Compare commits

...

1 Commits

Author SHA1 Message Date
ae4d211c11 test: drive the Settings screen in a browser and guard every popup element id (closes #229)
All checks were successful
check / check (push) Successful in 29s
Nothing exercised the Settings view in a browser, and jest runs in the
node environment with no DOM, so the densest run of $("...") lookups in
the codebase was unverified at runtime. A wrong id is valid JavaScript
naming a defined function: $() returns null and the next property access
throws, which inside a view's init() aborts the rest of the popup's
init() and leaves every screen blank.

Two halves, because they catch different things.

The e2e suite (tests/e2e/run.js) gains six cases between the address
removal and dust threshold sections. They assert the About well and the
wallet list were actually written — show() populates those last, so
reading them back proves the whole of show() ran rather than just enough
of it to unhide the section — that the four Token Spam Protection
controls are real input[type=checkbox] elements defaulted on, and that
the theme and network selectors offer exactly the choices
src/shared/networks.js and index.html define while carrying the
persisted value. One filter is then toggled off and back on across a
popup reopen each way, which runs the change handler, saveState(),
loadState() and the init() assignment rather than only looking at the
screen. Each group records a coverage key and a final case demands the
exact set, so a section that silently stopped running reddens the suite
instead of shrinking it.

tests/popupElementIds.test.js is the general half and needs no browser,
so jest picks it up and it runs in make check: every literal id reached
through $(), document.getElementById(), showError()/hideError() and
showView() must exist in src/popup/index.html, no id in index.html may
be defined twice, and the scan asserts it found the code and the markup
so it cannot pass by covering nothing. Only literal arguments are
resolvable statically; $(containerId) and a lookup naming the wrong
existing element are the browser suites' job, and README says so.

Demonstrated against three deliberate breaks. A typo'd id in
settings.js reddens both halves, the e2e run reporting
"pageerror: Cannot set properties of null (setting 'checked')" against
its first test. A handler bound to the wrong but existing element passes
the static guard and reddens only the new functional case. A typo in a
view no browser suite opens reddens only the static guard.
2026-08-17 06:21:27 +00:00
4 changed files with 477 additions and 0 deletions

View File

@@ -321,6 +321,24 @@ pick it up either. Neither is wired into the Gitea workflow yet —
docker-in-docker in CI is a separate question. Run them locally before changing docker-in-docker in CI is a separate question. Run them locally before changing
anything under `src/popup/views/`. anything under `src/popup/views/`.
### Element id guard (part of `make check`)
`tests/popupElementIds.test.js` asserts statically that every element id the
popup looks up — `$("...")`, `document.getElementById("...")`,
`showError()`/`hideError()`, and the `view-<name>` a literal `showView("...")`
resolves to — exists in `src/popup/index.html`, and that `index.html` defines no
id twice. A wrong id is valid JavaScript naming a defined function, so neither
jest (node environment, no DOM) nor a linter objects to it; at runtime `$()`
returns `null` and the next property access throws, which inside a view's
`init()` aborts the rest of `src/popup/index.js` `init()` and leaves the popup
blank.
It runs with no browser, so unlike the e2e suites it fits inside `make check`,
and it covers every view rather than the ones some test happens to open. It only
sees literal arguments: a call like `$(containerId)` is invisible to it, and a
lookup naming the wrong existing element is valid by construction. Both of those
are the browser suites' job.
## Rationale ## Rationale
Common popular EVM wallets have become bloated with swap UIs, portfolio Common popular EVM wallets have become bloated with swap UIs, portfolio

21
TODO.md
View File

@@ -45,6 +45,27 @@ undefined identifiers, which is how
# Completed Steps # Completed Steps
- 2026-08-17: The Settings screen is driven in a browser, and every element id
the popup looks up is checked statically. Nothing exercised Settings in the
e2e suite, and jest runs with no DOM, so the densest run of `$("...")` lookups
in the codebase was unverified at runtime. Six new cases in `tests/e2e/run.js`
reach Settings, assert the About well and the wallet list were actually
written, assert the four Token Spam Protection checkboxes are real checkboxes
defaulted on, assert the theme and network selectors offer the choices
`src/shared/networks.js` defines and carry the persisted value, and toggle one
filter off and back on across a popup reopen each way — which runs the change
handler, `saveState()`, `loadState()` and the `init()` assignment rather than
just looking at the screen. Each group records a coverage key and a final case
demands the exact set, so a shortened or skipped section reddens the run
instead of shrinking it. `tests/popupElementIds.test.js` is the general half
and runs in `make check` with no browser: every literal id reached through
`$()`, `document.getElementById()`, `showError()`/`hideError()` and
`showView()` must exist in `src/popup/index.html`, which no id in `index.html`
may define twice. Demonstrated on three deliberate breaks — a typo'd id (both
halves red), a handler bound to the wrong but existing element (only the
functional e2e case red), and a typo in a view no browser suite opens (only
the static guard red)
([#229](https://git.eeqj.de/sneak/AutistMask/issues/229)).
- 2026-08-17: `README.md` no longer advertises a defect the wallet does not - 2026-08-17: `README.md` no longer advertises a defect the wallet does not
have. The End-to-End Tests section listed the EIP-1193 code being dropped in have. The End-to-End Tests section listed the EIP-1193 code being dropped in
the last hop into the page as a standing limit of the dApp coverage; that the last hop into the page as a standing limit of the dApp coverage; that

View File

@@ -42,6 +42,7 @@ const {
STUB_TX_HASH, STUB_TX_HASH,
} = require("./network"); } = require("./network");
const { DUST_THRESHOLD_MESSAGE } = require("../../src/popup/dustThreshold"); const { DUST_THRESHOLD_MESSAGE } = require("../../src/popup/dustThreshold");
const { NETWORKS } = require("../../src/shared/networks");
const TEST_TIMEOUT_MS = 120000; const TEST_TIMEOUT_MS = 120000;
@@ -940,6 +941,256 @@ test("confirming removes the address and returns Home (#162)", async (env) => {
); );
}); });
// ------------------------------------------------- Settings screen (#229)
// Settings holds the densest run of $("...") lookups in the codebase, and
// until this section nothing drove it in a browser. One wrong id makes
// settings.init() throw, which aborts the rest of index.js init() before it
// renders anything at all — so a broken id does not degrade Settings, it
// leaves the whole popup blank. These tests assert the controls are there
// AND that they work, because "the view is visible" would still pass
// against a screen whose handlers were never wired.
// The four Token Spam Protection checkboxes, in markup order, with the
// src/shared/state.js key each one is bound to. All four default true.
const SPAM_FILTER_CHECKBOXES = [
{ id: "settings-hide-spoofed-symbols", key: "hideSpoofedSymbols" },
{ id: "settings-hide-low-holders", key: "hideLowHolderTokens" },
{ id: "settings-hide-fraud-contracts", key: "hideFraudContracts" },
{ id: "settings-hide-dust", key: "hideDustTransactions" },
];
// The one toggled through a reopen. Chosen because nothing later in this
// suite depends on it: the other three filter token and transaction lists
// that the ConfirmTx and dApp sections go on to drive.
const TOGGLED_FILTER = "settings-hide-dust";
// Everything the Settings assertions below must observe, recorded as each
// group of them completes. The final test demands the exact set.
//
// The point is that a green run cannot mean the assertions were skipped.
// Navigation that silently fails already fails a test — visible() throws
// on a timeout — but an early return, a deleted test, or a body that
// stopped being reached would otherwise shrink this section quietly
// instead of reddening the run.
const SETTINGS_COVERAGE = [
"about-well",
"spam-checkbox-defaults",
"theme-select",
"network-select",
"toggle-off-survives-reopen",
"toggle-on-survives-reopen",
"wallet-list",
];
// A control read as the DOM has it, not as a selector claims: tag name and
// type distinguish a real <input type="checkbox"> from a <div> that merely
// carries the id, and `checked` is the live property rather than the
// attribute, so it reflects what init() assigned.
function controlState(page, id) {
return page.evaluate((elementId) => {
const el = document.getElementById(elementId);
if (!el) return null;
return {
tag: el.tagName.toLowerCase(),
type: el.type || "",
checked: el.checked,
value: el.value,
options: Array.from(el.options || []).map((o) => o.value),
};
}, id);
}
async function checkboxStates(page) {
const out = {};
for (const { id } of SPAM_FILTER_CHECKBOXES) {
out[id] = await controlState(page, id);
}
return out;
}
function assertSpamCheckbox(st, id, expected, where) {
assert(st !== null, "no element with id " + id + " on Settings " + where);
assert(
st.tag === "input" && st.type === "checkbox",
id + " is a <" + st.tag + " type=" + st.type + ">, not a checkbox",
);
assert(
st.checked === expected,
id +
" reads " +
st.checked +
" " +
where +
", expected " +
expected +
" — the checkbox is on screen but not carrying the persisted value",
);
}
test("Settings renders with the whole screen populated (#229)", async (env) => {
await visible(env.page, "#view-main");
await openSettings(env.page);
// show() writes the About well last thing before showView(), so an id
// it cannot find aborts before Settings is ever displayed. Reading the
// values back proves the whole of show() ran, not just enough of it to
// unhide the section. These are filled from build-time constants that
// always have a value, so empty means the write did not happen.
const about = await env.page.evaluate(() => {
const out = {};
for (const id of [
"about-license",
"about-author",
"about-version",
"about-release-date",
"about-commit-link",
]) {
const el = document.getElementById(id);
out[id] = el === null ? null : el.textContent.trim();
}
return out;
});
for (const [id, text] of Object.entries(about)) {
assert(
text !== null && text.length > 0,
"the About well left #" +
id +
" unwritten: " +
JSON.stringify(about),
);
}
env.settingsCoverage.add("about-well");
// The wallet list is rendered by settings.js rather than authored in
// index.html, so an empty container means renderWalletListSettings()
// did not run even though the screen came up.
const wallets = await env.page
.locator("#settings-wallet-list .settings-wallet-name")
.count();
assert(
wallets >= 2,
"Settings lists " +
wallets +
" wallets; the suite created two by this point",
);
env.settingsCoverage.add("wallet-list");
});
test("the four Token Spam Protection checkboxes render, defaulted on (#229)", async (env) => {
await openSettings(env.page);
const states = await checkboxStates(env.page);
for (const { id } of SPAM_FILTER_CHECKBOXES) {
assertSpamCheckbox(states[id], id, true, "on first render");
}
env.settingsCoverage.add("spam-checkbox-defaults");
});
test("the theme and network selectors render their real choices (#229)", async (env) => {
await openSettings(env.page);
const theme = await controlState(env.page, "settings-theme");
assert(theme !== null, "no #settings-theme element on Settings");
assert(
theme.tag === "select",
"#settings-theme is a <" + theme.tag + ">, not a <select>",
);
assert(
theme.options.join(",") === "system,light,dark",
"the theme selector offers " + JSON.stringify(theme.options),
);
assert(
theme.value === "system",
"the theme selector shows " +
JSON.stringify(theme.value) +
", expected the persisted default 'system'",
);
env.settingsCoverage.add("theme-select");
const network = await controlState(env.page, "settings-network");
assert(network !== null, "no #settings-network element on Settings");
assert(
network.tag === "select",
"#settings-network is a <" + network.tag + ">, not a <select>",
);
const wantNetworks = Object.keys(NETWORKS).sort().join(",");
assert(
network.options.slice().sort().join(",") === wantNetworks,
"the network selector offers " +
JSON.stringify(network.options) +
", expected the networks in src/shared/networks.js: " +
wantNetworks,
);
assert(
network.value === "mainnet",
"the network selector shows " +
JSON.stringify(network.value) +
", expected the persisted default 'mainnet'",
);
env.settingsCoverage.add("network-select");
});
// The functional half. A checkbox that renders but is not wired looks
// identical on screen; only a value that survives being written to storage
// and read back by a fresh page load tells the two apart. That round trip
// runs through the change handler, saveState(), loadState() and the
// assignment init() makes — every part of the wiring at once.
test("a spam filter toggled in Settings survives a popup reopen (#229)", async (env) => {
await openSettings(env.page);
await env.page.click("#" + TOGGLED_FILTER);
const immediately = await controlState(env.page, TOGGLED_FILTER);
assert(
immediately.checked === false,
"clicking #" + TOGGLED_FILTER + " did not clear it",
);
await reopenPopup(env, "#view-settings");
const after = await checkboxStates(env.page);
for (const { id } of SPAM_FILTER_CHECKBOXES) {
assertSpamCheckbox(
after[id],
id,
id !== TOGGLED_FILTER,
"after reopening the popup",
);
}
env.settingsCoverage.add("toggle-off-survives-reopen");
});
test("turning the same filter back on survives a reopen too (#229)", async (env) => {
await openSettings(env.page);
await env.page.click("#" + TOGGLED_FILTER);
await reopenPopup(env, "#view-settings");
// Restores the fixture the later sections inherit, and rules out a
// checkbox that persists "off" only because it is stuck there.
const after = await checkboxStates(env.page);
for (const { id } of SPAM_FILTER_CHECKBOXES) {
assertSpamCheckbox(after[id], id, true, "after toggling back on");
}
env.settingsCoverage.add("toggle-on-survives-reopen");
await env.page.click("#btn-settings-back");
await visible(env.page, "#view-main");
});
test("the Settings assertions above all ran (#229)", async (env) => {
const seen = [...env.settingsCoverage].sort();
const want = SETTINGS_COVERAGE.slice().sort();
assert(
seen.join(",") === want.join(","),
"the Settings section covered " +
JSON.stringify(seen) +
" but must cover " +
JSON.stringify(want) +
" — a green run here would otherwise mean only that fewer " +
"assertions ran, not that they passed",
);
});
// ------------------------------------------------ dust threshold (#233) // ------------------------------------------------ dust threshold (#233)
// The popup size README documents the UI as designed for. Pages in this // The popup size README documents the UI as designed for. Pages in this
@@ -2668,6 +2919,11 @@ async function main() {
// The recovery phrase of the wallet created in test 2, so later // The recovery phrase of the wallet created in test 2, so later
// tests can assert on the real secret rather than its shape. // tests can assert on the real secret rather than its shape.
phrase: null, phrase: null,
// What the Settings section (#229) actually observed. A guard test
// at the end of that section demands the full set, so a skipped or
// silently shortened assertion reddens the run instead of shrinking
// it.
settingsCoverage: new Set(),
// Confirmation-screen heights, measured in the pending state and // Confirmation-screen heights, measured in the pending state and
// compared against every later state of the same screen. // compared against every later state of the same screen.
ethPendingHeight: null, ethPendingHeight: null,

View File

@@ -0,0 +1,182 @@
// Every element id the popup views look up must exist in the markup they
// look it up in.
//
// The failure this catches: `$("settings-hide-dsut")` is valid JavaScript
// referring to a defined function, so neither jest (node environment, no
// DOM) nor a linter has anything to object to. At runtime `$()` returns
// null and the next property access throws, which in `init()` aborts the
// rest of that view's wiring and takes the whole screen down. Settings is
// the densest concentration of these lookups in the codebase.
//
// This is the cheap general half of the guard: it runs in `make check`
// with no browser and covers every id in every view, not the ones some
// test happens to click. The expensive specific half is the Settings
// section of the end-to-end suite (tests/e2e/run.js), which proves the
// screen actually comes up and its controls work.
//
// Scope and limits, stated rather than implied:
// - Only literal string arguments are resolvable statically. A call
// like `$(containerId)` is invisible here; those are covered by the
// e2e run instead.
// - `document.getElementById()` is checked too, minus the ids listed in
// RUNTIME_CREATED_IDS, which name nodes the code creates itself and
// which are legitimately absent from the static markup.
"use strict";
const fs = require("fs");
const path = require("path");
const POPUP_DIR = path.join(__dirname, "..", "src", "popup");
const POPUP_HTML_PATH = path.join(POPUP_DIR, "index.html");
// Nodes built at runtime rather than authored in index.html. Each one must
// be created unconditionally by the code before it is ever looked up.
const RUNTIME_CREATED_IDS = new Set([
// Created by updateDebugBanner() in src/popup/views/helpers.js.
"debug-banner",
]);
// Every id lookup the popup performs with a literal argument, as
// {id, file, line, source} records.
//
// showView("x") is included because it resolves to the element id
// "view-x": a view name with no matching section is the same defect one
// indirection further out.
const PATTERNS = [
{ re: /\$\(\s*"([^"\n]+)"\s*\)/g, id: (m) => m[1], source: "$()" },
{
re: /document\.getElementById\(\s*"([^"\n]+)"\s*\)/g,
id: (m) => m[1],
source: "getElementById()",
},
{
re: /\b(?:showError|hideError)\(\s*"([^"\n]+)"/g,
id: (m) => m[1],
source: "showError()/hideError()",
},
{
re: /\bshowView\(\s*"([^"\n]+)"\s*\)/g,
id: (m) => "view-" + m[1],
source: "showView()",
},
];
function jsFilesUnder(dir) {
const out = [];
for (const entry of fs.readdirSync(dir, { withFileTypes: true })) {
const full = path.join(dir, entry.name);
if (entry.isDirectory()) {
out.push(...jsFilesUnder(full));
} else if (entry.name.endsWith(".js")) {
out.push(full);
}
}
return out.sort();
}
function lineOf(text, index) {
return text.slice(0, index).split("\n").length;
}
function collectReferences() {
const refs = [];
for (const file of jsFilesUnder(POPUP_DIR)) {
const text = fs.readFileSync(file, "utf8");
const rel = path.relative(path.join(__dirname, ".."), file);
for (const { re, id, source } of PATTERNS) {
re.lastIndex = 0;
let m;
while ((m = re.exec(text)) !== null) {
refs.push({
id: id(m),
file: rel,
line: lineOf(text, m.index),
source,
});
}
}
}
return refs;
}
function collectHtmlIds(html) {
const ids = [];
const re = /\bid="([^"]+)"/g;
let m;
while ((m = re.exec(html)) !== null) ids.push(m[1]);
return ids;
}
const HTML = fs.readFileSync(POPUP_HTML_PATH, "utf8");
const HTML_IDS = collectHtmlIds(HTML);
const HTML_ID_SET = new Set(HTML_IDS);
const REFERENCES = collectReferences();
describe("every element id the popup looks up exists in its markup", () => {
// A guard that found nothing to check would pass forever. If a
// refactor renames the directory, changes the helper, or moves the
// markup, this fails instead of quietly covering zero call sites.
// The floors are far below the counts measured when this was written
// (434 lookups across 20 of the 24 files under src/popup/, against 274
// ids in the markup), so ordinary churn does not trip them.
test("the scan actually found the code and the markup", () => {
const files = new Set(REFERENCES.map((r) => r.file));
expect(files.size).toBeGreaterThanOrEqual(15);
expect(REFERENCES.length).toBeGreaterThanOrEqual(300);
expect(HTML_IDS.length).toBeGreaterThanOrEqual(200);
// The densest screen, named explicitly: a scan that stopped
// covering src/popup/views/settings.js is the exact regression
// this file was written for.
expect(
files.has(path.join("src", "popup", "views", "settings.js")),
).toBe(true);
expect(
REFERENCES.some((r) => r.id === "settings-hide-spoofed-symbols"),
).toBe(true);
expect(REFERENCES.some((r) => r.id === "view-settings")).toBe(true);
});
test("no lookup names an id that src/popup/index.html does not define", () => {
const missing = REFERENCES.filter(
(r) => !HTML_ID_SET.has(r.id) && !RUNTIME_CREATED_IDS.has(r.id),
).map(
(r) =>
r.file +
":" +
r.line +
" " +
r.source +
' looks up id "' +
r.id +
'", which is not in src/popup/index.html',
);
expect(missing).toEqual([]);
});
test("every id excused as runtime-created is still looked up somewhere", () => {
// Otherwise the exception list becomes a place stale names
// accumulate, and the next real miss can be waved through by
// adding one more.
for (const id of RUNTIME_CREATED_IDS) {
expect(REFERENCES.some((r) => r.id === id)).toBe(true);
expect(HTML_ID_SET.has(id)).toBe(false);
}
});
test("index.html defines no id twice", () => {
// getElementById returns the first match, so a duplicate id means
// one of the two elements can never be reached by the code that
// thinks it owns it.
const seen = new Set();
const duplicated = [];
for (const id of HTML_IDS) {
if (seen.has(id)) duplicated.push(id);
seen.add(id);
}
expect(duplicated).toEqual([]);
});
});