Integrating third party libriaries into SPA frameworks could not be trivial at all. Mithril, instead, allows to easily integrate external code to manipulate the DOM outside the virtual DOM inner workings.

Let's take as example this website: we use Prism as syntax highlighter for all the code snippets. Prism applies the syntax highlighting by doing manipulations in the DOM: all styling is done through CSS, so it adds sensible class names to the HTML markup. Because of that it can't work on server side, so we need to require it only on client side, otherwise we will get errors in Node.js (there is no DOM on server side):

const Prism = process.browser ? require('prismjs') : null;

For the same reason, we want Prism run only after the content with the code snippets is rendered by Mithril. For that we simply use the Mithril's oncreate lifecycle method. The virtual node of our content in the view of our app/pages/Section.js main component will then look like:

vnode.state.loading || !vm.section.content ? m(LoadingDots) : m('div', {
    oncreate: () => Prism.highlightAll()
}, vm.section.content)

And that's it! In this way we have implemented the last point of the sequence:

  1. the async request fetches the content (markdown file) in the oninit
  2. the content replaces the loading animation (LoadingDots component) in the view
  3. Prism processes the real DOM and applies the syntax highlighting

Maybe you are questioning if Prism.highlightAll() will throw errors because Prism = null on the server. It wont, because the oncreate lifecycle method isn't called on the server by mithril-node-renderer (only the oninit and the view are).