ANTLR入门遇变量声明规则匹配错误求助：mismatched input 'var'问题排查

Hey there! Let's break down why your var test; input is throwing that confusing mismatched input error, and fix it step by step.

The Root Cause #

That error message looks contradictory at first ("expecting 'var' but got 'var'")—but the problem is happening before the parser even gets to work. ANTLR's tokenizer is misclassifying your var keyword as a STRING instead of the VAR token your parser expects.

Here's why:

• Your STRING rule is defined as [a-zA-Z]+—this matches any sequence of letters, including keywords like var and if.
• In ANTLR 4, when multiple token rules can match the same input, the first defined rule takes priority. Since STRING comes before VAR and IF in your grammar, those keywords get swallowed up as STRING tokens instead of their dedicated keyword tokens.
• On top of that, your NAME rule is also [a-zA-Z]+, which creates another conflict with both STRING and your keywords.

Step-by-Step Fixes #

1. Fix the STRING Token Definition #

First, strings should be enclosed in quotes (this is standard for most languages). Update your STRING rule to match quoted text instead of plain letters:

STRING: '"' [a-zA-Z0-9_]* '"' | '\'' [a-zA-Z0-9_]* '\'';

This will match strings like "test" or 'hello_world' instead of conflicting with identifiers and keywords.

2. Reorder Token Rules #

Move all keyword tokens (like VAR, IF, BOOLEAN) before your NAME rule. ANTLR will prioritize these exact matches over the general NAME rule:

/* Tokens - reordered to prioritize keywords first */
VAR : 'var';
IF: 'if';
BOOLEAN: 'true' | 'false';
NAME : [a-zA-Z_][a-zA-Z0-9_]*; // Updated to allow underscores and start with letter/underscore (standard for identifiers)
NUM : '0' | '-'?[1-9][0-9]*;
STRING: '"' [a-zA-Z0-9_]* '"' | '\'' [a-zA-Z0-9_]* '\'';
SEMICOLON : ';';
LPAREN: '(';
RPAREN: ')';
LCURLY: '{';
RCURLY: '}';
EQUALITY: '==' | '<' | '>' | '<=' | '>=' | '!=' ;
operation: '+' | '-' | '*' | '/';
operand: NUM | NAME; // Added NAME as a valid operand (makes sense for expressions)
WS : [ \t\r\n]+ -> skip;

3. Clean Up Ambiguous Rules #

Your term rule has an empty alternative (term: | ...), which can lead to parsing ambiguity. If you want to allow optional terms, use a ? modifier instead:

term: (NUM EQUALITY NUM | NAME EQUALITY NUM | NUM EQUALITY NAME | NAME EQUALITY NAME)?;

Or remove the empty option entirely if terms should always have a comparison.

Full Fixed Grammar #

Here's the complete corrected grammar:

grammar treetwo;

program : (declaration | statement)+ EOF;
declaration : variable_declaration | variable_assignment ;
statement: expression | ifstmnt ;
variable_declaration: VAR NAME SEMICOLON ;
variable_assignment: NAME '=' NUM SEMICOLON | NAME '=' STRING SEMICOLON | NAME '=' BOOLEAN SEMICOLON ;
expression: operand operation operand SEMICOLON 
          | expression operation expression SEMICOLON 
          | operand operation expression SEMICOLON 
          | expression operation operand SEMICOLON ;
ifstmnt: IF LPAREN term RPAREN LCURLY (declaration | statement)+ RCURLY ;
term: (NUM EQUALITY NUM | NAME EQUALITY NUM | NUM EQUALITY NAME | NAME EQUALITY NAME)?;

/* Tokens - reordered and fixed */
VAR : 'var';
IF: 'if';
BOOLEAN: 'true' | 'false';
NAME : [a-zA-Z_][a-zA-Z0-9_]*;
NUM : '0' | '-'?[1-9][0-9]*;
STRING: '"' [a-zA-Z0-9_]* '"' | '\'' [a-zA-Z0-9_]* '\'';
SEMICOLON : ';';
LPAREN: '(';
RPAREN: ')';
LCURLY: '{';
RCURLY: '}';
EQUALITY: '==' | '<' | '>' | '<=' | '>=' | '!=' ;
operation: '+' | '-' | '*' | '/';
operand: NUM | NAME;
WS : [ \t\r\n]+ -> skip;

Testing the Fix #

Now when you input var test;, the tokenizer will correctly identify var as the VAR token, and test as a NAME token. The parser will match this to your variable_declaration rule without errors.

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