net.sf.saxon.pattern
Class NodeTest

java.lang.Object
  net.sf.saxon.pattern.NodeTest

All Implemented Interfaces:

Serializable, ItemType

Direct Known Subclasses:

AnyChildNodePattern, AnyNodeTest, CombinedNodeTest, ContentTypeTest, DocumentNodeTest, EmptySequenceTest, LocalNameTest, NamespaceTest, NameTest, NodeKindTest, SubstitutionGroupTest

public abstract class NodeTest
extends Object
implements ItemType, Serializable

A NodeTest is a simple kind of pattern that enables a context-free test of whether a node has a particular name. There are several kinds of node test: a full name test, a prefix test, and an "any node of a given type" test, an "any node of any type" test, a "no nodes" test (used, e.g. for "@comment()").

As well as being used to support XSLT pattern matching, NodeTests act as predicates in axis steps, and also act as item types for type matching.

For use in user-written application calling NodeInfo.iterateAxis(byte, NodeTest), it is possible to write a user-defined subclass of NodeTest that implements a single method, matches(int, int, int)

Author:

Michael H. Kay

See Also:

Serialized Form

Constructor Summary

NodeTest()

Method Summary

AtomicType	getAtomizedItemType() Get the item type of the atomic values that will be produced when an item of this type is atomized (assuming that atomization succeeds)
SchemaType	getContentType() Get the content type allowed by this NodeTest (that is, the type annotation of the matched nodes).
abstract double	getDefaultPriority() Determine the default priority of this node test when used on its own as a Pattern
int	getFingerprint() Get the name of the nodes matched by this nodetest, if it matches a specific name.
int	getNodeKindMask() Get a mask indicating which kinds of nodes this NodeTest can match.
ItemType	getPrimitiveItemType() Get the primitive item type corresponding to this item type.
int	getPrimitiveType() Get the basic kind of object that this ItemType matches: for a NodeTest, this is the kind of node, or Type.Node if it matches different kinds of nodes.
IntHashSet	getRequiredNodeNames() Get the set of node names allowed by this NodeTest.
ItemType	getSuperType(TypeHierarchy th) Get the type from which this item type is derived by restriction.
boolean	isAtomicType() Determine whether this item type is atomic (that is, whether it can ONLY match atomic values)
boolean	isAtomizable() Ask whether values of this type are atomizable
boolean	isNillable() Determine whether the content type (if present) is nillable
abstract boolean	matches(int nodeKind, int fingerprint, int annotation) Test whether this node test is satisfied by a given node.
boolean	matches(NodeInfo node) Test whether this node test is satisfied by a given node.
boolean	matches(TinyTree tree, int nodeNr) Test whether this node test is satisfied by a given node on a TinyTree.
boolean	matchesItem(Item item, boolean allowURIPromotion, Configuration config) Test whether a given item conforms to this type.
String	toString(NamePool pool) Display the type descriptor for diagnostics

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

NodeTest

public NodeTest()

Method Detail

matchesItem

public boolean matchesItem(Item item,
                           boolean allowURIPromotion,
                           Configuration config)

Test whether a given item conforms to this type. This implements a method of the ItemType interface.

Specified by:

matchesItem in interface ItemType

Parameters:

item - The item to be tested

allowURIPromotion -

config -

Returns:

true if the item is an instance of this type; false otherwise

getSuperType

public ItemType getSuperType(TypeHierarchy th)

Get the type from which this item type is derived by restriction. This is the supertype in the XPath type heirarchy, as distinct from the Schema base type: this means that the supertype of xs:boolean is xs:anyAtomicType, whose supertype is item() (rather than xs:anySimpleType).

In fact the concept of "supertype" is not really well-defined, because the types form a lattice rather than a hierarchy. The only real requirement on this function is that it returns a type that strictly subsumes this type, ideally as narrowly as possible.

Specified by:

getSuperType in interface ItemType

Parameters:

th - the type hierarchy cache

Returns:

the supertype, or null if this type is item()

getDefaultPriority

public abstract double getDefaultPriority()

Determine the default priority of this node test when used on its own as a Pattern

getPrimitiveItemType

public ItemType getPrimitiveItemType()

Get the primitive item type corresponding to this item type. For item(), this is Type.ITEM. For node(), it is Type.NODE. For specific node kinds, it is the value representing the node kind, for example Type.ELEMENT. For anyAtomicValue it is Type.ATOMIC_VALUE. For numeric it is Type.NUMBER. For other atomic types it is the primitive type as defined in XML Schema, except that INTEGER is considered to be a primitive type.

Specified by:

getPrimitiveItemType in interface ItemType

getPrimitiveType

public int getPrimitiveType()

Get the basic kind of object that this ItemType matches: for a NodeTest, this is the kind of node, or Type.Node if it matches different kinds of nodes.

Specified by:

getPrimitiveType in interface ItemType

Returns:

the node kind matched by this node test

getFingerprint

public int getFingerprint()

Get the name of the nodes matched by this nodetest, if it matches a specific name. Return -1 if the node test matches nodes of more than one name

isAtomicType

public boolean isAtomicType()

Determine whether this item type is atomic (that is, whether it can ONLY match atomic values)

Specified by:

isAtomicType in interface ItemType

Returns:

false: this is not ANY_ATOMIC_TYPE or a subtype thereof

getAtomizedItemType

public AtomicType getAtomizedItemType()

Get the item type of the atomic values that will be produced when an item of this type is atomized (assuming that atomization succeeds)

Specified by:

getAtomizedItemType in interface ItemType

Returns:

the item type of the atomic values that will be produced when an item of this type is atomized

isAtomizable

public boolean isAtomizable()

Ask whether values of this type are atomizable

Specified by:

isAtomizable in interface ItemType

Returns:

true unless it is known that these items will be elements with element-only content, in which case return false

matches

public boolean matches(TinyTree tree,
                       int nodeNr)

Test whether this node test is satisfied by a given node on a TinyTree. The node must be a document, element, text, comment, or processing instruction node. This method is provided so that when navigating a TinyTree a node can be rejected without actually instantiating a NodeInfo object. The default implementation instantiates the node and then calls the method matches(NodeInfo)

Parameters:

tree - the TinyTree containing the node

nodeNr - the number of the node within the TinyTree

Returns:

true if the node matches the NodeTest, otherwise false

matches

public abstract boolean matches(int nodeKind,
                                int fingerprint,
                                int annotation)

Test whether this node test is satisfied by a given node. This method is only fully supported for a subset of NodeTests, because it doesn't provide all the information needed to evaluate all node tests. In particular (a) it can't be used to evaluate a node test of the form element(N,T) or schema-element(E) where it is necessary to know whether the node is nilled, and (b) it can't be used to evaluate a node test of the form document-node(element(X)). This in practice means that it is used (a) to evaluate the simple node tests found in the XPath 1.0 subset used in XML Schema, and (b) to evaluate node tests where the node kind is known to be an attribute.

Parameters:

nodeKind - The kind of node to be matched

fingerprint - identifies the expanded name of the node to be matched. The value should be -1 for a node with no name.

annotation - The actual content type of the node

matches

public boolean matches(NodeInfo node)

Test whether this node test is satisfied by a given node. This alternative method is used in the case of nodes where calculating the fingerprint is expensive, for example DOM or JDOM nodes. The default implementation calls the method matches(int, int, int)

Parameters:

node - the node to be matched

getNodeKindMask

public int getNodeKindMask()

Get a mask indicating which kinds of nodes this NodeTest can match. This is a combination of bits: 1<

getContentType

public SchemaType getContentType()

Get the content type allowed by this NodeTest (that is, the type annotation of the matched nodes). Return AnyType if there are no restrictions. The default implementation returns AnyType.

getRequiredNodeNames

public IntHashSet getRequiredNodeNames()

Get the set of node names allowed by this NodeTest. This is returned as a set of Integer fingerprints. A null value indicates that all names are permitted (i.e. that there are no constraints on the node name. The default implementation returns null.

isNillable

public boolean isNillable()

Determine whether the content type (if present) is nillable

Returns:

true if the content test (when present) can match nodes that are nilled

toString

public String toString(NamePool pool)

Display the type descriptor for diagnostics

Specified by:

toString in interface ItemType

Parameters:

pool - the name pool

Returns:

a string representation of the type, in notation resembling but not necessarily identical to XPath syntax