I'm curious, have you had this support checked by a native speaker of Chinese/Japanese/Korean? Due to Han unification (Unicode's biggest mistake) many code points require different glyphs depending on the target language, but Unifont only has a single glyph per code point--it's impossible for the one glyph to be correct in all CJK languages. It seems more likely that this "kinda works" for CJK but renders incorrect glyphs some of the time.