spec/unicode_emoji_spec.rb in unicode-emoji-3.8.0 vs spec/unicode_emoji_spec.rb in unicode-emoji-4.0.0

@@ -37,13 +37,24 @@
     it "matches textual singleton emoji in combination with emoji variation selector" do
       "▶\u{FE0F} play button" =~ Unicode::Emoji::REGEX
       assert_equal "▶\u{FE0F}", $&
     end
 
-    it "does not match singleton 'component' emoji codepoints" do
+    it "matches singleton skin tone modifiers and hair components" do
       "🏻 light skin tone" =~ Unicode::Emoji::REGEX
+      assert_equal "🏻", $&
+
+      "🦰 emoji component red hair" =~ Unicode::Emoji::REGEX
+      assert_equal "🦰", $&
+    end
+
+    it "does not match singleton components that are not skin tone modifiers or hair components" do
+      "1 digit one" =~ Unicode::Emoji::REGEX
       assert_nil $&
+
+      "🇦 regional indicator symbol letter a" =~ Unicode::Emoji::REGEX
+      assert_nil $&
     end
 
     it "matches modified emoji if modifier base emoji is used" do
       "🛌🏽 person in bed: medium skin tone" =~ Unicode::Emoji::REGEX
       assert_equal "🛌🏽", $&
@@ -92,16 +103,16 @@
       assert_equal "🤾🏽‍♀️", $&
     end
 
     it "does not match MQE zwj sequences" do
       "🤾🏽‍♀ woman playing handball: medium skin tone, missing VS16" =~ Unicode::Emoji::REGEX
-      refute_equal `🤾🏽‍♀`, $&
+      refute_equal "🤾🏽‍♀", $&
     end
 
     it "does not match UQE emoji" do
       "🏌‍♂️ man golfing, missing VS16" =~ Unicode::Emoji::REGEX
-      refute_equal `🏌‍♂️`, $&
+      refute_equal "🏌‍♂️", $&
     end
 
     it "does not match valid zwj sequences that are not recommended" do
       "🤠‍🤢 vomiting cowboy" =~ Unicode::Emoji::REGEX
       assert_equal "🤠", $&
@@ -151,28 +162,28 @@
   end
 
   describe "REGEX_INCLUDE_MQE" do
     it "matches MQE emoji" do
       "🤾🏽‍♀ woman playing handball: medium skin tone, missing VS16" =~ Unicode::Emoji::REGEX_INCLUDE_MQE
-      assert_equal `🤾🏽‍♀`, $&
+      assert_equal "🤾🏽‍♀", $&
     end
 
     it "does not match UQE emoji" do
       "🏌‍♂️ man golfing, missing VS16" =~ Unicode::Emoji::REGEX_INCLUDE_MQE
-      refute_equal `🏌‍♂️`, $&
+      refute_equal "🏌‍♂️", $&
     end
   end
 
   describe "REGEX_INCLUDE_MQE_UQE" do
     it "matches MQE emoji" do
       "🤾🏽‍♀ woman playing handball: medium skin tone, missing VS16" =~ Unicode::Emoji::REGEX_INCLUDE_MQE_UQE
-      assert_equal `🤾🏽‍♀`, $&
+      assert_equal "🤾🏽‍♀", $&
     end
 
     it "matches UQE emoji" do
       "🏌‍♂️ man golfing, missing VS16" =~ Unicode::Emoji::REGEX_INCLUDE_MQE_UQE
-      assert_equal `🏌‍♂️`, $&
+      assert_equal "🏌‍♂️", $&
     end
   end
 
   describe "REGEX_VALID" do
     it "matches most singleton emoji codepoints" do
@@ -198,13 +209,24 @@
     it "matches textual singleton emoji in combination with emoji variation selector" do
       "▶\u{FE0F} play button" =~ Unicode::Emoji::REGEX_VALID
       assert_equal "▶\u{FE0F}", $&
     end
 
-    it "does not match singleton 'component' emoji codepoints" do
+    it "matches singleton skin tone modifiers and hair components" do
       "🏻 light skin tone" =~ Unicode::Emoji::REGEX_VALID
+      assert_equal "🏻", $&
+
+      "🦰 emoji component red hair" =~ Unicode::Emoji::REGEX_VALID
+      assert_equal "🦰", $&
+    end
+
+    it "does not match singleton components that are not skin tone modifiers or hair components" do
+      "1 digit one" =~ Unicode::Emoji::REGEX_VALID
       assert_nil $&
+
+      "🇦 regional indicator symbol letter a" =~ Unicode::Emoji::REGEX_VALID
+      assert_nil $&
     end
 
     it "matches modified emoji if modifier base emoji is used" do
       "🛌🏽 person in bed: medium skin tone" =~ Unicode::Emoji::REGEX_VALID
       assert_equal "🛌🏽", $&
@@ -293,13 +315,24 @@
     it "matches textual singleton emoji in combination with emoji variation selector" do
       "▶\u{FE0F} play button" =~ Unicode::Emoji::REGEX_WELL_FORMED
       assert_equal "▶\u{FE0F}", $&
     end
 
-    it "does not match singleton 'component' emoji codepoints" do
+    it "matches singleton skin tone modifiers and hair components" do
       "🏻 light skin tone" =~ Unicode::Emoji::REGEX_WELL_FORMED
+      assert_equal "🏻", $&
+
+      "🦰 emoji component red hair" =~ Unicode::Emoji::REGEX_WELL_FORMED
+      assert_equal "🦰", $&
+    end
+
+    it "does not match singleton components that are not skin tone modifiers or hair components" do
+      "1 digit one" =~ Unicode::Emoji::REGEX_WELL_FORMED
       assert_nil $&
+
+      "🇦 regional indicator symbol letter a" =~ Unicode::Emoji::REGEX_WELL_FORMED
+      assert_nil $&
     end
 
     it "matches modified emoji if modifier base emoji is used" do
       "🛌🏽 person in bed: medium skin tone" =~ Unicode::Emoji::REGEX_WELL_FORMED
       assert_equal "🛌🏽", $&
@@ -393,15 +426,26 @@
     it "matches textual singleton emoji in combination with emoji variation selector" do
       "▶\u{FE0F} play button" =~ Unicode::Emoji::REGEX_POSSIBLE
       assert_equal "▶\u{FE0F}", $&
     end
 
-    it "matches singleton 'component' emoji codepoints" do
+    it "matches singleton skin tone modifiers and hair components" do
       "🏻 light skin tone" =~ Unicode::Emoji::REGEX_POSSIBLE
       assert_equal "🏻", $&
+
+      "🦰 emoji component red hair" =~ Unicode::Emoji::REGEX_POSSIBLE
+      assert_equal "🦰", $&
     end
 
+    it "matches singleton components that are not skin tone modifiers or hair components" do
+      "1 digit one" =~ Unicode::Emoji::REGEX_POSSIBLE
+      assert_equal "1", $&
+
+      "🇦 regional indicator symbol letter a" =~ Unicode::Emoji::REGEX_POSSIBLE
+      assert_equal "🇦", $&
+    end
+
     it "matches modified emoji if modifier base emoji is used" do
       "🛌🏽 person in bed: medium skin tone" =~ Unicode::Emoji::REGEX_POSSIBLE
       assert_equal "🛌🏽", $&
     end
 
@@ -484,22 +528,33 @@
       "😴\u{FE0E} sleeping face" =~ Unicode::Emoji::REGEX_BASIC
       assert_nil $&
     end
 
     it "does not match textual singleton emoji" do
-      "▶ play button" =~ Unicode::Emoji::REGEX
+      "▶ play button" =~ Unicode::Emoji::REGEX_BASIC
       assert_nil $&
     end
 
     it "matches textual singleton emoji in combination with emoji variation selector" do
-      "▶\u{FE0F} play button" =~ Unicode::Emoji::REGEX
+      "▶\u{FE0F} play button" =~ Unicode::Emoji::REGEX_BASIC
       assert_equal "▶\u{FE0F}", $&
     end
 
-    it "does not match singleton 'component' emoji codepoints" do
+    it "matches singleton skin tone modifiers and hair components" do
       "🏻 light skin tone" =~ Unicode::Emoji::REGEX_BASIC
+      assert_equal "🏻", $&
+
+      "🦰 emoji component red hair" =~ Unicode::Emoji::REGEX_BASIC
+      assert_equal "🦰", $&
+    end
+
+    it "does not match singleton components that are not skin tone modifiers or hair components" do
+      "1 digit one" =~ Unicode::Emoji::REGEX_BASIC
       assert_nil $&
+
+      "🇦 regional indicator symbol letter a" =~ Unicode::Emoji::REGEX_BASIC
+      assert_nil $&
     end
 
     it "does not match modified emoji" do
       "🛌🏽 person in bed: medium skin tone" =~ Unicode::Emoji::REGEX_BASIC
       assert_equal "🛌", $&
@@ -555,15 +610,26 @@
     it "does not match textual singleton emoji in combination with emoji modifiers" do
       "✌🏻 victory hand" =~ Unicode::Emoji::REGEX_TEXT
       assert_nil $&
     end
 
-    it "does not match singleton 'component' emoji codepoints" do
+    it "does not match singleton skin tone modifiers and hair components" do
       "🏻 light skin tone" =~ Unicode::Emoji::REGEX_TEXT
       assert_nil $&
+
+      "🦰 emoji component red hair" =~ Unicode::Emoji::REGEX_TEXT
+      assert_nil $&
     end
 
+    it "does not match singleton components that are not skin tone modifiers or hair components" do
+      "1 digit one" =~ Unicode::Emoji::REGEX_TEXT
+      assert_nil $&
+
+      "🇦 regional indicator symbol letter a" =~ Unicode::Emoji::REGEX_TEXT
+      assert_nil $&
+    end
+
     it "does not match modified emoji" do
       "🛌🏽 person in bed: medium skin tone" =~ Unicode::Emoji::REGEX_TEXT
       assert_nil $&
     end
 
@@ -594,13 +660,28 @@
       "🤾🏽‍♀️ woman playing handball: medium skin tone" =~ Unicode::Emoji::REGEX_TEXT
       assert_nil $&
     end
   end
 
-  describe "REGEX_ANY" do
+  describe "REGEX_PROP_EMOJI" do
     it "returns any emoji-related codepoint (but no variation selectors or tags)" do
-      matches = "1 string 😴\u{FE0F} sleeping face with 🇵 and modifier 🏾, also 🏴󠁧󠁢󠁳󠁣󠁴󠁿 Scotland".scan(Unicode::Emoji::REGEX_ANY)
+      matches = "1 string 😴\u{FE0F} sleeping face with 🇵 and modifier 🏾, also 🏴󠁧󠁢󠁳󠁣󠁴󠁿 Scotland".scan(Unicode::Emoji::REGEX_PROP_EMOJI)
       assert_equal ["1", "😴", "🇵", "🏾", "🏴"], matches
+    end
+  end
+
+  describe "REGEX_EMOJI_KEYCAP" do
+    it "matches emoji keycap sequences" do
+      "2️⃣ keycap: 2" =~ Unicode::Emoji::REGEX_EMOJI_KEYCAP
+      assert_equal "2️⃣", $&
+    end
+
+    it "does not match non-emoji keycap sequences" do
+      "8⃣ text keycap: 8" =~ Unicode::Emoji::REGEX_EMOJI_KEYCAP
+      assert_nil $&
+
+      "#⃣ text keycap: #" =~ Unicode::Emoji::REGEX_EMOJI_KEYCAP
+      assert_nil $&
     end
   end
 
   describe "REGEX_PICTO" do
     it "matches codepoints with Extended_Pictograph property (almost all emoji are, but also others)" do