oxjs/source/Ox/js/HTML.js

'use strict';

/*@
Ox.parseEmailAddresses <f> Takes HTML and turns e-mail addresses into links
@*/
// fixme: no tests
Ox.parseEmailAddresses = function(html) {
    return html.replace(
        /\b([0-9A-Z\.\+\-_]+@(?:[0-9A-Z\-]+\.)+[A-Z]{2,6})\b/gi,
        '<a href="mailto:$1" title="mailto:$1">$1</a>'
    );
};

/*@
Ox.parseHTML <f> Takes HTML from an untrusted source and returns something sane
    > Ox.parseHTML('http://foo.com, bar')
    '<a href="http://foo.com" title="http://foo.com">foo.com</a>, bar'
    > Ox.parseHTML('(see: www.foo.com)')
    '(see: <a href="http://www.foo.com" title="http://www.foo.com">www.foo.com</a>)'
    > Ox.parseHTML('foo@bar.com')
    '<a href="mailto:foo@bar.com" title="mailto:foo@bar.com">foo@bar.com</a>'
    > Ox.parseHTML('<a href="http://foo.com" onmouseover="alert()">foo</a>')
    '<a href="http://foo.com" title="http://foo.com">foo</a>'
    > Ox.parseHTML('<a href="javascript:alert()">foo</a>')
    '&lt;a href="javascript:alert()"&gt;foo'
    > Ox.parseHTML('[http://foo.com foo]')
    '<a href="http://foo.com" title="http://foo.com">foo</a>'
    > Ox.parseHTML('<rtl>foo</rtl>')
    '<div style="direction: rtl">foo</div>'
    > Ox.parseHTML('<script>alert()</script>')
    '&lt;script&gt;alert()&lt;/script&gt;'
    > Ox.parseHTML('\'foo\' < \'bar\' && "foo" > "bar"')
    '\'foo\' &lt; \'bar\' &amp;&amp; "foo" &gt; "bar"'
    > Ox.parseHTML('<b>foo')
    '<b>foo</b>'
    > Ox.parseHTML('<b>foo</b></b>')
    '<b>foo</b>'
@*/

Ox.parseHTML = (function() {
    var defaultTags = [
            // inline formatting
            'b', 'code', 'i', 's', 'sub', 'sup', 'u',
            // block formatting
            'blockquote', 'h1', 'p', 'pre',
            // lists
            'li', 'ol', 'ul',
            // tables
            'table', 'tbody', 'td', 'tfoot', 'th', 'thead', 'tr',
            // other
            'a', 'br', 'img',
            // special
            'rtl', '[]'
        ],
        parse = {
            a: {
                '<a [^<>]*?href="((https?:\/\/|\/).+?)".*?>': '<a href="{1}" title="{1}">',
                '<\/a>': '</a>'
            },
            img: {
                '<img [^<>]*?src="(https?:\/\/.+?)".*?>': '<img src="{1}">'
            },
            rtl: {
                '<rtl>': '<div style="direction: rtl">',
                '<\/rtl>': '</div>'
            },
            '*': function(tag) {
                var ret = {};
                ret['<(/?' + tag + ')>'] = '<{1}>';
                return ret;
            }
        },
        tab = '\t';

    return function(html, tags, wikilinks) {
        var matches = [],
            tags = tags || defaultTags;
        // html = Ox.clean(html); fixme: can this be a parameter?
        if (tags.indexOf('[]') > -1) {
            html = html.replace(/\[(https?:\/\/.+?) (.+?)\]/gi, '<a href="$1">$2</a>');
            tags = tags.filter(function(tag) {
                return tag != '[]';
            });
        }
        tags.forEach(function(tag) {
            var p = parse[tag] || parse['*'](tag);
            Ox.forEach(p, function(replace, regexp) {
                html = html.replace(new RegExp(regexp, 'gi'), function() {
                    matches.push(Ox.formatString(replace, arguments));
                    return tab + (matches.length - 1) + tab; 
                });
            });
        });
        html = Ox.encodeHTML(html);
        html = Ox.parseURLs(html);
        html = Ox.parseEmailAddresses(html);
        //Ox.print('Ox.parseHTML', html, 'matches', matches);
        matches.forEach(function(match, i) {
            html = html.replace(new RegExp(tab + i + tab, 'gi'), match);
        });
        //html = html.replace(/\n/g, '<br/>\n');
        html = html.replace(/\n\n/g, '<br/><br/>');
        // close extra opening (and remove extra closing) tags
        // note: this converts '&quot;' to '"' 
        return Ox.element('<div>').html(html).html();
    }

}());

/*@
Ox.parseURL <f> Takes a URL, returns its components
    (url) -> <o> URL components
    url <s> URL
    <script>
        Ox.test.object = Ox.parseURL('http://www.foo.com:8080/bar/index.html?a=0&b=1#c');
    </script>
    > Ox.test.object.hash
    '#c'
    > Ox.test.object.host
    'www.foo.com:8080'
    > Ox.test.object.hostname
    'www.foo.com'
    > Ox.test.object.origin
    'http://www.foo.com:8080'
    > Ox.test.object.pathname
    '/bar/index.html'
    > Ox.test.object.port
    '8080'
    > Ox.test.object.protocol
    'http:'
    > Ox.test.object.search
    '?a=0&b=1'
@*/
Ox.parseURL = (function() {
    // fixme: leak memory, like now, or create every time? ... benchmark??
    var a = document.createElement('a'),
        keys = ['hash', 'host', 'hostname', 'origin',
            'pathname', 'port', 'protocol', 'search'];
    return function(str) {
        var ret = {};
        a.href = str;
        keys.forEach(function(key) {
            ret[key] = a[key];
        });
        return ret;
    };
}());

/*@
Ox.parseURLs <f> Takes HTML and turns URLs into links
@*/
// fixme: is parseURLs the right name?
// fixme: no tests
Ox.parseURLs = function(html) {
    return html.replace(
        /\b((https?:\/\/|www\.).+?)([\.,:;!\?\)\]]*?(\s|$))/gi,
        function(str, url, pre, end) {
            url = (pre == 'www.' ? 'http://' : '' ) + url;
            return Ox.formatString(
                '<a href="{url}" title="{url}">{host}</a>{end}',
                {
                    end: end,
                    host: Ox.parseURL(url).hostname,
                    url: url
                }
            );
        }
    );
};