token.c

source code: Covert TXT to PDF

 
  /* Handle first non-zero digit */
  if ((digit=digit_value[ch]) < radix) {
    value = digit;
    save_ch(ch);
    ch = next_ch();
 
    /* Add digits until boundary case reached */
    while((digit=digit_value[ch]) < radix
            && value < (MAX_ULONG / radix)) {
      value = value * radix + digit;
      save_ch(ch);
      ch = next_ch();
    };
 
    /* Scan remaining digits */
    if ((digit=digit_value[ch]) < radix) {
 
      /* Examine boundary case ---
       *   radix*(MAX_ULONG/radix) <= number <= MAX_ULONG
       */
      if (value == (MAX_ULONG/radix) && digit <= MAX_ULONG%radix)
        value = value * radix + digit;
      else
        ++scale;
 
      /* Continue scanning digits, but can't store them */
      save_ch(ch);
      ch = next_ch();
      while(digit_value[ch] < radix) {
        ++scale;
        save_ch(ch);
        ch = next_ch();
      }
    }
  }
 
  /* Store result */
  r_value = (LONG) value; /* result is signed */
  r_scale = scale;
 
  return(ch);
}
 
/* -------------------------------------------------------------------
 * Complete a number; set token type and done flag.
 * Put current input character back, if it is not white space.
 */
 
/* Done: Radix Number */
static int RADIX_NUMBER(ch)
  int ch;
{
  back_ch_not_white(ch);
  if (r_scale == 0) {
    tokenValue.integer = r_value;
    tokenType = TOKEN_INTEGER;
  }
  else {
    tokenType = TOKEN_NAME;
  }
  return(DONE);
}
 
/* Done: Integer */
static int INTEGER(ch)
  int ch;
{
  back_ch_not_white(ch);
  if (m_scale == 0) {
    tokenValue.integer = m_value;
    tokenType = TOKEN_INTEGER;
  }
  else {
    tokenValue.real = (DOUBLE)(m_value) * Exp10(m_scale);
    tokenType = TOKEN_REAL;
  }
  return(DONE);
}
 
/* Done: Real */
static int REAL(ch)
  int ch;
{
  DOUBLE temp;
 
  back_ch_not_white(ch);
 
  /* NOTE: ignore e_scale, since e_value alone will cause
   *   exponent overflow if e_scale > 0.
   */
 
  /* HAZARD: exponent overflow of intermediate result
   * (e.g., in 370 floating point); this should not be a problem
   * with IEEE floating point.  Reduce exponent overflow hazard by
   * combining m_scale and e_value first, if they have different signs,
   * or multiplying m_value and one of the other factors, if both
   * m_scale and e_value are negative.
   */
  if ((m_scale >= 0 && e_value <= 0)
      || (m_scale <= 0 && e_value >= 0)) {
    tokenValue.real = (DOUBLE)(m_value) * Exp10(m_scale + e_value);
  }
  else {
    temp = (DOUBLE)(m_value) * Exp10(m_scale);
    tokenValue.real = temp * Exp10(e_value);
  }
 
  tokenType = TOKEN_REAL;
  return(DONE);
}
 
 
/* -------------------------------------------------------------------
 * Assemble a hex string; set token type and done flag.
 */
 
/* Done: Hex String */
static int HEX_STRING(ch)
  int ch;
{
  int value;
 
  while(TRUE) {
 
    /* Process odd digit */
    ch = next_ch();
    if (!isHEX_DIGIT(ch)) {
 
      /* Skip white space */
      while(isWHITE_SPACE(ch))
        ch = next_ch();
 
      /* Check for terminator */
      if (!isHEX_DIGIT(ch)) {
        break;
      }
    }
    value = digit_value[ch] << 4;
 
    /* Process even digit */
    ch = next_ch();
    if (!isHEX_DIGIT(ch)) {
 
      /* Skip white space */
      while(isWHITE_SPACE(ch))
        ch = next_ch();
 
      /* Check for terminator */
      if (!isHEX_DIGIT(ch)) {
        save_ch(value);
        break;
      }
    }
    save_ch(value + digit_value[ch]);
  }
 
  /* Classify result, based on why loop ended */
  if (ch == '>')
    tokenType = TOKEN_HEX_STRING;
  else {
    /* save the invalid character for error reporting */
    save_ch(ch);
    tokenType = TOKEN_INVALID;
  }
 
  return(DONE);
}
 
/* -------------------------------------------------------------------
 * Assemble a string; set token type and done flag
 */
 
/* Save a backslash-coded character in a string --
 *
 *   Store the proper character for special cases
 *   "\b", "\f", "\n", "\r", and "\t".
 *
 *   Decode and store octal-coded character, up to
 *   three octal digits, "\o", "\oo", and "\ooo".
 *
 *   The sequence "\<newline>" is a line continuation,
 *   so consume both without storing anything.
 *
 *   The sequence "\<EOF>" is an error; exit without
 *   storing anything and let the caller handle it.
 *
 *   For other characters, including the sequences
 *   "\\", "\(", and "\)", simply store the second
 *   character.
 */
static void save_digraph(ch)
  int ch;
{
  int value;
 
  switch (ch) {
 
    case 'b':   /* backspace */
      ch = '\b';
      break;
 
    case 'f':   /* formfeed */
      ch = '\f';
      break;
 
    case 'n':   /* newline */
      ch = '\n';
      break;
 
    case 'r':   /* carriage return */
      ch = '\r';
      break;
 
    case 't':   /* horizontal tab */
      ch = '\t';
      break;
 
    case '\n':  /* line continuation -- consume it */
      return;
 
    case '\r':  /* carriage return   -- consume it */
      ch = next_ch();   /* look at next character, is it \n?  */
      if (ch == '\n')  return;
      back_ch(ch);      /* if not a line feed, then return it */
      return;
 
    case EOF:   /* end of file -- forget it */
      return;
 
  default:
    /* scan up to three octal digits to get value */
    if (isOCTAL_DIGIT(ch)) {
      value = digit_value[ch];
      ch = next_ch();
      if (isOCTAL_DIGIT(ch)) {
        value = (value << 3) + digit_value[ch];
        ch = next_ch();
        if (isOCTAL_DIGIT(ch))
          value = (value << 3) + digit_value[ch];
        else
          back_ch(ch);
      }
      else
        back_ch(ch);
      ch = value;
    }
  }
 
  /* Found a character to save */
  save_ch(ch);
}
 
/* Done: String */
static int STRING(ch)
  int ch;
{
  int nest_level = 1;
 
  tokenType = TOKEN_STRING;
 
  do {
 
    ch = next_ch();
    while(!isSTRING_SPECIAL(ch)) {
      save_ch(ch);
      ch = next_ch();
    };
 
    switch (ch) {
 
     case '(':
       ++nest_level;
       save_ch(ch);
       break;
 
     case ')':
       if (--nest_level > 0)
         save_ch(ch);
       break;
 
     case '\\':
          save_digraph(next_ch());
        break;
 
     case '\r':
        /* All carriage returns (\r) are turned into linefeeds (\n)*/
          ch = next_ch();       /* get the next one, is it \n? */
          if (ch != '\n') {     /* if not, then put it back.   */
            back_ch(ch);
          }
          save_ch('\n');        /* in either case, save a linefeed */
        break;
 
 
     case EOF:
       tokenType = TOKEN_INVALID;  /* Unterminated string */
       nest_level = 0;
       break;
    }
 
  } while(nest_level > 0);
 
  return(DONE);
}
 
 
/* -------------------------------------------------------------------
 * Assemble a name; set token type and done flag.
 * Put current input character back, if it is not white space.
 */
 
/* Done: Name
 *  (Safe version used to complete name tokens that
 *   start out looking like something else).
 */
 
static int AAH_NAME(ch)
  int ch;
{
  do {
    save_ch(ch);
    ch = next_ch();
  } while(isNAME(ch));
 
  back_ch_not_white(ch);
  tokenType = TOKEN_NAME;
  return(DONE);
}
 
/* Done: Name */
static int NAME(ch)
  int ch;
{
  save_unsafe_ch(ch);
  ch = next_ch();
  if (isNAME(ch)) {
    save_unsafe_ch(ch);
    ch = next_ch();
    if (isNAME(ch)) {
      save_unsafe_ch(ch);
      ch = next_ch();
      if (isNAME(ch)) {
        save_unsafe_ch(ch);
        ch = next_ch();
        if (isNAME(ch)) {
          save_unsafe_ch(ch);
          ch = next_ch();
          if (isNAME(ch)) {
            save_unsafe_ch(ch);
            ch = next_ch();
            if (isNAME(ch)) {
              save_unsafe_ch(ch);
              ch = next_ch();
              while(isNAME(ch)) {
                save_ch(ch);
                ch = next_ch();
              }
            }
          }
        }
      }
    }
  }
 
  back_ch_not_white(ch);
  tokenType = TOKEN_NAME;
  return(DONE);
}
 
/* Done: Literal Name */
static int LITERAL_NAME(ch)
  int ch;
{
  if (isNAME(ch)) {
    save_unsafe_ch(ch);
    ch = next_ch();
    if (isNAME(ch)) {
      save_unsafe_ch(ch);
      ch = next_ch();
      if (isNAME(ch)) {
        save_unsafe_ch(ch);
        ch = next_ch();
        if (isNAME(ch)) {
          save_unsafe_ch(ch);
          ch = next_ch();
          if (isNAME(ch)) {
            save_unsafe_ch(ch);
            ch = next_ch();
            if (isNAME(ch)) {
              save_unsafe_ch(ch);
              ch = next_ch();
              while(isNAME(ch)) {
                save_ch(ch);
                ch = next_ch();
              }
            }
          }
        }
      }
    }
  }
 
  back_ch_not_white(ch);
  tokenType = TOKEN_LITERAL_NAME;
  return(DONE);
}
 
/* Done: immediate Name */
static int IMMED_NAME(ch)
  int ch;
{
  ch = next_ch();
  if (isNAME(ch)) {
    save_unsafe_ch(ch);
    ch = next_ch();
    if (isNAME(ch)) {
      save_unsafe_ch(ch);
      ch = next_ch();
      if (isNAME(ch)) {
        save_unsafe_ch(ch);
        ch = next_ch();
        if (isNAME(ch)) {
          save_unsafe_ch(ch);
          ch = next_ch();
          if (isNAME(ch)) {
            save_unsafe_ch(ch);
            ch = next_ch();
            if (isNAME(ch)) {
              save_unsafe_ch(ch);
              ch = next_ch();
              while(isNAME(ch)) {
                save_ch(ch);
                ch = next_ch();
              }
            }
          }
        }
      }
    }
  }
 
  back_ch_not_white(ch);
  tokenType = TOKEN_IMMED_NAME;
  return(DONE);
}
 
/* Done: Name found while looking for something else */
static int OOPS_NAME(ch)
  int ch;
{
  back_ch_not_white(ch);
  tokenType = TOKEN_NAME;
  return(DONE);
}
 
 
/* -------------------------------------------------------------------
 * Complete a miscellaneous token; set token type and done flag.
 */
 
/* Done: Unmatched Right Angle-Bracket */
static int RIGHT_ANGLE(ch)
  int ch;
{
  tokenType = TOKEN_RIGHT_ANGLE;
  return(DONE);
}
 
/* Done: Unmatched Right Parenthesis */
static int RIGHT_PAREN(ch)
  int ch;
{
  tokenType = TOKEN_RIGHT_PAREN;
  return(DONE);
}
 
/* Done: Left Brace */
static int LEFT_BRACE(ch)
  int ch;
{
  tokenType = TOKEN_LEFT_BRACE;
  return(DONE);
}
 
/* Done: Right Brace */
static int RIGHT_BRACE(ch)
  int ch;
{
  tokenType = TOKEN_RIGHT_BRACE;
  return(DONE);
}
 
/* Done: Left Bracket */
static int LEFT_BRACKET(ch)
  int ch;
{
  save_unsafe_ch(ch);
  tokenType = TOKEN_LEFT_BRACKET;
  return(DONE);
}
 
/* Done: Right Bracket */
static int RIGHT_BRACKET(ch)
  int ch;
{
  save_unsafe_ch(ch);
  tokenType = TOKEN_RIGHT_BRACKET;
  return(DONE);
}
 
/* Done: Break */
static int BREAK_SIGNAL(ch)
  int ch;
{
  tokenType = TOKEN_BREAK;
  return(DONE);
}
 
/* Done: No Token Found */
static int NO_TOKEN(ch)
  int ch;
{
  tokenType = TOKEN_EOF;
  return(DONE);
}
 
 
/*
 * -------------------------------------------------------------------
 *  scan_token -- scan one token from the input.  It uses a simple
 *    finite state machine to recognize token classes.
 *
 *  The input is from a file.
 *
 *  On entry --
 *
 *    inputP -> input PostScript object, a file.
 *    tokenStartP -> buffer in VM for accumulating the token.
 *    tokenMaxP -> last character in the token buffer
 *
 *  On exit --
 *
 *    tokenLength = number of characters in the token
 *    tokenTooLong = TRUE if the token did not fit in the buffer
 *    tokenType = code for the type of token parsed.
 *    tokenValue = converted value of a numeric token.
 *
 *
 * -------------------------------------------------------------------
 */
void scan_token(inputP)
  psobj *inputP;
{
  int ch;
  unsigned char *stateP = s0;
  unsigned char entry;
  int (*actionP)();
 
  /* Define input source */
  inputFileP = inputP->data.fileP;
  if (inputFileP == NULL)  {
    tokenType = TOKEN_EOF;
    return;
  }
 
  /* Ensure enough space for most cases
   * (so we don't have to keep checking)
   * The length needs to cover the maximum number
   * of save_unsafe_ch() calls that might be executed.
   * That number is 11 (a sign and 10 decimal digits, e.g.,
   * when scanning -2147483648), but use MAX_NAME_LEN
   * in case someone changes that without checking.
   */
  if (vm_free_bytes() < (MAX_NAME_LEN)) {
     if (!(vm_init())) {
        tokenLength = 0;
        tokenTooLong = TRUE;
        tokenType = TOKEN_NONE;
        tokenValue.integer = 0;
        return;
     }
  }
 
  tokenStartP = vm_next_byte();
 
  /* Reset token */
  tokenCharP = tokenStartP;
  tokenTooLong = FALSE;
 
  /* Scan one token */
  ch = next_ch();
  do {
    entry = stateP[ch];
    stateP = classActionTable[entry].nextStateP;
    actionP = classActionTable[entry].actionRoutineP;
    ch = (*actionP)(ch);
  } while(ch != DONE);
 
 
  /* Return results */
  tokenLength = tokenCharP - tokenStartP;
}