/*
   A Program to forcibly extract the text content of Microsoft Word 
     files which are unreadable by other methods. 

                 Jan 26, 2001
  ---------------------------------------------------------------
              usage: NUTCR_WD src dst
              
              where src, dst are filespecs
  ----------------------------------------------------------------
             environment: MS-DOS, using MIX Power C compiler
             (superset of Microsoft C, Borland Turbo C), however,
             should be extremely portable.
  -----------------------------------------------------------------
             This program is in the public domain, and may be 
             used for any purpose whatsover, without permission 
             or notification, but it is offered "as-is" without 
             any warrentee, express or implied.
  -----------------------------------------------------------------
                                   Andrew D. Todd
                                   1249 Pineview Dr., Apt 4
                                   Morgantown, WV 26505

                                   U46A8@WVNVM.WVNET.EDU

*/
#include <stdio.h>
#include <io.h>
#include <fcntl.h>
#include <sys\types.h>
#include <sys\stat.h>
main(int argc, char *argv[])
{
  int c_row, status, f_in, f_out;
  char c;
  char e_string[80];
  if( argc < 3 )
  {
    printf(" \nBoth Source and Destination Files Needed\n");
    _exit(3);
  }
  f_in = open(argv[1], O_RDONLY|O_BINARY);
  if(f_in == -1) 
  {
    printf(" \n%s Does Not Exist\n", argv[1]);
    _exit(4);
  }
  f_out = open(argv[2], O_WRONLY|O_CREAT|O_EXCL|O_BINARY, S_IREAD|S_IWRITE);
  if( f_out == -1) 
  {
    strcpy(&e_string, " \nDestination File ");
    strcat(&e_string,  argv[2]);
    strcat(&e_string, " Invalid\n");
    perror(&e_string);
    _exit(5);
  }
  c_row=0;
  for(;;)            /* Do for the whole file */
  {
    status = read(f_in, &c, 1);
    if(status != 1) break;
         /*PRINTABLE*/
    else if( isprint(c)) 
    {
     write(f_out, &c, 1);
     c_row++;
    }
          /*CARRIAGE RETURN*/
    else if( c == '\r') 
    {
     write_string(f_out, "\r\n", &c_row);
     c_row=0;
    }
    else if( c == '\t') write_string(f_out, "   ", &c_row); /* TAB */
    else if( (c == '\x93') || (c == '\x94') )
                 write_string(f_out, "\"", &c_row); /* 2- QUOTE MARK */
    else if( c == '\x92') 
                 write_string(f_out, "'", &c_row);  /* 1-QUOTE MARK */
    else if( (c == '\xff') || (c ==  '\x00') ) continue;
    else write_hex_rep(f_out, c, &c_row);
    /*always*/
    if( (c_row > 50) &&  isspace(c) )
    {
     write(f_out, "\x8d\n", 2);
     c_row=0;
    }
    else if(  c_row > 65 )
    {
     write(f_out, "\r\n", 2);
     c_row=0;
    }
  }
  close(f_in);
  close(f_out);
  _exit(0);
}

/*-------------------------------------------------------------*/
write_hex_rep(int f_out, char c, int *c_row)
{
  char hex_rep[20], hexrepi[17];
  strcpy(&hex_rep, "[");
  strcat(&hex_rep, itoa(c, &hexrepi,16));
  strcat(&hex_rep, "]");
  write(f_out, &hex_rep, strlen(&hex_rep));
  *c_row+=4;
}
/*-------------------------------------------------------------*/
write_string(int f_out, char *string_to_put, int *c_row)
{
  int string_size;
  string_size = strlen(string_to_put);
  write(f_out, string_to_put, string_size);
  *c_row += string_size;
}