Currently only ASCII-compatible. Would be good to have Binary reading in there, too; should be the same process, after all.