Re: [xml] Potential wrong usage of xmlIsID() in tree.c

[Kasimier Buchcik <K Buchcik 4commerce de>]

On Mon, 2006-02-20 at 13:56 -0500, Daniel Veillard wrote:
> On Mon, Feb 20, 2006 at 06:43:34PM +0100, Kasimier Buchcik wrote:
> > > line with how XInclude works and imo would make it easy for a user to 
> >
> > I took a look at the XInclude spec, but it seems - to me - to make no
> > explicit mention of what happens with ID attributes. It does say what
> > to do with IDREF attributes though.
> >
> > In the "Fragment Inclusion Example" of this spec, the included
> > fragment has some ID attributes, which are converted to non-ID
> > attributes in the resulting infoset. So I wonder if this example
> > is just broken, or is preserving IDness from XIncluded fragments
> > just a myth? Did I miss something?
> > http://www.w3.org/TR/xinclude/#fragment-example
>
>   Not an error. It is intended.XInclude works at the infoset level,
> i.e. IDNess is computed at parse time but a property of the node
> (or rather the attribute on the node), and that property is maintained 
> when genertaing the resulting XIncluded tree.
>
> Daniel

What confuses me in the XInclude example is the following sentence:

"The infoset resulting from resolving inclusions on this document is the
same (except for the include history and language properties) as that of
the following document:"
... and then the doc with no IDs is displayed.

I understand this as: if we parse the document and look at the created
infoset, then this (except for the include history and and language
properties) is what the result-infoset should look like.
The @id on <description id="w002-description" .../> does not have a
[attribute type] of ID.
If the example would use the xml:id,
i.e. <description xml:id="w002-description" .../>,
then it would be easier to understand. Hmm, but I don't know if xml:id
produces an [attribute type] of ID.

However, the XInclude spec also says:
"All properties of an information item described in
[XML InformationSet] other than those specifically modified by this
specification are preserved during inclusion"
... so since ID is not mentioned in the spec, the IDness should be
preserved.

Regards & thanks for the clarification,

Kasimier