remove emoji identifiers

Some unicode characters or character groups lead to a large increase
in parser size. This change halves the size of the generated parser
file.

Author: ChrHorn

grammar.js

@@ -894,7 +894,8 @@ module.exports = grammar({
       // Some symbols in Sm and So unicode categories that are identifiers
       const validMathSymbols = '°∀-∇∎-∑∫-∳';
 
-      const start = `[_\\p{XID_Start}${validMathSymbols}\\p{Emoji}&&[^0-9#*]]`;
+      // Emojis are currently not supported because they double the parser size
+      const start = `[_\\p{XID_Start}${validMathSymbols}&&[^0-9#*]]`;
       const rest = `[^"'\`\\s\\.\\-\\[\\]${nonIdentifierCharacters}]*`;
       return new RegExp(start + rest);
     },

test/corpus/expressions.txt

@@ -11,8 +11,6 @@ x′
 logŷ
 ϵ
 ŷ
-🙋
-🦀
 
 ---
 
@@ -27,8 +25,6 @@ logŷ
   (identifier)
   (identifier)
   (identifier)
-  (identifier)
-  (identifier)
   (identifier))