Check the behavior around sanitization

[shuymn]

I think the possibility is very low, but it is a possible bug part on HTMLRewriter side.
It is necessary to check if it reproduces with the minimum configuration.

wildebeest/backend/src/activitypub/objects/index.ts

Lines 271 to 276 in 97ca3d8

	export async function sanitizeContent(unsafeContent: string): Promise<string> {
	if (unsafeContent === '') {
	return ''
	}
	return await getContentRewriter().transform(new Response(unsafeContent)).text()
	}

I feel there is room for improvement in sanitization itself and would like to review the implementation and testing on the Mastodon side to see if the current implementation should be modified.

wildebeest/backend/src/activitypub/objects/index.ts

Lines 290 to 297 in 97ca3d8

	function getContentRewriter() {
	const contentRewriter = new HTMLRewriter()
	contentRewriter.on('*', {
	element(el) {
	if (!['p', 'span', 'br', 'a'].includes(el.tagName)) {
	const element = el as { tagName: string }
	element.tagName = 'p'
	}

ref: https://docs.joinmastodon.org/spec/activitypub/#sanitization