Mercurial > hg > orthanc
annotate OrthancFramework/Sources/Compression/GzipCompressor.cpp @ 5640:f7adfb22e20e
updated copyright, as Orthanc Team now replaces Osimis
| author | Sebastien Jodogne <s.jodogne@gmail.com> |
|---|---|
| date | Thu, 30 May 2024 21:19:57 +0200 |
| parents | 48b8dae6dc77 |
| children |
| 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 |
|
5640
f7adfb22e20e
updated copyright, as Orthanc Team now replaces Osimis
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
5485
diff
changeset
|
5 * Copyright (C) 2017-2023 Osimis S.A., Belgium |
|
f7adfb22e20e
updated copyright, as Orthanc Team now replaces Osimis
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
5485
diff
changeset
|
6 * Copyright (C) 2024-2024 Orthanc Team SRL, Belgium |
|
5485
48b8dae6dc77
upgrade to year 2024
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
5185
diff
changeset
|
7 * Copyright (C) 2021-2024 Sebastien Jodogne, ICTEAM UCLouvain, Belgium |
| 1513 | 8 * |
| 9 * 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
|
10 * 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
|
11 * 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
|
12 * the License, or (at your option) any later version. |
| 1513 | 13 * |
| 14 * This program is distributed in the hope that it will be useful, but | |
| 15 * WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 16 * 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
|
17 * Lesser General Public License for more details. |
| 1513 | 18 * |
|
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
|
19 * 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
|
20 * 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
|
21 * <http://www.gnu.org/licenses/>. |
| 1513 | 22 **/ |
| 23 | |
| 24 | |
| 25 #include "../PrecompiledHeaders.h" | |
| 26 #include "GzipCompressor.h" | |
| 27 | |
| 28 #include <stdio.h> | |
| 29 #include <string.h> | |
| 30 #include <zlib.h> | |
| 31 | |
| 32 #include "../OrthancException.h" | |
| 33 #include "../Logging.h" | |
| 34 | |
| 35 namespace Orthanc | |
| 36 { | |
| 37 uint64_t GzipCompressor::GuessUncompressedSize(const void* compressed, | |
| 38 size_t compressedSize) | |
| 39 { | |
| 40 /** | |
| 41 * "Is there a way to find out the size of the original file which | |
| 42 * is inside a GZIP file? [...] There is no truly reliable way, | |
| 43 * other than gunzipping the stream. You do not need to save the | |
| 44 * result of the decompression, so you can determine the size by | |
| 45 * simply reading and decoding the entire file without taking up | |
| 46 * space with the decompressed result. | |
| 47 * | |
| 48 * There is an unreliable way to determine the uncompressed size, | |
| 49 * which is to look at the last four bytes of the gzip file, which | |
| 50 * is the uncompressed length of that entry modulo 232 in little | |
| 51 * endian order. | |
| 52 * | |
| 53 * It is unreliable because a) the uncompressed data may be longer | |
| 54 * than 2^32 bytes, and b) the gzip file may consist of multiple | |
| 55 * gzip streams, in which case you would find the length of only | |
| 56 * the last of those streams. | |
| 57 * | |
| 58 * If you are in control of the source of the gzip files, you know | |
| 59 * that they consist of single gzip streams, and you know that | |
| 60 * they are less than 2^32 bytes uncompressed, then and only then | |
| 61 * can you use those last four bytes with confidence." | |
| 62 * | |
| 63 * http://stackoverflow.com/a/9727599/881731 | |
| 64 **/ | |
| 65 | |
| 66 if (compressedSize < 4) | |
| 67 { | |
| 68 throw OrthancException(ErrorCode_BadFileFormat); | |
| 69 } | |
| 70 | |
| 71 const uint8_t* p = reinterpret_cast<const uint8_t*>(compressed) + compressedSize - 4; | |
| 72 | |
| 73 return ((static_cast<uint32_t>(p[0]) << 0) + | |
| 74 (static_cast<uint32_t>(p[1]) << 8) + | |
| 75 (static_cast<uint32_t>(p[2]) << 16) + | |
| 76 (static_cast<uint32_t>(p[3]) << 24)); | |
| 77 } | |
| 78 | |
| 79 | |
| 4297 | 80 GzipCompressor::GzipCompressor() |
|
4296
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 SetPrefixWithUncompressedSize(false); |
|
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
83 } |
|
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
84 |
| 1513 | 85 |
| 86 void GzipCompressor::Compress(std::string& compressed, | |
| 87 const void* uncompressed, | |
| 88 size_t uncompressedSize) | |
| 89 { | |
|
3378
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
90 uLongf compressedSize = compressBound(static_cast<uLong>(uncompressedSize)) |
|
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
91 + 1024 /* security margin */; |
|
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
92 |
| 1513 | 93 if (compressedSize == 0) |
| 94 { | |
| 95 compressedSize = 1; | |
| 96 } | |
| 97 | |
| 98 uint8_t* target; | |
| 99 if (HasPrefixWithUncompressedSize()) | |
| 100 { | |
| 101 compressed.resize(compressedSize + sizeof(uint64_t)); | |
| 102 target = reinterpret_cast<uint8_t*>(&compressed[0]) + sizeof(uint64_t); | |
| 103 } | |
| 104 else | |
| 105 { | |
| 106 compressed.resize(compressedSize); | |
| 107 target = reinterpret_cast<uint8_t*>(&compressed[0]); | |
| 108 } | |
| 109 | |
| 110 z_stream stream; | |
| 111 memset(&stream, 0, sizeof(stream)); | |
| 112 | |
| 113 stream.next_in = const_cast<Bytef*>(reinterpret_cast<const Bytef*>(uncompressed)); | |
| 114 stream.next_out = reinterpret_cast<Bytef*>(target); | |
| 115 | |
| 116 stream.avail_in = static_cast<uInt>(uncompressedSize); | |
| 117 stream.avail_out = static_cast<uInt>(compressedSize); | |
| 118 | |
| 119 // Ensure no overflow (if the buffer is too large for the current archicture) | |
| 120 if (static_cast<size_t>(stream.avail_in) != uncompressedSize || | |
| 121 static_cast<size_t>(stream.avail_out) != compressedSize) | |
| 122 { | |
| 123 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 124 } | |
| 125 | |
| 126 // Initialize the compression engine | |
| 127 int error = deflateInit2(&stream, | |
| 128 GetCompressionLevel(), | |
| 129 Z_DEFLATED, | |
| 130 MAX_WBITS + 16, // ask for gzip output | |
| 131 8, // default memory level | |
| 132 Z_DEFAULT_STRATEGY); | |
| 133 | |
| 134 if (error != Z_OK) | |
| 135 { | |
| 136 // Cannot initialize zlib | |
| 137 compressed.clear(); | |
| 138 throw OrthancException(ErrorCode_InternalError); | |
| 139 } | |
| 140 | |
| 141 // Compress the input buffer | |
| 142 error = deflate(&stream, Z_FINISH); | |
| 143 | |
| 144 if (error != Z_STREAM_END) | |
| 145 { | |
| 146 deflateEnd(&stream); | |
| 147 compressed.clear(); | |
| 148 | |
| 149 switch (error) | |
| 150 { | |
| 151 case Z_MEM_ERROR: | |
| 152 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 153 | |
| 154 default: | |
| 155 throw OrthancException(ErrorCode_InternalError); | |
| 156 } | |
| 157 } | |
| 158 | |
| 159 size_t size = stream.total_out; | |
| 160 | |
| 161 if (deflateEnd(&stream) != Z_OK) | |
| 162 { | |
| 163 throw OrthancException(ErrorCode_InternalError); | |
| 164 } | |
| 165 | |
| 166 // The compression was successful | |
| 167 if (HasPrefixWithUncompressedSize()) | |
| 168 { | |
| 169 uint64_t s = static_cast<uint64_t>(uncompressedSize); | |
| 170 memcpy(&compressed[0], &s, sizeof(uint64_t)); | |
| 171 compressed.resize(size + sizeof(uint64_t)); | |
| 172 } | |
| 173 else | |
| 174 { | |
| 175 compressed.resize(size); | |
| 176 } | |
| 177 } | |
| 178 | |
| 179 | |
| 180 void GzipCompressor::Uncompress(std::string& uncompressed, | |
| 181 const void* compressed, | |
| 182 size_t compressedSize) | |
| 183 { | |
| 184 uint64_t uncompressedSize; | |
| 185 const uint8_t* source = reinterpret_cast<const uint8_t*>(compressed); | |
| 186 | |
| 187 if (HasPrefixWithUncompressedSize()) | |
| 188 { | |
| 189 uncompressedSize = ReadUncompressedSizePrefix(compressed, compressedSize); | |
| 190 source += sizeof(uint64_t); | |
| 191 compressedSize -= sizeof(uint64_t); | |
| 192 } | |
| 193 else | |
| 194 { | |
| 195 uncompressedSize = GuessUncompressedSize(compressed, compressedSize); | |
| 196 } | |
| 197 | |
| 198 try | |
| 199 { | |
| 1545 | 200 uncompressed.resize(static_cast<size_t>(uncompressedSize)); |
| 1513 | 201 } |
| 202 catch (...) | |
| 203 { | |
| 204 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 205 } | |
| 206 | |
| 207 z_stream stream; | |
| 208 memset(&stream, 0, sizeof(stream)); | |
| 209 | |
| 210 char dummy = '\0'; // zlib does not like NULL output buffers (even if the uncompressed data is empty) | |
| 211 stream.next_in = const_cast<Bytef*>(source); | |
| 212 stream.next_out = reinterpret_cast<Bytef*>(uncompressedSize == 0 ? &dummy : &uncompressed[0]); | |
| 213 | |
| 214 stream.avail_in = static_cast<uInt>(compressedSize); | |
| 215 stream.avail_out = static_cast<uInt>(uncompressedSize); | |
| 216 | |
| 217 // Ensure no overflow (if the buffer is too large for the current archicture) | |
| 218 if (static_cast<size_t>(stream.avail_in) != compressedSize || | |
| 219 static_cast<size_t>(stream.avail_out) != uncompressedSize) | |
| 220 { | |
| 221 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 222 } | |
| 223 | |
| 224 // Initialize the compression engine | |
| 225 int error = inflateInit2(&stream, | |
| 226 MAX_WBITS + 16); // this is a gzip input | |
| 227 | |
| 228 if (error != Z_OK) | |
| 229 { | |
| 230 // Cannot initialize zlib | |
| 231 uncompressed.clear(); | |
| 232 throw OrthancException(ErrorCode_InternalError); | |
| 233 } | |
| 234 | |
| 235 // Uncompress the input buffer | |
| 236 error = inflate(&stream, Z_FINISH); | |
| 237 | |
| 238 if (error != Z_STREAM_END) | |
| 239 { | |
| 240 inflateEnd(&stream); | |
| 241 uncompressed.clear(); | |
| 242 | |
| 243 switch (error) | |
| 244 { | |
| 245 case Z_MEM_ERROR: | |
| 246 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 247 | |
| 248 case Z_BUF_ERROR: | |
| 249 case Z_NEED_DICT: | |
| 250 throw OrthancException(ErrorCode_BadFileFormat); | |
| 251 | |
| 252 default: | |
| 253 throw OrthancException(ErrorCode_InternalError); | |
| 254 } | |
| 255 } | |
| 256 | |
| 257 size_t size = stream.total_out; | |
| 258 | |
| 259 if (inflateEnd(&stream) != Z_OK) | |
| 260 { | |
| 261 uncompressed.clear(); | |
| 262 throw OrthancException(ErrorCode_InternalError); | |
| 263 } | |
| 264 | |
| 265 if (size != uncompressedSize) | |
| 266 { | |
| 267 uncompressed.clear(); | |
| 268 | |
| 269 // The uncompressed size was not that properly guess, presumably | |
| 270 // because of a file size over 4GB. Should fallback to | |
| 271 // stream-based decompression. | |
|
2954
d924f9bb61cc
taking advantage of details in OrthancException
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
2447
diff
changeset
|
272 throw OrthancException(ErrorCode_NotImplemented, |
|
d924f9bb61cc
taking advantage of details in OrthancException
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
2447
diff
changeset
|
273 "The uncompressed size of a gzip-encoded buffer was not properly guessed"); |
| 1513 | 274 } |
| 275 } | |
| 276 } |