Mercurial > hg > orthanc
annotate OrthancFramework/Sources/Compression/GzipCompressor.cpp @ 4492:0b2484663233
Fix build on big-endian architectures
| author | Sebastien Jodogne <s.jodogne@gmail.com> |
|---|---|
| date | Sat, 30 Jan 2021 18:19:11 +0100 |
| parents | d9473bd5ed43 |
| children | 7053502fbf97 |
| rev | line source |
|---|---|
| 1513 | 1 /** |
| 2 * Orthanc - A Lightweight, RESTful DICOM Store | |
| 1900 | 3 * Copyright (C) 2012-2016 Sebastien Jodogne, Medical Physics |
| 1513 | 4 * Department, University Hospital of Liege, Belgium |
|
4437
d9473bd5ed43
upgrade to year 2021
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4297
diff
changeset
|
5 * Copyright (C) 2017-2021 Osimis S.A., Belgium |
| 1513 | 6 * |
| 7 * This program is free software: you can redistribute it and/or | |
|
4119
bf7b9edf6b81
re-licensing the OrthancFramework to LGPL, in order to license Stone of Orthanc under LGPL
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4044
diff
changeset
|
8 * modify it under the terms of the GNU Lesser General Public License |
|
bf7b9edf6b81
re-licensing the OrthancFramework to LGPL, in order to license Stone of Orthanc under LGPL
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4044
diff
changeset
|
9 * as published by the Free Software Foundation, either version 3 of |
|
bf7b9edf6b81
re-licensing the OrthancFramework to LGPL, in order to license Stone of Orthanc under LGPL
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4044
diff
changeset
|
10 * the License, or (at your option) any later version. |
| 1513 | 11 * |
| 12 * This program is distributed in the hope that it will be useful, but | |
| 13 * WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 14 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
|
4119
bf7b9edf6b81
re-licensing the OrthancFramework to LGPL, in order to license Stone of Orthanc under LGPL
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4044
diff
changeset
|
15 * Lesser General Public License for more details. |
| 1513 | 16 * |
|
4119
bf7b9edf6b81
re-licensing the OrthancFramework to LGPL, in order to license Stone of Orthanc under LGPL
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4044
diff
changeset
|
17 * You should have received a copy of the GNU Lesser General Public |
|
bf7b9edf6b81
re-licensing the OrthancFramework to LGPL, in order to license Stone of Orthanc under LGPL
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4044
diff
changeset
|
18 * License along with this program. If not, see |
|
bf7b9edf6b81
re-licensing the OrthancFramework to LGPL, in order to license Stone of Orthanc under LGPL
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4044
diff
changeset
|
19 * <http://www.gnu.org/licenses/>. |
| 1513 | 20 **/ |
| 21 | |
| 22 | |
| 23 #include "../PrecompiledHeaders.h" | |
| 24 #include "GzipCompressor.h" | |
| 25 | |
| 26 #include <stdio.h> | |
| 27 #include <string.h> | |
| 28 #include <zlib.h> | |
| 29 | |
| 30 #include "../OrthancException.h" | |
| 31 #include "../Logging.h" | |
| 32 | |
| 33 namespace Orthanc | |
| 34 { | |
| 35 uint64_t GzipCompressor::GuessUncompressedSize(const void* compressed, | |
| 36 size_t compressedSize) | |
| 37 { | |
| 38 /** | |
| 39 * "Is there a way to find out the size of the original file which | |
| 40 * is inside a GZIP file? [...] There is no truly reliable way, | |
| 41 * other than gunzipping the stream. You do not need to save the | |
| 42 * result of the decompression, so you can determine the size by | |
| 43 * simply reading and decoding the entire file without taking up | |
| 44 * space with the decompressed result. | |
| 45 * | |
| 46 * There is an unreliable way to determine the uncompressed size, | |
| 47 * which is to look at the last four bytes of the gzip file, which | |
| 48 * is the uncompressed length of that entry modulo 232 in little | |
| 49 * endian order. | |
| 50 * | |
| 51 * It is unreliable because a) the uncompressed data may be longer | |
| 52 * than 2^32 bytes, and b) the gzip file may consist of multiple | |
| 53 * gzip streams, in which case you would find the length of only | |
| 54 * the last of those streams. | |
| 55 * | |
| 56 * If you are in control of the source of the gzip files, you know | |
| 57 * that they consist of single gzip streams, and you know that | |
| 58 * they are less than 2^32 bytes uncompressed, then and only then | |
| 59 * can you use those last four bytes with confidence." | |
| 60 * | |
| 61 * http://stackoverflow.com/a/9727599/881731 | |
| 62 **/ | |
| 63 | |
| 64 if (compressedSize < 4) | |
| 65 { | |
| 66 throw OrthancException(ErrorCode_BadFileFormat); | |
| 67 } | |
| 68 | |
| 69 const uint8_t* p = reinterpret_cast<const uint8_t*>(compressed) + compressedSize - 4; | |
| 70 | |
| 71 return ((static_cast<uint32_t>(p[0]) << 0) + | |
| 72 (static_cast<uint32_t>(p[1]) << 8) + | |
| 73 (static_cast<uint32_t>(p[2]) << 16) + | |
| 74 (static_cast<uint32_t>(p[3]) << 24)); | |
| 75 } | |
| 76 | |
| 77 | |
| 4297 | 78 GzipCompressor::GzipCompressor() |
|
4296
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
79 { |
|
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
80 SetPrefixWithUncompressedSize(false); |
|
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
81 } |
|
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
82 |
| 1513 | 83 |
| 84 void GzipCompressor::Compress(std::string& compressed, | |
| 85 const void* uncompressed, | |
| 86 size_t uncompressedSize) | |
| 87 { | |
|
3378
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
88 uLongf compressedSize = compressBound(static_cast<uLong>(uncompressedSize)) |
|
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
89 + 1024 /* security margin */; |
|
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
90 |
| 1513 | 91 if (compressedSize == 0) |
| 92 { | |
| 93 compressedSize = 1; | |
| 94 } | |
| 95 | |
| 96 uint8_t* target; | |
| 97 if (HasPrefixWithUncompressedSize()) | |
| 98 { | |
| 99 compressed.resize(compressedSize + sizeof(uint64_t)); | |
| 100 target = reinterpret_cast<uint8_t*>(&compressed[0]) + sizeof(uint64_t); | |
| 101 } | |
| 102 else | |
| 103 { | |
| 104 compressed.resize(compressedSize); | |
| 105 target = reinterpret_cast<uint8_t*>(&compressed[0]); | |
| 106 } | |
| 107 | |
| 108 z_stream stream; | |
| 109 memset(&stream, 0, sizeof(stream)); | |
| 110 | |
| 111 stream.next_in = const_cast<Bytef*>(reinterpret_cast<const Bytef*>(uncompressed)); | |
| 112 stream.next_out = reinterpret_cast<Bytef*>(target); | |
| 113 | |
| 114 stream.avail_in = static_cast<uInt>(uncompressedSize); | |
| 115 stream.avail_out = static_cast<uInt>(compressedSize); | |
| 116 | |
| 117 // Ensure no overflow (if the buffer is too large for the current archicture) | |
| 118 if (static_cast<size_t>(stream.avail_in) != uncompressedSize || | |
| 119 static_cast<size_t>(stream.avail_out) != compressedSize) | |
| 120 { | |
| 121 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 122 } | |
| 123 | |
| 124 // Initialize the compression engine | |
| 125 int error = deflateInit2(&stream, | |
| 126 GetCompressionLevel(), | |
| 127 Z_DEFLATED, | |
| 128 MAX_WBITS + 16, // ask for gzip output | |
| 129 8, // default memory level | |
| 130 Z_DEFAULT_STRATEGY); | |
| 131 | |
| 132 if (error != Z_OK) | |
| 133 { | |
| 134 // Cannot initialize zlib | |
| 135 compressed.clear(); | |
| 136 throw OrthancException(ErrorCode_InternalError); | |
| 137 } | |
| 138 | |
| 139 // Compress the input buffer | |
| 140 error = deflate(&stream, Z_FINISH); | |
| 141 | |
| 142 if (error != Z_STREAM_END) | |
| 143 { | |
| 144 deflateEnd(&stream); | |
| 145 compressed.clear(); | |
| 146 | |
| 147 switch (error) | |
| 148 { | |
| 149 case Z_MEM_ERROR: | |
| 150 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 151 | |
| 152 default: | |
| 153 throw OrthancException(ErrorCode_InternalError); | |
| 154 } | |
| 155 } | |
| 156 | |
| 157 size_t size = stream.total_out; | |
| 158 | |
| 159 if (deflateEnd(&stream) != Z_OK) | |
| 160 { | |
| 161 throw OrthancException(ErrorCode_InternalError); | |
| 162 } | |
| 163 | |
| 164 // The compression was successful | |
| 165 if (HasPrefixWithUncompressedSize()) | |
| 166 { | |
| 167 uint64_t s = static_cast<uint64_t>(uncompressedSize); | |
| 168 memcpy(&compressed[0], &s, sizeof(uint64_t)); | |
| 169 compressed.resize(size + sizeof(uint64_t)); | |
| 170 } | |
| 171 else | |
| 172 { | |
| 173 compressed.resize(size); | |
| 174 } | |
| 175 } | |
| 176 | |
| 177 | |
| 178 void GzipCompressor::Uncompress(std::string& uncompressed, | |
| 179 const void* compressed, | |
| 180 size_t compressedSize) | |
| 181 { | |
| 182 uint64_t uncompressedSize; | |
| 183 const uint8_t* source = reinterpret_cast<const uint8_t*>(compressed); | |
| 184 | |
| 185 if (HasPrefixWithUncompressedSize()) | |
| 186 { | |
| 187 uncompressedSize = ReadUncompressedSizePrefix(compressed, compressedSize); | |
| 188 source += sizeof(uint64_t); | |
| 189 compressedSize -= sizeof(uint64_t); | |
| 190 } | |
| 191 else | |
| 192 { | |
| 193 uncompressedSize = GuessUncompressedSize(compressed, compressedSize); | |
| 194 } | |
| 195 | |
| 196 try | |
| 197 { | |
| 1545 | 198 uncompressed.resize(static_cast<size_t>(uncompressedSize)); |
| 1513 | 199 } |
| 200 catch (...) | |
| 201 { | |
| 202 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 203 } | |
| 204 | |
| 205 z_stream stream; | |
| 206 memset(&stream, 0, sizeof(stream)); | |
| 207 | |
| 208 char dummy = '\0'; // zlib does not like NULL output buffers (even if the uncompressed data is empty) | |
| 209 stream.next_in = const_cast<Bytef*>(source); | |
| 210 stream.next_out = reinterpret_cast<Bytef*>(uncompressedSize == 0 ? &dummy : &uncompressed[0]); | |
| 211 | |
| 212 stream.avail_in = static_cast<uInt>(compressedSize); | |
| 213 stream.avail_out = static_cast<uInt>(uncompressedSize); | |
| 214 | |
| 215 // Ensure no overflow (if the buffer is too large for the current archicture) | |
| 216 if (static_cast<size_t>(stream.avail_in) != compressedSize || | |
| 217 static_cast<size_t>(stream.avail_out) != uncompressedSize) | |
| 218 { | |
| 219 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 220 } | |
| 221 | |
| 222 // Initialize the compression engine | |
| 223 int error = inflateInit2(&stream, | |
| 224 MAX_WBITS + 16); // this is a gzip input | |
| 225 | |
| 226 if (error != Z_OK) | |
| 227 { | |
| 228 // Cannot initialize zlib | |
| 229 uncompressed.clear(); | |
| 230 throw OrthancException(ErrorCode_InternalError); | |
| 231 } | |
| 232 | |
| 233 // Uncompress the input buffer | |
| 234 error = inflate(&stream, Z_FINISH); | |
| 235 | |
| 236 if (error != Z_STREAM_END) | |
| 237 { | |
| 238 inflateEnd(&stream); | |
| 239 uncompressed.clear(); | |
| 240 | |
| 241 switch (error) | |
| 242 { | |
| 243 case Z_MEM_ERROR: | |
| 244 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 245 | |
| 246 case Z_BUF_ERROR: | |
| 247 case Z_NEED_DICT: | |
| 248 throw OrthancException(ErrorCode_BadFileFormat); | |
| 249 | |
| 250 default: | |
| 251 throw OrthancException(ErrorCode_InternalError); | |
| 252 } | |
| 253 } | |
| 254 | |
| 255 size_t size = stream.total_out; | |
| 256 | |
| 257 if (inflateEnd(&stream) != Z_OK) | |
| 258 { | |
| 259 uncompressed.clear(); | |
| 260 throw OrthancException(ErrorCode_InternalError); | |
| 261 } | |
| 262 | |
| 263 if (size != uncompressedSize) | |
| 264 { | |
| 265 uncompressed.clear(); | |
| 266 | |
| 267 // The uncompressed size was not that properly guess, presumably | |
| 268 // because of a file size over 4GB. Should fallback to | |
| 269 // stream-based decompression. | |
|
2954
d924f9bb61cc
taking advantage of details in OrthancException
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
2447
diff
changeset
|
270 throw OrthancException(ErrorCode_NotImplemented, |
|
d924f9bb61cc
taking advantage of details in OrthancException
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
2447
diff
changeset
|
271 "The uncompressed size of a gzip-encoded buffer was not properly guessed"); |
| 1513 | 272 } |
| 273 } | |
| 274 } |