GuideHow-to recipes

Minify HTML with a command step

Run html-minifier-terser on every page with a filter command step, and keep the result cached until the tool changes.

A filter command step runs a program once per file: the file's content goes to its stdin, and its stdout becomes the new content. Any minifier that reads stdin and writes stdout fits. This recipe uses html-minifier-terser, a Node package.

Install the minifierLink to this heading

Install it in the site's folder, so that the version is locked in package-lock.json:

npm install --save-dev html-minifier-terser

Add the stepLink to this heading

The step runs after layout, on the final pages, and write takes its output. Declaring steps replaces the default workflow, so douze.yml lists every step (douze --print-workflow prints the default ones):

douze.yml
steps:
  site-data:
    inputs: "_data/**/*.{yml,yaml,json,xml}"
    uses: data
    with:
      scope: site
  folder-data:
    inputs: "_content/**/*.{yml,yaml,json,xml}"
    uses: data
    with:
      scope: folder
      root: _content
  pages:
    inputs: ["_content/**", "!_content/**/*.{yml,yaml,json,xml}"]
    needs: [site-data, folder-data]
    uses: frontmatter
  permalink:
    needs: pages
    uses: permalink
    with:
      root: _content
  collections:
    needs: permalink
    uses: collect
  paginate:
    needs: [permalink, collections]
    uses: paginate
  render:
    needs: paginate
    uses: liquid
  markdown:
    needs: render
    uses: markdown
  layout:
    needs: markdown
    uses: layout
  minify:
    needs: layout
    mode: filter
    match: "**/*.html"
    run: "node_modules/.bin/html-minifier-terser --collapse-whitespace --remove-comments --minify-css true --minify-js true"
    sources: package-lock.json
  write:
    needs: minify
    uses: write
  • match: "**/*.html" sends only the pages to the minifier; the other files pass through the step unchanged.
  • The program of run is a path, looked up relative to the step's cwd (the source folder). Before the build, Douze checks that it exists, and douze doctor reports it.
  • Each page's result is cached. sources: package-lock.json adds the lock file to the step's cache key, so that updating the minifier runs the step again on every page.

A page such as:

<!doctype html>
<html lang="en">
<head>
  <meta charset="utf-8">
  <title>Home</title>
</head>
<body>
  <main>
    <h1>Home</h1>
    <p>Welcome to the site.</p>
  </main>
</body>
</html>

is written as:

<!doctype html><html lang="en"><head><meta charset="utf-8"><title>Home</title></head><body><main><h1>Home</h1><p>Welcome to the site.</p></main></body></html>

ErrorsLink to this heading

A non-zero exit code fails the page with the end of the program's stderr:

error: minify: /blog/a.html: command exited with code 2 for '/blog/a.html': <last 20 lines of stderr>

Output on stderr fails the page too, even with exit code 0. stderr: warn shows it as a warning instead, and stderr: ignore drops it. Node's own warnings (ExperimentalWarning) go away with env: { NODE_NO_WARNINGS: "1" }.

Many pagesLink to this heading

A filter step starts one process per page, up to -p at once. Starting Node takes about 50 ms, which adds up over thousands of pages. A worker step starts Node once and sends it every page: Write a Node worker step does the same minification in a worker.

See alsoLink to this heading

Esc to close