Parsing Large S3 XML Listings in MV3 with Offscreen

Last updated:

MV3 service workers do not have DOMParser. The recommended pattern is to spin up an offscreen document that runs in a DOM context and communicates with the worker via chrome.runtime.sendMessage. This enables fast and robust XML parsing for S3/R2 listings.

Create offscreen document

// utils/offscreen-xml.js (excerpt)
await chrome.offscreen.createDocument({
  url: chrome.runtime.getURL('offscreen.html'),
  reasons: ['DOM_PARSER'],
  justification: 'Parse XML responses from S3 API'
});

Parse XML with DOMParser

// offscreen.js (excerpt)
chrome.runtime.onMessage.addListener((message, sender, sendResponse) => {
  if (message.type === 'PARSE_XML') {
    const doc = new DOMParser().parseFromString(message.payload.xmlText, 'application/xml');
    const contents = Array.from(doc.querySelectorAll('Contents')).map(n => ({
      key: n.querySelector('Key')?.textContent || '',
      size: Number(n.querySelector('Size')?.textContent || '0'),
    }));
    sendResponse({ success: true, data: { items: contents } });
    return true;
  }
});

Fallbacks

When offscreen creation fails (e.g., limited environments), fall back to a regex parser with clear warnings.