Mercurial > hg > orthanc

comparison Core/Compression/GzipCompressor.cpp @ 1513:fe07f82d83d3

Find changesets by keywords (author, files, the commit message), revision number or hash, or revset expression.
gzip compression
author Sebastien Jodogne <s.jodogne@gmail.com>
date Mon, 10 Aug 2015 16:01:37 +0200
parents
children 33d34bc4ac15
comparison
equal deleted inserted replaced
1512:52dc56bcec7d 1513:fe07f82d83d3
1 /**
2 * Orthanc - A Lightweight, RESTful DICOM Store
3 * Copyright (C) 2012-2015 Sebastien Jodogne, Medical Physics
4 * Department, University Hospital of Liege, Belgium
5 *
6 * This program is free software: you can redistribute it and/or
7 * modify it under the terms of the GNU General Public License as
8 * published by the Free Software Foundation, either version 3 of the
9 * License, or (at your option) any later version.
10 *
11 * In addition, as a special exception, the copyright holders of this
12 * program give permission to link the code of its release with the
13 * OpenSSL project's "OpenSSL" library (or with modified versions of it
14 * that use the same license as the "OpenSSL" library), and distribute
15 * the linked executables. You must obey the GNU General Public License
16 * in all respects for all of the code used other than "OpenSSL". If you
17 * modify file(s) with this exception, you may extend this exception to
18 * your version of the file(s), but you are not obligated to do so. If
19 * you do not wish to do so, delete this exception statement from your
20 * version. If you delete this exception statement from all source files
21 * in the program, then also delete it here.
22 *
23 * This program is distributed in the hope that it will be useful, but
24 * WITHOUT ANY WARRANTY; without even the implied warranty of
25 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
26 * General Public License for more details.
27 *
28 * You should have received a copy of the GNU General Public License
29 * along with this program. If not, see <http://www.gnu.org/licenses/>.
30 **/
31
32
33 #include "../PrecompiledHeaders.h"
34 #include "GzipCompressor.h"
35
36 #include <stdio.h>
37 #include <string.h>
38 #include <zlib.h>
39
40 #include "../OrthancException.h"
41 #include "../Logging.h"
42
43 namespace Orthanc
44 {
45 uint64_t GzipCompressor::GuessUncompressedSize(const void* compressed,
46 size_t compressedSize)
47 {
48 /**
49 * "Is there a way to find out the size of the original file which
50 * is inside a GZIP file? [...] There is no truly reliable way,
51 * other than gunzipping the stream. You do not need to save the
52 * result of the decompression, so you can determine the size by
53 * simply reading and decoding the entire file without taking up
54 * space with the decompressed result.
55 *
56 * There is an unreliable way to determine the uncompressed size,
57 * which is to look at the last four bytes of the gzip file, which
58 * is the uncompressed length of that entry modulo 232 in little
59 * endian order.
60 *
61 * It is unreliable because a) the uncompressed data may be longer
62 * than 2^32 bytes, and b) the gzip file may consist of multiple
63 * gzip streams, in which case you would find the length of only
64 * the last of those streams.
65 *
66 * If you are in control of the source of the gzip files, you know
67 * that they consist of single gzip streams, and you know that
68 * they are less than 2^32 bytes uncompressed, then and only then
69 * can you use those last four bytes with confidence."
70 *
71 * http://stackoverflow.com/a/9727599/881731
72 **/
73
74 if (compressedSize < 4)
75 {
76 throw OrthancException(ErrorCode_BadFileFormat);
77 }
78
79 const uint8_t* p = reinterpret_cast<const uint8_t*>(compressed) + compressedSize - 4;
80
81 return ((static_cast<uint32_t>(p[0]) << 0) +
82 (static_cast<uint32_t>(p[1]) << 8) +
83 (static_cast<uint32_t>(p[2]) << 16) +
84 (static_cast<uint32_t>(p[3]) << 24));
85 }
86
87
88
89 void GzipCompressor::Compress(std::string& compressed,
90 const void* uncompressed,
91 size_t uncompressedSize)
92 {
93 uLongf compressedSize = compressBound(uncompressedSize) + 1024 /* security margin */;
94 if (compressedSize == 0)
95 {
96 compressedSize = 1;
97 }
98
99 uint8_t* target;
100 if (HasPrefixWithUncompressedSize())
101 {
102 compressed.resize(compressedSize + sizeof(uint64_t));
103 target = reinterpret_cast<uint8_t*>(&compressed[0]) + sizeof(uint64_t);
104 }
105 else
106 {
107 compressed.resize(compressedSize);
108 target = reinterpret_cast<uint8_t*>(&compressed[0]);
109 }
110
111 z_stream stream;
112 memset(&stream, 0, sizeof(stream));
113
114 stream.next_in = const_cast<Bytef*>(reinterpret_cast<const Bytef*>(uncompressed));
115 stream.next_out = reinterpret_cast<Bytef*>(target);
116
117 stream.avail_in = static_cast<uInt>(uncompressedSize);
118 stream.avail_out = static_cast<uInt>(compressedSize);
119
120 // Ensure no overflow (if the buffer is too large for the current archicture)
121 if (static_cast<size_t>(stream.avail_in) != uncompressedSize ||
122 static_cast<size_t>(stream.avail_out) != compressedSize)
123 {
124 throw OrthancException(ErrorCode_NotEnoughMemory);
125 }
126
127 // Initialize the compression engine
128 int error = deflateInit2(&stream,
129 GetCompressionLevel(),
130 Z_DEFLATED,
131 MAX_WBITS + 16, // ask for gzip output
132 8, // default memory level
133 Z_DEFAULT_STRATEGY);
134
135 if (error != Z_OK)
136 {
137 // Cannot initialize zlib
138 compressed.clear();
139 throw OrthancException(ErrorCode_InternalError);
140 }
141
142 // Compress the input buffer
143 error = deflate(&stream, Z_FINISH);
144
145 if (error != Z_STREAM_END)
146 {
147 deflateEnd(&stream);
148 compressed.clear();
149
150 switch (error)
151 {
152 case Z_MEM_ERROR:
153 throw OrthancException(ErrorCode_NotEnoughMemory);
154
155 default:
156 throw OrthancException(ErrorCode_InternalError);
157 }
158 }
159
160 size_t size = stream.total_out;
161
162 if (deflateEnd(&stream) != Z_OK)
163 {
164 throw OrthancException(ErrorCode_InternalError);
165 }
166
167 // The compression was successful
168 if (HasPrefixWithUncompressedSize())
169 {
170 uint64_t s = static_cast<uint64_t>(uncompressedSize);
171 memcpy(&compressed[0], &s, sizeof(uint64_t));
172 compressed.resize(size + sizeof(uint64_t));
173 }
174 else
175 {
176 compressed.resize(size);
177 }
178 }
179
180
181 void GzipCompressor::Uncompress(std::string& uncompressed,
182 const void* compressed,
183 size_t compressedSize)
184 {
185 uint64_t uncompressedSize;
186 const uint8_t* source = reinterpret_cast<const uint8_t*>(compressed);
187
188 if (HasPrefixWithUncompressedSize())
189 {
190 uncompressedSize = ReadUncompressedSizePrefix(compressed, compressedSize);
191 source += sizeof(uint64_t);
192 compressedSize -= sizeof(uint64_t);
193 }
194 else
195 {
196 uncompressedSize = GuessUncompressedSize(compressed, compressedSize);
197 }
198
199 try
200 {
201 uncompressed.resize(uncompressedSize);
202 }
203 catch (...)
204 {
205 throw OrthancException(ErrorCode_NotEnoughMemory);
206 }
207
208 z_stream stream;
209 memset(&stream, 0, sizeof(stream));
210
211 char dummy = '\0'; // zlib does not like NULL output buffers (even if the uncompressed data is empty)
212 stream.next_in = const_cast<Bytef*>(source);
213 stream.next_out = reinterpret_cast<Bytef*>(uncompressedSize == 0 ? &dummy : &uncompressed[0]);
214
215 stream.avail_in = static_cast<uInt>(compressedSize);
216 stream.avail_out = static_cast<uInt>(uncompressedSize);
217
218 // Ensure no overflow (if the buffer is too large for the current archicture)
219 if (static_cast<size_t>(stream.avail_in) != compressedSize ||
220 static_cast<size_t>(stream.avail_out) != uncompressedSize)
221 {
222 throw OrthancException(ErrorCode_NotEnoughMemory);
223 }
224
225 // Initialize the compression engine
226 int error = inflateInit2(&stream,
227 MAX_WBITS + 16); // this is a gzip input
228
229 if (error != Z_OK)
230 {
231 // Cannot initialize zlib
232 uncompressed.clear();
233 throw OrthancException(ErrorCode_InternalError);
234 }
235
236 // Uncompress the input buffer
237 error = inflate(&stream, Z_FINISH);
238
239 if (error != Z_STREAM_END)
240 {
241 inflateEnd(&stream);
242 uncompressed.clear();
243
244 switch (error)
245 {
246 case Z_MEM_ERROR:
247 throw OrthancException(ErrorCode_NotEnoughMemory);
248
249 case Z_BUF_ERROR:
250 case Z_NEED_DICT:
251 throw OrthancException(ErrorCode_BadFileFormat);
252
253 default:
254 throw OrthancException(ErrorCode_InternalError);
255 }
256 }
257
258 size_t size = stream.total_out;
259
260 if (inflateEnd(&stream) != Z_OK)
261 {
262 uncompressed.clear();
263 throw OrthancException(ErrorCode_InternalError);
264 }
265
266 if (size != uncompressedSize)
267 {
268 uncompressed.clear();
269
270 // The uncompressed size was not that properly guess, presumably
271 // because of a file size over 4GB. Should fallback to
272 // stream-based decompression.
273 LOG(ERROR) << "The uncompressed size of a gzip-encoded buffer was not properly guessed";
274 throw OrthancException(ErrorCode_NotImplemented);
275 }
276 }
277 }