All Downloads are FREE. Search and download functionalities are using the official Maven repository.

net.sf.saxon.tree.tiny.TinyBuilderCondensed Maven / Gradle / Ivy

There is a newer version: 12.5
Show newest version
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////
// Copyright (c) 2018-2022 Saxonica Limited
// This Source Code Form is subject to the terms of the Mozilla Public License, v. 2.0.
// If a copy of the MPL was not distributed with this file, You can obtain one at http://mozilla.org/MPL/2.0/.
// This Source Code Form is "Incompatible With Secondary Licenses", as defined by the Mozilla Public License, v. 2.0.
////////////////////////////////////////////////////////////////////////////////////////////////////////////////////

package net.sf.saxon.tree.tiny;

import net.sf.saxon.event.PipelineConfiguration;
import net.sf.saxon.om.AttributeInfo;
import net.sf.saxon.trans.XPathException;
import net.sf.saxon.z.IntHashMap;

/**
 * Variant of the TinyBuilder to create a tiny tree in which multiple text nodes or attribute
 * nodes sharing the same string value economize on space by only holding the value once.
 */
public class TinyBuilderCondensed extends TinyBuilder {

    public TinyBuilderCondensed(PipelineConfiguration pipe) {
        super(pipe);
    }

    // Keep a map from the hashcode of the string (calculated within this class) to the list
    // of node numbers of text nodes having that hashcode

    // Note from the specification of CharSequence:
    // "It is therefore inappropriate to use arbitrary CharSequence instances as elements in a set or as keys in
    // a map." We therefore take special care over the design of the map.
    // We rely on the fact that all Saxon implementations of CharSequence have a hashCode() that is compatible with String.
    // And we don't use equals() unless both CharSequences are of the same kind.

    public IntHashMap textValues = new IntHashMap<>(100);

    @Override
    public void endElement() throws XPathException {
        // When ending an element, consider whether the just-completed text node can be commoned-up with
        // any other text nodes. (Don't bother if its more than 256 chars, as it's then likely to be unique)
        // We do this at endElement() time because we need to make sure that adjacent text nodes are concatenated first.

        // First do the endElement() in the normal way
        TinyTree tree = getTree();
        super.endElement();

        // This might have created a TEXTUAL_ELEMENT node
        int last = tree.numberOfNodes - 1;
        boolean sameDepth = tree.depth[last] == getCurrentDepth();
//        if (sameDepth) {
//            boolean isTextualElement = tree.nodeKind[last] == Type.TEXTUAL_ELEMENT;
//
//            // Alternatively, see if there was a final text node
//            boolean hasFinalTextNode = tree.nodeKind[last] == Type.TEXT;
//            if ((isTextualElement || hasFinalTextNode) && (tree.beta[last] <= 256)) {
//                // Get the string value of this final node: works both for text nodes and textual element nodes
//                CodePoints chars = TinyTextImpl.getStringValue(tree, last);
//                // We rely on all relevant implementations of UnicodeString having compatible hashcodes (but if they
//                // don't, the only consequence is that we get less compression)
//                int hash = chars.hashCode();
//                int[] nodes = textValues.get(hash);
//                if (nodes != null) {
//                    // We've seen a previous node with this hash value
//                    int used = nodes[0];
//                    for (int i = 1; i < used; i++) {
//                        int nodeNr = nodes[i];
//                        if (nodeNr == 0) {
//                            break;
//                        } else if (isEqual(chars, TinyTextImpl.getStringValue(tree, nodeNr))) {
//                            // the latest text node is equal to some previous text node
//                            int length = tree.alpha[last];
//                            tree.alpha[last] = tree.alpha[nodeNr];
//                            tree.beta[last] = tree.beta[nodeNr];
//                            tree.getCharacterBuffer().setLength(length);
//                            return;
//                        }
//                    }
//                } else {
//                    // Haven't seen this value before; add an entry to the hash table
//                    nodes = new int[4];
//                    nodes[0] = 1;
//                    textValues.put(hash, nodes);
//                }
//                if (nodes[0] + 1 > nodes.length) {
//                    int[] n2 = Arrays.copyOf(nodes, nodes.length * 2);
//                    textValues.put(hash, n2);
//                    nodes = n2;
//                }
//                // Add this node to the list of distinct nodes with this hash code
//                nodes[nodes[0]++] = last;
//            }
//        }
    }

    /**
     * For attribute nodes, the commoning-up of stored values is achieved simply by calling intern() on the
     * string value of the attribute.
     */

    @Override
    protected String getAttValue(AttributeInfo att) {
        return super.getAttValue(att).intern();
    }


}





© 2015 - 2024 Weber Informatics LLC | Privacy Policy