Possibly a better alternative to parsing innerHTML

[treshugart]

See: https://github.com/webcomponents/webcomponentsjs/blob/3ea23df8bc0f45d44fc903b706d379d7d50aba29/src/ShadowDOM/wrappers/HTMLElement.js#L154. I found this from this comment: webcomponents/webcomponentsjs#511 (comment).

It seems that we should be able to use a similar algorithm to the one in the first link instead of how we're doing it.

This is where it's overridden in a subclass of HTMLElement: https://github.com/webcomponents/webcomponentsjs/blob/3ea23df8bc0f45d44fc903b706d379d7d50aba29/src/ShadowDOM/wrappers/HTMLElement.js#L179.