string_cleaner 0.2.0 → 0.2.1

data/.gitignore

@@ -3,4 +3,5 @@
 coverage
 rdoc
 pkg
-*.gem
+*.gem
+Gemfile.lock

data/README.rdoc

@@ -18,9 +18,15 @@ Ruby 1.9+ has native support for unicode and specs are 100% passing.
 
 Because Ruby 1.8.x has no native support for Unicode, you must install oniguruma and the jasherai-oniguruma gem.
 
+For example, using homebrew you would do:
+  
+    brew install oniguruma
+    bundle config build.jasherai-oniguruma --with-onig-dir=`brew --prefix oniguruma`
+    bundle install
+
 == Example usage
 
-  "\210\004".clean # => " "
+    "\210\004".clean # => " "
 
 == Copyright
 

data/Rakefile

@@ -1,3 +1,5 @@
+require 'rake/tasklib'
+require 'rake/rdoctask'
 require 'rubygems'
 begin
   require 'bundler/setup'

data/lib/string_cleaner.rb

@@ -14,9 +14,8 @@ module String::Cleaner
       unless utf8.valid_encoding? # if invalid UTF-8
         utf8 = utf8.force_encoding("ISO8859-15")
         utf8.encode!("UTF-8", :invalid => :replace, :undef => :replace, :replace => "")
-        utf8.gsub!("\xC2\x80", "€") # special case for euro sign from Windows-1252
-        utf8.force_encoding("UTF-8")
       end
+      utf8.gsub!("\u0080", "€") # special case for euro sign from Windows-1252
       utf8
     else
       require "iconv"
@@ -25,7 +24,7 @@ module String::Cleaner
         Iconv.new("UTF-8", "UTF-8").iconv(utf8)
       rescue
         utf8.gsub!(/\x80/n, "\xA4")
-        Iconv.new("UTF-8//IGNORE", "ISO8859-15").iconv(utf8)
+        Iconv.new("UTF-8//IGNORE", "ISO8859-1").iconv(utf8).gsub("¤", "€")
       end
     end
   end
@@ -67,7 +66,7 @@ module String::Cleaner
   end
 
   def to_permalink(separator="-")
-    fix_endlines.to_ascii(chartable).downcase.gsub(/[^a-z0-9]+/, separator).trim(separator)
+    clean.to_ascii(chartable).downcase.gsub(/[^a-z0-9]+/, separator).trim(separator)
   end
 
   def nl2br

data/spec/string_cleaner_spec.rb

@@ -141,7 +141,7 @@ describe String::Cleaner do
         @output = @input.clean
       end
       it "should wipe out the control characters" do
-        @output.should == "          \n  \n                   !\"\#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~ €                                ¡¢£€¥Š§š©ª«¬ ®¯°±²³Žµ¶·ž¹º»ŒœŸ¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿ"
+        @output.should == "          \n  \n                   !\"\#$%&'()*+,-./0123456789:;<=>?@ABCDEFGHIJKLMNOPQRSTUVWXYZ[\\]^_`abcdefghijklmnopqrstuvwxyz{|}~ €                                ¡¢£€¥¦§¨©ª«¬ ®¯°±²³´µ¶·¸¹º»¼½¾¿ÀÁÂÃÄÅÆÇÈÉÊËÌÍÎÏÐÑÒÓÔÕÖ×ØÙÚÛÜÝÞßàáâãäåæçèéêëìíîïðñòóôõö÷øùúûüýþÿ"
       end
     end
     describe "with invalid UTF-8 sequence" do

data/string_cleaner.gemspec

@@ -2,7 +2,7 @@
 
 Gem::Specification.new do |s|
   s.name = %q{string_cleaner}
-  s.version = "0.2.0"
+  s.version = "0.2.1"
 
   s.required_rubygems_version = Gem::Requirement.new(">= 0") if s.respond_to? :required_rubygems_version=
   s.authors = ["Joseph Halter"]

metadata

@@ -5,8 +5,8 @@ version: !ruby/object:Gem::Version
   segments: 
   - 0
   - 2
-  - 0
-  version: 0.2.0
+  - 1
+  version: 0.2.1
 platform: ruby
 authors: 
 - Joseph Halter