This would appear to violate the premise of this famous StackOverflow answer , however this is not parsing as such, only matching or heuristic identification.
Technically, all text is HTML if served to a browser in such a way that the browser chooses to interpret it that way, e.g. using a text/html Content-Type.