Files
bookmarkmagic/scripts/sanitize/html.mjs
T
poli0981andClaude Opus 5.5 434c05b85f fix(review): credit everything the build compiles in, close three sanitizer leaks, sync the spec
From the v1.1.0 adversarial review (security) and the docs pass:

- clsx (MIT) ships too — the Svelte runtime imports it — and wxt's browser
  module as well as @wxt-dev/browser. All four are now credited in About,
  PRIVACY, README and both notice files, and wxt.config.ts fails the build when
  code from any uncredited package would ship; tests/unit/third-party.test.ts
  keeps the notices in step with that list.
- Sanitizer, CSV: default-deny. A title with an unquoted comma — the row that
  makes the extension report CSV_ROW_MISMATCH — shifted a file: path and free
  text into columns no rule rewrote, and the host-only check passed them.
- Sanitizer, HTML: unknown attributes (Firefox POST_DATA, any DATA-*) are
  rewritten; only dates, root markers and the charset line are kept.
- Sanitizer: a value-level leak check over CSV fields, attribute values and
  JSON strings, and a refusal that prints counts only — it used to print the
  leaked words and ask for them in a public issue.
- BM_ENV_ERROR_CODES joins the per-locale message check, with the four fixed
  keys describeError maps; the popup no longer logs crashes nothing can read.
- docs 00/02/03/05/06/07/08/09/11/13 describe v1.1.0 (diagnostics moved to
  09 §7; 13 §3 points at store/listing.* instead of a drifting copy); tokens
  and gen-icons comments corrected.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
2026-10-29 03:22:56 +07:00

161 lines
5.7 KiB
JavaScript
Raw Blame History

This file contains ambiguous Unicode characters
This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
// @ts-check
/**
* Netscape HTML: rewrite addresses, titles, descriptions and user-authored
* attributes by span, leaving every other byte exactly as the browser wrote it.
*/
import { isAddressLike } from './mappers.mjs';
import {
ANY_ATTR,
attrRewriter,
attrSpans,
replaceSpans,
spansFrom,
TAG_INNARDS,
} from './spans.mjs';
/** @typedef {import('./spans.mjs').Span} Span */
/** @typedef {import('./mappers.mjs').Mapper} Mapper */
/** 1×1 transparent PNG. Keeps ICON= present, which is what the parser reacts to. */
const BLANK_ICON =
'data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAAEAAAABCAYAAAAfFcSJAAAAC0lEQVR42mNkYAAAAAYAAjCB0C8AAAAASUVORK5CYII=';
/** Attributes handled explicitly; the generic sweep must not touch them. */
const HANDLED_ATTRS = new Set(['href', 'icon', 'shortcuturl', 'tags']);
/**
* Attributes whose values are structure — dates, root markers, the charset
* line — and are kept as written. Exported for the leak check.
*/
export const STRUCTURAL_ATTRS = new Set([
'add_date',
'last_modified',
'last_visit',
'personal_toolbar_folder',
'unfiled_bookmarks_folder',
'mobile_bookmarks_folder',
'last_charset',
'folded',
'http-equiv',
'content',
'charset',
]);
/**
* Every OTHER attribute, default-deny: an address is mapped as one, a number
* or boolean is kept, anything else becomes `attrN`. The old sweep touched
* only address-shaped values, so Firefox's `POST_DATA="…&user=jane.doe"` on a
* keyword bookmark — and any exporter's `DATA-NOTE` — went out verbatim.
* `file:` counts as an address: `ICON_URI="file:///C:/Users/<name>/…"`.
*
* @param {string} text
* @param {Mapper} mapUrl
* @returns {Span[]}
*/
function otherAttrSpans(text, mapUrl) {
/** @type {Span[]} */
const spans = [];
const rewrite = attrRewriter(mapUrl);
let others = 0;
for (const m of text.matchAll(ANY_ATTR)) {
const name = (m[1] ?? '').toLowerCase();
const value = m[2] ?? m[3] ?? m[4] ?? '';
if (HANDLED_ATTRS.has(name) || STRUCTURAL_ATTRS.has(name) || value === '') continue;
if (/^(?:-?\d+(?:\.\d+)?|true|false)$/i.test(value)) continue;
const at = m.indices?.[2] ?? m.indices?.[3] ?? m.indices?.[4];
if (at === undefined) continue;
const mapped = isAddressLike(value) ? rewrite(value) : `attr${++others}`;
spans.push({ start: at[0], end: at[1], value: mapped });
}
return spans;
}
/**
* Every attribute value outside the structural set — what the leak check reads.
*
* @param {string} html
* @returns {string[]}
*/
export function attributeValues(html) {
/** @type {string[]} */
const values = [];
for (const m of html.matchAll(ANY_ATTR)) {
if (STRUCTURAL_ATTRS.has((m[1] ?? '').toLowerCase())) continue;
values.push(m[2] ?? m[3] ?? m[4] ?? '');
}
return values;
}
/** The next element that ends a description: `<DT>`, `<DD>`, `<DL>` or a close. */
const DESCRIPTION_END = /<\s*(?:\/\s*)?(?:dt|dd|dl)\b/gi;
/**
* `<DD>` descriptions, found in one linear pass.
*
* The description runs to the next `<DT>/<DD>/<DL>` (or a closing one), not to
* the end of the line — a wrapped description used to leak its second line —
* and trailing whitespace stays outside the span, because the newline and
* indentation before the next tag are byte-level shape this tool keeps.
*
* Replaces `/<dd>([\s\S]*?)(?=\s*<…>|\s*$)/`, which backtracked polynomially
* (SonarCloud S8786) and matched only a bare `<dd>` — `<DD class=n>Private
* note` passed through untouched.
*
* @param {string} text
* @param {() => string} next
* @returns {Span[]}
*/
function descriptionSpans(text, next) {
/** @type {Span[]} */
const spans = [];
for (const open of text.matchAll(/<dd\b[^>]*>/gi)) {
const start = (open.index ?? 0) + open[0].length;
DESCRIPTION_END.lastIndex = start;
const close = DESCRIPTION_END.exec(text);
let end = close === null ? text.length : close.index;
while (end > start && /\s/.test(text.charAt(end - 1))) end--;
if (end > start) spans.push({ start, end, value: next() });
}
return spans;
}
/** A document heading that says nothing about its owner stays as it is. */
const GENERIC_HEADING = /^\s*bookmarks(?:\s+menu)?\s*$/i;
/**
* @param {string} text
* @param {Mapper} mapUrl
* @param {Mapper} mapTitle
*/
export function sanitizeHtml(text, mapUrl, mapTitle) {
let descriptions = 0;
let keywords = 0;
/** @param {string} heading */
const mapHeading = (heading) => (GENERIC_HEADING.test(heading) ? heading : mapTitle(heading));
const spans = [
// HREF and ICON, at any quoting — attribute name case is left as found.
...attrSpans(text, /^href$/, attrRewriter(mapUrl)),
...attrSpans(text, /^icon$/, () => BLANK_ICON),
...otherAttrSpans(text, mapUrl),
// Firefox keywords and tags are user-authored text, so they are personal
// too. Tag count is preserved because it is the only part a parser sees.
...attrSpans(text, /^shortcuturl$/, () => `kw${++keywords}`),
...attrSpans(text, /^tags$/, (v) =>
v
.split(',')
.map((_, i) => `tag${i + 1}`)
.join(','),
),
// Bookmark and folder titles. `[\s\S]*?`, not `[^<]*`: a title holding a
// raw `<` made the whole match fail and stay in the output.
...spansFrom(text, new RegExp(`<a\\b${TAG_INNARDS}>([\\s\\S]*?)</a>`, 'gi'), mapTitle),
...spansFrom(text, new RegExp(`<h3\\b${TAG_INNARDS}>([\\s\\S]*?)</h3>`, 'gi'), mapTitle),
// `<TITLE>` / `<H1>` — usually "Bookmarks", but an exporter can write the
// owner's name there ("Jane Doe's bookmarks").
...spansFrom(text, /<title\b[^>]*>([^<]*)<\/title>/gi, mapHeading),
...spansFrom(text, /<h1\b[^>]*>([^<]*)<\/h1>/gi, mapHeading),
...descriptionSpans(text, () => `Description ${++descriptions}`),
];
return replaceSpans(text, spans);
}