Ignore some part of input when parsing with ANTLR

2020-07-27 03:23发布

问题:

I'm trying to parse a language by ANTLR (ANTLRWorks-3.5.2). The goal is to enter complete input but Antlr gives a parse tree of defined parts in grammar and ignore the rest of inputs, for example this is my grammar :

grammar asap;
project : '/begin PROJECT' name  module+ '/end PROJECT';    
module : '/begin MODULE'name '/end MODULE'; 
name :   IDENT ;              
IDENT :     ('a'..'z'|'A'..'Z')('a'..'z'|'A'..'Z'|'0'..'9'|'_'|'.'|':'|'-')*;

Given input:

/begin PROJECT HybridSailboat_2 
/begin MODULE engine

/begin A2ML
/include XCP_common_v1_0.aml
"XCP" struct {      
taggedstruct Common_Parameters ;
};
/end A2ML

/end MODULE
/end PROJECT

regarding to this input I just want the parse tree contains project and module and not A2ML part. Is it possible in antlr that it ignore some part of inputs? Can I specify start and end points of unimportant parts in grammar?

回答1:

Simply match the A2ML part as a single token in the lexer and skip() it:

grammar asap;

project
 : BEGIN_PROJECT name module* END_PROJECT EOF
 ;

module
 : BEGIN_MODULE name END_MODULE
 ;

name
 :   IDENT
 ;  

IDENT
 : ('a'..'z'|'A'..'Z') ('a'..'z'|'A'..'Z'|'0'..'9'|'_'|'.'|':'|'-')*
 ;

BEGIN_PROJECT
 : '/begin' S 'PROJECT'
 ;

END_PROJECT
 : '/end' S 'PROJECT'
 ;

BEGIN_MODULE
 : '/begin' S 'MODULE'
 ;

END_MODULE
 : '/end' S 'MODULE'
 ;

A2ML
 : '/begin' S 'A2ML' .* '/end' S 'A2ML' {skip();}
 ;

SPACES
 : S {skip();}
 ;

fragment S
 : (' ' | '\t' | '\r' | '\n')+
 ;