A zoning gate that checks a list against what its author declared, not against my intuition

Part 1 of making UPG-100 a reusable pattern (Driver via repo; dx-shepherd UPGPAT). Any project list —
7000 / 4700 / 3000 / 2000 / 700 — will be zoned by an agent who did not write this code. Zoning is human
judgement (geography first, balanced with language), so a gate cannot decide whether a zoning is GOOD. It
can decide whether it still matches what the author SAID, which turns "did I zone this right?" from a
browser question into an exit code.

★ EVERY OBVIOUS RULE I WOULD HAVE WRITTEN FAILED THE REFERENCE IMPLEMENTATION. Measured against the live
28 zones and 2,107 records BEFORE writing the assertions, which is the only reason they are right:
  "every group in exactly one zone"  → 112 groups are in two. 110 are the DEAF BLOC, and that is
                                       design-approved (legend-tree.js:151, slice 08): an affinity bloc's
                                       members also live in a geographic zone. So the rule is "at most one
                                       GEOGRAPHIC zone", with bloc zones declared as a second partition.
  "no group is unzoned"              → 11 are, all JPN/CHN/MNG, plausibly deliberate. A gate cannot tell a
                                       deliberate exclusion from a silent drop; only a declaration can.
  "no two zones overlap"             → 3 non-bloc overlaps exist, because "Scattered / Diaspora"
                                       (country_codes AND,CCK,GUF,USA) overlaps the regions it scatters
                                       across.
A gate that fails on the only correct example of the thing it checks teaches people to disable it.

⚠ AND THE GATE FOUND A THIRD OVERLAP MY OWN HAND ANALYSIS HAD MISSED: deaf-cocos-islanders, in both
oceania-indonesia and uncovered-scatter. I had filtered out every A017 group as "the design-approved deaf
double-count" — but its TWO GEOGRAPHIC claims (CCK is in both zones' country lists) are a SEPARATE fact
from its bloc membership. Conflating them is exactly what A2's bloc/geographic split exists to prevent,
and the gate was right where I was wrong. Best argument for it existing.

THE FIVE ASSERTIONS, and two would have caught this cycle's real defects before a browser did:
  A1 no zone resolves to zero groups     ← `deaf` + the 5 wagf zones were in exactly this state, found
                                           only by generating all 28 and reading the log
  A2 at most one GEOGRAPHIC zone/group   ← blocs declared; found the overlap above
  A3 orphan set EQUALS the declared set  ← and a RESOLVED orphan fails too, deliberately: leaving a
                                           now-zoned group listed means the next real drop passes
                                           silently. An unmaintained declaration stops being evidence.
  A4 every selector type is supported    ← the evaluator THROWS on an unknown type, at render time, on one
                                           zone's page; better as a build-time failure on the whole set
  A5 no zone over the declared budget    ← a wildly oversized zone is usually a selector matching too much

DECLARATIONS pre-filled with today's actuals and marked PENDING DRIVER, per UPGPAT: the 11 orphans and
the 3 overlaps are recorded so the gate is green on arrival and any CHANGE fails. Both are in
DRIVER-DECISIONS-DIGEST and neither membership was touched.

VERIFIED: --self-test PASSes 12 cases, each assertion firing on a violation AND staying quiet on the
declared exception (including the negative control that A2's bloc exemption is doing the work). Poison
test on REAL data: removing ainu-japan from the declaration makes A3 fail with that slug, so a green run
is meaningful rather than vacuous.

⚠ AND A package.json MISTAKE WORTH RECORDING. My first insertion put all three keys into the //scripts
DOC block, because re.finditer over the whole file matched check:zone-census there (//scripts comes later
in the file) — producing duplicate keys that json.load silently resolved last-wins, so the script COUNT
WENT DOWN and `npm run check:zoning-declared` reported "Missing script". Fixed by scoping each insertion
to its own block span. The diff is now exactly 3 added lines and reorders nothing: an earlier attempt that
re-serialised the sorted object moved gospel's check:embed-contract and check:engine-contract, which is
not mine to reshuffle in a shared file.
This commit is contained in:
Doxa Maps 2026-09-27 03:36:19 +00:00
parent 34b375ee16
commit a681fd3c65
3 changed files with 317 additions and 0 deletions

View file

@ -0,0 +1,250 @@
#!/usr/bin/env node
/**
* check-zoning-declared — a project list's ZONING is consistent with what its author DECLARED.
*
* WHY THIS EXISTS, and why it is a declaration-checker rather than a rule-checker. UPG-100 is becoming a
* PATTERN: any project list (7000 / 4700 / 3000 / 2000 / 700) gets its own ~100-group clusters, zoned by
* an agent who is not the author of this code. Zoning is human judgement — geography first, balanced with
* language — so a gate cannot decide whether a zoning is GOOD. What it can decide is whether the zoning
* matches what the author said it would be, which turns "did I zone this right?" from a browser question
* into an exit code.
*
* ★ EVERY OBVIOUS RULE I WOULD HAVE WRITTEN FROM INTUITION FAILED THE REFERENCE IMPLEMENTATION, measured
* against the live 28 zones and 2,107 records on 2026-09-27. That is the reason this file is shaped the
* way it is:
*
* "every group in exactly one zone" → 112 groups are in two. 110 are the DEAF BLOC, and that is
* design-approved (legend-tree.js:151, slice 08): an affinity
* bloc's members also live in a geographic zone. Per-zone
* equality holds; only a naive GLOBAL sum over-counts. So the
* rule is "at most one GEOGRAPHIC zone", with blocs declared.
* "no group is unzoned" → 11 are, all Japan/China/Mongolia, and plausibly deliberate.
* A gate cannot tell a deliberate exclusion from a silent drop.
* Hence expected_orphans: the author states them, and the gate
* fails on a CHANGE rather than on their existence.
* "no two zones overlap" → 2 non-bloc overlaps exist (cocos-malays, andorrans) because
* "Scattered / Diaspora" overlaps the regions it scatters
* across. Declared, pending the Driver.
*
* A gate that fails on the only correct example of the thing it checks teaches people to disable it.
*
* ⚠ AND TWO OF THESE ASSERTIONS WOULD HAVE CAUGHT THIS CYCLE'S REAL DEFECTS BEFORE A BROWSER DID.
* A1 (every zone resolves to ≥1 group) is exactly the state `deaf` and the five wagf zones were in when
* the packet generator skipped them with "no groups resolved" — six of 28 zones silently producing
* nothing, found only by generating all 28 and reading the log. A4 (every selector type is supported)
* is the other: selector-evaluator THROWS on an unknown type, which surfaces at render time on one
* zone's page rather than at build time on the whole set.
*
* Usage: node internal/check-zoning-declared.mjs [--list upg-100] [--records <file>]
* node internal/check-zoning-declared.mjs --self-test
*
* Exit 0 pass · 1 fail · 75 NOT MEASURED (no records source — "found nothing" is not "found nothing wrong")
*/
import { readFileSync, existsSync } from 'node:fs'
import { resolve, dirname, join } from 'node:path'
import { fileURLToPath, pathToFileURL } from 'node:url'
const HERE = dirname(fileURLToPath(import.meta.url))
const REPO = process.env.DOXA_REPO || resolve(HERE, '..')
const TAG = 'ZONING-DECLARED'
const argv = process.argv.slice(2)
const SELF_TEST = argv.includes('--self-test')
const flag = (name, dflt) => {
const i = argv.indexOf(`--${name}`)
return i >= 0 && argv[i + 1] && !argv[i + 1].startsWith('--') ? argv[i + 1] : dflt
}
const LIST = flag('list', 'upg-100')
// ── the five assertions, as pure functions so --self-test can exercise them without any data ──────────
/** A1 — a zone that resolves to nothing is a broken selector, not an empty region. */
export function findEmptyZones(zoneSlugs) {
return [...zoneSlugs].filter(([, slugs]) => !slugs || slugs.length === 0).map(([id]) => id)
}
/** A2 — at most one GEOGRAPHIC zone per group. Bloc zones are a declared second partition. */
export function findGeographicOverlaps(zoneSlugs, blocIds) {
const bloc = new Set(blocIds || [])
const owners = new Map()
for (const [id, slugs] of zoneSlugs) {
if (bloc.has(id)) continue
for (const s of slugs) {
if (!owners.has(s)) owners.set(s, [])
owners.get(s).push(id)
}
}
return [...owners].filter(([, zs]) => zs.length > 1).map(([slug, zones]) => ({ slug, zones }))
}
/** A3 — the orphan set must EQUAL the declared one. Both directions matter (see below). */
export function diffOrphans(actual, declared) {
const a = new Set(actual); const d = new Set(declared || [])
return {
// A new orphan is a group that silently stopped being zoned — the failure this gate exists for.
unexpected: [...a].filter((s) => !d.has(s)).sort(),
// ★ AND A *RESOLVED* ORPHAN FAILS TOO, which is deliberate and non-obvious. If a declared orphan is
// now zoned, that is good news and a STALE DECLARATION: leaving it listed means the next real drop of
// that same group passes silently. A declaration that is not maintained stops being evidence.
stale: [...d].filter((s) => !a.has(s)).sort(),
}
}
/** A4 — every selector type must be one the evaluator supports (it throws at render time otherwise). */
export function findUnsupportedSelectors(selectorsById, supported) {
const ok = new Set(supported)
const bad = []
const walk = (id, sel) => {
if (!sel || !sel.type) { bad.push({ id, type: '(none)' }); return }
if (!ok.has(sel.type)) bad.push({ id, type: sel.type })
if (sel.type === 'union') for (const s of sel.selectors || []) walk(id, s)
}
for (const [id, sel] of selectorsById) walk(id, sel)
return bad
}
/** A5 — a zone wildly over the page budget is usually a selector that matched too much. */
export function findOversizedZones(zoneSlugs, maxGroups) {
if (!(maxGroups > 0)) return []
return [...zoneSlugs]
.filter(([, slugs]) => (slugs?.length || 0) > maxGroups)
.map(([id, slugs]) => ({ id, count: slugs.length }))
}
// ── self-test ─────────────────────────────────────────────────────────────────────────────────────────
if (SELF_TEST) {
// ★ PROVE EACH ASSERTION FIRES AND THAT IT IGNORES THE DECLARED CASES. A gate whose assertions have
// never matched anything is indistinguishable from a gate with an inverted condition, and three of
// these five are subtle enough that "it passed on the real data" is not evidence they work.
const fails = []
const ok = (cond, msg) => { console.log(` ${cond ? 'ok ' : 'FAIL'} ${msg}`); if (!cond) fails.push(msg) }
ok(findEmptyZones([['a', ['x']], ['b', []]]).join() === 'b', 'A1 flags an empty zone, not a full one')
ok(findEmptyZones([['a', ['x']]]).length === 0, 'A1 passes when every zone resolves')
const zs = [['geo1', ['p', 'q']], ['geo2', ['q']], ['deaf', ['p', 'q']]]
ok(findGeographicOverlaps(zs, ['deaf']).map((o) => o.slug).join() === 'q',
'A2 flags the geographic overlap and EXEMPTS the declared bloc (this is the deaf double-count)')
ok(findGeographicOverlaps(zs, []).length === 2,
'A2 WITHOUT the bloc declaration flags both — so the exemption is doing the work, not the regex')
const d1 = diffOrphans(['new', 'known'], ['known'])
ok(d1.unexpected.join() === 'new' && d1.stale.length === 0, 'A3 flags a NEW orphan')
const d2 = diffOrphans([], ['known'])
ok(d2.stale.join() === 'known' && d2.unexpected.length === 0,
'A3 also flags a RESOLVED orphan as a stale declaration — an unmaintained declaration is not evidence')
ok(diffOrphans(['known'], ['known']).unexpected.length === 0, 'A3 passes on an exact match')
const sup = ['country_code', 'union']
ok(findUnsupportedSelectors([['z', { type: 'wagf_block' }]], sup)[0].type === 'wagf_block',
'A4 flags an unsupported type')
ok(findUnsupportedSelectors([['z', { type: 'union', selectors: [{ type: 'nope' }] }]], sup)
.some((b) => b.type === 'nope'), 'A4 descends INTO a union — southern-africa-islands is one')
ok(findUnsupportedSelectors([['z', null]], sup)[0].type === '(none)', 'A4 flags a missing selector')
ok(findOversizedZones([['big', new Array(300).fill('x')]], 200)[0].count === 300, 'A5 flags an oversized zone')
ok(findOversizedZones([['ok', new Array(120).fill('x')]], 200).length === 0, 'A5 passes a large-but-sane zone')
if (fails.length) {
console.error(`\n${TAG} --self-test: FAIL — ${fails.length} case(s) wrong.`)
process.exit(1)
}
console.log(`\n${TAG} --self-test: PASS — all five assertions fire on a real violation and stay quiet on `
+ 'the declared exceptions.')
process.exit(0)
}
// ── the run ───────────────────────────────────────────────────────────────────────────────────────────
const DATA = join(REPO, 'plugin-designs', 'my-upg-100-list', 'data')
const declPath = join(DATA, LIST === 'upg-100' ? 'zoning-declarations.json' : `zoning-declarations.${LIST}.json`)
if (!existsSync(declPath)) {
console.log(`${TAG}: NOT MEASURED — no declarations at ${declPath}.`)
console.log(' A list without declarations cannot be checked against them. Copy upg-100\'s and fill it in.')
process.exit(75)
}
const decl = JSON.parse(readFileSync(declPath, 'utf8'))
// Records: the same source everything else uses. NOT fetched here — a gate that needs the network is a
// gate that SKIPs in CI, which is how check:packets-all went unevaluated on every build for weeks.
const recordsPath = flag('records', join(REPO, 'library/scripts/posters/out/_records.json'))
if (!existsSync(recordsPath)) {
console.log(`${TAG}: NOT MEASURED — no records at ${recordsPath}.`)
console.log(' Generate them once (OFFLINE=1 or a packet run caches them), or pass --records <file>.')
console.log(' Exit 75 rather than 0: "found no records" is not "found nothing wrong".')
process.exit(75)
}
const records = JSON.parse(readFileSync(recordsPath, 'utf8'))
const clusters = JSON.parse(readFileSync(join(DATA, 'clusters.json'), 'utf8'))
const { buildLegendNodes, leafZones } = await import(pathToFileURL(join(DATA, 'legend-tree.js')).href)
const { evaluate } = await import(pathToFileURL(join(DATA, 'selector-evaluator.js')).href)
// Selectors by id, exactly as generate.mjs:2801-2808 and packet.mjs's allZones do it — leaves do not
// carry one, and `deaf`'s is implicit.
const selectorsById = new Map()
for (const k of ['lists', 'asia_lists', 'uncovered_lists']) {
for (const z of clusters[k] || []) selectorsById.set(z.id, z.selector)
}
selectorsById.set('deaf', { type: 'affinity_bloc', rop1: 'A017' })
const zones = leafZones(buildLegendNodes(clusters, { records: null }))
const zoneSlugs = []
for (const z of zones) {
const sel = selectorsById.get(z.id)
let slugs = []
if (sel?.type === 'pg_slugs') slugs = sel.values || []
else if (sel) { try { slugs = [...evaluate(sel, records)] } catch (_) { slugs = [] } }
zoneSlugs.push([z.id, slugs])
}
const SUPPORTED = ['country_code', 'country_codes', 'wagf_block', 'wagf_region', 'union', 'deaf_predicate',
'affinity_bloc', 'pg_slugs']
const covered = new Set(zoneSlugs.flatMap(([, s]) => s))
const actualOrphans = records.map((r) => r.slug).filter((s) => s && !covered.has(s))
const declaredOverlaps = new Set((decl.known_geographic_overlaps?.slugs || []).map((o) => o.slug || o))
const empty = findEmptyZones(zoneSlugs)
const overlaps = findGeographicOverlaps(zoneSlugs, decl.bloc_zones?.ids).filter((o) => !declaredOverlaps.has(o.slug))
const orphanDiff = diffOrphans(actualOrphans, decl.expected_orphans?.slugs)
const unsupported = findUnsupportedSelectors(selectorsById, SUPPORTED)
const oversized = findOversizedZones(zoneSlugs, decl.page_budget?.max_groups)
console.log(TAG)
console.log(` list ${LIST} · ${zones.length} zone(s) · ${records.length} record(s) · ${covered.size} covered`)
console.log(` declared: ${(decl.bloc_zones?.ids || []).length} bloc zone(s), `
+ `${(decl.expected_orphans?.slugs || []).length} expected orphan(s), `
+ `${declaredOverlaps.size} known overlap(s)`)
const problems = []
if (empty.length) problems.push(`A1 ${empty.length} zone(s) resolve to NO groups: ${empty.join(', ')}`)
if (overlaps.length) {
problems.push(`A2 ${overlaps.length} group(s) in two GEOGRAPHIC zones and not declared: `
+ overlaps.slice(0, 8).map((o) => `${o.slug} [${o.zones.join(' + ')}]`).join('; '))
}
if (orphanDiff.unexpected.length) {
problems.push(`A3 ${orphanDiff.unexpected.length} NEW orphan(s) — in the list, in no zone, not declared: `
+ orphanDiff.unexpected.slice(0, 12).join(', '))
}
if (orphanDiff.stale.length) {
problems.push(`A3 ${orphanDiff.stale.length} declared orphan(s) are now ZONED, so the declaration is `
+ `stale — remove them or the next real drop passes silently: ${orphanDiff.stale.slice(0, 12).join(', ')}`)
}
if (unsupported.length) {
problems.push(`A4 ${unsupported.length} zone(s) use a selector type the evaluator does not support: `
+ unsupported.map((b) => `${b.id}(${b.type})`).join(', '))
}
if (oversized.length) {
problems.push(`A5 ${oversized.length} zone(s) exceed the declared page budget `
+ `(${decl.page_budget?.max_groups}): ` + oversized.map((o) => `${o.id}=${o.count}`).join(', '))
}
if (problems.length) {
console.error(`\n${TAG}: FAIL — ${problems.length} problem(s).`)
for (const p of problems) console.error(` ✗ ${p}`)
console.error('\n Each of these is EITHER a zoning defect OR a declaration that needs updating, and the')
console.error(' gate cannot tell which — that is the author\'s call. What it can guarantee is that the')
console.error(' zoning and the declaration have not drifted apart silently.')
process.exit(1)
}
console.log(`${TAG}: PASS — the zoning matches every declaration: no empty zone, no undeclared geographic `
+ 'overlap, the orphan set is exactly as declared, every selector type is supported, no zone over budget.')

View file

@ -227,6 +227,8 @@
"check:vision-card-numbers": "node internal/check-vision-card-numbers.mjs",
"check:wp-embed": "php deployment/wordpress/tests/wp-embed-contract-test.php",
"check:zone-census": "node internal/check-zone-census.mjs",
"check:zoning-declared": "node internal/check-zoning-declared.mjs",
"check:zoning-declared:self-test": "node internal/check-zoning-declared.mjs --self-test",
"clean": "node internal/clean.mjs",
"deploy:dev": "bun run build && node internal/sync-dropin.mjs --from-config --prune",
"dev": "bun x vite",
@ -417,6 +419,7 @@
"check:url-resolver": "cap-5 resolve() must not prefix a rooted URL (the dev-surface 404 class)",
"check:vision-card-numbers": "the DOXA Vision Card must show the numbers the DATA gives: independently recomputes all four funnel steps (and DOXA-in-frontier) from people_groups.csv + the saved DOXA list, compares with vision-funnel.json, and with --url also with the numbers painted in the card. A typed-in number, a stale funnel file or a rendering bug fails it.",
"check:zone-census": "every zone a USER CAN SELECT must have a poster (asks the LEGEND, not clusters.json — two collapsed deaf lists are unreachable and not a gap). CONDITIONAL: SKIPs without out/posters/, which is gitignored — run `bun run posters` first. It cannot be fixtured like check:packets-all, because it asserts COVERAGE of real generated output and a fixture would only assert the fixture is complete.",
"check:zoning-declared": "A project list's ZONING matches what its author DECLARED (data/zoning-declarations.json). Zoning is human judgement — geography first, balanced with language — so a gate cannot decide whether a zoning is GOOD; it can decide whether it still matches what was stated, which is what makes UPG-100 reusable as a PATTERN for any project list (7000/4700/3000/2000/700) zoned by an agent who did not write this code. ★ EVERY OBVIOUS RULE FAILED THE REFERENCE IMPLEMENTATION when measured on the live 28 zones / 2107 records: 'every group in exactly one zone' is wrong (112 double-claims, 110 of them the design-approved deaf bloc double-count, legend-tree.js:151) and 'no group unzoned' is wrong (11 JPN/CHN/MNG groups, plausibly deliberate). So it asserts A1 no zone resolves to zero groups, A2 at most one GEOGRAPHIC zone per group with bloc zones declared as a second partition, A3 the orphan set EQUALS the declared set (a RESOLVED orphan fails too — a stale declaration stops being evidence), A4 every selector type is one selector-evaluator supports, A5 no zone over the declared page budget. A1 and A4 would both have caught this cycle's real defects (deaf + 5 wagf zones resolving to nothing) before a browser did, and A2 caught a third overlap my own hand analysis had missed. Exit 75 without records or declarations. Run :self-test to prove the five still fire.",
"deploy:dev": "destination + allowlist come from vite.doxa-maps-build-config.json (dropin.dest / dropin.include); DOXA_DROPIN_DEST=<dir> overrides the destination",
"gallery": "writes public/__gallery/index.html — the ARCHIVED bundle-index gallery, from public/doxa-maps/manifest.json. It is NO LONGER the root page: as of 2026-09-22 `/` mounts <plugin-studio> in BOTH trees (the built root page is written by `bun run tree`), and the gallery is the opt-in index at /__gallery. Archived, not deleted — the dev server imports renderGallery()/readBuildConfig() from this module.",
"internet:adm1": "regenerate data/tables/internet-by-adm1.json (internet use by first-level area; India from NFHS-5 2019-21, CC BY 4.0) from data/internet/ through a strict name crosswalk to the shipped gb-IND-ADM1 boundaries. Fails on an unmatched name, never guesses. Generic rows (iso3, adm1, measure, sex, area, value, source, year) so other plugins and go-int can reuse it.",

View file

@ -0,0 +1,64 @@
{
"_meta": {
"what": "Zoning declarations for the UPG-100 list: the facts about this list's zoning that a gate cannot infer and a reader cannot guess from the data.",
"why": "check-zoning-declared.mjs asserts the ACTUAL zoning against these DECLARATIONS. Without them a gate can only test self-consistency, which is satisfied by a zoning that silently dropped groups.",
"how_to_use": "A new project list ships its own <id>.json beside this one. Copy the shape, not the values.",
"list": "upg-100",
"generated_from": "clusters.json + the live records, 2026-09-27"
},
"bloc_zones": {
"_why": "Zones that partition by AFFINITY or LANGUAGE rather than geography. Their members ALSO belong to a geographic zone, so they are exempt from the one-geographic-zone rule. legend-tree.js:151 records the deaf case as design-approved (slice 08): per-zone equality holds; only a naive GLOBAL sum over-counts.",
"ids": [
"deaf"
]
},
"expected_orphans": {
"_why": "Groups in the list that belong to NO zone. An empty list here means 'every group must be zoned'. A gate cannot tell a deliberate exclusion from a silent drop; only this declaration can.",
"_status": "PENDING DRIVER — these 11 are TODAY'S ACTUAL orphans, recorded so the gate is green on arrival and any NEW orphan fails. They are all Japan / China / Mongolia and are plausibly deliberate (asia-china-tibet-autonomous-region is labelled 'Mainland China + Mongolia' and carries only 43 groups), but nobody has said so. In DRIVER-DECISIONS-DIGEST; do NOT change the memberships to make this list shorter.",
"slugs": [
"ainu-japan",
"jing",
"kalmyk-oirats",
"oki-no-erabu",
"pana-china",
"toku-no-shima",
"tsaatan",
"tuvinians",
"yaeyama",
"yonaguni",
"yoron"
]
},
"known_geographic_overlaps": {
"_why": "Groups that DO sit in two geographic zones, declared so the gate fails on a NEW overlap rather than on these. Not an endorsement — see _status.",
"_status": "PENDING DRIVER — uncovered-scatter is country_codes AND,CCK,GUF,USA ('Scattered / Diaspora'), so it overlaps the regional zones it scatters across. Either a scatter zone is a THIRD partition (like a bloc) or these two memberships are wrong. In DRIVER-DECISIONS-DIGEST; do NOT change either membership. ⚠ deaf-cocos-islanders was MISSED by my own hand analysis and found by this gate: I had filtered out every A017 group as 'the design-approved deaf double-count', but its TWO GEOGRAPHIC claims (CCK is in both zones' country lists) are a SEPARATE fact from its bloc membership. Conflating them is exactly what the bloc/geographic split in A2 exists to prevent, and the gate got it right where I did not — which is the best argument for the gate existing.",
"slugs": [
{
"slug": "deaf-cocos-islanders",
"zones": [
"oceania-indonesia",
"uncovered-scatter"
]
},
{
"slug": "cocos-malays",
"zones": [
"oceania-indonesia",
"uncovered-scatter"
]
},
{
"slug": "andorrans",
"zones": [
"uncovered-scatter",
"europe"
]
}
]
},
"page_budget": {
"_why": "Groups per printed page, from packet-render's PER_PAGE. A zone larger than max_groups is not an error — it just prints more sheets — but a zone WILDLY larger is usually a selector that matched too much, which is how the wagf and deaf selector bugs looked before they were found.",
"per_page": 20,
"max_groups": 200
}
}