Config


1   // Jericho HTML Parser - Java based library for analysing and manipulating HTML
2   // Version 2.2
3   // Copyright (C) 2006 Martin Jericho
4   // http://sourceforge.net/projects/jerichohtml/
5   //
6   // This library is free software; you can redistribute it and/or
7   // modify it under the terms of the GNU Lesser General Public
8   // License as published by the Free Software Foundation; either
9   // version 2.1 of the License, or (at your option) any later version.
10  // http://www.gnu.org/copyleft/lesser.html
11  //
12  // This library is distributed in the hope that it will be useful,
13  // but WITHOUT ANY WARRANTY; without even the implied warranty of
14  // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
15  // Lesser General Public License for more details.
16  //
17  // You should have received a copy of the GNU Lesser General Public
18  // License along with this library; if not, write to the Free Software
19  // Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
20  
21  package au.id.jericho.lib.html;
22  
23  import java.util.*;
24  
25  /**
26   * Encapsulates global configuration properties which determine the behaviour of various functions.
27   * <p>
28   * All of the properties in this class are static, affecting all objects and threads.
29   * Multiple concurrent configurations are not possible.
30   * <p>
31   * Properties that relate to <a target="_blank" HREF="http://www.w3.org/TR/html401/conform.html#didx-user_agent">user agent</a>
32   * compatibility issues are stored in instances of the {@link Config.CompatibilityMode} class.
33   * This allows all of the properties in the compatibility mode to be set as a block by setting the static
34   * {@link #CurrentCompatibilityMode} property to a different instance.
35   *
36   * @see Config.CompatibilityMode
37   */ 
38  public final class Config {
39      private Config() {}
40  
41      /**
42       * Determines the string used to separate a single column's multiple values in the output of the {@link FormFields#getColumnValues(Map)} method.
43       * <p>
44       * The situation where a single column has multiple values only arises if {@link FormField#getUserValueCount()}<code>&gt;1</code>
45       * on the relevant form field, which usually indicates a poorly designed form.
46       * <p>
47       * The default value is "<code>,</code>" (a comma, not including the quotes).
48       * <p>
49       * Must not be <code>null</code>.
50       */
51      public static String   ColumnMultipleValueSeparator=",";
52  
53      /**
54       * Determines the string that represents the value <code>true</code> in the output of the {@link FormFields#getColumnValues(Map)} method.
55       * <p>
56       * The default value is "<code>true</code>" (without the quotes).
57       * <p>
58       * Must not be <code>null</code>.
59       */
60      public static String   ColumnValueTrue=Boolean.toString(true);
61  
62      /**
63       * Determines the string that represents the value <code>false</code> in the output of the {@link FormFields#getColumnValues(Map)} method.
64       * <p>
65       * The default value is <code>null</code>, which represents no output at all.
66       */
67      public static String   ColumnValueFalse=null;
68  
69      /**
70       * Determines the currently active {@linkplain Config.CompatibilityMode compatibility mode}.
71       * <p>
72       * The default setting is {@link Config.CompatibilityMode#IE} (MS Internet Explorer 6.0).
73       * <p>
74       * Must not be <code>null</code>.
75       */
76      public static CompatibilityMode CurrentCompatibilityMode=CompatibilityMode.IE;
77  
78      /**
79       * Determines whether apostrophes are encoded when calling the {@link CharacterReference#encode(CharSequence)} method.
80       * <p>
81       * A value of <code>false</code> means {@linkplain CharacterEntityReference#_apos apostrophe}
82       * (U+0027) characters are not encoded.
83       * The only time apostrophes need to be encoded is within an attribute value delimited by
84       * single quotes (apostrophes), so in most cases ignoring apostrophes is perfectly safe and
85       * enhances the readability of the source document.
86       * <p>
87       * Note that apostrophes are always encoded as a {@linkplain NumericCharacterReference numeric character reference}, never as the
88       * character entity reference {@link CharacterEntityReference#_apos &amp;apos;}.
89       * <p>
90       * The default value is <code>false</code>.
91       */
92      public static boolean IsApostropheEncoded=false;
93  
94      /**
95       * Used in Element.getChildElements.
96       * Will only make this public if someone makes a convincing argument why you would ever need to include server tags in an element hierarchy.
97       */ 
98      static final boolean IncludeServerTagsInElementHierarchy=false;
99  
100     /**
101      * Represents a set of maximum unicode code points to be recognised for the three types of
102      * <a HREF="CharacterReference.html#Unterminated">unterminated</a> character reference in a given context.
103      * <p>
104      * The three types of character reference are:
105      * <ul>
106      *  <li>{@linkplain CharacterEntityReference Character entity reference}
107      *  <li><a HREF="NumericCharacterReference.html#DecimalCharacterReference">Decimal character reference</a>
108      *  <li><a HREF="NumericCharacterReference.html#HexadecimalCharacterReference">Hexadecimal character reference</a>
109      * </ul>
110      * <p>
111      * The two types of contexts used in this library are:
112      * <ul>
113      *  <li>Inside an attribute value
114      *  <li>Outside an attribute value
115      * </ul>
116      */ 
117     static class UnterminatedCharacterReferenceSettings {
118         // use volatile fields to make them thread safe
119         public volatile int characterEntityReferenceMaxCodePoint;
120         public volatile int decimalCharacterReferenceMaxCodePoint;
121         public volatile int hexadecimalCharacterReferenceMaxCodePoint;
122 
123         public static UnterminatedCharacterReferenceSettings ACCEPT_ALL=new UnterminatedCharacterReferenceSettings(CompatibilityMode.CODE_POINTS_ALL,CompatibilityMode.CODE_POINTS_ALL,CompatibilityMode.CODE_POINTS_ALL);
124 
125         public UnterminatedCharacterReferenceSettings() {
126             this(CompatibilityMode.CODE_POINTS_NONE,CompatibilityMode.CODE_POINTS_NONE,CompatibilityMode.CODE_POINTS_NONE);
127         }
128 
129         public UnterminatedCharacterReferenceSettings(final int characterEntityReferenceMaxCodePoint, final int decimalCharacterReferenceMaxCodePoint, final int hexadecimalCharacterReferenceMaxCodePoint) {
130             this.characterEntityReferenceMaxCodePoint=characterEntityReferenceMaxCodePoint;
131             this.decimalCharacterReferenceMaxCodePoint=decimalCharacterReferenceMaxCodePoint;
132             this.hexadecimalCharacterReferenceMaxCodePoint=hexadecimalCharacterReferenceMaxCodePoint;
133         }
134 
135         public String   toString() {
136             return "\n    Character entity reference: "+getDescription(characterEntityReferenceMaxCodePoint)
137                         +"\n    Decimal character reference: "+getDescription(decimalCharacterReferenceMaxCodePoint)
138                         +"\n    Haxadecimal character reference: "+getDescription(hexadecimalCharacterReferenceMaxCodePoint);
139         }
140 
141         private String   getDescription(final int codePoint) {
142             if (codePoint==CompatibilityMode.CODE_POINTS_NONE) return "None";
143             if (codePoint==CompatibilityMode.CODE_POINTS_ALL) return "All";
144             return "0x"+Integer.toString(codePoint,16);
145         }
146     }
147 
148     /**
149      * Represents a set of configuration parameters that relate to
150      * <a target="_blank" HREF="http://www.w3.org/TR/html401/conform.html#didx-user_agent">user agent</a> compatibility issues.
151      * <p>
152      * The predefined compatibility modes {@link #IE}, {@link #MOZILLA}, {@link #OPERA} and {@link #XHTML} provide an easy means of
153      * ensuring the library interprets the markup in a way consistent with some of the most commonly used browsers,
154      * at least in relation to the behaviour described by the properties in this class.
155      * <p>
156      * The properties of any <code>CompatibilityMode</code> object can be modified individually, including those in
157      * the predefined instances as well as newly constructed instances.
158      * Take note however that modifying the properties of the predefined instances has a global affect.
159      * <p>
160      * The currently active compatibility mode is stored in the static 
161      * {@link Config#CurrentCompatibilityMode} property.
162      * <p>
163      */
164     public static final class CompatibilityMode {
165         private String   name;
166         private volatile boolean formFieldNameCaseInsensitive;
167         volatile UnterminatedCharacterReferenceSettings unterminatedCharacterReferenceSettingsInsideAttributeValue;
168         volatile UnterminatedCharacterReferenceSettings unterminatedCharacterReferenceSettingsOutsideAttributeValue;
169 
170         /**
171          * Indicates the recognition of all unicode code points.
172          * <p>
173          * This value is used in properties which specify a maximum unicode code point to be recognised by the parser.
174          *
175          * @see #getUnterminatedCharacterEntityReferenceMaxCodePoint(boolean insideAttributeValue)
176          * @see #getUnterminatedDecimalCharacterReferenceMaxCodePoint(boolean insideAttributeValue)
177          * @see #getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(boolean insideAttributeValue)
178          */
179         public static final int CODE_POINTS_ALL=CharacterReference.MAX_CODE_POINT;
180 
181         /**
182          * Indicates the recognition of no unicode code points.
183          * <p>
184          * This value is used in properties which specify a maximum unicode code point to be recognised by the parser.
185          *
186          * @see #getUnterminatedCharacterEntityReferenceMaxCodePoint(boolean insideAttributeValue)
187          * @see #getUnterminatedDecimalCharacterReferenceMaxCodePoint(boolean insideAttributeValue)
188          * @see #getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(boolean insideAttributeValue)
189          */
190         public static final int CODE_POINTS_NONE=CharacterReference.INVALID_CODE_POINT;
191 
192         /**
193          * <a target="_blank" HREF="http://www.microsoft.com/windows/ie/">Microsoft Internet Explorer</a> compatibility mode.
194          * <p>
195          * <code>{@link #getName() Name} = IE</code><br />
196          * <code>{@link #isFormFieldNameCaseInsensitive() FormFieldNameCaseInsensitive} = true</code><br />
197          * <table cellspacing="0" cellpadding="0">
198          *  <tr><th>Recognition of unterminated character references:<th><th align="center">&nbsp; (inside attribute) &nbsp;<th align="center">&nbsp; (outside attribute) &nbsp;
199          *  <tr><td>{@link #getUnterminatedCharacterEntityReferenceMaxCodePoint(boolean) UnterminatedCharacterEntityReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">U+00FF<td align="center">U+00FF
200          *  <tr><td>{@link #getUnterminatedDecimalCharacterReferenceMaxCodePoint(boolean) UnterminatedDecimalCharacterReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_ALL All}<td align="center">{@linkplain #CODE_POINTS_ALL All}
201          *  <tr><td>{@link #getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(boolean) UnterminatedHexadecimalCharacterReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_ALL All}<td align="center">{@linkplain #CODE_POINTS_NONE None}
202          * </table>      
203          */
204         public static final CompatibilityMode IE=new CompatibilityMode("IE",true,
205             new UnterminatedCharacterReferenceSettings(0xFF, CODE_POINTS_ALL, CODE_POINTS_ALL), // inside attributes
206             new UnterminatedCharacterReferenceSettings(0xFF, CODE_POINTS_ALL, CODE_POINTS_NONE) // outside attributes
207         );
208 
209         /**
210          * <a target="_blank" HREF="http://www.mozilla.org/products/mozilla1.x/">Mozilla</a> / 
211          * <a target="_blank" HREF="http://www.mozilla.org/products/firefox/">Firefox</a> /
212          * <a target="_blank" HREF="http://browser.netscape.com/">Netscape</a> compatibility mode.
213          * <p>
214          * <code>{@link #getName() Name} = Mozilla</code><br />
215          * <code>{@link #isFormFieldNameCaseInsensitive() FormFieldNameCaseInsensitive} = false</code><br />
216          * <table cellspacing="0" cellpadding="0">
217          *  <tr><th>Recognition of unterminated character references:<th><th align="center">&nbsp; (inside attribute) &nbsp;<th align="center">&nbsp; (outside attribute) &nbsp;
218          *  <tr><td>{@link #getUnterminatedCharacterEntityReferenceMaxCodePoint(boolean) UnterminatedCharacterEntityReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">U+00FF<td align="center">{@linkplain #CODE_POINTS_ALL All}
219          *  <tr><td>{@link #getUnterminatedDecimalCharacterReferenceMaxCodePoint(boolean) UnterminatedDecimalCharacterReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_ALL All}<td align="center">{@linkplain #CODE_POINTS_ALL All}
220          *  <tr><td>{@link #getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(boolean) UnterminatedHexadecimalCharacterReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_ALL All}<td align="center">{@linkplain #CODE_POINTS_ALL All}
221          * </table>      
222          */
223         public static final CompatibilityMode MOZILLA=new CompatibilityMode("Mozilla",false,
224             new UnterminatedCharacterReferenceSettings(0xFF,            CODE_POINTS_ALL, CODE_POINTS_ALL), // inside attributes
225             new UnterminatedCharacterReferenceSettings(CODE_POINTS_ALL, CODE_POINTS_ALL, CODE_POINTS_ALL) // outside attributes
226         );
227 
228         /**
229          * Opera compatibility mode.
230          * <p>
231          * <code>{@link #getName() Name} = Opera</code><br />
232          * <code>{@link #isFormFieldNameCaseInsensitive() FormFieldNameCaseInsensitive} = true</code><br />
233          * <table cellspacing="0" cellpadding="0">
234          *  <tr><th>Recognition of unterminated character references:<th><th align="center">&nbsp; (inside attribute) &nbsp;<th align="center">&nbsp; (outside attribute) &nbsp;
235          *  <tr><td>{@link #getUnterminatedCharacterEntityReferenceMaxCodePoint(boolean) UnterminatedCharacterEntityReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">U+003E<td align="center">{@linkplain #CODE_POINTS_ALL All}
236          *  <tr><td>{@link #getUnterminatedDecimalCharacterReferenceMaxCodePoint(boolean) UnterminatedDecimalCharacterReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_ALL All}<td align="center">{@linkplain #CODE_POINTS_ALL All}
237          *  <tr><td>{@link #getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(boolean) UnterminatedHexadecimalCharacterReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_ALL All}<td align="center">{@linkplain #CODE_POINTS_ALL All}
238          * </table>      
239          */
240         public static final CompatibilityMode OPERA=new CompatibilityMode("Opera",true,
241             new UnterminatedCharacterReferenceSettings(0x3E,            CODE_POINTS_ALL, CODE_POINTS_ALL), // inside attributes
242             new UnterminatedCharacterReferenceSettings(CODE_POINTS_ALL, CODE_POINTS_ALL, CODE_POINTS_ALL) // outside attributes
243         );
244 
245         /**
246          * <a target="_blank" HREF="http://www.w3.org/TR/xhtml1/#xhtml">XHTML</a> compatibility mode.
247          * <p>
248          * <code>{@link #getName() Name} = XHTML</code><br />
249          * <code>{@link #isFormFieldNameCaseInsensitive() FormFieldNameCaseInsensitive} = false</code><br />
250          * <table cellspacing="0" cellpadding="0">
251          *  <tr><th>Recognition of unterminated character references:<th><th align="center">&nbsp; (inside attribute) &nbsp;<th align="center">&nbsp; (outside attribute) &nbsp;
252          *  <tr><td>{@link #getUnterminatedCharacterEntityReferenceMaxCodePoint(boolean) UnterminatedCharacterEntityReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_NONE None}<td align="center">{@linkplain #CODE_POINTS_NONE None}
253          *  <tr><td>{@link #getUnterminatedDecimalCharacterReferenceMaxCodePoint(boolean) UnterminatedDecimalCharacterReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_NONE None}<td align="center">{@linkplain #CODE_POINTS_NONE None}
254          *  <tr><td>{@link #getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(boolean) UnterminatedHexadecimalCharacterReferenceMaxCodePoint}<td><code>&nbsp;=</code><td align="center">{@linkplain #CODE_POINTS_NONE None}<td align="center">{@linkplain #CODE_POINTS_NONE None}
255          * </table>      
256          */
257         public static final CompatibilityMode XHTML=new CompatibilityMode("XHTML");
258 
259         /**
260          * Constructs a new <code>CompatibilityMode</code> with the given {@linkplain #getName() name}.
261          * <p>
262          * All properties in the new instance are initially assigned their default values, which are the same as the strict
263          * rules of the {@link #XHTML} compatibility mode.
264          *
265          * @param name  the {@linkplain #getName() name} of the new compatibility mode
266          */
267         public CompatibilityMode(final String   name) {
268             this(name,false,new UnterminatedCharacterReferenceSettings(),new UnterminatedCharacterReferenceSettings());
269         }
270 
271         private CompatibilityMode(final String   name, final boolean formFieldNameCaseInsensitive, final UnterminatedCharacterReferenceSettings unterminatedCharacterReferenceSettingsInsideAttributeValue, final UnterminatedCharacterReferenceSettings unterminatedCharacterReferenceSettingsOutsideAttributeValue) {
272             this.name=name;
273             this.formFieldNameCaseInsensitive=formFieldNameCaseInsensitive;
274             this.unterminatedCharacterReferenceSettingsInsideAttributeValue=unterminatedCharacterReferenceSettingsInsideAttributeValue;
275             this.unterminatedCharacterReferenceSettingsOutsideAttributeValue=unterminatedCharacterReferenceSettingsOutsideAttributeValue;
276         }
277 
278         /**
279          * Returns the name of this compatibility mode.
280          * @return the name of this compatibility mode.
281          */
282         public String   getName() {
283             return name;
284         }
285 
286         /**
287          * Indicates whether {@linkplain FormField#getName() form field names} are treated as case insensitive.
288          * <p>
289          * Microsoft Internet Explorer treats field names as case insensitive,
290          * while Mozilla treats them as case sensitive.
291          * <p>
292          * The value of this property in the {@linkplain Config#CurrentCompatibilityMode current compatibility mode}
293          * affects all instances of the {@link FormFields} class.
294          * It should be set to the desired configuration before any instances of <code>FormFields</code> are created.
295          *
296          * @return <code>true</code> if {@linkplain FormField#getName() form field names} are treated as case insensitive, otherwise <code>false</code>.
297          * @see #setFormFieldNameCaseInsensitive(boolean)
298          */
299         public boolean isFormFieldNameCaseInsensitive() {
300             return formFieldNameCaseInsensitive;
301         }
302 
303         /**
304          * Sets whether {@linkplain FormField#getName() form field names} are treated as case insensitive.
305          * <p>
306          * See {@link #isFormFieldNameCaseInsensitive()} for the documentation of this property.
307          *
308          * @param value  the new value of the property
309          */
310         public void setFormFieldNameCaseInsensitive(final boolean value) {
311             formFieldNameCaseInsensitive=value;
312         }
313 
314         /**
315          * Returns the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a>
316          * {@linkplain CharacterEntityReference character entity reference} which is to be recognised in the specified context.
317          * <p>
318          * For example, if <code>getUnterminatedCharacterEntityReferenceMaxCodePoint(true)</code> has the value <code>0xFF</code> (U+00FF)
319          * in the {@linkplain Config#CurrentCompatibilityMode current compatibility mode}, then:
320          * <ul>
321          *  <li>{@link CharacterReference#decode(CharSequence,boolean) CharacterReference.decode("&amp;gt",true)}
322          *   returns "<code>&gt;</code>".<br />
323          *   The string is recognised as the character entity reference {@link CharacterEntityReference#_gt &amp;gt;}
324          *   despite the fact that it is <a HREF="CharacterReference.html#Unterminated">unterminated</a>,
325          *   because its unicode code point U+003E is below the maximum of U+00FF set by this property.
326          *  <li>{@link CharacterReference#decode(CharSequence,boolean) CharacterReference.decode("&amp;euro",true)}
327          *   returns "<code>&amp;euro</code>".<br />
328          *   The string is not recognised as the character entity reference {@link CharacterEntityReference#_euro &amp;euro;}
329          *   because it is <a HREF="CharacterReference.html#Unterminated">unterminated</a>
330          *   and its unicode code point U+20AC is above the maximum of U+00FF set by this property.
331          * </ul>
332          * <p>
333          * See the documentation of the {@link Attribute#getValue()} method for further discussion.
334          * 
335          * @param insideAttributeValue  the context within an HTML document - <code>true</code> if inside an attribute value or <code>false</code> if outside an attribute value.
336          * @return the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a> {@linkplain CharacterEntityReference character entity reference} which is to be recognised in the specified context.
337          * @see #setUnterminatedCharacterEntityReferenceMaxCodePoint(boolean insideAttributeValue, int maxCodePoint)
338          */
339         public int getUnterminatedCharacterEntityReferenceMaxCodePoint(final boolean insideAttributeValue) {
340             return getUnterminatedCharacterReferenceSettings(insideAttributeValue).characterEntityReferenceMaxCodePoint;
341         }
342 
343         /**
344          * Sets the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a>
345          * {@linkplain CharacterEntityReference character entity reference} which is to be recognised in the specified context.
346          * <p>
347          * See {@link #getUnterminatedCharacterEntityReferenceMaxCodePoint(boolean insideAttributeValue)} for the documentation of this property.
348          *
349          * @param insideAttributeValue  the context within an HTML document - <code>true</code> if inside an attribute value or <code>false</code> if outside an attribute value.
350          * @param maxCodePoint  the maximum unicode code point.
351          */
352         public void setUnterminatedCharacterEntityReferenceMaxCodePoint(final boolean insideAttributeValue, final int maxCodePoint) {
353             getUnterminatedCharacterReferenceSettings(insideAttributeValue).characterEntityReferenceMaxCodePoint=maxCodePoint;
354         }
355 
356         /**
357          * Returns the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a>
358          * <a HREF="NumericCharacterReference.html#DecimalCharacterReference">decimal character reference</a> which is to be recognised in the specified context.
359          * <p>
360          * For example, if <code>getUnterminatedDecimalCharacterReferenceMaxCodePoint(true)</code> had the hypothetical value <code>0xFF</code> (U+00FF)
361          * in the {@linkplain Config#CurrentCompatibilityMode current compatibility mode}, then:
362          * <ul>
363          *  <li>{@link CharacterReference#decode(CharSequence,boolean) CharacterReference.decode("&amp;#62",true)}
364          *   returns "<code>&gt;</code>".<br />
365          *   The string is recognised as the numeric character reference <code>&amp;#62;</code>
366          *   despite the fact that it is <a HREF="CharacterReference.html#Unterminated">unterminated</a>,
367          *   because its unicode code point U+003E is below the maximum of U+00FF set by this property.
368          *  <li>{@link CharacterReference#decode(CharSequence,boolean) CharacterReference.decode("&amp;#8364",true)}
369          *   returns "<code>&amp;#8364</code>".<br />
370          *   The string is not recognised as the numeric character reference <code>&amp;#8364;</code>
371          *   because it is <a HREF="CharacterReference.html#Unterminated">unterminated</a>
372          *   and its unicode code point U+20AC is above the maximum of U+00FF set by this property.
373          * </ul>
374          * 
375          * @param insideAttributeValue  the context within an HTML document - <code>true</code> if inside an attribute value or <code>false</code> if outside an attribute value.
376          * @return the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a> <a HREF="NumericCharacterReference.html#DecimalCharacterReference">decimal character reference</a> which is to be recognised in the specified context.
377          * @see #setUnterminatedDecimalCharacterReferenceMaxCodePoint(boolean insideAttributeValue, int maxCodePoint)
378          */
379         public int getUnterminatedDecimalCharacterReferenceMaxCodePoint(final boolean insideAttributeValue) {
380             return getUnterminatedCharacterReferenceSettings(insideAttributeValue).decimalCharacterReferenceMaxCodePoint;
381         }
382 
383         /**
384          * Sets the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a>
385          * <a HREF="NumericCharacterReference.html#DecimalCharacterReference">decimal character reference</a> which is to be recognised in the specified context.
386          * <p>
387          * See {@link #getUnterminatedDecimalCharacterReferenceMaxCodePoint(boolean insideAttributeValue)} for the documentation of this property.
388          *
389          * @param insideAttributeValue  the context within an HTML document - <code>true</code> if inside an attribute value or <code>false</code> if outside an attribute value.
390          * @param maxCodePoint  the maximum unicode code point.
391          */
392         public void setUnterminatedDecimalCharacterReferenceMaxCodePoint(final boolean insideAttributeValue, final int maxCodePoint) {
393             getUnterminatedCharacterReferenceSettings(insideAttributeValue).decimalCharacterReferenceMaxCodePoint=maxCodePoint;
394         }
395 
396         /**
397          * Returns the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a>
398          * <a HREF="NumericCharacterReference.html#HexadecimalCharacterReference">hexadecimal character reference</a> which is to be recognised in the specified context.
399          * <p>
400          * For example, if <code>getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(true)</code> had the hypothetical value <code>0xFF</code> (U+00FF)
401          * in the {@linkplain Config#CurrentCompatibilityMode current compatibility mode}, then:
402          * <ul>
403          *  <li>{@link CharacterReference#decode(CharSequence,boolean) CharacterReference.decode("&amp;#x3e",true)}
404          *   returns "<code>&gt;</code>".<br />
405          *   The string is recognised as the numeric character reference <code>&amp;#x3e;</code>
406          *   despite the fact that it is <a HREF="CharacterReference.html#Unterminated">unterminated</a>,
407          *   because its unicode code point U+003E is below the maximum of U+00FF set by this property.
408          *  <li>{@link CharacterReference#decode(CharSequence,boolean) CharacterReference.decode("&amp;#x20ac",true)}
409          *   returns "<code>&amp;#x20ac</code>".<br />
410          *   The string is not recognised as the numeric character reference <code>&amp;#20ac;</code>
411          *   because it is <a HREF="CharacterReference.html#Unterminated">unterminated</a>
412          *   and its unicode code point U+20AC is above the maximum of U+00FF set by this property.
413          * </ul>
414          * 
415          * @param insideAttributeValue  the context within an HTML document - <code>true</code> if inside an attribute value or <code>false</code> if outside an attribute value.
416          * @return the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a> <a HREF="NumericCharacterReference.html#HexadecimalCharacterReference">hexadecimal character reference</a> which is to be recognised in the specified context.
417          * @see #setUnterminatedHexadecimalCharacterReferenceMaxCodePoint(boolean insideAttributeValue, int maxCodePoint)
418          */
419         public int getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(final boolean insideAttributeValue) {
420             return getUnterminatedCharacterReferenceSettings(insideAttributeValue).hexadecimalCharacterReferenceMaxCodePoint;
421         }
422 
423         /**
424          * Sets the maximum unicode code point of an <a HREF="CharacterReference.html#Unterminated">unterminated</a>
425          * <a HREF="NumericCharacterReference.html#HexadecimalCharacterReference">headecimal character reference</a> which is to be recognised in the specified context.
426          * <p>
427          * See {@link #getUnterminatedHexadecimalCharacterReferenceMaxCodePoint(boolean insideAttributeValue)} for the documentation of this property.
428          *
429          * @param insideAttributeValue  the context within an HTML document - <code>true</code> if inside an attribute value or <code>false</code> if outside an attribute value.
430          * @param maxCodePoint  the maximum unicode code point.
431          */
432         public void setUnterminatedHexadecimalCharacterReferenceMaxCodePoint(final boolean insideAttributeValue, final int maxCodePoint) {
433             getUnterminatedCharacterReferenceSettings(insideAttributeValue).hexadecimalCharacterReferenceMaxCodePoint=maxCodePoint;
434         }
435 
436         /**
437          * Returns a string representation of this object useful for debugging purposes.
438          * @return a string representation of this object useful for debugging purposes.
439          */
440         public String   getDebugInfo() {
441             return "Form field name case insensitive: "+formFieldNameCaseInsensitive
442                 +"\nMaximum codepoints in unterminated character references:"
443                 +"\n  Inside attribute values:"
444                 +unterminatedCharacterReferenceSettingsInsideAttributeValue
445                 +"\n  Outside attribute values:"
446                 +unterminatedCharacterReferenceSettingsOutsideAttributeValue;
447         }
448     
449         /**
450          * Returns the {@linkplain #getName() name} of this compatibility mode.
451          * @return the {@linkplain #getName() name} of this compatibility mode.
452          */
453         public String   toString() {
454             return getName();
455         }
456 
457         UnterminatedCharacterReferenceSettings getUnterminatedCharacterReferenceSettings(final boolean insideAttributeValue) {
458             return insideAttributeValue ? unterminatedCharacterReferenceSettingsInsideAttributeValue : unterminatedCharacterReferenceSettingsOutsideAttributeValue;
459         }
460     }
461 }
462
A to Z: JavaDoc & Examples Daily Java News & Articles Open Source Projects Open Source Codes Free Computer Books Remove Frame
Popular Tags