/*
 * Copyright (c) 1996-2007, OpenFWTK Development Group
 * All rights reserved. See LICENSE.
 */

/* maketable.c */

/* Copyright 1997-1999 by Eberhard Mattes <em-gw@windhager.de>
   Donated to the public domain.  No warranty.

   1997-07-19 Initial version
   1997-09-06 Lower-case stuff moved to libem
   1999-01-01 Accept '&' in attribute values (CTYPE_VALUE) */

#include <assert.h>
#include <stdio.h>
#include <stdlib.h>
#include <limits.h>
#include <string.h>
#define SELECT_CTYPE_HTML
#define SELECT_CTYPE_HTTP
#define SELECT_CTYPE_URL
#include "tables.h"

enum target
{
  TARGET_HTTP, TARGET_HTML, TARGET_URL
};

static void usage (void)
{
  puts ("Usage:");
  puts ("  maketable -t <type>         Make character-type table for http, html, or url");
  exit (1);
}

static void do_not_edit (void)
{
  puts ("/* DO NOT EDIT THIS FILE -- "
	"it is automatically generated by maketable */\n");
}

/* Build a table for character classification.  The table maps an
   unsigned char to an unsigned char containing a set of bits.  The
   output contains just the initialization data, without braces. */

static void make_ctype (enum target type)
{
  int i;
  unsigned c;

  do_not_edit ();
  i = 0;
  for (;;)
    {
      c = 0;
      if (i >= 'A' && i <= 'Z')
	c |= CTYPE_UPALPHA;
      if (i >= 'a' && i <= 'z')
	c |= CTYPE_LOALPHA;
      if (i >= '0' && i <= '9')
	c |= CTYPE_DIGIT;
      if (i == SP || i == HT)	/* Be tolerant: treat HT as space for HTTP */
	c |= CTYPE_WHITE;
      switch (type)
	{
	case TARGET_HTML:
	  if (i == LF || i == CR)
	    c |= CTYPE_WHITE;
	  if (i == '.' || i == '-')
	    c |= CTYPE_I_NAME;

	  if (i != HT && i != LF && i != CR
	      && (i < 0x20 || (i >= 0x7f && i <= 0x9a) || i == '<' || i == '>'
		  || i == '&' || i == '"' || i == '\''))
	    c |= CTYPE_ESCAPE;
	  /* Note that we cannot yet deal with quotes and "&" in
	     parse_attributes(). */
	  if (i >= 0x21 && i != 0x7f && i != '"' && i != '\''
	      && i != '<' && i != '>')
	    c |= CTYPE_VALUE;
	  break;
	case TARGET_HTTP:
	  /* See `token' in RFC 2068.  This is used for cookies (RFC
	     2109). */
	  if (i >= 32 && i != '(' && i != ')' && i != '@'
	      && i != ',' && i != ';' && i != ':' && i != '\\' && i != '"'
	      && i != '/' && i != '[' && i != ']' && i != '?' && i != '='
	      && i != '{' && i != '}' && i != SP && i != HT)
	    c |= CTYPE_TOKEN;
	  break;
	case TARGET_URL:
	  /* Oops, the BNF in RFC 1630 does not allow "+" and "." in
             schemes!  Go with RFC 1738.  Add '_' for Squid's
             "cache_object". */

	  if (i == '+' || i == '-' || i == '.' || i == '_')
	    c |= CTYPE_I_SCHEME;
	  if ((i >= 'a' && i <= 'f') || (i >= 'A' && i <= 'F'))
	    c |= CTYPE_I_HEX;

	  /* "hsegment" of RFC 1738.  Of course, RFC 1738 and RFC 2068
             contradict.  In fact, we implement "uchar" of RFC 2068,
             plus "&", "/", "+", ":", ";", "<", "=", ">", and "@". */

	  if (i >= 33 && i != '"' && i != '#' && i != '%' && i != '?')
	    c |= CTYPE_I_HPATH;

	  /* "fsegment" of RFC 1738.  Of course, RFC 1738 and RFC 2068
	     contradict.  In fact, we implement "uchar" of RFC 2068,
	     plus "&", "/", "+", ":", "<", "=", ">", and "@", minus
	     ";". */

	  if (i >= 33 && i != '"' && i != '#' && i != '%'
	      && i != ';' && i != '?')
	    c |= CTYPE_I_FPATH;

	  /* "user" of RFC 1738, sans "alpha", "digit", and
             "escape". */

	  if (/* "user" */
	      i == ';' || i == '?' || i == '&' || i == '='
	      /* "safe" */
	      || i == '$' || i == '-' || i == '_' || i == '.' || i == '+'
	      /* "extra" */
	      || i == '!' || i == '*' || i == '\'' || i == '(' || i == ')'
	      || i == ',')
	    c |= CTYPE_I_USER;
	  break;
	}
      printf ("0x%.2x", c);
      if (i == UCHAR_MAX)
	break;
      ++i;
      if (i % 8 == 0)
	fputs (",\n", stdout);
      else
	fputs (", ", stdout);
    }
  putchar ('\n');
}

int main (int argc, char *argv[])
{
  if (argc == 3 && strcmp (argv[1], "-t") == 0)
    {
      if (strcmp (argv[2], "http") == 0)
	make_ctype (TARGET_HTTP);
      else if (strcmp (argv[2], "html") == 0)
	make_ctype (TARGET_HTML);
      else if (strcmp (argv[2], "url") == 0)
	make_ctype (TARGET_URL);
      else
	usage ();
    }
  else
    usage ();
  if (fflush (stdout) != 0)
    {
      perror ("maketable");
      exit (2);
    }
  return 0;
}