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 import java.util.BitSet;
25
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
31 /**
32 * Similar to the Quoted-Printable content-transfer-encoding defined in
33 * <a href="https://www.ietf.org/rfc/rfc1521.txt">RFC 1521</a> and designed to allow text containing mostly ASCII
34 * characters to be decipherable on an ASCII terminal without decoding.
35 * <p>
36 * <a href="https://www.ietf.org/rfc/rfc1522.txt">RFC 1522</a> describes techniques to allow the encoding of non-ASCII
37 * text in various portions of a RFC 822 [2] message header, in a manner which is unlikely to confuse existing message
38 * handling software.
39 * </p>
40 * <p>
41 * This class is conditionally thread-safe.
42 * The instance field for encoding blanks is mutable {@link #setEncodeBlanks(boolean)}
43 * but is not volatile, and accesses are not synchronized.
44 * If an instance of the class is shared between threads, the caller needs to ensure that suitable synchronization
45 * is used to ensure safe publication of the value between threads, and must not invoke
46 * {@link #setEncodeBlanks(boolean)} after initial setup.
47 * </p>
48 *
49 * @see <a href="https://www.ietf.org/rfc/rfc1522.txt">MIME (Multipurpose Internet Mail Extensions) Part Two: Message
50 * Header Extensions for Non-ASCII Text</a>
51 *
52 * @since 1.3
53 */
54 public class QCodec extends RFC1522Codec implements StringEncoder, StringDecoder {
55
56 /**
57 * BitSet of printable characters as defined in RFC 1522.
58 */
59 private static final BitSet PRINTABLE_CHARS = new BitSet(256);
60
61 // Static initializer for printable chars collection
62 static {
63 // alpha characters
64 PRINTABLE_CHARS.set(' ');
65 PRINTABLE_CHARS.set('!');
66 PRINTABLE_CHARS.set('"');
67 PRINTABLE_CHARS.set('#');
68 PRINTABLE_CHARS.set('$');
69 PRINTABLE_CHARS.set('%');
70 PRINTABLE_CHARS.set('&');
71 PRINTABLE_CHARS.set('\'');
72 PRINTABLE_CHARS.set('(');
73 PRINTABLE_CHARS.set(')');
74 PRINTABLE_CHARS.set('*');
75 PRINTABLE_CHARS.set('+');
76 PRINTABLE_CHARS.set(',');
77 PRINTABLE_CHARS.set('-');
78 PRINTABLE_CHARS.set('.');
79 PRINTABLE_CHARS.set('/');
80 for (int i = '0'; i <= '9'; i++) {
81 PRINTABLE_CHARS.set(i);
82 }
83 PRINTABLE_CHARS.set(':');
84 PRINTABLE_CHARS.set(';');
85 PRINTABLE_CHARS.set('<');
86 PRINTABLE_CHARS.set('>');
87 PRINTABLE_CHARS.set('@');
88 for (int i = 'A'; i <= 'Z'; i++) {
89 PRINTABLE_CHARS.set(i);
90 }
91 PRINTABLE_CHARS.set('[');
92 PRINTABLE_CHARS.set('\\');
93 PRINTABLE_CHARS.set(']');
94 PRINTABLE_CHARS.set('^');
95 PRINTABLE_CHARS.set('`');
96 for (int i = 'a'; i <= 'z'; i++) {
97 PRINTABLE_CHARS.set(i);
98 }
99 PRINTABLE_CHARS.set('{');
100 PRINTABLE_CHARS.set('|');
101 PRINTABLE_CHARS.set('}');
102 PRINTABLE_CHARS.set('~');
103 }
104
105 private static final byte UNDERSCORE = 95;
106
107 private boolean encodeBlanks;
108
109 /**
110 * Constructs a new instance.
111 */
112 public QCodec() {
113 this(StandardCharsets.UTF_8);
114 }
115
116 /**
117 * Constructs a new instance for the selection of a default Charset.
118 *
119 * @param charset
120 * the default string Charset to use.
121 *
122 * @see Charset
123 * @since 1.7
124 */
125 public QCodec(final Charset charset) {
126 super(charset);
127 }
128
129 /**
130 * Constructs a new instance for the selection of a default Charset.
131 *
132 * @param charsetName
133 * the Charset to use.
134 * @throws java.nio.charset.UnsupportedCharsetException
135 * Thrown if the named Charset is unavailable.
136 * @since 1.7 throws UnsupportedCharsetException if the named Charset is unavailable
137 * @see Charset
138 */
139 public QCodec(final String charsetName) {
140 this(Charset.forName(charsetName));
141 }
142
143 /**
144 * Decodes a quoted-printable object into its original form. Escaped characters are converted back to their original
145 * representation.
146 *
147 * @param obj
148 * quoted-printable object to convert into its original form.
149 * @return original object.
150 * @throws DecoderException
151 * Thrown if the argument is not a {@code String}. Thrown if a failure condition is encountered
152 * during the decode process.
153 */
154 @Override
155 public Object decode(final Object obj) throws DecoderException {
156 if (obj == null) {
157 return null;
158 }
159 if (obj instanceof String) {
160 return decode((String) obj);
161 }
162 throw new DecoderException("Objects of type " + obj.getClass().getName() + " cannot be decoded using Q codec");
163 }
164
165 /**
166 * Decodes a quoted-printable string into its original form. Escaped characters are converted back to their original
167 * representation.
168 *
169 * <p>
170 * Uses {@link QuotedPrintableCodec#decodeQuotedPrintable(byte[])} to decode the encoded text. Since 1.23.0, unescaped CR and LF bytes in malformed
171 * encoded words are preserved rather than discarded, and {@code =CR} without a following LF is rejected. This lenient handling does not make such
172 * encoded words valid under RFC 2047.
173 * </p>
174 *
175 * @param str
176 * quoted-printable string to convert into its original form.
177 * @return original string.
178 * @throws DecoderException
179 * Thrown if a failure condition is encountered during the decoding process.
180 */
181 @Override
182 public String decode(final String str) throws DecoderException {
183 try {
184 return decodeText(str);
185 } catch (final UnsupportedEncodingException e) {
186 throw new DecoderException(e.getMessage(), e);
187 }
188 }
189
190 @Override
191 protected byte[] doDecoding(final byte[] bytes) throws DecoderException {
192 if (bytes == null) {
193 return null;
194 }
195 boolean hasUnderscores = false;
196 for (final byte b : bytes) {
197 if (b == UNDERSCORE) {
198 hasUnderscores = true;
199 break;
200 }
201 }
202 if (hasUnderscores) {
203 final byte[] tmp = new byte[bytes.length];
204 for (int i = 0; i < bytes.length; i++) {
205 final byte b = bytes[i];
206 if (b != UNDERSCORE) {
207 tmp[i] = b;
208 } else {
209 tmp[i] = Utils.SPACE;
210 }
211 }
212 return QuotedPrintableCodec.decodeQuotedPrintable(tmp);
213 }
214 return QuotedPrintableCodec.decodeQuotedPrintable(bytes);
215 }
216
217 @Override
218 protected byte[] doEncoding(final byte[] bytes) {
219 if (bytes == null) {
220 return null;
221 }
222 final byte[] data = QuotedPrintableCodec.encodeQuotedPrintable(PRINTABLE_CHARS, bytes);
223 if (this.encodeBlanks) {
224 for (int i = 0; i < data.length; i++) {
225 if (data[i] == Utils.SPACE) {
226 data[i] = UNDERSCORE;
227 }
228 }
229 }
230 return data;
231 }
232
233 /**
234 * Encodes an object into its quoted-printable form using the default Charset. Unsafe characters are escaped.
235 *
236 * @param obj
237 * object to convert to quoted-printable form.
238 * @return quoted-printable object.
239 * @throws EncoderException
240 * Thrown if a failure condition is encountered during the encoding process.
241 */
242 @Override
243 public Object encode(final Object obj) throws EncoderException {
244 if (obj == null) {
245 return null;
246 }
247 if (obj instanceof String) {
248 return encode((String) obj);
249 }
250 throw new EncoderException("Objects of type " + obj.getClass().getName() + " cannot be encoded using Q codec");
251 }
252
253 /**
254 * Encodes a string into its quoted-printable form using the default Charset. Unsafe characters are escaped.
255 *
256 * @param sourceStr
257 * string to convert to quoted-printable form.
258 * @return quoted-printable string.
259 * @throws EncoderException
260 * Thrown if a failure condition is encountered during the encoding process.
261 */
262 @Override
263 public String encode(final String sourceStr) throws EncoderException {
264 return encode(sourceStr, getCharset());
265 }
266
267 /**
268 * Encodes a string into its quoted-printable form using the specified Charset. Unsafe characters are escaped.
269 *
270 * @param sourceStr
271 * string to convert to quoted-printable form.
272 * @param sourceCharset
273 * the Charset for sourceStr.
274 * @return quoted-printable string.
275 * @throws EncoderException
276 * Thrown if a failure condition is encountered during the encoding process.
277 * @since 1.7
278 */
279 public String encode(final String sourceStr, final Charset sourceCharset) throws EncoderException {
280 return encodeText(sourceStr, sourceCharset);
281 }
282
283 /**
284 * Encodes a string into its quoted-printable form using the specified Charset. Unsafe characters are escaped.
285 *
286 * @param sourceStr
287 * string to convert to quoted-printable form.
288 * @param sourceCharset
289 * the Charset for sourceStr.
290 * @return quoted-printable string.
291 * @throws EncoderException
292 * Thrown if a failure condition is encountered during the encoding process.
293 */
294 public String encode(final String sourceStr, final String sourceCharset) throws EncoderException {
295 try {
296 return encodeText(sourceStr, sourceCharset);
297 } catch (final UnsupportedCharsetException e) {
298 throw new EncoderException(e.getMessage(), e);
299 }
300 }
301
302 @Override
303 protected String getEncoding() {
304 return "Q";
305 }
306
307 /**
308 * Tests whether the optional transformation of SPACE characters is to be used.
309 *
310 * @return {@code true} if SPACE characters are to be transformed, {@code false} otherwise.
311 */
312 public boolean isEncodeBlanks() {
313 return this.encodeBlanks;
314 }
315
316 /**
317 * Sets whether the optional transformation of SPACE characters is to be used.
318 *
319 * @param b
320 * {@code true} if SPACE characters are to be transformed, {@code false} otherwise.
321 */
322 public void setEncodeBlanks(final boolean b) {
323 this.encodeBlanks = b;
324 }
325 }