Specify charset for importscripts in webworkers

Question

I am using webworkers for executing a script seperately apart from my main js thread. While importing the script inside the webworker how to specify the charset?

Answer 1

There is no defined way to do it.

According to specs, fetch a classic worker imported script must read the fetched data as UTF-8:

5. Let source text be the result of UTF-8 decoding response's body.

(Same applies for fetch a classic worker script b.t.w.)

This leaves no room for doubt. Even setting the charset in the Content-Type header wouldn't help here (while it should for fetching classic scripts).

Now, it seems Chrome doesn't honor this rule (at least not for data:// URLs, I didn't try with real file served, and for blob:// URLs they don't even look at the defined charset...), but Firefox does.

// we store the text encoded as base64 to preserve the encoding
// original text is `postMessage("é and ü")`
const script_url = "data:text/javascript;charset=windows-1252;base64,cG9zdE1lc3NhZ2UoIukgYW5kIPwiKQ==";

// importScript() test
{
  const worker_script = `importScripts( "${ script_url }" );`;
  const import_worker = new Worker( URL.createObjectURL( new Blob( [ worker_script ], { type: "text/javascript" } ) ) );
  import_worker.onmessage = (evt) => {
    console.log( 'importScripts:', evt.data );
  };
}

// Worker script test
{
  const direct_worker = new Worker( script_url );
  direct_worker.onmessage = (evt) => {
    console.log( 'worker script:', evt.data );
  };
}

// inline script test
{
  const script = document.createElement( 'script' );
  script.src = script_url;
  document.head.append( script );
  onmessage = (evt) => {
    console.log( 'inline-script:', evt.data );
  };
}

---

Of course, there are ways around that:

• Definitely the best solution is to re-encode your scripts as UTF-8.
• Another completely hacky solution is to fetch your script as a Blob, read it as a DOMString using the correct encoding, and then generate a new UTF-8 Blob that you'd serve to importScripts. Doable, but really ugly.