[antlr-interest] about range float and stuff

Fabien Hermenier hermenierfabien at gmail.com
Fri Nov 4 07:04:49 PDT 2011


Thanks Bart, I think I have understand your approach and indeed, it 
seems beautiful and simple.
I will try your solution during the week-end.

Fabien.

Le 04/11/11 02:48, Bart Kiers a écrit :
> Hi Fabien,
>
> Handling this in the parser will make your life much harder than it 
> has to. Doing it in the lexer, you will need a bit of custom code, but 
> I'd go for something similar to this (something like it is on the WIki 
> somewhere, but can't find it...):
>
> grammar RangeDemo;
>
> @lexer::members {
>
>   java.util.Queue<Token> tokens = new java.util.LinkedList<Token>();
>
>   public void offer(int ttype, String ttext) {
>     emit(new CommonToken(ttype, ttext));
>   }
>   @Override
>   public void emit(Token t) {
>     state.token = t;
>     tokens.offer(t);
>   }
>   @Override
>   public Token nextToken() {
>     super.nextToken();
>     return tokens.isEmpty() ? Token.EOF_TOKEN : tokens.poll();
>   }
> }
>
> parse
>   :  (t=. {System.out.printf("\%-10s \%s\n", tokenNames[$t.type], 
> $t.text);})* EOF
>   ;
>
> FLOAT
>   :  INT '..'   {offer(INT, $INT.text); offer(RANGE, "..");}
>   |  OCTAL '..' {offer(OCTAL, $OCTAL.text); offer(RANGE, "..");}
>   |  '.' DIGITS
>   |  DIGITS '.' DIGITS?
>   ;
>
> RANGE
>   :  '..'
>   ;
>
> INT
>   :  '1'..'9' DIGIT*
>   |  '0'
>   ;
>
> OCTAL
>   :  '0' ('0'..'7')+
>   ;
>
> fragment DIGITS : DIGIT+;
> fragment DIGIT  : '0'..'9';
>
> SPACE
>   :  (' ' | '\t' | '\r' | '\n') {skip();}
>   ;
>
> And if you run the class:
>
> import org.antlr.runtime.*;
>
> public class Main {
>   public static void main(String[] args) throws Exception {
>     String src = "..07..8.5 1.9..02 1..3.4";
>     RangeDemoLexer lexer = new RangeDemoLexer(new ANTLRStringStream(src));
>     RangeDemoParser parser = new RangeDemoParser(new 
> CommonTokenStream(lexer));
>     System.out.println("Parsing: '" + src + "'");
>     parser.parse();
>   }
> }
>
> You'll see the following being printed to the console:
>
> Parsing: '..07..8.5 1.9..02 1..3.4'
> RANGE      ..
> OCTAL      07
> RANGE      ..
> FLOAT      8.5
> FLOAT      1.9
> RANGE      ..
> OCTAL      02
> INT        1
> RANGE      ..
> FLOAT      3.4
>
>
> Regards,
>
> Bart.
>
>
>
> On Fri, Nov 4, 2011 at 7:28 AM, Fabien Hermenier 
> <hermenierfabien at gmail.com <mailto:hermenierfabien at gmail.com>> wrote:
>
>     Hi
>
>     In an earlier version of my language, I had to parse range of integers
>     in various base. Now I want to include float. I have read
>     http://www.antlr.org/wiki/display/ANTLR3/Lexer+grammar+for+floating+point,+dot,+range,+time+specs
>     but I've still got some questions.
>
>     All the work seems to be done at the lexer level so the type of the
>     following tokens will be as example:
>     5 : DECIMAL_LITTERAL
>     07 : OCTAL_LITTERAL
>     7.5: FLOATING_POINT_LITTERAL
>     5..7 : DOTDOT
>
>     In the last example, the result is not very convenient because I will
>     still have to extract the bounds
>     and compute their type by myself which seems quite redundant with the
>     job performed by the lexer.
>     May be I am missing something ?
>
>     I would rather be able to express the range at the parser level which
>     seems much more convenient to me:
>     range: FLOATING_POINT_LITTERAL DOTDOT FLOATING_POINT_LITTERAL.
>     In this way, I will also be able to manage the possible spaces between
>     the bounds and the DOTDOT.
>
>     So, am I right to try to parse range at the parser level ? Or is
>     there a
>     solution to extract easily the bounds with their type if I am
>     doing the
>     job at the lexer level ?
>
>     Thanks in advance,
>     Fabien.
>
>
>
>
>     List: http://www.antlr.org/mailman/listinfo/antlr-interest
>     Unsubscribe:
>     http://www.antlr.org/mailman/options/antlr-interest/your-email-address
>
>



More information about the antlr-interest mailing list