On Jan 31, 2010, at 2:32 PM, Kay Röpke wrote:
> i always try to push as much code into separate classes as possible, if only
> for the reason that having it in actions makes it really hard to use an IDE.
> plus it makes the grammar itself much more readable, especially if the helper
> method names are appropriately named.
>
>> Other option is to put error conditions in grammar with predicates and
>> repeat syntactic tree pattern.
>
> "repeat" somehow raises a red flag on this side of the pond ;)
Yeah, decided it's a good idea to separate. I have
rule: ^( RULE r=ID .*) {BasicSemanticChecks.checkInvalidRuleDef(gtype,
fileName, $r.token);}
;
in BasicSemanticTriggers.g and then matching class:
public class BasicSemanticChecks {
protected static void checkInvalidRuleDef(int gtype, String fileName, Token
ruleID) {
if ( gtype== ANTLRParser.LEXER_GRAMMAR &&
Character.isLowerCase(ruleID.getText().charAt(0)) ) {
ErrorManager.grammarError(ErrorType.PARSER_RULES_NOT_ALLOWED,
fileName, ruleID, ruleID.getText());
}
if (
(gtype==ANTLRParser.PARSER_GRAMMAR||gtype==ANTLRParser.TREE_GRAMMAR) &&
Character.isUpperCase(ruleID.getText().charAt(0)) )
{
ErrorManager.grammarError(ErrorType.LEXER_RULES_NOT_ALLOWED,
fileName, ruleID, ruleID.getText());
}
}
// todo: get filename from stream via token?
protected static void checkInvalidRuleRef(int gtype, String fileName, Token
ruleID) {}
}
who knows? ANTLR v4 might even be readable!!! nah...why should i change my
ways after 20 years? ;)
Ter
_______________________________________________
antlr-dev mailing list
[email protected]
http://www.antlr.org/mailman/listinfo/antlr-dev
