digest-murmurhash 1.1.0 → 1.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: cd9a5f608b1cac914b791b811bc75eba224ca841
-  data.tar.gz: ed6533e449b132c385c3457708c52d556d47114d
+  metadata.gz: b61d7d502181711a8b3b63e9b5609274e56c3b27
+  data.tar.gz: 66e152971a285ad197563d14dd480a8e03b5095a
 SHA512:
-  metadata.gz: 54555dc73a82614b6fe9ba109662e96d4dad18b8e9ed30a073d9cde49d611b256158381d9c61280a9dbe7783e8db4967ccc8de2af9a2a507c71bdca87c99730b
-  data.tar.gz: 4b47b42486c884ddb3edee9b7c4f1ec11529657baf471492e45c113abc6e7dba622a5e7e7cc664480c9790b139fe49942e091c3257f0134635316075f8ce3791
+  metadata.gz: 1a761f71b68f236f70acdd525322e99a1f0e7bb304fb22c164dd557a461963b9a62c85d017a509f7409c47ab8377f3475f4f0a248ff2d39e4ccd7eee4ffd60e0
+  data.tar.gz: 074fafcd53b409d3596af8d16bdf51ad2a8756c10d65974ec3a9365537d7fb6375ba6e163cc1938d99d18dbce1cd624e72cbf9b9bfcbc5f4ab09ac186d8f175e

data/.travis.yml

@@ -1,6 +1,8 @@
 language: ruby
 rvm:
-  - 1.9.3
   - 2.0.0
-  - 2.1.3
-  - 2.2.0
+  - 2.1.8
+  - 2.2.4
+  - 2.3.0
+notifications:
+  email: false

data/README.md

@@ -50,50 +50,35 @@ p Digest::MurmurHash1.file("./LICENSE.txt").hexdigest #=> "41962e71"
 - You can try *all* Hurmurhash algorithms
 - You can use same interface built-in Digest::{MD5,SH1,...} classes.
 
-## Simple benchmark
-
-```ruby
-#! /usr/bin/env ruby
-
-require 'benchmark'
-require 'digest/md5'
-require 'digest/sha1'
-require 'digest/murmurhash'
-require 'digest/siphash'
-
-include Digest
-
-n = 1000
-str = "teststrings" * 1024 * 10
-
-Benchmark.bm do |f|
-  [MurmurHash1, MurmurHash2, MurmurHash2A, MurmurHash64A, MurmurHash64B, MurmurHashNeutral2, MurmurHashAligned2, MurmurHash3_x86_32, MurmurHash3_x86_128, MurmurHash3_x64_128, SipHash, MD5, SHA1].each do |klass|
-    f.report(klass.to_s) {
-      i = 0
-      while i < n
-        klass.digest(str)
-        i += 1
-      end
-    }
-  end
-end
-```
+## Benchmark
 
 ```
-       user     system      total        real
-Digest::MurmurHash1  0.050000   0.010000   0.060000 (  0.047889)
-Digest::MurmurHash2  0.030000   0.000000   0.030000 (  0.034564)
-Digest::MurmurHash2A  0.030000   0.010000   0.040000 (  0.031808)
-Digest::MurmurHash64A  0.010000   0.000000   0.010000 (  0.018400)
-Digest::MurmurHash64B  0.030000   0.000000   0.030000 (  0.027818)
-Digest::MurmurHashNeutral2  0.040000   0.000000   0.040000 (  0.041021)
-Digest::MurmurHashAligned2  0.020000   0.010000   0.030000 (  0.030409)
-Digest::MurmurHash3_x86_32  0.130000   0.010000   0.140000 (  0.139622)
-Digest::MurmurHash3_x86_128  0.120000   0.020000   0.140000 (  0.143768)
-Digest::MurmurHash3_x64_128  0.070000   0.010000   0.080000 (  0.072687)
-Digest::SipHash  0.060000   0.010000   0.070000 (  0.068243)
-Digest::MD5  0.130000   0.010000   0.140000 (  0.153793)
-Digest::SHA1  0.130000   0.020000   0.150000 (  0.137686)
+$ bundle ex ruby spec/bench.rb
+### condition
+
+    RUBY_VERSION = 2.2.2
+    count = 100
+    data size = 1024 KB
+
+### benchmark
+
+MurmurHash1	0.05145844700746238
+MurmurHash2	0.03399810096016154
+MurmurHash2A	0.032202123024035245
+MurmurHash64A	0.01880742999492213
+MurmurHash64B	0.02227930899243802
+MurmurHashNeutral2	0.04134096298366785
+MurmurHashAligned2	0.03514020598959178
+MurmurHash3_x86_32	0.1174831380485557
+MurmurHash3_x86_128	0.1244338889955543
+MurmurHash3_x64_128	0.07122775202151388
+Digest::MD5	0.1401238819817081
+Digest::SHA1	0.09980463003739715
+Digest::SHA256	0.2248609460075386
+Digest::SHA2	0.2245309569989331
+OpenSSL::HMAC(sha256)	0.23656713595846668
+Base64	0.1765239799860865
+Zlib.crc32	0.009485946036875248
 ```
 
 ## Installation

data/digest-murmurhash.gemspec

@@ -2,7 +2,7 @@
 
 Gem::Specification.new do |spec|
   spec.name          = "digest-murmurhash"
-  spec.version       = "1.1.0"
+  spec.version       = "1.1.1"
   spec.authors       = ["ksss"]
   spec.email         = ["co000ri@gmail.com"]
   spec.description   = %q{digest-murmurhash is class collections of use algorithm MurmurHash desiged by Austin Appleby.}

data/ext/digest/murmurhash/1.c

@@ -45,7 +45,7 @@ VALUE
 murmur1_finish(VALUE self)
 {
   uint8_t digest[4];
-  uint64_t h;
+  uint32_t h;
 
   h = _murmur_finish32(self, murmur_hash_process1);
   assign_by_endian_32(digest, h);

data/ext/digest/murmurhash/2.c

@@ -44,7 +44,7 @@ VALUE
 murmur2_finish(VALUE self)
 {
   uint8_t digest[4];
-  uint64_t h;
+  uint32_t h;
 
   h = _murmur_finish32(self, murmur_hash_process2);
   assign_by_endian_32(digest, h);

data/ext/digest/murmurhash/2a.c

@@ -45,7 +45,7 @@ VALUE
 murmur2a_finish(VALUE self)
 {
   uint8_t digest[4];
-  uint64_t h;
+  uint32_t h;
 
   h = _murmur_finish32(self, murmur_hash_process2a);
   assign_by_endian_32(digest, h);

data/ext/digest/murmurhash/aligned2.c

@@ -16,7 +16,7 @@ murmur_hash_process_aligned2(const char * key, uint32_t len, uint32_t seed)
 
   uint32_t h = seed ^ len;
 
-  int align = (uint64_t)data & 3;
+  uint32_t align = (uint64_t)data & 3;
 
   if(align && (len >= 4))
   {
@@ -150,7 +150,7 @@ VALUE
 murmur_aligned2_s_digest(int argc, VALUE *argv, VALUE klass)
 {
   uint8_t digest[4];
-  uint64_t h;
+  uint32_t h;
   h = _murmur_s_digest32(argc, argv, klass, murmur_hash_process_aligned2);
   assign_by_endian_32(digest, h);
   return rb_str_new((const char*) digest, 4);

data/ext/digest/murmurhash/init.c

@@ -111,13 +111,22 @@ assign_by_endian_128(uint8_t *digest, void *out)
   }
 }
 
+static uint32_t
+rstring2uint32_t(VALUE str)
+{
+  long len = RSTRING_LEN(str);
+  if (UINT32_MAX < len) {
+    rb_raise(rb_eRangeError, "String length=%ld will overflow from long to uint32_t", len);
+  }
+  return (uint32_t)len;
+}
 
 uint32_t
 _murmur_finish32(VALUE self, uint32_t (*process)(const char*, uint32_t, uint32_t))
 {
   const char *seed = RSTRING_PTR(rb_ivar_get(self, iv_seed));
   VALUE buffer = rb_ivar_get(self, iv_buffer);
-  return process(RSTRING_PTR(buffer), RSTRING_LEN(buffer), *(uint32_t*)seed);
+  return process(RSTRING_PTR(buffer), rstring2uint32_t(buffer), *(uint32_t*)seed);
 }
 
 uint64_t
@@ -125,7 +134,7 @@ _murmur_finish64(VALUE self, uint64_t (*process)(const char*, uint32_t, uint64_t
 {
   const char *seed = RSTRING_PTR(rb_ivar_get(self, iv_seed));
   VALUE buffer = rb_ivar_get(self, iv_buffer);
-  return process(RSTRING_PTR(buffer), RSTRING_LEN(buffer), *(uint64_t*)seed);
+  return process(RSTRING_PTR(buffer), rstring2uint32_t(buffer), *(uint64_t*)seed);
 }
 
 void
@@ -133,7 +142,7 @@ _murmur_finish128(VALUE self, void *out, void (*process)(const char*, uint32_t,
 {
   const char *seed = RSTRING_PTR(rb_ivar_get(self, iv_seed));
   VALUE buffer = rb_ivar_get(self, iv_buffer);
-  process(RSTRING_PTR(buffer), RSTRING_LEN(buffer), *(uint32_t*)seed, out);
+  process(RSTRING_PTR(buffer), rstring2uint32_t(buffer), *(uint32_t*)seed, out);
 }
 
 uint32_t
@@ -159,7 +168,7 @@ _murmur_s_digest32(int argc, VALUE *argv, VALUE klass, uint32_t (*process)(const
     seed = RSTRING_PTR(rb_const_get(klass, id_DEFAULT_SEED));
   }
 
-  return process(RSTRING_PTR(str), RSTRING_LEN(str), *(uint32_t*)seed);
+  return process(RSTRING_PTR(str), rstring2uint32_t(str), *(uint32_t*)seed);
 }
 
 uint64_t
@@ -185,7 +194,7 @@ _murmur_s_digest64(int argc, VALUE *argv, VALUE klass, uint64_t (*process)(const
     seed = RSTRING_PTR(rb_const_get(klass, id_DEFAULT_SEED));
   }
 
-  return process(RSTRING_PTR(str), RSTRING_LEN(str), *(uint64_t*)seed);
+  return process(RSTRING_PTR(str), rstring2uint32_t(str), *(uint64_t*)seed);
 }
 
 void
@@ -212,7 +221,7 @@ _murmur_s_digest128(int argc, VALUE *argv, VALUE klass, void *out, void (*proces
     seed = RSTRING_PTR(rb_const_get(klass, id_DEFAULT_SEED));
   }
 
-  process(RSTRING_PTR(str), RSTRING_LEN(str), *(uint32_t*)seed, out);
+  process(RSTRING_PTR(str), rstring2uint32_t(str), *(uint32_t*)seed, out);
 }
 
 

data/ext/digest/murmurhash/neutral2.c

@@ -62,7 +62,7 @@ VALUE
 murmur_neutral2_s_digest(int argc, VALUE *argv, VALUE klass)
 {
   uint8_t digest[4];
-  uint64_t h;
+  uint32_t h;
   h = _murmur_s_digest32(argc, argv, klass, murmur_hash_process_neutral2);
   assign_by_endian_32(digest, h);
   return rb_str_new((const char*) digest, 4);

data/lib/digest/murmurhash/base.rb

@@ -17,7 +17,7 @@ module Digest
   }.each do |name, s|
     class_eval %Q{
       class MurmurHash#{name} < Digest::Class
-        DEFAULT_SEED = "#{"\x00" * s.seed_length}".encode('ASCII-8BIT')
+        DEFAULT_SEED = "#{"\x00" * s.seed_length}".b
 
         def initialize
           @buffer = ""
@@ -25,7 +25,7 @@ module Digest
         end
 
         def update(str)
-          @buffer += str
+          @buffer << str
           self
         end
         alias << update
@@ -41,7 +41,7 @@ module Digest
         end
 
         def seed=(s)
-          fail ArgumentError, "seed string should be #{s.seed_length} length" if #{s.seed_length} != s.length
+          raise ArgumentError, "seed string should be #{s.seed_length} length" if #{s.seed_length} != s.length
           @seed = s
         end
 

data/spec/bench.rb

@@ -4,6 +4,10 @@ lib = File.expand_path('../../lib', __FILE__)
 $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 
 require 'digest/murmurhash'
+require 'digest/md5'
+require 'digest/sha1'
+require 'openssl'
+require 'zlib'
 require 'benchmark'
 
 @rands = "abcdefghijklmnopqrstuvwxyzABCDEFGHIJKLMNOPQRSTUVWXYZ0123456789".split(//)
@@ -15,53 +19,6 @@ def rand_str length
   rand
 end
 
-class Integer
-  def to_32
-    self & 0xffffffff
-  end
-end
-
-def murmur_hash str, seed
-  data = str.dup.unpack("C*")
-  m = 0x5bd1e995
-  r = 16
-  length = str.bytesize
-  h = (seed ^ (length * m).to_32).to_32
-
-  while 4 <= length
-    d = data.shift(4).pack("C*").unpack("I")[0]
-    h = (h + d).to_32
-    h = (h * m).to_32
-    h ^= h >> r
-    length -= 4
-  end
-
-  if 2 < length
-    h = (h + (data[2] << 16).to_32).to_32
-  end
-  if 1 < length
-    h = (h + (data[1] << 8).to_32).to_32
-  end
-  if 0 < length
-    h = (h + data[0]).to_32
-    h = (h * m).to_32
-    h ^= h >> r
-  end
-
-  h = (h * m).to_32
-  h ^= h >> 10
-  h = (h * m).to_32
-  h ^= h >> 17
-
-  h
-end
-
-n = 100000
-
-a = Array.new(n, 0)
-n.times do |i|
-  a[i] = rand_str 20
-end
 seed = rand(2**32)
 seed_str32 = [seed].pack("L")
 seed_str64 = [seed].pack("Q")
@@ -69,7 +26,6 @@ c = Struct.new "Cases",
                :name,
                :func
 cases = [
-  c.new("pureRuby", proc{|x| murmur_hash x, seed }),
   c.new("MurmurHash1", proc{|x| Digest::MurmurHash1.digest x, seed_str32 }),
   c.new("MurmurHash2", proc{|x| Digest::MurmurHash2.digest x, seed_str32 }),
   c.new("MurmurHash2A", proc{|x| Digest::MurmurHash2A.digest x, seed_str32 }),
@@ -80,53 +36,60 @@ cases = [
   c.new("MurmurHash3_x86_32", proc{|x| Digest::MurmurHash3_x86_32.digest x, seed_str32 }),
   c.new("MurmurHash3_x86_128", proc{|x| Digest::MurmurHash3_x86_128.digest x, seed_str32 }),
   c.new("MurmurHash3_x64_128", proc{|x| Digest::MurmurHash3_x64_128.digest x, seed_str32 }),
+  c.new("Digest::MD5", proc{|x| Digest::MD5.digest x }),
+  c.new("Digest::SHA1", proc{|x| Digest::SHA1.digest x }),
+  c.new("Digest::SHA256", proc{|x| Digest::SHA256.digest x }),
+  c.new("Digest::SHA2", proc{|x| Digest::SHA2.digest x }),
+  c.new("OpenSSL::HMAC(sha256)", proc{|x| OpenSSL::HMAC.digest "sha256", seed_str32, x }),
+  c.new("Base64", proc{|x| [x].pack("m0") }),
+  c.new("Zlib.crc32", proc{|x| Zlib.crc32(x) }),
 ]
-reals = {}
+
 confrict = {}
 confricts = {}
 
+n = 100
+a = Array.new(n, 0)
+n.times do |i|
+  a[i] = rand_str(1024*1024)
+end
+
 puts "### condition"
 puts
 puts "    RUBY_VERSION = #{RUBY_VERSION}"
 puts "    count = #{n}"
+puts "    data size = #{a[0].length / 1024} KB"
 puts
 puts "### benchmark"
 puts
 puts "```"
-Benchmark.bm do |x|
-  cases.each do |c|
-    i = 0
-    z = x.report c.name do
-      while i < n
-        c.func.call(a[i])
-        i += 1
-      end
-    end
-
-    confrict.clear
 
-    i = 0
+GC.start
+cases.each do |c|
+  i = 0
+  b = Benchmark.realtime do
     while i < n
-      rethash = c.func.call(a[i])
-      if confrict[rethash].nil?
-        confrict[rethash] = 0
-      else
-        confrict[rethash] += 1
-      end
+      c.func.call(a[i])
       i += 1
     end
-    reals[c.name] = z.real
-    confricts[c.name] = confrict.count{|hash, count| 0 < count}
   end
-end
-puts "```"
+  puts "#{c.name}\t#{b}"
 
-puts
-puts "### real second rate (pureRuby/)"
-puts
-reals.each do |name, real|
-  puts "    " + (reals["pureRuby"] / real).to_s + "/" + name
+  confrict.clear
+
+  i = 0
+  while i < n
+    rethash = c.func.call(a[i])
+    if confrict[rethash].nil?
+      confrict[rethash] = 0
+    else
+      confrict[rethash] += 1
+    end
+    i += 1
+  end
+  confricts[c.name] = confrict.count{|hash, count| 0 < count}
 end
+puts "```"
 
 puts
 puts "### confrict count (/#{n})"

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: digest-murmurhash
 version: !ruby/object:Gem::Version
-  version: 1.1.0
+  version: 1.1.1
 platform: ruby
 authors:
 - ksss
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-10-16 00:00:00.000000000 Z
+date: 2016-01-05 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -122,7 +122,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.4.1
+rubygems_version: 2.5.1
 signing_key: 
 specification_version: 4
 summary: digest-murmurhash is class collections of use algorithm MurmurHash desiged