2 false 2008-11-14T14:45:10+01:00 16084 9616 443 objectisdate-method 20 8886 enhancement patched 2009-11-16T16:50:18+01:00 32948 T.J. Crowder Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 1.7 @@@ javascript Object.extend(Object, { /** * Returns true if object is instance of Date. */ isDate: function(object) { return object.constructor == Date } }) javascript @@@ javascript Object.extend(Object, { /** * Returns true if object is instance of Date. */ isDate: function(object) { return object.constructor == Date } }) javascript <div><pre><code class="javascript"> Object.extend(Object, { /** * Returns true if object is instance of Date. */ isDate: function(object) { return object.constructor == Date } }) javascript@@@ </code></pre></div> 0 @@@ javascript Object.extend(Object, { /** * Returns true if object is instance of Date. */ isDate: function(object) { return object.constructor == Date } }) @@@ <div><pre><code class="javascript"> Object.extend(Object, { /** * Returns true if object is instance of Date. */ isDate: function(object) { return object.constructor == Date } }) </code></pre></div> false 2008-11-14T14:45:11+01:00 16084 --- {} 443 objectisdate-method 0 8886 new 2009-01-05T23:39:18+01:00 16084 Nesterenko Dmitry Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 0 Since Date is standard object in javascript, prototype should have method to detect it. <div><p>Since Date is standard object in javascript, prototype should have method to detect it.</p></div> false 2008-11-14T14:46:24+01:00 16084 --- {} 443 objectisdate-method 0 8886 new 2008-11-14T14:46:26+01:00 16084 Nesterenko Dmitry Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 0 To make it into core, this would need a solution that works cross-frame. <div><p>To make it into core, this would need a solution that works cross-frame.</p></div> false 2008-11-14T17:51:35+01:00 16084 --- {} 443 objectisdate-method 0 8886 new 2008-11-14T17:51:39+01:00 14168 Tobie Langel Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 14168 0 false 2008-11-14T17:53:25+01:00 16084 --- :tag: :state: new :assigned_user: 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2008-11-14T17:53:27+01:00 14168 Tobie Langel Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 0 what does mean «cross-frame»? <div><p>what does mean &laquo;cross-frame&raquo;?</p></div> false 2008-11-14T18:41:00+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2008-11-14T18:41:02+01:00 16084 Nesterenko Dmitry Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 0 Dmitry, @@@ javascript document.body.appendChild(document.createElement('iframe')); var xDate = frames[frames.length-1].Date; xDate === Date; // false (new xDate()).constructor === Date; // false (new xDate()) instanceof Date; // false @@@ <div><p>Dmitry,</p> <pre><code class="javascript"> document.body.appendChild(document.createElement('iframe')); var xDate = frames[frames.length-1].Date; xDate === Date; // false (new xDate()).constructor === Date; // false (new xDate()) instanceof Date; // false </code></pre></div> false 2008-11-14T20:42:57+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2008-11-14T20:44:18+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 0 One way to "solve" x-frame issue is to use `Object.prototype.toString` which returns `this` object's internal [[Class]] (as part of the resulting string - see 15.2.4.2). Since Date's [[Class]] is specified to be that of "Date" (see 15.9.3.3), I'm pretty sure this solution is viable as long as browser conforms to specification: @@@ javascript function isDate(o) { return Object.prototype.toString.call(o).indexOf('Date') != -1; } // test document.body.appendChild(document.createElement('iframe')); var xDate = frames[frames.length-1].Date; isDate(new Date()); // true isDate(new xDate()); // true isDate({}); // false // etc. @@@ <div><p>One way to "solve" x-frame issue is to use <code>Object.prototype.toString</code> which returns <code>this</code> object's internal [[Class]] (as part of the resulting string - see 15.2.4.2). Since Date's [[Class]] is specified to be that of "Date" (see 15.9.3.3), I'm pretty sure this solution is viable as long as browser conforms to specification:</p> <pre><code class="javascript"> function isDate(o) { return Object.prototype.toString.call(o).indexOf('Date') != -1; } // test document.body.appendChild(document.createElement('iframe')); var xDate = frames[frames.length-1].Date; isDate(new Date()); // true isDate(new xDate()); // true isDate({}); // false // etc. </code></pre></div> false 2008-11-14T20:52:56+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2008-11-14T20:52:57+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 0 Another option is to test for a couple of `Date` methods — of which there are a crapload. Most have very specific names that are unlikely to cause false positives. <div><p>Another option is to test for a couple of <code>Date</code> methods &#8212; of which there are a crapload. Most have very specific names that are unlikely to cause false positives.</p></div> false 2008-11-14T20:58:06+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2008-11-14T20:58:09+01:00 15556 Andrew Dupont Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 0 So which one do you think we should go with? Test methods presence or checking [[Class]] value? Thoughts, anyone? <div><p>So which one do you think we should go with? Test methods presence or checking [[Class]] value?</p> <p>Thoughts, anyone?</p></div> false 2008-12-13T07:01:13+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2008-12-13T07:01:15+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 0 @kangax: the former. <div><p>@kangax: the former.</p></div> false 2008-12-13T16:46:19+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2008-12-13T16:46:23+01:00 14168 Tobie Langel Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 0 Ok. I guess something along these lines then: @@@ javascript function isDate(o) { return !!(o && typeof o == 'object' && 'getDate' in o && 'getFullYear' in o && 'setMilliseconds' in o); } @@@ <div><p>Ok. I guess something along these lines then:</p> <pre><code class="javascript"> function isDate(o) { return !!(o &amp;&amp; typeof o == 'object' &amp;&amp; 'getDate' in o &amp;&amp; 'getFullYear' in o &amp;&amp; 'setMilliseconds' in o); } </code></pre></div> false 2008-12-14T03:11:47+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2008-12-14T03:11:53+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 Your code returns xDate being undefined in IE6/7. On line: @@@javascript xDate = frames[frames.length-1].Date; @@@ <div><p>Your code returns xDate being undefined in IE6/7. On line: @@@javascript xDate = frames[frames.length-1].Date;</p> <pre><code> </code></pre></div> false 2009-01-05T23:32:12+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-05T23:32:14+01:00 16084 Nesterenko Dmitry Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 Why don't we use toString() method on object's constructor and match it with Date.toString @@@ javascript Object.isDate = function(object){ return object.constructor && object.constructor.toString() === Date.toString(); } @@@ <div><p>Why don't we use toString() method on object's constructor and match it with Date.toString</p> <pre><code class="javascript"> Object.isDate = function(object){ return object.constructor &amp;&amp; object.constructor.toString() === Date.toString(); } </code></pre></div> false 2009-01-06T00:53:05+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-06T00:53:10+01:00 29239 Radoslav Stankov Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 @Radoslav Stankov: that's relying on function de-compilation, which isn't ECMAScript compliant, isn't supported on all browsers (notably Opera mini)... and can be costly depending on the implementation. <div><p>@Radoslav Stankov: that's relying on function de-compilation, which isn't ECMAScript compliant, isn't supported on all browsers (notably Opera mini)... and can be costly depending on the implementation.</p></div> false 2009-01-06T01:24:54+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-06T01:24:58+01:00 14168 Tobie Langel Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 @Dmitry IE seems to "populate" a proper context (iframe's DOM) only when frame's content is finished loading. It's easy to modify the test to set iframe's `src` attribute, then poll for iframe DOM to be ready and - once finished - use its `Array` in a test. <div><p>@Dmitry</p> <p>IE seems to "populate" a proper context (iframe's DOM) only when frame's content is finished loading. It's easy to modify the test to set iframe's <code>src</code> attribute, then poll for iframe DOM to be ready and - once finished - use its <code>Array</code> in a test.</p></div> false 2009-01-06T03:24:13+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-06T03:24:14+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 Juriy, why don't you want to use kangax's method, i found in his protolicious: http://github.com/kangax/protolicious/tree/master/experimental/__getClass.js @@@ javascript function __getClass(object){ return Object.prototype.toString.call(object) .match(/^\[object\s(.*)\]$/)[1]; }; @@@ it passes all your test cases in isDate.js <div><p>Juriy, why don't you want to use kangax's method, i found in his protolicious: <a href="http://github.com/kangax/protolicious/tree/master/experimental/__getClass.js"> http://github.com/kangax/protoli...</a></p> <pre><code class="javascript"> function __getClass(object){ return Object.prototype.toString.call(object) .match(/^\[object\s(.*)\]$/)[1]; }; </code></pre> <p>it passes all your test cases in isDate.js</p></div> false 2009-01-06T09:50:28+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-06T09:50:30+01:00 16084 Nesterenko Dmitry Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 @Nesterenko Dmitry: Juriy Zaytsev === kangax :) <div><p>@Nesterenko Dmitry: Juriy Zaytsev === kangax :)</p></div> false 2009-01-06T12:25:04+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-06T12:25:06+01:00 14168 Tobie Langel Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 sorry, i realized later that it's the same person reading kangax's profile )) <div><p>sorry, i realized later that it's the same person reading kangax's profile ))</p></div> false 2009-01-06T12:32:10+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-06T12:32:12+01:00 16084 Nesterenko Dmitry Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 I actually do want to take advantage of [[Class]] (see my previous comment in this thread). `Object.prototype.toString` behavior is unambigous and (contrary to `Function.prototype.toString`) always guarantees to include value of [[Class]] at a specified position in a string. `Object.isArray` is another method that could benefit from `getClass`. Looking at ES3.1 spec draft, it seems like their `Array.isArray` is defined very similar - by checking whether object's [[Class]] has a value of "Array": > 15.4.3.2 Array.isArray ( arg ) The isArray function takes one argument arg, and returns the Boolean value true if the argument is an object whose [[Class]] internal property has the string value "Array"; otherwise it return false. When I get I chance, I'll run tests on more browsers than those covered by prototype (just to check how compliant they are in this regard) <div><p>I actually do want to take advantage of [[Class]] (see my previous comment in this thread). <code>Object.prototype.toString</code> behavior is unambigous and (contrary to <code>Function.prototype.toString</code>) always guarantees to include value of [[Class]] at a specified position in a string.</p> <p><code>Object.isArray</code> is another method that could benefit from <code>getClass</code>. Looking at ES3.1 spec draft, it seems like their <code>Array.isArray</code> is defined very similar - by checking whether object's [[Class]] has a value of "Array":</p> <blockquote> <p>15.4.3.2 Array.isArray ( arg ) The isArray function takes one argument arg, and returns the Boolean value true if the argument is an object whose [[Class]] internal property has the string value "Array"; otherwise it return false.</p> </blockquote> <p>When I get I chance, I'll run tests on more browsers than those covered by prototype (just to check how compliant they are in this regard)</p></div> false 2009-01-06T14:42:23+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-06T14:42:26+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 Tested positive on: Desktop: * Firefox 2.0.0.18, 3.0.5 * Safari 2.0.2 - 3.2.1 * Chrome 1.0.154.36 * Opera 9.27 - 10a * IE 5.5 - 8RC1 Mobile: * Opera Mini 4.2 * MSIEMobile 4.01, 6.0 * Blackberry 8100, 8320, 9000, 9530 * Android 1.0 * Fennec 1.0a1 * Safari Mobile 3.1.1 Which one did I miss? <div><p>Tested positive on:</p> <p>Desktop:</p> <ul> <li>Firefox 2.0.0.18, 3.0.5</li> <li>Safari 2.0.2 - 3.2.1</li> <li>Chrome 1.0.154.36</li> <li>Opera 9.27 - 10a</li> <li>IE 5.5 - 8RC1</li> </ul> <p>Mobile:</p> <ul> <li>Opera Mini 4.2</li> <li>MSIEMobile 4.01, 6.0</li> <li>Blackberry 8100, 8320, 9000, 9530</li> <li>Android 1.0</li> <li>Fennec 1.0a1</li> <li>Safari Mobile 3.1.1</li> </ul> <p>Which one did I miss?</p></div> false 2009-01-07T02:04:50+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-07T02:07:13+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 er... Camino, flock, Avant Browser, Konqueror, Lynx, Flock, Arachne, Epiphany, K-Meleon... ;) Awesome find kangax! <div><p>er... Camino, flock, Avant Browser, Konqueror, Lynx, Flock, Arachne, Epiphany, K-Meleon... ;)</p> <p>Awesome find kangax!</p></div> false 2009-01-07T02:25:25+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-07T02:25:29+01:00 14168 Tobie Langel Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 Have you checked what was rendered for host objects in the different browsers ? <div><p>Have you checked what was rendered for host objects in the different browsers ?</p></div> false 2009-01-07T02:28:21+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-07T02:28:23+01:00 14168 Tobie Langel Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 Tobie, I remember host objects yielding pretty strange values in some of the browsers : ) You can see some of them in comments of the original `__getClass` [1] I'm a bit worried about host objects throwing errors when asked for [[Class]] (if it is, say, unimplemented). This is unlikely, of course, but is possible (IE already throws errors for [[Get]] on some objects) and would be pretty disturbing : ) There was also a discussion on ES4 mailing list regarding requirements for host objects' [[Class]] values (in ES3.1). Perhaps that discussion could shed more light on whether we could/should use it. [1] http://github.com/kangax/protolicious/tree/master/experimental/__getClass.js <div><p>Tobie,</p> <p>I remember host objects yielding pretty strange values in some of the browsers : ) You can see some of them in comments of the original <code>__getClass</code> <a href="/projects/8886/changesets/1" title="Changeset #1">[1]</a></p> <p>I'm a bit worried about host objects throwing errors when asked for [[Class]] (if it is, say, unimplemented). This is unlikely, of course, but is possible (IE already throws errors for [[Get]] on some objects) and would be pretty disturbing : )</p> <p>There was also a discussion on ES4 mailing list regarding requirements for host objects' [[Class]] values (in ES3.1). Perhaps that discussion could shed more light on whether we could/should use it.</p> <p><a href="/projects/8886/changesets/1" title="Changeset #1">[1]</a> <a href="http://github.com/kangax/protolicious/tree/master/experimental/__getClass.js"> http://github.com/kangax/protoli...</a></p></div> false 2009-01-07T02:46:33+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-07T02:47:51+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 Another batch (mostly *nix this time) - all good to go : ) * Camino 1.6.6 * SeaMonkey 1.1.14 * Flock 2.0.2 * Galeon 2.0.6 * Epiphany 2.22 * Konqueror 4.1.2 * Arora 0.4 * OmniWeb 622.3.0.105198 <div><p>Another batch (mostly *nix this time) - all good to go : )</p> <ul> <li>Camino 1.6.6</li> <li>SeaMonkey 1.1.14</li> <li>Flock 2.0.2</li> <li>Galeon 2.0.6</li> <li>Epiphany 2.22</li> <li>Konqueror 4.1.2</li> <li>Arora 0.4</li> <li>OmniWeb 622.3.0.105198</li> </ul></div> false 2009-01-07T03:05:09+01:00 16084 --- {} 443 objectisdate-method 0 8886 enhancement needs_patch needs_tests 2009-01-07T03:05:10+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 1 Here is a patch implementing `Object.isDate` using `getClass`, with documentation and corresponding unit tests. <div><p>Here is a patch implementing <code>Object.isDate</code> using <code>getClass</code>, with documentation and corresponding unit tests.</p></div> false 2009-02-25T00:34:49+01:00 16084 --- :tag: needs_patch needs_tests 443 objectisdate-method 0 8886 enhancement patched tested 2009-02-25T00:34:52+01:00 17913 Samuel Lebeau Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 Tobie Langel 14168 2 Samuel, thanks, looks good. The only thing is that in documentation, "Returns `true` if `object` is an instance of the [[Date]] class" is a bit misleading. There are really no instances of [[Date]] class. Rather, there are objects with internal [[Class]] property having a value of "Date" : ) In other words, those are native `Date` objects created (explicitly or implicitly) with a native `Date` constructor (in any DOM context). I would also be in favor of avoiding an extra function call in `Object.is*` methods (for perf. reasons) Maybe replace it with a simple string comparison? @@@ javascript (function(){ var _toString = Object.prototype.toString; ... function isDate(o) { return _toString.call(o) === '[object Date]'; } ... Object.isDate = isDate; })(); @@@ Marking as 1.6.1 <div><p>Samuel, thanks, looks good. The only thing is that in documentation, "Returns <code>true</code> if <code>object</code> is an instance of the [[Date]] class" is a bit misleading.</p> <p>There are really no instances of [[Date]] class. Rather, there are objects with internal [[Class]] property having a value of "Date" : ) In other words, those are native <code>Date</code> objects created (explicitly or implicitly) with a native <code>Date</code> constructor (in any DOM context).</p> <p>I would also be in favor of avoiding an extra function call in <code>Object.is*</code> methods (for perf. reasons)</p> <p>Maybe replace it with a simple string comparison?</p> <pre><code class="javascript"> (function(){ var _toString = Object.prototype.toString; ... function isDate(o) { return _toString.call(o) === '[object Date]'; } ... Object.isDate = isDate; })(); </code></pre> <p>Marking as 1.6.1</p></div> false 2009-02-25T02:06:17+01:00 16084 --- :milestone: 9616 443 objectisdate-method 0 8886 enhancement patched tested 2009-02-25T02:06:22+01:00 15559 Juriy Zaytsev Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 1.7 Tobie Langel 2 [responsible:none bulk edit command] <div><p>[responsible:none bulk edit command]</p></div> false 2009-11-16T16:50:18+01:00 16084 --- :tag: patched tested :assigned_user: 14168 9616 443 objectisdate-method 0 8886 enhancement patched 2009-11-16T16:50:18+01:00 32948 T.J. Crowder Nesterenko Dmitry http://prototype.lighthouseapp.com/projects/8886/tickets/443 1.7 f75317c2394544dc95abe433e4c36f62db17cf7a text/javascript 2008-12-24T18:33:15+01:00 isDate.js 73800 505 15559 http://prototype.lighthouseapp.com/attachments/73800/isDate.js 706bf39bee3c4e4e862b36c0bfc305443882b08b text/plain 2009-02-25T00:34:49+01:00 object_isDate.patch 92581 1515 17913 http://prototype.lighthouseapp.com/attachments/92581/object_isDate.patch