Re: [xsl] Problem with Entities

["Colin Adams" <colinpauladams@xxxxxxxxxxx>]

> From: David Carlisle <davidc@xxxxxxxxx>
>
> > Will the output of this always be relative to the location of the source 
> xml?
> I think in practice it just uses whatever string you put in the DTD
> SYSTEM identifier.
>
> Unlike other URI it doesn't seem to be automatically resolved angainst
> any base.
>
> XSLT2 says it returns the system id from the data model and teh data
> model says
>
>
> http://www.w3.org/TR/xpath-datamodel/#dm-unparsed-entity-system-id
>
> The dm:unparsed-entity-system-id accessor returns the system identifier
> of an unparsed external entity declared in the specified document. I
>
> Note that this is _not_ defined to be a URI, but defined to be a "system 
> identifier"
>
> XDM defines "system identifier" in section 6.5.1 to be "The system 
> identifier of the entity."
> which is getting a bit circular,
>
>
> but infoset defines it
> http://www.w3.org/TR/xml-infoset/#infoitem.rse
> here as
>
> [system identifier] The system identifier of the entity, as it appears
> in the declaration of the entity, without any additional URI escaping
> applied by the processor.
>
> so I think that confirms that you should get back the DTD-provided
> string.
The infoset also defines a declaration base uri for it to be resolved 
against.
It looks to me like the DM should be using that.

On practical grounds, since SAX will be doing just that (effectively), it 
would be better if the DM defined that, so we get consistency across 
processors. Perhaps it can be considered for the errata?

_________________________________________________________________
MSN Hotmail is evolving  check out the new Windows Live Mail 
http://ideas.live.com