characteristics 0.3.1 → 0.4.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: de978fd64342c2dd2b12cee0a86b0091481435fb
-  data.tar.gz: 829b05c91801308ce59a65380d6e0a4894a82ce5
+  metadata.gz: e75050c917591b71a2bec6839ba773ce8db0b0a5
+  data.tar.gz: 932dbfc7d19c456d58d59403e8e4f79462f56373
 SHA512:
-  metadata.gz: 7c1c6cc3e2f27120f9d8fed885be4947f50084eb454e318fa449f5e2c4133aa7fe8323c0255cdf7071a9b69e4087d620109f730f02fc3722787d6a6e727c48b8
-  data.tar.gz: 3946512a880937497f8173e1b2ea6086480518beddc41d6ec83f2406036055e3098a9d0836bb06006f7657c763a6d9dd3c72a1524c26692d8a8a7746a365dc36
+  metadata.gz: 492d9986683f98be2f0e268445ddecbc976a372d06854b1a2858b9556f9627ea149a2fd39866fe0ec6e0d062c715e20824eee95ba0c8e89f958e1d3964722ba0
+  data.tar.gz: d7bd3e2ae36cd69be1693d617ba859ceb66c580f8dfb0d7bc719d089e140be749e7249f219c0bc2d3087daf071c3f64cdaaa9f08804868a42dd94be0e1e79cde

data/CHANGELOG.md

@@ -1,5 +1,9 @@
 ## CHANGELOG
 
+### 0.4.0
+
+* Support Japanese Emojis (KDDI / SoftBank / DoCoMo) for Unicode
+
 ### 0.3.1
 
 * Add two Khmer blanks (U+17B4, U+17B5)

data/Gemfile.lock

@@ -1,7 +1,7 @@
 PATH
   remote: .
   specs:
-    characteristics (0.3.1)
+    characteristics (0.4.0)
       unicode-categories (~> 1.1, >= 1.1.2)
 
 GEM

data/lib/characteristics/ascii.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 class AsciiCharacteristics < Characteristics
   BLANKS = [
     0x9,

data/lib/characteristics/binary.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 class BinaryCharacteristics < Characteristics
   BLANKS = [
     0x9,

data/lib/characteristics/byte.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 class ByteCharacteristics < Characteristics
   HAS_C1 = /^(ISO-8859-|TIS-620)/
 
@@ -151,18 +153,6 @@ class ByteCharacteristics < Characteristics
     @ord = char.ord
   end
 
-  def encoding_has_c0?
-    true
-  end
-
-  def encoding_has_delete?
-    true
-  end
-
-  def encoding_has_c1?
-    !!(HAS_C1 =~ @encoding_name)
-  end
-
   def unicode?
     false
   end
@@ -196,4 +186,10 @@ class ByteCharacteristics < Characteristics
   def format?
     FORMATS[@ord] =~ @encoding_name
   end
+
+  private
+
+  def encoding_has_c1?
+    !!(HAS_C1 =~ @encoding_name)
+  end
 end

data/lib/characteristics/unicode.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require "unicode/categories"
 
 class UnicodeCharacteristics < Characteristics
@@ -111,4 +113,28 @@ class UnicodeCharacteristics < Characteristics
   def format?
     @is_valid && @category == "Cf"
   end
+
+  def kddi?
+    encoding_has_kddi? && ( @ord >= 0xE468 && @ord <= 0xE5DF ||
+                            @ord >= 0xEA80 && @ord <= 0xEB8E )
+  end
+
+  def softbank?
+    encoding_has_softbank? && ( @ord >= 0xE001 && @ord <= 0xE05A ||
+                                @ord >= 0xE101 && @ord <= 0xE15A ||
+                                @ord >= 0xE201 && @ord <= 0xE25A ||
+                                @ord >= 0xE301 && @ord <= 0xE34D ||
+                                @ord >= 0xE401 && @ord <= 0xE44C ||
+                                @ord >= 0xE501 && @ord <= 0xE53E )
+  end
+
+  def docomo?
+    encoding_has_docomo? && ( @ord >= 0xE63E && @ord <= 0xE757 )
+  end
+
+  private
+
+  def encoding_has_c1?
+    true
+  end
 end

data/lib/characteristics/version.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 class Characteristics
-  VERSION = "0.3.1".freeze
+  VERSION = "0.4.0".freeze
 end

data/lib/characteristics.rb

@@ -1,3 +1,5 @@
+# frozen_string_literal: true
+
 require_relative "characteristics/version"
 
 require_relative "characteristics/ascii"
@@ -12,7 +14,7 @@ class Characteristics
       :ascii
     when "ASCII-8BIT"
       :binary
-    when /^UTF-/
+    when /^UTF-?/
       :unicode
     when /^ISO-8859-/, /^Windows-125/, /^(IBM|CP85)/, /^mac/, 'TIS-620', 'Windows-874', /^KOI8-/
       :byte
@@ -66,4 +68,42 @@ class Characteristics
 
   def format?
   end
+
+  # private use emojis
+  def kddi?
+  end
+
+  # private use emojis
+  def softbank?
+  end
+
+  # private use emojis
+  def docomo?
+  end
+
+  private
+
+  def encoding_has_kddi?
+    @encoding_name.end_with? "KDDI"
+  end
+
+  def encoding_has_softbank?
+    @encoding_name.end_with? "SoftBank"
+  end
+
+  def encoding_has_docomo?
+    @encoding_name.end_with? "DoCoMo"
+  end
+
+  def encoding_has_c0?
+    true
+  end
+
+  def encoding_has_delete?
+    true
+  end
+
+  def encoding_has_c1?
+    false
+  end
 end

data/spec/characteristics_spec.rb

@@ -23,7 +23,7 @@ describe Characteristics do
   end
 
   describe UnicodeCharacteristics do
-    describe "UTF-*" do
+    describe "UTF*" do
       let(:encoding) { "UTF-8" }
 
       it "is valid or not" do
@@ -53,6 +53,26 @@ describe Characteristics do
         refute format? "\x21"
       end
     end
+
+    describe "Japanese Emojis" do
+      it "can be a KDDI emoji" do
+        encoding = "UTF8-KDDI"
+        assert Characteristics.create("\uE468".force_encoding(encoding)).kddi?
+        refute Characteristics.create("A".force_encoding(encoding)).kddi?
+      end
+
+      it "can be a SoftBank emoji" do
+        encoding = "UTF8-SoftBank"
+        assert Characteristics.create("\uE001".force_encoding(encoding)).softbank?
+        refute Characteristics.create("A".force_encoding(encoding)).softbank?
+      end
+
+      it "can be a DoCoMo emoji" do
+        encoding = "UTF8-DoCoMo"
+        assert Characteristics.create("\uE63E".force_encoding(encoding)).docomo?
+        refute Characteristics.create("A".force_encoding(encoding)).docomo?
+      end
+    end
   end
 
   describe BinaryCharacteristics do

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: characteristics
 version: !ruby/object:Gem::Version
-  version: 0.3.1
+  version: 0.4.0
 platform: ruby
 authors:
 - Jan Lelis