%token_table and stuff

[Michael (Micksa) Slade]

This is more of a bug report than anything but hey

I managed to get bison and flex to cooperate so I could put tokens in 
the grammar in the style of '"foo"' without having to define macros for 
token numbers and all that.  The extra code required to do this isn't 
documented, so I thought I'd show what I did:

int token_value(char * token_buffer) {
  int i;
  int token_len = strlen(token_buffer);

  for (i = 0; i < YYNTOKENS; i++) {
     if (yytname[i] != 0
     && yytname[i][0] == '"'
    && strncmp (yytname[i] + 1, token_buffer,token_len) == 0
        && yytname[i][token_len + 1] == '"'
        && yytname[i][token_len + 2] == 0)
   break;
  }
  if(i == YYNTOKENS) return 2;
  return yytoknum[i];
}

Then in the lexer, I just went:

{WORD} {
  return token_value(yytext);
}

(With the prototype for token_value up the top as well)

Hope this helps,

Mick.