A reference-set contains references that may be accessed by other mapping rules through the <refexpr> expression wherever an expression sequence is allowed. A reference-set is defined in a reference mapping rule by the <references> element.
The <references> element may contain zero or more <reference> elements. Each <reference> element defines a single reference in the set. In addition, each reference may belong to a particular named group, so that several references in a set may have the same name as long as they belong to different groups. An optional group-name attribute in a <reference> element names the group for a reference. References defined without a group-name belong in the default group of a reference-set.
<mapping match=”...”> <references> <!-- Defined 2 references in group-1 --> <reference group-name=”group-1”> ... </reference> <reference group-name=”group-1”> ... </reference> ... <!-- Define a reference in the default group --> <reference> </reference> </references> </mapping> A reference has a name and a value. The name and value for each reference are specified through expression sequences. <reference> <name> ... some expression sequence ... </name> <value> ... some expression sequence ... </value> </reference>
A reference has a name and a value. The name and value for each reference are specified through expression sequences.
<reference> <name> ... some expression sequence ... </name> <value> ... some expression sequence ... </value> </reference>
The name along with the reference’s group-name together form the handle for the interested value we wish to store. The value of a reference can be accessed with an refexpr expression, that is, the value for a stored reference can be accessed wherever an expression sequence is allowed.
The r-stack, the stack of constructed reference-sets, is searched from top to bottom for stored references. Therefore, newer reference-sets having the same reference handles as previous constructed and non-destroyed references-sets always override older references.