Mercurial > hg > orthanc
annotate OrthancFramework/Sources/Compression/GzipCompressor.cpp @ 4917:95378de340d4
merge
| author | Sebastien Jodogne <s.jodogne@gmail.com> |
|---|---|
| date | Wed, 23 Feb 2022 06:36:29 +0100 |
| parents | 43e613a7756b |
| children | 0ea402b4d901 |
| 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 |
|
4870
43e613a7756b
upgrade to year 2022
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4831
diff
changeset
|
5 * Copyright (C) 2017-2022 Osimis S.A., Belgium |
|
43e613a7756b
upgrade to year 2022
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4831
diff
changeset
|
6 * Copyright (C) 2021-2022 Sebastien Jodogne, ICTEAM UCLouvain, Belgium |
| 1513 | 7 * |
| 8 * 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
|
9 * 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
|
10 * 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
|
11 * the License, or (at your option) any later version. |
| 1513 | 12 * |
| 13 * This program is distributed in the hope that it will be useful, but | |
| 14 * WITHOUT ANY WARRANTY; without even the implied warranty of | |
| 15 * 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
|
16 * Lesser General Public License for more details. |
| 1513 | 17 * |
|
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
|
18 * 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
|
19 * 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
|
20 * <http://www.gnu.org/licenses/>. |
| 1513 | 21 **/ |
| 22 | |
| 23 | |
| 24 #include "../PrecompiledHeaders.h" | |
| 25 #include "GzipCompressor.h" | |
| 26 | |
| 27 #include <stdio.h> | |
| 28 #include <string.h> | |
| 29 #include <zlib.h> | |
| 30 | |
| 31 #include "../OrthancException.h" | |
| 32 #include "../Logging.h" | |
| 33 | |
| 34 namespace Orthanc | |
| 35 { | |
| 36 uint64_t GzipCompressor::GuessUncompressedSize(const void* compressed, | |
| 37 size_t compressedSize) | |
| 38 { | |
| 39 /** | |
| 40 * "Is there a way to find out the size of the original file which | |
| 41 * is inside a GZIP file? [...] There is no truly reliable way, | |
| 42 * other than gunzipping the stream. You do not need to save the | |
| 43 * result of the decompression, so you can determine the size by | |
| 44 * simply reading and decoding the entire file without taking up | |
| 45 * space with the decompressed result. | |
| 46 * | |
| 47 * There is an unreliable way to determine the uncompressed size, | |
| 48 * which is to look at the last four bytes of the gzip file, which | |
| 49 * is the uncompressed length of that entry modulo 232 in little | |
| 50 * endian order. | |
| 51 * | |
| 52 * It is unreliable because a) the uncompressed data may be longer | |
| 53 * than 2^32 bytes, and b) the gzip file may consist of multiple | |
| 54 * gzip streams, in which case you would find the length of only | |
| 55 * the last of those streams. | |
| 56 * | |
| 57 * If you are in control of the source of the gzip files, you know | |
| 58 * that they consist of single gzip streams, and you know that | |
| 59 * they are less than 2^32 bytes uncompressed, then and only then | |
| 60 * can you use those last four bytes with confidence." | |
| 61 * | |
| 62 * http://stackoverflow.com/a/9727599/881731 | |
| 63 **/ | |
| 64 | |
| 65 if (compressedSize < 4) | |
| 66 { | |
| 67 throw OrthancException(ErrorCode_BadFileFormat); | |
| 68 } | |
| 69 | |
| 70 const uint8_t* p = reinterpret_cast<const uint8_t*>(compressed) + compressedSize - 4; | |
| 71 | |
| 72 return ((static_cast<uint32_t>(p[0]) << 0) + | |
| 73 (static_cast<uint32_t>(p[1]) << 8) + | |
| 74 (static_cast<uint32_t>(p[2]) << 16) + | |
| 75 (static_cast<uint32_t>(p[3]) << 24)); | |
| 76 } | |
| 77 | |
| 78 | |
| 4297 | 79 GzipCompressor::GzipCompressor() |
|
4296
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
80 { |
|
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
81 SetPrefixWithUncompressedSize(false); |
|
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
82 } |
|
3b70a2e6a06c
moving inline methods to source files for ABI compatibility
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
4119
diff
changeset
|
83 |
| 1513 | 84 |
| 85 void GzipCompressor::Compress(std::string& compressed, | |
| 86 const void* uncompressed, | |
| 87 size_t uncompressedSize) | |
| 88 { | |
|
3378
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
89 uLongf compressedSize = compressBound(static_cast<uLong>(uncompressedSize)) |
|
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
90 + 1024 /* security margin */; |
|
596cfabd72c5
Fixed a couple of truncation warnings
Benjamin Golinvaux <bgo@osimis.io>
parents:
3060
diff
changeset
|
91 |
| 1513 | 92 if (compressedSize == 0) |
| 93 { | |
| 94 compressedSize = 1; | |
| 95 } | |
| 96 | |
| 97 uint8_t* target; | |
| 98 if (HasPrefixWithUncompressedSize()) | |
| 99 { | |
| 100 compressed.resize(compressedSize + sizeof(uint64_t)); | |
| 101 target = reinterpret_cast<uint8_t*>(&compressed[0]) + sizeof(uint64_t); | |
| 102 } | |
| 103 else | |
| 104 { | |
| 105 compressed.resize(compressedSize); | |
| 106 target = reinterpret_cast<uint8_t*>(&compressed[0]); | |
| 107 } | |
| 108 | |
| 109 z_stream stream; | |
| 110 memset(&stream, 0, sizeof(stream)); | |
| 111 | |
| 112 stream.next_in = const_cast<Bytef*>(reinterpret_cast<const Bytef*>(uncompressed)); | |
| 113 stream.next_out = reinterpret_cast<Bytef*>(target); | |
| 114 | |
| 115 stream.avail_in = static_cast<uInt>(uncompressedSize); | |
| 116 stream.avail_out = static_cast<uInt>(compressedSize); | |
| 117 | |
| 118 // Ensure no overflow (if the buffer is too large for the current archicture) | |
| 119 if (static_cast<size_t>(stream.avail_in) != uncompressedSize || | |
| 120 static_cast<size_t>(stream.avail_out) != compressedSize) | |
| 121 { | |
| 122 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 123 } | |
| 124 | |
| 125 // Initialize the compression engine | |
| 126 int error = deflateInit2(&stream, | |
| 127 GetCompressionLevel(), | |
| 128 Z_DEFLATED, | |
| 129 MAX_WBITS + 16, // ask for gzip output | |
| 130 8, // default memory level | |
| 131 Z_DEFAULT_STRATEGY); | |
| 132 | |
| 133 if (error != Z_OK) | |
| 134 { | |
| 135 // Cannot initialize zlib | |
| 136 compressed.clear(); | |
| 137 throw OrthancException(ErrorCode_InternalError); | |
| 138 } | |
| 139 | |
| 140 // Compress the input buffer | |
| 141 error = deflate(&stream, Z_FINISH); | |
| 142 | |
| 143 if (error != Z_STREAM_END) | |
| 144 { | |
| 145 deflateEnd(&stream); | |
| 146 compressed.clear(); | |
| 147 | |
| 148 switch (error) | |
| 149 { | |
| 150 case Z_MEM_ERROR: | |
| 151 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 152 | |
| 153 default: | |
| 154 throw OrthancException(ErrorCode_InternalError); | |
| 155 } | |
| 156 } | |
| 157 | |
| 158 size_t size = stream.total_out; | |
| 159 | |
| 160 if (deflateEnd(&stream) != Z_OK) | |
| 161 { | |
| 162 throw OrthancException(ErrorCode_InternalError); | |
| 163 } | |
| 164 | |
| 165 // The compression was successful | |
| 166 if (HasPrefixWithUncompressedSize()) | |
| 167 { | |
| 168 uint64_t s = static_cast<uint64_t>(uncompressedSize); | |
| 169 memcpy(&compressed[0], &s, sizeof(uint64_t)); | |
| 170 compressed.resize(size + sizeof(uint64_t)); | |
| 171 } | |
| 172 else | |
| 173 { | |
| 174 compressed.resize(size); | |
| 175 } | |
| 176 } | |
| 177 | |
| 178 | |
| 179 void GzipCompressor::Uncompress(std::string& uncompressed, | |
| 180 const void* compressed, | |
| 181 size_t compressedSize) | |
| 182 { | |
| 183 uint64_t uncompressedSize; | |
| 184 const uint8_t* source = reinterpret_cast<const uint8_t*>(compressed); | |
| 185 | |
| 186 if (HasPrefixWithUncompressedSize()) | |
| 187 { | |
| 188 uncompressedSize = ReadUncompressedSizePrefix(compressed, compressedSize); | |
| 189 source += sizeof(uint64_t); | |
| 190 compressedSize -= sizeof(uint64_t); | |
| 191 } | |
| 192 else | |
| 193 { | |
| 194 uncompressedSize = GuessUncompressedSize(compressed, compressedSize); | |
| 195 } | |
| 196 | |
| 197 try | |
| 198 { | |
| 1545 | 199 uncompressed.resize(static_cast<size_t>(uncompressedSize)); |
| 1513 | 200 } |
| 201 catch (...) | |
| 202 { | |
| 203 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 204 } | |
| 205 | |
| 206 z_stream stream; | |
| 207 memset(&stream, 0, sizeof(stream)); | |
| 208 | |
| 209 char dummy = '\0'; // zlib does not like NULL output buffers (even if the uncompressed data is empty) | |
| 210 stream.next_in = const_cast<Bytef*>(source); | |
| 211 stream.next_out = reinterpret_cast<Bytef*>(uncompressedSize == 0 ? &dummy : &uncompressed[0]); | |
| 212 | |
| 213 stream.avail_in = static_cast<uInt>(compressedSize); | |
| 214 stream.avail_out = static_cast<uInt>(uncompressedSize); | |
| 215 | |
| 216 // Ensure no overflow (if the buffer is too large for the current archicture) | |
| 217 if (static_cast<size_t>(stream.avail_in) != compressedSize || | |
| 218 static_cast<size_t>(stream.avail_out) != uncompressedSize) | |
| 219 { | |
| 220 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 221 } | |
| 222 | |
| 223 // Initialize the compression engine | |
| 224 int error = inflateInit2(&stream, | |
| 225 MAX_WBITS + 16); // this is a gzip input | |
| 226 | |
| 227 if (error != Z_OK) | |
| 228 { | |
| 229 // Cannot initialize zlib | |
| 230 uncompressed.clear(); | |
| 231 throw OrthancException(ErrorCode_InternalError); | |
| 232 } | |
| 233 | |
| 234 // Uncompress the input buffer | |
| 235 error = inflate(&stream, Z_FINISH); | |
| 236 | |
| 237 if (error != Z_STREAM_END) | |
| 238 { | |
| 239 inflateEnd(&stream); | |
| 240 uncompressed.clear(); | |
| 241 | |
| 242 switch (error) | |
| 243 { | |
| 244 case Z_MEM_ERROR: | |
| 245 throw OrthancException(ErrorCode_NotEnoughMemory); | |
| 246 | |
| 247 case Z_BUF_ERROR: | |
| 248 case Z_NEED_DICT: | |
| 249 throw OrthancException(ErrorCode_BadFileFormat); | |
| 250 | |
| 251 default: | |
| 252 throw OrthancException(ErrorCode_InternalError); | |
| 253 } | |
| 254 } | |
| 255 | |
| 256 size_t size = stream.total_out; | |
| 257 | |
| 258 if (inflateEnd(&stream) != Z_OK) | |
| 259 { | |
| 260 uncompressed.clear(); | |
| 261 throw OrthancException(ErrorCode_InternalError); | |
| 262 } | |
| 263 | |
| 264 if (size != uncompressedSize) | |
| 265 { | |
| 266 uncompressed.clear(); | |
| 267 | |
| 268 // The uncompressed size was not that properly guess, presumably | |
| 269 // because of a file size over 4GB. Should fallback to | |
| 270 // stream-based decompression. | |
|
2954
d924f9bb61cc
taking advantage of details in OrthancException
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
2447
diff
changeset
|
271 throw OrthancException(ErrorCode_NotImplemented, |
|
d924f9bb61cc
taking advantage of details in OrthancException
Sebastien Jodogne <s.jodogne@gmail.com>
parents:
2447
diff
changeset
|
272 "The uncompressed size of a gzip-encoded buffer was not properly guessed"); |
| 1513 | 273 } |
| 274 } | |
| 275 } |