corefines 1.3.0 → 1.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: de04dfa5cd06fc172ea71b0f428f04b76b6b0527
-  data.tar.gz: 0deb05f19ce1dddd0ca60d13e8f9101dee5d49c0
+  metadata.gz: 85b1a470338e303705ff7caf0c4cfb7dd36c4cb6
+  data.tar.gz: 5fbaac4aa91cd5b8d36ed2889bb1296deb5d2b6f
 SHA512:
-  metadata.gz: 078f83c0f36e11534267b88cb445aad00f8483c6ddf9e1e3c2f804044d09f619655b75560038bc2f99cafccf577f0468ea0e32fca166069ed2c5ff7bcb78e5e2
-  data.tar.gz: 366f0084b4bac8eec6666fc40fb6091896be88c1d3d90f9c59047227e354bcd99ce82b991a26935639c35f1b483e59471ea0a717492dd94918d281846cbc72a5
+  metadata.gz: 3f2dc602df43dcfdcdadc87d9bcd551f59a9e9836fe0ded9609dfeb1ed1c33fbba0afeb09dea1259675769e959e23fc80f8f65200736a3f3a470a8b8b47832d5
+  data.tar.gz: ba274790cd0239df3761c3d231d92e9acf0d6158e96889d4b7016349e0c4ac242194c10532391f4ff11999bf5d7d23d3f42087bee4a6cc10e56d3e1ad621e7ea

data/CHANGELOG.adoc

@@ -3,6 +3,12 @@
 :doc-base-url: http://www.rubydoc.info/github/jirutka/corefines/Corefines
 :issue-uri: {repo-uri}/issues
 
+
+== 1.4.0 (2015-05-03)
+
+* Add new refinement {doc-base-url}/String/ForceUTF8[String#force_utf8].
+
+
 == 1.3.0 (2015-04-29)
 
 * Add new refinement {doc-base-url}/String/ToRegexp[String#to_regexp].

data/README.adoc

@@ -41,12 +41,12 @@ TODO
 Add this line to your application’s Gemfile:
 
 [source]
-gem 'corefines', '~> 1.3'
+gem 'corefines', '~> 1.4'
 
 or to your gemspec:
 
 [source]
-s.add_runtime_dependency 'corefines', '~> 1.3'
+s.add_runtime_dependency 'corefines', '~> 1.4'
 
 and then execute:
 
@@ -165,6 +165,8 @@ Not ideal indeed, but probably the best of what we can achieve.
 ** {doc-base-url}/String/Color[#color]
 ** {doc-base-url}/String/Concat[#concat!]
 ** {doc-base-url}/String/Decolor[#decolor]
+** {doc-base-url}/String/ForceUTF8[#force_utf8]
+** {doc-base-url}/String/ForceUTF8[#force_utf8!]
 ** {doc-base-url}/String/Indent[#indent]
 ** {doc-base-url}/String/RelativePathFrom[#relative_path_from]
 ** {doc-base-url}/String/Remove[#remove]

data/lib/corefines/string.rb

@@ -127,6 +127,56 @@ module Corefines
       end
     end
 
+    ##
+    # @!method force_utf8
+    #   Returns a copy of _str_ with encoding changed to UTF-8 and all invalid
+    #   byte sequences replaced with the Unicode Replacement Character (U+FFFD).
+    #
+    #   If _str_ responds to +#scrub!+ (Ruby >=2.1), then it's used for
+    #   replacing invalid bytes. Otherwise a simple custom implementation is
+    #   used (may not return the same result as +#scrub!+).
+    #
+    #   @return [String] a valid UTF-8 string.
+    #
+    # @!method force_utf8!
+    #   Changes the encoding to UTF-8, replaces all invalid byte sequences with
+    #   the Unicode Replacement Character (U+FFFD) and returns self.
+    #   This is same as {#force_utf8}, except it indents the receiver in-place.
+    #
+    #   @return (see #force_utf8)
+    #
+    module ForceUTF8
+      refine ::String do
+        def force_utf8
+          dup.force_utf8!
+        end
+
+        def force_utf8!
+          str = force_encoding(Encoding::UTF_8)
+
+          if str.respond_to? :scrub!
+            str.scrub!
+
+          else
+            result = ''.force_encoding('BINARY')
+            invalid = false
+
+            str.chars.each do |c|
+              if c.valid_encoding?
+                result << c
+                invalid = false
+              elsif !invalid
+                result << "\uFFFD"
+                invalid = true
+              end
+            end
+
+            replace result.force_encoding(Encoding::UTF_8)
+          end
+        end
+      end
+    end
+
     ##
     # @!method indent(amount, indent_str = nil, indent_empty_lines = false)
     #   Returns an indented copy of this string.
@@ -365,6 +415,7 @@ module Corefines
 
     class << self
       alias_method :concat!, :concat
+      alias_method :force_utf8!, :force_utf8
       alias_method :indent!, :indent
       alias_method :remove!, :remove
     end

data/lib/corefines/version.rb

@@ -1,3 +1,3 @@
 module Corefines
-  VERSION = '1.3.0'
+  VERSION = '1.4.0'
 end

data/spec/string/force_utf8_spec.rb

@@ -0,0 +1,68 @@
+# coding: utf-8
+
+describe String do
+  using Corefines::String::force_utf8
+
+  describe '#force_utf8!' do
+
+    context "string encoded in ISO-8859-1" do
+      subject(:str) { 'foo'.encode(Encoding::ISO_8859_1) }
+
+      it "sets encoding to UTF-8" do
+        expect( str.force_utf8!.encoding ).to eql Encoding::UTF_8
+        expect( str.encoding ).to eql Encoding::UTF_8
+      end
+    end
+
+    context "string with bad bytes" do
+      {
+        "\xE3llons-y!" => "�llons-y!",
+        "Al\xE9ons-y!" => "Al�ons-y!",
+        "Al\xE9ons-y\x80" => "Al�ons-y�"
+      }
+      .each do |input, expected|
+        context input.inspect do
+
+          subject(:str) { input.dup }  # defreeze!
+
+          it "replaces bad bytes with the replacement char" do
+            expect( str.force_utf8! ).to eq expected
+            expect( str ).to eq expected
+          end
+
+          it "produces a valid UTF-8 string" do
+            str.force_utf8!
+            expect( str.valid_encoding? ).to be true
+            expect( str.encoding ).to eql Encoding::UTF_8
+          end
+        end
+      end
+
+      it "collapses multiple consecutive bad bytes into one replacement" do
+        str = "abc\u3042\xE3\x80"
+        expect( str.force_utf8! ).to eq "abc\u3042�"
+        expect( str ).to eq "abc\u3042�"
+        expect( str.valid_encoding? ).to be true
+      end
+    end
+  end
+
+
+  # force_utf8 just duplicates the string and calls force_utf8!, so there's no
+  # need to test it thoroughly.
+  describe '#force_utf8' do
+
+    it "replaces bad bytes with the replacement char" do
+      expect( "Al\xE9ons-y\x80".force_utf8 ).to eq "Al�ons-y�"
+    end
+
+    it "returns a copy of the string and keeps original unaffected" do
+      str = '\xE3llons-y!'.encode(Encoding::ISO_8859_1)
+      copy = str.dup
+      str.force_utf8
+
+      expect( str ).to eql copy
+      expect( str.encoding ).to eql Encoding::ISO_8859_1
+    end
+  end
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: corefines
 version: !ruby/object:Gem::Version
-  version: 1.3.0
+  version: 1.4.0
 platform: ruby
 authors:
 - Jakub Jirutka
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2015-04-28 00:00:00.000000000 Z
+date: 2015-05-02 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: asciidoctor
@@ -144,6 +144,7 @@ files:
 - spec/string/color_spec.rb
 - spec/string/concat_spec.rb
 - spec/string/decolor_spec.rb
+- spec/string/force_utf8_spec.rb
 - spec/string/indent_spec.rb
 - spec/string/relative_path_from_spec.rb
 - spec/string/remove_spec.rb