http://gwt-code-reviews.appspot.com/516801/diff/1/4 File user/super/com/google/gwt/emul/java/lang/String.java (right):
http://gwt-code-reviews.appspot.com/516801/diff/1/4#newcode458 user/super/com/google/gwt/emul/java/lang/String.java:458: int n = str.length(); There's actually a much easier and faster way, if you can somehow cast a String to a byte[]: unescape(encodeURIComponent(bytes)) encodeURIComponent will %-escape each char using UTF-8, and unescape will unescape each %hh sequence. This would turn, e.g. "\u00E9" into "\u00C3\u00A9". You could then get the bytes just like getChars(), which is still probably faster than all those if/else, bit-shifts et al. to do UTF-8 encoding yourself. http://gwt-code-reviews.appspot.com/516801/diff/1/4#newcode515 user/super/com/google/gwt/emul/java/lang/String.java:515: private static String utf8ToString(byte[] bytes, int ofs, int len) { Same as above re. getBytesUtf8: decodeURIComponent(escape(bytes)) would turn "\xC3\xA9" into "\u00E9". Building the "\xC3\xA9" string from a byte[] should be as easy as String.fromCharCode.apply(null, bytes.splice(ofs, len)), as in valueOf. In case the String-constructed-from-bytes contains chars out of the 0-255 range (such as when a byte[] array contain non-byte values, which could happen in web mode), then escape() will turn them into %uHHHH escapes and decodeURIComponent then throws; this can be fixed upfront as in getBytesLatin1 or afterwards with a .replace(/%u([0-9a-f]){4}/gi, function(... I don't have numbers but I think it'd be faster than the "manual UTF-8" algorithm as done here. http://gwt-code-reviews.appspot.com/516801/show -- http://groups.google.com/group/Google-Web-Toolkit-Contributors
