|
There's no way to bypass it. Tokenization of text into words id done at indexing time, and once done its done. calibre uses the ICU library to do this tokenization and that uses language sensitive rules, for a number of languages including european ones.
|