Comments on: Anatomy of a JavaScript object http://blog.mozilla.com/jorendorff/2008/11/17/anatomy-of-a-javascript-object/ where we're building a better SpiderMonkey from parts Tue, 30 Jun 2009 14:40:54 -0700 http://wordpress.org/?v=2.8.6 hourly 1 By: jorendorff http://blog.mozilla.com/jorendorff/2008/11/17/anatomy-of-a-javascript-object/comment-page-1/#comment-21026 jorendorff Tue, 18 Nov 2008 21:33:11 +0000 http://blog.mozilla.com/jorendorff/?p=28#comment-21026 Malte: I think it was originally a coincidence, but retaining the insertion order is now considered a feature. Occasionally there is talk of standardizing it. In any case I doubt we will ever drop it. This is a constraint on storage. In the case of SpiderMonkey, the linked list of property descriptors is in exactly the reverse of property insertion order. Malte: I think it was originally a coincidence, but retaining the insertion order is now considered a feature. Occasionally there is talk of standardizing it. In any case I doubt we will ever drop it.

This is a constraint on storage. In the case of SpiderMonkey, the linked list of property descriptors is in exactly the reverse of property insertion order.

]]> By: jorendorff http://blog.mozilla.com/jorendorff/2008/11/17/anatomy-of-a-javascript-object/comment-page-1/#comment-21025 jorendorff Tue, 18 Nov 2008 21:05:16 +0000 http://blog.mozilla.com/jorendorff/?p=28#comment-21025 "Are we going to see the same thing once Mozilla goes JIT?" Yes, we're already doing this in existing beta releases of Firefox 3.1. (The JIT is off by default in those releases. It will be on by default in Firefox 3.1 beta 2, which is coming soon.) “Are we going to see the same thing once Mozilla goes JIT?”

Yes, we’re already doing this in existing beta releases of Firefox 3.1. (The JIT is off by default in those releases. It will be on by default in Firefox 3.1 beta 2, which is coming soon.)

]]> By: Ilya http://blog.mozilla.com/jorendorff/2008/11/17/anatomy-of-a-javascript-object/comment-page-1/#comment-21016 Ilya Tue, 18 Nov 2008 11:53:59 +0000 http://blog.mozilla.com/jorendorff/?p=28#comment-21016 About the last bullet, Chrome's V8 also already "classifies" JS objects _and_ benefits from it on the JIT level. In fact, Google recommends developers to create their objects always in the same layout (e.g. by initializing with a constructor) to reap the benefits of it. Are we going to see the same thing once Mozilla goes JIT? About the last bullet, Chrome’s V8 also already “classifies” JS objects _and_ benefits from it on the JIT level. In fact, Google recommends developers to create their objects always in the same layout (e.g. by initializing with a constructor) to reap the benefits of it.

Are we going to see the same thing once Mozilla goes JIT?

]]> By: Malte http://blog.mozilla.com/jorendorff/2008/11/17/anatomy-of-a-javascript-object/comment-page-1/#comment-21012 Malte Tue, 18 Nov 2008 08:05:07 +0000 http://blog.mozilla.com/jorendorff/?p=28#comment-21012 Most JS-engines (notable exception being Rhino) iterate over properties in insertion oder (when iterated through for(var foo in object)). Is this a coincidence of early JS implementations and does this place constraints on the storage of properties? Most JS-engines (notable exception being Rhino) iterate over properties in insertion oder (when iterated through for(var foo in object)). Is this a coincidence of early JS implementations and does this place constraints on the storage of properties?

]]> By: Blake Kaplan http://blog.mozilla.com/jorendorff/2008/11/17/anatomy-of-a-javascript-object/comment-page-1/#comment-21004 Blake Kaplan Tue, 18 Nov 2008 04:13:19 +0000 http://blog.mozilla.com/jorendorff/?p=28#comment-21004 mixedpuppy, the brilliance of this stuff is that it's based on the JavaScript code that people already write. For example, the property tree (and by extension, the property cache) take advantage of the fact that even though JavaScript has no named types or structures, existing code already creates many similar objects. mixedpuppy, the brilliance of this stuff is that it’s based on the JavaScript code that people already write. For example, the property tree (and by extension, the property cache) take advantage of the fact that even though JavaScript has no named types or structures, existing code already creates many similar objects.

]]> By: jorendorff http://blog.mozilla.com/jorendorff/2008/11/17/anatomy-of-a-javascript-object/comment-page-1/#comment-21000 jorendorff Mon, 17 Nov 2008 23:49:20 +0000 http://blog.mozilla.com/jorendorff/?p=28#comment-21000 You don't need to know anything about this stuff to write good, fast JavaScript code. This is important only if you want to work on the JavaScript engine itself. You don’t need to know anything about this stuff to write good, fast JavaScript code. This is important only if you want to work on the JavaScript engine itself.

]]> By: mixedpuppy http://blog.mozilla.com/jorendorff/2008/11/17/anatomy-of-a-javascript-object/comment-page-1/#comment-20994 mixedpuppy Mon, 17 Nov 2008 21:47:59 +0000 http://blog.mozilla.com/jorendorff/?p=28#comment-20994 I think I kind of get what's going on. How do I actually take advantage of this in JavaScript itself? When would it be advantageous to do so? I think I kind of get what’s going on. How do I actually take advantage of this in JavaScript itself? When would it be advantageous to do so?

]]>