Lists Home |
Date Index |
Henry Thompson wrote:
> > If it
> > sits better with you, I'll add the word "non-redundant": "exactly where
> > every [non-redundant] xmlns declaration is and what prefix it uses".
> It's the _exactly where_ part I'm disputing. I couldn't care less
> where the namespaces which are in scope are declared, and the
> [in-scope namespaces] property doesn't record that.
You're quite right that on the single-element level, the [in-scope
namespaces] property does not indicate where the namespace was declared.
However, I can determine that information by recursively traversing the
[parent] property. In other words, for a given Infoset (i.e. the entire
document), I indeed can tell you _exactly where_ all the non-redundant
namespace declarations occurred in the XML source (modulo attribute order).
Individual xmlns declarations are more significant than you're implying. The
extra level of indirection/abstraction provided by the Infoset does not take
away from the fact that this information needs to be preserved.