1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
|
/**
* @file hbxxh.h
* @brief High performances vectorized hashing based on xxHash.
*
* $LicenseInfo:firstyear=2023&license=viewerlgpl$
* Second Life Viewer Source Code
* Copyright (c) 2023, Henri Beauchamp.
*
* This library is free software; you can redistribute it and/or
* modify it under the terms of the GNU Lesser General Public
* License as published by the Free Software Foundation;
* version 2.1 of the License only.
*
* This library is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
* Lesser General Public License for more details.
*
* You should have received a copy of the GNU Lesser General Public
* License along with this library; if not, write to the Free Software
* Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301 USA
*
* Linden Research, Inc., 945 Battery Street, San Francisco, CA 94111 USA
* $/LicenseInfo$
*/
#ifndef LL_HBXXH_H
#define LL_HBXXH_H
#include "lluuid.h"
// HBXXH* classes are to be used where speed matters and cryptographic quality
// is not required (no "one-way" guarantee, though they are likely not worst in
// this respect than MD5 which got busted and is now considered too weak). The
// xxHash code they are built upon is vectorized and about 50 times faster than
// MD5. A 64 bits hash class is also provided for when 128 bits of entropy are
// not needed. The hashes collision rate is similar to MD5's.
// See https://github.com/Cyan4973/xxHash#readme for details.
// 64 bits hashing class
class HBXXH64
{
friend std::ostream& operator<<(std::ostream&, HBXXH64);
protected:
LOG_CLASS(HBXXH64);
public:
inline HBXXH64() { init(); }
// Constructors for special circumstances; they all digest the first passed
// parameter. Set 'do_finalize' to false if you do not want to finalize the
// context, which is useful/needed when you want to update() it afterwards.
// Ideally, the compiler should be smart enough to get our clue and
// optimize out the const bool test during inlining...
inline HBXXH64(const void* buffer, size_t len,
const bool do_finalize = true)
{
init();
update(buffer, len);
if (do_finalize)
{
finalize();
}
}
inline HBXXH64(const std::string& str, const bool do_finalize = true)
{
init();
update(str);
if (do_finalize)
{
finalize();
}
}
inline HBXXH64(std::istream& s, const bool do_finalize = true)
{
init();
update(s);
if (do_finalize)
{
finalize();
}
}
inline HBXXH64(FILE* file, const bool do_finalize = true)
{
init();
update(file);
if (do_finalize)
{
finalize();
}
}
~HBXXH64();
void update(const void* buffer, size_t len);
void update(const std::string& str);
void update(std::istream& s);
void update(FILE* file);
// Note that unlike what happens with LLMD5, you do not need to finalize()
// HBXXH64 before using digest(), and you may keep updating() it even after
// you got a first digest() (the next digest would of course change after
// any update). It is still useful to use finalize() when you do not want
// to store a final digest() result in a separate U64; after this method
// has been called, digest() simply returns mDigest value.
void finalize();
U64 digest() const;
// Fast static methods. Use them when hashing just one contiguous block of
// data.
static U64 digest(const void* buffer, size_t len);
static U64 digest(const char* str); // str must be NUL-terminated
static U64 digest(const std::string& str);
private:
void init();
private:
// We use a void pointer to avoid including xxhash.h here for XXH3_state_t
// (which cannot either be trivially forward-declared, due to complex API
// related pre-processor macros in xxhash.h).
void* mState;
U64 mDigest;
};
inline bool operator==(const HBXXH64& a, const HBXXH64& b)
{
return a.digest() == b.digest();
}
inline bool operator!=(const HBXXH64& a, const HBXXH64& b)
{
return a.digest() != b.digest();
}
// 128 bits hashing class
class HBXXH128
{
friend std::ostream& operator<<(std::ostream&, HBXXH128);
protected:
LOG_CLASS(HBXXH128);
public:
inline HBXXH128() { init(); }
// Constructors for special circumstances; they all digest the first passed
// parameter. Set 'do_finalize' to false if you do not want to finalize the
// context, which is useful/needed when you want to update() it afterwards.
// Ideally, the compiler should be smart enough to get our clue and
// optimize out the const bool test during inlining...
inline HBXXH128(const void* buffer, size_t len,
const bool do_finalize = true)
{
init();
update(buffer, len);
if (do_finalize)
{
finalize();
}
}
inline HBXXH128(const std::string& str, const bool do_finalize = true)
{
init();
update(str);
if (do_finalize)
{
finalize();
}
}
inline HBXXH128(std::istream& s, const bool do_finalize = true)
{
init();
update(s);
if (do_finalize)
{
finalize();
}
}
inline HBXXH128(FILE* file, const bool do_finalize = true)
{
init();
update(file);
if (do_finalize)
{
finalize();
}
}
~HBXXH128();
void update(const void* buffer, size_t len);
void update(const std::string& str);
void update(std::istream& s);
void update(FILE* file);
// Note that unlike what happens with LLMD5, you do not need to finalize()
// HBXXH128 before using digest(), and you may keep updating() it even
// after you got a first digest() (the next digest would of course change
// after any update). It is still useful to use finalize() when you do not
// want to store a final digest() result in a separate LLUUID; after this
// method has been called, digest() simply returns a reference on mDigest.
void finalize();
// We use an LLUUID for the digest, since this is a 128 bits wide native
// type available in the viewer code, making it easy to manipulate. It also
// allows to use HBXXH128 efficiently in LLUUID generate() and combine()
// methods.
const LLUUID& digest() const;
// Here, we avoid an LLUUID copy whenever we already got one to store the
// result *and* we did not yet call finalize().
void digest(LLUUID& result) const;
// Fast static methods. Use them when hashing just one contiguous block of
// data.
static LLUUID digest(const void* buffer, size_t len);
static LLUUID digest(const char* str); // str must be NUL-terminated
static LLUUID digest(const std::string& str);
// Same as above, but saves you from an LLUUID copy when you already got
// one for storage use.
static void digest(LLUUID& result, const void* buffer, size_t len);
static void digest(LLUUID& result, const char* str); // str NUL-terminated
static void digest(LLUUID& result, const std::string& str);
private:
void init();
private:
// We use a void pointer to avoid including xxhash.h here for XXH3_state_t
// (which cannot either be trivially forward-declared, due to complex API
// related pre-processor macros in xxhash.h).
void* mState;
LLUUID mDigest;
};
inline bool operator==(const HBXXH128& a, const HBXXH128& b)
{
return a.digest() == b.digest();
}
inline bool operator!=(const HBXXH128& a, const HBXXH128& b)
{
return a.digest() != b.digest();
}
#endif // LL_HBXXH_H
|