Class Expression

  • All Implemented Interfaces:
    java.io.Serializable
    Direct Known Subclasses:
    AttributeExp, BinaryExp, DataExp, ElementExp, Expression.AnyStringExpression, Expression.EpsilonExpression, Expression.NullSetExpression, OtherExp, ReferenceExp, UnaryExp, ValueExp

    public abstract class Expression
    extends java.lang.Object
    implements java.io.Serializable
    Primitive of the tree regular expression. most of the derived class is immutable (except ReferenceExp, ElementExp and OtherExp).

    By making it immutable, it becomes possible to share subexpressions among expressions. This is very important for regular-expression-derivation based validation algorithm, as well as for smaller memory footprint. This sharing is automatically achieved by ExpressionPool.

    ReferebceExp, ElementExp, and OtherExp are also placed in the pool, but these are not unified. Since they are not unified, application can derive classes from these expressions and mix them into AGM. This technique is heavily used to introduce schema language specific primitives into AGM. See various sub-packages of this package for examples.

    The equals method must be implemented by the derived type. equals method will be used to unify the expressions. equals method can safely assume that its children are already unified (therefore == can be used to test the equality, rather than equals method).

    To achieve unification, we overload the equals method so that o1.equals(o2) is true if o1 and o2 are identical. There, those two objects must return the same hash code. For this purpose, the hash code is calculated statically and cached internally.

    See Also:
    Serialized Form
    • Field Detail

      • epsilonReducibility

        private java.lang.Boolean epsilonReducibility
        cached value of epsilon reducibility. Epsilon reducibility can only be calculated after parsing the entire expression, because of forward reference to other pattern.
      • expandedExp

        private Expression expandedExp
        Cached value of the expression after ReferenceExps are removed. This value is computed on demand.
      • verifierTag

        public transient java.lang.Object verifierTag
        this field can be used by Verifier implementation to speed up validation. Due to its nature, this field is not serialized. TODO: revisit this decision of not to serialize this field.
      • cachedHashCode

        private transient int cachedHashCode
        Hash code of this object.

        To memorize every sub expression, hash code is frequently used. And computation of the hash code requires full-traversal of the expression. Therefore, hash code is computed when the object is constructed, and kept cached thereafter.

        This field is essentially final, but because of the serialization support, we cannot declare it as such.

      • epsilon

        public static final Expression epsilon
        Special expression object that represents epsilon (ε). This expression matches to "empty". Epsilon can be thought as an empty sequence.
      • nullSet

        public static final Expression nullSet
        special expression object that represents the empty set (Φ). This expression doesn't match to anything. NullSet can be thought as an empty choice.
      • anyString

        public static final Expression anyString
        special expression object that represents "any string". It is close to xsd:string datatype, but they have different semantics in several things.

        This object is used as <anyString/> pattern of TREX and <text/> pattern of RELAX NG.

    • Constructor Detail

      • Expression

        protected Expression​(int hashCode)
      • Expression

        protected Expression()
        this constructor can be used for the ununified expressions. the only reason there are two parameters is to prevent unintentional use of the default constructor.
    • Method Detail

      • isEpsilonReducible

        public boolean isEpsilonReducible()
        returns true if this expression accepts empty sequence.

        If this method is called while creating Expressions, then this method may return approximated value. When this method is used while validation, this method is guaranteed to return the correct value.

      • calcEpsilonReducibility

        protected abstract boolean calcEpsilonReducibility()
        computes epsilon reducibility
      • getExpandedExp

        public Expression getExpandedExp​(ExpressionPool pool)
        Gets the expression after removing all ReferenceExps, until child AttributeExp or ElementExp.
      • peelOccurence

        public final Expression peelOccurence()
        Peels the occurence expressions from this expression.

        In AGM, 'X?','X+' and 'X*' are represented by using other primitives. This method returns the 'X' part by removing occurence related expressions.

      • hashCode

        public final int hashCode()
        Overrides:
        hashCode in class java.lang.Object
      • setHashCode

        private final void setHashCode​(int hashCode)
      • calcHashCode

        protected abstract int calcHashCode()
        Computes the hashCode again.

        This method and the parameter to the constructor has to be the same. This method is used when the object is being read from the stream.

      • equals

        public abstract boolean equals​(java.lang.Object o)
        Overrides:
        equals in class java.lang.Object
      • hashCode

        protected static int hashCode​(java.lang.Object o1,
                                      java.lang.Object o2,
                                      int hashKey)
      • hashCode

        protected static int hashCode​(java.lang.Object o,
                                      int hashKey)
      • readResolve

        protected java.lang.Object readResolve()