#! /usr/bin/gawk -f

BEGIN {
  # OK is set to indicate normal exit
  ok = false;
  
  # Get command line args:
  if (ARGC != 3) { printf "usage: %s file1 file2\n" > "/dev/stderr"; exit 1; }
  file1 = ARGV[1]
  file2 = ARGV[2]
  ARGC = 1;
  
  # Table of changed characters:
  split("", change)
  split("", ch1)
  split("", ch2)
  
  # Loop on records:
  linect=0
  while (1)
    {
      eof1 = ((getline rec1 < file1) <= 0);
      eof2 = ((getline rec2 < file2) <= 0);
      if (eof1 && eof2) { ok = 1; exit 0; }
      if (eof1 || eof2) 
        { printf "unpaired lines\n" > "/dev/stderr"; exit 1; }
      linect++;
      n = length(rec1);
      if (length(rec2) != n)
        { printf "line lengths differ, line %d\n", linect > "/dev/stderr"; exit 1; }
      for (i=1;i<=n;i++)
        { c1 = substr(rec1, i, 1);
          c2 = substr(rec2, i, 1);
          changes[c1,c2]++;
          ch1[c1]++;
          ch2[c2]++;
          if ((c1 == " ") && (c2 == ":"))
            { printf "+" }
          else if ((c1 == ":") && (c2 == " "))
            { printf "-" }
          else
            { printf "%s", c1 }
        }
      printf "\n"
    }
}

END {
  if (! ok) { exit 1; }
  e = "/dev/stderr";

  for (c1 in ch1) { if (ch1[c1] == changes[c1,c1]) delete ch1[c1]; }
  for (c2 in ch2) { if (ch2[c2] == changes[c2,c2]) delete ch2[c2]; }

  printf "   " > e;
  for (c2 in ch2) { printf " %5.5s", c2 > e; }
  printf "\n" > e;

  printf "   " > e;
  for (c2 in ch2) { printf " %5.5s", "-----------" > e; }
  printf "\n" > e;
  
  for (c1 in ch1)
    { printf "%s |", c1 > e;
      for (c2 in ch2) { printf " %5d", changes[c1,c2] > e; }
      printf "\n" > e;
    }
}