Re: [xsl] Schema awareness, normalizedString, token

Subject: Re: [xsl] Schema awareness, normalizedString, token
From: David Carlisle <davidc@xxxxxxxxx>
Date: Wed, 21 Feb 2007 15:04:02 GMT
the data model maintains two valued for such a type annotated element,
the string value and the typed value.

It has to be the case that the string value is in the lexical space for
the type as allowed by the schema, but it is explicitly system dependent
whether an implementation of the data model stores the original content
to derive the string value, or if it derives the string value by taking
a canonical serialisation of the typed value.

so xsl:copy which ultimately copies the text nodes in the data model may
or may not preserve the white space and comments that were originally
there (but probably will).

on the other hand xsl:value-of invokes atomization

so that always goes via the typed value so will always return a
canonical form for the data.


Current Thread