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