Re: RDF Containers in XUL Trees

R_O_O_K wrote:
> Hi!
> 
> Well, Mozilla doesn't seem to go by the current spec very closely :)
> (this rdf:instanceOf for example)
> I figure out that the builder checks on the type assertion and the 
> existence of nextVal predicate to decide whether or not he current 
> object is a collection.
> 
> I have found no trace of "nextVal" in the spec.
> 
> The spec itself, on the other hand (AFAIK), says nothing about the case:
> <rdf:Seq>
>  <rdf:li/>
>  <rdf:li/>
>  <rdf:_1/>
> </rdf:Seq>
> 
> Should it be interpreted as
> <rdf:Seq>
>  <rdf:_2/>
>  <rdf:_3/>
>  <rdf:_1/>
> </rdf:Seq>
> 
> or
> <rdf:Seq>
>  <rdf:_1/>
>  <rdf:_2/>
>  <rdf:_1/>
> </rdf:Seq>

This one. And the spec does in fact say that it is document order that 
specifies the _n. Note that the _n is neither guaranteed to be unique 
nor continuous. It's just arcs.

> The Moz's RDF flavor is somewhat different, but I believe fully usable.
> Maybe there should be some bug to bring it up to date with the spec.
> But for the time being, maybe it'd be better to simply document the 
> differences.

That bug exists, and there are a few bugs it depends on. I'm not sure 
this particular issue is filed, so please do.

Axel