Comments on: Using E4X With XHTML? Watch Your Namespaces!

By: links for 2009-01-22 « Jeffrey Isham

links for 2009-01-22 « Jeffrey Isham — Fri, 23 Jan 2009 00:01:48 +0000

[...] dispatchEvent()™ » Blog Archive » Using E4X With XHTML? Watch Your Namespaces! (tags: actionscript) [...]

By: Derek Basch

Derek Basch — Thu, 22 Jan 2009 07:25:22 +0000

Thanks,

This was invaluable. I was stuck on this for hours.

By: Monokai » Blog Archive » Flash SEO: graceful degradation to add meaning and structure

Mon, 20 Oct 2008 08:40:13 +0000

[...] Actionscript 3 E4X rundown Using E4X with XHTML? Watch your namespaces! [...]

By: Philip Bulley

Philip Bulley — Wed, 02 Apr 2008 16:33:30 +0000

Great post… figured I had a namespace issue, but had no idea of how to reference it! Thanks mate.

By: Mitch

Mitch — Tue, 17 Jul 2007 12:27:29 +0000

This is invaluable. Thank you for publishing this.

By: Mark

Mark — Thu, 05 Jul 2007 14:52:57 +0000

Don’t forget:
if (myXhtmlDocument.namespace(”") != undefined){
default xml namespace = myXhtmlDocument.namespace(”");
}

If your xml has a default namespace it’ll set it as your default, so your calls won’t need the namespace prepended

By: dispatchEvent » AS3 E4X Rundown

dispatchEvent » AS3 E4X Rundown — Tue, 22 May 2007 01:07:43 +0000

[...] I can write more about namespaces in a future revision, but for now check out the good discussion started in my previous post, Using E4X? Watch Your Namespaces. [...]

By: Peter Hall

Peter Hall — Wed, 28 Mar 2007 11:50:11 +0000

It’s not really that toXMLString() is limited. e4x works on a canonicalised data model. There are several ways to achieve the same XML object from different source strings – using default vs explicit namespaces is just one example. Canonicalisation makes XML simpler to work with, at the expense of losing how it was originally declared.

By: Roger Braunstein

Roger Braunstein — Fri, 23 Mar 2007 00:13:37 +0000

No, thank you for making things more interesting! Not to mention, I learned something about XML-ID!
I also think it’s weird that just because the node exists in a namespace, it’s not considered as declaring that namespace…

By: Claus Wahlers

Claus Wahlers — Fri, 23 Mar 2007 00:09:21 +0000

Yeah something like that, i don’t know. What worries me more though is the buggy (?) behavior of namespaceDeclarations() in that case.

Sorry btw for hijacking your post ;)

By: Roger Braunstein

Roger Braunstein — Thu, 22 Mar 2007 23:25:13 +0000

It seems like a limitation that toXmlString() can't handle default namespaces or implicitly defined namespaces. For instance, when you use xml:id, you're using a namespace xml that you don't declare, and when you're using the default namespace http://example.com that namespace doesn't have a binding. toXmlString() apparently wants everything to be explicit, so it makes up bindings for the two namespaces — aaa and aab — and uses them to explicitly namespace the id attribute as well as the node. I think the output XML string is functionally equivalent, but it's clearly not literally equivalent :(

By: Claus Wahlers

Claus Wahlers — Thu, 22 Mar 2007 20:09:53 +0000

Note that you didn’t have to declare xmlns:xml=”http://www.w3.org/XML/1998/namespace” in the node, but the attribute was clearly scoped to that namespace. That means AS3 respects its implicit existence. :)

Yep that’s the way it’s supposed to be. However, toXMLString() seems to behave wrong as it’s adding the xml namespace declaration (which is not valid).
Also namespaceDeclarations() seems to have a general problem when used with XMLLists downcasted to XML (or something)

By: Roger Braunstein

Roger Braunstein — Thu, 22 Mar 2007 18:50:23 +0000

Interesting! I didn't know about xml:id. It appears that for processors that support it, the xml namespace is bound implicitly to http://www.w3.org/XML/1998/namespace, and id is an attribute scoped to this namespace. AS3 seems to support this. Hurray! Here are some ways you could extract the id attribute. This shows how you can use namespaced attributes as well.

var b:XML = XML('');
trace(b.attributes()[0]); //wtf
trace(b.@id); //null
trace(b.@*::id); //wtf
namespace xml="http://www.w3.org/XML/1998/namespace";
trace(b.@xml::id); //wtf

Note that you didn't have to declare xmlns:xml="http://www.w3.org/XML/1998/namespace" in the node, but the attribute was clearly scoped to that namespace. That means AS3 respects its implicit existence. :)

By: Claus Wahlers

Claus Wahlers — Thu, 22 Mar 2007 18:24:19 +0000

What’s up with the semicolon

Whoops, that must have been an email typo, sorry for that. There’s no lost semicolon in my original code.

and xml: should be xmlns:

Nope, see here: http://www.w3.org/TR/xml-id/

Shouldn’t b.namespaceDeclarations() also include the default namespace declaration (xmlns=”http://example.com/”), or does that get stripped out because it’s already accessible via b.namespace()?

By: Roger Braunstein

Roger Braunstein — Thu, 22 Mar 2007 17:08:32 +0000

I think in this example the processing instruction is a red herring. You were testing:

right? But I think that's not well formed. What's up with the semicolon, and xml: should be xmlns: Try this:

var a:XMLList = new XMLList(
   '' +
   '');
var b:XML = a[1];
trace(b.namespaceDeclarations()); //wtf
trace(b.namespace()); //http://example.com/

namespaceDeclarations() gets the namespaces the node defines, and namespace() gets the namespace the node is in. You can see that the processing instruction has no impact on whether this works.