1 /*
2 * ====================================================================
3 * Licensed to the Apache Software Foundation (ASF) under one
4 * or more contributor license agreements. See the NOTICE file
5 * distributed with this work for additional information
6 * regarding copyright ownership. The ASF licenses this file
7 * to you under the Apache License, Version 2.0 (the
8 * "License"); you may not use this file except in compliance
9 * with the License. You may obtain a copy of the License at
10 *
11 * http://www.apache.org/licenses/LICENSE-2.0
12 *
13 * Unless required by applicable law or agreed to in writing,
14 * software distributed under the License is distributed on an
15 * "AS IS" BASIS, WITHOUT WARRANTIES OR CONDITIONS OF ANY
16 * KIND, either express or implied. See the License for the
17 * specific language governing permissions and limitations
18 * under the License.
19 * ====================================================================
20 *
21 * This software consists of voluntary contributions made by many
22 * individuals on behalf of the Apache Software Foundation. For more
23 * information on the Apache Software Foundation, please see
24 * <http://www.apache.org/>.
25 *
26 */
27
28 package org.apache.hc.core5.http.io.entity;
29
30 import java.io.IOException;
31 import java.io.InputStream;
32 import java.io.InputStreamReader;
33 import java.io.Reader;
34 import java.io.UnsupportedEncodingException;
35 import java.nio.charset.Charset;
36 import java.nio.charset.StandardCharsets;
37 import java.nio.charset.UnsupportedCharsetException;
38 import java.util.Collections;
39 import java.util.HashMap;
40 import java.util.List;
41 import java.util.Map;
42
43 import org.apache.hc.core5.http.ContentType;
44 import org.apache.hc.core5.http.HttpEntity;
45 import org.apache.hc.core5.http.NameValuePair;
46 import org.apache.hc.core5.http.ParseException;
47 import org.apache.hc.core5.io.Closer;
48 import org.apache.hc.core5.net.WWWFormCodec;
49 import org.apache.hc.core5.util.Args;
50 import org.apache.hc.core5.util.ByteArrayBuffer;
51 import org.apache.hc.core5.util.CharArrayBuffer;
52
53 /**
54 * Support methods for {@link HttpEntity}.
55 *
56 * @since 4.0
57 */
58 public final class EntityUtils {
59
60 // TODO Consider using a sane value, but what is sane? 1 GB? 100 MB? 10 MB?
61 private static final int DEFAULT_ENTITY_RETURN_MAX_LENGTH = Integer.MAX_VALUE;
62 private static final Charset DEFAULT_CHARSET = StandardCharsets.ISO_8859_1;
63 private static final int DEFAULT_CHAR_BUFFER_SIZE = 1024;
64 private static final int DEFAULT_BYTE_BUFFER_SIZE = 4096;
65
66 private EntityUtils() {
67 // NoOp
68 }
69
70 /**
71 * Ensures that the entity content is fully consumed and the content stream, if exists,
72 * is closed. The process is done, <i>quietly</i> , without throwing any IOException.
73 *
74 * @param entity the entity to consume.
75 *
76 * @since 4.2
77 */
78 public static void consumeQuietly(final HttpEntity entity) {
79 try {
80 consume(entity);
81 } catch (final IOException ignore) {
82 // Ignore exception
83 }
84 }
85
86 /**
87 * Ensures that the entity content is fully consumed and the content stream, if exists,
88 * is closed.
89 *
90 * @param entity the entity to consume.
91 * @throws IOException if an error occurs reading the input stream
92 *
93 * @since 4.1
94 */
95 public static void consume(final HttpEntity entity) throws IOException {
96 if (entity == null) {
97 return;
98 }
99 if (entity.isStreaming()) {
100 Closer.close(entity.getContent());
101 }
102 }
103
104 /**
105 * Gets a usable content length value for the given candidate.
106 *
107 * @param contentLength an integer.
108 * @return The given content length or {@value #DEFAULT_BYTE_BUFFER_SIZE} if it is < 0.
109 */
110 private static int toContentLength(final int contentLength) {
111 return contentLength < 0 ? DEFAULT_BYTE_BUFFER_SIZE : contentLength;
112 }
113
114 /**
115 * Reads the contents of an entity and return it as a byte array.
116 *
117 * @param entity the entity to read from=
118 * @return byte array containing the entity content. May be null if
119 * {@link HttpEntity#getContent()} is null.
120 * @throws IOException if an error occurs reading the input stream
121 * @throws IllegalArgumentException if entity is null or if content length > Integer.MAX_VALUE
122 */
123 public static byte[] toByteArray(final HttpEntity entity) throws IOException {
124 Args.notNull(entity, "HttpEntity");
125 final int contentLength = toContentLength((int) Args.checkContentLength(entity));
126 try (final InputStream inStream = entity.getContent()) {
127 if (inStream == null) {
128 return null;
129 }
130 final ByteArrayBuffereArrayBuffer.html#ByteArrayBuffer">ByteArrayBuffer buffer = new ByteArrayBuffer(contentLength);
131 final byte[] tmp = new byte[DEFAULT_BYTE_BUFFER_SIZE];
132 int l;
133 while ((l = inStream.read(tmp)) != -1) {
134 buffer.append(tmp, 0, l);
135 }
136 return buffer.toByteArray();
137 }
138 }
139
140 /**
141 * Reads the contents of an entity and return it as a byte array.
142 *
143 * @param entity the entity to read from=
144 * @return byte array containing the entity content. May be null if
145 * {@link HttpEntity#getContent()} is null.
146 * @param maxResultLength
147 * The maximum size of the String to return; use it to guard against unreasonable or malicious processing.
148 * @throws IOException if an error occurs reading the input stream
149 * @throws IllegalArgumentException if entity is null or if content length > Integer.MAX_VALUE
150 */
151 public static byte[] toByteArray(final HttpEntity entity, final int maxResultLength) throws IOException {
152 Args.notNull(entity, "HttpEntity");
153 final int contentLength = toContentLength((int) Args.checkContentLength(entity));
154 try (final InputStream inStream = entity.getContent()) {
155 if (inStream == null) {
156 return null;
157 }
158 final ByteArrayBuffereArrayBuffer.html#ByteArrayBuffer">ByteArrayBuffer buffer = new ByteArrayBuffer(Math.min(maxResultLength, contentLength));
159 final byte[] tmp = new byte[DEFAULT_BYTE_BUFFER_SIZE];
160 int l;
161 while ((l = inStream.read(tmp, 0, Math.min(DEFAULT_BYTE_BUFFER_SIZE, buffer.capacity() - buffer.length()))) > 0) {
162 buffer.append(tmp, 0, l);
163 }
164 return buffer.toByteArray();
165 }
166 }
167
168 private static CharArrayBuffer toCharArrayBuffer(final InputStream inStream, final int contentLength,
169 final Charset charset, final int maxResultLength) throws IOException {
170 Args.notNull(inStream, "InputStream");
171 Args.positive(maxResultLength, "maxResultLength");
172 final Charset actualCharset = charset == null ? DEFAULT_CHARSET : charset;
173 final CharArrayBufferCharArrayBuffer.html#CharArrayBuffer">CharArrayBuffer buf = new CharArrayBuffer(
174 Math.min(maxResultLength, contentLength > 0 ? contentLength : DEFAULT_CHAR_BUFFER_SIZE));
175 final Reader reader = new InputStreamReader(inStream, actualCharset);
176 final char[] tmp = new char[DEFAULT_CHAR_BUFFER_SIZE];
177 int chReadCount;
178 while ((chReadCount = reader.read(tmp)) != -1 && buf.length() < maxResultLength) {
179 buf.append(tmp, 0, chReadCount);
180 }
181 buf.setLength(Math.min(buf.length(), maxResultLength));
182 return buf;
183 }
184
185 private static final Map<String, ContentType> CONTENT_TYPE_MAP;
186 static {
187 final ContentType[] contentTypes = {
188 ContentType.APPLICATION_ATOM_XML,
189 ContentType.APPLICATION_FORM_URLENCODED,
190 ContentType.APPLICATION_JSON,
191 ContentType.APPLICATION_SVG_XML,
192 ContentType.APPLICATION_XHTML_XML,
193 ContentType.APPLICATION_XML,
194 ContentType.MULTIPART_FORM_DATA,
195 ContentType.TEXT_HTML,
196 ContentType.TEXT_PLAIN,
197 ContentType.TEXT_XML };
198 final HashMap<String, ContentType> map = new HashMap<>();
199 for (final ContentType contentType: contentTypes) {
200 map.put(contentType.getMimeType(), contentType);
201 }
202 CONTENT_TYPE_MAP = Collections.unmodifiableMap(map);
203 }
204
205 private static String toString(final HttpEntity entity, final ContentType contentType, final int maxResultLength)
206 throws IOException {
207 Args.notNull(entity, "HttpEntity");
208 final int contentLength = toContentLength((int) Args.checkContentLength(entity));
209 try (final InputStream inStream = entity.getContent()) {
210 if (inStream == null) {
211 return null;
212 }
213 Charset charset = null;
214 if (contentType != null) {
215 charset = contentType.getCharset();
216 if (charset == null) {
217 final ContentType defaultContentType = CONTENT_TYPE_MAP.get(contentType.getMimeType());
218 charset = defaultContentType != null ? defaultContentType.getCharset() : null;
219 }
220 }
221 return toCharArrayBuffer(inStream, contentLength, charset, maxResultLength).toString();
222 }
223 }
224
225 /**
226 * Gets the entity content as a String, using the provided default character set
227 * if none is found in the entity.
228 * If defaultCharset is null, the default "ISO-8859-1" is used.
229 *
230 * @param entity must not be null
231 * @param defaultCharset character set to be applied if none found in the entity,
232 * or if the entity provided charset is invalid or not available.
233 * @return the entity content as a String. May be null if
234 * {@link HttpEntity#getContent()} is null.
235 * @throws ParseException if header elements cannot be parsed
236 * @throws IllegalArgumentException if entity is null or if content length > Integer.MAX_VALUE
237 * @throws IOException if an error occurs reading the input stream
238 * @throws java.nio.charset.UnsupportedCharsetException Thrown when the named entity's charset is not available in
239 * this instance of the Java virtual machine and no defaultCharset is provided.
240 */
241 public static String toString(
242 final HttpEntity entity, final Charset defaultCharset) throws IOException, ParseException {
243 return toString(entity, defaultCharset, DEFAULT_ENTITY_RETURN_MAX_LENGTH);
244 }
245
246 /**
247 * Gets the entity content as a String, using the provided default character set
248 * if none is found in the entity.
249 * If defaultCharset is null, the default "ISO-8859-1" is used.
250 *
251 * @param entity must not be null
252 * @param defaultCharset character set to be applied if none found in the entity,
253 * or if the entity provided charset is invalid or not available.
254 * @param maxResultLength
255 * The maximum size of the String to return; use it to guard against unreasonable or malicious processing.
256 * @return the entity content as a String. May be null if
257 * {@link HttpEntity#getContent()} is null.
258 * @throws ParseException if header elements cannot be parsed
259 * @throws IllegalArgumentException if entity is null or if content length > Integer.MAX_VALUE
260 * @throws IOException if an error occurs reading the input stream
261 * @throws java.nio.charset.UnsupportedCharsetException Thrown when the named entity's charset is not available in
262 * this instance of the Java virtual machine and no defaultCharset is provided.
263 */
264 public static String toString(
265 final HttpEntity entity, final Charset defaultCharset, final int maxResultLength) throws IOException, ParseException {
266 Args.notNull(entity, "HttpEntity");
267 ContentType contentType = null;
268 try {
269 contentType = ContentType.parse(entity.getContentType());
270 } catch (final UnsupportedCharsetException ex) {
271 if (defaultCharset == null) {
272 throw new UnsupportedEncodingException(ex.getMessage());
273 }
274 }
275 if (contentType != null) {
276 if (contentType.getCharset() == null) {
277 contentType = contentType.withCharset(defaultCharset);
278 }
279 } else {
280 contentType = ContentType.DEFAULT_TEXT.withCharset(defaultCharset);
281 }
282 return toString(entity, contentType, maxResultLength);
283 }
284
285 /**
286 * Gets the entity content as a String, using the provided default character set
287 * if none is found in the entity.
288 * If defaultCharset is null, the default "ISO-8859-1" is used.
289 *
290 * @param entity must not be null
291 * @param defaultCharset character set to be applied if none found in the entity
292 * @return the entity content as a String. May be null if
293 * {@link HttpEntity#getContent()} is null.
294 * @throws ParseException if header elements cannot be parsed
295 * @throws IllegalArgumentException if entity is null or if content length > Integer.MAX_VALUE
296 * @throws IOException if an error occurs reading the input stream
297 * @throws java.nio.charset.UnsupportedCharsetException Thrown when the named charset is not available in
298 * this instance of the Java virtual machine
299 */
300 public static String toString(
301 final HttpEntity entity, final String defaultCharset) throws IOException, ParseException {
302 return toString(entity, defaultCharset, DEFAULT_ENTITY_RETURN_MAX_LENGTH);
303 }
304
305 /**
306 * Gets the entity content as a String, using the provided default character set
307 * if none is found in the entity.
308 * If defaultCharset is null, the default "ISO-8859-1" is used.
309 *
310 * @param entity must not be null
311 * @param defaultCharset character set to be applied if none found in the entity
312 * @param maxResultLength
313 * The maximum size of the String to return; use it to guard against unreasonable or malicious processing.
314 * @return the entity content as a String. May be null if
315 * {@link HttpEntity#getContent()} is null.
316 * @throws ParseException if header elements cannot be parsed
317 * @throws IllegalArgumentException if entity is null or if content length > Integer.MAX_VALUE
318 * @throws IOException if an error occurs reading the input stream
319 * @throws java.nio.charset.UnsupportedCharsetException Thrown when the named charset is not available in
320 * this instance of the Java virtual machine
321 */
322 public static String toString(
323 final HttpEntity entity, final String defaultCharset, final int maxResultLength) throws IOException, ParseException {
324 return toString(entity, defaultCharset != null ? Charset.forName(defaultCharset) : null, maxResultLength);
325 }
326
327 /**
328 * Reads the contents of an entity and return it as a String.
329 * The content is converted using the character set from the entity (if any),
330 * failing that, "ISO-8859-1" is used.
331 *
332 * @param entity the entity to convert to a string; must not be null
333 * @return String containing the content.
334 * @throws ParseException if header elements cannot be parsed
335 * @throws IllegalArgumentException if entity is null or if content length > Integer.MAX_VALUE
336 * @throws IOException if an error occurs reading the input stream
337 * @throws java.nio.charset.UnsupportedCharsetException Thrown when the named charset is not available in
338 * this instance of the Java virtual machine
339 */
340 public static String toString(final HttpEntity entity) throws IOException, ParseException {
341 return toString(entity, DEFAULT_ENTITY_RETURN_MAX_LENGTH);
342 }
343
344 /**
345 * Reads the contents of an entity and return it as a String.
346 * The content is converted using the character set from the entity (if any),
347 * failing that, "ISO-8859-1" is used.
348 *
349 * @param entity the entity to convert to a string; must not be null
350 * @param maxResultLength
351 * The maximum size of the String to return; use it to guard against unreasonable or malicious processing.
352 * @return String containing the content.
353 * @throws ParseException if header elements cannot be parsed
354 * @throws IllegalArgumentException if entity is null or if content length > Integer.MAX_VALUE
355 * @throws IOException if an error occurs reading the input stream
356 * @throws java.nio.charset.UnsupportedCharsetException Thrown when the named charset is not available in
357 * this instance of the Java virtual machine
358 */
359 public static String toString(final HttpEntity entity, final int maxResultLength) throws IOException, ParseException {
360 Args.notNull(entity, "HttpEntity");
361 return toString(entity, ContentType.parse(entity.getContentType()), maxResultLength);
362 }
363
364 /**
365 * Returns a list of {@link NameValuePair NameValuePairs} as parsed from an {@link HttpEntity}.
366 * The encoding is taken from the entity's Content-Encoding header.
367 * <p>
368 * This is typically used while parsing an HTTP POST.
369 * </p>
370 *
371 * @param entity
372 * The entity to parse
373 * @return a list of {@link NameValuePair} as built from the URI's query portion.
374 * @throws IOException
375 * If there was an exception getting the entity's data.
376 */
377 public static List<NameValuePair> parse(final HttpEntity entity) throws IOException {
378 return parse(entity, DEFAULT_ENTITY_RETURN_MAX_LENGTH);
379 }
380
381 /**
382 * Returns a list of {@link NameValuePair NameValuePairs} as parsed from an {@link HttpEntity}.
383 * The encoding is taken from the entity's Content-Encoding header.
384 * <p>
385 * This is typically used while parsing an HTTP POST.
386 * </p>
387 *
388 * @param entity
389 * The entity to parse
390 * @param maxStreamLength
391 * The maximum size of the stream to read; use it to guard against unreasonable or malicious processing.
392 * @return a list of {@link NameValuePair} as built from the URI's query portion.
393 * @throws IOException
394 * If there was an exception getting the entity's data.
395 */
396 public static List<NameValuePair> parse(final HttpEntity entity, final int maxStreamLength) throws IOException {
397 Args.notNull(entity, "HttpEntity");
398 final int contentLength = toContentLength((int) Args.checkContentLength(entity));
399 final ContentType contentType = ContentType.parse(entity.getContentType());
400 if (!ContentType.APPLICATION_FORM_URLENCODED.isSameMimeType(contentType)) {
401 return Collections.emptyList();
402 }
403 final Charset charset = contentType.getCharset() != null ? contentType.getCharset()
404 : DEFAULT_CHARSET;
405 final CharArrayBuffer buf;
406 try (final InputStream inStream = entity.getContent()) {
407 if (inStream == null) {
408 return Collections.emptyList();
409 }
410 buf = toCharArrayBuffer(inStream, contentLength, charset, maxStreamLength);
411
412 }
413 if (buf.isEmpty()) {
414 return Collections.emptyList();
415 }
416 return WWWFormCodec.parse(buf, charset);
417 }
418
419 }