Ugg. Nobody has even mentioned the accent problem to me, even after a month running XF.
Super120, I will be finding solutions to these problems over the next month or so. (My forums are multilingual. I absolutely have to.) If you write me at the end of summer, I should be able to share the code.
This particular issue is independent of accent related things. This example related to how words are split -- you want ' to be a word splitter when it's in the middle of a word, but that is not really globally applicable. "Don't" should not necessarily tokenize to "don" and "t"; the use of the apostrophe means it's a single word.
MySQL full text doesn't split on this either for example. Unlike with MySQL, there is the ability to manipulate the tokenizing process in Elasticsearch.