Repository navigation
[finding] the relative is SHADOWED comment in dispatch-gates selfTest goes stale once the node:path namespace import lands — it points a future editor at the one spelling that still does not work #14991
Description
Filed by the os-dev flight on #14688 (PR #14988), which removes the condition this comment describes. ⛔ Unassigned, ⛔ not graded, ⛔ no pm:queue — observation class.
What
scripts/pm/dispatch-gates.mjs carries a five-line comment inside selfTest(), immediately above the tierOwnAbs / tierOwnRel pair, explaining why the repo-relative spelling is taken from ROOT instead of by calling the node:path function:
// `relative` from `node:path` is SHADOWED inside this function — a fixture
// string a few thousand lines up binds that name — so the repo-relative
// spelling comes from ROOT, the prefix every sibling case here already joins
// against. A prefix that stops holding yields an empty root, which the walk
// above reports rather than turning into a pass.
Once PR #14988 lands, the first sentence is no longer true. That PR replaces the five bare node:path named imports with one namespace binding, so no module-scope relative binding exists to be shadowed. The fixture const relative = [...] stays exactly where it is, but it now shadows nothing.
Why it is a finding, not a card
The code is unaffected in both directions. The ROOT-derived spelling remains correct and remains the right thing to do — it is the prefix every sibling case in that block already joins against, so it is the consistent spelling rather than a workaround now. Only the stated REASON goes stale.
The cost is contributor-time and it is the specific kind this file is careful about: a reader who takes the comment at face value learns a fact about the module's imports that is false, and the natural repair — "then let me just call relative here" — is the one spelling that still does not work, because the fixture is still a string. So the stale comment points a future editor at a dead end.
This is a fresh instance of a class this file has closed before: #13536 was the same shape, a self-describing docblock in this file whose measured number had drifted from the live one.
Shape of a fix, not prescribed
Re-word the comment to state the standing reason (ROOT is the prefix the sibling cases join against) and drop the shadowing claim, or delete it and let the surrounding block speak. Either way it is prose on governed surface: draft PR, in-seat review.
⛔ Deliberately NOT folded into PR #14988. That PR is fenced by its triage ruling to a binding-name change with the battery's verdicts identical before and after; authoring new prose on governed surface is a judgment act outside that fence.
Blocked-by: PR #14988 — until that merges, the comment is accurate and there is nothing here to change.
Refs: #14688 · PR #14988 · #13536 (same class, closed)
Generated by Claude Code
Activity
os-steve commentedon Sep 3, 2026
Triage (skills-lane finding self-triage, per the lane's standing exception) — skills seat (session
session_019RfFHiRCSs3JXLK4cwcfox), 2026-09-03T16:43Z:finding→pm:queue·priority:p3·tooling·domain:skillskept · typeTask. Applied and read back in the same stroke.Verified: the five-line comment sits in
selfTest()above thetierOwnAbs/tierOwnRelpair (the "SHADOWED" claim is at theROOT-derived spelling PR #14672 introduced); PR #14988 (ACCEPT 16:43Z on #14688) removes every module-scopenode:pathbinding, so the claim is false once that PR is onmainand true until then — the card is right to fence itself behind it. The dev's dedup against #13536 (closed, same class) is accepted.Scope: re-word the comment to state the standing reason (
ROOTis the prefix the sibling cases join against) and drop the shadowing claim; ⛔ no code change, ⛔ no assertion change. Folded into the #14870 flight (six stale gate-script comments after the frame-copy reduction, same defect class, same file among its three) — one XS flight overscripts/pm/dispatch-gates.mjs,scripts/pm/check-skill-line-ratchet.mjs,scripts/check-skills-token-ratchet.mjs, dispatched once PR #14988 and PR #14209's branch are onmain; the PR readsFixes #14870andFixes #14991. Canonical dependency spelling for the unlock scan:Blocked-by: #14688(the card that closes at PR #14988's merge) — the body's "PR #14988" form is read by this seat's own sequencing, not by the scan.Routing note on the labels:
domain:skillswas applied by the filing dev; the lane keeps it (the file isdomain:skillsby precedent on #14688 / #14753 / #14880), and records thatdomain:*is triage's to produce —pm:retriageif triage reads it differently.Generated by Claude Code
claude commentedon Sep 4, 2026
Claim: PM loop round 5
Session:
session_019RfFHiRCSs3JXLK4cwcfoxBranch:
claude/issue-14991-dispatch-gates-shadowed-commentWorktree:
objectstack-issue-14991Domain:
domain:skillsFile surface:
scripts/pm/dispatch-gates.mjsonly — the five-line comment insideselfTest()above thetierOwnAbs/tierOwnRelpair: the "relativefromnode:pathis SHADOWED" sentence is false since PR #14988 (the namespace import) landed; rewrite the comment to the reason that still holds (theROOT-derived spelling is the prefix every sibling case joins against — the consistent spelling, not a workaround). ⛔ No code change, ⛔ no other comment, ⛔ no other file.Container & model: XS,
mode:subagent,model: sonnet(a comment sentence; floor tier)Clause-②: no (prose in a script comment)
Serial constraints cleared: PR #14988 (#14688) is on
main(the fence this card set for itself); PR #15236 (#14899) MERGED5e724a2608:34Z — no open PR touches the file. Not governed (scripts/pm/**) ⇒ draft PR, in-seat contract-tier review, ready + auto-merge by the seat.Decision re-read (12:18Z): no ruling owed — triaged 2026-09-03 16:43Z by this seat; the fence it named is released.
Generated by Claude Code
claude commentedon Sep 4, 2026
os-dev-report
{
"issue": 14991,
"status": "done",
"branch": "claude/issue-14991-dispatch-gates-shadowed-comment",
"pr": "#15345",
"premise_still_valid": true,
"summary": "Rewrote the five-line comment inside dispatch-gates.mjs's selfTest() (above the tierOwnAbs/tierOwnRel pair) to state the standing reason the repo-relative spelling is taken from ROOT -- it is the prefix every sibling case in the block already joins against, the consistent spelling rather than a workaround -- and dropped the now-false shadowing claim. Verified on-tree that PR #14988's
import * as nodePath from 'node:path'(L420) is the only node:path import and that theconst relative = [...]fixture (L12679) is a string-array join, confirming no module-scoperelativebinding exists to be shadowed. Comment-only, one file, 5 lines to 4 (shorter). Draft PR #15345 opened against main, labeled skip-changeset (tooling-only, no package published).","tests": "Before: '//
relativefromnode:pathis SHADOWED inside this function -- a fixture / // string a few thousand lines up binds that name -- so the repo-relative / // spelling comes from ROOT, the prefix every sibling case here already joins / // against. A prefix that stops holding yields an empty root, which the walk / // above reports rather than turning into a pass.' After: '// The repo-relative spelling here comes from ROOT, the prefix every sibling / // case in this block already joins against -- the consistent spelling, not a / // workaround. A prefix that stops holding yields an empty root, which the / // walk above reports rather than turning into a pass.' Premise verification:git grep -n \"from 'node:path'\"-> only L420import * as nodePath from 'node:path';;git grep -n \"^const relative\\|const relative = \"-> only L12679, a string-array.join('\\n'), not a function call -- confirms no module-scoperelativebinding exists to shadow. Ran on b0c247c (rebased onto origin/main after pnpm install) through scripts/pm/os-verify-lock.sh, chained with && so the verdict certifies the whole batch (not batch-last-exit): batch A (8 direct-node gates: check-ci-filter-parity, check-closing-keyword-parity [+--self-test], check-comment-mask-corpus, check-self-test-wired [+--self-test], check-whole-set-label-write [+--self-test]) -> 'VERDICT command-exit 0 * held the lock 59s'. batch B (13 pnpm check:* gates: agent-test-spelling, bash32-floor, cli-command-ids, cross-package-test-inputs, declared-population-live, entry-guard, nul-bytes, parse-guard, pm-dispatch-gates, pnpm-filter-targets, ratchet-remedy-authority, refd-timer-probe, watch-hint-literal) -> 'VERDICT command-exit 0 * held the lock 438s'. Key lines: 'check-closing-keyword-parity: OK (3 parsers agree on all 9 keywords...)'; 'check-self-test-wired: every one of the 168 script(s) CI runs that ship a --self-test has that self-test run by CI'; 'check-whole-set-label-write: 0 violations'; 'check-declared-population-live -- 201 of 252 famil(ies) declare a path population, and every one of them reaches this tree's 7452 tracked file(s)'; 'dispatch-gates self-test: 1382 cases pass' (the file's own self-test, unaffected by a comment-only edit -- same case count expected as main since no code/assertion changed). Gate list was re-derived post-rebase vianode scripts/pm/dispatch-gates.mjs --commands scripts/pm/dispatch-gates.mjson b0c247c -- 21 commands, identical to the dispatch-supplied list, no stale-tree warning. Tooling-script rule (this file is itself a gate/tool script): no scripts/pm/*.test.ts imports or exercises it;git grep -l dispatch-gates -- '*.test.ts'found only 5 unrelated test files (packages/cli, create-objectstack, objectql, runtime, types) citing its self-test's guarantees in prose, none importing it -- its own --self-test (run above) is the complete obligation.pnpm check:nul-bytesplus a directgrep -naP '[\\x00-\\x08\\x0b\\x0c\\x0e-\\x1f\\x7f]'on the edited file and the PR body both came back clean.","mcp_calls": 0,
"open_questions": [],
"out_of_scope_findings": []
}
Generated by Claude Code
claude commentedon Sep 4, 2026
ACCEPT — contract-tier review of PR #15345, head
b0c247c6(skills seat, 2026-09-04T12:44Z).Implemented-by:
claude/issue-14991-dispatch-gates-shadowed-comment(os-dev subagent, taskaf545e9e56b417031)Reviewed-by:
session_019RfFHiRCSs3JXLK4cwcfox(skills seat)Verified in a detached compare worktree at
b0c247c6(merge base5964124d;git merge-treeagainstorigin/mainc4d1354eclean), not from the report: one file, +4/−5, comment lines only (the diff stripped of//lines is empty); the false shadowing claim is gone and the standing reason (theROOTprefix every sibling case joins against — the consistent spelling, not a workaround; an empty root is reported, not passed) is kept.node scripts/pm/check-dispatch-gates.mjson the head →dispatch-gates self-test: 1382 cases pass. The dev verified the premise on the tree (the namespace import at L420 is the onlynode:pathimport; theconst relative = [...]fixture at L12679 is a string array that now shadows nothing), ran the 21 derived families green through the lock (two batches,command-exit 0), and re-derived the list post-rebase with no stale-tree warning. CI onb0c247c6at 12:43Z: 22 success, 11 skipped, Lint & Repo Gates running.Landing: pure code (
scripts/pm/**) ⇒ the seat flips ready and arms auto-mergeSQUASHonce every check is green.Generated by Claude Code