For those not familiar with Han unification, Unicode uses the same codepoints to encode CJK characters as used in Chinese and Japanese, even when they look slightly different. This has spurred some controversy, since the characters look "wrong" if the default font of the user doesn't match the intended language, unless one uses metadata to specify the language (e.g. the HTML "lang" attribute).
Leaving pros and cons of Han unification aside, are there any other (non-CJK) examples of glyphs that look (or should look) different depending on the language? Failing that, perhaps there are historical examples where that used to be the case?