hash_digest 1.1.2 → 1.1.3

checksums.yaml

@@ -1,15 +1,15 @@
 ---
 !binary "U0hBMQ==":
   metadata.gz: !binary |-
-    ZGNkZjg5MjcyZmI4ODJiNzFiYTNkNmI3MmE5MzcyZjNiZjRiZDRiMg==
+    NzI4MWEwNmNhZGM2NWQ3M2E5ZWRiZTgyODU1OGQ3NjNjMTMzYWQyNg==
   data.tar.gz: !binary |-
-    NDExNjc1YzI5OGJlNDQ5ZjUzYzM0MTU1YTYwNGQ2Mjg1MjI0YTdmNg==
+    MDhlNjIwYTRjNWY0ZTI1NDlkNDg5NWZkOGZjNzkwMGVmOTJhNDBlNw==
 SHA512:
   metadata.gz: !binary |-
-    MTExOWU5MTkyNWRjY2JjYjBlMDRhY2ZhOGU5YmNmNGY0OGUzYjg5ZjlkZTRk
-    NWU1YTdmMzAxMTkxYzQ4MjBiMDIzODc4MGRiZGUyNTNjZmI1MTgzOGM2M2U5
-    ODY5YzkyOTcwNGZjMzM4ZGIzMWJmNTU1ZTQ5ZWQzNWI3NDY0YmI=
+    YTQyMWEzNzIzNWRmMWIyZWYyZTVjMjdiMDJlMDQ4ODUwYjVhMDc0MmIwNWIx
+    MGM0ZmMxYzZjMmZkMDY0MjJmMTE4OWY5YmEzOTQyYjRlMDMxMzA2NjNiM2Fj
+    ZGI5Y2VkMjhiNGYxMDk3ODlkOTZkNmFhY2ZmNzY1ZTRmYmVkMmU=
   data.tar.gz: !binary |-
-    MTU0NDNjMzg0M2Q2NmEyZWM0OWFlODU3NjU5NDRkMTY0NjlkZjQwOTBiNTVl
-    MzU3MjY0MDYxNjcyZDdlMTcxOWE4MzJhNWMxYzYwZTgwODM5MDY0N2U1ZmQ4
-    MjdkOTE4YTg0YTc4YmNmZTgyYTRmNzQ4YmZjODZkZTkxYWFlMmM=
+    NjI4MTM2ZDhmMGJiNTI3NzNiN2RlNzg5ZDU2OGRmZTEyOGM2Y2ViOWNiNzgw
+    ZWIyZjE2Yzk3MGNmZTUwOThiYTRiODk2NmI5ZDY2Y2YxYmQ3NTQyMTdiNDEw
+    MzdkYWJiMDkyMWQ0ODVkZGNiZmNiOTAwNTkxYWY4M2VjMmJmNTg=

data/CHANGELOG

@@ -1,3 +1,10 @@
+1.1.3 / 2014-01-14
+
+* Enhancements
+
+  * digest2 is deprecated - too many collisions because we were using MurmurHash3 V32 under the hood
+  * digest3 is the replacement - uses SHA1 hexdigest
+
 1.1.2 / 2014-01-06
 
 * Bug fixes

data/README.markdown

@@ -6,51 +6,42 @@ Extracted from [`RemoteTable`](https://github.com/seamusabshere/remote_table).
 
 ## Note plz
 
-1. You should use `HashDigest.digest2` for new applications. `hexdigest` is legacy.
-2. `digest2`, unlike its predecessor, is CASE SENSITIVE.
+1. You should use `HashDigest.digest3` for new applications. `hexdigest` and `digest2` are legacy.
+2. `digest2` is CASE SENSITIVE and has relatively high level of collisions - not recommended.
 
 ## Example
 
 ### Indifferent to key type
 
-    >> HashDigest.digest2(:a => 1)
-    => "1s05qdo"
-    >> HashDigest.digest2('a' => 1)
-    => "1s05qdo"
+    >> HashDigest.digest3(:a => 1)
+    => "86eda770a6060824b090dd4df091e3bd4121279c"
+    >> HashDigest.digest3('a' => 1)
+    => "86eda770a6060824b090dd4df091e3bd4121279c"
 
 ### Indifferent to key order
 
-    >> HashDigest.digest2(:a => 1, 'b' => 2)
-    => "fkqncr"
-    >> HashDigest.digest2(:b => 2, 'a' => 1)
-    => "fkqncr"
+    >> HashDigest.digest3(:a => 1, 'b' => 2)
+    => "d53cf64e768f4ef09c806bbe12258c78211b2690"
+    >> HashDigest.digest3(:b => 2, 'a' => 1)
+    => "d53cf64e768f4ef09c806bbe12258c78211b2690"
 
 ## Speed
 
 If you're **not** on JRuby, having [`EscapeUtils`](https://github.com/brianmario/escape_utils) in your `Gemfile` will make things much faster.
 
-### With EscapeUtils (MRI, doesn't work on JRuby)
-
-      HashDigest.digest2      2492 i/100ms
-    HashDigest.hexdigest      2276 i/100ms
-    
-### With stdlib's CGI (JRuby)
-
-      HashDigest.digest2       645 i/100ms
-    HashDigest.hexdigest       213 i/100ms
-
-### With stdlib's CGI (MRI)
+## Algorithm
 
-      HashDigest.digest2       531 i/100ms
-    HashDigest.hexdigest       513 i/100ms
+### digest3
 
-## Algorithm
+1. Represent the hash as a URL querystring
+2. Sort by key
+3. SHA1 hexdigest
 
-### digest2 (current)
+### digest2 (deprecated and not recommended)
 
 1. Represent the hash as a URL querystring
 2. Sort by key
-3. [MurmurHash3](http://en.wikipedia.org/wiki/MurmurHash)
+3. [MurmurHash3](http://en.wikipedia.org/wiki/MurmurHash) V32 (this turned out to have too many collisions)
 4. Convert to base 36 to save space
 
 Note: non-cryptographic and variable length. CASE SENSITIVE.

data/lib/hash_digest.rb

@@ -1,4 +1,5 @@
 require 'digest/md5'
+require 'digest/sha1'
 require 'murmurhash3'
 # see below for requiring either cgi (on JRuby) or escape_utils (if available)
 
@@ -6,17 +7,26 @@ require "hash_digest/version"
 
 module HashDigest
   # CURRENT
+  def self.as_digest3(obj)
+    obj.to_hash_digest_query
+  end
+
+  # CURRENT
+  def self.digest3(obj)
+    ::Digest::SHA1.hexdigest as_digest2(obj)
+  end
 
+  # LEGACY
   def self.as_digest2(obj)
     obj.to_hash_digest_query
   end
 
+  # LEGACY
   def self.digest2(obj)
     ::MurmurHash3::V32.str_hash(as_digest2(obj)).to_s 36
   end
 
   # LEGACY
-
   def self.as_digest1(obj)
     ordered_list = case obj
     when ::Hash
@@ -35,6 +45,7 @@ module HashDigest
     ordered_list.join '&'
   end
 
+  # LEGACY
   def self.hexdigest(obj)
     ::Digest::MD5.hexdigest as_digest1(obj)
   end

data/lib/hash_digest/version.rb

@@ -1,3 +1,3 @@
 module HashDigest
-  VERSION = "1.1.2"
+  VERSION = "1.1.3"
 end

data/test/test_hash_digest.rb

@@ -5,8 +5,8 @@ require 'stringio'
 describe HashDigest do
   include TestHelper
 
-  describe '.digest2' do
-    it "is much faster than .hexdigest" do
+  describe 'speed' do
+    it "improves as we get older" do
       hsh = { 'a' => 1, :b => { foo: :bar, zoo: 'animal' }, ";b\n`&" => { foo: :bar, zoo: "==&ani!!.;&mal\n" } }
       begin
         old_stdout = $stdout
@@ -14,6 +14,7 @@ describe HashDigest do
         Benchmark.ips do |x|
           x.report("HashDigest.hexdigest")  { HashDigest.hexdigest(hsh) }
           x.report("HashDigest.digest2")    { HashDigest.digest2(hsh)   }
+          x.report("HashDigest.digest3")    { HashDigest.digest3(hsh)   }
         end
         $stdout.rewind
         result = $stdout.read
@@ -56,11 +57,11 @@ describe HashDigest do
 
   describe 'indifference to' do
     it 'key type' do
-      HashDigest.as_digest2(:a => 1, 'b' => 2).must_equal(HashDigest.as_digest2('a' => 1, :b => 2))
+      HashDigest.as_digest3(:a => 1, 'b' => 2).must_equal(HashDigest.as_digest3('a' => 1, :b => 2))
     end
     
     it 'key order' do
-      HashDigest.as_digest2(:a => 1, 'b' => 2).must_equal(HashDigest.as_digest2(:b => 2, 'a' => 1))
+      HashDigest.as_digest3(:a => 1, 'b' => 2).must_equal(HashDigest.as_digest3(:b => 2, 'a' => 1))
     end
 
     [
@@ -72,7 +73,7 @@ describe HashDigest do
       it "trivial difference between #{a} and #{b}" do
         assert_same_as_old a
         assert_same_as_old b
-        HashDigest.as_digest2(b).must_equal(HashDigest.as_digest2(b))
+        HashDigest.as_digest3(b).must_equal(HashDigest.as_digest3(b))
       end
     end
   end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: hash_digest
 version: !ruby/object:Gem::Version
-  version: 1.1.2
+  version: 1.1.3
 platform: ruby
 authors:
 - Seamus Abshere
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-01-06 00:00:00.000000000 Z
+date: 2014-01-14 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: murmurhash3