Hi,
I wonder if readline can change the current word bondary algorithm by an algorithm following Unicode UAX TR29 [1].
Current readline algorithm is too simple and it doesn't work fine for some languages. For instance, in Catalan we use MIDDLEDOT U+00B7 as inner-word punctuation, according to UAX TR29, it's a MidLetter char and it doesn't break words, but current readline alghorithm splits words like "goril·la" (gorilla) or "paral·lel" (parallel)