For: a fast, honest suite (#1242): scripts/test-times.json records test/crew-check-failed.test.ts at 12.7 s, over the 10 s file budget.
Done:
[ ] bun test test/crew-check-failed.test.ts finishes under 10 s on the vxrail and every test in it under 1 s; scripts/test-times.json records the new time.
[ ] Every behaviour the file proves today is still proved: no test is deleted without naming, in the pull, the test that now covers it.
[ ] Crew runs are driven by a fake harness and a fake clock, the forge fixture is built once per file, and nothing sleeps.
Files: test/crew-check-failed.test.ts, scripts/test-times.json; production code only if a wait there is the cause, said in the pull.
Risk: a faster test that proves less. Run it ten times in a row green before pushing.
Not this issue: any other test file.
For: a fast, honest suite (#1242):
scripts/test-times.jsonrecords test/crew-check-failed.test.ts at 12.7 s, over the 10 s file budget.Done:
bun test test/crew-check-failed.test.tsfinishes under 10 s on the vxrail and every test in it under 1 s;scripts/test-times.jsonrecords the new time.Files:
test/crew-check-failed.test.ts,scripts/test-times.json; production code only if a wait there is the cause, said in the pull. Risk: a faster test that proves less. Run it ten times in a row green before pushing. Not this issue: any other test file.