roman_numerals.rl

/** Convert Roman numerals given on the command line to decimals.

  $ ragel -F0 -o roman_numerals.c roman_numerals.rl
  $ gcc -finput-charset=utf-8 -fexec-charset=utf-8 -o roman_numerals roman_numerals.c
  $ ./roman_numerals IX Ⅻ CCCC
  9
  12
  400
  $ ./roman_numerals IIX && exit 1 || echo invalid
  invalid
*/
#include <string.h>
#include <stdio.h>
#include <stdlib.h>

%%{
  machine roman_numerals;
  write data;
}%%

int
from_roman_to_int(const char* s)
{
    int n, cs;
    const char *p = s, *pe = s + strlen(s), *eof = pe;

  %%{
      # units
      I = ('I' | 'i' |  "Ⅰ" | "ⅰ") %{ n += 1; };
      II =       I I | ("Ⅱ" | "ⅱ") %{ n += 2; };
      III =    I I I | ("Ⅲ" | "ⅲ") %{ n += 3; };
      V = ('V' | 'v' | "Ⅴ" | "ⅴ")  %{ n += 5; };
      VI =       V I | ("Ⅵ" | "ⅵ") %{ n += 6; };
      VII =    V I I | ("Ⅶ" | "ⅶ") %{ n += 7; };
      VIII = V I I I | ("Ⅷ" | "ⅷ") %{ n += 8; };
      # tens
      X = ('X' | 'x' |  "Ⅹ" | "ⅹ") %{ n += 10; };
      XI =             ("Ⅺ" | "ⅺ") %{ n += 11; };
      XII =            ("Ⅻ" | "ⅻ") %{ n += 12; };
      L = ('L' | 'l' |  "Ⅼ" | "ⅼ") %{ n += 50; };
      # hundreds
      C = ('C' | 'c' |  "Ⅽ" | "ⅽ") %{ n += 100; };
      D = ('D' | 'd' |  "Ⅾ" | "ⅾ") %{ n += 500; };
      # thousands
      M = ('M' | 'm' |  "Ⅿ" | "ⅿ") %{ n += 1000;};

      # subtractive principle
      IV = I V %{ n -= 2;   } | I{4}   | ("Ⅳ" | "ⅳ") %{ n += 4; };
      IX = I X %{ n -= 2;   } | V I{4} | ("Ⅸ" | "ⅸ") %{ n += 9; };
      XL = X L %{ n -= 20;  };
      XC = X C %{ n -= 20;  };
      CD = C D %{ n -= 200; };
      CM = C M %{ n -= 200; };

    thousands =             M{,4};
    hundreds = CM | CD | D? C{,4};
    tens     = XC | XL | L? X{,4};
    units    = IX | VIII | VII | VI | V | IV | III | II | I;
    numeral = thousands? hundreds? (XII? | XI? | tens? units?);

    main := numeral > { n = 0; } ;

    write init;
    write exec;

  }%%
     return (cs == roman_numerals_error || cs == roman_numerals_start) ? -1 : n ;
}



int main(int argc, char* argv[]) {
  if (argc < 2) {
    fputs("Usage: roman_numerals <text>...\n", stderr);
    exit(EXIT_FAILURE);
  }

  for (int i = 1; i < argc; ++i) {
    int n = from_roman_to_int(argv[i]);
    if (n < 0)
      exit(EXIT_FAILURE);

    printf("%d\n", n); /* print result */
  }
  return 0;
}