1 /*
2 * Licensed to the Apache Software Foundation (ASF) under one or more
3 * contributor license agreements. See the NOTICE file distributed with
4 * this work for additional information regarding copyright ownership.
5 * The ASF licenses this file to You under the Apache License, Version 2.0
6 * (the "License"); you may not use this file except in compliance with
7 * the License. You may obtain a copy of the License at
8 *
9 * https://www.apache.org/licenses/LICENSE-2.0
10 *
11 * Unless required by applicable law or agreed to in writing, software
12 * distributed under the License is distributed on an "AS IS" BASIS,
13 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 * See the License for the specific language governing permissions and
15 * limitations under the License.
16 */
17
18 package org.apache.commons.codec.net;
19
20 import java.io.UnsupportedEncodingException;
21 import java.nio.charset.Charset;
22 import java.nio.charset.StandardCharsets;
23 import java.nio.charset.UnsupportedCharsetException;
24
25 import org.apache.commons.codec.CodecPolicy;
26 import org.apache.commons.codec.DecoderException;
27 import org.apache.commons.codec.EncoderException;
28 import org.apache.commons.codec.StringDecoder;
29 import org.apache.commons.codec.StringEncoder;
30 import org.apache.commons.codec.binary.Base64;
31 import org.apache.commons.codec.binary.BaseNCodec;
32
33 /**
34 * Identical to the Base64 encoding defined by <a href="https://www.ietf.org/rfc/rfc1521.txt">RFC 1521</a>
35 * and allows a character set to be specified.
36 * <p>
37 * <a href="https://www.ietf.org/rfc/rfc1522.txt">RFC 1522</a> describes techniques to allow the encoding of non-ASCII
38 * text in various portions of a RFC 822 [2] message header, in a manner which is unlikely to confuse existing message
39 * handling software.
40 * </p>
41 * <p>
42 * This class is immutable and thread-safe.
43 * </p>
44 *
45 * <p>
46 * Decoding is lenient by default: the Base64 payload can contain ignored characters, noncanonical padding or trailing bits, and data after padding.
47 * Different encoded words can therefore decode to the same text. To require a canonical Base64 payload, select {@link CodecPolicy#STRICT}:
48 * </p>
49 *
50 * <pre>
51 * BCodec codec = new BCodec(StandardCharsets.UTF_8, CodecPolicy.STRICT);
52 * </pre>
53 *
54 * <p>
55 * Strict decoding requires the standard Base64 alphabet, padding for partial blocks, and no whitespace within the payload. Invalid payloads cause a
56 * {@link DecoderException}. This validates the Base64 payload only; it does not establish a unique representation of the complete encoded word or message
57 * header, including its charset label. Applications comparing header values for security decisions must use a consistent representation, and signature
58 * verification must follow the signing protocol.
59 * </p>
60 *
61 * @see <a href="https://www.ietf.org/rfc/rfc1522.txt">MIME (Multipurpose Internet Mail Extensions) Part Two: Message
62 * Header Extensions for Non-ASCII Text</a>
63 *
64 * @since 1.3
65 */
66 public class BCodec extends RFC1522Codec implements StringEncoder, StringDecoder {
67
68 /**
69 * The default decoding policy is lenient.
70 */
71 private static final CodecPolicy DECODING_POLICY_DEFAULT = CodecPolicy.LENIENT;
72
73 /**
74 * Decoding policy for the Base64 payload. The default is lenient; strict decoding requires a canonical payload.
75 */
76 private final CodecPolicy decodingPolicy;
77
78 /**
79 * Constructs a new instance.
80 */
81 public BCodec() {
82 this(StandardCharsets.UTF_8);
83 }
84
85 /**
86 * Constructs a new instance for the selection of a default Charset.
87 *
88 * @param charset
89 * the default string Charset to use.
90 *
91 * @see Charset
92 * @since 1.7
93 */
94 public BCodec(final Charset charset) {
95 this(charset, DECODING_POLICY_DEFAULT);
96 }
97
98 /**
99 * Constructs a new instance for the selection of a default Charset.
100 *
101 * <p>
102 * Use {@link CodecPolicy#STRICT} to require canonical standard Base64 payloads. The other constructors use {@link CodecPolicy#LENIENT}.
103 * This policy applies to the Base64 payload, not the complete encoded word; see the class documentation.
104 * </p>
105 *
106 * @param charset
107 * the default string Charset to use.
108 * @param decodingPolicy The decoding policy.
109 * @see Charset
110 * @since 1.15
111 */
112 public BCodec(final Charset charset, final CodecPolicy decodingPolicy) {
113 super(charset);
114 this.decodingPolicy = decodingPolicy;
115 }
116
117 /**
118 * Constructs a new instance for the selection of a default Charset.
119 *
120 * @param charsetName
121 * the default Charset to use.
122 * @throws java.nio.charset.UnsupportedCharsetException
123 * Thrown if the named Charset is unavailable.
124 * @since 1.7 throws UnsupportedCharsetException if the named Charset is unavailable
125 * @see Charset
126 */
127 public BCodec(final String charsetName) {
128 this(Charset.forName(charsetName));
129 }
130
131 /**
132 * Decodes a Base64 object into its original form. Escaped characters are converted back to their original
133 * representation.
134 *
135 * <p>
136 * Uses the decoding policy selected at construction. The default is lenient and does not require a canonical Base64 payload. Use
137 * {@link #BCodec(Charset, CodecPolicy)} with {@link CodecPolicy#STRICT} for canonical payload validation.
138 * </p>
139 *
140 * @param value
141 * Base64 object to convert into its original form.
142 * @return original object.
143 * @throws DecoderException
144 * Thrown if the argument is not a {@code String}. Thrown if a failure condition is encountered
145 * during the decode process.
146 */
147 @Override
148 public Object decode(final Object value) throws DecoderException {
149 if (value == null) {
150 return null;
151 }
152 if (value instanceof String) {
153 return decode((String) value);
154 }
155 throw new DecoderException("Objects of type " + value.getClass().getName() + " cannot be decoded using BCodec");
156 }
157
158 /**
159 * Decodes a Base64 string into its original form. Escaped characters are converted back to their original
160 * representation.
161 *
162 * <p>
163 * Uses the decoding policy selected at construction. The default is lenient and does not require a canonical Base64 payload. Use
164 * {@link #BCodec(Charset, CodecPolicy)} with {@link CodecPolicy#STRICT} for canonical payload validation.
165 * </p>
166 *
167 * @param value
168 * Base64 string to convert into its original form.
169 * @return original string.
170 * @throws DecoderException
171 * Thrown if a failure condition is encountered during the decoding process.
172 */
173 @Override
174 public String decode(final String value) throws DecoderException {
175 try {
176 return decodeText(value);
177 } catch (final UnsupportedEncodingException | IllegalArgumentException e) {
178 throw new DecoderException(e.getMessage(), e);
179 }
180 }
181
182 /**
183 * {@inheritDoc}
184 *
185 * @throws IllegalArgumentException Thrown when a problem is detected processing data.
186 */
187 @Override
188 protected byte[] doDecoding(final byte[] bytes) throws DecoderException {
189 if (bytes == null) {
190 return null;
191 }
192 // @formatter:off
193 try {
194 return Base64.builder()
195 .setLineLength(0)
196 .setLineSeparator(BaseNCodec.getChunkSeparator())
197 .setUrlSafe(false)
198 .setDecodingPolicy(decodingPolicy)
199 .get()
200 .decode(bytes);
201 } catch (final IllegalArgumentException e) {
202 throw new DecoderException(e.getMessage(), e);
203 }
204 // @formatter:on
205 }
206
207 @Override
208 protected byte[] doEncoding(final byte[] bytes) {
209 if (bytes == null) {
210 return null;
211 }
212 return Base64.encodeBase64(bytes);
213 }
214
215 /**
216 * Encodes an object into its Base64 form using the default Charset. Unsafe characters are escaped.
217 *
218 * @param value
219 * object to convert to Base64 form.
220 * @return Base64 object.
221 * @throws EncoderException
222 * Thrown if a failure condition is encountered during the encoding process.
223 */
224 @Override
225 public Object encode(final Object value) throws EncoderException {
226 if (value == null) {
227 return null;
228 }
229 if (value instanceof String) {
230 return encode((String) value);
231 }
232 throw new EncoderException("Objects of type " + value.getClass().getName() + " cannot be encoded using BCodec");
233 }
234
235 /**
236 * Encodes a string into its Base64 form using the default Charset. Unsafe characters are escaped.
237 *
238 * @param strSource
239 * string to convert to Base64 form.
240 * @return Base64 string.
241 * @throws EncoderException
242 * Thrown if a failure condition is encountered during the encoding process.
243 */
244 @Override
245 public String encode(final String strSource) throws EncoderException {
246 return encode(strSource, getCharset());
247 }
248
249 /**
250 * Encodes a string into its Base64 form using the specified Charset. Unsafe characters are escaped.
251 *
252 * @param strSource
253 * string to convert to Base64 form.
254 * @param sourceCharset
255 * the Charset for {@code value}.
256 * @return Base64 string.
257 * @throws EncoderException
258 * Thrown if a failure condition is encountered during the encoding process.
259 * @since 1.7
260 */
261 public String encode(final String strSource, final Charset sourceCharset) throws EncoderException {
262 return encodeText(strSource, sourceCharset);
263 }
264
265 /**
266 * Encodes a string into its Base64 form using the specified Charset. Unsafe characters are escaped.
267 *
268 * @param strSource
269 * string to convert to Base64 form.
270 * @param sourceCharset
271 * the Charset for {@code value}.
272 * @return Base64 string.
273 * @throws EncoderException
274 * Thrown if a failure condition is encountered during the encoding process.
275 */
276 public String encode(final String strSource, final String sourceCharset) throws EncoderException {
277 try {
278 return encodeText(strSource, sourceCharset);
279 } catch (final UnsupportedCharsetException e) {
280 throw new EncoderException(e.getMessage(), e);
281 }
282 }
283
284 @Override
285 protected String getEncoding() {
286 return "B";
287 }
288
289 /**
290 * Tests whether decoding requires a canonical Base64 payload.
291 *
292 * <p>
293 * Strict decoding raises {@link DecoderException} for a noncanonical Base64 payload, including invalid alphabet characters, padding, or trailing bits.
294 * The default is lenient. This policy does not establish a canonical representation of the complete encoded word.
295 * </p>
296 *
297 * @return true if using strict decoding.
298 * @since 1.15
299 */
300 public boolean isStrictDecoding() {
301 return decodingPolicy == CodecPolicy.STRICT;
302 }
303 }