The failure this catches is a short count with no error. generate.mjs used to
abort on the first bad zone — 16 of 29 written, the rest silently unbuilt — and
separately, tooling that enumerated clusters.json rather than the LEGEND warned
two zones were "MISSING" when nothing was. Both look like success from a
distance; only a census shows them.
★ ASK THE LEGEND, NOT THE DATA FILE. clusters.json carries 30 list definitions;
the legend exposes 28 leaves. deaf-old-world and deaf-new-world-asia-pacific are
collapsed into the unified `deaf` row (affinity bloc A017) — VERIFIED against
live records rather than taken from a comment: 63 + 60 = 123 = A017 exactly,
with zero groups on either side of the difference. They remain in clusters.json
deliberately, because three isolation-proof scripts assert
deaf_lists.length === 2, but no user can select them.
So they are reported as CONTEXT ("superseded, not a gap"), never as a failure.
Counting the file instead of the legend manufactures two phantom failures and
then hides real ones behind them — which is how a genuinely missing poster would
get lost in noise everyone had learned to ignore.
Passing: 28 selectable zones, 28 posters. Confirmed to fail before being
accepted: removing one poster reports that zone by name.
Driver: "make sure poster generation works and the aspect ratio of the print
dialog can be custom based on your print point, so print dialog works for
poster, for vertical packet, and for horizontal cards... all working in the
jamstack maps."
One gate over all four paths, because they share one failure: silently falling
back to letter. Chromium does that for ANY @page size it does not recognise
('A6 portrait' -> 612x792pt, no error, no warning), so every assertion here is
on the MEASURED MediaBox from page.pdf(), never on the stylesheet. A gate that
grepped the CSS would pass a deck that prints as paper.
Measured, all from the shared renderers the UPG-100 page itself imports:
packet (vertical) 612 x 792 pt letter portrait
cards (poker) 180 x 252 pt 2.5 x 3.5in
cards (landscape) 252 x 180 pt 3.5 x 2.5in
poster 1500 x 2000 px rasterised, painted
Orientation is asserted separately from size, so a transposed stock cannot pass
by accident, and the poster is checked for ink rather than dimensions because it
rasterises rather than paginating.
Gates the client-side poster fallback, and it measures PIXELS rather than
bytes — because the failure mode here is silent and total.
⚠ An SVG loaded through <img> is an ISOLATED DOCUMENT. It cannot see the host
page's CSS, so every var() falls back to its initial value, and `fill`'s initial
value is BLACK. A poster that loses its stylesheet downloads as a black
rectangle with no error, no exception, and a perfectly normal file size. The
prayer packet shipped exactly that bug once already (660 of 661 map elements
painting black, nothing in the console).
So the gate asserts mean luminance and colour variety, not "did it produce
bytes", and it checks the stylesheet actually travelled into the inlined SVG.
Confirmed to FAIL on the real regression before being accepted: with the
<style> block stripped, luminance drops 229 -> 13 and distinct colours 54 -> 2,
and all three assertions fire.
Passing: laos rasterises to 3000x4000, 725 KB, stylesheet inline, luminance 229,
54 distinct colours.
Gates the prayer-card deck at the rendered layer, and asserts two things a
plausible-looking check would miss.
1. THE MEASURED MediaBox, not the CSS. Chromium SILENTLY falls back to letter
for a named @page size it does not know — `size: A6 portrait` yields
612x792pt with no error, while `size: 2.5in 3.5in` yields exactly 180x252pt.
A gate that grepped the stylesheet would pass a deck that prints as paper,
which is the entire bug being fixed. It also names the failure: if the page
comes back 612pt wide it says "this is LETTER: the @page size silently fell
back".
2. THE VALUES ARE HUMAN-READABLE, not merely present. The first deck rendered
"LAO", "thm", "EAN" because normalizeRecords() flattens { value, label } to
the VALUE and parks the label under <field>_label. Counting rows passed it —
the rows existed, they were just codes on a card meant for praying. The gate
now flags any field showing a bare 2-4 character token.
Also asserts one card per page, portraits actually loading, a real QR symbol per
card (not an empty path), and no card overflowing its own box.
Expected page sizes are DERIVED from the stock's inches rather than hand-copied:
an earlier constant said A6 was 396.5x559.7pt and FAILED a correct render — the
gate's own arithmetic was wrong, not the renderer.
Verified across all three stocks: poker 180x252pt, postcard 288x432pt, A6
297.12x420pt, 95 cards / 95 pages each.
check:packet-render verifies a packet FILE from the build-time generator. This
verifies the BROWSER path — the legend's Export > Prayer packet action, which
imports library/packet/packet-render.js, fetches live records and prints
in-place.
Both callers now share that one renderer, and this gate is what proves the
shared module actually runs in a tab. A module can be perfectly correct under
Node and still fail in a browser: a bare process.env reference, a Node-only
import, a JSON import assertion. Node passing says nothing about that.
It drives the real thing: the renderer over HTTP, the app's own normalizer
(the API returns { posts, total } — hand-rolling that extraction is how a probe
ends up testing something the product never does), live records, then
page.pdf() and a PDF page count.
Measured on the browser path: 8 sheets -> 8 pages at 612x792pt, 95 people
cards, 12 wheel segments, Mark 5:1-20 in full, exactly 2 metacamp.org
attributions, 0 page errors.
I18N is asserted here because it is a product requirement, not a nicety: pass a
language and the gate reports which headings came back translated. Verified with
a partial Spanish block — "12 Puntos de Oración" and "Rueda de Oración" rendered
while Mark 5 and both attributions correctly fell back to English. Fallback is
PER KEY, so a half-finished translation renders its finished keys and never
blanks. The stub block was removed after verifying; shipping it would imply a
Spanish translation that does not exist.
Parameterised rather than pinned: PACKET_ORIGIN / PACKET_API / PACKET_ISO3, and
the language as argv. Needs a static server over the repo root.
Every bug this gate exists for produced a file of the right size, with no
error, no broken image, and a generator reporting success. They were only
visible once a browser laid the document out, so this drives a real browser
and reads computed style rather than inspecting the HTML source.
★ The law it encodes: never print a number without asserting on it. The
throwaway probe this replaces PRINTED "PDF pages: 16" for several runs while
reporting "packet verified", because the failure counter never looked at that
value. A number that is displayed but not asserted on is decoration.
Two regressions it catches, both confirmed to FAIL before this was accepted:
transform instead of zoom -> 16 PDF pages for 8 sheets
map stylesheet removed -> unresolved --paper/--sea/--frameline;
660/661 map elements fill BLACK
The second is the subtler one. The map SVG carries almost no presentation
attributes; it is styled from the poster document's stylesheet. Lift the
<svg> without those rules and SVG fill falls back to its initial value, which
is black — the whole map paints as a solid rectangle with nothing to see in
the DOM, no console error, and no broken image. It is asserted as a
WHOLE-MAP property because the first attempt tested one hand-picked <rect>
(the one just repaired) and went green while 647 others were still black.
Page count is load-bearing rather than cosmetic: the packet is printed
front-and-back, so "one sheet is exactly one page" is a correctness property,
not a tidiness one.
Counts are relative (sheets vs pages, cards vs photos) so the gate is not
pinned to one zone's group count, and the packet path is an argument or
PACKET_HTML rather than a hardcoded path. With no packet given it SKIPS
rather than failing, since the packet is generated on demand.
Main moved 20+ commits while the regions work was on its own branch. Every
conflict was a GENERATED file — manifest.json, the root page, and the
regions-dashboard bundle's .js and leaf pages. Zero source conflicts: all six
of my source files (the two regions .vue, the new map child, fix-sri.mjs,
sri-consistency-check.mjs, package.json) auto-merged.
Generated files are not hand-merged. They were resolved to main's side and then
REGENERATED from the merged source — BUNDLE=regions-dashboard vite build, then
bun run tree, then bun run fix:sri. `bun run tree` rewrote every bundle's pages
and produced byte-identical output to main's committed ones for all twenty
other bundles, so the only files this commit changes beyond the merge are the
six that had to be rebuilt. That agreement is also evidence main's committed
tree matches main's source.
The pinHint:false regression I saw before is gone: main's 566ad46 committed the
two metas that were previously only on disk, which was the root cause — a tree
regeneration now preserves them instead of dropping them.
Re-verified against the merged library, which is the part that mattered: main
changed useMapInstance.js, StackLayout.vue and ViewSelect.vue, all of which the
regions element sits on. All three suites still pass — the map draws 934 pins /
51 outline features for a region and 61 / 5 for a country, the pin mode still
follows the subject, the country page still renders CountryDetail's shape, and
every width still matches DX-life's container formula at six viewports.
23 of 25 gates green. The two that fail are pre-existing and not in this diff:
check:library-index is an mtime artifact (regenerating the index produces
byte-identical content), and check:backlog-status reports 6 of 546 non-enum
statuses in the shared _qa-cycle board, which lives outside every repo.
The 29 gates in build:unlocked cost 84.5s chained with &&, but 21 of them
finish in under 700ms — they were queueing behind check:mjs-parse (26.9s)
and check:parse (18.7s) on an otherwise idle 8-core box. The 22 vite builds
cost 129.0s and were never ordered at all: vite.config.js gives each bundle
its own outDir with emptyOutDir:false and emits nothing outside it, so they
only ever needed to not overwrite each other.
internal/run-parallel.mjs runs a set of independent steps concurrently and
grades the set by exit code. build:unlocked now reads:
gates:pre -> library-index -> clean -> token -> build:bundles -> tree
-> gates:post -> gallery
Measured end-to-end, both chains green, two rounds each, clean+token
excluded as common to both: 165.7s -> 65.3s mean (170.7/160.8 vs 60.2/70.5).
No gate was removed, weakened, sampled, or made conditional on what changed.
The step set is identical — 31 in, 31 out, asserted at edit time — and each
step runs as the same command graded by the same exit code. Only start times
moved. The && BETWEEN phases is load-bearing and stays: a source gate still
blocks the build, and the build still blocks every gate reading its output.
Within a phase the runner reports ALL failures rather than stopping at the
first, so three broken gates cost one fix-and-rerun cycle instead of three.
It imposes no timeout of its own, because a command killed by a short
timeout is not a failure.
Negative-tested three ways, each required to exit 1: a failing gate inside a
phase, a red phase halting the chain, and a failing vite build inside
build:bundles. The bundle case also confirmed the runner enumerates
plugin-designs live (it picked up 23 dirs, not a cached 22).
check:sri failed whenever a single-bundle `BUNDLE=x bun x vite build` left a
leaf's baked integrity behind its bundle — and the gate then told the agent the
fix was `bun run build`, whose first step deletes public/doxa-maps/ and tears
the tree out from under every concurrent reader.
Adds `bun run fix:sri`: rewrites the disagreeing hashes in the built tree and
nothing else — leaf <script integrity>, manifest bundles[].integrity, and the
paste-verbatim embed snippets. 0.4s, writes only the offending pages, and the
bytes it writes are identical to what the full build wrote.
check:sri's failure text now prescribes that instead of a rebuild, and its
header no longer claims generate-tree runs only inside the full build (it does
not; `bun run tree` re-bakes the hashes too, at the cost of rewriting 1234
files under public/ and re-reading .vue source every run).
Negative-tested: `node internal/fix-sri.mjs --self-test` builds a poisoned
synthetic tree, requires check:sri to go RED on it first, then green, then a
second repair to write zero files.
A single-bundle build regenerates the JS but not the leaf pages, so four
big-picture leaves carried a stale sha384 and the browser refused to execute the
script at all. That is why a nav probe measured 0 sections on a shell whose
source was correct, and it is the likeliest cause of a teammate's 'pggMounted:
false on big-picture everywhere' reading.
check:sri named both the cause and the fix; ran the full build it prescribes on
a settled tree. Now PASS: 44 script references across 45 leaf pages, every
leaf's integrity matching the bundle it loads.
THE RAMP. The Driver reported that 60-149 reads as the same red as the 150+
heartbeat bin. The old ramp was mathematically even and he was still right --
that tension was the finding, not a contradiction. Measured in plain sRGB
distance the top step (60.0) looked like the LARGEST in the ramp; that
measurement was lying, because RGB distance is not perceptual. Re-measured as
deltaE the old steps were 15.5 / 15.8 / 16.7 / 17.7 -- nearly flat, while chroma
collapses at the dark end so hue and saturation stop carrying information there.
The new ramp deliberately ACCELERATES toward the dark end: 14.1 / 17.9 / 21.4 /
24.9. Top-bin separation 17.67 -> 24.94, a 41% improvement on exactly the pair
he named, and total lightness range widens from 37.2 to 49.0 L*.
A named greyscale ramp ships alongside it, grey to black, as he asked ("just in
case i need to swap from red to black scale"). Its steps measure 22.8 / 22.4 /
22.4 / 22.8 -- near-perfectly even, every step well above the ~10 threshold for
clear distinguishability. It is an OPTION, selected by name; the red ramp
remains the default, so no prior decision is silently reversed.
Bin EDGES are untouched -- check:heartbeat passes, which replays the bin
assignment and fails if a default moves.
THE INTEROP GATE. verify:interop drives the actual compiled cap-4 composites in
a real browser against live data, closing the gap verify:gate1 leaves open. On
its first run it PASSED both desktop directions: a facet-rail click narrowed the
grid to "Showing 1,309", and selecting three rows moved the camera from
28.000,18.000@1.45 to 73.666,13.162@2.48 -- read back off the live map, not
inferred. It asserts the grid's live count line rather than card count, because
card count is a proven false-negative (always 24).
It ships with verify:interop:self-test, a two-mode negative test -- grid-prop
freeze and camera freeze -- each required to fire its own combination-named
failure. A harness that has only ever reported PASS is unproven.
Gates: sfc-parse, template-literal, one-palette, synced-css, mjs-parse,
heartbeat, doc-scripts all pass.
Every gate in this repo verifies INTERNAL consistency. check:facet-changes-map
proves the dots change and that the categories sum to 2,107 — but it asserts
that inside the page, against data the page already loaded. If the upstream
corpus shifted, the page would stay perfectly self-consistent while its legend
quietly stopped describing reality. Nothing fetched the live endpoint: not the
six data guards, not check:label-joins, not any browser gate.
That is the worst failure this product can ship, because it is invisible to
every structural check. The element mounts, the canvas paints, the counts are
formatted correctly, the partition sums. Only comparison against the SOURCE
catches a legend that is beautifully wrong — and these maps exist to say where
the gospel has not reached, so a wrong number is not a cosmetic defect.
Measured by hand first, then encoded. The cap-2 engagement legend shows
Unengaged 2,080 / Engaged 27; the live API returns 2,107 rows splitting exactly
2,080 / 27, with no third bucket and no remainder. The cap-4 dashboard KPI strip
prints "1% ADOPTED"; the live API gives 21 of 2,107, which is 1.0%. Three
independent surfaces agreeing to the unit.
The gate also asserts the partition is COMPLETE and that no category exists which
the two-row legend does not show — a third engagement value would hide rows from
the reader while every count still looked right.
The expected values are a baseline, not a law, and the failure message says so:
the corpus is living data and these numbers will move. A mismatch is a prompt to
update the legend, the KPI and the baseline TOGETHER, which is the coupling that
was missing. Updating the baseline alone would silence the gate and leave the
maps wrong.
Offline SKIPs at exit 0 rather than failing, because reporting a mismatch that
may not exist is the fail-open mistake in reverse. Negative-tested both ways:
drifting the baseline fails naming the exact category and both numbers;
an unreachable host skips and says why.
The browser user-agent is deliberate — pray.doxa.life answers 403 to a default
script UA, a trap already recorded here, and one that presents as "the data is
gone".
research-map-plus-grid had the same defect fixed yesterday on its desktop
sibling: sharing the composite emitted a link to doxa-research-map, the child it
nests. The gate said it exactly — "share URL reopened <doxa-research-map>, not
<research-map-plus-grid>" — and a recipient got the map alone, losing the grid.
Checked all four cap-4 composites rather than assuming the fix generalised.
my-upg-100-list owns its Share button directly and is fine. big-picture composes
a header plus the grid, owns no map and ships no Share button at all; the gate
reports "window.mapboxgl UNDEFINED" against it, which is its map precondition,
not a share failure — that bundle is out of this gate's scope and I put it in
the probe by mistake. So one real instance, now fixed.
Tier 1 again, for the same reason: this child also hardcodes a prop
(research-map.vue:2149 `:embed-path="'doxa-research-map/research-map/index.html'"`)
and the ladder is `embedPath || cfg?.shareEmbedPath`, so a prop outranks host
config and tier 2 is unreachable. Delivered through the profile-config attribute
rather than Vue provide, because this composite mounts its children as real
custom elements via src — there is no provide chain to reach them.
The url had to be ABSOLUTE, and the first version was not. resolveShareUrl
returns a tier-1 shareUrl verbatim (embedBase.js:95, `return whole.trim()`, no
resolution), and ShareButton feeds the same string into its embed snippet:
<iframe src="{shareUrl}">. A relative '../../…' would have resolved correctly in
the share LINK, because the browser resolves it against this page, and then
404'd in the EMBED, because it resolves against the partner's page instead. That
would have passed the gate — the gate navigates the link — while breaking the
deployment method the link exists for. Tier 1 is documented as "a WHOLE url" for
precisely this reason. Now computed with resolveEmbedBase() so it stays correct
on doxa.life, on nuxtbase.app and in a dropped tree.
Verified the desktop sibling did not have the same latent flaw: it produces a
fully absolute url, confirmed in the live gate output.
smoke:share-all now carries eleven targets. Full suite 11/11 pass · 0 fail.
Eleven gates pass, including sri, composite, composite-srcs and synced-css.
Sharing from dashboard-plus-grid produced a link to
research-dashboard/research-dashboard/index.html — its own nested child. A
recipient opened the inner dashboard alone and silently lost the grid half of
the composite. The gate said it precisely: "share URL reopened
<research-dashboard-map>, not <dashboard-plus-grid>".
The nested dashboard's DashboardTopBar owns the Share button, and inside the
composite that button had no idea it was a passenger, so it shared its own
address.
Two attempts, and the first one is worth keeping in the record because it built
and shipped and changed nothing. I set profileConfig.shareEmbedPath — tier 2 of
ShareButton's documented ladder, and the natural-looking channel. The ladder is
`embedPath || cfg?.shareEmbedPath`, and DashboardTopBar passes an explicit
prop (`<ShareButton :embed-path="SHARE_EMBED_PATH" />`), so a child that has
already made up its mind always wins. The value was baked into the bundle,
resolveShareUrl reads that key, and the rebuilt bundle failed the gate
identically — a fix that is present in the artifact and unreachable at runtime.
Tier 1, `shareUrl`, is the only rung above a prop, and its docs say exactly what
it is for: "a WHOLE url the HOST supplies wins".
Tier 1 takes a whole url, so it is computed rather than hardcoded, via the same
resolveEmbedBase() helper ShareButton itself uses: profileConfig.embedBaseUrl →
window.MAP_EMBED_BASE_URL → the root the built bundle derives from its own
script url. The link therefore stays correct on doxa.life, on nuxtbase.app and
in a dropped tree with no configuration. An absolute url would have pinned every
share to whichever host happened to build it.
Also adds the two uncovered bundles to smoke:share-all. Nine bundles ship a
Share affordance and the hardcoded --url list covered seven; research-dashboard
and dashboard-plus-grid were absent, which is why this survived. The list is now
ten targets over nine bundles. Note the structural difference from
smoke:embeds, which enumerates from manifest.json and therefore cannot omit a
bundle that ships later — a hardcoded list silently omits whatever arrives after
it was written, and that is what happened here.
This is the two-host defect class exactly: research-dashboard standalone PASSES
this gate, and the identical component inside the composite FAILED. No
single-host gate can tell those apart.
Verified: the composite now reopens <dashboard-plus-grid> with a canvas, and the
full suite is 10/10 pass · 0 fail, so nothing regressed. Eleven gates pass,
including sri, composite, composite-srcs, synced-css and one-palette. The share
gate aborts rather than reporting a false failure when handed Firefox, which has
no WebGL here — it told me to re-run with chromium, and it was right.
package.json described verify:browser-all as running "ALL 7 browser gates". The
GATES array in internal/verify-browser-all.mjs holds nine. Two were appended
below the count — check:fixed-anchoring and check:shell-scroll, both added the
same week — and the prose was never touched, so a reader who ran the command and
tallied seven would have concluded two gates had not run.
This is the same shape as the runbook's "all 6 DX guards PASS" corrected earlier
today, and the fix is the same: stop stating a number that lives in another
file. A prose count of a list that grows elsewhere rots on the next append, and
it rots silently, because nothing compares the two. The description now says it
runs every gate in the array and explains why it refuses to give a figure. The
script already prints its own GATES.length at runtime, which is the number that
cannot be wrong.
Also repairs a garbled sentence in the same string — "Only smoke:share the
`verify:browser-all:plugin-gallery` script mode is out of scope" appears to be
two edits collided. It now says plainly that the plugin-gallery mode is separate
because it needs the vite dev server.
One line changed; check:quoted-scripts and check:doc-scripts both pass, which
matters because those gates compare this file against the docs that quote it.
check:library-index existed, was correct, and had been failing for about 19.6
hours. Nothing noticed because it runs only when a person types it:
build:unlocked chains 24 gates and this was not one of them.
check:no-orphan-gates passes on it, which is worth understanding rather than
treating as a gap in that gate. It asks whether every gate is REACHABLE — "named
in package.json or imported by something that is" — and check:library-index is
named there. Reachable and wired are different properties. 64 gate-shaped files
exist; 24 run on every build; the rest only run on request.
The fix is to regenerate rather than to check harder. check:composite is also
mtime-based and is in the chain, but it guards build output whose freshness
depends on build ORDER — a real invariant only a human can resolve.
reference-library-index.md is DERIVED from library/, so a staleness failure has
exactly one remedy: run the generator. Wiring the check alone would fail builds
and hand the author a chore; wiring the generator makes the staleness
impossible, which is the only way the file's own header — "DERIVED — cannot
drift from the code" — can be true.
Both are now in the pre-build group, generator then check, beside the other
source-only gates and before `clean`. Safe to run unconditionally: measured
deterministic (byte-identical on re-run) and 237ms. The check is kept after it
so a generator that silently writes nothing still fails the build rather than
passing quietly.
Updating package.json made check:quoted-scripts fail, which is the gate doing
its job, and it surfaced that the quoted copy in README-SECTION-3 was ALREADY
stale before this change: it omitted check:tab-radius, check:synced-css,
check:wp-embed and check:doc-scripts. An omitted step in a quoted list is
invisible — there is no wrong character to notice, only a check the reader
believes is not in the chain. The doc now matches what runs.
Verified by a full build: the generator ran, the check passed, vite built every
bundle, exit 0. The index is not dirty afterwards, which is the determinism
claim holding inside the real chain rather than in an isolated test. Five lines
in the build log match /error|FAIL/ and all five are gate DESCRIPTIONS
containing those words, not results.
Applying last tick's lesson -- grep for the class, not the instance -- found a
third surface teaching the old serve command. README-SECTION-3-cli-commands.md
transcribes the package.json scripts so a reader need not open the file, and it
still showed python3 -m http.server two commits after serve became
serve-strict.mjs. An inheritor following the CLI reference would have started the
permissive server, the one whose forgiveness hid a URL bug in two gates for a
whole flight.
Measured before fixing: ten quoted scripts, nine matching, exactly one drifted. I
had expected build:unlocked to be stale too and it is not -- the doc truncates it
rather than misquoting it, which is the right call for a seven-hundred-character
line.
Hand-fixing the one line would leave the next script change to rot the same way,
so the doc is now checked. The comparison is loose in the three ways a
human-facing doc needs: package.json's trailing hash comments are ignored, the
doc's own slash-slash annotations are ignored, and truncation passes on a prefix
match either way. A quoted script that does not exist in package.json is a hard
failure, since that is a command someone would type and get "script not found".
Completeness is deliberately not asserted; the doc is a curated subset and
demanding otherwise would make it churn on every internal gate.
Self-test poisons the doc in memory two ways, a changed body and a nonexistent
script name, and requires both to be caught. Both are. Full build passes with the
gate in the chain.
check:wp-embed existed as a script and nothing invoked it -- the same orphan
shape as check:synced-css, which had been red and unrun for weeks. This one is
green and unrun, which is quieter but no better: a gate the build does not
consult is documentation.
It is worth running. It renders the real shortcode against stubbed WordPress
functions and asserts the output against the real built manifest, so it needs
neither MySQL nor Docker. Its load-bearing assertion is that the wrapper emits
the same SRI integrity hash the catalog ships, which is what gives the WordPress
path the same tamper protection as the direct-script embed. It also carries a
negative case: a malformed integrity value must be rejected.
Full build passes with it in the chain.
bun run serve was python3 -m http.server, which resolves a directory request to
its index.html. The live static layer returns 404 for the same request. So a gate
could pass locally on a URL that does not work in production, and nothing local
could ever catch it.
That is not a harness quirk, it is a blind spot with a green light on it: every
defect inside the permissive gap passes locally by construction and can only be
found in production, which is the worst place to find it. It cost a false
"the publish failed, zero filters live" reading on a publish that was
byte-for-byte correct.
serve-strict.mjs returns 404 for a directory, exactly like production, with no
index fallback and no directory listing. It also refuses to read outside its
root. Loopback only, inherited from the script it replaces, whose own comment
records that the bind was once 0.0.0.0 and answered on the box's public IP while
a gate was running. The permissive server is kept as serve:permissive rather
than deleted, so the difference can be demonstrated.
It earned itself immediately: verify-shell-scroll held two bare-directory URLs
and failed the first time it ran against the strict server. Both now ask for
index.html, which is the form the manifest already ships for all 28 profiles,
and the gate passes again -- both wheel gestures moving the grid 1500px onto
screen.
Fixing the server rather than the URLs is the point. The repo was already
production-correct: 28 of 28 manifest URLs and 82 gate references use
index.html. What was missing was anything that could fail when one did not. A
rule you have to remember fails under time pressure; a server that returns 404
does not.
The gate reported 14 conflicts across two files. Thirteen were false and it was
blind to real ones, and both came from a single cause: it keyed a rule on the
last class before the brace, so any qualified rule was filed under its own base
rule. A dark-mode override, a .dsm-dark variant and a :host(:fullscreen) rule all
merged into the unconditional rules they were meant to override, and the pair was
then reported as the file disagreeing with itself. In every case both copies were
identical and the code correct.
Merging is last-write-wins, so the same bug also HID conflicts. Poisoning
.dsm-root inset from 0 to 7px in the shadow copy produced no finding at all,
because the fullscreen rule's own inset:0 was merged in afterwards and restored
it. A property could be shared, genuinely divergent, and invisible.
The key is now the full selector, and at-rules are part of it. The header already
claimed media-query variants were treated as distinct rules; nothing implemented
it, and PeopleGroupDetail's 18px name with a 16px @media override read as the
file contradicting itself when the shadow copy simply has no mobile rule -- which
this gate's own contract explicitly allows. At-rule bodies are brace-matched
rather than regex-matched, since a lazy pattern stops at the first inner brace.
The shadow literal is also sliced AFTER its opening backtick. Slicing from the
anchor left useShadowStyles(` at the head of the CSS, which a full-selector
pattern absorbs into the first rule -- .dsm-root was captured as
"useShadowStyles(` .dsm-root" and silently dropped.
That leaves one true finding, and it was real: .action-buttons padding-top was
12px as a Vue child and 4px in a shadow root. Both values date from the initial
commit, so there is no later intent to infer; aligned to the shadow copy's 4px
because that is the copy that reaches the shipped bundles, so it preserves what
users have been seeing rather than changing it.
check:synced-css:self-test poisons three properties in memory and requires each
to be caught. inset is there deliberately -- it is the one the old implementation
missed, so it is the case that proves the fix. Now in build:unlocked, which it
could never be while it was red.
The 999px stadium was a real instruction that shipped, and its rationale is still
in the layout file as history. A superseded rationale that still reads as current
is how a reversed decision gets undone -- a reader trusting it would restore the
pill believing they were fixing a regression. The comment now says not to; this
makes that enforceable.
It asserts the TOKEN rather than the value, because the requirement is that the
tabs, the collapse button and the filter toggles agree. A shared knob guarantees
that; three matching literals only coincidentally satisfy it. The expected
fallback is READ FROM .vtl-collapse rather than hard-coded, so a gate asserting
8px cannot pass while the collapse button moves to 10px and the two surfaces
silently diverge. Both synced copies are checked, so a fix applied to one host
fails here rather than shipping correct in one and wrong in the other.
.vtl-cap keeps its pill and is deliberately not asserted: the complaint was the
ratio of radius to size, and a gate banning 999px everywhere would over-apply the
correction.
check:tab-radius:self-test poisons the source in memory and requires both
regressions to be caught -- the pill restored, and a bare literal beside the
token. Both caught, tree untouched. Wired into build:unlocked beside the other
source-level checks so it fails in a second rather than after the vite loop.
Noted while wiring: check:synced-css is NOT in the build chain, which is how two
files have stayed in conflict. Left alone here -- turning it on is its own change
with its own failures to fix first.
check:heartbeat proved the PANEL FALLBACK was intact and called that "default
ramp unchanged". It is not the same property. The running app resolves the ramp
with dashboard.config.js's map.* layered on top of that fallback, and the gate
never read the real config: assertion F checked only that heartbeatThreshold was
DECLARED, never what it was SET TO.
Measured consequence, reproduced before fixing:
sed -i 's/heartbeatThreshold: 150,/heartbeatThreshold: 999,/' \
plugin-designs/research-dashboard/dashboard.config.js
node internal/check-heartbeat-threshold.mjs # exit 0
The shipped map could silently change its ramp while the gate reported it
unchanged. Now exits 1, naming the key and its value.
Adds assertion A2: resolve the ramp the way the running app does — panel
defaults with the shipped map.* applied — and require it to equal the historical
5/20/60/150/Infinity with its labels. A (fallback) and A2 (shipped) share one
comparison helper so they cannot drift. A2 also rejects a shipped heartbeatMode
other than 'fixed': a data-derived ramp is right for an unknown dataset and
wrong for this one, where 150 is deliberate. Changing the shipped ramp is now a
deliberate act that must update HISTORICAL in the same commit.
dashboard.config.js is read by brace-matching its `map:` literal rather than
import(): it pulls in @map/* aliases and a JSON module, so importing it would
need Vite's resolver — a dependency that buys nothing for a bag of scalars.
The self-test now runs three cases, each required to fire the assertion that
exists for it, and poisons in memory so a crash cannot leave the tree edited:
[literal] resolver replaced by the old array -> must fire B
[shipped-default] shipped heartbeatThreshold changed -> must fire A2
[shipped-mode] shipped heartbeatMode flipped to quantile -> must fire A2
The header previously claimed the gate was "negative-tested" on the strength of
the first case alone. True of that case, misleading overall, so it now states
per-case coverage and what is NOT covered (colour contrast, and whether the map
paints at all). Testing that the FEATURE works — a configured threshold moves
the edge — is not testing that the REGRESSION is caught, and only the second
protects the requirement that today's map must not change.
The prayer filter was reported as "not applying" to the map. Measuring it first
changed the fix: the filter was never broken. Selecting "Has Prayer" already
narrowed the rd-bubbles source 2107 -> 689 features, exactly matching the
dropdown's own tally, disjoint from "Needs Prayer" (1418), summing to 2107 and
round-tripping on clear.
What was broken is that circle-color was the flat literal '#ef6c00' — read back
off the live layer, the bare string, not an expression — so every dot painted the
same orange whether it had prayer or not. A prayer question got an answer with no
prayer information in it, and the surviving dots looked identical to the ones the
user started with. That is why a working filter reads as "not applying".
This gate asserts all three halves, because any one alone is a false pass:
1. the selection changes WHICH dots are on the map, by count AND by identity
2. every surviving dot satisfies the level that was selected
3. while a prayer filter is active the colour is a data-driven expression
reading peoplePraying, not a flat literal
Assertion 3 is the one that matters: 1 and 2 were already green while the bug was
visible. Identity is checked as well as count because the filter removes from the
tail of an alphabetical list, so the first five slugs are byte-identical before
and after — a gate reading the head of the list would report "unchanged".
NEGATIVE-TESTED in four modes, each isolating one assertion, all verified caught:
colour reverts only the paint (the reported bug) -> caught by assertion 3
props strips peoplePraying -> caught by assertion 2
frozen blocks the source update entirely -> caught by assertion 1
both reverts the lot -> caught
A gate that cannot fail is worse than no gate.
Verified on both hosts against the BUILT tree: the cap-5 shell and the standalone
leaf. Browser gate, so not in build:unlocked.
The repo is handed to the Doxa team, and a code name from its build history
means nothing to an inheritor. Two places carried it, both regenerating:
- public/.build.lock was TRACKED. internal/build-lock.sh stamps it with
`by=$(whoami)` under flock on every build, so each rebuild committed the
building machine's username. Scrubbing the string would have been useless;
the file is pure runtime state nothing reads but the lock, so it is now
ignored via public/.gitignore instead.
- check-no-absolute-paths.mjs quoted a real home path in its header to
explain the rule. The prose works the same with /home/<user>/.
check:no-codename makes it stick. It scans tracked files only (the set a clone
receives), names every offending file:line in one pass, and refuses to pass
vacuously on a short file list.
Two things it has to get right or it becomes noise. "10x" is ordinary English
for a multiplier and library/scripts/research uses it correctly, so only the
full two-part name matches. And the same letters appear as the unix username
inside /home/ paths, which is a portability defect check:no-abs-paths already
reports better — those are listed, not failed, so the two gates cannot
disagree about one string.
It carries no ALLOW entry for itself. A gate must never be its own witness,
and an exemption is a hole a real leak could hide in, so the pattern is
assembled from fragments at run time and the file scans itself on the same
terms as every other.
Negative-tested: planting the name in README.md fails naming README.md:2,
exit 1; removing it passes. All five spellings (LK10X, LK 10X, lk10x, LK-10X,
lk_10x) planted across five different shipped surfaces are reported together
in one run, exit 1; reverting returns exit 0.
Also quotes check:heartbeat into the CLI doc, which check:quoted-scripts
requires and which landed in package.json without it.
A `position:fixed; inset:0` scrim is only viewport-relative while nothing
establishes a containing block for fixed descendants. VerticalTabsLayout sets
`contain:layout` on .vtl-content, so inside the cap-5 studio every such scrim
covered the CONTENT PANE instead. Measured at a 1280x900 viewport:
.dt-scrim / .modal-scrim / .lt-overlay ... 1024x793 @ (256,107)
captured by div.vtl-content [contain:layout]
256 is the rail width, 107 the header height — left uncovered and CLICKABLE
behind what still looks like a modal. The fix (library/utils/viewportScrim.js)
measures the real origin and subtracts it, rather than removing the shell's
contain, which is load-bearing for a different reason.
This adds the gate that keeps it fixed. A source grep cannot: every one of these
scrims greps as a correct inset:0, because the variable is the ANCESTOR, not the
declaration — and which ancestors exist depends on the HOST, not the component.
check:scrim-viewport:self-test reverts each scrim to plain inset:0 and requires
EVERY case to fail; a gate that cannot fail is worse than no gate, and this one
measures a property that looks fine in a screenshot either way. The self-test
also surfaced that .lt-overlay is captured TWICE — div.lt-root [transform] as
well — collapsing it to 233x30 rather than merely to the pane.
SemanticTreeLegend's .stl-export-backdrop is fixed the same way but marked as
NOT verified through the UI: it sits behind SHOW_EXPORT_BUTTON = false, so the
menu that owns it cannot be opened. Correct-if-enabled, not observed.
Five lanes had 465 insertions across 18 files sitting uncommitted, some for
15 minutes. That is the state that silently lost the shell-scroll fix earlier
today: a verified fix existed only in the working tree, a second agent's edit
to the same file overwrote it, and the reverted version got committed and
described as the fix.
Nothing here is finished and no card is being closed on it. This is a
snapshot so the work survives an overwrite, taken at a moment when every file
had been quiet 15s or more, so no agent was mid-write.
Verified coherent before staging rather than committed blind: template-literal,
mjs-parse, sfc-parse, no-orphan-gates, kpi-unit, composite-reg, script-flags,
quoted-scripts and backlog-status all PASS. check:synced-css fails on
PeopleGroupDetail.vue and doxa-simple-map.vue only -- the same two files and
the same counts as before the swarm started, both untouched by any agent, both
already carded.
Two untracked files included because both are load-bearing: internal/
check-facet-one-field.mjs is wired into package.json with a self-test, and
library/utils/viewportScrim.js is imported by DataTable, SemanticTreeLegend
and PosterDialog -- a shared helper for the modal-scrim fix rather than six
copies of the same workaround.
Two different agents broke the build the same way within ten minutes: a
backtick inside a // comment sitting inside a template literal of generated
browser JS. The backtick closes the literal and the parser reads prose as code,
so it fails hundreds of lines from the edit -- one break reported at :386 with
the cause in a comment above it.
check:template-literal did not catch either, and that is not a bug in it: it
scans backticks inside CSS comments, the shadow-CSS shape. A // comment inside
a generated-JS literal is a different grammar. Proven by controlled negative
test -- injecting a backticked term into a // comment in generate-tree.mjs, a
file that gate explicitly lists, left it reporting PASS while node --check
failed immediately. The rule was right; the implementation covered a narrower
case than the rule stated.
Nothing parsed the 70 .mjs build scripts. check:parse covers inline scripts in
built pages and check:sfc-parse covers .vue, so the scripts the build is made
of were the one unchecked surface.
Uses the parser rather than a wider regex. A backtick is legal inside a string,
a regex, a nested template, or a comment inside a template; a pattern precise
enough to catch the bad case would false-positive on the legal ones, and a gate
that cries wolf gets skipped -- worse than the gap. The parser already knows
all of that.
Parses, never imports: these scripts read files, spawn browsers and hit the
network at import time, and a syntax check must have no side effects.
Runs as build step 2, beside check:template-literal, so a broken script fails
in a second rather than after minutes of compilation.
The failure message points at the real cause rather than the reported line, and
tells the reader to grep for every instance in one pass -- node --check reports
only the first, and each rebuild to surface the next costs minutes, which is
how one of today's breaks was being fixed one at a time.
Negative-tested side by side: exit 1 with "SyntaxError: Unexpected identifier
'TREE'" at generate-tree.mjs:369, while check:template-literal reports PASS on
the same file in the same moment.
Cards were being written with explanatory statuses -- "done — reconciled: …",
"blocked — Steps 1+3 DONE and re-verified…", "ready". The prose is valuable
and several of those strings carry measurements, but it made status
unqueryable: every count used status !== 'done', which is FALSE for
"done — reconciled: …", so finished cards kept counting as open.
Measured across the three phase backlogs: 130 of 395 cards carried a prose
status and 71 of those were already done. Reported open was 176 against a real
105. The P0 figure quoted every tick was 21; it was 14.
That is worse than an off-by-N because a backlog is a decision instrument. It
answers "what is left" and "is the rest human-gated", and inflating the open
count makes a drained board look stalled, which pushes toward manufacturing
work instead of holding honestly. The error also hides: every card reads
correctly to a human, because the prose says exactly what is true. Only the
aggregate is wrong.
Migration preserved every explanation in fields.status_detail -- 130 of 130,
none lost -- and normalised status to the enum. Six cards whose head-word was
not in the map were left alone and resolved by hand against reality rather than
guessed, since a wrong normalisation hides a card instead of counting it.
The gate does not require status_detail. Most cards need no explanation and
demanding one turns a real check into paperwork satisfied with filler.
Negative-tested: restoring one prose status fails by name and prints it.
The Driver named the sequence: Big Picture, UPG-100, Prayer Progress, Gospel
Resources. Tab 1 is the landing surface -- the first thing anyone opening the
product SPA sees, and the opening step of the onboarding ramp. That ordering
was held only by the order of entries in NAV_MODEL and by comments explaining
why, with nothing failing if someone reordered them.
Easy to break silently: NAV_MODEL is also where sections, badges and
narrow-screen flags live, so it gets edited for reasons unrelated to ordering,
and a row moved while adding a badge produces no error and a SPA that still
works. It just opens somewhere else. That file's own comments record Big
Picture once being painted twice in the rail, from a one-row section that
re-declared a row below.
Asserts the RENDERED rail, not the source array. Parsing NAV_MODEL was the
first attempt and returned an empty list -- the entries do not carry the id
shape I assumed, and a regex over a 300-line commented array is guesswork
either way. The rail is what the user sees and what the Driver described, and
a source parse could pass while a rendering bug shows something else.
Order, not membership: checking that four names are present would pass on a
rail that opens on Gospel Resources. Prefix-matched so a row may carry its
subtitle.
Negative-tested: swapping the first two expected entries fails by name and
prints want versus got for every row.
.filter-dropdown was flex:1 1 160px with min-width:160px and max-width:240px
-- a range, not a width -- so flex-grow renegotiated the moment the selected
label appeared. The row grew 960 to 1294px and triggers reached 240px.
Not cosmetic. The selections row with Clear All sits underneath the filters,
so a filter that grows on selection re-flows the row below it and the band
stops being a fixed block. The stable width is what lets the centred stack
hold still.
Independently reproduced before gating rather than taken from the agent
report: 168px across all six triggers, identical after selecting the longest
option in each. The negative test then reproduced the broken widths
[240,240,240,184,171,219] -- the same array the original report cited, which
confirms that measurement was real.
Asserts the observable outcome, not the CSS. A grep for --dx-filter-w would
pass while a missing min-width:0 on the label still pushed the box open: as a
flex item without it the label is floored at min-content and refuses to
ellipsise. Both halves are needed, so the gate measures the width.
Selecting the LONGEST option is load-bearing. A short option fits inside the
old range too, so a probe that clicks the first option passes on a build that
still has the bug.
Two hosts, and an unmeasurable host fails rather than skips -- three defects
this week were invisible standalone and appeared only in the shell.
Negative-tested: restoring the flex range fails both hosts by name, naming the
before and after arrays and the total growth.
In the cap-5 shell the Region donut, country bars and Place tree rendered
entirely grey while the identical build was correct standalone. A cap-4
composite imports the sibling component directly
(dashboard-plus-grid.vue:18), which takes the COMPONENT and skips
research-dashboard/index.js -- the file that actually assembles the bundle:
Pinia, i18n, colour strategies, customElements.define. The composite
re-declares all of it by hand, a manual copy nothing keeps in sync, which is
how the strategies leg was lost. That file already documented this bypass for
i18n and fixed only that one leg.
It is silent by construction: getColorStrategy() falls back to language-family
for an unregistered mode instead of throwing, so the result is a plausible
wrong picture -- grey charts that read as a data problem rather than a wiring
one. Found by reading an image; every number passed.
The gate compares the leaf's registerStrategies keys against the composite's.
A composite registering nothing while its leaf registers nothing is fine: the
rule is about drift between two lists meant to agree. big-picture also bakes a
sibling and registers zero strategies, and that is correct -- neither it nor
people-groups-grid references colorFor, getColorStrategy or mapbox. A first
pass called it a second instance; measuring usage showed it is not one.
Four rounds to get working, each a real bug found by testing the wrong half:
- pre-stripping comments deleted 2,953 of 4,102 characters of
research-dashboard/index.js, including the call being searched for, so the
gate claimed "no composite bakes a sibling that registers strategies" --
a confident claim about the repo derived from parsing nothing. Not stripping
is also correct: the call is located by requiring "(", so a mention in prose
cannot match.
- indexOf('registerStrategies') found the IMPORT line first and brace-matched
its own braces.
- a non-greedy brace regex stopped at the first inner "}" and the key pattern
missed quoted path keys.
Every fix was verified against dashboard-plus-grid, which survives stripping,
and never against research-dashboard, which does not. A comparison needs both
halves asserted before any fix is believed.
Negative-tested: dropping doxa-region from the composite fails by name and
prints both lists.
A shared component declares the same selectors twice: once in a scoped <style>
block for when it is imported as a Vue child, and once in a useShadowStyles
literal for when it is mounted inside a shadow root, where scoped CSS does not
reach. Both are load-bearing and the standing rule is that they stay
identical. Nothing enforced it.
library/components/PeopleGroupDetail.vue disagreed with itself on eleven
properties, and the shadow copy is a coherent DARK theme under a light-only
rule: .detail-header background white vs #16213e, .image-container #f5f5f5 vs
#2d2d4e, and .people-group-name color #333 vs #fff -- white text, legible only
on the dark background, so fixing the background alone would make the name
invisible. Six importers; the dark hex reached four shipped bundles.
It was invisible because it is a two-host bug: the component looks correct in
whichever host you tested. Fourth of that shape this week.
Three findings while building this, each of which had the gate reporting
success about something it never compared:
- a loose /<style[^>]*>/ matched the words "<style scoped>" inside a JS
comment 43 characters before useShadowStyles( in SemanticTreeLegend.vue, then
ran to the next real </style> and swallowed the whole shadow literal into the
<style> text. Both maps read the same bytes, so that file could not fail on
any of its 51 apparent shared selectors. Anchored to line start, since an SFC
block starts at column 0. That also raised PeopleGroupDetail 8 -> 11.
- assigning m[sel] = props let the last declaration win, discarding every
property that appeared only in a base rule when a media query redeclared the
selector. Rules are merged now.
- comparing raw text reported comma-spacing as divergence: 5 findings for
doxa-simple-map of which 2 were formatting. Compared as property maps, so the
rule is about meaning.
Negative-tested by change rather than by exit code, which with pre-existing
failures proves nothing: fixing .detail-header takes PeopleGroupDetail 11 -> 10
and restoring it returns 11.
Standalone rather than in build:unlocked, because it fails on live defects
today.
The dashboard rendered "99% % UNREACHED" and "1% ADOPTED %". kpiText returns
k.value unchanged for kind:'pct' and that value already carries its percent
sign, while the labels in dashboard.config.js carried a second one.
Each half is correct alone. A label reading '% unreached' is a reasonable
label; a formatter returning a ready-formatted percentage is a reasonable
formatter. Only the composition is wrong, and neither file mentions the other,
which is why it was found by reading the rendered screenshot rather than by
reading either source.
check:kpi-unit asserts no kind:'pct' label contains a '%'. The value owns the
unit because it is produced by the formatter that knows the number; the label
says what is being measured.
Parses the config as text rather than importing it -- the research-dashboard
module graph reaches library/colors/_registry.js, which uses import.meta.glob
and throws under plain node. Text parsing keeps it hermetic enough for
build:unlocked. Comments are stripped first, because the config now documents
this defect right above the fixed line and a naive scan fails on the
explanation of the fix.
It also fails when it finds no pct entries at all, rather than printing a
green PASS: if the config is refactored so the parse stops matching, silence
would read as success forever.
Negative-tested: restoring either label fails by name with exit 1.
The dashboard aggregations module used to hold two private palettes the map
had never seen, and picked chart colours with
CATEGORICAL[i % CATEGORICAL.length]. Two failures came out of that: near-miss
regions (Africa #e74c3c in a chart, #ff2e2e on the map) and religion coloured
by array index. No religion key existed in the private table, so every
religion fell through to its position in a list sorted BY COUNT -- and counts
move when you cross-filter, so a religion changed colour as you filtered.
Islam and Hinduism swapped, Buddhism and Christianity swapped.
That is already fixed: both private palettes deleted, every hex now read from
the same strategy modules the map paints with. But the fix is a deletion plus
a convention, and nothing enforces a convention. The next person adding a
facet writes a lookup table because it is the obvious thing to do, and the
divergence returns with no error and no crash -- just two surfaces disagreeing
about what blue means.
Reads the BUILT bundles rather than the source because the registry discovers
strategies with import.meta.glob, which plain node cannot evaluate. That makes
it a test of the glob too: if the registry ever silently discovers nothing,
the bundles lose the palette and this fails, where a source-only check would
still pass.
Comments are stripped before scanning, because both consumer files describe
this defect at length in their headers and a naive scan matches the
documentation of the fix -- the same self-reference that made two gates report
on their own prose this week.
Negative-tested three ways: a reintroduced private palette, an [i % len]
lookup, and corrupting four of ten religion hexes in a built bundle, which
fails as a partial palette and names the missing hexes.
verify-shell-scroll.mjs was written and negative-tested, then left with no
package.json entry and no importer, so nothing could ever run it again. It
also defaulted to localhost:8899 -- an ad-hoc port from an earlier session --
so it failed with ERR_CONNECTION_REFUSED for anyone else and read as a broken
gate rather than a missing server. Now check:shell-scroll, defaulting to
BASE_URL or 127.0.0.1:4173 like every other browser gate, and listed in
verify-browser-all's GATES array.
check:no-orphan-gates fails any gate-shaped file in internal/ that nothing can
invoke. The repo carded this once already; it recurs because writing a gate
and registering one are separate acts and only the first feels like work.
Not every unreferenced file is a lost gate, so the rule is not "everything
must be wired": theme-flash-probe is a dated diagnostic for two fixed bugs,
worth keeping as a reproduction recipe. It now carries an @one-shot tag saying
so, which makes the author state which kind of thing they wrote.
Two self-referential defects in the first version of that gate, both from
reading evidence out of documentation:
- it cited theme-flash-probe.mjs in its own header as an example, then grepped
for importers, found itself, and declared the orphan reachable. The gate
written to catch orphans made an orphan pass by mentioning it. A gate must
never be its own witness, so it now excludes its own file.
- it matched @one-shot anywhere in prose, including its own paragraph
explaining the tag, and reported "1 tagged" when nothing had been tagged.
The tag must now be a marker at line start.
Negative-tested: removing check:shell-scroll from package.json fails by name.
The scroll fix itself is NOT done. contain:layout and .vtl-host overflow-y are
confirmed in the built bundle, and a real wheel gesture still moves the grid
0px with 2586px mounted and unreachable. The gate is what caught that; the
diff looked correct.
bun run appends arguments to the command string, so a script ending in an
inline # comment swallows them. Proven side by side:
"node argtest.mjs" bun run _t1 --self-test -> ["--self-test"]
"node argtest.mjs # a comment" bun run _t2 --self-test -> []
No error, no warning: the script runs in its default mode and exits 0.
Four scripts advertised a flag that could never arrive, and two of them are
--self-test. A --self-test IS the proof that a gate can fail, so running
check:facet-cascade --self-test printed the ordinary PASS and exited 0 --
which reads as "the negative test passed" when the negative test never ran.
The gate itself was fine: invoking node directly reverted the cascade and
failed 13 assertions correctly, including the real symptom (switching to
another language family leaves the map empty because the orphan still
filters). Only the invocation was broken, so the mechanism that certifies a
gate can fail was a gate certifying itself.
Each flag now gets a dedicated entry that hardcodes it, which also makes the
mode discoverable in bun run output instead of documented only in prose
nobody greps.
check:script-flags fails any script whose comment advertises a flag the
command cannot receive. Negative-tested: adding such a comment back fails by
name with exit 1.
serve used --bind 0.0.0.0, and every browser gate starts it. A live audit of
this box found nineteen publicly bound http.server processes on 4820-4970 plus
one on 8899, all rooted at public/, with curl to the box's public IP returning
200. Nobody noticed, because a listening socket produces no output and no
failure: the gate it served passed and the process outlived it.
The standing rule "keep the box closed" lived in prose. The behaviour lived in
a flag. Prose cannot constrain a flag, so the rule was never violated by a
decision -- only by a default nobody re-read.
Loopback is sufficient for the whole suite: verify-browser-all,
check:a11y-button-names, check:nav-reachable and smoke:share-all already
default to 127.0.0.1. The wide bind bought nothing.
check:no-public-binds fails any script with an explicit all-interface bind, a
bare vite --host, or an http.server with no --bind at all (its default is also
all interfaces, just less visible in review). It reads the repo rather than the
machine's live sockets, because a socket check would pass or fail on whatever
happens to be running and could not be fixed by editing anything here.
Negative-tested: restoring --bind 0.0.0.0 fails by name with exit 1.
The menu is position:fixed with viewport coordinates. That resolves against the
viewport only while nothing establishes a containing block for fixed
descendants. VerticalTabsLayout -- the cap-5 shell that hosts every plugin as a
tab -- sets contain:layout paint on .vtl-content, inside .vtl-main{left:256px}.
Measured at 1600x900, same component, two hosts:
standalone research-dashboard ...... dLeft=0 dTop=4 six of six
inside plugin-studio dashboard tab . dLeft=256 dTop=111 every one
256 is the rail width and 111 the header height. Reported as the menu landing
"in a random spot further down to the right"; it was never random.
An earlier probe measured six triggers at four viewport widths and passed, so
the card was closed while the bug was still there. Every number was accurate.
They were answers about a surface the user never looks at -- widening across
more widths could not have found it, because the variable was the host.
Fix measures the containing block instead of assuming it: a fixed probe at 0,0
inserted in the menu's own parent reports the origin the menu resolves against.
When nothing captures, that origin is 0,0 and the standalone numbers are
unchanged bit-for-bit.
Not fixed by dropping contain from the layout: it is load-bearing there,
stopping a rail collapse from reflowing a heavy mounted map.
check:fixed-anchoring measures both hosts and treats an unmeasured host as a
failure. Negative-tested: reverting the subtraction fails the studio host with
dLeft=256 while standalone still passes -- it fails exactly where the old probe
passed.
Also cards a second symptom of the same cause, not yet fixed: fixed-position
modal scrims written inset:0 stretch to the content pane instead of the
viewport, leaving the rail clickable behind an open modal.
A card was closed on an agent measurement while the bug was still visible to
the Driver, and the card had no field in which that disagreement could live.
The board said done and the product said broken, and nothing in the data knew.
Adds fields.verification (agent|human|both) plus verification_kind
(visual|state) and a conflict flag, back-filled across 179 closed cards, and
a gate that refuses an anonymous close.
The gate does not require human sign-off. Gating the board on human review was
tried and it stalled. What it forbids is losing the distinction, so the
question "which of these did a human actually look at?" stays answerable.
The visual/state split encodes the asymmetry the Driver named: human
verification is more reliable on visual bugs, agent verification on non-visual
ones. A headless probe reads geometry, not perception -- it can measure a menu
at dTop=4 and still be blind to it feeling wrong. So a visual card closed on
agent evidence is a hypothesis; a state card closed on agent evidence is a
result. The counts are reported separately.
Negative-tested three ways: stripping a tag, claiming human verification with
no quote, and marking a conflict card done each exit 1 by name.
The export carried its provenance comment above the header, and every spreadsheet
importer treats line one as the header — so the scope sentence became the column names
and the real header became the first data row. A single-column line above a sixty-eight
column table is what "glitched" looks like on screen.
Three other explanations were plausible and all three are wrong, measured before
anything changed: the file is comma-separated, not tab-separated, so the tabs in the
report were terminal rendering rather than bytes; the escaper is genuinely RFC-4180, and
embedded commas, quotes and newlines all survive a strict parser round trip; and
injection-prefixed cells were already neutralised. Only the position was wrong, and
chasing the other three would have produced churn and no fix.
The provenance line moves to the end rather than away. It records which filters produced
the file, which is the difference between a dataset and an anecdote, and a trailing
single-column row lands as one populated cell instead of destroying the column mapping.
The field counter that read line two moved with it — left alone it would have counted
the first data row, usually the same number, so it would have looked right and been
wrong.
A new check parses the export the way a spreadsheet would, which no existing gate did:
forty-six checks and not one read the file. It extracts the shipped escaper rather than
restating the rule, so it cannot agree with itself while the product disagrees.
Negative-tested by putting the comment back on top.
A bundle can be compiled and present on disk while its profiles never reach the
manifest, and nothing noticed: the leaf floor counts index.html recursively, so one
bundle losing every leaf while another keeps extras stays above the floor; the
embed-target check iterates what the manifest lists, so an omission is invisible to it;
and a bundle count passes too, because one bundle can be swapped for another and the
totals still match. Counting is not covering. The only way to see an omission is to
compare the manifest against the scanner rather than against itself.
The first version of this check cried wolf on its first real run. It reported key-cities
as missing — discovered with four profiles, absent from the manifest, and still absent
after a full rebuild, which made it look persistent and real. It is neither: the build
config gates that bundle by name and generate-tree filters on it, so the bundle is
legitimately discovered and legitimately unpublished. The scanner answers what exists;
the config answers what ships. Asserting the first against the manifest without
consulting the second asserts an invariant that was never true.
That failure mattered more than the bug it was chasing. A gate that reports a deliberate
configuration as a defect gets switched off, and a switched-off gate protects nothing —
so the check now reads the same config the build reads, and names the gated bundles in
its passing output so the exclusion stays visible rather than becoming folklore.
A manifest entry the scanner does not return is reported as a note rather than a
failure: stale build output heals on the next full build, while a missing plugin is
silently absent from the studio for everyone reviewing the product.
Negative-tested: removing any published bundle from the manifest fails with exit 1
naming it; restoring passes; the gated bundle never trips it.
One commit shipped the requested change and a regression together: a base prefix applied
to an already-rooted URL 404'd all 29 studio tabs, and finding it took 32 minutes. The
root-cause analysis attributes the largest single share of that to there being no gate
that exercises the dev server at all — of 46 gate-ish scripts and 13 browser gates, none
did. The built manifest hands relative refs and the dev manifest hands rooted ones, so a
base-join bug is invisible to every build-time gate and fatal to every developer.
The check finishes in about four seconds against an already-running dev server, which is
the point: the full sweep costs about seventy-five, so it only ever gets run at the end
of a change, which is exactly how a regression rides along unnoticed. It activates one
tab rather than twenty-nine because a single mount already exercises the whole resolver
path — manifest fetch, script resolve, module load, define, mount. If the resolver is
wrong, the first tab fails like all of them.
It asserts the five things that have actually broken: that the root serves the studio and
not a staging page, that the element both mounts and registers, that the tab list is not
empty, that no request failed, and that the activated pane holds real content — because
an element can register and attach while its content never arrives, which is what a
missing API base looks like.
An unreachable dev server exits 2 rather than 1. Conflating "no server" with "broken
product" is how a gate earns a reputation for flakiness and gets switched off.
Negative-tested: removing the tree base makes it fail on the empty tab list, and an
unreachable URL exits 2; restoring returns to green.
Prefixing a base onto an already-rooted URL 404'd all 29 studio tabs, and no existing
gate could have seen it: the built manifest hands relative refs, so concatenation is
correct there and the assumption is never exercised. Only the dev server hands rooted
refs, so the entire build-time suite passes green while every developer's studio is
broken. A product that has to work on ten laptops cannot be verified by the build alone.
The check discovers every shell that defines a resolve(ref) rather than naming them,
because the two capability-5 shells carried byte-identical implementations and one
latent assumption therefore produced two separate outages that had to be found
separately. A third shell copied from either is now caught on its first run.
It evaluates the shipped function against a stubbed host instead of re-implementing the
rule, so the gate cannot quietly agree with itself while the product disagrees.
Negative-tested: removing the rooted-ref guard fails 4 of 8 cases and reproduces the
exact double-slash path from the outage; restoring it returns to green.
A backtick in live shadow CSS breaks the build exactly as one in a CSS comment
does, and check:template-literal only reads the comments. Proven both directions
last tick: the planted defect left that gate at pass while the build died with a
Babel stack trace pointing far from the edit.
The card concluded this was not achievable, on good evidence — three lookahead
heuristics were tried and reverted, scoring 98 and then 136 false positives on a
healthy tree, because the real terminator of a template literal is also a
backtick and no window of surrounding characters separates the two. The evidence
was right and the inference one step too strong. No heuristic works is not
nothing works, and the card named the answer in its last line: needs a tokenizer.
A parser does not guess where a literal ends; ending it is what parsing means.
Measured against the same tree the heuristics failed on: zero false positives
across 83 files. Splitting the SFC first means it sees exactly the script text
the build compiles, and using the parser the build reports errors from means its
verdict matches what actually breaks rather than approximating it.
Negative-tested with the exact case the other gate misses: a backtick planted in
live CSS passes check:template-literal and fails this one, naming the file and
the line that was edited rather than a stack trace somewhere else. That line
number is the point — the build was always the backstop, just an unhelpful one.
I excluded smoke:share from the aggregate two ticks ago and the reason was wrong.
Every clause of it was true — the script takes required arguments, and its
gallery mode does need the vite dev server — but the conclusion did not follow. I
had tried exactly two invocations, bare and --plugin-gallery, and never --url,
which the script's own header documents first. It runs fine against the static
tree: 8 of 8 pass, each Share button reopening the correct custom element with a
live canvas.
That is the same shape I logged as an anti-pattern last tick, now found in my own
work: a not-applicable verdict is more durable than a wrong bug report, because
it closes the question. It cost the suite its highest-consequence surface for two
ticks. A share URL is the one artefact a partner pastes into an email, and it
fails silently.
It runs as smoke:share-all, carrying its own eight --url flags and --browser
chromium, since Firefox headless has no WebGL on this box and the default engine
stops. The eight leaves are the set derived by grepping for ShareButton, not a
guess. The simple-map case is the discriminating one: it shares to a different
leaf than the one opened, so the deep link is really being followed rather than
an SPA fallback rendering something plausible.
Two shell traps on the way. A VAR=x prefix does not apply to $VAR on the same
line, because the shell expands the argument before the assignment scopes to the
command, so the URLs came out with an empty base; deferring through sh -c fixes
it. And a glob in a block comment closed the comment early.
It showed an empty pane reading "Choose an element from the list." The README's
headline image for the highest capability level demonstrated nothing, and every
shot gate passed it, because freshness, distinctness and element counts are all
satisfied by an empty pane. The shot now selects the first rail row before
capturing, so the picture shows what the studio is for: a nested SPA rendering
inside the shell that lists every element by capability. The file went from 167KB
to just over a megabyte, in line with its sibling.
Found by reading the image, the same way the mobile error card was found
yesterday. Reading one picture has now caught two defects that eleven gates
passed.
Add check:shot-fonts. The studio labels each row with a capability emoji, and
this machine has no emoji-capable font, so all 28 rendered as tofu boxes in the
committed PNG. The markup is right and the product is fine in a real browser —
the shot machine could not draw it — so the fix is not to strip the emoji. The
gate asks fontconfig whether this system can render the glyphs the UI actually
uses, and names the font install rather than the markup.
It is deliberately not in build:unlocked: it reports an environment condition
that is currently true here, and a gate that always fails in the build makes the
build useless. Its first version also flagged the star this repo uses as a
comment marker, which every text font renders; the range is now narrowed to the
pictographic planes.
All seven browser gates were orphans: each had a script, and not one was invoked
by anything else. They were reachable only by someone who already knew the name
to type, which is exactly the person who does not need the list. This repo has
been bitten by this before — the Gate-1 acceptance test existed, worked, and was
invoked by nothing, so it could never report anything. An orphan gate is worse
than a missing one because it looks like coverage.
The data guards already solved this; verify-all.mjs says it plainly — a guard
nobody runs is documentation. That file has no reference to playwright or to the
built tree. It is the data half. This is the browser half.
Every gate listed here was run before it was listed, following the rule the
earlier audit set: listing a program you have not executed puts unvalidated
entries in the hand-off surface. All six pass, and the aggregate now runs them
end to end: capabilities, embeds, map paint, the Gate-1 chain, accessible names,
and phone navigation.
smoke:share is excluded on purpose and says so. It takes required arguments and
its gallery mode reads a registry only the vite dev server serves, so listing it
would make the suite fail for something that is not a defect.
Preconditions are asserted up front and exit 2, so a missing server reads as a
setup problem rather than six mystery failures.
Yesterday the cap-5 shell shipped unusable on a phone: the rail was an off-canvas
drawer, the toggle lived inside a header the shell does not render, and the only
other control sat inside the drawer itself at x:-244. Every tab but the one that
happened to open was unreachable, with no error and a page that painted fine.
That survived because it is an intersection bug. Header-off is correct, and an
off-canvas mobile drawer is correct; the defect exists only where they overlap.
So this checks the intersection directly — each cap-5 shell as it ships, at a
phone viewport — and asks whether any on-screen control opens the rail.
It then presses that control and requires the rail to arrive. Presence is not
function: a present-but-inert opener is the same dead end wearing a disguise, and
asserting only presence is how the original bug survived a card marked verified.
This is not covered by the accessible-name gate. Neither property implies the
other: the off-canvas button had a perfectly good name while being unreachable,
and reachability is geometry, not markup, so no source grep can answer it.
Negative-tested by re-creating the real bug — removing the floating opener makes
it fail naming that shell, while the header-bearing studio still passes.
A cap-5 shell nests its children as iframes, and window.MAP_TOKEN does not cross
an iframe boundary — each frame is its own window. A nested map cannot inherit
the parent's token, so it works only because the child leaf carries its own baked
copy. Proven in a browser: all thirteen nested frames in the cap-5 shell hold a
real token, including a map two levels deep. That was a one-off run; this makes
it standing.
The existing smokes cannot catch this. Both inject window.MAP_TOKEN with
addInitScript before the page runs, which is right for what they test but means
the test supplies the thing under test. A user opening a shipped leaf has nothing
injecting a token for them.
Also asserts the reverse: no token may be compiled into a .js bundle. Bundles are
cached, SRI-pinned and reused across hosts, so a token baked into one cannot be
rotated without a rebuild and leaks into every embed.
Reads the built manifest rather than walking directories, so it checks the set
that actually ships. Counted 43 leaves, matching an independent filesystem count.
Negative-tested both directions: stripping a leaf's token and planting one in a
bundle each fail with the artifact named.
The desktop composite pairs the research dashboard with the grid. This is its
mobile sibling: the monolithic 7-tab research map on top, the same grid beneath,
stacked to scroll. Which one a shell shows is an aspect-ratio choice made at the
SPA level, so neither plugin knows about the other and both stay embeddable.
The grid's own search and filters are ON here. The desktop composite turns them
off because the dashboard header owns those controls and drives the grid through
props; the research map has no such header, so switching them off would leave the
grid with no way to be searched at all.
Composes its children by tag and bundle URL rather than importing their .vue
files. Importing the research map directly would bypass its bundle entry, which
is where its three bundle-private colour strategies, its profile registry, its
Pinia and i18n pair and the RTL text plugin are installed — and losing those is a
runtime defect in a green build, because a thrown setup() removes the subtree
with an empty error.
Add check:composite-srcs, which resolves every nested src against the built tree
and asserts every nested tag is registered. A wrong relative depth is a 404, and
the stack layout swallows a failed section on purpose so one broken child cannot
take the whole stack down — which means the cost of that resilience is a silently
empty band. Negative-tested both ways: breaking the depth and renaming a tag each
fail with the file and path named; restoring returns to pass.
