gigo 1.3.0 → 1.4.0

data/gemfiles/activesupport30.gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: /Users/kencollins/Repositories/customink/gigo
   specs:
-    gigo (1.2.0)
+    gigo (1.4.0)
       activesupport (>= 3.0)
       ensure_valid_encoding (~> 0.5.3)
 

data/gemfiles/activesupport31.gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: /Users/kencollins/Repositories/customink/gigo
   specs:
-    gigo (1.2.0)
+    gigo (1.4.0)
       activesupport (>= 3.0)
       ensure_valid_encoding (~> 0.5.3)
 

data/gemfiles/activesupport32.gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: /Users/kencollins/Repositories/customink/gigo
   specs:
-    gigo (1.2.0)
+    gigo (1.4.0)
       activesupport (>= 3.0)
       ensure_valid_encoding (~> 0.5.3)
 

data/gemfiles/activesupport40.gemfile.lock

@@ -12,7 +12,7 @@ GIT
 PATH
   remote: /Users/kencollins/Repositories/customink/gigo
   specs:
-    gigo (1.2.0)
+    gigo (1.4.0)
       activesupport (>= 3.0)
       ensure_valid_encoding (~> 0.5.3)
 

data/lib/gigo/transcoders/rchardet.rb

@@ -14,7 +14,7 @@ module GIGO
       private
 
       def self.detect_encoding(data)
-        CharDet.detect(data.dup)['encoding']
+        GIGO::CharDet.detect(data.dup)['encoding']
       end
       
     end

data/lib/gigo/version.rb

@@ -1,3 +1,3 @@
 module GIGO
-  VERSION = "1.3.0"
+  VERSION = "1.4.0"
 end

data/test/cases/gigo_test.rb

@@ -11,6 +11,10 @@ module GIGO
     let(:data_bin_apos)     { "won\x92t".force_encoding('binary') }
     let(:data_really_bad)   { "ed.Ã\u0083Ã\u0083\xC3" }
 
+    let(:data_medico_utf8)     { "Med\u00EDco".force_encoding('UTF-8') }
+    let(:data_medico_iso88591) { "Med\xEDco".force_encoding('iso8859-1') }
+    let(:data_medico_unknown)  { "Med\uFFFDco".force_encoding('UTF-8') }
+
 
     describe '.encoding' do
 
@@ -42,7 +46,7 @@ module GIGO
         GIGO.load(data_bin_apos).must_equal "won’t"
       end
 
-      it 'should allows properly encoded and marked strings to be passed thru' do
+      it 'should allow properly encoded and marked strings to be passed thru' do
         GIGO.load(data_utf8).must_equal data_utf8
         GIGO.load(data_utf8_emoji).must_equal data_utf8_emoji
       end
@@ -52,7 +56,11 @@ module GIGO
       end
 
       it 'allows really bad data to be encoded using default replace and question marks' do
-        GIGO.load(data_utf8_emoji.force_encoding('ASCII-8BIT')).must_equal data_utf8_emoji
+        GIGO.load(data_medico_unknown).must_equal "Med�co"
+      end
+
+      it 'makes sure UTF-8 data read in as US-ASCII us fixed' do
+        GIGO.load(data_medico_utf8.force_encoding('US-ASCII')).must_equal 'Medíco'
       end
 
       it 'converts windows codepages that are poorly marked as another encoding' do
@@ -66,7 +74,12 @@ module GIGO
         GIGO.load(db_data4).must_equal data_utf8
       end
 
-      it 'can make sure to it is really a valid encoding afteward' do
+      it 'converts iso8859 when poorly marked as another encoding' do
+        GIGO.load(data_medico_iso88591).must_equal 'Medíco'
+        GIGO.load(data_medico_iso88591.force_encoding('US-ASCII')).must_equal 'Medíco'
+      end
+
+      it 'can make sure to it is really a valid encoding afterward' do
         html_escape GIGO.load(data_really_bad)
       end
 

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: gigo
 version: !ruby/object:Gem::Version
-  version: 1.3.0
+  version: 1.4.0
   prerelease: 
 platform: ruby
 authors:
@@ -186,7 +186,7 @@ required_ruby_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -1915512776307670961
+      hash: 1872930804050042885
 required_rubygems_version: !ruby/object:Gem::Requirement
   none: false
   requirements:
@@ -195,7 +195,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
       segments:
       - 0
-      hash: -1915512776307670961
+      hash: 1872930804050042885
 requirements: []
 rubyforge_project: 
 rubygems_version: 1.8.25