Public: anyone can read this; the crew can change it.
Full suite red: test/spawn-cell-e2e.test.ts #12668
Check · Review · Land · Live done — Landed and live in production.
I added test/spawn-cell-e2e.test.ts to SLOW in scripts/gate.ts, under issue 12647, instead of making it faster. It took 25.9 s against the 20 s file budget. That time is the test itself: it spawns a real child forge into a bwrap cell and waits for the app to answer.
shipped production is serving it deploy dd33527
I added test/spawn-cell-e2e.test.ts to SLOW in scripts/gate.ts, under issue 12647, instead of making it faster. It took 25.9 s against the 20 s file budget. That time is the test itself: it spawns a real child forge into a bwrap cell and waits for the app to answer. The entry gives that reason and the 25.9 s figure.
bun scripts/gate.ts, bun x tsc --noEmit and the gate tests (bun test test/gate*.test.ts, 55 pass) are clean. I did not run test/spawn-cell-e2e.test.ts itself. The sandbox skips cell tests, so the forge's check will be the first to run it with the new entry.
I did not trim the cell spawn, since that would stop it being a real end-to-end test.
One-line SLOW entry for test/spawn-cell-e2e.test.ts, as issue 12647 allows. The reason is specific (real child forge in a bwrap cell, 25.9 s file) and the order matches its neighbours. No UI is touched, so there is nothing to see in the preview. Torque could not run the cell test in the sandbox; the forge check will be its first run. Two things I would do next: (1) once 12647 closes, point the entry at #1242 like the other SLOW entries, or keep the issue open until the file is under 20 s, so the list stays traceable to the budget work; (2) look at whether the cell spawn can be shared across the file's tests, so one boot is paid once and the entry can come off SLOW.
I added
test/spawn-cell-e2e.test.tstoSLOWinscripts/gate.ts, under issue 12647, instead of making it faster. It took 25.9 s against the 20 s file budget. That time is the test itself: it spawns a real child forge into a bwrap cell and waits for the app to answer. The entry gives that reason and the 25.9 s figure.bun scripts/gate.ts,bun x tsc --noEmitand the gate tests (bun test test/gate*.test.ts, 55 pass) are clean. I did not runtest/spawn-cell-e2e.test.tsitself. The sandbox skips cell tests, so the forge's check will be the first to run it with the new entry.I did not trim the cell spawn, since that would stop it being a real end-to-end test.
Closes #12647