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 <cstddef> |
43 | #include <string> |
44 | |
45 | #include "snappy-stubs-public.h" |
46 | |
47 | namespace snappy { |
48 | class Source; |
49 | class Sink; |
50 | |
51 | // ------------------------------------------------------------------------ |
52 | // Generic compression/decompression routines. |
53 | // ------------------------------------------------------------------------ |
54 | |
55 | // Compress the bytes read from "*source" and append to "*sink". Return the |
56 | // number of bytes written. |
57 | size_t Compress(Source* source, Sink* sink); |
58 | |
59 | // Find the uncompressed length of the given stream, as given by the header. |
60 | // Note that the true length could deviate from this; the stream could e.g. |
61 | // be truncated. |
62 | // |
63 | // Also note that this leaves "*source" in a state that is unsuitable for |
64 | // further operations, such as RawUncompress(). You will need to rewind |
65 | // or recreate the source yourself before attempting any further calls. |
66 | bool GetUncompressedLength(Source* source, uint32* result); |
67 | |
68 | // ------------------------------------------------------------------------ |
69 | // Higher-level string based routines (should be sufficient for most users) |
70 | // ------------------------------------------------------------------------ |
71 | |
72 | // Sets "*compressed" to the compressed version of "input[0,input_length-1]". |
73 | // Original contents of *compressed are lost. |
74 | // |
75 | // REQUIRES: "input[]" is not an alias of "*compressed". |
76 | size_t Compress(const char* input, size_t input_length, |
77 | std::string* compressed); |
78 | |
79 | // Decompresses "compressed[0,compressed_length-1]" to "*uncompressed". |
80 | // Original contents of "*uncompressed" are lost. |
81 | // |
82 | // REQUIRES: "compressed[]" is not an alias of "*uncompressed". |
83 | // |
84 | // returns false if the message is corrupted and could not be decompressed |
85 | bool Uncompress(const char* compressed, size_t compressed_length, |
86 | std::string* uncompressed); |
87 | |
88 | // Decompresses "compressed" to "*uncompressed". |
89 | // |
90 | // returns false if the message is corrupted and could not be decompressed |
91 | bool Uncompress(Source* compressed, Sink* uncompressed); |
92 | |
93 | // This routine uncompresses as much of the "compressed" as possible |
94 | // into sink. It returns the number of valid bytes added to sink |
95 | // (extra invalid bytes may have been added due to errors; the caller |
96 | // should ignore those). The emitted data typically has length |
97 | // GetUncompressedLength(), but may be shorter if an error is |
98 | // encountered. |
99 | size_t UncompressAsMuchAsPossible(Source* compressed, Sink* uncompressed); |
100 | |
101 | // ------------------------------------------------------------------------ |
102 | // Lower-level character array based routines. May be useful for |
103 | // efficiency reasons in certain circumstances. |
104 | // ------------------------------------------------------------------------ |
105 | |
106 | // REQUIRES: "compressed" must point to an area of memory that is at |
107 | // least "MaxCompressedLength(input_length)" bytes in length. |
108 | // |
109 | // Takes the data stored in "input[0..input_length]" and stores |
110 | // it in the array pointed to by "compressed". |
111 | // |
112 | // "*compressed_length" is set to the length of the compressed output. |
113 | // |
114 | // Example: |
115 | // char* output = new char[snappy::MaxCompressedLength(input_length)]; |
116 | // size_t output_length; |
117 | // RawCompress(input, input_length, output, &output_length); |
118 | // ... Process(output, output_length) ... |
119 | // delete [] output; |
120 | void RawCompress(const char* input, |
121 | size_t input_length, |
122 | char* compressed, |
123 | size_t* compressed_length); |
124 | |
125 | // Given data in "compressed[0..compressed_length-1]" generated by |
126 | // calling the Snappy::Compress routine, this routine |
127 | // stores the uncompressed data to |
128 | // uncompressed[0..GetUncompressedLength(compressed)-1] |
129 | // returns false if the message is corrupted and could not be decrypted |
130 | bool RawUncompress(const char* compressed, size_t compressed_length, |
131 | char* uncompressed); |
132 | |
133 | // Given data from the byte source 'compressed' generated by calling |
134 | // the Snappy::Compress routine, this routine stores the uncompressed |
135 | // data to |
136 | // uncompressed[0..GetUncompressedLength(compressed,compressed_length)-1] |
137 | // returns false if the message is corrupted and could not be decrypted |
138 | bool RawUncompress(Source* compressed, char* uncompressed); |
139 | |
140 | // Given data in "compressed[0..compressed_length-1]" generated by |
141 | // calling the Snappy::Compress routine, this routine |
142 | // stores the uncompressed data to the iovec "iov". The number of physical |
143 | // buffers in "iov" is given by iov_cnt and their cumulative size |
144 | // must be at least GetUncompressedLength(compressed). The individual buffers |
145 | // in "iov" must not overlap with each other. |
146 | // |
147 | // returns false if the message is corrupted and could not be decrypted |
148 | bool RawUncompressToIOVec(const char* compressed, size_t compressed_length, |
149 | const struct iovec* iov, size_t iov_cnt); |
150 | |
151 | // Given data from the byte source 'compressed' generated by calling |
152 | // the Snappy::Compress routine, this routine stores the uncompressed |
153 | // data to the iovec "iov". The number of physical |
154 | // buffers in "iov" is given by iov_cnt and their cumulative size |
155 | // must be at least GetUncompressedLength(compressed). The individual buffers |
156 | // in "iov" must not overlap with each other. |
157 | // |
158 | // returns false if the message is corrupted and could not be decrypted |
159 | bool RawUncompressToIOVec(Source* compressed, const struct iovec* iov, |
160 | size_t iov_cnt); |
161 | |
162 | // Returns the maximal size of the compressed representation of |
163 | // input data that is "source_bytes" bytes in length; |
164 | size_t MaxCompressedLength(size_t source_bytes); |
165 | |
166 | // REQUIRES: "compressed[]" was produced by RawCompress() or Compress() |
167 | // Returns true and stores the length of the uncompressed data in |
168 | // *result normally. Returns false on parsing error. |
169 | // This operation takes O(1) time. |
170 | bool GetUncompressedLength(const char* compressed, size_t compressed_length, |
171 | size_t* result); |
172 | |
173 | // Returns true iff the contents of "compressed[]" can be uncompressed |
174 | // successfully. Does not return the uncompressed data. Takes |
175 | // time proportional to compressed_length, but is usually at least |
176 | // a factor of four faster than actual decompression. |
177 | bool IsValidCompressedBuffer(const char* compressed, |
178 | size_t compressed_length); |
179 | |
180 | // Returns true iff the contents of "compressed" can be uncompressed |
181 | // successfully. Does not return the uncompressed data. Takes |
182 | // time proportional to *compressed length, but is usually at least |
183 | // a factor of four faster than actual decompression. |
184 | // On success, consumes all of *compressed. On failure, consumes an |
185 | // unspecified prefix of *compressed. |
186 | bool IsValidCompressed(Source* compressed); |
187 | |
188 | // The size of a compression block. Note that many parts of the compression |
189 | // code assumes that kBlockSize <= 65536; in particular, the hash table |
190 | // can only store 16-bit offsets, and EmitCopy() also assumes the offset |
191 | // is 65535 bytes or less. Note also that if you change this, it will |
192 | // affect the framing format (see framing_format.txt). |
193 | // |
194 | // Note that there might be older data around that is compressed with larger |
195 | // block sizes, so the decompression code should not rely on the |
196 | // non-existence of long backreferences. |
197 | static constexpr int kBlockLog = 16; |
198 | static constexpr size_t kBlockSize = 1 << kBlockLog; |
199 | |
200 | static constexpr int kMinHashTableBits = 8; |
201 | static constexpr size_t kMinHashTableSize = 1 << kMinHashTableBits; |
202 | |
203 | static constexpr int kMaxHashTableBits = 14; |
204 | static constexpr size_t kMaxHashTableSize = 1 << kMaxHashTableBits; |
205 | } // end namespace snappy |
206 | |
207 | #endif // THIRD_PARTY_SNAPPY_SNAPPY_H__ |
208 | |