2.1 KiB
Row 97 (#1591) round 3: check run on the landing base
Owner: Researcher. Run at 2026-10-13T14:36Z, in the slot Sage gave for this check. build.patch and
candidate-manifest.sha256 are untouched by this run.
Base
Fresh detached worktree at bb45f75e (origin/refactor; HEAD and origin/refactor both read
bb45f75ed2). That is two commits past the cd976b09 the patch was
built on, with row 98 landed on top. git apply --3way r3/build.patch applied all three files
cleanly. node_modules was linked from the canonical checkout.
Manifest
sha256sum -c r3/candidate-manifest.sha256 in that worktree: OK for all three paths
(packages/webui/README.md, tests/board-held.test.mjs, tests/browser-edge.test.mjs).
Webui suite
node --test tests/ in packages/webui: 107 tests, 107 pass, 0 fail, 0 cancelled, 0 skipped.
grep -c '✖' on the output: 0. Round 2 had 101; the rest is row 94's tests and the base's other
landings, not this candidate.
Mutants
r2/mutants.py restricted to 26 and 27, the two that edit bus.js. The runner starts each
node --test in its own session and kills the process group afterwards.
| mutant | outcome |
|---|---|
| 27 bus schedule polls signed out | killed (exit 1, 4.2 s) |
| 26 bus render without its signed-out guard | killed (exit 1, 4.1 s) |
Both kills are an AssertionError in one test, "a redirect answer on a Board read, a bus read, a
conversation read, a board POST, a module read and a send signs the page out: one line, nothing
stale, no further reads", not a crash or a timeout. I confirmed that with a second run that captured
the failing lines. bus.js was restored byte for byte after both runs (sha256 checked). No
survivors, no failing tests.
The row 94 /api/bus/messages reads in render() did not change either outcome, as the build note
expected: render() returns on out() before them.
Cleanup
After the runs no process had a cwd under my scratch (/proc check, 0 found). The check worktree is
removed after this note. Not run here: control-board and the scripts/test-*.sh suites; Sage's
landing gate covers them.