View Javadoc
1   /*
2    * Licensed to the Apache Software Foundation (ASF) under one   *
3    * or more contributor license agreements.  See the NOTICE file *
4    * distributed with this work for additional information        *
5    * regarding copyright ownership.  The ASF licenses this file   *
6    * to you under the Apache License, Version 2.0 (the            *
7    * "License"); you may not use this file except in compliance   *
8    * with the License.  You may obtain a copy of the License at   *
9    *                                                              *
10   *   http://www.apache.org/licenses/LICENSE-2.0                 *
11   *                                                              *
12   * Unless required by applicable law or agreed to in writing,   *
13   * software distributed under the License is distributed on an  *
14   * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY       *
15   * KIND, either express or implied.  See the License for the    *
16   * specific language governing permissions and limitations      *
17   * under the License.                                           *
18   */
19  package org.apache.rat.report.xml.writer.impl.base;
20  
21  import java.io.IOException;
22  import java.io.Writer;
23  import java.util.ArrayDeque;
24  import java.util.Arrays;
25  import java.util.HashSet;
26  import java.util.Set;
27  
28  import org.apache.rat.report.xml.writer.IXmlWriter;
29  import org.apache.rat.report.xml.writer.InvalidXmlException;
30  import org.apache.rat.report.xml.writer.OperationNotAllowedException;
31  
32  /**
33   * <p>
34   * Lightweight {@link IXmlWriter} implementation.
35   * </p>
36   * <p>
37   * Requires a wrapper to be used safely in a multithreaded environment.
38   * </p>
39   * <p>
40   * Not intended to be subclassed. Please copy and hack!
41   * </p>
42   */
43  public final class XmlWriter implements IXmlWriter {
44  
45      private static final byte NAME_START_MASK = 1 << 1;
46      private static final byte NAME_MASK = 1 << 2;
47      private static final byte NAME_BODY_CHAR = NAME_MASK;
48      private static final byte NAME_START_OR_BODY_CHAR = NAME_MASK | NAME_START_MASK;
49  
50      private final static boolean[] ALLOWED_CHARACTERS = new boolean[1 << 16];
51  
52      static {
53          Arrays.fill(ALLOWED_CHARACTERS, false);
54          ALLOWED_CHARACTERS[0x9] = true;
55          ALLOWED_CHARACTERS[0xA] = true;
56          ALLOWED_CHARACTERS[0xD] = true;
57          Arrays.fill(ALLOWED_CHARACTERS, 0x20, 0xD7FF, true);
58          Arrays.fill(ALLOWED_CHARACTERS, 0xE000, 0xFFFD, true);
59      }
60  
61      private final static byte[] CHARACTER_CODES = new byte[1 << 16];
62  
63      static {
64          // Name ::= (Letter | '_' | ':') (NameChar)*
65          CHARACTER_CODES['_'] = NAME_START_OR_BODY_CHAR;
66          CHARACTER_CODES[':'] = NAME_START_OR_BODY_CHAR;
67          // Letter ::= BaseChar | Ideographic
68          // BaseChar
69          Arrays.fill(CHARACTER_CODES, 0x0041, 0x005A, NAME_START_OR_BODY_CHAR);
70          Arrays.fill(CHARACTER_CODES, 0x0061, 0x007A, NAME_START_OR_BODY_CHAR);
71          Arrays.fill(CHARACTER_CODES, 0x00C0, 0x00D6, NAME_START_OR_BODY_CHAR);
72          Arrays.fill(CHARACTER_CODES, 0x00D8, 0x00F6, NAME_START_OR_BODY_CHAR);
73          Arrays.fill(CHARACTER_CODES, 0x00F8, 0x00FF, NAME_START_OR_BODY_CHAR);
74          Arrays.fill(CHARACTER_CODES, 0x0100, 0x0131, NAME_START_OR_BODY_CHAR);
75          Arrays.fill(CHARACTER_CODES, 0x0134, 0x013E, NAME_START_OR_BODY_CHAR);
76          Arrays.fill(CHARACTER_CODES, 0x0141, 0x0148, NAME_START_OR_BODY_CHAR);
77          Arrays.fill(CHARACTER_CODES, 0x014A, 0x017E, NAME_START_OR_BODY_CHAR);
78          Arrays.fill(CHARACTER_CODES, 0x0180, 0x01C3, NAME_START_OR_BODY_CHAR);
79          Arrays.fill(CHARACTER_CODES, 0x01CD, 0x01F0, NAME_START_OR_BODY_CHAR);
80          Arrays.fill(CHARACTER_CODES, 0x01F4, 0x01F5, NAME_START_OR_BODY_CHAR);
81          Arrays.fill(CHARACTER_CODES, 0x01FA, 0x0217, NAME_START_OR_BODY_CHAR);
82          Arrays.fill(CHARACTER_CODES, 0x0250, 0x02A8, NAME_START_OR_BODY_CHAR);
83          Arrays.fill(CHARACTER_CODES, 0x02BB, 0x02C1, NAME_START_OR_BODY_CHAR);
84          CHARACTER_CODES[0x0386] = NAME_START_OR_BODY_CHAR;
85          Arrays.fill(CHARACTER_CODES, 0x0388, 0x038A, NAME_START_OR_BODY_CHAR);
86          CHARACTER_CODES[0x038C] = NAME_START_OR_BODY_CHAR;
87          Arrays.fill(CHARACTER_CODES, 0x038E, 0x03A1, NAME_START_OR_BODY_CHAR);
88          Arrays.fill(CHARACTER_CODES, 0x03A3, 0x03CE, NAME_START_OR_BODY_CHAR);
89          Arrays.fill(CHARACTER_CODES, 0x03D0, 0x03D6, NAME_START_OR_BODY_CHAR);
90          CHARACTER_CODES[0x03DA] = NAME_START_OR_BODY_CHAR;
91          CHARACTER_CODES[0x03DC] = NAME_START_OR_BODY_CHAR;
92          CHARACTER_CODES[0x03DE] = NAME_START_OR_BODY_CHAR;
93          CHARACTER_CODES[0x03E0] = NAME_START_OR_BODY_CHAR;
94          Arrays.fill(CHARACTER_CODES, 0x03E2, 0x03F3, NAME_START_OR_BODY_CHAR);
95          Arrays.fill(CHARACTER_CODES, 0x0401, 0x040C, NAME_START_OR_BODY_CHAR);
96          Arrays.fill(CHARACTER_CODES, 0x040E, 0x044F, NAME_START_OR_BODY_CHAR);
97          Arrays.fill(CHARACTER_CODES, 0x0451, 0x045C, NAME_START_OR_BODY_CHAR);
98          Arrays.fill(CHARACTER_CODES, 0x045E, 0x0481, NAME_START_OR_BODY_CHAR);
99          Arrays.fill(CHARACTER_CODES, 0x0490, 0x04C4, NAME_START_OR_BODY_CHAR);
100         Arrays.fill(CHARACTER_CODES, 0x04C7, 0x04C8, NAME_START_OR_BODY_CHAR);
101         Arrays.fill(CHARACTER_CODES, 0x04CB, 0x04CC, NAME_START_OR_BODY_CHAR);
102         Arrays.fill(CHARACTER_CODES, 0x04D0, 0x04EB, NAME_START_OR_BODY_CHAR);
103         Arrays.fill(CHARACTER_CODES, 0x04EE, 0x04F5, NAME_START_OR_BODY_CHAR);
104         Arrays.fill(CHARACTER_CODES, 0x04F8, 0x04F9, NAME_START_OR_BODY_CHAR);
105         Arrays.fill(CHARACTER_CODES, 0x0531, 0x0556, NAME_START_OR_BODY_CHAR);
106         CHARACTER_CODES[0x0559] = NAME_START_OR_BODY_CHAR;
107         Arrays.fill(CHARACTER_CODES, 0x0561, 0x0586, NAME_START_OR_BODY_CHAR);
108         Arrays.fill(CHARACTER_CODES, 0x05D0, 0x05EA, NAME_START_OR_BODY_CHAR);
109         Arrays.fill(CHARACTER_CODES, 0x05F0, 0x05F2, NAME_START_OR_BODY_CHAR);
110         Arrays.fill(CHARACTER_CODES, 0x0621, 0x063A, NAME_START_OR_BODY_CHAR);
111         Arrays.fill(CHARACTER_CODES, 0x0641, 0x064A, NAME_START_OR_BODY_CHAR);
112         Arrays.fill(CHARACTER_CODES, 0x0671, 0x06B7, NAME_START_OR_BODY_CHAR);
113         Arrays.fill(CHARACTER_CODES, 0x06BA, 0x06BE, NAME_START_OR_BODY_CHAR);
114         Arrays.fill(CHARACTER_CODES, 0x06C0, 0x06CE, NAME_START_OR_BODY_CHAR);
115         Arrays.fill(CHARACTER_CODES, 0x06D0, 0x06D3, NAME_START_OR_BODY_CHAR);
116         CHARACTER_CODES[0x06D5] = NAME_START_OR_BODY_CHAR;
117         Arrays.fill(CHARACTER_CODES, 0x06E5, 0x06E6, NAME_START_OR_BODY_CHAR);
118         Arrays.fill(CHARACTER_CODES, 0x0905, 0x0939, NAME_START_OR_BODY_CHAR);
119         CHARACTER_CODES[0x093D] = NAME_START_OR_BODY_CHAR;
120         Arrays.fill(CHARACTER_CODES, 0x0958, 0x0961, NAME_START_OR_BODY_CHAR);
121         Arrays.fill(CHARACTER_CODES, 0x0985, 0x098C, NAME_START_OR_BODY_CHAR);
122         Arrays.fill(CHARACTER_CODES, 0x098F, 0x0990, NAME_START_OR_BODY_CHAR);
123         Arrays.fill(CHARACTER_CODES, 0x0993, 0x09A8, NAME_START_OR_BODY_CHAR);
124         Arrays.fill(CHARACTER_CODES, 0x09AA, 0x09B0, NAME_START_OR_BODY_CHAR);
125         CHARACTER_CODES[0x09B2] = NAME_START_OR_BODY_CHAR;
126         Arrays.fill(CHARACTER_CODES, 0x09B6, 0x09B9, NAME_START_OR_BODY_CHAR);
127         Arrays.fill(CHARACTER_CODES, 0x09DC, 0x09DD, NAME_START_OR_BODY_CHAR);
128         Arrays.fill(CHARACTER_CODES, 0x09DF, 0x09E1, NAME_START_OR_BODY_CHAR);
129         Arrays.fill(CHARACTER_CODES, 0x09F0, 0x09F1, NAME_START_OR_BODY_CHAR);
130         Arrays.fill(CHARACTER_CODES, 0x0A05, 0x0A0A, NAME_START_OR_BODY_CHAR);
131         Arrays.fill(CHARACTER_CODES, 0x0A0F, 0x0A10, NAME_START_OR_BODY_CHAR);
132         Arrays.fill(CHARACTER_CODES, 0x0A13, 0x0A28, NAME_START_OR_BODY_CHAR);
133         Arrays.fill(CHARACTER_CODES, 0x0A2A, 0x0A30, NAME_START_OR_BODY_CHAR);
134         Arrays.fill(CHARACTER_CODES, 0x0A32, 0x0A33, NAME_START_OR_BODY_CHAR);
135         Arrays.fill(CHARACTER_CODES, 0x0A35, 0x0A36, NAME_START_OR_BODY_CHAR);
136         Arrays.fill(CHARACTER_CODES, 0x0A38, 0x0A39, NAME_START_OR_BODY_CHAR);
137         Arrays.fill(CHARACTER_CODES, 0x0A59, 0x0A5C, NAME_START_OR_BODY_CHAR);
138         CHARACTER_CODES[0x0A5E] = NAME_START_OR_BODY_CHAR;
139         Arrays.fill(CHARACTER_CODES, 0x0A72, 0x0A74, NAME_START_OR_BODY_CHAR);
140         Arrays.fill(CHARACTER_CODES, 0x0A85, 0x0A8B, NAME_START_OR_BODY_CHAR);
141         CHARACTER_CODES[0x0A8D] = NAME_START_OR_BODY_CHAR;
142         Arrays.fill(CHARACTER_CODES, 0x0A8F, 0x0A91, NAME_START_OR_BODY_CHAR);
143         Arrays.fill(CHARACTER_CODES, 0x0A93, 0x0AA8, NAME_START_OR_BODY_CHAR);
144         Arrays.fill(CHARACTER_CODES, 0x0AAA, 0x0AB0, NAME_START_OR_BODY_CHAR);
145         Arrays.fill(CHARACTER_CODES, 0x0AB2, 0x0AB3, NAME_START_OR_BODY_CHAR);
146         Arrays.fill(CHARACTER_CODES, 0x0AB5, 0x0AB9, NAME_START_OR_BODY_CHAR);
147         CHARACTER_CODES[0x0ABD] = NAME_START_OR_BODY_CHAR;
148         CHARACTER_CODES[0x0AE0] = NAME_START_OR_BODY_CHAR;
149         Arrays.fill(CHARACTER_CODES, 0x0B05, 0x0B0C, NAME_START_OR_BODY_CHAR);
150         Arrays.fill(CHARACTER_CODES, 0x0B0F, 0x0B10, NAME_START_OR_BODY_CHAR);
151         Arrays.fill(CHARACTER_CODES, 0x0B13, 0x0B28, NAME_START_OR_BODY_CHAR);
152         Arrays.fill(CHARACTER_CODES, 0x0B2A, 0x0B30, NAME_START_OR_BODY_CHAR);
153         Arrays.fill(CHARACTER_CODES, 0x0B32, 0x0B33, NAME_START_OR_BODY_CHAR);
154         Arrays.fill(CHARACTER_CODES, 0x0B36, 0x0B39, NAME_START_OR_BODY_CHAR);
155         CHARACTER_CODES[0x0B3D] = NAME_START_OR_BODY_CHAR;
156         Arrays.fill(CHARACTER_CODES, 0x0B5C, 0x0B5D, NAME_START_OR_BODY_CHAR);
157         Arrays.fill(CHARACTER_CODES, 0x0B5F, 0x0B61, NAME_START_OR_BODY_CHAR);
158         Arrays.fill(CHARACTER_CODES, 0x0B85, 0x0B8A, NAME_START_OR_BODY_CHAR);
159         Arrays.fill(CHARACTER_CODES, 0x0B8E, 0x0B90, NAME_START_OR_BODY_CHAR);
160         Arrays.fill(CHARACTER_CODES, 0x0B92, 0x0B95, NAME_START_OR_BODY_CHAR);
161         Arrays.fill(CHARACTER_CODES, 0x0B99, 0x0B9A, NAME_START_OR_BODY_CHAR);
162         CHARACTER_CODES[0x0B9C] = NAME_START_OR_BODY_CHAR;
163         Arrays.fill(CHARACTER_CODES, 0x0B9E, 0x0B9F, NAME_START_OR_BODY_CHAR);
164         Arrays.fill(CHARACTER_CODES, 0x0BA3, 0x0BA4, NAME_START_OR_BODY_CHAR);
165         Arrays.fill(CHARACTER_CODES, 0x0BA8, 0x0BAA, NAME_START_OR_BODY_CHAR);
166         Arrays.fill(CHARACTER_CODES, 0x0BAE, 0x0BB5, NAME_START_OR_BODY_CHAR);
167         Arrays.fill(CHARACTER_CODES, 0x0BB7, 0x0BB9, NAME_START_OR_BODY_CHAR);
168         Arrays.fill(CHARACTER_CODES, 0x0C05, 0x0C0C, NAME_START_OR_BODY_CHAR);
169         Arrays.fill(CHARACTER_CODES, 0x0C0E, 0x0C10, NAME_START_OR_BODY_CHAR);
170         Arrays.fill(CHARACTER_CODES, 0x0C12, 0x0C28, NAME_START_OR_BODY_CHAR);
171         Arrays.fill(CHARACTER_CODES, 0x0C2A, 0x0C33, NAME_START_OR_BODY_CHAR);
172         Arrays.fill(CHARACTER_CODES, 0x0C35, 0x0C39, NAME_START_OR_BODY_CHAR);
173         Arrays.fill(CHARACTER_CODES, 0x0C60, 0x0C61, NAME_START_OR_BODY_CHAR);
174         Arrays.fill(CHARACTER_CODES, 0x0C85, 0x0C8C, NAME_START_OR_BODY_CHAR);
175         Arrays.fill(CHARACTER_CODES, 0x0C8E, 0x0C90, NAME_START_OR_BODY_CHAR);
176         Arrays.fill(CHARACTER_CODES, 0x0C92, 0x0CA8, NAME_START_OR_BODY_CHAR);
177         Arrays.fill(CHARACTER_CODES, 0x0CAA, 0x0CB3, NAME_START_OR_BODY_CHAR);
178         Arrays.fill(CHARACTER_CODES, 0x0CB5, 0x0CB9, NAME_START_OR_BODY_CHAR);
179         CHARACTER_CODES[0x0CDE] = NAME_START_OR_BODY_CHAR;
180         Arrays.fill(CHARACTER_CODES, 0x0CE0, 0x0CE1, NAME_START_OR_BODY_CHAR);
181         Arrays.fill(CHARACTER_CODES, 0x0D05, 0x0D0C, NAME_START_OR_BODY_CHAR);
182         Arrays.fill(CHARACTER_CODES, 0x0D0E, 0x0D10, NAME_START_OR_BODY_CHAR);
183         Arrays.fill(CHARACTER_CODES, 0x0D12, 0x0D28, NAME_START_OR_BODY_CHAR);
184         Arrays.fill(CHARACTER_CODES, 0x0D2A, 0x0D39, NAME_START_OR_BODY_CHAR);
185         Arrays.fill(CHARACTER_CODES, 0x0D60, 0x0D61, NAME_START_OR_BODY_CHAR);
186         Arrays.fill(CHARACTER_CODES, 0x0E01, 0x0E2E, NAME_START_OR_BODY_CHAR);
187         CHARACTER_CODES[0x0E30] = NAME_START_OR_BODY_CHAR;
188         Arrays.fill(CHARACTER_CODES, 0x0E32, 0x0E33, NAME_START_OR_BODY_CHAR);
189         Arrays.fill(CHARACTER_CODES, 0x0E40, 0x0E45, NAME_START_OR_BODY_CHAR);
190         Arrays.fill(CHARACTER_CODES, 0x0E81, 0x0E82, NAME_START_OR_BODY_CHAR);
191         CHARACTER_CODES[0x0E84] = NAME_START_OR_BODY_CHAR;
192         Arrays.fill(CHARACTER_CODES, 0x0E87, 0x0E88, NAME_START_OR_BODY_CHAR);
193         CHARACTER_CODES[0x0E8A] = NAME_START_OR_BODY_CHAR;
194         CHARACTER_CODES[0x0E8D] = NAME_START_OR_BODY_CHAR;
195         Arrays.fill(CHARACTER_CODES, 0x0E94, 0x0E97, NAME_START_OR_BODY_CHAR);
196         Arrays.fill(CHARACTER_CODES, 0x0E99, 0x0E9F, NAME_START_OR_BODY_CHAR);
197         Arrays.fill(CHARACTER_CODES, 0x0EA1, 0x0EA3, NAME_START_OR_BODY_CHAR);
198         CHARACTER_CODES[0x0EA5] = NAME_START_OR_BODY_CHAR;
199         CHARACTER_CODES[0x0EA7] = NAME_START_OR_BODY_CHAR;
200         Arrays.fill(CHARACTER_CODES, 0x0EAA, 0x0EAB, NAME_START_OR_BODY_CHAR);
201         Arrays.fill(CHARACTER_CODES, 0x0EAD, 0x0EAE, NAME_START_OR_BODY_CHAR);
202         CHARACTER_CODES[0x0EB0] = NAME_START_OR_BODY_CHAR;
203         Arrays.fill(CHARACTER_CODES, 0x0EB2, 0x0EB3, NAME_START_OR_BODY_CHAR);
204         CHARACTER_CODES[0x0EBD] = NAME_START_OR_BODY_CHAR;
205         Arrays.fill(CHARACTER_CODES, 0x0EC0, 0x0EC4, NAME_START_OR_BODY_CHAR);
206         Arrays.fill(CHARACTER_CODES, 0x0F40, 0x0F47, NAME_START_OR_BODY_CHAR);
207         Arrays.fill(CHARACTER_CODES, 0x0F49, 0x0F69, NAME_START_OR_BODY_CHAR);
208         Arrays.fill(CHARACTER_CODES, 0x10A0, 0x10C5, NAME_START_OR_BODY_CHAR);
209         Arrays.fill(CHARACTER_CODES, 0x10D0, 0x10F6, NAME_START_OR_BODY_CHAR);
210         CHARACTER_CODES[0x1100] = NAME_START_OR_BODY_CHAR;
211         Arrays.fill(CHARACTER_CODES, 0x1102, 0x1103, NAME_START_OR_BODY_CHAR);
212         Arrays.fill(CHARACTER_CODES, 0x1105, 0x1107, NAME_START_OR_BODY_CHAR);
213         CHARACTER_CODES[0x1109] = NAME_START_OR_BODY_CHAR;
214         Arrays.fill(CHARACTER_CODES, 0x110B, 0x110C, NAME_START_OR_BODY_CHAR);
215         Arrays.fill(CHARACTER_CODES, 0x110E, 0x1112, NAME_START_OR_BODY_CHAR);
216         CHARACTER_CODES[0x113C] = NAME_START_OR_BODY_CHAR;
217         CHARACTER_CODES[0x113E] = NAME_START_OR_BODY_CHAR;
218         CHARACTER_CODES[0x1140] = NAME_START_OR_BODY_CHAR;
219         CHARACTER_CODES[0x114C] = NAME_START_OR_BODY_CHAR;
220         CHARACTER_CODES[0x114E] = NAME_START_OR_BODY_CHAR;
221         CHARACTER_CODES[0x1150] = NAME_START_OR_BODY_CHAR;
222         Arrays.fill(CHARACTER_CODES, 0x1154, 0x1155, NAME_START_OR_BODY_CHAR);
223         CHARACTER_CODES[0x1159] = NAME_START_OR_BODY_CHAR;
224         Arrays.fill(CHARACTER_CODES, 0x115F, 0x1161, NAME_START_OR_BODY_CHAR);
225         CHARACTER_CODES[0x1163] = NAME_START_OR_BODY_CHAR;
226         CHARACTER_CODES[0x1165] = NAME_START_OR_BODY_CHAR;
227         CHARACTER_CODES[0x1167] = NAME_START_OR_BODY_CHAR;
228         CHARACTER_CODES[0x1169] = NAME_START_OR_BODY_CHAR;
229         Arrays.fill(CHARACTER_CODES, 0x116D, 0x116E, NAME_START_OR_BODY_CHAR);
230         Arrays.fill(CHARACTER_CODES, 0x1172, 0x1173, NAME_START_OR_BODY_CHAR);
231         CHARACTER_CODES[0x1175] = NAME_START_OR_BODY_CHAR;
232         CHARACTER_CODES[0x119E] = NAME_START_OR_BODY_CHAR;
233         CHARACTER_CODES[0x11A8] = NAME_START_OR_BODY_CHAR;
234         CHARACTER_CODES[0x11AB] = NAME_START_OR_BODY_CHAR;
235         Arrays.fill(CHARACTER_CODES, 0x11AE, 0x11AF, NAME_START_OR_BODY_CHAR);
236         Arrays.fill(CHARACTER_CODES, 0x11B7, 0x11B8, NAME_START_OR_BODY_CHAR);
237         CHARACTER_CODES[0x11BA] = NAME_START_OR_BODY_CHAR;
238         Arrays.fill(CHARACTER_CODES, 0x11BC, 0x11C2, NAME_START_OR_BODY_CHAR);
239         CHARACTER_CODES[0x11EB] = NAME_START_OR_BODY_CHAR;
240         CHARACTER_CODES[0x11F0] = NAME_START_OR_BODY_CHAR;
241         CHARACTER_CODES[0x11F9] = NAME_START_OR_BODY_CHAR;
242         Arrays.fill(CHARACTER_CODES, 0x1E00, 0x1E9B, NAME_START_OR_BODY_CHAR);
243         Arrays.fill(CHARACTER_CODES, 0x1EA0, 0x1EF9, NAME_START_OR_BODY_CHAR);
244         Arrays.fill(CHARACTER_CODES, 0x1F00, 0x1F15, NAME_START_OR_BODY_CHAR);
245         Arrays.fill(CHARACTER_CODES, 0x1F18, 0x1F1D, NAME_START_OR_BODY_CHAR);
246         Arrays.fill(CHARACTER_CODES, 0x1F20, 0x1F45, NAME_START_OR_BODY_CHAR);
247         Arrays.fill(CHARACTER_CODES, 0x1F48, 0x1F4D, NAME_START_OR_BODY_CHAR);
248         Arrays.fill(CHARACTER_CODES, 0x1F50, 0x1F57, NAME_START_OR_BODY_CHAR);
249         CHARACTER_CODES[0x1F59] = NAME_START_OR_BODY_CHAR;
250         CHARACTER_CODES[0x1F5B] = NAME_START_OR_BODY_CHAR;
251         CHARACTER_CODES[0x1F5D] = NAME_START_OR_BODY_CHAR;
252         Arrays.fill(CHARACTER_CODES, 0x1F5F, 0x1F7D, NAME_START_OR_BODY_CHAR);
253         Arrays.fill(CHARACTER_CODES, 0x1F80, 0x1FB4, NAME_START_OR_BODY_CHAR);
254         Arrays.fill(CHARACTER_CODES, 0x1FB6, 0x1FBC, NAME_START_OR_BODY_CHAR);
255         CHARACTER_CODES[0x1FBE] = NAME_START_OR_BODY_CHAR;
256         Arrays.fill(CHARACTER_CODES, 0x1FC2, 0x1FC4, NAME_START_OR_BODY_CHAR);
257         Arrays.fill(CHARACTER_CODES, 0x1FC6, 0x1FCC, NAME_START_OR_BODY_CHAR);
258         Arrays.fill(CHARACTER_CODES, 0x1FD0, 0x1FD3, NAME_START_OR_BODY_CHAR);
259         Arrays.fill(CHARACTER_CODES, 0x1FD6, 0x1FDB, NAME_START_OR_BODY_CHAR);
260         Arrays.fill(CHARACTER_CODES, 0x1FE0, 0x1FEC, NAME_START_OR_BODY_CHAR);
261         Arrays.fill(CHARACTER_CODES, 0x1FF2, 0x1FF4, NAME_START_OR_BODY_CHAR);
262         Arrays.fill(CHARACTER_CODES, 0x1FF6, 0x1FFC, NAME_START_OR_BODY_CHAR);
263         CHARACTER_CODES[0x2126] = NAME_START_OR_BODY_CHAR;
264         Arrays.fill(CHARACTER_CODES, 0x212A, 0x212B, NAME_START_OR_BODY_CHAR);
265         CHARACTER_CODES[0x212E] = NAME_START_OR_BODY_CHAR;
266         Arrays.fill(CHARACTER_CODES, 0x2180, 0x2182, NAME_START_OR_BODY_CHAR);
267         Arrays.fill(CHARACTER_CODES, 0x3041, 0x3094, NAME_START_OR_BODY_CHAR);
268         Arrays.fill(CHARACTER_CODES, 0x30A1, 0x30FA, NAME_START_OR_BODY_CHAR);
269         Arrays.fill(CHARACTER_CODES, 0x3105, 0x312C, NAME_START_OR_BODY_CHAR);
270         Arrays.fill(CHARACTER_CODES, 0xAC00, 0xD7A3, NAME_START_OR_BODY_CHAR);
271         // Ideographic
272         Arrays.fill(CHARACTER_CODES, 0x4E00, 0x9FA5, NAME_START_OR_BODY_CHAR);
273         CHARACTER_CODES[0x3007] = NAME_START_OR_BODY_CHAR;
274         Arrays.fill(CHARACTER_CODES, 0x3021, 0x3029, NAME_START_OR_BODY_CHAR);
275         // NameChar ::= Letter | Digit | '.' | '-' | '_' | ':' | CombiningChar |
276         // Extender
277         CHARACTER_CODES['.'] = NAME_BODY_CHAR;
278         CHARACTER_CODES['-'] = NAME_BODY_CHAR;
279         // CombiningChar
280         Arrays.fill(CHARACTER_CODES, 0x0300, 0x0345, NAME_BODY_CHAR);
281         Arrays.fill(CHARACTER_CODES, 0x0360, 0x0361, NAME_BODY_CHAR);
282         Arrays.fill(CHARACTER_CODES, 0x0483, 0x0486, NAME_BODY_CHAR);
283         Arrays.fill(CHARACTER_CODES, 0x0591, 0x05A1, NAME_BODY_CHAR);
284         Arrays.fill(CHARACTER_CODES, 0x05A3, 0x05B9, NAME_BODY_CHAR);
285         Arrays.fill(CHARACTER_CODES, 0x05BB, 0x05BD, NAME_BODY_CHAR);
286         CHARACTER_CODES[0x05BF] = NAME_BODY_CHAR;
287         Arrays.fill(CHARACTER_CODES, 0x05C1, 0x05C2, NAME_BODY_CHAR);
288         CHARACTER_CODES[0x05C4] = NAME_BODY_CHAR;
289         Arrays.fill(CHARACTER_CODES, 0x064B, 0x0652, NAME_BODY_CHAR);
290         CHARACTER_CODES[0x0670] = NAME_BODY_CHAR;
291         Arrays.fill(CHARACTER_CODES, 0x06D6, 0x06DC, NAME_BODY_CHAR);
292         Arrays.fill(CHARACTER_CODES, 0x06DD, 0x06DF, NAME_BODY_CHAR);
293         Arrays.fill(CHARACTER_CODES, 0x06E0, 0x06E4, NAME_BODY_CHAR);
294         Arrays.fill(CHARACTER_CODES, 0x06E7, 0x06E8, NAME_BODY_CHAR);
295         Arrays.fill(CHARACTER_CODES, 0x06EA, 0x06ED, NAME_BODY_CHAR);
296         Arrays.fill(CHARACTER_CODES, 0x0901, 0x0903, NAME_BODY_CHAR);
297         CHARACTER_CODES[0x093C] = NAME_BODY_CHAR;
298         Arrays.fill(CHARACTER_CODES, 0x093E, 0x094C, NAME_BODY_CHAR);
299         CHARACTER_CODES[0x094D] = NAME_BODY_CHAR;
300         Arrays.fill(CHARACTER_CODES, 0x0951, 0x0954, NAME_BODY_CHAR);
301         Arrays.fill(CHARACTER_CODES, 0x0962, 0x0963, NAME_BODY_CHAR);
302         Arrays.fill(CHARACTER_CODES, 0x0981, 0x0983, NAME_BODY_CHAR);
303         CHARACTER_CODES[0x09BC] = NAME_BODY_CHAR;
304         CHARACTER_CODES[0x09BE] = NAME_BODY_CHAR;
305         CHARACTER_CODES[0x09BF] = NAME_BODY_CHAR;
306         Arrays.fill(CHARACTER_CODES, 0x09C0, 0x09C4, NAME_BODY_CHAR);
307         Arrays.fill(CHARACTER_CODES, 0x09C7, 0x09C8, NAME_BODY_CHAR);
308         Arrays.fill(CHARACTER_CODES, 0x09CB, 0x09CD, NAME_BODY_CHAR);
309         CHARACTER_CODES[0x09D7] = NAME_BODY_CHAR;
310         Arrays.fill(CHARACTER_CODES, 0x09E2, 0x09E3, NAME_BODY_CHAR);
311         CHARACTER_CODES[0x0A02] = NAME_BODY_CHAR;
312         CHARACTER_CODES[0x0A3C] = NAME_BODY_CHAR;
313         CHARACTER_CODES[0x0A3E] = NAME_BODY_CHAR;
314         CHARACTER_CODES[0x0A3F] = NAME_BODY_CHAR;
315         Arrays.fill(CHARACTER_CODES, 0x0A40, 0x0A42, NAME_BODY_CHAR);
316         Arrays.fill(CHARACTER_CODES, 0x0A47, 0x0A48, NAME_BODY_CHAR);
317         Arrays.fill(CHARACTER_CODES, 0x0A4B, 0x0A4D, NAME_BODY_CHAR);
318         Arrays.fill(CHARACTER_CODES, 0x0A70, 0x0A71, NAME_BODY_CHAR);
319         Arrays.fill(CHARACTER_CODES, 0x0A81, 0x0A83, NAME_BODY_CHAR);
320         CHARACTER_CODES[0x0ABC] = NAME_BODY_CHAR;
321         Arrays.fill(CHARACTER_CODES, 0x0ABE, 0x0AC5, NAME_BODY_CHAR);
322         Arrays.fill(CHARACTER_CODES, 0x0AC7, 0x0AC9, NAME_BODY_CHAR);
323         Arrays.fill(CHARACTER_CODES, 0x0ACB, 0x0ACD, NAME_BODY_CHAR);
324         Arrays.fill(CHARACTER_CODES, 0x0B01, 0x0B03, NAME_BODY_CHAR);
325         CHARACTER_CODES[0x0B3C] = NAME_BODY_CHAR;
326         Arrays.fill(CHARACTER_CODES, 0x0B3E, 0x0B43, NAME_BODY_CHAR);
327         Arrays.fill(CHARACTER_CODES, 0x0B47, 0x0B48, NAME_BODY_CHAR);
328         Arrays.fill(CHARACTER_CODES, 0x0B4B, 0x0B4D, NAME_BODY_CHAR);
329         Arrays.fill(CHARACTER_CODES, 0x0B56, 0x0B57, NAME_BODY_CHAR);
330         Arrays.fill(CHARACTER_CODES, 0x0B82, 0x0B83, NAME_BODY_CHAR);
331         Arrays.fill(CHARACTER_CODES, 0x0BBE, 0x0BC2, NAME_BODY_CHAR);
332         Arrays.fill(CHARACTER_CODES, 0x0BC6, 0x0BC8, NAME_BODY_CHAR);
333         Arrays.fill(CHARACTER_CODES, 0x0BCA, 0x0BCD, NAME_BODY_CHAR);
334         CHARACTER_CODES[0x0BD7] = NAME_BODY_CHAR;
335         Arrays.fill(CHARACTER_CODES, 0x0C01, 0x0C03, NAME_BODY_CHAR);
336         Arrays.fill(CHARACTER_CODES, 0x0C3E, 0x0C44, NAME_BODY_CHAR);
337         Arrays.fill(CHARACTER_CODES, 0x0C46, 0x0C48, NAME_BODY_CHAR);
338         Arrays.fill(CHARACTER_CODES, 0x0C4A, 0x0C4D, NAME_BODY_CHAR);
339         Arrays.fill(CHARACTER_CODES, 0x0C55, 0x0C56, NAME_BODY_CHAR);
340         Arrays.fill(CHARACTER_CODES, 0x0C82, 0x0C83, NAME_BODY_CHAR);
341         Arrays.fill(CHARACTER_CODES, 0x0CBE, 0x0CC4, NAME_BODY_CHAR);
342         Arrays.fill(CHARACTER_CODES, 0x0CC6, 0x0CC8, NAME_BODY_CHAR);
343         Arrays.fill(CHARACTER_CODES, 0x0CCA, 0x0CCD, NAME_BODY_CHAR);
344         Arrays.fill(CHARACTER_CODES, 0x0CD5, 0x0CD6, NAME_BODY_CHAR);
345         Arrays.fill(CHARACTER_CODES, 0x0D02, 0x0D03, NAME_BODY_CHAR);
346         Arrays.fill(CHARACTER_CODES, 0x0D3E, 0x0D43, NAME_BODY_CHAR);
347         Arrays.fill(CHARACTER_CODES, 0x0D46, 0x0D48, NAME_BODY_CHAR);
348         Arrays.fill(CHARACTER_CODES, 0x0D4A, 0x0D4D, NAME_BODY_CHAR);
349         CHARACTER_CODES[0x0D57] = NAME_BODY_CHAR;
350         CHARACTER_CODES[0x0E31] = NAME_BODY_CHAR;
351         Arrays.fill(CHARACTER_CODES, 0x0E34, 0x0E3A, NAME_BODY_CHAR);
352         Arrays.fill(CHARACTER_CODES, 0x0E47, 0x0E4E, NAME_BODY_CHAR);
353         CHARACTER_CODES[0x0EB1] = NAME_BODY_CHAR;
354         Arrays.fill(CHARACTER_CODES, 0x0EB4, 0x0EB9, NAME_BODY_CHAR);
355         Arrays.fill(CHARACTER_CODES, 0x0EBB, 0x0EBC, NAME_BODY_CHAR);
356         Arrays.fill(CHARACTER_CODES, 0x0EC8, 0x0ECD, NAME_BODY_CHAR);
357         Arrays.fill(CHARACTER_CODES, 0x0F18, 0x0F19, NAME_BODY_CHAR);
358         CHARACTER_CODES[0x0F35] = NAME_BODY_CHAR;
359         CHARACTER_CODES[0x0F37] = NAME_BODY_CHAR;
360         CHARACTER_CODES[0x0F39] = NAME_BODY_CHAR;
361         CHARACTER_CODES[0x0F3E] = NAME_BODY_CHAR;
362         CHARACTER_CODES[0x0F3F] = NAME_BODY_CHAR;
363         Arrays.fill(CHARACTER_CODES, 0x0F71, 0x0F84, NAME_BODY_CHAR);
364         Arrays.fill(CHARACTER_CODES, 0x0F86, 0x0F8B, NAME_BODY_CHAR);
365         Arrays.fill(CHARACTER_CODES, 0x0F90, 0x0F95, NAME_BODY_CHAR);
366         CHARACTER_CODES[0x0F97] = NAME_BODY_CHAR;
367         Arrays.fill(CHARACTER_CODES, 0x0F99, 0x0FAD, NAME_BODY_CHAR);
368         Arrays.fill(CHARACTER_CODES, 0x0FB1, 0x0FB7, NAME_BODY_CHAR);
369         CHARACTER_CODES[0x0FB9] = NAME_BODY_CHAR;
370         Arrays.fill(CHARACTER_CODES, 0x20D0, 0x20DC, NAME_BODY_CHAR);
371         CHARACTER_CODES[0x20E1] = NAME_BODY_CHAR;
372         Arrays.fill(CHARACTER_CODES, 0x302A, 0x302F, NAME_BODY_CHAR);
373         CHARACTER_CODES[0x3099] = NAME_BODY_CHAR;
374         CHARACTER_CODES[0x309A] = NAME_BODY_CHAR;
375         // Digit
376         Arrays.fill(CHARACTER_CODES, 0x0030, 0x0039, NAME_BODY_CHAR);
377         Arrays.fill(CHARACTER_CODES, 0x0660, 0x0669, NAME_BODY_CHAR);
378         Arrays.fill(CHARACTER_CODES, 0x06F0, 0x06F9, NAME_BODY_CHAR);
379         Arrays.fill(CHARACTER_CODES, 0x0966, 0x096F, NAME_BODY_CHAR);
380         Arrays.fill(CHARACTER_CODES, 0x09E6, 0x09EF, NAME_BODY_CHAR);
381         Arrays.fill(CHARACTER_CODES, 0x0A66, 0x0A6F, NAME_BODY_CHAR);
382         Arrays.fill(CHARACTER_CODES, 0x0AE6, 0x0AEF, NAME_BODY_CHAR);
383         Arrays.fill(CHARACTER_CODES, 0x0B66, 0x0B6F, NAME_BODY_CHAR);
384         Arrays.fill(CHARACTER_CODES, 0x0BE7, 0x0BEF, NAME_BODY_CHAR);
385         Arrays.fill(CHARACTER_CODES, 0x0C66, 0x0C6F, NAME_BODY_CHAR);
386         Arrays.fill(CHARACTER_CODES, 0x0CE6, 0x0CEF, NAME_BODY_CHAR);
387         Arrays.fill(CHARACTER_CODES, 0x0D66, 0x0D6F, NAME_BODY_CHAR);
388         Arrays.fill(CHARACTER_CODES, 0x0E50, 0x0E59, NAME_BODY_CHAR);
389         Arrays.fill(CHARACTER_CODES, 0x0ED0, 0x0ED9, NAME_BODY_CHAR);
390         Arrays.fill(CHARACTER_CODES, 0x0F20, 0x0F29, NAME_BODY_CHAR);
391         // Extender
392         CHARACTER_CODES[0x00B7] = NAME_BODY_CHAR;
393         CHARACTER_CODES[0x02D0] = NAME_BODY_CHAR;
394         CHARACTER_CODES[0x02D1] = NAME_BODY_CHAR;
395         CHARACTER_CODES[0x0387] = NAME_BODY_CHAR;
396         CHARACTER_CODES[0x0640] = NAME_BODY_CHAR;
397         CHARACTER_CODES[0x0E46] = NAME_BODY_CHAR;
398         CHARACTER_CODES[0x0EC6] = NAME_BODY_CHAR;
399         CHARACTER_CODES[0x3005] = NAME_BODY_CHAR;
400         Arrays.fill(CHARACTER_CODES, 0x3031, 0x3035, NAME_BODY_CHAR);
401         Arrays.fill(CHARACTER_CODES, 0x309D, 0x309E, NAME_BODY_CHAR);
402         Arrays.fill(CHARACTER_CODES, 0x30FC, 0x30FE, NAME_BODY_CHAR);
403 
404     }
405 
406     private final Writer writer;
407     private final ArrayDeque<CharSequence> elementNames;
408     private final Set<CharSequence> currentAttributes = new HashSet<>();
409 
410     boolean elementsWritten = false;
411     boolean inElement = false;
412     boolean prologWritten = false;
413 
414     /**
415      * Constructs an XmlWriter with the specified writer for output.
416      * @param writer the writer to write to.
417      */
418     public XmlWriter(final Writer writer) {
419         this.writer = writer;
420         this.elementNames = new ArrayDeque<>();
421     }
422 
423     /**
424      * Starts a document by writing a prolog. Calling this method is optional. When
425      * writing a document fragment, it should <em>not</em> be called.
426      *
427      * @return this object
428      * @throws OperationNotAllowedException if called after the first element has
429      * been written or once a prolog has already been written
430      */
431     @Override
432     public IXmlWriter startDocument() throws IOException {
433         if (elementsWritten) {
434             throw new OperationNotAllowedException("Document already started");
435         }
436         if (prologWritten) {
437             throw new OperationNotAllowedException("Only one prolog allowed");
438         }
439         writer.write("<?xml version='1.0'?>");
440         prologWritten = true;
441         return this;
442     }
443 
444     /**
445      * Writes the start of an element.
446      *
447      * @param elementName the name of the element, not null
448      * @return this object
449      * @throws InvalidXmlException if the name is not valid for an xml element
450      * @throws OperationNotAllowedException if called after the first element has
451      * been closed
452      */
453     @Override
454     public IXmlWriter openElement(final CharSequence elementName) throws IOException {
455         if (elementsWritten && elementNames.isEmpty()) {
456             throw new OperationNotAllowedException("Root element already closed. Cannot open new element.");
457         }
458         if (isInvalidName(elementName)) {
459             throw new InvalidXmlException("'" + elementName + "' is not a valid element name");
460         }
461         elementsWritten = true;
462         if (inElement) {
463             writer.write('>');
464         }
465         writer.write('<');
466         rawWrite(elementName);
467         inElement = true;
468         elementNames.push(elementName);
469         currentAttributes.clear();
470         return this;
471     }
472 
473     /**
474      * Writes an attribute of an element. Note that this is only allowed directly
475      * after {@link #openElement(CharSequence)} or {@link #attribute}.
476      *
477      * @param name the attribute name, not null
478      * @param value the attribute value, not null
479      * @return this object
480      * @throws InvalidXmlException if the name is not valid for an xml attribute or
481      * if a value for the attribute has already been written
482      * @throws OperationNotAllowedException if called after
483      * {@link #content(CharSequence)} or {@link #closeElement()} or before any call
484      * to {@link #openElement(CharSequence)}
485      */
486     @Override
487     public IXmlWriter attribute(CharSequence name, CharSequence value) throws IOException {
488         if (elementNames.isEmpty()) {
489             if (elementsWritten) {
490                 throw new OperationNotAllowedException("Root element has already been closed.");
491             }
492             throw new OperationNotAllowedException("Close called before an element has been opened.");
493         }
494         if (isInvalidName(name)) {
495             throw new InvalidXmlException("'" + name + "' is not a valid attribute name.");
496         }
497         if (!inElement) {
498             throw new InvalidXmlException("Attributes can only be written in elements");
499         }
500         if (currentAttributes.contains(name)) {
501             throw new InvalidXmlException("Each attribute can only be written once");
502         }
503         writer.write(' ');
504         rawWrite(name);
505         writer.write('=');
506         writer.write('\'');
507         writeAttributeContent(value);
508         writer.write('\'');
509         currentAttributes.add(name);
510         return this;
511     }
512 
513     private void writeAttributeContent(CharSequence content) throws IOException {
514         writeEscaped(content, true);
515     }
516 
517     /**
518      * Writes content. Calling this method will automatically Note that this method
519      * does not use CDATA.
520      *
521      * @param content the content to write
522      * @return this object
523      * @throws OperationNotAllowedException if called before any call to
524      * {@link #openElement} or after the first element has been closed
525      */
526     @Override
527     public IXmlWriter content(CharSequence content) throws IOException {
528         if (elementNames.isEmpty()) {
529             if (elementsWritten) {
530                 throw new OperationNotAllowedException("Root element has already been closed.");
531             } 
532             throw new OperationNotAllowedException("An element must be opened before content can be written.");
533         }
534         if (inElement) {
535             writer.write('>');
536         }
537         writeBodyContent(content);
538         inElement = false;
539         return this;
540     }
541 
542     private void writeBodyContent(final CharSequence content) throws IOException {
543         writeEscaped(content, false);
544     }
545 
546     private void writeEscaped(final CharSequence content, boolean isAttributeContent) throws IOException {
547         final int length = content.length();
548         for (int i = 0; i < length; i++) {
549             char character = content.charAt(i);
550             if (character == '&') {
551                 writer.write("&amp;");
552             } else if (character == '<') {
553                 writer.write("&lt;");
554             } else if (character == '>') {
555                 writer.write("&gt;");
556             } else if (isAttributeContent && character == '\'') {
557                 writer.write("&apos;");
558             } else if (isAttributeContent && character == '\"') {
559                 writer.write("&quot;");
560             } else if (isOutOfRange(character)) {
561                 writer.write('?');
562             } else {
563                 writer.write(character);
564             }
565         }
566     }
567 
568     private boolean isOutOfRange(final char character) {
569         return !ALLOWED_CHARACTERS[character];
570     }
571 
572     /**
573      * Closes the last element written.
574      *
575      * @return this object
576      * @throws OperationNotAllowedException if called before any call to
577      * {@link #openElement} or after the first element has been closed
578      */
579     @Override
580     public IXmlWriter closeElement() throws IOException {
581         if (elementNames.isEmpty()) {
582             if (elementsWritten) {
583                 throw new OperationNotAllowedException("Root element has already been closed.");
584             } 
585             throw new OperationNotAllowedException("Close called before an element has been opened.");
586         }
587         final CharSequence elementName = elementNames.pop();
588         if (inElement) {
589             writer.write('/');
590             writer.write('>');
591         } else {
592             writer.write('<');
593             writer.write('/');
594             rawWrite(elementName);
595             writer.write('>');
596         }
597         writer.flush();
598         inElement = false;
599         return this;
600     }
601 
602     /**
603      * Closes all pending elements. When appropriate, resources are also flushed and
604      * closed. No exception is raised when called upon a document whose root element
605      * has already been closed.
606      *
607      * @return this object
608      * @throws OperationNotAllowedException if called before any call to
609      * {@link #openElement}
610      */
611     @Override
612     public IXmlWriter closeDocument() throws IOException {
613         if (elementNames.isEmpty() && !elementsWritten) {
614             throw new OperationNotAllowedException("Close called before an element has been opened.");
615         }
616         while (!elementNames.isEmpty()) {
617             closeElement();
618         }
619         writer.flush();
620         return this;
621     }
622 
623     private void rawWrite(final CharSequence sequence) throws IOException {
624         for (int i = 0; i < sequence.length(); i++) {
625             final char charAt = sequence.charAt(i);
626             writer.write(charAt);
627         }
628     }
629 
630     private boolean isInvalidName(final CharSequence sequence) {
631         boolean result = true;
632         final int length = sequence.length();
633         for (int i = 0; i < length; i++) {
634             char character = sequence.charAt(i);
635             if (i == 0) {
636                 if (!isValidNameStart(character)) {
637                     result = false;
638                     break;
639                 }
640             } else {
641                 if (!isValidNameBody(character)) {
642                     result = false;
643                     break;
644                 }
645             }
646         }
647         return !result;
648     }
649 
650     private boolean isValidNameStart(final char character) {
651         final byte code = CHARACTER_CODES[character];
652         return (code & NAME_START_MASK) > 0;
653     }
654 
655     private boolean isValidNameBody(final char character) {
656         final byte code = CHARACTER_CODES[character];
657         return (code & NAME_MASK) > 0;
658     }
659 
660     @Override
661     public void close() throws Exception {
662         closeDocument();
663     }
664 }