RE: parseType="literal"

Nice example.

It's a bug.
It's a bug with the spec (M&S); well not really a bug, M&S is deliberately
and consciously vague.
It's a bug with the implementation (was that ARP?).

I am very pleased to see an example with processing instructions in.

The WG is currently working on specifying the bahaviour more clearly.

I'll add some in-line comments ...

>   <tp:representacionXul rdf:parseType="Literal">
>       <?xml-stylesheet href="/sistema-acceso/css/bindings.css"
>                        type="text/css"?>
>       <?xul-overlay href="/sistema-acceso/xul/c4d_common.xul"?>
Those processing instructions are difficult.
An XML based processor will prcoess them through a different path than most
of the rest of XML content.
I think some processing instructions might be expected to be processed
before RDF processing, whereas these want to wait until after RDF

>       <xul:window id="Simulacion"
>                   title="Pantalla de Simulación"
> xmlns="
> y.xul"
> xmlns:xul=""
> xmlns:c4d="">
>                 <vbox></vbox>
>       </xul:window>
>   </tp:representacionXul>
>    so, why this is converted to this?
>    <tp:representacionXul rdf:parseType='Literal'>
>       	<xul:window
> xmlns:xul="
> ly.xul"
> xmlns=""
> id="Simulacion"

Reordering of attributes is permitted because it is XML.
The algorithm seems to be:
 - the element has visibly used the xul namespace, so let's declare it.
 - the next attribute uses the default namespace, so let's declare it.
> xmlns=""
Repeating this default namespace declaration is an error.
 - the element has visibly used the xul namespace, so let's declare it.

> title="Pantalla de Simulación">
>   	   <vbox

Repeating the namespace declaration is unnecessary, but not IMO wrong.
M&S does not specify.
Repeating the namespace declaration changes the XML Infoset but does not
change the XPath nodeset.

> xmlns="">
>            </vbox>

Have you transcribed this into e-mail incorrectly: the whitespace element
content of the vbox element is wrong. It is not the sort of bug I would

>         </xul:window>
>    </tp:representacionXul>
>    Anybody knows why???

FIX: wait a few weeks for the WG to determine what the "correct" behaviour
should be, wait a few weeks (or months) more for your implementor (me?) to
fix it.
(Was that really ARP? It's worse than I expected :( .)

>    Thanks,
>           Marc

Received on Tuesday, 12 March 2002 00:25:34 UTC