No, that's the reason for my comment#3 - as soon as the Encoding errors made it into the semantic data, only full re-indexing helps, as there is no way for Nepomuk to determine, which data are correctly encoded and which not.
No, that's the reason for my comment#3 - as soon as the Encoding errors made it into the semantic data, only full re-indexing helps, as there is no way for Nepomuk to determine, which data are correctly encoded and which not.