Using the OPFS API to determine which mime type a browser infers based on file extension.

mimedb.js

/*

This script runs a Deno server that serves an html page that
uses the the Origin Private File System API to determine which
mime type a browser infers based on file extension. It uses
the mime-db database to enumerate all candidate extensions.

For background, see https://mastodon.social/@jed/110533898437100502

1. Run deno run -A --watch mimedb.js
2. Open http://localhost:8000 in Chrome, Edge, Safari, and Firefox
3. Check the last downloaded CSV for the results.

*/

import db from 'https://esm.sh/mime-db'

let data = [
	['extension', 'mimedb', 'edg', 'chrome', 'safari', 'firefox']
]

let map = new Map()
for (let [type, {extensions = []}] of Object.entries(db)) {
	for (let extension of extensions) map.set(extension, type)
}

data.push(...map)

let onload = async () => {
	let dir = await navigator.storage.getDirectory()
	let form = document.forms[0]
	for (let element of form.elements) try {
		let handle = await dir.getFileHandle(`test.${element.name}`, {create: true})
		let {name, type} = await handle.getFile()
		element.value = type
		dir.removeEntry(name)
	} catch (e) {}

	form.submit()
}

Deno.serve(async request => {
	if (request.method === 'POST') {
		let header = request.headers.get('user-agent')
		let index = data[0].findIndex(a => RegExp(a, 'i').test(header))
		let formData = await request.formData()
		for (let [ext, type] of formData) {
			data.find(r => r[0] === ext)[index] = type
		}

		let csv = data.reduce((a, b) => `${a}\n${b.join(',')}`)
		return new Response(csv, {headers: {'content-type': 'text/csv'}})
	}

	let inputs = data.slice(1).map(r => `<input type=hidden name=${r[0]}>`).join('')
	let html = `<form method=post>${inputs}</form><script>onload=${onload}</script>`
	return new Response(html, {headers: {'content-type': 'text/html'}})
})

results.csv