Truncating a string at word break

truncate.js

/**
 * Assumptions:
 * Text is a UTF8 string
 * Text is without HTML entities and HTML tags
 * Suffix is added at nearest word break, excluding punctuation
 * Length of the text string is longer than the suffix
 * Arguments are assumed present and of correct type
 */
(function (exports) {

    // \s: Whitespace
    // \u0021-\u002F: ! " # $ % & ' ( ) * + , - . /
    // \u003A-\u0040: : ; < = ? @
    // \u005B-\u0060: [ \ ] ^ _ '
    // \u007B-\u007F: { | } ~
    // \u00A0-\u00BF: Inverted punctuations, fractions, glyphs, etc

    // \b boundary not safe: http://blog.stevenlevithan.com/archives/javascript-regex-and-unicode
    // unicode ranges: http://www.utf8-chartable.de/unicode-utf8-table.pl?view=3

    var lastword = /\w+$/,
        blacklist = /[\s\u0021-\u002F\u003A-\u0040\u005B-\u0060\u007B-\u007F\u00A0-\u00BF]$/,
        hellip = '\u2026',
        empty = '';

    exports.truncate = function (text, length, suffix) {
        suffix = suffix || hellip;
        text = text.slice(0, length - suffix.length).replace(lastword, empty);
        while (text && blacklist.test(text)) {
            text = text.replace(blacklist, empty);
        }
        return text ? text + suffix : empty;
    };

}(this));

var s = 'Hello ! " # $ % & \' ( ) * + , - . / : ; < = ? @ [ \ ] ^ _ { | } ~ world';
// Truncate mid-way through last word
console.log(this.truncate(s, s.indexOf('r')));

Note to self, \w is also ASCII-only; same flaw as \b

You people & your fake ellipses ;-)

If you're assuming UTF8, why not go with … (\u2026)?

Thinks of all the trees you'll save by omitting those two extra characters!

Logical change to me. Unicode or bust.