我是很大的字……我居然比旁边的字小我最小...啊啊啊......居然可以设置不同的字体字号
我先来个居中对齐!
我是来右对齐的!
我可以设置很多不同的字体风格,比如:加粗、斜体、下划线。
我可以设置一个超链接,牛逼吗 快戳我看看
我是很大的字……我居然比旁边的字小我最小...啊啊啊......居然可以设置不同的字体字号
我先来个居中对齐!
我是来右对齐的!
我可以设置很多不同的字体风格,比如:加粗、斜体、下划线。
我可以设置一个超链接,牛逼吗 快戳我看看
SAXParser using the currently
65 | * configured factory parameters.
66 | */
67 | public SAXParser newSAXParser()
68 | throws ParserConfigurationException {
69 | try {
70 | return SAXParserImpl.newInstance(features);
71 | } catch (SAXException se) {
72 | // Translate to ParserConfigurationException
73 | throw new ParserConfigurationException(se.getMessage());
74 | }
75 | }
76 |
77 | /**
78 | * Defines that the specified feature is to enabled/disabled (as
79 | * per second argument) on reader instances created by this
80 | * factory.
81 | */
82 | public void setFeature(String name, boolean value)
83 | throws ParserConfigurationException, SAXNotRecognizedException,
84 | SAXNotSupportedException {
85 | // First, let's see if it's a valid call
86 | getPrototype().setFeature(name, value);
87 |
88 | // If not, exception was thrown: so we are good now:
89 | if (features == null) {
90 | // Let's retain the ordering as well
91 | features = new LinkedHashMap();
92 | }
93 | features.put(name, value ? Boolean.TRUE : Boolean.FALSE);
94 | }
95 |
96 | /**
97 | * Returns whether the specified property will be enabled or disabled
98 | * on reader instances constructed by this factory.
99 | */
100 | public boolean getFeature(String name)
101 | throws ParserConfigurationException, SAXNotRecognizedException,
102 | SAXNotSupportedException {
103 | return getPrototype().getFeature(name);
104 | }
105 |
106 | // // // Internal methods
107 |
108 | private SAXParserImpl getPrototype() {
109 | if (prototypeParser == null) {
110 | prototypeParser = new SAXParserImpl();
111 | }
112 | return prototypeParser;
113 | }
114 | }
115 |
--------------------------------------------------------------------------------
/lib_tv_fullhtml/src/main/java/com/xys/lib_tv_fullhtml/tagsoup/PYXScanner.java:
--------------------------------------------------------------------------------
1 | // This file is part of TagSoup and is Copyright 2002-2008 by John Cowan.
2 | //
3 | // TagSoup is licensed under the Apache License,
4 | // Version 2.0. You may obtain a copy of this license at
5 | // http://www.apache.org/licenses/LICENSE-2.0 . You may also have
6 | // additional legal rights not granted by this license.
7 | //
8 | // TagSoup is distributed in the hope that it will be useful, but
9 | // unless required by applicable law or agreed to in writing, TagSoup
10 | // is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS
11 | // OF ANY KIND, either express or implied; not even the implied warranty
12 | // of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
13 | //
14 | //
15 | // This file is part of TagSoup.
16 | //
17 | // This program is free software; you can redistribute it and/or modify
18 | // it under the terms of the GNU General Public License as published by
19 | // the Free Software Foundation; either version 2 of the License, or
20 | // (at your option) any later version. You may also distribute
21 | // and/or modify it under version 2.1 of the Academic Free License.
22 | //
23 | // This program is distributed in the hope that it will be useful,
24 | // but WITHOUT ANY WARRANTY; without even the implied warranty of
25 | // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
26 | //
27 | //
28 | // PYX Scanner
29 |
30 | package com.xys.lib_tv_fullhtml.tagsoup;
31 |
32 | import org.xml.sax.SAXException;
33 |
34 | import java.io.BufferedReader;
35 | import java.io.BufferedWriter;
36 | import java.io.IOException;
37 | import java.io.InputStreamReader;
38 | import java.io.OutputStreamWriter;
39 | import java.io.Reader;
40 | import java.io.Writer;
41 |
42 | /**
43 | * A Scanner that accepts PYX format instead of HTML.
44 | * Useful primarily for debugging.
45 | **/
46 | public class PYXScanner implements Scanner {
47 |
48 | public static void main(String[] argv) throws IOException, SAXException {
49 | Scanner s = new PYXScanner();
50 | Reader r = new InputStreamReader(System.in, "UTF-8");
51 | Writer w = new BufferedWriter(new OutputStreamWriter(System.out, "UTF-8"));
52 | s.scan(r, new PYXWriter(w));
53 | }
54 |
55 | public void resetDocumentLocator(String publicid, String systemid) {
56 | // Need this method for interface compatibility, but note
57 | // that PyxScanner does not implement Locator.
58 | }
59 |
60 | public void scan(Reader r, ScanHandler h) throws IOException, SAXException {
61 | BufferedReader br = new BufferedReader(r);
62 | String s;
63 | char[] buff = null;
64 | boolean instag = false;
65 | while ((s = br.readLine()) != null) {
66 | int size = s.length();
67 | if (buff == null || buff.length < size) {
68 | buff = new char[size];
69 | }
70 | s.getChars(0, size, buff, 0);
71 | switch (buff[0]) {
72 | case '(':
73 | if (instag) {
74 | h.stagc(buff, 0, 0);
75 | instag = false;
76 | }
77 | h.gi(buff, 1, size - 1);
78 | instag = true;
79 | break;
80 | case ')':
81 | if (instag) {
82 | h.stagc(buff, 0, 0);
83 | instag = false;
84 | }
85 | h.etag(buff, 1, size - 1);
86 | break;
87 | case '?':
88 | if (instag) {
89 | h.stagc(buff, 0, 0);
90 | instag = false;
91 | }
92 | h.pi(buff, 1, size - 1);
93 | break;
94 | case 'A':
95 | int sp = s.indexOf(' ');
96 | h.aname(buff, 1, sp - 1);
97 | h.aval(buff, sp + 1, size - sp - 1);
98 | break;
99 | case '-':
100 | if (instag) {
101 | h.stagc(buff, 0, 0);
102 | instag = false;
103 | }
104 | if (s.equals("-\\n")) {
105 | buff[0] = '\n';
106 | h.pcdata(buff, 0, 1);
107 | } else {
108 | // FIXME:
109 | // Does not decode \t and \\ in input
110 | h.pcdata(buff, 1, size - 1);
111 | }
112 | break;
113 | case 'E':
114 | if (instag) {
115 | h.stagc(buff, 0, 0);
116 | instag = false;
117 | }
118 | h.entity(buff, 1, size - 1);
119 | break;
120 | default:
121 | // System.err.print("Gotcha ");
122 | // System.err.print(s);
123 | // System.err.print('\n');
124 | break;
125 | }
126 | }
127 | h.eof(buff, 0, 0);
128 | }
129 |
130 | public void startCDATA() {
131 | }
132 | }
133 |
--------------------------------------------------------------------------------
/gradlew:
--------------------------------------------------------------------------------
1 | #!/usr/bin/env bash
2 |
3 | ##############################################################################
4 | ##
5 | ## Gradle start up script for UN*X
6 | ##
7 | ##############################################################################
8 |
9 | # Add default JVM options here. You can also use JAVA_OPTS and GRADLE_OPTS to pass JVM options to this script.
10 | DEFAULT_JVM_OPTS=""
11 |
12 | APP_NAME="Gradle"
13 | APP_BASE_NAME=`basename "$0"`
14 |
15 | # Use the maximum available, or set MAX_FD != -1 to use that value.
16 | MAX_FD="maximum"
17 |
18 | warn ( ) {
19 | echo "$*"
20 | }
21 |
22 | die ( ) {
23 | echo
24 | echo "$*"
25 | echo
26 | exit 1
27 | }
28 |
29 | # OS specific support (must be 'true' or 'false').
30 | cygwin=false
31 | msys=false
32 | darwin=false
33 | case "`uname`" in
34 | CYGWIN* )
35 | cygwin=true
36 | ;;
37 | Darwin* )
38 | darwin=true
39 | ;;
40 | MINGW* )
41 | msys=true
42 | ;;
43 | esac
44 |
45 | # Attempt to set APP_HOME
46 | # Resolve links: $0 may be a link
47 | PRG="$0"
48 | # Need this for relative symlinks.
49 | while [ -h "$PRG" ] ; do
50 | ls=`ls -ld "$PRG"`
51 | link=`expr "$ls" : '.*-> \(.*\)$'`
52 | if expr "$link" : '/.*' > /dev/null; then
53 | PRG="$link"
54 | else
55 | PRG=`dirname "$PRG"`"/$link"
56 | fi
57 | done
58 | SAVED="`pwd`"
59 | cd "`dirname \"$PRG\"`/" >/dev/null
60 | APP_HOME="`pwd -P`"
61 | cd "$SAVED" >/dev/null
62 |
63 | CLASSPATH=$APP_HOME/gradle/wrapper/gradle-wrapper.jar
64 |
65 | # Determine the Java command to use to start the JVM.
66 | if [ -n "$JAVA_HOME" ] ; then
67 | if [ -x "$JAVA_HOME/jre/sh/java" ] ; then
68 | # IBM's JDK on AIX uses strange locations for the executables
69 | JAVACMD="$JAVA_HOME/jre/sh/java"
70 | else
71 | JAVACMD="$JAVA_HOME/bin/java"
72 | fi
73 | if [ ! -x "$JAVACMD" ] ; then
74 | die "ERROR: JAVA_HOME is set to an invalid directory: $JAVA_HOME
75 |
76 | Please set the JAVA_HOME variable in your environment to match the
77 | location of your Java installation."
78 | fi
79 | else
80 | JAVACMD="java"
81 | which java >/dev/null 2>&1 || die "ERROR: JAVA_HOME is not set and no 'java' command could be found in your PATH.
82 |
83 | Please set the JAVA_HOME variable in your environment to match the
84 | location of your Java installation."
85 | fi
86 |
87 | # Increase the maximum file descriptors if we can.
88 | if [ "$cygwin" = "false" -a "$darwin" = "false" ] ; then
89 | MAX_FD_LIMIT=`ulimit -H -n`
90 | if [ $? -eq 0 ] ; then
91 | if [ "$MAX_FD" = "maximum" -o "$MAX_FD" = "max" ] ; then
92 | MAX_FD="$MAX_FD_LIMIT"
93 | fi
94 | ulimit -n $MAX_FD
95 | if [ $? -ne 0 ] ; then
96 | warn "Could not set maximum file descriptor limit: $MAX_FD"
97 | fi
98 | else
99 | warn "Could not query maximum file descriptor limit: $MAX_FD_LIMIT"
100 | fi
101 | fi
102 |
103 | # For Darwin, add options to specify how the application appears in the dock
104 | if $darwin; then
105 | GRADLE_OPTS="$GRADLE_OPTS \"-Xdock:name=$APP_NAME\" \"-Xdock:icon=$APP_HOME/media/gradle.icns\""
106 | fi
107 |
108 | # For Cygwin, switch paths to Windows format before running java
109 | if $cygwin ; then
110 | APP_HOME=`cygpath --path --mixed "$APP_HOME"`
111 | CLASSPATH=`cygpath --path --mixed "$CLASSPATH"`
112 | JAVACMD=`cygpath --unix "$JAVACMD"`
113 |
114 | # We build the pattern for arguments to be converted via cygpath
115 | ROOTDIRSRAW=`find -L / -maxdepth 1 -mindepth 1 -type d 2>/dev/null`
116 | SEP=""
117 | for dir in $ROOTDIRSRAW ; do
118 | ROOTDIRS="$ROOTDIRS$SEP$dir"
119 | SEP="|"
120 | done
121 | OURCYGPATTERN="(^($ROOTDIRS))"
122 | # Add a user-defined pattern to the cygpath arguments
123 | if [ "$GRADLE_CYGPATTERN" != "" ] ; then
124 | OURCYGPATTERN="$OURCYGPATTERN|($GRADLE_CYGPATTERN)"
125 | fi
126 | # Now convert the arguments - kludge to limit ourselves to /bin/sh
127 | i=0
128 | for arg in "$@" ; do
129 | CHECK=`echo "$arg"|egrep -c "$OURCYGPATTERN" -`
130 | CHECK2=`echo "$arg"|egrep -c "^-"` ### Determine if an option
131 |
132 | if [ $CHECK -ne 0 ] && [ $CHECK2 -eq 0 ] ; then ### Added a condition
133 | eval `echo args$i`=`cygpath --path --ignore --mixed "$arg"`
134 | else
135 | eval `echo args$i`="\"$arg\""
136 | fi
137 | i=$((i+1))
138 | done
139 | case $i in
140 | (0) set -- ;;
141 | (1) set -- "$args0" ;;
142 | (2) set -- "$args0" "$args1" ;;
143 | (3) set -- "$args0" "$args1" "$args2" ;;
144 | (4) set -- "$args0" "$args1" "$args2" "$args3" ;;
145 | (5) set -- "$args0" "$args1" "$args2" "$args3" "$args4" ;;
146 | (6) set -- "$args0" "$args1" "$args2" "$args3" "$args4" "$args5" ;;
147 | (7) set -- "$args0" "$args1" "$args2" "$args3" "$args4" "$args5" "$args6" ;;
148 | (8) set -- "$args0" "$args1" "$args2" "$args3" "$args4" "$args5" "$args6" "$args7" ;;
149 | (9) set -- "$args0" "$args1" "$args2" "$args3" "$args4" "$args5" "$args6" "$args7" "$args8" ;;
150 | esac
151 | fi
152 |
153 | # Split up the JVM_OPTS And GRADLE_OPTS values into an array, following the shell quoting and substitution rules
154 | function splitJvmOpts() {
155 | JVM_OPTS=("$@")
156 | }
157 | eval splitJvmOpts $DEFAULT_JVM_OPTS $JAVA_OPTS $GRADLE_OPTS
158 | JVM_OPTS[${#JVM_OPTS[*]}]="-Dorg.gradle.appname=$APP_BASE_NAME"
159 |
160 | exec "$JAVACMD" "${JVM_OPTS[@]}" -classpath "$CLASSPATH" org.gradle.wrapper.GradleWrapperMain "$@"
161 |
--------------------------------------------------------------------------------
/lib_tv_fullhtml/src/main/java/com/xys/lib_tv_fullhtml/tagsoup/Schema.java:
--------------------------------------------------------------------------------
1 | // This file is part of TagSoup and is Copyright 2002-2008 by John Cowan.
2 | //
3 | // TagSoup is licensed under the Apache License,
4 | // Version 2.0. You may obtain a copy of this license at
5 | // http://www.apache.org/licenses/LICENSE-2.0 . You may also have
6 | // additional legal rights not granted by this license.
7 | //
8 | // TagSoup is distributed in the hope that it will be useful, but
9 | // unless required by applicable law or agreed to in writing, TagSoup
10 | // is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS
11 | // OF ANY KIND, either express or implied; not even the implied warranty
12 | // of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
13 | //
14 | //
15 | // Model of document
16 |
17 | package com.xys.lib_tv_fullhtml.tagsoup;
18 |
19 | import java.util.HashMap;
20 |
21 | /**
22 | * Abstract class representing a TSSL schema.
23 | * Actual TSSL schemas are compiled into concrete subclasses of this class.
24 | **/
25 |
26 | public abstract class Schema {
27 |
28 | public static final int M_ANY = 0xFFFFFFFF;
29 | public static final int M_EMPTY = 0;
30 | public static final int M_PCDATA = 1 << 30;
31 | public static final int M_ROOT = 1 << 31;
32 |
33 |
34 | public static final int F_RESTART = 1;
35 | public static final int F_CDATA = 2;
36 | public static final int F_NOFORCE = 4;
37 |
38 | private HashMap theEntities =
39 | new HashMap(); // String -> Character
40 | private HashMap theElementTypes =
41 | new HashMap(); // String -> ElementType
42 |
43 | private String theURI = "";
44 | private String thePrefix = "";
45 | private ElementType theRoot = null;
46 |
47 | /**
48 | * Add or replace an element type for this schema.
49 | *
50 | * @param name Name (Qname) of the element
51 | * @param model Models of the element's content as a vector of bits
52 | * @param memberOf Models the element is a member of as a vector of bits
53 | * @param flags Flags for the element
54 | **/
55 |
56 | public void elementType(String name, int model, int memberOf, int flags) {
57 | ElementType e = new ElementType(name, model, memberOf, flags, this);
58 | theElementTypes.put(name.toLowerCase(), e);
59 | if (memberOf == M_ROOT) theRoot = e;
60 | }
61 |
62 | /**
63 | * Get the root element of this schema
64 | **/
65 |
66 | public ElementType rootElementType() {
67 | return theRoot;
68 | }
69 |
70 | /**
71 | * Add or replace a default attribute for an element type in this schema.
72 | *
73 | * @param elemName Name (Qname) of the element type
74 | * @param attrName Name (Qname) of the attribute
75 | * @param type Type of the attribute
76 | * @param value Default value of the attribute; null if no default
77 | **/
78 |
79 | public void attribute(String elemName, String attrName,
80 | String type, String value) {
81 | ElementType e = getElementType(elemName);
82 | if (e == null) {
83 | throw new Error("Attribute " + attrName +
84 | " specified for unknown element type " +
85 | elemName);
86 | }
87 | e.setAttribute(attrName, type, value);
88 | }
89 |
90 | /**
91 | * Specify natural parent of an element in this schema.
92 | *
93 | * @param name Name of the child element
94 | * @param parentName Name of the parent element
95 | **/
96 |
97 | public void parent(String name, String parentName) {
98 | ElementType child = getElementType(name);
99 | ElementType parent = getElementType(parentName);
100 | if (child == null) {
101 | throw new Error("No child " + name + " for parent " + parentName);
102 | }
103 | if (parent == null) {
104 | throw new Error("No parent " + parentName + " for child " + name);
105 | }
106 | child.setParent(parent);
107 | }
108 |
109 | /**
110 | * Add to or replace a character entity in this schema.
111 | *
112 | * @param name Name of the entity
113 | * @param value Value of the entity
114 | **/
115 |
116 | public void entity(String name, int value) {
117 | theEntities.put(name, new Integer(value));
118 | }
119 |
120 | /**
121 | * Get an ElementType by name.
122 | *
123 | * @param name Name (Qname) of the element type
124 | * @return The corresponding ElementType
125 | **/
126 |
127 | public ElementType getElementType(String name) {
128 | return (ElementType) (theElementTypes.get(name.toLowerCase()));
129 | }
130 |
131 | /**
132 | * Get an entity value by name.
133 | *
134 | * @param name Name of the entity
135 | * @return The corresponding character, or 0 if none
136 | **/
137 |
138 | public int getEntity(String name) {
139 | // System.err.println("%% Looking up entity " + name);
140 | Integer ch = (Integer) theEntities.get(name);
141 | if (ch == null) return 0;
142 | return ch.intValue();
143 | }
144 |
145 | /**
146 | * Return the URI (namespace name) of this schema.
147 | **/
148 |
149 | public String getURI() {
150 | return theURI;
151 | }
152 |
153 | /**
154 | * Change the URI (namespace name) of this schema.
155 | **/
156 |
157 | public void setURI(String uri) {
158 | theURI = uri;
159 | }
160 |
161 | /**
162 | * Return the prefix of this schema.
163 | **/
164 |
165 | public String getPrefix() {
166 | return thePrefix;
167 | }
168 |
169 | /**
170 | * Change the prefix of this schema.
171 | **/
172 |
173 | public void setPrefix(String prefix) {
174 | thePrefix = prefix;
175 | }
176 |
177 | }
178 |
--------------------------------------------------------------------------------
/lib_tv_fullhtml/src/main/java/com/xys/lib_tv_fullhtml/tagsoup/jaxp/SAX1ParserAdapter.java:
--------------------------------------------------------------------------------
1 | // This file is part of TagSoup and is Copyright 2002-2008 by John Cowan.
2 | //
3 | // TagSoup is licensed under the Apache License,
4 | // Version 2.0. You may obtain a copy of this license at
5 | // http://www.apache.org/licenses/LICENSE-2.0 . You may also have
6 | // additional legal rights not granted by this license.
7 | //
8 | // TagSoup is distributed in the hope that it will be useful, but
9 | // unless required by applicable law or agreed to in writing, TagSoup
10 | // is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS
11 | // OF ANY KIND, either express or implied; not even the implied warranty
12 | // of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
13 |
14 | package com.xys.lib_tv_fullhtml.tagsoup.jaxp;
15 |
16 | import org.xml.sax.AttributeList;
17 | import org.xml.sax.Attributes;
18 | import org.xml.sax.ContentHandler;
19 | import org.xml.sax.DTDHandler;
20 | import org.xml.sax.DocumentHandler;
21 | import org.xml.sax.EntityResolver;
22 | import org.xml.sax.ErrorHandler;
23 | import org.xml.sax.InputSource;
24 | import org.xml.sax.Locator;
25 | import org.xml.sax.SAXException;
26 | import org.xml.sax.SAXNotSupportedException;
27 | import org.xml.sax.XMLReader;
28 |
29 | import java.io.IOException;
30 |
31 | /**
32 | * This is a simpler adapter class that allows using SAX1 interface on top
33 | * of basic SAX2 implementation, such as TagSoup.
34 | *
35 | * @author Tatu Saloranta (cowtowncoder@yahoo.com)
36 | * @deprecated
37 | */
38 | public class SAX1ParserAdapter
39 | implements org.xml.sax.Parser {
40 | final XMLReader xmlReader;
41 |
42 | public SAX1ParserAdapter(XMLReader xr) {
43 | xmlReader = xr;
44 | }
45 |
46 | // Sax1 API impl
47 |
48 | public void parse(InputSource source)
49 | throws SAXException {
50 | try {
51 | xmlReader.parse(source);
52 | } catch (IOException ioe) {
53 | throw new SAXException(ioe);
54 | }
55 | }
56 |
57 | public void parse(String systemId)
58 | throws SAXException {
59 | try {
60 | xmlReader.parse(systemId);
61 | } catch (IOException ioe) {
62 | throw new SAXException(ioe);
63 | }
64 | }
65 |
66 | /**
67 | * @deprecated
68 | */
69 | public void setDocumentHandler(DocumentHandler h) {
70 | xmlReader.setContentHandler(new DocHandlerWrapper(h));
71 | }
72 |
73 | public void setDTDHandler(DTDHandler h) {
74 | xmlReader.setDTDHandler(h);
75 | }
76 |
77 | public void setEntityResolver(EntityResolver r) {
78 | xmlReader.setEntityResolver(r);
79 | }
80 |
81 | public void setErrorHandler(ErrorHandler h) {
82 | xmlReader.setErrorHandler(h);
83 | }
84 |
85 | public void setLocale(java.util.Locale locale)
86 | throws SAXException {
87 | /* I have no idea what this is supposed to do... so let's
88 | * throw an exception
89 | */
90 | throw new SAXNotSupportedException("TagSoup does not implement setLocale() method");
91 | }
92 |
93 | // Helper classes:
94 |
95 | /**
96 | * We need another helper class to deal with differences between
97 | * Sax2 handler (content handler), and Sax1 handler (document handler)
98 | *
99 | * @deprecated
100 | */
101 | final static class DocHandlerWrapper
102 | implements ContentHandler {
103 | final DocumentHandler docHandler;
104 |
105 | final AttributesWrapper mAttrWrapper = new AttributesWrapper();
106 |
107 | /**
108 | * @deprecated
109 | */
110 | DocHandlerWrapper(DocumentHandler h) {
111 | docHandler = h;
112 | }
113 |
114 | public void characters(char[] ch, int start, int length)
115 | throws SAXException {
116 | docHandler.characters(ch, start, length);
117 | }
118 |
119 | public void endDocument()
120 | throws SAXException {
121 | docHandler.endDocument();
122 | }
123 |
124 | public void endElement(String uri, String localName, String qName)
125 | throws SAXException {
126 | if (qName == null) {
127 | qName = localName;
128 | }
129 | docHandler.endElement(qName);
130 | }
131 |
132 | public void endPrefixMapping(String prefix) {
133 | // no equivalent in SAX1, ignore
134 | }
135 |
136 | public void ignorableWhitespace(char[] ch, int start, int length)
137 | throws SAXException {
138 | docHandler.ignorableWhitespace(ch, start, length);
139 | }
140 |
141 | public void processingInstruction(String target, String data)
142 | throws SAXException {
143 | docHandler.processingInstruction(target, data);
144 | }
145 |
146 | public void setDocumentLocator(Locator locator) {
147 | docHandler.setDocumentLocator(locator);
148 | }
149 |
150 | public void skippedEntity(String name) {
151 | // no equivalent in SAX1, ignore
152 | }
153 |
154 | public void startDocument()
155 | throws SAXException {
156 | docHandler.startDocument();
157 | }
158 |
159 | public void startElement(String uri, String localName, String qName,
160 | Attributes attrs)
161 | throws SAXException {
162 | if (qName == null) {
163 | qName = localName;
164 | }
165 | // Also, need to wrap Attributes to look like AttributeLost
166 | mAttrWrapper.setAttributes(attrs);
167 | docHandler.startElement(qName, mAttrWrapper);
168 | }
169 |
170 | public void startPrefixMapping(String prefix, String uri) {
171 | // no equivalent in SAX1, ignore
172 | }
173 | }
174 |
175 | /**
176 | * And one more helper to deal with attribute access differences
177 | *
178 | * @deprecated
179 | */
180 | final static class AttributesWrapper
181 | implements AttributeList {
182 | Attributes attrs;
183 |
184 | public AttributesWrapper() {
185 | }
186 |
187 | public void setAttributes(Attributes a) {
188 | attrs = a;
189 | }
190 |
191 | public int getLength() {
192 | return attrs.getLength();
193 | }
194 |
195 | public String getName(int i) {
196 | String n = attrs.getQName(i);
197 | return (n == null) ? attrs.getLocalName(i) : n;
198 | }
199 |
200 | public String getType(int i) {
201 | return attrs.getType(i);
202 | }
203 |
204 | public String getType(String name) {
205 | return attrs.getType(name);
206 | }
207 |
208 | public String getValue(int i) {
209 | return attrs.getValue(i);
210 | }
211 |
212 | public String getValue(String name) {
213 | return attrs.getValue(name);
214 | }
215 | }
216 | }
217 |
--------------------------------------------------------------------------------
/lib_tv_fullhtml/src/main/java/com/xys/lib_tv_fullhtml/tagsoup/Element.java:
--------------------------------------------------------------------------------
1 | // This file is part of TagSoup and is Copyright 2002-2008 by John Cowan.
2 | //
3 | // TagSoup is licensed under the Apache License,
4 | // Version 2.0. You may obtain a copy of this license at
5 | // http://www.apache.org/licenses/LICENSE-2.0 . You may also have
6 | // additional legal rights not granted by this license.
7 | //
8 | // TagSoup is distributed in the hope that it will be useful, but
9 | // unless required by applicable law or agreed to in writing, TagSoup
10 | // is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS
11 | // OF ANY KIND, either express or implied; not even the implied warranty
12 | // of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
13 |
14 | package com.xys.lib_tv_fullhtml.tagsoup;
15 |
16 | /**
17 | * The internal representation of an actual element (not an element type).
18 | * An Element has an element type, attributes, and a successor Element
19 | * for use in constructing stacks and queues of Elements.
20 | *
21 | * @see ElementType
22 | * @see AttributesImpl
23 | */
24 | public class Element {
25 |
26 |
27 | private ElementType theType; // type of element
28 | private AttributesImpl theAtts; // attributes of element
29 | private Element theNext; // successor of element
30 | private boolean preclosed; // this element has been preclosed
31 |
32 | /**
33 | * Return an Element from a specified ElementType.
34 | *
35 | * @param type The element type of the newly constructed element
36 | * @param defaultAttributes True if default attributes are wanted
37 | */
38 |
39 | public Element(ElementType type, boolean defaultAttributes) {
40 | theType = type;
41 | if (defaultAttributes) theAtts = new AttributesImpl(type.atts());
42 | else theAtts = new AttributesImpl();
43 | theNext = null;
44 | preclosed = false;
45 | }
46 |
47 | /**
48 | * Return the element type.
49 | *
50 | * @return The element type.
51 | */
52 |
53 | public ElementType type() {
54 | return theType;
55 | }
56 |
57 | /**
58 | * Return the attributes as an AttributesImpl object.
59 | * Returning an AttributesImpl makes the attributes mutable.
60 | *
61 | * @return The attributes
62 | * @see AttributesImpl
63 | */
64 | public AttributesImpl atts() {
65 | return theAtts;
66 | }
67 |
68 | /**
69 | * Return the next element in an element stack or queue.
70 | *
71 | * @return The next element
72 | */
73 |
74 | public Element next() {
75 | return theNext;
76 | }
77 |
78 | /**
79 | * Change the next element in an element stack or queue.
80 | *
81 | * @param next The new next element
82 | */
83 |
84 | public void setNext(Element next) {
85 | theNext = next;
86 | }
87 |
88 | /**
89 | * Return the name of the element's type.
90 | * Convenience method.
91 | *
92 | * @return The element type name
93 | */
94 |
95 | public String name() {
96 | return theType.name();
97 | }
98 |
99 | /**
100 | * Return the namespace name of the element's type.
101 | * Convenience method.
102 | *
103 | * @return The element type namespace name
104 | */
105 |
106 | public String namespace() {
107 | return theType.namespace();
108 | }
109 |
110 | /**
111 | * Return the local name of the element's type.
112 | * Convenience method.
113 | *
114 | * @return The element type local name
115 | */
116 |
117 | public String localName() {
118 | return theType.localName();
119 | }
120 |
121 | /**
122 | * Return the content model vector of the element's type.
123 | * Convenience method.
124 | *
125 | * @return The content model vector
126 | */
127 |
128 | public int model() {
129 | return theType.model();
130 | }
131 |
132 | /**
133 | * Return the member-of vector of the element's type.
134 | * Convenience method.
135 | *
136 | * @return The member-of vector
137 | */
138 |
139 | public int memberOf() {
140 | return theType.memberOf();
141 | }
142 |
143 | /**
144 | * Return the flags vector of the element's type.
145 | * Convenience method.
146 | *
147 | * @return The flags vector
148 | */
149 |
150 | public int flags() {
151 | return theType.flags();
152 | }
153 |
154 | /**
155 | * Return the parent element type of the element's type.
156 | * Convenience method.
157 | *
158 | * @return The parent element type
159 | */
160 |
161 | public ElementType parent() {
162 | return theType.parent();
163 | }
164 |
165 | /**
166 | * Return true if the type of this element can contain the type of
167 | * another element.
168 | * Convenience method.
169 | *
170 | * @param other The other element
171 | */
172 |
173 | public boolean canContain(Element other) {
174 | return theType.canContain(other.theType);
175 | }
176 |
177 |
178 | /**
179 | * Set an attribute and its value into this element.
180 | *
181 | * @param name The attribute name (Qname)
182 | * @param type The attribute type
183 | * @param value The attribute value
184 | */
185 |
186 | public void setAttribute(String name, String type, String value) {
187 | theType.setAttribute(theAtts, name, type, value);
188 | }
189 |
190 | /**
191 | * Make this element anonymous.
192 | * Remove any id or name attribute present
193 | * in the element's attributes.
194 | */
195 |
196 | public void anonymize() {
197 | for (int i = theAtts.getLength() - 1; i >= 0; i--) {
198 | if (theAtts.getType(i).equals("ID") ||
199 | theAtts.getQName(i).equals("name")) {
200 | theAtts.removeAttribute(i);
201 | }
202 | }
203 | }
204 |
205 | /**
206 | * Clean the attributes of this element.
207 | * Attributes with null name (the name was ill-formed)
208 | * or null value (the attribute was present in the element type but
209 | * not in this actual element) are removed.
210 | */
211 |
212 | public void clean() {
213 | for (int i = theAtts.getLength() - 1; i >= 0; i--) {
214 | String name = theAtts.getLocalName(i);
215 | if (theAtts.getValue(i) == null || name == null ||
216 | name.length() == 0) {
217 | theAtts.removeAttribute(i);
218 | continue;
219 | }
220 | }
221 | }
222 |
223 | /**
224 | * Force this element to preclosed status, meaning that an end-tag has
225 | * been seen but the element cannot yet be closed for structural reasons.
226 | */
227 |
228 | public void preclose() {
229 | preclosed = true;
230 | }
231 |
232 | /**
233 | * Return true if this element has been preclosed.
234 | */
235 |
236 | public boolean isPreclosed() {
237 | return preclosed;
238 | }
239 |
240 | }
241 |
--------------------------------------------------------------------------------
/lib_tv_fullhtml/lib_tv_fullhtml.iml:
--------------------------------------------------------------------------------
1 |
2 | 25 | * This module, both source code and documentation, is in the 26 | * Public Domain, and comes with NO WARRANTY. 27 | * See http://www.saxproject.org 28 | * for further information. 29 | *30 | * 31 | *
This class provides a default implementation of the SAX2 32 | * {@link Attributes Attributes} interface, with the 33 | * addition of manipulators so that the list can be modified or 34 | * reused.
35 | * 36 | *There are two typical uses of this class:
37 | * 38 | *This class replaces the now-deprecated SAX1 {@link 45 | * org.xml.sax.helpers.AttributeListImpl AttributeListImpl} 46 | * class; in addition to supporting the updated Attributes 47 | * interface rather than the deprecated {@link org.xml.sax.AttributeList 48 | * AttributeList} interface, it also includes a much more efficient 49 | * implementation using a single array rather than a set of Vectors.
50 | * 51 | * @author David Megginson 52 | * @version 2.0.1 (sax2r2) 53 | * @since SAX 2.0 54 | */ 55 | public class AttributesImpl implements Attributes { 56 | 57 | 58 | //////////////////////////////////////////////////////////////////// 59 | // Constructors. 60 | //////////////////////////////////////////////////////////////////// 61 | 62 | 63 | int length; 64 | String data[]; 65 | 66 | 67 | //////////////////////////////////////////////////////////////////// 68 | // Implementation of org.xml.sax.Attributes. 69 | //////////////////////////////////////////////////////////////////// 70 | 71 | 72 | /** 73 | * Construct a new, empty AttributesImpl object. 74 | */ 75 | public AttributesImpl() { 76 | length = 0; 77 | data = null; 78 | } 79 | 80 | 81 | /** 82 | * Copy an existing Attributes object. 83 | * 84 | *This constructor is especially useful inside a 85 | * {@link org.xml.sax.ContentHandler#startElement startElement} event.
86 | * 87 | * @param atts The existing Attributes object. 88 | */ 89 | public AttributesImpl(Attributes atts) { 90 | setAttributes(atts); 91 | } 92 | 93 | /** 94 | * Return the number of attributes in the list. 95 | * 96 | * @return The number of attributes in the list. 97 | * @see Attributes#getLength 98 | */ 99 | public int getLength() { 100 | return length; 101 | } 102 | 103 | /** 104 | * Return an attribute's Namespace URI. 105 | * 106 | * @param index The attribute's index (zero-based). 107 | * @return The Namespace URI, the empty string if none is 108 | * available, or null if the index is out of range. 109 | * @see Attributes#getURI 110 | */ 111 | public String getURI(int index) { 112 | if (index >= 0 && index < length) { 113 | return data[index * 5]; 114 | } else { 115 | return null; 116 | } 117 | } 118 | 119 | /** 120 | * Return an attribute's local name. 121 | * 122 | * @param index The attribute's index (zero-based). 123 | * @return The attribute's local name, the empty string if 124 | * none is available, or null if the index if out of range. 125 | * @see Attributes#getLocalName 126 | */ 127 | public String getLocalName(int index) { 128 | if (index >= 0 && index < length) { 129 | return data[index * 5 + 1]; 130 | } else { 131 | return null; 132 | } 133 | } 134 | 135 | /** 136 | * Return an attribute's qualified (prefixed) name. 137 | * 138 | * @param index The attribute's index (zero-based). 139 | * @return The attribute's qualified name, the empty string if 140 | * none is available, or null if the index is out of bounds. 141 | * @see Attributes#getQName 142 | */ 143 | public String getQName(int index) { 144 | if (index >= 0 && index < length) { 145 | return data[index * 5 + 2]; 146 | } else { 147 | return null; 148 | } 149 | } 150 | 151 | /** 152 | * Return an attribute's type by index. 153 | * 154 | * @param index The attribute's index (zero-based). 155 | * @return The attribute's type, "CDATA" if the type is unknown, or null 156 | * if the index is out of bounds. 157 | * @see Attributes#getType(int) 158 | */ 159 | public String getType(int index) { 160 | if (index >= 0 && index < length) { 161 | return data[index * 5 + 3]; 162 | } else { 163 | return null; 164 | } 165 | } 166 | 167 | /** 168 | * Return an attribute's value by index. 169 | * 170 | * @param index The attribute's index (zero-based). 171 | * @return The attribute's value or null if the index is out of bounds. 172 | * @see Attributes#getValue(int) 173 | */ 174 | public String getValue(int index) { 175 | if (index >= 0 && index < length) { 176 | return data[index * 5 + 4]; 177 | } else { 178 | return null; 179 | } 180 | } 181 | 182 | /** 183 | * Look up an attribute's index by Namespace name. 184 | * 185 | *In many cases, it will be more efficient to look up the name once and 186 | * use the index query methods rather than using the name query methods 187 | * repeatedly.
188 | * 189 | * @param uri The attribute's Namespace URI, or the empty 190 | * string if none is available. 191 | * @param localName The attribute's local name. 192 | * @return The attribute's index, or -1 if none matches. 193 | * @see Attributes#getIndex(String, String) 194 | */ 195 | public int getIndex(String uri, String localName) { 196 | int max = length * 5; 197 | for (int i = 0; i < max; i += 5) { 198 | if (data[i].equals(uri) && data[i + 1].equals(localName)) { 199 | return i / 5; 200 | } 201 | } 202 | return -1; 203 | } 204 | 205 | /** 206 | * Look up an attribute's index by qualified (prefixed) name. 207 | * 208 | * @param qName The qualified name. 209 | * @return The attribute's index, or -1 if none matches. 210 | * @see Attributes#getIndex(String) 211 | */ 212 | public int getIndex(String qName) { 213 | int max = length * 5; 214 | for (int i = 0; i < max; i += 5) { 215 | if (data[i + 2].equals(qName)) { 216 | return i / 5; 217 | } 218 | } 219 | return -1; 220 | } 221 | 222 | /** 223 | * Look up an attribute's type by Namespace-qualified name. 224 | * 225 | * @param uri The Namespace URI, or the empty string for a name 226 | * with no explicit Namespace URI. 227 | * @param localName The local name. 228 | * @return The attribute's type, or null if there is no 229 | * matching attribute. 230 | * @see Attributes#getType(String, String) 231 | */ 232 | public String getType(String uri, String localName) { 233 | int max = length * 5; 234 | for (int i = 0; i < max; i += 5) { 235 | if (data[i].equals(uri) && data[i + 1].equals(localName)) { 236 | return data[i + 3]; 237 | } 238 | } 239 | return null; 240 | } 241 | 242 | /** 243 | * Look up an attribute's type by qualified (prefixed) name. 244 | * 245 | * @param qName The qualified name. 246 | * @return The attribute's type, or null if there is no 247 | * matching attribute. 248 | * @see Attributes#getType(String) 249 | */ 250 | public String getType(String qName) { 251 | int max = length * 5; 252 | for (int i = 0; i < max; i += 5) { 253 | if (data[i + 2].equals(qName)) { 254 | return data[i + 3]; 255 | } 256 | } 257 | return null; 258 | } 259 | 260 | 261 | //////////////////////////////////////////////////////////////////// 262 | // Manipulators. 263 | //////////////////////////////////////////////////////////////////// 264 | 265 | /** 266 | * Look up an attribute's value by Namespace-qualified name. 267 | * 268 | * @param uri The Namespace URI, or the empty string for a name 269 | * with no explicit Namespace URI. 270 | * @param localName The local name. 271 | * @return The attribute's value, or null if there is no 272 | * matching attribute. 273 | * @see Attributes#getValue(String, String) 274 | */ 275 | public String getValue(String uri, String localName) { 276 | int max = length * 5; 277 | for (int i = 0; i < max; i += 5) { 278 | if (data[i].equals(uri) && data[i + 1].equals(localName)) { 279 | return data[i + 4]; 280 | } 281 | } 282 | return null; 283 | } 284 | 285 | /** 286 | * Look up an attribute's value by qualified (prefixed) name. 287 | * 288 | * @param qName The qualified name. 289 | * @return The attribute's value, or null if there is no 290 | * matching attribute. 291 | * @see Attributes#getValue(String) 292 | */ 293 | public String getValue(String qName) { 294 | int max = length * 5; 295 | for (int i = 0; i < max; i += 5) { 296 | if (data[i + 2].equals(qName)) { 297 | return data[i + 4]; 298 | } 299 | } 300 | return null; 301 | } 302 | 303 | /** 304 | * Clear the attribute list for reuse. 305 | * 306 | *Note that little memory is freed by this call: 307 | * the current array is kept so it can be 308 | * reused.
309 | */ 310 | public void clear() { 311 | if (data != null) { 312 | for (int i = 0; i < (length * 5); i++) 313 | data[i] = null; 314 | } 315 | length = 0; 316 | } 317 | 318 | /** 319 | * Copy an entire Attributes object. 320 | * 321 | *It may be more efficient to reuse an existing object 322 | * rather than constantly allocating new ones.
323 | * 324 | * @param atts The attributes to copy. 325 | */ 326 | public void setAttributes(Attributes atts) { 327 | clear(); 328 | length = atts.getLength(); 329 | if (length > 0) { 330 | data = new String[length * 5]; 331 | for (int i = 0; i < length; i++) { 332 | data[i * 5] = atts.getURI(i); 333 | data[i * 5 + 1] = atts.getLocalName(i); 334 | data[i * 5 + 2] = atts.getQName(i); 335 | data[i * 5 + 3] = atts.getType(i); 336 | data[i * 5 + 4] = atts.getValue(i); 337 | } 338 | } 339 | } 340 | 341 | /** 342 | * Add an attribute to the end of the list. 343 | * 344 | *For the sake of speed, this method does no checking 345 | * to see if the attribute is already in the list: that is 346 | * the responsibility of the application.
347 | * 348 | * @param uri The Namespace URI, or the empty string if 349 | * none is available or Namespace processing is not 350 | * being performed. 351 | * @param localName The local name, or the empty string if 352 | * Namespace processing is not being performed. 353 | * @param qName The qualified (prefixed) name, or the empty string 354 | * if qualified names are not available. 355 | * @param type The attribute type as a string. 356 | * @param value The attribute value. 357 | */ 358 | public void addAttribute(String uri, String localName, String qName, 359 | String type, String value) { 360 | ensureCapacity(length + 1); 361 | data[length * 5] = uri; 362 | data[length * 5 + 1] = localName; 363 | data[length * 5 + 2] = qName; 364 | data[length * 5 + 3] = type; 365 | data[length * 5 + 4] = value; 366 | length++; 367 | } 368 | 369 | /** 370 | * Set an attribute in the list. 371 | * 372 | *For the sake of speed, this method does no checking 373 | * for name conflicts or well-formedness: such checks are the 374 | * responsibility of the application.
375 | * 376 | * @param index The index of the attribute (zero-based). 377 | * @param uri The Namespace URI, or the empty string if 378 | * none is available or Namespace processing is not 379 | * being performed. 380 | * @param localName The local name, or the empty string if 381 | * Namespace processing is not being performed. 382 | * @param qName The qualified name, or the empty string 383 | * if qualified names are not available. 384 | * @param type The attribute type as a string. 385 | * @param value The attribute value. 386 | * @throws ArrayIndexOutOfBoundsException When the 387 | * supplied index does not point to an attribute 388 | * in the list. 389 | */ 390 | public void setAttribute(int index, String uri, String localName, 391 | String qName, String type, String value) { 392 | if (index >= 0 && index < length) { 393 | data[index * 5] = uri; 394 | data[index * 5 + 1] = localName; 395 | data[index * 5 + 2] = qName; 396 | data[index * 5 + 3] = type; 397 | data[index * 5 + 4] = value; 398 | } else { 399 | badIndex(index); 400 | } 401 | } 402 | 403 | /** 404 | * Remove an attribute from the list. 405 | * 406 | * @param index The index of the attribute (zero-based). 407 | * @throws ArrayIndexOutOfBoundsException When the 408 | * supplied index does not point to an attribute 409 | * in the list. 410 | */ 411 | public void removeAttribute(int index) { 412 | if (index >= 0 && index < length) { 413 | if (index < length - 1) { 414 | System.arraycopy(data, (index + 1) * 5, data, index * 5, 415 | (length - index - 1) * 5); 416 | } 417 | index = (length - 1) * 5; 418 | data[index++] = null; 419 | data[index++] = null; 420 | data[index++] = null; 421 | data[index++] = null; 422 | data[index] = null; 423 | length--; 424 | } else { 425 | badIndex(index); 426 | } 427 | } 428 | 429 | /** 430 | * Set the Namespace URI of a specific attribute. 431 | * 432 | * @param index The index of the attribute (zero-based). 433 | * @param uri The attribute's Namespace URI, or the empty 434 | * string for none. 435 | * @throws ArrayIndexOutOfBoundsException When the 436 | * supplied index does not point to an attribute 437 | * in the list. 438 | */ 439 | public void setURI(int index, String uri) { 440 | if (index >= 0 && index < length) { 441 | data[index * 5] = uri; 442 | } else { 443 | badIndex(index); 444 | } 445 | } 446 | 447 | /** 448 | * Set the local name of a specific attribute. 449 | * 450 | * @param index The index of the attribute (zero-based). 451 | * @param localName The attribute's local name, or the empty 452 | * string for none. 453 | * @throws ArrayIndexOutOfBoundsException When the 454 | * supplied index does not point to an attribute 455 | * in the list. 456 | */ 457 | public void setLocalName(int index, String localName) { 458 | if (index >= 0 && index < length) { 459 | data[index * 5 + 1] = localName; 460 | } else { 461 | badIndex(index); 462 | } 463 | } 464 | 465 | /** 466 | * Set the qualified name of a specific attribute. 467 | * 468 | * @param index The index of the attribute (zero-based). 469 | * @param qName The attribute's qualified name, or the empty 470 | * string for none. 471 | * @throws ArrayIndexOutOfBoundsException When the 472 | * supplied index does not point to an attribute 473 | * in the list. 474 | */ 475 | public void setQName(int index, String qName) { 476 | if (index >= 0 && index < length) { 477 | data[index * 5 + 2] = qName; 478 | } else { 479 | badIndex(index); 480 | } 481 | } 482 | 483 | 484 | //////////////////////////////////////////////////////////////////// 485 | // Internal methods. 486 | //////////////////////////////////////////////////////////////////// 487 | 488 | /** 489 | * Set the type of a specific attribute. 490 | * 491 | * @param index The index of the attribute (zero-based). 492 | * @param type The attribute's type. 493 | * @throws ArrayIndexOutOfBoundsException When the 494 | * supplied index does not point to an attribute 495 | * in the list. 496 | */ 497 | public void setType(int index, String type) { 498 | if (index >= 0 && index < length) { 499 | data[index * 5 + 3] = type; 500 | } else { 501 | badIndex(index); 502 | } 503 | } 504 | 505 | /** 506 | * Set the value of a specific attribute. 507 | * 508 | * @param index The index of the attribute (zero-based). 509 | * @param value The attribute's value. 510 | * @throws ArrayIndexOutOfBoundsException When the 511 | * supplied index does not point to an attribute 512 | * in the list. 513 | */ 514 | public void setValue(int index, String value) { 515 | if (index >= 0 && index < length) { 516 | data[index * 5 + 4] = value; 517 | } else { 518 | badIndex(index); 519 | } 520 | } 521 | 522 | 523 | //////////////////////////////////////////////////////////////////// 524 | // Internal state. 525 | //////////////////////////////////////////////////////////////////// 526 | 527 | /** 528 | * Ensure the internal array's capacity. 529 | * 530 | * @param n The minimum number of attributes that the array must 531 | * be able to hold. 532 | */ 533 | private void ensureCapacity(int n) { 534 | if (n <= 0) { 535 | return; 536 | } 537 | int max; 538 | if (data == null || data.length == 0) { 539 | max = 25; 540 | } else if (data.length >= n * 5) { 541 | return; 542 | } else { 543 | max = data.length; 544 | } 545 | while (max < n * 5) { 546 | max *= 2; 547 | } 548 | 549 | String newData[] = new String[max]; 550 | if (length > 0) { 551 | System.arraycopy(data, 0, newData, 0, length * 5); 552 | } 553 | data = newData; 554 | } 555 | 556 | /** 557 | * Report a bad array index in a manipulator. 558 | * 559 | * @param index The index to report. 560 | * @throws ArrayIndexOutOfBoundsException Always. 561 | */ 562 | private void badIndex(int index) 563 | throws ArrayIndexOutOfBoundsException { 564 | String msg = 565 | "Attempt to modify attribute at illegal index: " + index; 566 | throw new ArrayIndexOutOfBoundsException(msg); 567 | } 568 | 569 | } 570 | 571 | // end of AttributesImpl.java 572 | 573 | -------------------------------------------------------------------------------- /lib_tv_fullhtml/src/main/java/com/xys/lib_tv_fullhtml/tagsoup/HTMLScanner.java: -------------------------------------------------------------------------------- 1 | // This file is part of TagSoup and is Copyright 2002-2008 by John Cowan. 2 | // 3 | // TagSoup is licensed under the Apache License, 4 | // Version 2.0. You may obtain a copy of this license at 5 | // http://www.apache.org/licenses/LICENSE-2.0 . You may also have 6 | // additional legal rights not granted by this license. 7 | // 8 | // TagSoup is distributed in the hope that it will be useful, but 9 | // unless required by applicable law or agreed to in writing, TagSoup 10 | // is distributed on an "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS 11 | // OF ANY KIND, either express or implied; not even the implied warranty 12 | // of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. 13 | // 14 | // 15 | package com.xys.lib_tv_fullhtml.tagsoup; 16 | 17 | import org.xml.sax.Locator; 18 | import org.xml.sax.SAXException; 19 | 20 | import java.io.BufferedReader; 21 | import java.io.IOException; 22 | import java.io.InputStreamReader; 23 | import java.io.OutputStreamWriter; 24 | import java.io.PushbackReader; 25 | import java.io.Reader; 26 | import java.io.Writer; 27 | 28 | /** 29 | * This class implements a table-driven scanner for HTML, allowing for lots of 30 | * defects. It implements the Scanner interface, which accepts a Reader 31 | * object to fetch characters from and a ScanHandler object to report lexical 32 | * events to. 33 | */ 34 | 35 | public class HTMLScanner implements Scanner, Locator { 36 | 37 | // Start of state table 38 | private static final int S_ANAME = 1; 39 | private static final int S_APOS = 2; 40 | private static final int S_AVAL = 3; 41 | private static final int S_BB = 4; 42 | private static final int S_BBC = 5; 43 | private static final int S_BBCD = 6; 44 | private static final int S_BBCDA = 7; 45 | private static final int S_BBCDAT = 8; 46 | private static final int S_BBCDATA = 9; 47 | private static final int S_CDATA = 10; 48 | private static final int S_CDATA2 = 11; 49 | private static final int S_CDSECT = 12; 50 | private static final int S_CDSECT1 = 13; 51 | private static final int S_CDSECT2 = 14; 52 | private static final int S_COM = 15; 53 | private static final int S_COM2 = 16; 54 | private static final int S_COM3 = 17; 55 | private static final int S_COM4 = 18; 56 | private static final int S_DECL = 19; 57 | private static final int S_DECL2 = 20; 58 | private static final int S_DONE = 21; 59 | private static final int S_EMPTYTAG = 22; 60 | private static final int S_ENT = 23; 61 | private static final int S_EQ = 24; 62 | private static final int S_ETAG = 25; 63 | private static final int S_GI = 26; 64 | private static final int S_NCR = 27; 65 | private static final int S_PCDATA = 28; 66 | private static final int S_PI = 29; 67 | private static final int S_PITARGET = 30; 68 | private static final int S_QUOT = 31; 69 | private static final int S_STAGC = 32; 70 | private static final int S_TAG = 33; 71 | private static final int S_TAGWS = 34; 72 | private static final int S_XNCR = 35; 73 | private static final int A_ADUP = 1; 74 | private static final int A_ADUP_SAVE = 2; 75 | private static final int A_ADUP_STAGC = 3; 76 | private static final int A_ANAME = 4; 77 | private static final int A_ANAME_ADUP = 5; 78 | private static final int A_ANAME_ADUP_STAGC = 6; 79 | private static final int A_AVAL = 7; 80 | private static final int A_AVAL_STAGC = 8; 81 | private static final int A_CDATA = 9; 82 | private static final int A_CMNT = 10; 83 | private static final int A_DECL = 11; 84 | private static final int A_EMPTYTAG = 12; 85 | private static final int A_ENTITY = 13; 86 | private static final int A_ENTITY_START = 14; 87 | private static final int A_ETAG = 15; 88 | private static final int A_GI = 16; 89 | private static final int A_GI_STAGC = 17; 90 | private static final int A_LT = 18; 91 | private static final int A_LT_PCDATA = 19; 92 | private static final int A_MINUS = 20; 93 | private static final int A_MINUS2 = 21; 94 | private static final int A_MINUS3 = 22; 95 | private static final int A_PCDATA = 23; 96 | private static final int A_PI = 24; 97 | private static final int A_PITARGET = 25; 98 | private static final int A_PITARGET_PI = 26; 99 | private static final int A_SAVE = 27; 100 | private static final int A_SKIP = 28; 101 | private static final int A_SP = 29; 102 | private static final int A_STAGC = 30; 103 | private static final int A_UNGET = 31; 104 | private static final int A_UNSAVE_PCDATA = 32; 105 | private static final String[] debug_actionnames = {"", "A_ADUP", "A_ADUP_SAVE", "A_ADUP_STAGC", "A_ANAME", 106 | "A_ANAME_ADUP", "A_ANAME_ADUP_STAGC", "A_AVAL", "A_AVAL_STAGC", "A_CDATA", "A_CMNT", "A_DECL", 107 | "A_EMPTYTAG", "A_ENTITY", "A_ENTITY_START", "A_ETAG", "A_GI", "A_GI_STAGC", "A_LT", "A_LT_PCDATA", 108 | "A_MINUS", "A_MINUS2", "A_MINUS3", "A_PCDATA", "A_PI", "A_PITARGET", "A_PITARGET_PI", "A_SAVE", "A_SKIP", 109 | "A_SP", "A_STAGC", "A_UNGET", "A_UNSAVE_PCDATA"}; 110 | private static final String[] debug_statenames = {"", "S_ANAME", "S_APOS", "S_AVAL", "S_BB", "S_BBC", "S_BBCD", 111 | "S_BBCDA", "S_BBCDAT", "S_BBCDATA", "S_CDATA", "S_CDATA2", "S_CDSECT", "S_CDSECT1", "S_CDSECT2", "S_COM", 112 | "S_COM2", "S_COM3", "S_COM4", "S_DECL", "S_DECL2", "S_DONE", "S_EMPTYTAG", "S_ENT", "S_EQ", "S_ETAG", 113 | "S_GI", "S_NCR", "S_PCDATA", "S_PI", "S_PITARGET", "S_QUOT", "S_STAGC", "S_TAG", "S_TAGWS", "S_XNCR"}; 114 | private static int[] statetable = { 115 | S_ANAME, '/', A_ANAME_ADUP, S_EMPTYTAG, 116 | S_ANAME, '=', A_ANAME, S_AVAL, 117 | S_ANAME, '>', A_ANAME_ADUP_STAGC, S_PCDATA, 118 | S_ANAME, 0, A_SAVE, S_ANAME, 119 | S_ANAME, -1, A_ANAME_ADUP_STAGC, S_DONE, 120 | S_ANAME, ' ', A_ANAME, S_EQ, 121 | S_ANAME, '\n', A_ANAME, S_EQ, 122 | S_ANAME, '\t', A_ANAME, S_EQ, 123 | S_APOS, '\'', A_AVAL, S_TAGWS, 124 | S_APOS, 0, A_SAVE, S_APOS, 125 | S_APOS, -1, A_AVAL_STAGC, S_DONE, 126 | S_APOS, ' ', A_SP, S_APOS, 127 | S_APOS, '\n', A_SP, S_APOS, 128 | S_APOS, '\t', A_SP, S_APOS, 129 | S_AVAL, '\'', A_SKIP, S_APOS, 130 | S_AVAL, '"', A_SKIP, S_QUOT, 131 | S_AVAL, '>', A_AVAL_STAGC, S_PCDATA, 132 | S_AVAL, 0, A_SAVE, S_STAGC, 133 | S_AVAL, -1, A_AVAL_STAGC, S_DONE, 134 | S_AVAL, ' ', A_SKIP, S_AVAL, 135 | S_AVAL, '\n', A_SKIP, S_AVAL, 136 | S_AVAL, '\t', A_SKIP, S_AVAL, 137 | S_BB, 'C', A_SKIP, S_BBC, 138 | S_BB, 0, A_SKIP, S_DECL, 139 | S_BB, -1, A_SKIP, S_DONE, 140 | S_BBC, 'D', A_SKIP, S_BBCD, 141 | S_BBC, 0, A_SKIP, S_DECL, 142 | S_BBC, -1, A_SKIP, S_DONE, 143 | S_BBCD, 'A', A_SKIP, S_BBCDA, 144 | S_BBCD, 0, A_SKIP, S_DECL, 145 | S_BBCD, -1, A_SKIP, S_DONE, 146 | S_BBCDA, 'T', A_SKIP, S_BBCDAT, 147 | S_BBCDA, 0, A_SKIP, S_DECL, 148 | S_BBCDA, -1, A_SKIP, S_DONE, 149 | S_BBCDAT, 'A', A_SKIP, S_BBCDATA, 150 | S_BBCDAT, 0, A_SKIP, S_DECL, 151 | S_BBCDAT, -1, A_SKIP, S_DONE, 152 | S_BBCDATA, '[', A_SKIP, S_CDSECT, 153 | S_BBCDATA, 0, A_SKIP, S_DECL, 154 | S_BBCDATA, -1, A_SKIP, S_DONE, 155 | S_CDATA, '<', A_SAVE, S_CDATA2, 156 | S_CDATA, 0, A_SAVE, S_CDATA, 157 | S_CDATA, -1, A_PCDATA, S_DONE, 158 | S_CDATA2, '/', A_UNSAVE_PCDATA, S_ETAG, 159 | S_CDATA2, 0, A_SAVE, S_CDATA, 160 | S_CDATA2, -1, A_UNSAVE_PCDATA, S_DONE, 161 | S_CDSECT, ']', A_SAVE, S_CDSECT1, 162 | S_CDSECT, 0, A_SAVE, S_CDSECT, 163 | S_CDSECT, -1, A_SKIP, S_DONE, 164 | S_CDSECT1, ']', A_SAVE, S_CDSECT2, 165 | S_CDSECT1, 0, A_SAVE, S_CDSECT, 166 | S_CDSECT1, -1, A_SKIP, S_DONE, 167 | S_CDSECT2, '>', A_CDATA, S_PCDATA, 168 | S_CDSECT2, 0, A_SAVE, S_CDSECT, 169 | S_CDSECT2, -1, A_SKIP, S_DONE, 170 | S_COM, '-', A_SKIP, S_COM2, 171 | S_COM, 0, A_SAVE, S_COM2, 172 | S_COM, -1, A_CMNT, S_DONE, 173 | S_COM2, '-', A_SKIP, S_COM3, 174 | S_COM2, 0, A_SAVE, S_COM2, 175 | S_COM2, -1, A_CMNT, S_DONE, 176 | S_COM3, '-', A_SKIP, S_COM4, 177 | S_COM3, 0, A_MINUS, S_COM2, 178 | S_COM3, -1, A_CMNT, S_DONE, 179 | S_COM4, '-', A_MINUS3, S_COM4, 180 | S_COM4, '>', A_CMNT, S_PCDATA, 181 | S_COM4, 0, A_MINUS2, S_COM2, 182 | S_COM4, -1, A_CMNT, S_DONE, 183 | S_DECL, '-', A_SKIP, S_COM, 184 | S_DECL, '[', A_SKIP, S_BB, 185 | S_DECL, '>', A_SKIP, S_PCDATA, 186 | S_DECL, 0, A_SAVE, S_DECL2, 187 | S_DECL, -1, A_SKIP, S_DONE, 188 | S_DECL2, '>', A_DECL, S_PCDATA, 189 | S_DECL2, 0, A_SAVE, S_DECL2, 190 | S_DECL2, -1, A_SKIP, S_DONE, 191 | S_EMPTYTAG, '>', A_EMPTYTAG, S_PCDATA, 192 | S_EMPTYTAG, 0, A_SAVE, S_ANAME, 193 | S_EMPTYTAG, ' ', A_SKIP, S_TAGWS, 194 | S_EMPTYTAG, '\n', A_SKIP, S_TAGWS, 195 | S_EMPTYTAG, '\t', A_SKIP, S_TAGWS, 196 | S_ENT, 0, A_ENTITY, S_ENT, 197 | S_ENT, -1, A_ENTITY, S_DONE, 198 | S_EQ, '=', A_SKIP, S_AVAL, 199 | S_EQ, '>', A_ADUP_STAGC, S_PCDATA, 200 | S_EQ, 0, A_ADUP_SAVE, S_ANAME, 201 | S_EQ, -1, A_ADUP_STAGC, S_DONE, 202 | S_EQ, ' ', A_SKIP, S_EQ, 203 | S_EQ, '\n', A_SKIP, S_EQ, 204 | S_EQ, '\t', A_SKIP, S_EQ, 205 | S_ETAG, '>', A_ETAG, S_PCDATA, 206 | S_ETAG, 0, A_SAVE, S_ETAG, 207 | S_ETAG, -1, A_ETAG, S_DONE, 208 | S_ETAG, ' ', A_SKIP, S_ETAG, 209 | S_ETAG, '\n', A_SKIP, S_ETAG, 210 | S_ETAG, '\t', A_SKIP, S_ETAG, 211 | S_GI, '/', A_SKIP, S_EMPTYTAG, 212 | S_GI, '>', A_GI_STAGC, S_PCDATA, 213 | S_GI, 0, A_SAVE, S_GI, 214 | S_GI, -1, A_SKIP, S_DONE, 215 | S_GI, ' ', A_GI, S_TAGWS, 216 | S_GI, '\n', A_GI, S_TAGWS, 217 | S_GI, '\t', A_GI, S_TAGWS, 218 | S_NCR, 0, A_ENTITY, S_NCR, 219 | S_NCR, -1, A_ENTITY, S_DONE, 220 | S_PCDATA, '&', A_ENTITY_START, S_ENT, 221 | S_PCDATA, '<', A_PCDATA, S_TAG, 222 | S_PCDATA, 0, A_SAVE, S_PCDATA, 223 | S_PCDATA, -1, A_PCDATA, S_DONE, 224 | S_PI, '>', A_PI, S_PCDATA, 225 | S_PI, 0, A_SAVE, S_PI, 226 | S_PI, -1, A_PI, S_DONE, 227 | S_PITARGET, '>', A_PITARGET_PI, S_PCDATA, 228 | S_PITARGET, 0, A_SAVE, S_PITARGET, 229 | S_PITARGET, -1, A_PITARGET_PI, S_DONE, 230 | S_PITARGET, ' ', A_PITARGET, S_PI, 231 | S_PITARGET, '\n', A_PITARGET, S_PI, 232 | S_PITARGET, '\t', A_PITARGET, S_PI, 233 | S_QUOT, '"', A_AVAL, S_TAGWS, 234 | S_QUOT, 0, A_SAVE, S_QUOT, 235 | S_QUOT, -1, A_AVAL_STAGC, S_DONE, 236 | S_QUOT, ' ', A_SP, S_QUOT, 237 | S_QUOT, '\n', A_SP, S_QUOT, 238 | S_QUOT, '\t', A_SP, S_QUOT, 239 | S_STAGC, '>', A_AVAL_STAGC, S_PCDATA, 240 | S_STAGC, 0, A_SAVE, S_STAGC, 241 | S_STAGC, -1, A_AVAL_STAGC, S_DONE, 242 | S_STAGC, ' ', A_AVAL, S_TAGWS, 243 | S_STAGC, '\n', A_AVAL, S_TAGWS, 244 | S_STAGC, '\t', A_AVAL, S_TAGWS, 245 | S_TAG, '!', A_SKIP, S_DECL, 246 | S_TAG, '?', A_SKIP, S_PITARGET, 247 | S_TAG, '/', A_SKIP, S_ETAG, 248 | S_TAG, '<', A_SAVE, S_TAG, 249 | S_TAG, 0, A_SAVE, S_GI, 250 | S_TAG, -1, A_LT_PCDATA, S_DONE, 251 | S_TAG, ' ', A_LT, S_PCDATA, 252 | S_TAG, '\n', A_LT, S_PCDATA, 253 | S_TAG, '\t', A_LT, S_PCDATA, 254 | S_TAGWS, '/', A_SKIP, S_EMPTYTAG, 255 | S_TAGWS, '>', A_STAGC, S_PCDATA, 256 | S_TAGWS, 0, A_SAVE, S_ANAME, 257 | S_TAGWS, -1, A_STAGC, S_DONE, 258 | S_TAGWS, ' ', A_SKIP, S_TAGWS, 259 | S_TAGWS, '\n', A_SKIP, S_TAGWS, 260 | S_TAGWS, '\t', A_SKIP, S_TAGWS, 261 | S_XNCR, 0, A_ENTITY, S_XNCR, 262 | S_XNCR, -1, A_ENTITY, S_DONE, 263 | 264 | }; 265 | 266 | 267 | // End of state table 268 | int theState; // Current state 269 | int theNextState; // Next state 270 | char[] theOutputBuffer = new char[200]; // Output buffer 271 | int theSize; // Current buffer size 272 | int[] theWinMap = { // Windows chars map 273 | 0x20AC, 0xFFFD, 0x201A, 0x0192, 0x201E, 0x2026, 0x2020, 0x2021, 274 | 0x02C6, 0x2030, 0x0160, 0x2039, 0x0152, 0xFFFD, 0x017D, 0xFFFD, 275 | 0xFFFD, 0x2018, 0x2019, 0x201C, 0x201D, 0x2022, 0x2013, 0x2014, 276 | 0x02DC, 0x2122, 0x0161, 0x203A, 0x0153, 0xFFFD, 0x017E, 0x0178}; 277 | private String thePublicid; // Locator state 278 | private String theSystemid; 279 | private int theLastLine; 280 | private int theLastColumn; 281 | private int theCurrentLine; 282 | private int theCurrentColumn; 283 | 284 | /** 285 | * Test procedure. Reads HTML from the standard input and writes 286 | * PYX to the standard output. 287 | */ 288 | 289 | public static void main(String[] argv) throws IOException, SAXException { 290 | Scanner s = new HTMLScanner(); 291 | Reader r = new InputStreamReader(System.in, "UTF-8"); 292 | Writer w = new OutputStreamWriter(System.out, "UTF-8"); 293 | PYXWriter pw = new PYXWriter(w); 294 | s.scan(r, pw); 295 | w.close(); 296 | } 297 | 298 | // Locator implementation 299 | 300 | private static String nicechar(int in) { 301 | if (in == '\n') return "\\n"; 302 | if (in < 32) return "0x" + Integer.toHexString(in); 303 | return "'" + ((char) in) + "'"; 304 | } 305 | 306 | // Compensate for bug in PushbackReader that allows 307 | // pushing back EOF. 308 | private void unread(PushbackReader r, int c) throws IOException { 309 | if (c != -1) r.unread(c); 310 | } 311 | 312 | public int getLineNumber() { 313 | return theLastLine; 314 | } 315 | 316 | public int getColumnNumber() { 317 | return theLastColumn; 318 | } 319 | 320 | 321 | // Scanner implementation 322 | 323 | public String getPublicId() { 324 | return thePublicid; 325 | } 326 | 327 | public String getSystemId() { 328 | return theSystemid; 329 | } 330 | 331 | /** 332 | * Reset document locator, supplying systemid and publicid. 333 | * 334 | * @param systemid System id 335 | * @param publicid Public id 336 | */ 337 | 338 | public void resetDocumentLocator(String publicid, String systemid) { 339 | thePublicid = publicid; 340 | theSystemid = systemid; 341 | theLastLine = theLastColumn = theCurrentLine = theCurrentColumn = 0; 342 | } 343 | 344 | /** 345 | * Scan HTML source, reporting lexical events. 346 | * 347 | * @param r0 Reader that provides characters 348 | * @param h ScanHandler that accepts lexical events. 349 | */ 350 | 351 | public void scan(Reader r0, ScanHandler h) throws IOException, SAXException { 352 | theState = S_PCDATA; 353 | PushbackReader r; 354 | if (r0 instanceof PushbackReader) { 355 | r = (PushbackReader) r0; 356 | } else if (r0 instanceof BufferedReader) { 357 | r = new PushbackReader(r0); 358 | } else { 359 | r = new PushbackReader(new BufferedReader(r0, 200)); 360 | } 361 | 362 | int firstChar = r.read(); // Remove any leading BOM 363 | if (firstChar != '\uFEFF') unread(r, firstChar); 364 | 365 | while (theState != S_DONE) { 366 | int ch = r.read(); 367 | 368 | // Process control characters 369 | if (ch >= 0x80 && ch <= 0x9F) ch = theWinMap[ch - 0x80]; 370 | 371 | if (ch == '\r') { 372 | ch = r.read(); // expect LF next 373 | if (ch != '\n') { 374 | unread(r, ch); // nope 375 | ch = '\n'; 376 | } 377 | } 378 | 379 | if (ch == '\n') { 380 | theCurrentLine++; 381 | theCurrentColumn = 0; 382 | } else { 383 | theCurrentColumn++; 384 | } 385 | 386 | if (!(ch >= 0x20 || ch == '\n' || ch == '\t' || ch == -1)) continue; 387 | 388 | // Search state table 389 | int action = 0; 390 | for (int i = 0; i < statetable.length; i += 4) { 391 | if (theState != statetable[i]) { 392 | if (action != 0) break; 393 | continue; 394 | } 395 | if (statetable[i + 1] == 0) { 396 | action = statetable[i + 2]; 397 | theNextState = statetable[i + 3]; 398 | } else if (statetable[i + 1] == ch) { 399 | action = statetable[i + 2]; 400 | theNextState = statetable[i + 3]; 401 | break; 402 | } 403 | } 404 | // System.err.println("In " + debug_statenames[theState] + " got " + nicechar(ch) + " doing " + 405 | // debug_actionnames[action] + " then " + debug_statenames[theNextState]); 406 | switch (action) { 407 | case 0: 408 | throw new Error( 409 | "HTMLScanner can't cope with " + Integer.toString(ch) + " in state " + 410 | Integer.toString(theState)); 411 | case A_ADUP: 412 | h.adup(theOutputBuffer, 0, theSize); 413 | theSize = 0; 414 | break; 415 | case A_ADUP_SAVE: 416 | h.adup(theOutputBuffer, 0, theSize); 417 | theSize = 0; 418 | save(ch, h); 419 | break; 420 | case A_ADUP_STAGC: 421 | h.adup(theOutputBuffer, 0, theSize); 422 | theSize = 0; 423 | h.stagc(theOutputBuffer, 0, theSize); 424 | break; 425 | case A_ANAME: 426 | h.aname(theOutputBuffer, 0, theSize); 427 | theSize = 0; 428 | break; 429 | case A_ANAME_ADUP: 430 | h.aname(theOutputBuffer, 0, theSize); 431 | theSize = 0; 432 | h.adup(theOutputBuffer, 0, theSize); 433 | break; 434 | case A_ANAME_ADUP_STAGC: 435 | h.aname(theOutputBuffer, 0, theSize); 436 | theSize = 0; 437 | h.adup(theOutputBuffer, 0, theSize); 438 | h.stagc(theOutputBuffer, 0, theSize); 439 | break; 440 | case A_AVAL: 441 | h.aval(theOutputBuffer, 0, theSize); 442 | theSize = 0; 443 | break; 444 | case A_AVAL_STAGC: 445 | h.aval(theOutputBuffer, 0, theSize); 446 | theSize = 0; 447 | h.stagc(theOutputBuffer, 0, theSize); 448 | break; 449 | case A_CDATA: 450 | mark(); 451 | // suppress the final "]]" in the buffer 452 | if (theSize > 1) theSize -= 2; 453 | h.pcdata(theOutputBuffer, 0, theSize); 454 | theSize = 0; 455 | break; 456 | case A_ENTITY_START: 457 | h.pcdata(theOutputBuffer, 0, theSize); 458 | theSize = 0; 459 | save(ch, h); 460 | break; 461 | case A_ENTITY: 462 | mark(); 463 | char ch1 = (char) ch; 464 | // System.out.println("Got " + ch1 + " in state " + ((theState == S_ENT) ? "S_ENT" : ((theState == S_NCR) 465 | // ? "S_NCR" : "UNK"))); 466 | if (theState == S_ENT && ch1 == '#') { 467 | theNextState = S_NCR; 468 | save(ch, h); 469 | break; 470 | } else if (theState == S_NCR && (ch1 == 'x' || ch1 == 'X')) { 471 | theNextState = S_XNCR; 472 | save(ch, h); 473 | break; 474 | } else if (theState == S_ENT && Character.isLetterOrDigit(ch1)) { 475 | save(ch, h); 476 | break; 477 | } else if (theState == S_NCR && Character.isDigit(ch1)) { 478 | save(ch, h); 479 | break; 480 | } else if (theState == S_XNCR && (Character.isDigit(ch1) || "abcdefABCDEF".indexOf(ch1) != -1)) { 481 | save(ch, h); 482 | break; 483 | } 484 | 485 | // The whole entity reference has been collected 486 | // System.err.println("%%" + new String(theOutputBuffer, 0, theSize)); 487 | h.entity(theOutputBuffer, 1, theSize - 1); 488 | int ent = h.getEntity(); 489 | // System.err.println("%% value = " + ent); 490 | if (ent != 0) { 491 | theSize = 0; 492 | if (ent >= 0x80 && ent <= 0x9F) { 493 | ent = theWinMap[ent - 0x80]; 494 | } 495 | if (ent < 0x20) { 496 | // Control becomes space 497 | ent = 0x20; 498 | } else if (ent >= 0xD800 && ent <= 0xDFFF) { 499 | // Surrogates get dropped 500 | ent = 0; 501 | } else if (ent <= 0xFFFF) { 502 | // BMP character 503 | save(ent, h); 504 | } else { 505 | // Astral converted to two surrogates 506 | ent -= 0x10000; 507 | save((ent >> 10) + 0xD800, h); 508 | save((ent & 0x3FF) + 0xDC00, h); 509 | } 510 | if (ch != ';') { 511 | unread(r, ch); 512 | theCurrentColumn--; 513 | } 514 | } else { 515 | unread(r, ch); 516 | theCurrentColumn--; 517 | } 518 | theNextState = S_PCDATA; 519 | break; 520 | case A_ETAG: 521 | h.etag(theOutputBuffer, 0, theSize); 522 | theSize = 0; 523 | break; 524 | case A_DECL: 525 | h.decl(theOutputBuffer, 0, theSize); 526 | theSize = 0; 527 | break; 528 | case A_GI: 529 | h.gi(theOutputBuffer, 0, theSize); 530 | theSize = 0; 531 | break; 532 | case A_GI_STAGC: 533 | h.gi(theOutputBuffer, 0, theSize); 534 | theSize = 0; 535 | h.stagc(theOutputBuffer, 0, theSize); 536 | break; 537 | case A_LT: 538 | mark(); 539 | save('<', h); 540 | save(ch, h); 541 | break; 542 | case A_LT_PCDATA: 543 | mark(); 544 | save('<', h); 545 | h.pcdata(theOutputBuffer, 0, theSize); 546 | theSize = 0; 547 | break; 548 | case A_PCDATA: 549 | mark(); 550 | h.pcdata(theOutputBuffer, 0, theSize); 551 | theSize = 0; 552 | break; 553 | case A_CMNT: 554 | mark(); 555 | h.cmnt(theOutputBuffer, 0, theSize); 556 | theSize = 0; 557 | break; 558 | case A_MINUS3: 559 | save('-', h); 560 | save(' ', h); 561 | break; 562 | case A_MINUS2: 563 | save('-', h); 564 | save(' ', h); 565 | // fall through into A_MINUS 566 | case A_MINUS: 567 | save('-', h); 568 | save(ch, h); 569 | break; 570 | case A_PI: 571 | mark(); 572 | h.pi(theOutputBuffer, 0, theSize); 573 | theSize = 0; 574 | break; 575 | case A_PITARGET: 576 | h.pitarget(theOutputBuffer, 0, theSize); 577 | theSize = 0; 578 | break; 579 | case A_PITARGET_PI: 580 | h.pitarget(theOutputBuffer, 0, theSize); 581 | theSize = 0; 582 | h.pi(theOutputBuffer, 0, theSize); 583 | break; 584 | case A_SAVE: 585 | save(ch, h); 586 | break; 587 | case A_SKIP: 588 | break; 589 | case A_SP: 590 | save(' ', h); 591 | break; 592 | case A_STAGC: 593 | h.stagc(theOutputBuffer, 0, theSize); 594 | theSize = 0; 595 | break; 596 | case A_EMPTYTAG: 597 | mark(); 598 | // System.err.println("%%% Empty tag seen"); 599 | if (theSize > 0) h.gi(theOutputBuffer, 0, theSize); 600 | theSize = 0; 601 | h.stage(theOutputBuffer, 0, theSize); 602 | break; 603 | case A_UNGET: 604 | unread(r, ch); 605 | theCurrentColumn--; 606 | break; 607 | case A_UNSAVE_PCDATA: 608 | if (theSize > 0) theSize--; 609 | h.pcdata(theOutputBuffer, 0, theSize); 610 | theSize = 0; 611 | break; 612 | default: 613 | throw new Error("Can't process state " + action); 614 | } 615 | theState = theNextState; 616 | } 617 | h.eof(theOutputBuffer, 0, 0); 618 | } 619 | 620 | /** 621 | * Mark the current scan position as a "point of interest" - start of a tag, 622 | * cdata, processing instruction etc. 623 | */ 624 | 625 | private void mark() { 626 | theLastColumn = theCurrentColumn; 627 | theLastLine = theCurrentLine; 628 | } 629 | 630 | /** 631 | * A callback for the ScanHandler that allows it to force 632 | * the lexer state to CDATA content (no markup is recognized except 633 | * the end of element. 634 | */ 635 | 636 | public void startCDATA() { 637 | theNextState = S_CDATA; 638 | } 639 | 640 | private void save(int ch, ScanHandler h) throws IOException, SAXException { 641 | if (theSize >= theOutputBuffer.length - 20) { 642 | if (theState == S_PCDATA || theState == S_CDATA) { 643 | // Return a buffer-sized chunk of PCDATA 644 | h.pcdata(theOutputBuffer, 0, theSize); 645 | theSize = 0; 646 | } else { 647 | // Grow the buffer size 648 | char[] newOutputBuffer = new char[theOutputBuffer.length * 2]; 649 | System.arraycopy(theOutputBuffer, 0, newOutputBuffer, 0, theSize + 1); 650 | theOutputBuffer = newOutputBuffer; 651 | } 652 | } 653 | theOutputBuffer[theSize++] = (char) ch; 654 | } 655 | 656 | } 657 | --------------------------------------------------------------------------------