simhash 0.2.4 → 0.2.5

data/lib/simhash.rb

@@ -14,7 +14,11 @@ end
 
 module Simhash  
   DEFAULT_STRING_HASH_METHOD = String.public_instance_methods.include?("hash_vl") ? :hash_vl : :hash_vl_rb
-  PUNCTUATION_REGEXP = /(\s|\d|\W|\302\240| *— *|[«»…\-–—]| )+/u
+  PUNCTUATION_REGEXP = if RUBY_VERSION >= "1.9"
+    /(\s|\d|[^\p{L}]|\302\240| *— *|[«»…\-–—]| )+/u
+  else
+    /(\s|\d|\W|\302\240| *— *|[«»…\-–—]| )+/u
+  end
   
   
   def self.hash(tokens, options={})

metadata

@@ -1,13 +1,13 @@
 --- !ruby/object:Gem::Specification 
 name: simhash
 version: !ruby/object:Gem::Version 
-  hash: 31
+  hash: 29
   prerelease: 
   segments: 
   - 0
   - 2
-  - 4
-  version: 0.2.4
+  - 5
+  version: 0.2.5
 platform: ruby
 authors: 
 - Alex Gusev
@@ -15,8 +15,7 @@ autorequire:
 bindir: bin
 cert_chain: []
 
-date: 2011-04-06 00:00:00 +04:00
-default_executable: 
+date: 2011-10-28 00:00:00 Z
 dependencies: 
 - !ruby/object:Gem::Dependency 
   name: unicode
@@ -68,7 +67,6 @@ files:
 - lib/string.rb
 - ext/string_hashing/extconf.rb
 - ext/string_hashing/string_hashing.c
-has_rdoc: true
 homepage: http://github.com/bookmate/simhash
 licenses: []
 
@@ -98,7 +96,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
 requirements: []
 
 rubyforge_project: simhash
-rubygems_version: 1.6.2
+rubygems_version: 1.8.10
 signing_key: 
 specification_version: 3
 summary: "Gives you possbility to convert string into simhashes to futher use: finding near-duplicates, similar strings, etc."