If the text finishes on a word without further punctuation, the last word of the text is ignored in the parsing. It doesn't count in the statistics of the text, can't be added to terms or edited and so on.
Adding a point at the end of the text fix it. So that is a minor problem. Maybe it makes sense for some languages but for european languages it's strange.
Tested in LWT 2.8.1