1 /*
2 * reserved comment block
3 * DO NOT REMOVE OR ALTER!
4 */
5 /*
6 * Copyright 2001-2005 The Apache Software Foundation.
7 *
8 * Licensed under the Apache License, Version 2.0 (the "License");
9 * you may not use this file except in compliance with the License.
10 * You may obtain a copy of the License at
11 *
12 * http://www.apache.org/licenses/LICENSE-2.0
13 *
14 * Unless required by applicable law or agreed to in writing, software
15 * distributed under the License is distributed on an "AS IS" BASIS,
16 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
17 * See the License for the specific language governing permissions and
18 * limitations under the License.
19 */
20
21 package com.sun.org.apache.xerces.internal.impl.dv.xs;
22
23 import com.sun.org.apache.xerces.internal.impl.dv.InvalidDatatypeValueException;
24 import com.sun.org.apache.xerces.internal.util.URI;
25 import com.sun.org.apache.xerces.internal.impl.dv.ValidationContext;
26
27 /**
28 * Represent the schema type "anyURI"
29 *
30 * @xerces.internal
31 *
32 * @author Neeraj Bajaj, Sun Microsystems, inc.
33 * @author Sandy Gao, IBM
34 *
35 */
36 public class AnyURIDV extends TypeValidator {
37
38 private static final URI BASE_URI;
39 static {
40 URI uri = null;
41 try {
42 uri = new URI("abc://def.ghi.jkl");
43 } catch (URI.MalformedURIException ex) {
44 }
45 BASE_URI = uri;
46 }
47
48 public short getAllowedFacets(){
49 return (XSSimpleTypeDecl.FACET_LENGTH | XSSimpleTypeDecl.FACET_MINLENGTH | XSSimpleTypeDecl.FACET_MAXLENGTH | XSSimpleTypeDecl.FACET_PATTERN | XSSimpleTypeDecl.FACET_ENUMERATION | XSSimpleTypeDecl.FACET_WHITESPACE );
50 }
51
52 // before we return string we have to make sure it is correct URI as per spec.
53 // for some types (string and derived), they just return the string itself
54 public Object getActualValue(String content, ValidationContext context) throws InvalidDatatypeValueException {
55 // check 3.2.17.c0 must: URI (rfc 2396/2723)
56 try {
57 if( content.length() != 0 ) {
58 // encode special characters using XLink 5.4 algorithm
59 final String encoded = encode(content);
60 // Support for relative URLs
61 // According to Java 1.1: URLs may also be specified with a
62 // String and the URL object that it is related to.
63 new URI(BASE_URI, encoded );
64 }
65 } catch (URI.MalformedURIException ex) {
66 throw new InvalidDatatypeValueException("cvc-datatype-valid.1.2.1", new Object[]{content, "anyURI"});
67 }
68
69 // REVISIT: do we need to return the new URI object?
70 return content;
71 }
72
73 // which ASCII characters need to be escaped
74 private static boolean gNeedEscaping[] = new boolean[128];
75 // the first hex character if a character needs to be escaped
76 private static char gAfterEscaping1[] = new char[128];
77 // the second hex character if a character needs to be escaped
78 private static char gAfterEscaping2[] = new char[128];
79 private static char[] gHexChs = {'0', '1', '2', '3', '4', '5', '6', '7',
80 '8', '9', 'A', 'B', 'C', 'D', 'E', 'F'};
81 // initialize the above 3 arrays
82 static {
83 for (int i = 0; i <= 0x1f; i++) {
84 gNeedEscaping[i] = true;
85 gAfterEscaping1[i] = gHexChs[i >> 4];
86 gAfterEscaping2[i] = gHexChs[i & 0xf];
87 }
88 gNeedEscaping[0x7f] = true;
89 gAfterEscaping1[0x7f] = '7';
90 gAfterEscaping2[0x7f] = 'F';
91 char[] escChs = {' ', '<', '>', '"', '{', '}',
92 '|', '\\', '^', '~', '`'};
93 int len = escChs.length;
94 char ch;
95 for (int i = 0; i < len; i++) {
96 ch = escChs[i];
97 gNeedEscaping[ch] = true;
98 gAfterEscaping1[ch] = gHexChs[ch >> 4];
99 gAfterEscaping2[ch] = gHexChs[ch & 0xf];
100 }
101 }
102
103 // To encode special characters in anyURI, by using %HH to represent
104 // special ASCII characters: 0x00~0x1F, 0x7F, ' ', '<', '>', etc.
105 // and non-ASCII characters (whose value >= 128).
106 private static String encode(String anyURI){
107 int len = anyURI.length(), ch;
108 StringBuffer buffer = new StringBuffer(len*3);
109
110 // for each character in the anyURI
111 int i = 0;
112 for (; i < len; i++) {
113 ch = anyURI.charAt(i);
114 // if it's not an ASCII character, break here, and use UTF-8 encoding
115 if (ch >= 128)
116 break;
117 if (gNeedEscaping[ch]) {
118 buffer.append('%');
119 buffer.append(gAfterEscaping1[ch]);
120 buffer.append(gAfterEscaping2[ch]);
121 }
122 else {
123 buffer.append((char)ch);
124 }
125 }
126
127 // we saw some non-ascii character
128 if (i < len) {
129 // get UTF-8 bytes for the remaining sub-string
130 byte[] bytes = null;
131 byte b;
132 try {
133 bytes = anyURI.substring(i).getBytes("UTF-8");
134 } catch (java.io.UnsupportedEncodingException e) {
135 // should never happen
136 return anyURI;
137 }
138 len = bytes.length;
139
140 // for each byte
141 for (i = 0; i < len; i++) {
142 b = bytes[i];
143 // for non-ascii character: make it positive, then escape
144 if (b < 0) {
145 ch = b + 256;
146 buffer.append('%');
147 buffer.append(gHexChs[ch >> 4]);
148 buffer.append(gHexChs[ch & 0xf]);
149 }
150 else if (gNeedEscaping[b]) {
151 buffer.append('%');
152 buffer.append(gAfterEscaping1[b]);
153 buffer.append(gAfterEscaping2[b]);
154 }
155 else {
156 buffer.append((char)b);
157 }
158 }
159 }
160
161 // If encoding happened, create a new string;
162 // otherwise, return the orginal one.
163 if (buffer.length() != len) {
164 return buffer.toString();
165 }
166 else {
167 return anyURI;
168 }
169 }
170
171 } // class AnyURIDV