Comments for jstenback

Comment on New DOM bindings by jstenback

jstenback — Sat, 02 Feb 2013 07:42:49 +0000

Yes, significant parts of these new bindings are still under development, and the foundation of these bindings is already part of Firefox OS v1.0 as well. What’s left is largely the long tail of converting more objects to use the new bindings (actively worked on), and some further performance improvements that are now possible due to how we implemented the new bindings (also very actively worked on, some landed in nightly builds just days ago).

Comment on New DOM bindings by Erik Harrison

Erik Harrison — Fri, 01 Feb 2013 20:32:44 +0000

A lot of the public chatter coming out of Mozilla these days relates to Firefox OS. Are these bindings still under development?

Comment on New DOM bindings by Boris

Boris — Mon, 30 Apr 2012 01:07:30 +0000

Ah, interesting. See, the blog post here is in fact talking about the equivalent glue layer for C++. The actual C++ impl is pretty simple (again, if you exclude the hard parts like firing the relevant events and sync XHR tha).

Comment on New DOM bindings by Brandon Benvie

Brandon Benvie — Sun, 29 Apr 2012 20:08:06 +0000

I forgot that stuff is generated and not included in the repo. The XHR interfaces look like this https://gist.github.com/2552989

Comment on New DOM bindings by Brandon Benvie

Brandon Benvie — Sun, 29 Apr 2012 19:52:12 +0000

Most of those issues are dealt with at the interface layer (generated from IDL) that dom.js uses. It enforces types and does translations using things like OptrionalStringOrNull, OptionalBoolean, ToUint32, etc. So anything so that makes it to the impl has already been vetted for types. There are some shortcuts taken, and the implementation isn’t complete, but it’s much closer to complete than it appears just in that code.

Comment on New DOM bindings by Boris

Boris — Sun, 29 Apr 2012 19:47:37 +0000

It’s straightforward if you don’t actually implement what the WebIDL spec says. Which the implementation you link to doesn’t, in all sorts of ways. Most notably, as far as I can tell it doesn’t actually enforce that its methods are called on the right sorts of objects, its readonly attributes are not readonly, it’s not implementing some parts of the interface at all. I’ll assume the underlying Node impl that it’s using for send() actually handles all the various things send() can take correctly, and that Node’s implementation of the user/password stuff is compatible with what the XHR spec actually says (which I rather doubt, a priori).

Nothing wrong with cutting out the edge cases that make things difficult to implement if they’re not relevant to your problem domain, but unfortunately we have to deal with those edge cases in a browser.

Comment on History of Mozilla’s DOM bindings by HTML5 bullets: innovative ClojureScript IDE, CSS filter effects, and more | Share Blog

Sun, 29 Apr 2012 19:06:35 +0000

[...] engineering director Johny Stenback wrote an extremely detailed two-part article about the past and future of Mozilla’s DOM [...]

Comment on New DOM bindings by Brandon Benvie

Brandon Benvie — Sun, 29 Apr 2012 16:20:43 +0000

XMLHttpRequest is also the IDL interface I implemented for dom.js. (though the implementation relies on support from Node.js since I don’t know of any other low level support for networking in JS). https://github.com/Benvie/dom.js/blob/master/src/impl/XMLHttpRequest.js

It’s interesting how a lot of the constraints you work with aren’t necessarily from IDL itself, but rather having to expose it to multiple consumers at once. Between XPCOM stuff and JS itself. Implementing it in JS for JS consumers is relatively straightforward!

Comment on New DOM bindings by Aurora 14 is out! What’s new in it? ✩ Mozilla Hacks – the Web developer blog

Fri, 27 Apr 2012 22:17:03 +0000

[...] with added performance, DOM bindings for list-like objects have landed: these are often called the “Paris DOM bindings” as [...]

Comment on New DOM bindings by DPS911 Summary « diogogmt

DPS911 Summary « diogogmt — Wed, 18 Apr 2012 20:52:14 +0000

[...] readings that I really recomened: History of Mozilla’s DOM bindings New DOM bindings Like this:LikeBe the first to like this [...]