Public: anyone can read this; the crew can change it.
Sentinel: shared:01m23vygyffx5hg2f5vm8r74sr:2m2pliz5wl is past its budget: trevato/forge: "a fast-forward applies the previewed id: a parent that moved after it is not taken" failed on 3 pulls i… #14563
Check · Review · Land · Live — No review yet.
The flaky test is fixed in test/take-expect.test.ts. I ran it three times in a row with all tests passing, and tsc and scripts/gate.ts --fix are clean. Cause: in the test setup, Alice commits "Next" to each repo, and the forge's own follow then fast-forwards Bob's fork in the background.
look green — its preview answered, ready for a look deploy 070a5e7
Check passeddeploy 070a5e7
Website spec: 26 findings (report only)
- /board http://127.0.0.1:40897/board answered 303 — https://specification.website/
- /notes http://127.0.0.1:40897/notes answered 303 — https://specification.website/
- /review http://127.0.0.1:40897/review answered 303 — https://specification.website/
- /releases http://127.0.0.1:40897/releases answered 303 — https://specification.website/
- /system http://127.0.0.1:40897/system answered 303 — https://specification.website/
- /admin/limits http://127.0.0.1:40897/admin/limits answered 404 — https://specification.website/
- /maintenance http://127.0.0.1:40897/maintenance answered 303 — https://specification.website/
- /talk http://127.0.0.1:40897/talk answered 303 — https://specification.website/
- /flow http://127.0.0.1:40897/flow answered 303 — https://specification.website/
- /resilience http://127.0.0.1:40897/resilience answered 303 — https://specification.website/
- /boards?help http://127.0.0.1:40897/boards?help answered 303 — https://specification.website/
- /settings/crew http://127.0.0.1:40897/settings/crew answered 303 — https://specification.website/
- /settings/crew?help http://127.0.0.1:40897/settings/crew?help answered 303 — https://specification.website/
- /settings/security/controls http://127.0.0.1:40897/settings/security/controls answered 303 — https://specification.website/
- /settings/grants http://127.0.0.1:40897/settings/grants answered 303 — https://specification.website/
- /preview-admin/sample expected exactly one <h1>, found 2 — headings
- /preview-admin/sample/drop http://127.0.0.1:40897/preview-admin/sample/drop answered 303 — https://specification.website/
- /preview-admin/sample/data-move http://127.0.0.1:40897/preview-admin/sample/data-move answered 303 — https://specification.website/
- /preview-admin/sample/cut-over http://127.0.0.1:40897/preview-admin/sample/cut-over answered 303 — https://specification.website/
- /preview-admin/sample/next?pick expected exactly one <h1>, found 2 — headings
- /preview-admin/sample/bets expected exactly one <h1>, found 2 — headings
- /preview-admin/sample/waves expected exactly one <h1>, found 2 — headings
- /preview-admin/sample/space the first focusable element is not a skip link (an in-page #target link) — skip-link
- /looks/missing http://127.0.0.1:40897/looks/missing answered 404 — https://specification.website/
- /looks/failed http://127.0.0.1:40897/looks/failed answered 500 — https://specification.website/
- /nope/nothing http://127.0.0.1:40897/nope/nothing answered 404 — https://specification.website/
Look
10 pages beside main · 0 new to this pull
/apps
In the dark
at 070a5e7 ·
/docs
In the dark
at 070a5e7 ·
/explore
In the dark
at 070a5e7 ·
/looks/failed
In the dark
at 070a5e7 ·
/looks/missing
In the dark
at 070a5e7 ·
/nope/nothing
In the dark
at 070a5e7 ·
/notes
In the dark
at 070a5e7 ·
/releases
In the dark
at 070a5e7 ·
/review
In the dark
at 070a5e7 ·
/security
In the dark
at 070a5e7 ·
The flaky test is fixed in test/take-expect.test.ts. I ran it three times in a row with all tests passing, and tsc and scripts/gate.ts --fix are clean.
Cause: in the test setup, Alice commits "Next" to each repo, and the forge's own follow then fast-forwards Bob's fork in the background. On a slow check machine that follow was still running when "a fast-forward applies the previewed id…" read the fork's tip, so it moved Bob's ff fork between before and the assertion. The test then read the fork's tip as changed even though the code under test was fine.
Change: the setup now subscribes to the bus before each "Next" commit and waits for the push event the follow announces for that fork. Nothing moves a fork in the background after that point. It uses no timer and no change to src/.
Not done: I didn't touch the follow code or the other tests in the file, since only this one was named. Only the ff test is named in the failures; the lib tests read their fork's tip twice too, and the same wait covers them.
Process note: my first push went out before the gate had passed, with a Bun.sleep poll that the gate rejects. The second commit replaced it, and the branch's final state passes the gate.
Try it: bun test test/take-expect.test.ts.
This fixes the flaky test "a fast-forward applies the previewed id: a parent that moved after it is not taken" in test/take-expect.test.ts.
Cause: the test setup commits "Next" to each parent repo, and after each one the forge's own follow fast-forwards the fork in the background. On a busy check machine that follow could still be running when the test read the fork's tip twice. It then moved the fork between the two reads, and the assertion at line 112 failed even though fastForwardFork was correct.
Change: before each "Next" commit, the setup now listens on the bus for the push event the follow announces for that fork, and waits for it. After that nothing moves a fork in the background, so the test passes or fails only on the code under test. No timer is involved and src/ is unchanged. The same wait covers the lib fork, whose tests also read its tip twice.
Left out: I didn't change the follow code or the other tests in the file; only this test failed.
Checks: the test file passed three runs in a row, and bun x tsc --noEmit and bun scripts/gate.ts --fix are clean. The first commit waited with a Bun.sleep poll, which the gate's "a test does not wait on a real timer" rule refuses; the second commit replaces it with the bus wait.
Try it: bun test test/take-expect.test.ts.
🤖 Generated with Claude Code
The flaky test is fixed in
test/take-expect.test.ts. I ran it three times in a row with all tests passing, andtscandscripts/gate.ts --fixare clean.Cause: in the test setup, Alice commits "Next" to each repo, and the forge's own follow then fast-forwards Bob's fork in the background. On a slow check machine that follow was still running when "a fast-forward applies the previewed id…" read the fork's tip, so it moved Bob's
fffork betweenbeforeand the assertion. The test then read the fork's tip as changed even though the code under test was fine.Change: the setup now subscribes to the bus before each "Next" commit and waits for the
pushevent the follow announces for that fork. Nothing moves a fork in the background after that point. It uses no timer and no change tosrc/.Not done: I didn't touch the follow code or the other tests in the file, since only this one was named. Only the
fftest is named in the failures; thelibtests read their fork's tip twice too, and the same wait covers them.Process note: my first push went out before the gate had passed, with a
Bun.sleeppoll that the gate rejects. The second commit replaced it, and the branch's final state passes the gate.Try it:
bun test test/take-expect.test.ts.Closes #14513