tools/generate-xhtml-entities.js

#!/usr/bin/env node

var https = require('https');

function get(uri, callback) {
    var content = '';
    // console.log('Downloading', uri);
    https.request(uri, function (response) {
        response.on('data', function (chunk) {
            content += chunk;
        });
        response.on('end', function () {
            callback(content);
        });
    }).end();
}

function process(uri, callback) {
    var entities = [];
    get(uri, function (content) {
        var lines = content.split('\n');
        lines.forEach(function (line) {
            var name, code;
            if (line.indexOf('<!ENTITY') !== 0) {
                return;
            }
            name = /<!ENTITY\s([a-zA-Z0-4]+)/g.exec(line)[1];
            code = /"&#x([0-9A-Fa-f]+);/g.exec(line);
            if (code === null) {
                code = /"&#([0-9]+);/g.exec(line);
                code = parseInt(code[1], 10);
                entities.push({ name: name, code: code });
            } else {
                code = parseInt(code[1], 16);
                entities.push({ name: name, code: code });
            }
        });
        callback(entities);
    });
}

function collect(callback) {
    var entities = [];
    process('https://www.w3.org/2003/entities/2007/xhtml1-special.ent', function (e) {
        entities = entities.concat(e);
        process('https://www.w3.org/2003/entities/2007/xhtml1-lat1.ent', function (e) {
            entities = entities.concat(e);
            process('https://www.w3.org/2003/entities/2007/xhtml1-symbol.ent', function (e) {
                entities = entities.concat(e);
                process('https://www.w3.org/2003/entities/2007/predefined.ent', function (e) {
                    entities = entities.concat(e);
                    callback(entities);
                });
            });
        });
    });
}

function generate() {
    var entities = {};

    function cp(n) {
        var str = n.toString(16).toUpperCase();
        while (str.length < 4) {
            str = '0' + str;
        }
        return '\'\\u' + str + '\'';
    }

    collect(function (data) {
        var codes = [], lines = [];

        data.sort(function (p, q) {
            return p.code - q.code;
        });
        data.forEach(function (u) {
            if (codes.indexOf(u.code) < 0) {
                lines.push('    ' + u.name + ': ' + cp(u.code));
            }
            codes.push(u.code);
        });
        console.log('// Generated by generate-xhtml-entities.js. DO NOT MODIFY!');
        console.log();
        console.log('export const XHTMLEntities = {');
        console.log(lines.join(',\n'));
        console.log('};');
        console.log();
    });
}

generate();