1 /**
2 * This file Copyright (c) 2009-2014 Magnolia International
3 * Ltd. (http://www.magnolia-cms.com). All rights reserved.
4 *
5 *
6 * This file is dual-licensed under both the Magnolia
7 * Network Agreement and the GNU General Public License.
8 * You may elect to use one or the other of these licenses.
9 *
10 * This file is distributed in the hope that it will be
11 * useful, but AS-IS and WITHOUT ANY WARRANTY; without even the
12 * implied warranty of MERCHANTABILITY or FITNESS FOR A
13 * PARTICULAR PURPOSE, TITLE, or NONINFRINGEMENT.
14 * Redistribution, except as permitted by whichever of the GPL
15 * or MNA you select, is prohibited.
16 *
17 * 1. For the GPL license (GPL), you can redistribute and/or
18 * modify this file under the terms of the GNU General
19 * Public License, Version 3, as published by the Free Software
20 * Foundation. You should have received a copy of the GNU
21 * General Public License, Version 3 along with this program;
22 * if not, write to the Free Software Foundation, Inc., 51
23 * Franklin St, Fifth Floor, Boston, MA 02110-1301 USA.
24 *
25 * 2. For the Magnolia Network Agreement (MNA), this file
26 * and the accompanying materials are made available under the
27 * terms of the MNA which accompanies this distribution, and
28 * is available at http://www.magnolia-cms.com/mna.html
29 *
30 * Any modifications to this file must keep this entire header
31 * intact.
32 *
33 */
34 package info.magnolia.cms.util;
35
36 import info.magnolia.cms.core.SystemProperty;
37 import info.magnolia.objectfactory.Components;
38
39 import java.lang.reflect.InvocationTargetException;
40 import java.lang.reflect.Method;
41 import javax.inject.Singleton;
42
43 /**
44 * A wrapper around java.text.Normalizer
45 *
46 * <strong>note:</strong> If needed, one can use their own implementation,
47 * by setting the info.magnolia.cms.util.UnicodeNormalizer$Normalizer system property.
48 *
49 * @see java.text.Normalizer
50 * @see <a href="http://en.wikipedia.org/wiki/Unicode_equivalence#Normal_forms">http://en.wikipedia.org/wiki/Unicode_equivalence#Normal_forms</a> for more information.
51 *
52 * @author gjoseph
53 * @version $Revision: $ ($Author: $)
54 */
55 public class UnicodeNormalizer {
56 private static final org.slf4j.Logger log = org.slf4j.LoggerFactory.getLogger(UnicodeNormalizer.class);
57
58 private static final String JAVA6_NORMALIZER_CLASS = "java.text.Normalizer";
59 private static final String JAVA6_FORMPARAM_CLASS = "java.text.Normalizer$Form";
60
61 private static final Normalizer normalizer = Components.getSingleton(Normalizer.class);
62
63 public static String[] normalizeNFC(String[] values) {
64 if (values == null) {
65 return null;
66 }
67 for (int i = 0; i < values.length; i++) {
68 values[i] = normalizeNFC(values[i]);
69 }
70 return values;
71 }
72
73 /**
74 * Normalizes the given String to the NFC form.
75 */
76 public static String normalizeNFC(String in) {
77 if (in == null) {
78 return null;
79 }
80 return normalizer.normalizeNFC(in);
81 /* if you're in dire need to debug:
82 try {
83 log.debug("not normalized: " + Arrays.toString(in.getBytes("UTF-8")) + " (" + in + ")");
84 String out = normalizer.normalizeNFC(in);
85 log.debug(" normalized: " + Arrays.toString(out.getBytes("UTF-8")) + " (" + out + ")");
86 return out;
87 } catch (UnsupportedEncodingException e) {
88 // do nothing
89 }
90 return in;
91 */
92 }
93
94 /**
95 * Used to normalize a String.
96 */
97 public interface Normalizer {
98 String normalizeNFC(String in);
99 }
100
101 /**
102 * Java 6 Normalizer wrapper.
103 */
104 @Singleton
105 public static final class Java6Normalizer implements Normalizer {
106 private final Method normalize;
107 private final Object nfc;
108
109 public Java6Normalizer() {
110 try {
111 final Class<?> normalizer = Class.forName(JAVA6_NORMALIZER_CLASS);
112 final Class<?> form = Class.forName(JAVA6_FORMPARAM_CLASS);
113 normalize = normalizer.getMethod("normalize", CharSequence.class, form);
114 nfc = form.getField("NFC").get(null);
115 } catch (ClassNotFoundException e) {
116 throw new RuntimeException(e);
117 } catch (IllegalAccessException e) {
118 throw new RuntimeException(e);
119 } catch (NoSuchFieldException e) {
120 throw new RuntimeException(e);
121 } catch (NoSuchMethodException e) {
122 throw new RuntimeException(e);
123 }
124
125 }
126
127 @Override
128 public String normalizeNFC(String in) {
129 try {
130 return (String) normalize.invoke(null, in, nfc);
131 } catch (IllegalAccessException e) {
132 throw new RuntimeException(e);
133 } catch (InvocationTargetException e) {
134 throw new RuntimeException(e);
135 }
136 }
137 }
138
139 /**
140 * Returns the original value unchanged.
141 */
142 @Singleton
143 public static final class NonNormalizer implements UnicodeNormalizer.Normalizer {
144 @Override
145 public String normalizeNFC(String in) {
146 return in;
147 }
148 }
149
150 /**
151 * Tries to load the normalizer dynamically and respects the property {@link SystemProperty#MAGNOLIA_UTF8_ENABLED}.
152 */
153 @Singleton
154 public static final class AutoDetectNormalizer implements Normalizer {
155 private final Normalizer delegate;
156
157 public AutoDetectNormalizer() {
158 Normalizer candidate;
159 if (SystemProperty.getBooleanProperty(SystemProperty.MAGNOLIA_UTF8_ENABLED)) {
160 candidate = new Java6Normalizer();
161
162 } else {
163 candidate = new NonNormalizer();
164 }
165 this.delegate = candidate;
166 }
167
168 @Override
169 public String normalizeNFC(String in) {
170 return delegate.normalizeNFC(in);
171 }
172 }
173
174 }