Repository navigation
ci(id): shfmt-check fails on main — pre-existing, unrelated to PR diffs touching it #3208
Description
Found by queue worker lane B while shepherding PR #3184 (task 4622, 2026-10-31).
Evidence:
- id workflow run 41221436449 on main @ b3b9ad3 (2026-10-31T16:48Z): job "basic: nix flake check (shfmt-check)" concluded FAILURE.
- The same job failed in dispatched id run 41231798279 for PR fix(id/web): encode xterm modifiers for special keys in world keyboard encoder (#732) #3184 @ 64e63c0, whose diff touches only pkgs/id/web/src/world-keyboard.ts and world-keyboard.test.ts (TypeScript) — shfmt formats shell scripts, so the PR cannot be the cause.
- Other jobs in the main run were still in progress at filing time; job logs for the PR run are unavailable until that run completes, so the specific shell file(s) failing shfmt are not yet identified.
Acceptance: identify the shell file(s) failing shfmt on current main (run the shfmt flake check on a clean main checkout), fix the formatting or the check wiring, and verify a main id run with shfmt-check green. Until then, PR shepherds should treat shfmt-check failure as pre-existing and not block unrelated id PRs on it.
Metadata
Metadata
Assignees
Labels
Package/app: id (pkgs/ or apps/ in this repo)Trivial: expected under 30 minutes of workCI/CD pipelines, workflows, and build automationIntake review done: acceptance criteria, links, and normalized labels addedNice to have; do when convenientHas a queued task in the Supabase agent queueFiled by the new-issue-creator jobQueue packet kind: verify, reproduce, and route this item
Activity
dezren39 commentedon Nov 1, 2026
Intake review
Related: Filed from PR #3184 shepherding (task 4622, since merged for #732); the acceptance text already tells shepherds to treat this failure as pre-existing on unrelated id PRs.
Notes: Evidence verified at filing: main run 41221436449 @ b3b9ad3 has the shfmt-check job FAILURE, and PR #3184's diff (TypeScript only) cannot be its cause. The specific failing shell file(s) are not yet identified — the PR-run job logs were unavailable at filing. The body already carries acceptance criteria, so no body change was made.
Risks: Low direct risk, but a permanently-red main check erodes the id workflow's signal for every PR that touches it.
Suggested approach: Run the shfmt flake check on a clean main checkout (CI or laptop — no sandbox builds), name the failing file(s) on this issue, fix formatting or check wiring, and confirm a green main id run.
— fleet monitor intake, 2026-10-31 fleet-monitor run (19:12 CDT)
2 remaining items
dezren39 commentedon Nov 3, 2026
🤖 Agent note · picked up
9aca43b4-5439-4939-9f71-323f5d32521b(OpenChamber scheduled lane)main, name the failing shell file(s), scope the fix or close with evidence.dezren39 commentedon Nov 3, 2026
🤖 Agent note · done
9aca43b4-5439-4939-9f71-323f5d32521b(OpenChamber scheduled lane)mainand name the failing file(s).26c972e36Verdict: fixed on main — closing as no longer reproducible.
origin/main@1c1aa4a79(2026-11-03), running the check's own command verbatim —cd pkgs/id && find . -name '*.sh' -not -path '*/node_modules/*' -exec shfmt -d -i 2 -s {} +— exits 0 with an empty diff. (
pkgs/id/flake.nix:557defines the check this way.)26c972e36"style(id): fix shfmt drift in recipe checks" (on main, 2026-11-03T01:08Z) reformattedpkgs/id/scripts/just-recipes-check.shandjust-recipes-check-test.sh. Feeding those two files as they were at26c972e36^through the sameshfmt -d -i 2 -sinvocation reproduces the drift the check reports (e.g.> "$tmp/fresh.json"→>"$tmp/fresh.json"), which identifies the formerly-failing files precisely.b3b9ad33) has job basic: nix flake check (shfmt-check) = failure. Every completedidrun on main since the fix — 41637749940 (12:40Z), 41639624589, 41639779752, 41640197850 (12:58Z) — has that job = success. Runs after 12:58Z are still queued behind the merge backlog, not failing.Two adjacent facts so nobody re-opens this by mistake: the
idcheck (pkgs/id/flake.nix) is enforced (nonzero exit on drift), while the repo-wide one (lib/default.nix:945) ends in|| trueand is report-only — so drift outsidepkgs/id(currently visible ininfra/worktree-reaper/systemd/install-units.shandpkgs/opencode-config-normalize/normalize_test.sh) fails nothing and is out of scope for this issue. Shepherds can stop treating shfmt-check as pre-existing red on unrelatedidPRs.developing-today commentedon Nov 3, 2026
agent-note:v1
state: triaged (coverage pass)
Coverage pass 2026-11-03: reviewed in the fleet-wide per-issue pass (drift / overlap / already-handled / stale / blockers / ordering). Verdict: still live; queued in repo-health-code-pilot as triage at priority 30 (priority/low). If this is already handled at current main, the triage closes it with evidence; if it needs splitting, the triage files the split.
Found via: coverage pass 2026-11-03