By default it seems that html.parser.HTMLParser cannot handle self closing tags correctly, if they are not terminated using /. E.g. it handles <img src="asfd"/> fine, but it incorrectly handles <img scr="asdf"> by considering it as a non-self closing tag.