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