创造一种迅速而又随性的(quick and dirty)xml解释器

核心提示：From javaworld.com 翻译by matrix开源技术-chris xml是一种当前很受欢迎的数据格式, 它的优点在于: 人性化,自述性以及使用的方便性.但是,不幸的是,基于java的xml解释器往往太大了,比如sun的jaXP.jar 和 parser.jar 每个都达到了1.4mb. 假如你

　　From javaworld.com 翻译by matrix开源技术-chris

xml是一种当前很受欢迎的数据格式, 它的优点在于: 人性化,自述性以及使用的方便性.但是,不幸的是,基于java的xml解释器往往太大了,比如sun的jaXP.jar 和 parser.jar 每个都达到了1.4mb. 假如你要在只有有限的内存容量的运行环境里运行你的程序,比如j2me的环境.或者说带宽很有限的运行环境里,比如applet,这些大的package不应该成为你的选择对象.
注重:本篇的所有所需要的所有代码你可以通过此链接下载:
http://www.matrix.org.cn/down_view.asp?id=67
下面是QDParser的代码:
package qdxml;
import java.io.*;
import java.util.*;

/** Quick and Dirty xml parser. This parser is, like the SAX parser,
an event based parser, but with mUCh less functionality. */
public class QDParser {
PRivate static int popMode(Stack st) {
if(!st.empty())
return ((Integer)st.pop()).intvalue();
else
return PRE;
}
private final static int
TEXT = 1,
ENTITY = 2,
OPEN_TAG = 3,
CLOSE_TAG = 4,
START_TAG = 5,
ATTRIBUTE_Lvalue = 6,
ATTRIBUTE_EQUAL = 9,
ATTRIBUTE_Rvalue = 10,
QUOTE = 7,
IN_TAG = 8,
SINGLE_TAG = 12,
COMMENT = 13,
DONE = 11,
DOCTYPE = 14,
PRE = 15,
CDATA = 16;
public static void parse(DocHandler doc,Reader r) throws Exception {
Stack st = new Stack();
int depth = 0;
int mode = PRE;
int c = 0;
int quotec = ´"´;
depth = 0;
StringBuffer sb = new StringBuffer();
StringBuffer etag = new StringBuffer();
String tagName = null;
String lvalue = null;
String rvalue = null;
Hashtable attrs = null;
st = new Stack();
doc.startdocument．);
int line=1, col=0;
boolean eol = false;
while((c = r.read()) != -1) {

// We need to map , , and to
// See XML spec section 2.11
if(c == ´ ´ && eol) {
eol = false;
continue;
} else if(eol) {
eol = false;
} else if(c == ´ ´) {
line++;
col=0;
} else if(c == ´´) {
eol = true;
c = ´ ´;
line++;
col=0;
} else {
col++;
}

if(mode == DONE) {
doc.enddocument．);
return;

// We are between tags collecting text.
} else if(mode == TEXT) {
if(c == ´<´) {
st.push(new Integer(mode));
mode = START_TAG;
if(sb.length() > 0) {
doc.text(sb.toString());
sb.setLength(0);
}
} else if(c == ´&´) {
st.push(new Integer(mode));
mode = ENTITY;
etag.setLength(0);
} else
sb.append((char)c);

// we are processing a closing tag: e.g. </foo>
} else if(mode == CLOSE_TAG) {
if(c == ´>´) {
mode = popMode(st);
tagName = sb.toString();
sb.setLength(0);
depth--;
if(depth==0)
mode = DONE;
doc.endElement(tagName);
} else {
sb.append((char)c);
}

// we are processing CDATA
} else if(mode == CDATA) {
if(c == ´>´
&& sb.toString().endsWith("]]")) {
sb.setLength(sb.length()-2);
doc.text(sb.toString());
sb.setLength(0);
mode = popMode(st);
} else
sb.append((char)c);

// we are processing a comment. We are inside
// the <!-- .... --> looking for the -->.
} else if(mode == COMMENT) {
if(c == ´>´
&& sb.toString().endsWith("--")) {
sb.setLength(0);
mode = popMode(st);
} else
sb.append((char)c);

// We are outside the root tag element
} else if(mode == PRE) {
if(c == ´<´) {
mode = TEXT;
st.push(new Integer(mode));
mode = START_TAG;
}

// We are inside one of these <? ... ?>
// or one of these <!DOCTYPE ... >
} else if(mode == DOCTYPE) {
if(c == ´>´) {
mode = popMode(st);
if(mode == TEXT) mode = PRE;
}

// we have just seen a < and
// are wondering what we are looking at
// <foo>, </foo>, <!-- ... --->, etc.
} else if(mode == START_TAG) {
mode = popMode(st);
if(c == ´/´) {
st.push(new Integer(mode));
mode = CLOSE_TAG;
} else if (c == ´?´) {
mode = DOCTYPE;
} else {
st.push(new Integer(mode));
mode = OPEN_TAG;
tagName = null;
attrs = new Hashtable();
sb.append((char)c);
}

// we are processing an entity, e.g. <, », etc.
} else if(mode == ENTITY) {
if(c == ´;´) {
mode = popMode(st);
String cent = etag.toString();
etag.setLength(0);
if(cent.equals("lt"))
sb.append(´<´);
else if(cent.equals("gt"))
sb.append(´>´);
else if(cent.equals("amp"))
sb.append(´&´);
else if(cent.equals("quot"))
sb.append(´"´);
else if(cent.equals("apos"))
sb.append(´´´);
// Could parse hex entities if we wanted to
//else if(cent.startsWith("#x"))
//sb.append((char)Integer.parseInt(cent.substring(2),16));
else if(cent.startsWith("#"))
sb.append((char)Integer.parseInt(cent.substring(1)));
// Insert custom entity definitions here
else
exc("Unknown entity: &"+cent+";",line,col);
} else {
etag.append((char)c);
}

// we have just seen something like this:
// <foo a="b"/
// and are looking for the final >.
} else if(mode == SINGLE_TAG) {
if(tagName == null)
tagName = sb.toString();
if(c != ´>´)
exc("Expected > for tag: <"+tagName+"/>",line,col);
doc.startElement(tagName,attrs);
doc.endElement(tagName);
if(depth==0) {
doc.enddocument．);
return;
}
sb.setLength(0);
attrs = new Hashtable();
tagName = null;
mode = popMode(st);

// we are processing something
// like this <foo ... >. It could
// still be a <!-- ... --> or something.
} else if(mode == OPEN_TAG) {
if(c == ´>´) {
if(tagName == null)
tagName = sb.toString();
sb.setLength(0);
depth++;
doc.startElement(tagName,attrs);
tagName = null;
attrs = new Hashtable();
mode = popMode(st);
} else if(c == ´/´) {
mode = SINGLE_TAG;
} else if(c == ´-´ && sb.toString().equals("!-")) {
mode = COMMENT;
} else if(c == ´[´ && sb.toString().equals("![CDATA")) {
mode = CDATA;
sb.setLength(0);
} else if(c == ´E´ && sb.toString().equals("!DOCTYP")) {
sb.setLength(0);
mode = DOCTYPE;
} else if(Character.isWhitespace((char)c)) {
tagName = sb.toString();
sb.setLength(0);
mode = IN_TAG;
} else {
sb.append((char)c);
}

// We are processing the quoted right-hand side
// of an element´s attribute.
} else if(mode == QUOTE) {
if(c == quotec) {
rvalue = sb.toString();
sb.setLength(0);
attrs.put(lvalue,rvalue);
mode = IN_TAG;
// See section the XML spec, section 3.3.3
// on normalization processing.
} else if("