digest-murmurhash 0.0.1 → 0.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 6eb394fa67dc2adc52961e3d8de2faeef0826f68
-  data.tar.gz: 1ee189aa44955238877fcfd5978ae13fe2007185
+  metadata.gz: 5f158d21ca468ee7cccb9c39ba6fc57660850edd
+  data.tar.gz: 82730d511f8136b2ff960c0d2d99a62c9a78c339
 SHA512:
-  metadata.gz: 438420bdde5f3cbd0082274133c1dcdf61970a427d06592be020b93120cc8b9dcf8b87f21b93647aef26bbd9b59b65ad072fbaa8984b5886a7e20b6dd06d9c1c
-  data.tar.gz: 06d713bb5248de042c7539f99258b4d642bc4d941145ccf07a4cd08e8d62eb4e96c14031b8eb44eb516cb9a234dcccc23906158afad8ef40f76b9295818809eb
+  metadata.gz: 4494bc8ee54eb5bc525c9c8d5a8f275c299b73350fd0f6fcb638e7272fb546cab8d5bd5867cab5f388845ddc2eeec11302c8d982ea1e0e5bb33cc2b50b127d28
+  data.tar.gz: 20cc99086cc26e5ba616a9d179aa11121f88704a8effb6a417dd3cc58032b8c0fb89a21818c1bf790bfd7dde3b94a0477cd67d74d0fe860c882e339ddfd3c1b8

data/README.md

@@ -2,7 +2,7 @@
 
 [![Build Status](https://travis-ci.org/ksss/digest-murmurhash.png?branch=master)](https://travis-ci.org/ksss/digest-murmurhash)
 
-Digest::MurmurHash is a class use algorithm of MurmurHash(MurmurHash1 32-bit) desiged by Austin Appleby.
+Digest::MurmurHash is a class of use algorithm MurmurHash(MurmurHash1 32-bit) desiged by Austin Appleby.
 
 Digest::MurmurHash compliance Digest API of Ruby.
 
@@ -16,7 +16,7 @@ You can use same interface built in Digest::XXX classes.
 require 'digest/murmurhash'
 
 p Digest::MurmurHash.hexdigest('murmurhash') #=> 'c709abd5'
-p Digest::MurmurHash.file("./LICENSE.txt") #=> '712e9641'
+p Digest::MurmurHash.file("./LICENSE.txt").hexdigest #=> '712e9641'
 ```
 
 ## Class tree

data/digest-murmurhash.gemspec

@@ -6,11 +6,11 @@ $LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
 Gem::Specification.new do |spec|
   spec.name          = "digest-murmurhash"
   # FIXME replace Digest::MurmurHash::VERSION but raise NameError
-  spec.version       = "0.0.1"
+  spec.version       = "0.1.0"
   spec.author        = "ksss"
   spec.email         = "co000ri@gmail.com"
-  spec.description   = %q{The MurmurHash.}
-  spec.summary       = %q{The MurmurHash create by Austin Appleby.}
+  spec.description   = %q{Digest::MurmurHash is a class of use algorithm MurmurHash(MurmurHash1 32-bit) desiged by Austin Appleby.}
+  spec.summary       = %q{Digest::MurmurHash is a class of use algorithm MurmurHash(MurmurHash1 32-bit) desiged by Austin Appleby.}
   spec.homepage      = ""
   spec.license       = "MIT"
 

data/ext/digest/murmurhash/murmurhash.c

@@ -12,7 +12,9 @@
 ID id_concat;
 
 typedef struct {
-	VALUE data;
+	char* data;
+	char* p;
+	size_t memsize;
 } murmur_t;
 
 #define MURMURHASH(self, name) \
@@ -25,19 +27,20 @@ typedef struct {
 static void
 murmur_init(murmur_t* ptr)
 {
-	ptr->data = rb_str_new("", 0);
+	ptr->data = (char*) malloc(sizeof(char) * 64);
+	ptr->p = ptr->data;
+	ptr->memsize = 64;
 }
 
 static void
 murmur_mark(murmur_t* ptr)
 {
-	rb_gc_mark(ptr->data);
 }
 
 static void
 murmur_free(murmur_t* ptr)
 {
-	xfree(ptr);
+	free(ptr->data);
 }
 
 static VALUE
@@ -49,50 +52,72 @@ murmur_alloc(VALUE self)
 }
 
 static VALUE
-murmur_initialize_copy(VALUE self, VALUE obj)
+murmur_initialize_copy(VALUE copy, VALUE origin)
 {
-	if (self == obj) return self;
+	murmur_t *ptr_copy, *ptr_origin;
+	size_t data_len;
 
-	MURMURHASH(self, pctx1);
-	MURMURHASH(obj, pctx2);
+	if (copy == origin) return copy;
 
-	rb_check_frozen(self);
+	rb_check_frozen(copy);
 
-	memcpy(pctx1, pctx2, sizeof(murmur_t));
+	Data_Get_Struct(copy, murmur_t, ptr_copy);
+	Data_Get_Struct(origin, murmur_t, ptr_origin);
 
-	return self;
+	data_len = ptr_origin->p - ptr_origin->data;
+	ptr_copy->data = (char*) malloc(sizeof(char) * ptr_origin->memsize);
+	memcpy(ptr_copy->data, ptr_origin->data, data_len);
+	ptr_copy->p = ptr_copy->data + data_len;
+	ptr_copy->memsize = ptr_origin->memsize;
+
+	return copy;
 }
 
 static VALUE
 murmur_reset(VALUE self)
 {
 	MURMURHASH(self, ptr);
-	murmur_init(ptr);
+	ptr->p = ptr->data;
 	return self;
 }
 
 static VALUE
 murmur_update(VALUE self, VALUE str)
 {
+	size_t data_len, str_len, require, newsize;
+	const char* str_p;
 	MURMURHASH(self, ptr);
+
 	StringValue(str);
-	rb_funcall(ptr->data, id_concat, 1, str);
+	str_p = RSTRING_PTR(str);
+	str_len = RSTRING_LEN(str);
+	data_len = (ptr->p - ptr->data);
+	require = data_len + str_len;
+	if (ptr->memsize < require) {
+		newsize = ptr->memsize;
+		while (newsize < require) {
+			newsize *= 2;
+		}
+		ptr->data = realloc(ptr->data, sizeof(char) * newsize);
+		ptr->p = ptr->data + data_len;
+		ptr->memsize = newsize;
+	}
+	memcpy(ptr->p, str_p, str_len);
+	ptr->p += str_len;
+
 	return self;
 }
 
-static VALUE
-murmur_finish(VALUE self)
+static uint32_t
+murmur_hash_process(murmur_t* ptr)
 {
 	const uint32_t m = 0x5bd1e995;
 	const uint8_t r = 16;
 	uint32_t length, h;
 	const char* p;
-	uint8_t digest[MURMURHASH_DIGEST_LENGTH];
-
-	MURMURHASH(self, ptr);
 
-	p = RSTRING_PTR(ptr->data);
-	length = RSTRING_LEN(ptr->data);
+	p = ptr->data;
+	length = ptr->p - ptr->data;
 	h = length * m;
 
 	while (4 <= length) {
@@ -119,6 +144,18 @@ murmur_finish(VALUE self)
 	h *= m;
 	h ^= h >> 17;
 
+	return h;
+}
+
+static VALUE
+murmur_finish(VALUE self)
+{
+	uint32_t h;
+	uint8_t digest[MURMURHASH_DIGEST_LENGTH];
+	MURMURHASH(self, ptr);
+
+	h = murmur_hash_process(ptr);
+
 	digest[0] = (h >> 24);
 	digest[1] = (h >> 16);
 	digest[2] = (h >> 8);
@@ -139,6 +176,13 @@ murmur_block_length(VALUE self)
 	return INT2NUM(MURMURHASH_BLOCK_LENGTH);
 }
 
+static VALUE
+murmur_to_i(VALUE self)
+{
+	MURMURHASH(self, ptr);
+	return UINT2NUM(murmur_hash_process(ptr));
+}
+
 void
 Init_murmurhash()
 {
@@ -160,4 +204,6 @@ Init_murmurhash()
 	rb_define_private_method(cDigest_MurmurHash, "finish", murmur_finish, 0);
 	rb_define_method(cDigest_MurmurHash, "digest_length", murmur_digest_length, 0);
 	rb_define_method(cDigest_MurmurHash, "block_length", murmur_block_length, 0);
+
+	rb_define_method(cDigest_MurmurHash, "to_i", murmur_to_i, 0);
 }

data/lib/digest/murmurhash/version.rb

@@ -1,5 +1,5 @@
 module Digest
   class MurmurHash
-    VERSION = "0.0.1"
+    VERSION = "0.1.0"
   end
 end

data/spec/bench.rb

@@ -0,0 +1,73 @@
+#! /usr/bin/env ruby
+
+lib = File.expand_path('../../lib', __FILE__)
+$LOAD_PATH.unshift(lib) unless $LOAD_PATH.include?(lib)
+
+require 'digest/murmurhash'
+require 'benchmark'
+
+class Integer
+  def to_32
+    self & 0xffffffff
+  end
+end
+
+def murmur_hash str
+  data = str.dup.unpack("C*")
+  m = 0x5bd1e995
+  r = 16
+  length = str.bytesize
+  h = (length * m).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.to_32
+end
+
+def rand_str
+  rand = "";
+  File.open("/dev/urandom").read(20).each_byte{|x| rand << sprintf("%02x",x)}
+  rand
+end
+
+s = rand_str
+p [murmur_hash(s)].pack("N")
+p Digest::MurmurHash.digest(s)
+
+Benchmark.bm do |x|
+  n = 10000
+  a = []
+  n.times { |i|
+   a[i] = rand_str
+  }
+
+  x.report {n.times{ |i|
+    [murmur_hash(a[i])].pack("N")
+  }}
+  x.report {n.times{ |i|
+    Digest::MurmurHash.digest(a[i])
+  }}
+end

data/spec/digest_spec.rb

@@ -53,4 +53,8 @@ describe Digest::MurmurHash do
     # Therefore `block_length` return chunk size for calculate MurmurHash (equal 4) 
     expect(murmur.block_length).to eq(4);
   end
+
+  it "to_i" do
+    expect(murmur.update("murmurhash").to_i).to eq(0xc709abd5);
+  end
 end

data/spec/mem_spec.rb

@@ -1,3 +1,5 @@
+require 'spec_helper'
+
 describe Digest::MurmurHash do
   it "gc safe" do
     murmur = Digest::MurmurHash.new

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: digest-murmurhash
 version: !ruby/object:Gem::Version
-  version: 0.0.1
+  version: 0.1.0
 platform: ruby
 authors:
 - ksss
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2013-12-01 00:00:00.000000000 Z
+date: 2013-12-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -66,7 +66,8 @@ dependencies:
     - - ~>
       - !ruby/object:Gem::Version
         version: 0.8.3
-description: The MurmurHash.
+description: Digest::MurmurHash is a class of use algorithm MurmurHash(MurmurHash1
+  32-bit) desiged by Austin Appleby.
 email: co000ri@gmail.com
 executables: []
 extensions:
@@ -83,6 +84,7 @@ files:
 - ext/digest/murmurhash/extconf.rb
 - ext/digest/murmurhash/murmurhash.c
 - lib/digest/murmurhash/version.rb
+- spec/bench.rb
 - spec/digest_spec.rb
 - spec/mem_spec.rb
 - spec/spec_helper.rb
@@ -109,8 +111,10 @@ rubyforge_project:
 rubygems_version: 2.1.6
 signing_key: 
 specification_version: 4
-summary: The MurmurHash create by Austin Appleby.
+summary: Digest::MurmurHash is a class of use algorithm MurmurHash(MurmurHash1 32-bit)
+  desiged by Austin Appleby.
 test_files:
+- spec/bench.rb
 - spec/digest_spec.rb
 - spec/mem_spec.rb
 - spec/spec_helper.rb