Question

I have been working on a grammar for a long time but now when I’m trying it out on a big code base I get this annoying problem. When I’m trying to parse an ID, for example “INDEX”, and I have another rule that looks for ‘INDEX’ the parser fails. I have created an example grammar for illustrating the problem:

public variable : '@' ID '=' STRING;
index : INDEX;
WS : (' '|'\r'|'\t'|'\u000C'|'\n') {Skip();};
INDEX : 'INDEX';    
ID  :   ('a'..'z'|'A'..'Z'|'_') ('a'..'z'|'A'..'Z'|'0'..'9'|'_')*;
STRING : '\'' ( ESC_SEQ | ~('\\'|'\'') )* '\'';

So when I’m trying to parse “@some = 'some'” it works fine, but “@index = 'some'” gives me {<mismatched token: [@1,1:5='index',<8>,1:1], resync=@index='some'>}

Output is an AST in C#.
Any ideas on how to solve this?

Answer 1

Could you create a variable name lexer rule?

public VARIABLENAME : '@' ID

This way the keywords would still work, but woudln’t interfere with the variable names.

The variable rule would be modified to:

public variable : VARIABLENAME '=' STRING;