Hi,
sorry for late reply...though....
Try this approach
#free spacing regex
</[a-z][0-9]?> #closing html tag, used just as a milestone
(.(?<!\<))+ #any char but not the one that has "<" to
#the left of it, repeated many times, thus
#making a loooong word
<[a-z][0-9]?> # opening html tag, used just as a milestone
this one catches this snippet from your example
</h2>yeeeeeeeeeeeeyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyyy
which destroys <i>
Note: i deleted line breaks from your text before applying my regex
--
Regards, Eugeny
--
best regards, Eugeny