1 /**
2 * This file Copyright (c) 2009-2015 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
42 import javax.inject.Singleton;
43
44 /**
45 * A wrapper around java.text.Normalizer
46 *
47 * <strong>note:</strong> If needed, one can use their own implementation,
48 * by setting the info.magnolia.cms.util.UnicodeNormalizer$Normalizer system property.
49 *
50 * @see java.text.Normalizer
51 * @see <a href="http://en.wikipedia.org/wiki/Unicode_equivalence#Normal_forms">http://en.wikipedia.org/wiki/Unicode_equivalence#Normal_forms</a> for more information.
52 */
53 public class UnicodeNormalizer {
54 private static final org.slf4j.Logger log = org.slf4j.LoggerFactory.getLogger(UnicodeNormalizer.class);
55
56 private static final String JAVA6_NORMALIZER_CLASS = "java.text.Normalizer";
57 private static final String JAVA6_FORMPARAM_CLASS = "java.text.Normalizer$Form";
58
59 private static final Normalizer normalizer = Components.getSingleton(Normalizer.class);
60
61 public static String[] normalizeNFC(String[] values) {
62 if (values == null) {
63 return null;
64 }
65 for (int i = 0; i < values.length; i++) {
66 values[i] = normalizeNFC(values[i]);
67 }
68 return values;
69 }
70
71 /**
72 * Normalizes the given String to the NFC form.
73 */
74 public static String normalizeNFC(String in) {
75 if (in == null) {
76 return null;
77 }
78 return normalizer.normalizeNFC(in);
79 /* if you're in dire need to debug:
80 try {
81 log.debug("not normalized: " + Arrays.toString(in.getBytes("UTF-8")) + " (" + in + ")");
82 String out = normalizer.normalizeNFC(in);
83 log.debug(" normalized: " + Arrays.toString(out.getBytes("UTF-8")) + " (" + out + ")");
84 return out;
85 } catch (UnsupportedEncodingException e) {
86 // do nothing
87 }
88 return in;
89 */
90 }
91
92 /**
93 * Used to normalize a String.
94 */
95 public interface Normalizer {
96 String normalizeNFC(String in);
97 }
98
99 /**
100 * Java 6 Normalizer wrapper.
101 */
102 @Singleton
103 public static final class Java6Normalizer implements Normalizer {
104 private final Method normalize;
105 private final Object nfc;
106
107 public Java6Normalizer() {
108 try {
109 final Class<?> normalizer = Class.forName(JAVA6_NORMALIZER_CLASS);
110 final Class<?> form = Class.forName(JAVA6_FORMPARAM_CLASS);
111 normalize = normalizer.getMethod("normalize", CharSequence.class, form);
112 nfc = form.getField("NFC").get(null);
113 } catch (ClassNotFoundException e) {
114 throw new RuntimeException(e);
115 } catch (IllegalAccessException e) {
116 throw new RuntimeException(e);
117 } catch (NoSuchFieldException e) {
118 throw new RuntimeException(e);
119 } catch (NoSuchMethodException e) {
120 throw new RuntimeException(e);
121 }
122
123 }
124
125 @Override
126 public String normalizeNFC(String in) {
127 try {
128 return (String) normalize.invoke(null, in, nfc);
129 } catch (IllegalAccessException e) {
130 throw new RuntimeException(e);
131 } catch (InvocationTargetException e) {
132 throw new RuntimeException(e);
133 }
134 }
135 }
136
137 /**
138 * Returns the original value unchanged.
139 */
140 @Singleton
141 public static final class NonNormalizer implements UnicodeNormalizer.Normalizer {
142 @Override
143 public String normalizeNFC(String in) {
144 return in;
145 }
146 }
147
148 /**
149 * Tries to load the normalizer dynamically and respects the property {@link SystemProperty#MAGNOLIA_UTF8_ENABLED}.
150 */
151 @Singleton
152 public static final class AutoDetectNormalizer implements Normalizer {
153 private final Normalizer delegate;
154
155 public AutoDetectNormalizer() {
156 Normalizer candidate;
157 if (SystemProperty.getBooleanProperty(SystemProperty.MAGNOLIA_UTF8_ENABLED)) {
158 candidate = new Java6Normalizer();
159
160 } else {
161 candidate = new NonNormalizer();
162 }
163 this.delegate = candidate;
164 }
165
166 @Override
167 public String normalizeNFC(String in) {
168 return delegate.normalizeNFC(in);
169 }
170 }
171
172 }