Patrick R. Michaud via RT wrote: > (3) Case conversions on a string *can* cause its length to change -- in > particular, the character "ß" (U+00DF) becomes "SS" when converted to > uppercase. (I'm not sure that we have any tests for this at present, > and it probably doesn't work when ICU isn't present.)
How can you doubt, when a German hacker takes care of the test suite? after all we're about the only ones who use that weird character ;-) It's in t/spec/S32-str/uc.t, line 45, for regexes this is tested in S05-modifier/ignorecase.t. Actually there might be codepoints that turn into multiple codepoints on conversion to upper case; in particular if there's a precomposed character of a lower case letter and some diacritics, but no upper case equivalent precomposed character exists in the Unicode repertoire. (I don't know if such a thing actually exists, but it's entirely possible). Cheers, Moritz
