This can be caused by invisible characters that can't be displayed in the editor. I've seen this happen with strange quote characters that just don't show up in the editor (some bizarre Unicode symbols that not all fonts can represent). One way to see if that's the problem is to copy the text into an editor that doesn't understand Unicode (e.g. Textpad on Windows), and any non-8-bit characters will show up as "?" or some such. You can also use a regex-function to convert characters above U+00FF to entities:
Code:
def replace(match, number, file_name, metadata, dictionaries, data, functions, *args, **kwargs):
result = ''
for c in match.group():
if ord(c) > 255:
result += '&#%d;' % ord(c)
else:
result += c
return result
Select (mark) the text that looks funny, and do a replace-all on ".*" in the marked text using this function. Then look for any strange "&#nnn;" entities.
Note that the editor is showing you the raw HTML, so it's not an HTML formatting problem -- it's something that's giving the editor indigestion.