// Copyright (c) 2011 Google, Inc. // // Permission is hereby granted, free of charge, to any person obtaining a copy // of this software and associated documentation files (the "Software"), to deal // in the Software without restriction, including without limitation the rights // to use, copy, modify, merge, publish, distribute, sublicense, and/or sell // copies of the Software, and to permit persons to whom the Software is // furnished to do so, subject to the following conditions: // // The above copyright notice and this permission notice shall be included in // all copies or substantial portions of the Software. // // THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR // IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, // FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE // AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER // LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, // OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN // THE SOFTWARE. // // CityHash, by Geoff Pike and Jyrki Alakuijala // // http://code.google.com/p/cityhash/ // // This file provides a few functions for hashing strings. All of them are // high-quality functions in the sense that they pass standard tests such // as Austin Appleby's SMHasher. They are also fast. // // For 64-bit x86 code, on short strings, we don't know of anything faster than // CityHash64 that is of comparable quality. We believe our nearest competitor // is Murmur3. For 64-bit x86 code, CityHash64 is an excellent choice for hash // tables and most other hashing (excluding cryptography). // // For 64-bit x86 code, on long strings, the picture is more complicated. // On many recent Intel CPUs, such as Nehalem, Westmere, Sandy Bridge, etc., // CityHashCrc128 appears to be faster than all competitors of comparable // quality. CityHash128 is also good but not quite as fast. We believe our // nearest competitor is Bob Jenkins' Spooky. We don't have great data for // other 64-bit CPUs, but for long strings we know that Spooky is slightly // faster than CityHash on some relatively recent AMD x86-64 CPUs, for example. // Note that CityHashCrc128 is declared in citycrc.h. // // For 32-bit x86 code, we don't know of anything faster than CityHash32 that // is of comparable quality. We believe our nearest competitor is Murmur3A. // (On 64-bit CPUs, it is typically faster to use the other CityHash variants.) // // Functions in the CityHash family are not suitable for cryptography. // // Please see CityHash's README file for more details on our performance // measurements and so on. // // WARNING: This code has been only lightly tested on big-endian platforms! // It is known to work well on little-endian platforms that have a small penalty // for unaligned reads, such as current Intel and AMD moderate-to-high-end CPUs. // It should work on all 32-bit and 64-bit platforms that allow unaligned reads; // bug reports are welcome. // // By the way, for some hash functions, given strings a and b, the hash // of a+b is easily derived from the hashes of a and b. This property // doesn't hold for any hash functions in this file. #ifndef BASE_THIRD_PARTY_CITYHASH_CITY_H_ #define BASE_THIRD_PARTY_CITYHASH_CITY_H_ #include <stdint.h> #include <stdlib.h> // for size_t. #include <utility> // XXX(cavalcantii): Declaring it inside of the 'base' namespace allows to // handle linker symbol clash error with deprecated CityHash from // third_party/smhasher in a few unit tests. namespace base { namespace internal { namespace cityhash_v111 { uint8; uint32; uint64; uint128; inline uint64 Uint128Low64(const uint128& x) { … } inline uint64 Uint128High64(const uint128& x) { … } // Hash function for a byte array. uint64 CityHash64(const char* buf, size_t len); // Hash function for a byte array. For convenience, a 64-bit seed is also // hashed into the result. uint64 CityHash64WithSeed(const char* buf, size_t len, uint64 seed); // Hash function for a byte array. For convenience, two seeds are also // hashed into the result. uint64 CityHash64WithSeeds(const char* buf, size_t len, uint64 seed0, uint64 seed1); // Hash function for a byte array. uint128 CityHash128(const char* s, size_t len); // Hash function for a byte array. For convenience, a 128-bit seed is also // hashed into the result. uint128 CityHash128WithSeed(const char* s, size_t len, uint128 seed); // Hash function for a byte array. Most useful in 32-bit binaries. uint32 CityHash32(const char* buf, size_t len); // Hash 128 input bits down to 64 bits of output. // This is intended to be a reasonably good hash function. inline uint64 Hash128to64(const uint128& x) { … } } // namespace cityhash_v111 } // namespace internal } // namespace base #endif // CITY_HASH_H_