www/check.js

import { execFileSync } from 'node:child_process'
import { readdir, readFile, stat } from 'node:fs/promises'
import { join } from 'node:path'

const generatedRoot = '.generated'

const walk = async (directory) => {
  const files = []
  for (const entry of await readdir(directory, { withFileTypes: true })) {
    const path = join(directory, entry.name)
    if (entry.isDirectory()) files.push(...(await walk(path)))
    else files.push(path)
  }
  return files
}

const exists = async (path) => {
  try {
    await stat(path)
    return true
  } catch {
    return false
  }
}

const errors = []
const htmlFiles = (await walk(generatedRoot)).filter((file) => file.endsWith('.html'))
const noindexFiles = new Set()
const clientSource = await readFile('src/main.js', 'utf8')
const builtInPrismLanguages = new Set([
  'markup',
  'html',
  'xml',
  'svg',
  'css',
  'clike',
  'javascript',
  'js',
  'plain',
  'plaintext',
  'text',
  'txt',
])
const prismLanguageAliases = { sh: 'bash', shell: 'bash' }

for (const file of htmlFiles) {
  const html = await readFile(file, 'utf8')
  const ids = [...html.matchAll(/\sid="([^"]+)"/g)].map((match) => match[1])
  const duplicateIds = ids.filter((id, index) => ids.indexOf(id) !== index)
  if (duplicateIds.length > 0) errors.push(`${file}: duplicate IDs: ${[...new Set(duplicateIds)].join(', ')}`)
  if (!html.includes('<meta name="viewport"')) errors.push(`${file}: missing viewport metadata`)
  if (html.includes('<heading')) errors.push(`${file}: contains a nonstandard <heading> element`)
  if (html.includes('name="robots"') && html.includes('noindex')) {
    noindexFiles.add(file)
    errors.push(`${file}: draft page must not be in production output`)
  }

  for (const [, fragment] of html.matchAll(/href="#([^"]+)"/g)) {
    if (!ids.includes(fragment)) errors.push(`${file}: missing fragment target #${fragment}`)
  }

  for (const [, href] of html.matchAll(/href="(\/[^"]*)"/g)) {
    if (href === '/') continue
    const pathname = href.split(/[?#]/, 1)[0]
    const relative = pathname.replace(/^\//, '')
    let target
    if (pathname.startsWith('/src/')) target = `.${pathname}`
    else if (pathname.startsWith('/assets/') || pathname === '/favicon.ico') target = `public${pathname}`
    else target = pathname.endsWith('/') ? join(generatedRoot, relative, 'index.html') : join(generatedRoot, relative)
    if (!(await exists(target))) errors.push(`${file}: broken internal link ${href}`)
  }

  for (const [, language] of html.matchAll(/class="language-([\w-]+)"/g)) {
    if (builtInPrismLanguages.has(language)) continue
    const component = prismLanguageAliases[language] ?? language
    if (!clientSource.includes(`prism-${component}`))
      errors.push(`${file}: Prism grammar is not imported for ${language}`)
  }
}

const siteUrl = 'https://paulw.xyz'
const generatedFile = (pathname) => {
  const relative = pathname.replace(/^\//, '')
  return pathname.endsWith('/') ? join(generatedRoot, relative, 'index.html') : join(generatedRoot, relative)
}

for (const name of ['feed.xml', 'sitemap.xml']) {
  if (!(await exists(join(generatedRoot, name)))) errors.push(`missing ${name}`)
}

if (await exists(join(generatedRoot, 'feed.xml'))) {
  const feed = await readFile(join(generatedRoot, 'feed.xml'), 'utf8')
  if (!feed.includes('<rss version="2.0"')) errors.push('feed.xml: missing RSS 2.0 channel')
  if (feed.includes('noindex')) errors.push('feed.xml: unexpected noindex marker')
  for (const [, loc] of feed.matchAll(/<guid isPermaLink="true">([^<]+)<\/guid>/g)) {
    const { pathname } = new URL(loc)
    if (!(await exists(generatedFile(pathname)))) errors.push(`feed.xml: missing page for ${loc}`)
  }
}

if (await exists(join(generatedRoot, 'sitemap.xml'))) {
  const sitemap = await readFile(join(generatedRoot, 'sitemap.xml'), 'utf8')
  if (!sitemap.includes('http://www.sitemaps.org/schemas/sitemap/0.9'))
    errors.push('sitemap.xml: missing urlset namespace')

  const locs = [...sitemap.matchAll(/<loc>([^<]+)<\/loc>/g)].map((match) => match[1])
  for (const loc of locs) {
    const { origin, pathname } = new URL(loc)
    if (origin !== siteUrl) errors.push(`sitemap.xml: unexpected origin ${loc}`)
    if (!(await exists(generatedFile(pathname))) && pathname !== '/sitemap.xml')
      errors.push(`sitemap.xml: missing file for ${loc}`)
  }

  const locSet = new Set(locs)
  for (const file of htmlFiles) {
    if (file.endsWith('404.html') || noindexFiles.has(file)) continue
    const relative = file.slice(generatedRoot.length + 1).replaceAll('\\', '/')
    const pathname = relative === 'index.html' ? '/' : `/${relative.replace(/\/index\.html$/, '/')}`
    const loc = `${siteUrl}${pathname}`
    if (!locSet.has(loc)) errors.push(`sitemap.xml: missing ${loc}`)
  }
}

if (!(await exists('public/robots.txt'))) errors.push('missing public/robots.txt')

try {
  const tracked = execFileSync('git', ['ls-files', '-z'], { encoding: 'utf8' })
  for (const file of tracked.split('\0').filter(Boolean)) {
    if (/(?:^|\/)\.[^/]*\.md$/.test(file)) errors.push(`${file}: hidden markdown must not be tracked`)
  }
} catch {}

if (errors.length > 0) {
  console.error(errors.join('\n'))
  process.exitCode = 1
} else {
  console.log(`Checked ${htmlFiles.length} generated pages.`)
}