Nice, browsers do interpret overlapping elements in non well-formed HTML correctly.
This is the simple test HTML from http://stamm-wilbrandt.de/en/twitter/ub.html:
All big5 browsers handle overlapping tags well on Windows/Linux (click image for 1:1):
[just learned that rfc3092.txt is on "foo" and "bar"]
"circle graphs" from 1994 research publication are related to these non well-formed but "nearly" well-formed constructs.