store, search, snapshot-destination, standards, reviews-states and signin-throttle. I ran policy, prompts, public-url, parity and qr three times each. The others I ran once.
  • Sandbox network failure: scaffold failed one case, "it typechecks under strict settings", with ConnectionRefused downloading package manifest typescript.
  • What this means: these look like limits of the sandbox, not flaky tests, so I did not mark them red: …. Another sweep on a machine that allows sockets and registry access can prove them.

Try it. FORGE_CHECK_SHARDS=1 bun scripts/test-shards.ts --timeout 60000 test/pacing.test.ts test/palette.test.ts test/qr.test.ts

🤖 Generated with Claude Code

Closes #1141

Approved

Three files (pacing, palette, qr) move from EXCLUDED to GATE; the diff is exactly that, with no stray edits. I ran the three together three times locally: 37 pass, 0 fail each time. This adds no behaviour or UI, so the preview is unaffected. Two things I would do next: (1) admit the next batch of the test/[p-s] files still excluded under issue 261, and (2) record the measured gate load in the pull so the three-green claim is checkable.

Sign in to join this conversation.

↑↓ navigate ↵ open ⌘↵ its app esc close

Keyboard shortcuts

Docs