1 /**
2 * This file Copyright (c) 2009-2018 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.init.MagnoliaConfigurationProperties;
38 import info.magnolia.objectfactory.Components;
39
40 import java.lang.reflect.InvocationTargetException;
41 import java.lang.reflect.Method;
42
43 import javax.inject.Singleton;
44
45 /**
46 * A wrapper around java.text.Normalizer
47 *
48 * <strong>note:</strong> If needed, one can use their own implementation,
49 * by setting the info.magnolia.cms.util.UnicodeNormalizer$Normalizer system property.
50 *
51 * @see java.text.Normalizer
52 * @see <a href="http://en.wikipedia.org/wiki/Unicode_equivalence#Normal_forms">http://en.wikipedia.org/wiki/Unicode_equivalence#Normal_forms</a> for more information.
53 */
54 public class UnicodeNormalizer {
55 private static final org.slf4j.Logger log = org.slf4j.LoggerFactory.getLogger(UnicodeNormalizer.class);
56
57 private static final String JAVA6_NORMALIZER_CLASS = "java.text.Normalizer";
58 private static final String JAVA6_FORMPARAM_CLASS = "java.text.Normalizer$Form";
59
60 private static final Normalizer normalizer = Components.getSingleton(Normalizer.class);
61
62 public static String[] normalizeNFC(String[] values) {
63 if (values == null) {
64 return null;
65 }
66 for (int i = 0; i < values.length; i++) {
67 values[i] = normalizeNFC(values[i]);
68 }
69 return values;
70 }
71
72 /**
73 * Normalizes the given String to the NFC form.
74 */
75 public static String normalizeNFC(String in) {
76 if (in == null) {
77 return null;
78 }
79 return normalizer.normalizeNFC(in);
80 /* if you're in dire need to debug:
81 try {
82 log.debug("not normalized: " + Arrays.toString(in.getBytes("UTF-8")) + " (" + in + ")");
83 String out = normalizer.normalizeNFC(in);
84 log.debug(" normalized: " + Arrays.toString(out.getBytes("UTF-8")) + " (" + out + ")");
85 return out;
86 } catch (UnsupportedEncodingException e) {
87 // do nothing
88 }
89 return in;
90 */
91 }
92
93 /**
94 * Used to normalize a String.
95 */
96 public interface Normalizer {
97 String normalizeNFC(String in);
98 }
99
100 /**
101 * Java 6 Normalizer wrapper.
102 */
103 @Singleton
104 public static final class Java6Normalizer implements Normalizer {
105 private final Method normalize;
106 private final Object nfc;
107
108 public Java6Normalizer() {
109 try {
110 final Class<?> normalizer = Class.forName(JAVA6_NORMALIZER_CLASS);
111 final Class<?> form = Class.forName(JAVA6_FORMPARAM_CLASS);
112 normalize = normalizer.getMethod("normalize", CharSequence.class, form);
113 nfc = form.getField("NFC").get(null);
114 } catch (ClassNotFoundException e) {
115 throw new RuntimeException(e);
116 } catch (IllegalAccessException e) {
117 throw new RuntimeException(e);
118 } catch (NoSuchFieldException e) {
119 throw new RuntimeException(e);
120 } catch (NoSuchMethodException e) {
121 throw new RuntimeException(e);
122 }
123
124 }
125
126 @Override
127 public String normalizeNFC(String in) {
128 try {
129 return (String) normalize.invoke(null, in, nfc);
130 } catch (IllegalAccessException e) {
131 throw new RuntimeException(e);
132 } catch (InvocationTargetException e) {
133 throw new RuntimeException(e);
134 }
135 }
136 }
137
138 /**
139 * Returns the original value unchanged.
140 */
141 @Singleton
142 public static final class NonNormalizer implements UnicodeNormalizer.Normalizer {
143 @Override
144 public String normalizeNFC(String in) {
145 return in;
146 }
147 }
148
149 /**
150 * Tries to load the normalizer dynamically and respects the property {@link info.magnolia.init.MagnoliaConfigurationProperties#MAGNOLIA_UTF8_ENABLED}.
151 */
152 @Singleton
153 public static final class AutoDetectNormalizer implements Normalizer {
154 private final Normalizer delegate;
155
156 public AutoDetectNormalizer() {
157 Normalizer candidate;
158 if (SystemProperty.getBooleanProperty(MagnoliaConfigurationProperties.MAGNOLIA_UTF8_ENABLED)) {
159 candidate = new Java6Normalizer();
160
161 } else {
162 candidate = new NonNormalizer();
163 }
164 this.delegate = candidate;
165 }
166
167 @Override
168 public String normalizeNFC(String in) {
169 return delegate.normalizeNFC(in);
170 }
171 }
172
173 }