S9API interface

The Processor class now has a zero-argument constructor (new Processor()) which delivers the most capable processor available, based on the software and license key found on the classpath. For example if Saxon-EE is on the classpath and a license key with schema validation capability is located, then a Saxon-EE configuration is created; while if Saxon-PE is on the classpath, but no license key is found, it will deliver a Saxon-HE configuration. It is possible to interrogate the returned Processor to determine its capabilities.

A number of changes have been made to simplify the use of item types and sequence types (particularly useful when defining extension functions):

• Added the method ItemTypeFactory.parseItemType(String) as a convenient way of constructing an item type such as map(xs:string, node()*) from its XPath textual representation.

• Added methods to ItemType to construct a SequenceType by adding an occurrence indicator: one(), zeroOrOne(), zeroOrMore(), oneOrMore(). For example ItemType.STRING.zeroOrMore() represents the type xs:string*.

• Added the method SequenceType.matches(XdmValue) to test whether a value matches a sequence type.

• The method ItemTypeFactory.getItemType(item), when applied to a function item, now returns an item type that retains the full function signature, rather than simply returning function(*). This provides a convenient way to create a function item type by copying the signature of an existing function item.

• The results of ItemType.toString() have been tidied up and are better documented.

• The class ItemType now implements Predicate<XdmItem>, so an item type can be used wherever a predicate is expected (such as in a call to Step.where()).

• A new method XdmValue.where() is provided to filter an XdmValue on any predicate — including an ItemType. For example, value.where(ItemType.INTEGER) selects those items that are instances of xs:integer.

• XdmNode has a new method getTypeAnnotationName() which returns the name of the type annotation of the node (defaulting to xs:untyped and xs:untypedAtomic for elements and attributes that have not been schema-validated).

The XsltCompiler has two new convenience methods: the compile() and compilePackage() methods now have overloads that take a java.io.File as their argument.

The SchemaManager has a new convenience method: the load() method now has an overload that takes a java.io.File as its argument.

The output of the XdmNode method toString() has changed so it is now consistent with the output of XdmValue.toString() in the case where the XdmValue comprises a single node. That is, it is now output using the adaptive serialization method.

The class XdmNode has a new convenience method getOutermostElement() to find the outermost element (also known as the document element) of a tree.