Ignore tokens in the token characters?

问题

I have the following token definition in my lexer defining a CharacterString (e.g. 'abcd'):

CharacterString:
  Apostrophe
  (Alphanumeric)*
  Apostrophe
;

Is it possible to ignore the two apostrophes to then be able to get the token string without them in the lexer (via $CharacterString.text->chars)?

I tried ...

CharacterString:
  Apostrophe { $channel = HIDDEN; }
  (Alphanumeric)*
  Apostrophe { $channel = HIDDEN; }
;

... without success... This case does not even match my string anymore (e.g. 'oiu' will fail in the parser - Missmatched Set Exception).

Thank you :)

回答1:

The inline code {$channel=HIDDEN;} affects the entire CharacterString, so you can't do it like the way you tried.

You will need to add some custom code and remove the quotes yourself. Here's a small C demo:

grammar T;

options {
  language=C;
}

parse
  :  (t=. {printf(">\%s<\n", $t.text->chars);})+ EOF
  ;

CharacterString
  :  '\'' ~'\''* '\''
     {
       pANTLR3_STRING quoted = GETTEXT();
       SETTEXT(quoted->subString(quoted, 1, quoted->len-1));
     }
  ;

Any
  :  .
  ;

and a little test function:

#include "TLexer.h"
#include "TParser.h"

int main(int argc, char *argv[])
{
  pANTLR3_UINT8 fName = (pANTLR3_UINT8)"input.txt";
  pANTLR3_INPUT_STREAM input = antlr3AsciiFileStreamNew(fName);

  if(input == NULL)
  {
    fprintf(stderr, "Failed to open file %s\n", (char *)fName);
    exit(1);
  }

  pTLexer lexer = TLexerNew(input);

  if(lexer == NULL)
  {
    fprintf(stderr, "Unable to create the lexer due to malloc() failure1\n");
    exit(1);
  }

  pANTLR3_COMMON_TOKEN_STREAM tstream = antlr3CommonTokenStreamSourceNew(ANTLR3_SIZE_HINT, TOKENSOURCE(lexer));

  if(tstream == NULL)
  {
    fprintf(stderr, "Out of memory trying to allocate token stream\n");
    exit(1);
  }

  pTParser parser = TParserNew(tstream);

  if(parser == NULL)
  {
    fprintf(stderr, "Out of memory trying to allocate parser\n");
    exit(ANTLR3_ERR_NOMEM);
  }

  parser->parse(parser);

  parser->free(parser);   parser = NULL;
  tstream->free(tstream); tstream = NULL;
  lexer->free(lexer);     lexer = NULL;
  input->close(input);    input = NULL;

  return 0;
}

and the test input.txt file contains:

'abc'

If you now 1) generate the lexer and parser, 2) compile all .c source files, and 3) run main:

# 1
java -cp antlr-3.3.jar org.antlr.Tool T.g

# 2
gcc -Wall main.c TLexer.c TParser.c -l antlr3c -o main

# 3
./main

you'll see that abc (without the quotes) is being printed to the console.

回答2:

You can influence token construction via RecognizerSharedState state attribute of your lexer:

CharacterString:
  Apostrophe
  CharSequence
  Apostrophe
  { state.text = $CharSequence.text; }
;

fragment CharSequence:
  Alphanumeric+
;

来源：https://stackoverflow.com/questions/6047127/ignore-tokens-in-the-token-characters

标签

parsing

antlr

antlr3