Merge pull request #68 from oncletom/fix-66

Add tld.validHosts

README.md

@@ -135,6 +135,26 @@ tld.isValid('https://user:password@example.co.uk:8080/some/path?and&query#hash')
 
 # Troubleshouting
 
+## Retrieving subdomain of `localhost` and custom hostnames
+
+`tld.js` methods `getDomain` and `getSubdomain` are designed to **work only with *valid* TLDs**.
+This way, you can trust what a domain is.
+
+Unfortunately, `localhost` is a valid hostname but it is not a TLD.
+`tld.js` has a concept of `validHosts` you declare
+
+```js
+var tld = require('tldjs');
+
+tld.getDomain('localhost');           // returns null
+tld.getSubdomain('vhost.localhost');  // returns null
+
+tld.validHosts = ['localhost'];
+
+tld.getDomain('localhost');           // returns 'localhost'
+tld.getSubdomain('vhost.localhost');  // returns 'vhost'
+```
+
 ## Updating the TLDs List
 
 Many libraries offer a list of TLDs. But, are they up-to-date? And how to update them?

lib/rule.js

@@ -6,6 +6,7 @@ function Rule (data){
   this.exception = data.exception || false;
   this.firstLevel = data.firstLevel || '';
   this.secondLevel = data.secondLevel || null;
+  this.isHost = data.isHost || false;
   this.source = data.source || '';
   this.wildcard = data.wildcard || false;
 }
@@ -62,16 +63,19 @@ Rule.prototype.getPattern = function getPattern(before, after){
   var pattern = '';
 
   before = (before === undefined) ? '(': before+'';
-  after = (after === undefined) ? ')$': before+'';
+  after = (after === undefined) ? ')$': after+'';
 
   if (this.exception === true){
     pattern = this.getExceptionPattern();
   }
+  else if (this.isHost === true) {
+    pattern = this.firstLevel;
+  }
   else{
     pattern = '[^\\.]+' + (this.wildcard ? this.getWildcardPattern() : this.getNormalPattern());
   }
 
   return before + pattern + after;
 };
 
-module.exports = Rule;
+module.exports = Rule;

lib/tld.js

@@ -11,6 +11,7 @@ var URL = require('url');
  */
 function tld () {
   /* jshint validthis: true */
+  this.validHosts = [];
   this.rules = [];
 }
 
@@ -91,9 +92,9 @@ tld.getCandidateRule = function getCandidateRule (host, rules, options) {
   _someFunction(rules, function (r) {
     var pattern;
 
-    // sld matching? escape the loop immediately (except if it's an exception)
+    // sld matching or validHost? escape the loop immediately (except if it's an exception)
     if ('.' + host === r.getNormalXld()) {
-      if (options.lazy || r.exception === true) {
+      if (options.lazy || r.exception || r.isHost) {
         rule.normal = r;
       }
 
@@ -243,7 +244,7 @@ tld.prototype.getDomain = function getDomain (host) {
 
   host = tld.cleanHostValue(host);
   hostTld = tld.extractTldFromHost(host);
-  rules = this.getRulesForTld(hostTld, new Rule({"firstLevel": hostTld}));
+  rules = this.getRulesForTld(hostTld, new Rule({"firstLevel": hostTld, "isHost": this.validHosts.indexOf(hostTld) !== -1}));
   rule = tld.getCandidateRule(host, rules);
 
   if (rule === null) {
@@ -288,12 +289,11 @@ tld.prototype.getSubdomain = function getSubdomain(host){
  * Beware: it does not check if the TLD exists.
  *
  * @api
- * @todo handle localhost, local etc.
  * @param host {String}
  * @return {Boolean}
  */
 tld.prototype.isValid = function isValid (host) {
-  return !(typeof host !== 'string' || host.indexOf('.') === -1 || host[0] === '.');
+  return typeof host === 'string' && (this.validHosts.indexOf(host) !== -1 || (host.indexOf('.') !== -1 && host[0] !== '.'));
 };
 
 /**

test/tld.js

@@ -17,6 +17,7 @@ describe('tld.js', function () {
   describe('isValid method', function () {
     it('should detect valid hostname', function () {
       expect(tld.isValid('')).to.be(false);
+      expect(tld.isValid('localhost')).to.be(false);
       expect(tld.isValid('google.com')).to.be(true);
       expect(tld.isValid('miam.google.com')).to.be(true);
       expect(tld.isValid('miam.miam.google.com')).to.be(true);
@@ -33,6 +34,7 @@ describe('tld.js', function () {
     });
 
     it('should be falsy on invalid domain syntax', function () {
+      expect(tld.isValid('.localhost')).to.be(false);
       expect(tld.isValid('.google.com')).to.be(false);
       expect(tld.isValid('.com')).to.be(false);
     });
@@ -205,18 +207,19 @@ describe('tld.js', function () {
     it('should return www.nytimes.com even with an URL as a parameter', function(){
       expect(tldLib.cleanHostValue('http://www.nytimes.com/glogin?URI=http://www.notnytimes.com/2010/03/26/us/politics/26court.html&OQ=_rQ3D1Q26&OP=45263736Q2FKgi!KQ7Dr!K@@@Ko!fQ24KJg(Q3FQ5Cgg!Q60KQ60W.WKWQ22KQ60IKyQ3FKigQ24Q26!Q26(Q3FKQ60I(gyQ5C!Q2Ao!fQ24')).to.equal('www.nytimes.com');
     });
-    
+
     it('should return punycode for international hostnames', function() {
       expect(tldLib.cleanHostValue('台灣')).to.equal('xn--kpry57d');
     });
   });
 
   describe('getSubdomain method', function(){
     it('should return null if the domain cannot be found', function(){
-      expect(tld.getSubdomain('localhost')).to.equal(null);
+      expect(tld.getSubdomain('not-a-validHost')).to.equal(null);
     });
 
     it('should return the relevant subdomain of a hostname', function(){
+      expect(tld.getSubdomain('localhost')).to.equal(null);
       expect(tld.getSubdomain('google.com')).to.equal('');
       expect(tld.getSubdomain('fr.google.com')).to.equal('fr');
       expect(tld.getSubdomain('random.fr.google.com')).to.equal('random.fr');
@@ -267,11 +270,33 @@ describe('tld.js', function () {
       var domain = tld.getSubdomain('http://cdn.jsdelivr.net/g/jquery@1.8.2,jquery.waypoints@2.0.2,qtip2@2.2.1,typeahead.js@0.9.3,sisyphus@0.1,jquery.slick@1.3.15,fastclick@1.0.3');
       expect(domain).to.equal('cdn');
     });
-    
+
     //@see https://github.com/oncletom/tld.js/issues/35
     it('should provide consistent results', function(){
       expect(tld.getSubdomain('www.bl.uk')).to.equal('www');
       expect(tld.getSubdomain('www.majestic12.co.uk')).to.equal('www');
     });
   });
+
+  describe('validHosts', function(){
+    before(function(){
+      tld.validHosts = ['localhost'];
+    });
+
+    it('should now be a valid host', function(){
+      expect(tld.isValid('localhost')).to.be(true);
+    });
+
+    it('should return the known valid host', function () {
+      expect(tld.getDomain('localhost')).to.equal('localhost');
+      expect(tld.getDomain('subdomain.localhost')).to.equal('localhost');
+      expect(tld.getDomain('subdomain.notlocalhost')).to.equal('subdomain.notlocalhost');
+      expect(tld.getDomain('subdomain.not-localhost')).to.equal('subdomain.not-localhost');
+    });
+
+    //@see https://github.com/oncletom/tld.js/issues/66
+    it('should return the subdomain of a validHost', function(){
+      expect(tld.getSubdomain('vhost.localhost')).to.equal('vhost');
+    });
+  });
 });