/*
 * tokenize.C: Tokenization routines.
 *
 * Copyright © 2007 Pippijn van Steenhoven / The Ermyth Team
 * Rights to this code are as documented in COPYING.
 *
 *
 * Portions of this file were derived from sources bearing the following license:
 * Rights to this code are documented in doc/pod/license.pod.
 * Copyright © 2005-2007 Atheme Project (http://www.atheme.org)
 */

static char const rcsid[] = "$Id: tokenize.C,v 1.6 2007/09/16 18:54:45 pippijn Exp $";

#include "atheme.h"
#include "pmodule.h"

int
sjtoken (char *message, char delimiter, char **parv, int limit)
{
  char *next;
  int count;

  if (!message)
    return 0;

  /* now we take the beginning of the message and find all the spaces...
   * set them to \0 and use 'next' to go through the string
   */
  next = message;

  /* eat any additional delimiters */
  while (*next == delimiter)
    next++;

  parv[0] = next;
  count = 1;

  while (*next)
    {
      /* this is fine here, since we don't have a :delimited
       * parameter like tokenize
       */

      if (count == limit)
	{
	  /* we've reached our limit */
          if (count == 256)
            slog (LG_DEBUG, "sjtoken(): reached param limit");
	  return count;
	}

      if (*next == delimiter)
	{
	  *next = '\0';
	  next++;
	  /* eat any additional delimiters */
	  while (*next == delimiter)
	    next++;
	  /* if it's the end of the string, it's simply
	   ** an extra space at the end.  here we break.
	   */
	  if (*next == '\0')
	    break;

	  /* if it happens to be a stray \r, break too */
	  if (*next == '\r')
	    break;

	  parv[count] = next;
	  count++;
	}
      else
	next++;
    }

  return count;
}

/* this splits apart a message with origin and command picked off already */
int
tokenize (char *message, char **parv)
{
  char *pos = NULL;
  char *next;
  unsigned int count = 0;

  if (!message)
    return -1;

  /* first we fid out of there's a : in the message, save that string
   * somewhere so we can set it to the last param in parv
   * also make sure there's a space before it... if not then we're screwed
   */
  pos = message;
  while (true)
    {
      if ((pos = strchr (pos, ':')))
	{
	  pos--;
	  if (*pos != ' ')
	    {
	      pos += 2;
	      continue;
	    }
	  *pos = '\0';
	  pos++;
	  *pos = '\0';
	  pos++;
	  break;
	}
      else
	break;
    }

  /* now we take the beginning of the message and find all the spaces...
   * set them to \0 and use 'next' to go through the string
   */

  next = message;
  parv[0] = message;
  count = 1;

  while (*next)
    {
      if (count == MAXPARC)
	{
	  /* we've reached one less than our max limit
	   * to handle the parameter we already ripped off
	   */
	  slog (LG_DEBUG, "tokenize(): reached para limit");
	  return count;
	}
      if (*next == ' ')
	{
	  *next = '\0';
	  next++;
	  /* eat any additional spaces */
	  while (*next == ' ')
	    next++;
	  /* if it's the end of the string, it's simply
	   * an extra space before the :parameter. break.
	   */
	  if (*next == '\0')
	    break;
	  parv[count] = next;
	  count++;
	}
      else
	next++;
    }

  if (pos)
    {
      parv[count] = pos;
      count++;
    }

  return count;
}