open/ktg-plugin-marketplace
1
0
Fork 0
Code Issues Releases Activity

feat(okf): enforce §3 okf_version shape, bump convention 0.1 -> 0.2

Browse source 
The 2026-07-23 reservation said enforcement waits until "the emitters have
migrated". Measured rather than assumed, and the emitter set turned out to be
TWO, not three:

  okr @0059da7       okf_version: 0.1 + okf_layout: kb-layout-2026-06. Migrated.
  linkedin-studio    scaffold.ts:38 -> okf_version: 0.1. Never carried a layout
                     string; has no okf_layout and needs none.
  ms-ai-architect    NOT an emitter. Zero index.md in the repo; okf_version
                     appears in one planning doc. The status table already said
                     "designed, not built" -- only the word "emitters" implied it.

Swept the whole marketplace plus the sibling consumer repos: every live
okf_version value is 0.1. The flip is a no-op today and locks the invariant.

The gate checks SHAPE (/^\d+(\.\d+)*$/), never membership: the upstream value
set is Google's (spec §12), so a bundle targeting a newer version passes. A
membership check would be the convention claiming a set it says it does not own
-- the over-reach class this round has corrected three times. Presence stays
unenforced (reported, not failed) as a separate §3 MUST.

Convention bumped 0.1 -> 0.2 by this log's own criterion: the set of conforming
bundles changed (the spec explicitly said a layout string "conforms today", and
rollout rule 6 defines conformance as gate output). The §12 per-plugin re-check
is pre-measured as finding zero violations.

Also closes a parity blind spot found while landing this: check-okf-parity
compared the okf_version VALUE, which is identical exactly when two impls
disagree about whether it is acceptable -- a false "agree" on the axis the gate
exists to watch. The signature now carries a boolean shape verdict, with
red-marker-layout as the committed red proof. Its "diverge" expectation encodes
okr's lag (they still pure-echo) and flips to "agree" when they mirror it.

And a correction the round earned: portfolio-optimiser enumerated their own
bundle surface and it is four, not three. The dormant fourth
(reference_domain.py:49 -> package data at :70) has no flag form at all, so our
published claim that the re-measurement surfaced "one" unreported entry was
itself one-of-two. Sharpens the standing rule: a search shaped like one entry
type cannot see another; install-vs-fixture is one instance, not the class.

Suite 73 -> 78. coord: PO replied, okr + llm-ingestion-okf notified (the latter
under the standing promise to flag any okf-check.mjs change).
This commit is contained in:
Kjell Tore Guttormsen 2026-07-25 20:32:30 +02:00
commit 6a72b26985
9 changed files with 227 additions and 11 deletions
Download patch file Download diff file Expand all files Collapse all files

17
scripts/check-okf-parity.mjs
View file

@ -16,7 +16,8 @@
//
// CONTRACT: comparison unit is the PER-FILE outcome, not the bundle verdict — two
// checkers can both say "OK" while disagreeing on what IS a concept file. Each impl's
// existing checkBundle(root) return is normalized to { conceptCount, untyped[], okfVersion };
// existing checkBundle(root) return is normalized to
// { conceptCount, untyped[], okfVersion, okfVersionAccepted };
// the gate diverges iff those normalized outcomes are not identical across the
// available impls. Defined over DEFAULT read-mode (strictIngest OFF, not files-scoped).
//
@ -59,12 +60,19 @@ export async function normalizedOutcome(impl, root) {
conceptCount: r.scanned,
untyped: [...(r.missingType ?? [])].sort(),
okfVersion: r.okfVersion ?? null,
// spec §3 marker axis. Normalized to a BOOLEAN verdict, not the message: two impls that
// both enforce must agree even if they word the rejection differently. An impl that does
// not check at all reports no error, i.e. "accepts" — which is exactly the signal we want
// when one enforces and another still pure-echoes.
okfVersionAccepted: !(r.okfVersionError ?? null),
};
}
// Signature over the fields that define per-file parity. Distinct signatures => divergence.
// The okf_version VALUE alone could not carry the marker axis: when impls disagree about
// whether a value is acceptable, the value itself is identical, so the axis read as "agree".
function signature(o) {
return `${o.conceptCount}|${o.untyped.join(',')}|${o.okfVersion}`;
return `${o.conceptCount}|${o.untyped.join(',')}|${o.okfVersion}|${o.okfVersionAccepted}`;
}
// Run all impls on one bundle root; decide agree vs diverge among the comparable ones.
@ -145,7 +153,10 @@ if (isMain) {
r.status === 'PASS' && r.expected === 'diverge' ? 'PASS (red-proof)' : r.status;
out.push(`${r.fixture.padEnd(22)} [${(r.axis || '').padEnd(16)}] expect=${(r.expected || '').padEnd(8)} ${tag}`);
for (const o of r.outcomes ?? []) {
out.push(` ${o.name.padEnd(10)} count=${o.conceptCount} untyped=[${o.untyped.join(', ')}] okf=${o.okfVersion}`);
// The shape verdict is printed because the value alone cannot show this axis: on a
// marker divergence both impls print the SAME okf= value and only the verdict differs.
const verdict = o.okfVersionAccepted ? 'accepted' : 'REJECTED';
out.push(` ${o.name.padEnd(10)} count=${o.conceptCount} untyped=[${o.untyped.join(', ')}] okf=${o.okfVersion} (${verdict})`);
}
if (r.skipped?.length) out.push(` (skipped, unavailable: ${r.skipped.join(', ')})`);
}