xom-interest AT lists.ibiblio.org

Subject: XOM API for Processing XML with Java

List archive

[XOM-interest] [PATCH] Text using java.nio.charset.Charset

From: Wolfgang Hoschek <whoschek AT lbl.gov>
To: xom-interest AT lists.ibiblio.org
Subject: [XOM-interest] [PATCH] Text using java.nio.charset.Charset
Date: Wed, 30 Jun 2004 14:05:31 -0700

Elliotte,

I heard that the ByteToCharConverter used internally by java.lang.String when converting from and to bytes is rather inefficient, and that the new jdk-1.4 Charset coders are much more efficient.

So if you are currently doing benchmarking as seems to be indicated in the last release notes you might want to check if the attached patch or a derivative of it does any good in terms of performance. Unfortunately I'm too busy with tons of other stuff to try and see myself.
Search for changed lines ending with "// WH".

(Note that Charset.encode and decode are thread-safe).

Wolfgang.

// Copyright 2002-2004 Elliotte Rusty Harold
//
// This library is free software; you can redistribute
// it and/or modify it under the terms of version 2.1 of
// the GNU Lesser General Public License as published by
// the Free Software Foundation.
//
// This library is distributed in the hope that it will be useful,
// but WITHOUT ANY WARRANTY; without even the implied warranty of
// MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
// GNU Lesser General Public License for more details.
//
// You should have received a copy of the GNU Lesser General
// Public License along with this library; if not, write to the
// Free Software Foundation, Inc., 59 Temple Place, Suite 330,
// Boston, MA 02111-1307 USA
//
// You can contact Elliotte Rusty Harold by sending e-mail to
// elharo AT metalab.unc.edu. Please include the word "XOM" in the
// subject line. The XOM home page is temporarily located at
// http://www.cafeconleche.org/XOM/ but will eventually move
// to http://www.xom.nu/

package nu.xom;

//import java.io.UnsupportedEncodingException; // WH
import java.nio.ByteBuffer; // WH
import java.nio.charset.Charset; // WH

/**
* 
* This class represents a run of text.
* CDATA sections are not treated differently than
* normal text.
* 

* @author Elliotte Rusty Harold
* @version 1.0a1
*
*/
public class Text extends Node {

private ByteBuffer data; // WH
private static final Charset CHARSET = Charset.forName("UTF-8"); //
WH
//private byte[] data;

/**
* 
* This constructor creates a new <code>Text</code> object.
* The data is checked for legality according to XML 1.0 rules.
* Characters that can be serialized by escaping them
* such as < and & are allowed. However, characters
* such as the form feed, null, vertical tab,
* unmatched halves of surrogate pairs,
* and 0xFFFE and 0xFFFF are not allowed.
* 
*
* @param data the initial text of the object
*
* @throws IllegalCharacterDataException if data contains any
* characters which are illegal in well-formed XML 1.0 such as
* null, vertical tab, or unmatched halves of surrogate pairs
*/
public Text(String data) {
_setValue(data);
}

/**
* 
* Creates a copy of the specified <code>Text</code> object.
* 
*
* @param text the <code>Text</code> object to copy
*/
public Text(Text text) {
// I'm relying here on the data array being immutable.
// If this ever changes, e.g. by adding an append method,
// this method needs to change too.
this.data = text.data;
}

private Text() {}

static Text build(String data) {
Text result = new Text();
result.data = CHARSET.encode(data); // WH
// try {
// result.data = data.getBytes("UTF8");
// }
// catch (UnsupportedEncodingException ex) {
// throw new RuntimeException(
// "Bad VM! Does not support UTF-8"
// );
// }
return result;
}

/**
* 
* This sets the content of the <code>Text</code> object
* to the specified data. The data is checked for
* legality according to XML 1.0 rules. Characters that
* can be serialized such as < and & are allowed.
* However, characters such as the form feed, null,
* vertical tab, unmatched halves of surrogate pairs,
* and 0xFFFE and 0xFFFF are not allowed.
* 
*
* @param data the text to install in the object
*
* @throws IllegalCharacterDataException if data contains any
* characters which are illegal in well-formed XML 1.0 such as
* null, vertical tab, or unmatched halves of surrogate pairs
*/
public void setValue(String data) {
_setValue(data);
}

private void _setValue(String data) {
if (data == null) data = "";
else Verifier.checkPCDATA(data);
this.data = CHARSET.encode(data); // WH
// try {
// this.data = data.getBytes("UTF8");
// }
// catch (UnsupportedEncodingException ex) {
// throw new RuntimeException(
// "Bad VM! Does not support UTF-8"
// );
// }
}

/**
* 
* Returns the XPath 1.0 string-value of this <code>Text</code>
* node. The XPath string-value of a text node is the same as
* the text of the node.
* 
*
* @return The content of the node.
*
* @see nu.xom.Node#getValue()
*/
public final String getValue() {

return CHARSET.decode(data.duplicate()).toString(); // WH
// try {
// return new String(data, "UTF8");
// }
// catch (UnsupportedEncodingException ex) {
// throw new RuntimeException(
// "Bad VM! Does not support UTF-8"
// );
// }
}

/**
* 
* Throws <code>IndexOutOfBoundsException</code> because
* texts do not have children.
* 
*
* @return never returns because texts do not have children;
* always throws an exception.
*
* @param position the index of the child node to return
*
* @throws IndexOutOfBoundsException because texts
* do not have children
*/
public final Node getChild(int position) {
throw new IndexOutOfBoundsException(
"LeafNodes do not have children");
}

/**
* 
* Returns 0 because texts do not have children.
* 
*
* @return zero
*
* @see nu.xom.Node#getChildCount()
*/
public final int getChildCount() {
return 0;
}

/**
* 
* Returns a deep copy of this <code>Text</code> with no parent,
* that can be added to this document or a different one.
* 
*
* @return a deep copy of this text node with no parent
*
* @see nu.xom.Node#copy()
*/
public Node copy() {
if (isCDATASection()) {
return new CDATASection(this);
}
else {
return new Text(this);
}
}

/**
* 
* Returns a string containing the XML serialization of this text
* node. Unlike <code>getValue</code>, this method escapes
* characters such as & and < using entity references such
* as <code>&amp;</code> and <code>&lt;</code>.
* It escapes the carriage return (\r) as <code>&#x0D;</code>.
* 
*
* @return the string form of this text node
*
* @see nu.xom.Node#toXML()
*/
public final String toXML() {
return escapeText(getValue());
}

private static String escapeText(String s) {

StringBuffer result = new StringBuffer(s.length());
for (int i = 0; i < s.length(); i++) {
char c = s.charAt(i);
switch (c) {
case '<':
result.append("<");
break;
case '>':
result.append(">");
break;
case '&':
result.append("&");
break;
case '\r':
result.append("");
break;
default:
result.append(c);
}
}
return result.toString();
}

boolean isText() {
return true;
}

/**
* 
* Returns a <code>String</code>
* representation of this <code>Text</code> suitable for
* debugging and diagnosis. This is not
* the XML representation of this <code>Text</code> node.
* 
*
* @return a non-XML string representation of this node
*
* @see java.lang.Object#toString()
*/
public final String toString() {

String value = getValue();
if (value.length() <= 40) {
return "[" + getClass().getName() + ": " + value + "]";
}

return "[" + getClass().getName() + ": "
+ value.substring(0, 35) + "...]";

}

boolean isCDATASection() {
return false;
}

}

[XOM-interest] [PATCH] Text using java.nio.charset.Charset, Wolfgang Hoschek, 06/30/2004
- [XOM-interest] Re: [PATCH] Text using java.nio.charset.Charset, Wolfgang Hoschek, 06/30/2004