Antlr4语法修复：适配含空格值列表的内部语言解析需求

Got it, the problem here is your current ANTLR4 grammar can't handle list items or identifiers that include spaces—like Mechanical Instability in your $IF clause or Type grammar Trigger after $FROM. Let's walk through why this breaks and how to fix it:

Why Your Current Grammar Fails #

Your existing rules have two key limitations:

• The IDENTIFIER rule only allows [a-zA-Z_.'0-9]+, which excludes spaces. This means phrases with spaces get split into multiple IDENTIFIER tokens instead of being treated as a single value.
• LIST_VALUES is built on IDENTIFIER, so it can't recognize list items with spaces either.

Modified ANTLR4 Grammar #

Here's the updated grammar that fixes these issues, while preserving all your original language features:

prog: expr EOF;

expr : IF (ANY|EMPTY|NOT|comparator|count|complex|count_complex|LIST_ITEM|LIST_VALUES) FROM IDENTIFIER_WITH_SPACES THEN 
     | expr binary expr 
     | LPAREN expr RPAREN 
     ;

comparator : (GT | GE | LT | LE | EQ) (INT|IDENTIFIER_WITH_SPACES);
binary : AND | OR;
count : COUNT LPAREN (IDENTIFIER_WITH_SPACES | DISTINCT) RPAREN comparator;
complex : COMPLEX LPAREN expr RPAREN;
count_complex : COUNT_COMPLEX LPAREN (expr | DISTINCT IDENTIFIER_WITH_SPACES expr) RPAREN comparator;

// Keyword rules (higher priority than text tokens)
IF : '$IF';
FROM : '$FROM';
THEN : '$THEN';
AND : '$AND' ;
ANY : '$ANY';
EMPTY : '$EMPTY';
DISTINCT : '$DISTINCT';
COUNT : '$COUNT';
COUNT_COMPLEX : '$COUNT_COMPLEX';
COMPLEX : '$COMPLEX';
OR : '$OR' ;
NOT : '$NOT' LPAREN LIST_ITEM (',' LIST_ITEM)*? RPAREN;

// Comparison operators
GT : '>' ;
GE : '>=' ;
LT : '<' ;
LE : '<=' ;
EQ : '=' ;

// Punctuation
LPAREN : '(' ;
RPAREN : ')' ;

// Numeric values
INT : '-'?[0-9]+;

// Identifiers that can include spaces (for $FROM targets)
IDENTIFIER_WITH_SPACES : ~[$ \r\t\u000C\n]+ ( [ ] ~[$ \r\t\u000C\n]+ )*;

// List items that support spaces (stops at commas or $-keywords)
LIST_ITEM : ~[,$ \r\t\u000C\n]+ ( [ ] ~[,$ \r\t\u000C\n]+ )*;

// Comma-separated list values
LIST_VALUES : LIST_ITEM ( ',' LIST_ITEM )*;

// Skip whitespace
WS : [ \r\t\u000C\n]+ -> skip;

Key Changes Explained #

• IDENTIFIER_WITH_SPACES: Replaces your original IDENTIFIER for $FROM targets. It matches text with spaces, but stops at $ to avoid swallowing keywords like $THEN.
• LIST_ITEM: Designed to match list entries with spaces, stopping at commas or $-prefixed keywords. This lets Mechanical Instability be treated as a single list item.
• Updated LIST_VALUES: Now uses LIST_ITEM instead of IDENTIFIER, so it correctly parses comma-separated entries with spaces.
• Adjusted NOT Rule: Swapped CHAR for LIST_ITEM to let $NOT handle values with spaces too (if your language requires it).
• Keyword Priority: All $-prefixed keywords are defined before text tokens, ensuring ANTLR4 prioritizes them over longer text matches (so $FROM isn't mistaken for part of an identifier).

Test with Your Example #

Your sample statement $IF Mechanical Instability,Deformity $FROM Type grammar Trigger $THEN will now parse correctly:

• Mechanical Instability and Deformity are recognized as two distinct LIST_ITEM tokens in a LIST_VALUES
• Type grammar Trigger is parsed as a single IDENTIFIER_WITH_SPACES

内容的提问来源于stack exchange，提问作者IuryRibeiro