How this mixed-character string split on unicode word boundaries

Question

Consider the string "abc를". According to unicode\'s demo implementation of word segmentation, this string should be split into two words, "abc&