Files
proxy/nodejs/routes/docs.js
wmantly 6ede072213 security: swap sanitizer to xss for Jest-compatible SSR output
- Replace isomorphic-dompurify with xss to avoid ESM-only jsdom/htmlparser2
  transitive deps that break the existing Jest test suite.
- Keep rendered docs HTML sanitized via xss() in routes/docs.js.

Co-Authored-By: Claude <noreply@anthropic.com>
2026-07-18 23:02:11 -04:00

152 lines
6.7 KiB
JavaScript

'use strict';
const fs = require('fs');
const path = require('path');
const router = require('express').Router();
const {rateLimit} = require('express-rate-limit');
const {marked} = require('marked');
const xss = require('xss');
const conf = require('@simpleworkjs/conf');
const buildInfo = require('../utils/build_info');
// Public, unauthenticated, and reads from disk on every request -- throttle
// per IP so it can't be used to hammer the filesystem (mirrors the pattern
// in routes/auth.js/routes/host.js), generous since this is just docs.
const docsLimiter = rateLimit({
windowMs: 60 * 1000,
max: 120,
standardHeaders: true,
legacyHeaders: false,
message: {name: 'TooManyRequests', message: 'Too many requests, please try again later.'},
});
const values = {
title: conf.environment !== 'production' ? `dev` : '',
titleIcon: conf.environment !== 'production' ? `<i class="fa-brands fa-dev"></i>` : '',
name: conf.name,
logo: conf.logo,
...buildInfo,
};
// Full local copy of the project's documentation, rendered server-side --
// so an operator running air-gapped (no route to GitHub Pages, where this
// content otherwise only lives) can still read it from the running app.
// An explicit slug -> file allowlist, never a user-suppliable path, so
// there's no way to make this read outside the doc set below.
const DOCS = {
// Plain-language "what is this and why would I use it" guides -- linked
// directly from the relevant card in the UI (see the help icon on each
// card). Each links onward to the deeper technical doc below for readers
// who want the system-design/protocol-level detail.
hosts: {title: 'Hosts & HTTPS', file: path.join(__dirname, '../../docs/concepts-hosts.md')},
dns: {title: 'DNS Providers', file: path.join(__dirname, '../../docs/concepts-dns.md')},
access: {title: 'Users, Groups & Permissions', file: path.join(__dirname, '../../docs/concepts-access.md')},
'api-tokens': {title: 'API Tokens', file: path.join(__dirname, '../../docs/concepts-api-tokens.md')},
overview: {title: 'Overview', file: path.join(__dirname, '../../README.md')},
changelog: {title: 'Changelog', file: path.join(__dirname, '../../CHANGELOG.md')},
deployment: {title: 'Deployment', file: path.join(__dirname, '../../DEPLOYMENT.md')},
api: {title: 'API Reference', file: path.join(__dirname, '../api.md')},
installation: {title: 'Installation', file: path.join(__dirname, '../../docs/installation.md')},
architecture: {title: 'Architecture', file: path.join(__dirname, '../../docs/architecture.md')},
docker: {title: 'Docker', file: path.join(__dirname, '../../docs/docker.md')},
contributing: {title: 'Contributing', file: path.join(__dirname, '../../docs/contributing.md')},
};
const docList = Object.entries(DOCS).map(([slug, d]) => ({slug, title: d.title}));
// README.md links its screenshots as repo-relative "docs/images/...", which
// only resolves correctly on GitHub. Serve that same folder here and rewrite
// the rendered markup to point at it absolutely, so the images work when
// read from /docs/overview too.
router.use('/images', require('express').static(path.join(__dirname, '../../docs/images')));
function fixImagePaths(html) {
return html.replace(/(["(])docs\/images\//g, '$1/docs/images/');
}
// docs/*.md files (not the repo-root README/CHANGELOG/api.md) carry Jekyll
// front matter for the GitHub Pages build and a "← Back to Home" link back
// to that site's index -- both meaningless here (this viewer has its own
// doc-list sidebar, docs_page.ejs) and, worse, marked() doesn't know front
// matter isn't regular markdown: it rendered as a garbled heading + stray
// <hr> at the top of every page. Strip both before rendering.
function stripJekyllCruft(content) {
return content
.replace(/^---\n[\s\S]*?\n---\n/, '')
.replace(/^\s*\[← Back to Home\]\([^)]*\)\s*\n/m, '');
}
// Docs cross-link each other as "<slug>.html" (correct for the Jekyll/GitHub
// Pages build, which is what these same .md files also feed) and
// "index.html" for the docs home -- neither resolves here, where a doc lives
// at /docs/<slug> with no .html suffix. Rewrite known doc links to the
// in-app route, same idea as fixImagePaths() above. Only touches slugs that
// actually exist, so an unrelated "foo.html" link is left alone.
// Docs are also linked by their real filename stem (e.g. "concepts-hosts.html"
// for docs/concepts-hosts.md) -- the correct, working link on the Jekyll/
// GitHub Pages build, where the URL IS the filename stem. That doesn't match
// this viewer's own short slugs (DOCS keys, e.g. "hosts"), so also resolve by
// filename as a fallback -- one link written in a doc works correctly on
// both targets, rather than needing two different link forms.
const slugByFilename = Object.fromEntries(
Object.entries(DOCS).map(([slug, d]) => [path.basename(d.file, '.md'), slug])
);
function fixDocLinks(html) {
return html
.replace(/href="index\.html"/g, 'href="/docs"')
.replace(/href="([a-z0-9-]+)\.html"/g, (match, name) => {
const slug = DOCS[name] ? name : slugByFilename[name];
return slug ? `href="/docs/${slug}"` : match;
});
}
router.use(docsLimiter);
router.get('/', function(req, res) {
res.render('docs_index', {...values, docs: docList});
});
// Plain, dependency-free line-substring search over the same allowlisted
// doc set -- no separate index to build/maintain, no new dependency, and it
// keeps working with no internet access (same reasoning as the rest of this
// route). Must be registered before the /:slug catch-all below, or "search"
// would be treated as a (nonexistent) doc slug and 404.
router.get('/search', function(req, res) {
const q = (req.query.q || '').trim();
if (!q) return res.json({results: []});
const qLower = q.toLowerCase();
const results = [];
for (const [slug, doc] of Object.entries(DOCS)) {
try {
const content = stripJekyllCruft(fs.readFileSync(doc.file, 'utf8'));
const matchLine = content.split('\n').find(line => line.toLowerCase().includes(qLower));
if (matchLine) {
results.push({slug, title: doc.title, snippet: matchLine.trim().slice(0, 200)});
}
} catch (error) { /* unreadable doc file -- skip it */ }
}
res.json({results});
});
router.get('/:slug', function(req, res, next) {
const doc = DOCS[req.params.slug];
if (!doc) return next({status: 404, message: 'Doc not found'});
try {
const content = stripJekyllCruft(fs.readFileSync(doc.file, 'utf8'));
res.render('docs_page', {
...values,
docs: docList,
currentSlug: req.params.slug,
docTitle: doc.title,
docHtml: xss(fixDocLinks(fixImagePaths(marked(content)))),
});
} catch (error) {
next(error);
}
});
module.exports = router;