1// Copyright 2005 and onwards Google Inc.
2//
3// Redistribution and use in source and binary forms, with or without
4// modification, are permitted provided that the following conditions are
5// met:
6//
7// * Redistributions of source code must retain the above copyright
8// notice, this list of conditions and the following disclaimer.
9// * Redistributions in binary form must reproduce the above
10// copyright notice, this list of conditions and the following disclaimer
11// in the documentation and/or other materials provided with the
12// distribution.
13// * Neither the name of Google Inc. nor the names of its
14// contributors may be used to endorse or promote products derived from
15// this software without specific prior written permission.
16//
17// THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
18// "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
19// LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
20// A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
21// OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
22// SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
23// LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
24// DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
25// THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
26// (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
27// OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
28//
29// A light-weight compression algorithm. It is designed for speed of
30// compression and decompression, rather than for the utmost in space
31// savings.
32//
33// For getting better compression ratios when you are compressing data
34// with long repeated sequences or compressing data that is similar to
35// other data, while still compressing fast, you might look at first
36// using BMDiff and then compressing the output of BMDiff with
37// Snappy.
38
39#ifndef THIRD_PARTY_SNAPPY_SNAPPY_H__
40#define THIRD_PARTY_SNAPPY_SNAPPY_H__
41
42#include <stddef.h>
43#include <stdint.h>
44
45#include <string>
46
47#include "snappy-stubs-public.h"
48
49namespace snappy {
50 class Source;
51 class Sink;
52
53 // ------------------------------------------------------------------------
54 // Generic compression/decompression routines.
55 // ------------------------------------------------------------------------
56
57 // Compress the bytes read from "*source" and append to "*sink". Return the
58 // number of bytes written.
59 size_t Compress(Source* source, Sink* sink);
60
61 // Find the uncompressed length of the given stream, as given by the header.
62 // Note that the true length could deviate from this; the stream could e.g.
63 // be truncated.
64 //
65 // Also note that this leaves "*source" in a state that is unsuitable for
66 // further operations, such as RawUncompress(). You will need to rewind
67 // or recreate the source yourself before attempting any further calls.
68 bool GetUncompressedLength(Source* source, uint32_t* result);
69
70 // ------------------------------------------------------------------------
71 // Higher-level string based routines (should be sufficient for most users)
72 // ------------------------------------------------------------------------
73
74 // Sets "*compressed" to the compressed version of "input[0,input_length-1]".
75 // Original contents of *compressed are lost.
76 //
77 // REQUIRES: "input[]" is not an alias of "*compressed".
78 size_t Compress(const char* input, size_t input_length,
79 std::string* compressed);
80
81 // Decompresses "compressed[0,compressed_length-1]" to "*uncompressed".
82 // Original contents of "*uncompressed" are lost.
83 //
84 // REQUIRES: "compressed[]" is not an alias of "*uncompressed".
85 //
86 // returns false if the message is corrupted and could not be decompressed
87 bool Uncompress(const char* compressed, size_t compressed_length,
88 std::string* uncompressed);
89
90 // Decompresses "compressed" to "*uncompressed".
91 //
92 // returns false if the message is corrupted and could not be decompressed
93 bool Uncompress(Source* compressed, Sink* uncompressed);
94
95 // This routine uncompresses as much of the "compressed" as possible
96 // into sink. It returns the number of valid bytes added to sink
97 // (extra invalid bytes may have been added due to errors; the caller
98 // should ignore those). The emitted data typically has length
99 // GetUncompressedLength(), but may be shorter if an error is
100 // encountered.
101 size_t UncompressAsMuchAsPossible(Source* compressed, Sink* uncompressed);
102
103 // ------------------------------------------------------------------------
104 // Lower-level character array based routines. May be useful for
105 // efficiency reasons in certain circumstances.
106 // ------------------------------------------------------------------------
107
108 // REQUIRES: "compressed" must point to an area of memory that is at
109 // least "MaxCompressedLength(input_length)" bytes in length.
110 //
111 // Takes the data stored in "input[0..input_length]" and stores
112 // it in the array pointed to by "compressed".
113 //
114 // "*compressed_length" is set to the length of the compressed output.
115 //
116 // Example:
117 // char* output = new char[snappy::MaxCompressedLength(input_length)];
118 // size_t output_length;
119 // RawCompress(input, input_length, output, &output_length);
120 // ... Process(output, output_length) ...
121 // delete [] output;
122 void RawCompress(const char* input,
123 size_t input_length,
124 char* compressed,
125 size_t* compressed_length);
126
127 // Given data in "compressed[0..compressed_length-1]" generated by
128 // calling the Snappy::Compress routine, this routine
129 // stores the uncompressed data to
130 // uncompressed[0..GetUncompressedLength(compressed)-1]
131 // returns false if the message is corrupted and could not be decrypted
132 bool RawUncompress(const char* compressed, size_t compressed_length,
133 char* uncompressed);
134
135 // Given data from the byte source 'compressed' generated by calling
136 // the Snappy::Compress routine, this routine stores the uncompressed
137 // data to
138 // uncompressed[0..GetUncompressedLength(compressed,compressed_length)-1]
139 // returns false if the message is corrupted and could not be decrypted
140 bool RawUncompress(Source* compressed, char* uncompressed);
141
142 // Given data in "compressed[0..compressed_length-1]" generated by
143 // calling the Snappy::Compress routine, this routine
144 // stores the uncompressed data to the iovec "iov". The number of physical
145 // buffers in "iov" is given by iov_cnt and their cumulative size
146 // must be at least GetUncompressedLength(compressed). The individual buffers
147 // in "iov" must not overlap with each other.
148 //
149 // returns false if the message is corrupted and could not be decrypted
150 bool RawUncompressToIOVec(const char* compressed, size_t compressed_length,
151 const struct iovec* iov, size_t iov_cnt);
152
153 // Given data from the byte source 'compressed' generated by calling
154 // the Snappy::Compress routine, this routine stores the uncompressed
155 // data to the iovec "iov". The number of physical
156 // buffers in "iov" is given by iov_cnt and their cumulative size
157 // must be at least GetUncompressedLength(compressed). The individual buffers
158 // in "iov" must not overlap with each other.
159 //
160 // returns false if the message is corrupted and could not be decrypted
161 bool RawUncompressToIOVec(Source* compressed, const struct iovec* iov,
162 size_t iov_cnt);
163
164 // Returns the maximal size of the compressed representation of
165 // input data that is "source_bytes" bytes in length;
166 size_t MaxCompressedLength(size_t source_bytes);
167
168 // REQUIRES: "compressed[]" was produced by RawCompress() or Compress()
169 // Returns true and stores the length of the uncompressed data in
170 // *result normally. Returns false on parsing error.
171 // This operation takes O(1) time.
172 bool GetUncompressedLength(const char* compressed, size_t compressed_length,
173 size_t* result);
174
175 // Returns true iff the contents of "compressed[]" can be uncompressed
176 // successfully. Does not return the uncompressed data. Takes
177 // time proportional to compressed_length, but is usually at least
178 // a factor of four faster than actual decompression.
179 bool IsValidCompressedBuffer(const char* compressed,
180 size_t compressed_length);
181
182 // Returns true iff the contents of "compressed" can be uncompressed
183 // successfully. Does not return the uncompressed data. Takes
184 // time proportional to *compressed length, but is usually at least
185 // a factor of four faster than actual decompression.
186 // On success, consumes all of *compressed. On failure, consumes an
187 // unspecified prefix of *compressed.
188 bool IsValidCompressed(Source* compressed);
189
190 // The size of a compression block. Note that many parts of the compression
191 // code assumes that kBlockSize <= 65536; in particular, the hash table
192 // can only store 16-bit offsets, and EmitCopy() also assumes the offset
193 // is 65535 bytes or less. Note also that if you change this, it will
194 // affect the framing format (see framing_format.txt).
195 //
196 // Note that there might be older data around that is compressed with larger
197 // block sizes, so the decompression code should not rely on the
198 // non-existence of long backreferences.
199 static constexpr int kBlockLog = 16;
200 static constexpr size_t kBlockSize = 1 << kBlockLog;
201
202 static constexpr int kMinHashTableBits = 8;
203 static constexpr size_t kMinHashTableSize = 1 << kMinHashTableBits;
204
205 static constexpr int kMaxHashTableBits = 14;
206 static constexpr size_t kMaxHashTableSize = 1 << kMaxHashTableBits;
207} // end namespace snappy
208
209#endif // THIRD_PARTY_SNAPPY_SNAPPY_H__
210