HTMLIndenter


1   package net.sf.saxon.event;
2   import net.sf.saxon.trans.XPathException;
3   
4   import java.util.Properties  ;
5   
6   /**
7   * HTMLIndenter: This ProxyEmitter indents HTML elements, by adding whitespace
8   * character data where appropriate.
9   * The character data is never added when within an inline element.
10  * The string used for indentation defaults to four spaces, but may be set using the
11  * indent-chars property
12  *
13  * @author Michael Kay
14  */
15  
16  
17  public class HTMLIndenter extends ProxyReceiver {
18  
19      private int level = 0;
20      private int indentSpaces = 3;
21      private String   indentChars = "                                                          ";
22      private boolean sameLine = false;
23      private boolean isInlineTag = false;
24      private boolean inFormattedTag = false;
25      private boolean afterInline = false;
26      private boolean afterFormatted = true;    // to prevent a newline at the start
27      private int[] propertyStack = new int[20];
28  
29  
30      // the list of inline tags is from the HTML 4.0 (loose) spec. The significance is that we
31      // mustn't add spaces immediately before or after one of these elements.
32  
33      protected static String  [] inlineTags = {
34          "tt", "i", "b", "u", "s", "strike", "big", "small", "em", "strong", "dfn", "code", "samp",
35           "kbd", "var", "cite", "abbr", "acronym", "a", "img", "applet", "object", "font",
36           "basefont", "br", "script", "map", "q", "sub", "sup", "span", "bdo", "iframe", "input",
37           "select", "textarea", "label", "button", "ins", "del" };
38  
39          // INS and DEL are not actually inline elements, but the SGML DTD for HTML
40          // (apparently) permits them to be used as if they were.
41  
42      private static HTMLTagHashSet inlineTable = new HTMLTagHashSet(101);
43  
44      static {
45          for (int j=0; j<inlineTags.length; j++) {
46              inlineTable.add(inlineTags[j]);
47          }
48      }
49  
50      protected static final int IS_INLINE = 1;
51      protected static final int IS_FORMATTED = 2;
52  
53      /**
54       * Classify an element name as inline, formatted, or both or neither.
55       * This method is overridden in the XHTML indenter
56       * @param nameCode the element name
57       * @return a bit-significant integer containing flags IS_INLINE and/or IS_FORMATTED
58       */
59      protected int classifyTag(int nameCode) {
60          int r = 0;
61          String   tag = getNamePool().getDisplayName(nameCode);
62          if (inlineTable.contains(tag)) {
63              r |= IS_INLINE;
64          }
65          if (formattedTable.contains(tag)) {
66              r |= IS_FORMATTED;
67          }
68          return r;
69      }
70  
71      // Table of preformatted elements
72  
73      private static HTMLTagHashSet formattedTable = new HTMLTagHashSet(23);
74      protected static String  [] formattedTags = {"pre", "script", "style", "textarea", "xmp"};
75                                      // "xmp" is obsolete but still encountered!
76  
77      static {
78          for (int i=0; i<formattedTags.length; i++) {
79              formattedTable.add(formattedTags[i]);
80          }
81      }
82  
83      public HTMLIndenter() {}
84  
85      /**
86      * Set the properties for this indenter
87      */
88  
89      public void setOutputProperties(Properties   props) {
90          String   s = props.getProperty(SaxonOutputKeys.INDENT_SPACES);
91          if (s==null) {
92              indentSpaces = 3;
93          } else {
94              try {
95                  indentSpaces = Integer.parseInt(s);
96              } catch (NumberFormatException   err) {
97                  indentSpaces = 3;
98              }
99          }
100     }
101 
102     /**
103     * Output element start tag
104     */
105 
106     public void startElement(int nameCode, int typeCode, int locationId, int properties) throws XPathException {
107         int tagProps = classifyTag(nameCode);
108         if (level >= propertyStack.length) {
109             int[] p2 = new int[level*2];
110             System.arraycopy(propertyStack, 0, p2, 0, propertyStack.length);
111             propertyStack = p2;
112         }
113         propertyStack[level] = tagProps;
114         isInlineTag = (tagProps & IS_INLINE) != 0;
115         inFormattedTag = inFormattedTag || ((tagProps & IS_FORMATTED) != 0);
116         if (!isInlineTag && !inFormattedTag &&
117              !afterInline && !afterFormatted) {
118             indent();
119         }
120 
121         super.startElement(nameCode, typeCode, locationId, properties);
122         level++;
123         sameLine = true;
124         afterInline = false;
125         afterFormatted = false;
126     }
127 
128     /**
129     * Output element end tag
130     */
131 
132     public void endElement() throws XPathException {
133         level--;
134         boolean thisInline = (propertyStack[level] & IS_INLINE) != 0;
135         boolean thisFormatted = (propertyStack[level] & IS_FORMATTED) != 0;
136         if (!thisInline && !thisFormatted && !afterInline &&
137                  !sameLine && !afterFormatted && !inFormattedTag) {
138             indent();
139             afterInline = false;
140             afterFormatted = false;
141         } else {
142             afterInline = thisInline;
143             afterFormatted = thisFormatted;
144         }
145         super.endElement();
146         inFormattedTag = inFormattedTag && !thisFormatted;
147         sameLine = false;
148     }
149 
150     /**
151     * Output character data
152     */
153 
154     public void characters(CharSequence   chars, int locationId, int properties) throws XPathException {
155         if (inFormattedTag) {
156             super.characters(chars, locationId, properties);
157         } else {
158             int lastNL = 0;
159 
160             for (int i=0; i<chars.length(); i++) {
161                 if (chars.charAt(i)=='\n' || (i-lastNL > 120 && chars.charAt(i)==' ')) {
162                     sameLine = false;
163                     super.characters(chars.subSequence(lastNL, i), locationId, properties);
164                     indent();
165                     lastNL = i+1;
166                     while (lastNL<chars.length() && chars.charAt(lastNL)==' ') {
167                         lastNL++;
168                     }
169                 }
170             }
171             if (lastNL < chars.length()) {
172                 super.characters(chars.subSequence(lastNL, chars.length()), locationId, properties);
173             }
174         }
175         afterInline = false;
176     }
177 
178     /**
179     * Output a comment
180     */
181 
182     public void comment(CharSequence   chars, int locationId, int properties) throws XPathException {
183         indent();
184         super.comment(chars, locationId, properties);
185     }
186 
187     /**
188     * Output white space to reflect the current indentation level
189     */
190 
191     private void indent() throws XPathException {
192         int spaces = level * indentSpaces;
193         while (spaces > indentChars.length()) {
194             indentChars += indentChars;
195         }
196         super.characters("\n", 0, 0);
197         super.characters(indentChars.subSequence(0, spaces), 0, 0);
198         sameLine = false;
199     }
200 
201 };
202 
203 //
204 // The contents of this file are subject to the Mozilla Public License Version 1.0 (the "License");
205 // you may not use this file except in compliance with the License. You may obtain a copy of the
206 // License at http://www.mozilla.org/MPL/
207 //
208 // Software distributed under the License is distributed on an "AS IS" basis,
209 // WITHOUT WARRANTY OF ANY KIND, either express or implied.
210 // See the License for the specific language governing rights and limitations under the License.
211 //
212 // The Original Code is: all this file.
213 //
214 // The Initial Developer of the Original Code is Michael H. Kay.
215 //
216 // Portions created by (your name) are Copyright (C) (your legal entity). All Rights Reserved.
217 //
218 // Contributor(s): none.
219 //
220 
221
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags