c-lex.c

gcc库的原代码,对编程有很大帮助.

  if (c == '\n')
    return c;

  /* Something follows the #; read a token.  */

  ungetc (c, finput);
  token = yylex ();

  if (token == CONSTANT
      && TREE_CODE (yylval.ttype) == INTEGER_CST)
    {
      int old_lineno = lineno;
      int used_up = 0;
      /* subtract one, because it is the following line that
	 gets the specified number */

      int l = TREE_INT_CST_LOW (yylval.ttype) - 1;

      /* Is this the last nonwhite stuff on the line?  */
      c = getc (finput);
      while (c == ' ' || c == '\t')
	c = getc (finput);
      if (c == '\n')
	{
	  /* No more: store the line number and check following line.  */
	  lineno = l;
	  return c;
	}
      ungetc (c, finput);

      /* More follows: it must be a string constant (filename).  */

      /* Read the string constant.  */
      token = yylex ();

      if (token != STRING || TREE_CODE (yylval.ttype) != STRING_CST)
	{
	  error ("invalid #line");
	  goto skipline;
	}

      input_filename
	= (char *) permalloc (TREE_STRING_LENGTH (yylval.ttype) + 1);
      strcpy (input_filename, TREE_STRING_POINTER (yylval.ttype));
      lineno = l;

      /* Each change of file name
	 reinitializes whether we are now in a system header.  */
      in_system_header = 0;

      if (main_input_filename == 0)
	main_input_filename = input_filename;

      /* Is this the last nonwhite stuff on the line?  */
      c = getc (finput);
      while (c == ' ' || c == '\t')
	c = getc (finput);
      if (c == '\n')
	{
	  /* Update the name in the top element of input_file_stack.  */
	  if (input_file_stack)
	    input_file_stack->name = input_filename;

	  return c;
	}
      ungetc (c, finput);

      token = yylex ();
      used_up = 0;

      /* `1' after file name means entering new file.
	 `2' after file name means just left a file.  */

      if (token == CONSTANT
	  && TREE_CODE (yylval.ttype) == INTEGER_CST)
	{
	  if (TREE_INT_CST_LOW (yylval.ttype) == 1)
	    {
	      /* Pushing to a new file.  */
	      struct file_stack *p
		= (struct file_stack *) xmalloc (sizeof (struct file_stack));
	      input_file_stack->line = old_lineno;
	      p->next = input_file_stack;
	      p->name = input_filename;
	      input_file_stack = p;
	      input_file_stack_tick++;
#ifdef DWARF_DEBUGGING_INFO
	      if (debug_info_level == DINFO_LEVEL_VERBOSE
		  && write_symbols == DWARF_DEBUG)
		dwarfout_start_new_source_file (input_filename);
#endif /* DWARF_DEBUGGING_INFO */

	      used_up = 1;
	    }
	  else if (TREE_INT_CST_LOW (yylval.ttype) == 2)
	    {
	      /* Popping out of a file.  */
	      if (input_file_stack->next)
		{
		  struct file_stack *p = input_file_stack;
		  input_file_stack = p->next;
		  free (p);
		  input_file_stack_tick++;
#ifdef DWARF_DEBUGGING_INFO
		  if (debug_info_level == DINFO_LEVEL_VERBOSE
		      && write_symbols == DWARF_DEBUG)
		    dwarfout_resume_previous_source_file (input_file_stack->line);
#endif /* DWARF_DEBUGGING_INFO */
		}
	      else
		error ("#-lines for entering and leaving files don't match");

	      used_up = 1;
	    }
	}

      /* Now that we've pushed or popped the input stack,
	 update the name in the top element.  */
      if (input_file_stack)
	input_file_stack->name = input_filename;

      /* If we have handled a `1' or a `2',
	 see if there is another number to read.  */
      if (used_up)
	{
	  /* Is this the last nonwhite stuff on the line?  */
	  c = getc (finput);
	  while (c == ' ' || c == '\t')
	    c = getc (finput);
	  if (c == '\n')
	    return c;
	  ungetc (c, finput);

	  token = yylex ();
	  used_up = 0;
	}

      /* `3' after file name means this is a system header file.  */

      if (token == CONSTANT
	  && TREE_CODE (yylval.ttype) == INTEGER_CST
	  && TREE_INT_CST_LOW (yylval.ttype) == 3)
	in_system_header = 1, used_up = 1;

      if (used_up)
	{
	  /* Is this the last nonwhite stuff on the line?  */
	  c = getc (finput);
	  while (c == ' ' || c == '\t')
	    c = getc (finput);
	  if (c == '\n')
	    return c;
	  ungetc (c, finput);
	}

      warning ("unrecognized text at end of #line");
    }
  else
    error ("invalid #-line");

  /* skip the rest of this line.  */
 skipline:
  if (c == '\n')
    return c;
  while ((c = getc (finput)) != EOF && c != '\n');
  return c;
}

#ifdef HANDLE_SYSV_PRAGMA

/* Handle a #pragma directive.  INPUT is the current input stream,
   and C is a character to reread.  Processes the entire input line
   and returns a character for the caller to reread: either \n or EOF.  */

/* This function has to be in this file, in order to get at
   the token types.  */

int
handle_sysv_pragma (input, c)
     FILE *input;
     int c;
{
  for (;;)
    {
      while (c == ' ' || c == '\t')
	c = getc (input);
      if (c == '\n' || c == EOF)
	{
	  handle_pragma_token (0, 0);
	  return c;
	}
      ungetc (c, input);
      switch (yylex ())
	{
	case IDENTIFIER:
	case TYPENAME:
	case STRING:
	case CONSTANT:
	  handle_pragma_token (token_buffer, yylval.ttype);
	  break;
	default:
	  handle_pragma_token (token_buffer, 0);
	}
      if (nextchar >= 0)
	c = nextchar, nextchar = -1;
      else
	c = getc (input);
    }
}

#endif /* HANDLE_SYSV_PRAGMA */

#define ENDFILE -1  /* token that represents end-of-file */

/* Read an escape sequence, returning its equivalent as a character,
   or store 1 in *ignore_ptr if it is backslash-newline.  */

static int
readescape (ignore_ptr)
     int *ignore_ptr;
{
  register int c = getc (finput);
  register int code;
  register unsigned count;
  unsigned firstdig = 0;
  int nonnull;

  switch (c)
    {
    case 'x':
      if (warn_traditional)
	warning ("the meaning of `\\x' varies with -traditional");

      if (flag_traditional)
	return c;

      code = 0;
      count = 0;
      nonnull = 0;
      while (1)
	{
	  c = getc (finput);
	  if (!(c >= 'a' && c <= 'f')
	      && !(c >= 'A' && c <= 'F')
	      && !(c >= '0' && c <= '9'))
	    {
	      ungetc (c, finput);
	      break;
	    }
	  code *= 16;
	  if (c >= 'a' && c <= 'f')
	    code += c - 'a' + 10;
	  if (c >= 'A' && c <= 'F')
	    code += c - 'A' + 10;
	  if (c >= '0' && c <= '9')
	    code += c - '0';
	  if (code != 0 || count != 0)
	    {
	      if (count == 0)
		firstdig = code;
	      count++;
	    }
	  nonnull = 1;
	}
      if (! nonnull)
	error ("\\x used with no following hex digits");
      else if (count == 0)
	/* Digits are all 0's.  Ok.  */
	;
      else if ((count - 1) * 4 >= TYPE_PRECISION (integer_type_node)
	       || (count > 1
		   && ((1 << (TYPE_PRECISION (integer_type_node) - (count - 1) * 4))
		       <= firstdig)))
	pedwarn ("hex escape out of range");
      return code;

    case '0':  case '1':  case '2':  case '3':  case '4':
    case '5':  case '6':  case '7':
      code = 0;
      count = 0;
      while ((c <= '7') && (c >= '0') && (count++ < 3))
	{
	  code = (code * 8) + (c - '0');
	  c = getc (finput);
	}
      ungetc (c, finput);
      return code;

    case '\\': case '\'': case '"':
      return c;

    case '\n':
      lineno++;
      *ignore_ptr = 1;
      return 0;

    case 'n':
      return TARGET_NEWLINE;

    case 't':
      return TARGET_TAB;

    case 'r':
      return TARGET_CR;

    case 'f':
      return TARGET_FF;

    case 'b':
      return TARGET_BS;

    case 'a':
      if (warn_traditional)
	warning ("the meaning of `\\a' varies with -traditional");

      if (flag_traditional)
	return c;
      return TARGET_BELL;

    case 'v':
#if 0 /* Vertical tab is present in common usage compilers.  */
      if (flag_traditional)
	return c;
#endif
      return TARGET_VT;

    case 'e':
    case 'E':
      if (pedantic)
	pedwarn ("non-ANSI-standard escape sequence, `\\%c'", c);
      return 033;

    case '?':
      return c;

      /* `\(', etc, are used at beginning of line to avoid confusing Emacs.  */
    case '(':
    case '{':
    case '[':
      /* `\%' is used to prevent SCCS from getting confused.  */
    case '%':
      if (pedantic)
	pedwarn ("non-ANSI escape sequence `\\%c'", c);
      return c;
    }
  if (c >= 040 && c < 0177)
    pedwarn ("unknown escape sequence `\\%c'", c);
  else
    pedwarn ("unknown escape sequence: `\\' followed by char code 0x%x", c);
  return c;
}

void
yyerror (string)
     char *string;
{
  char buf[200];

  strcpy (buf, string);

  /* We can't print string and character constants well
     because the token_buffer contains the result of processing escapes.  */
  if (end_of_file)
    strcat (buf, " at end of input");
  else if (token_buffer[0] == 0)
    strcat (buf, " at null character");
  else if (token_buffer[0] == '"')
    strcat (buf, " before string constant");
  else if (token_buffer[0] == '\'')
    strcat (buf, " before character constant");
  else if (token_buffer[0] < 040 || (unsigned char) token_buffer[0] >= 0177)
    sprintf (buf + strlen (buf), " before character 0%o",
	     (unsigned char) token_buffer[0]);
  else
    strcat (buf, " before `%s'");

  error (buf, token_buffer);
}

#if 0

struct try_type
{
  tree *node_var;
  char unsigned_flag;
  char long_flag;
  char long_long_flag;
};

struct try_type type_sequence[] = 
{
  { &integer_type_node, 0, 0, 0},
  { &unsigned_type_node, 1, 0, 0},
  { &long_integer_type_node, 0, 1, 0},
  { &long_unsigned_type_node, 1, 1, 0},
  { &long_long_integer_type_node, 0, 1, 1},
  { &long_long_unsigned_type_node, 1, 1, 1}
};
#endif /* 0 */

int
yylex ()
{
  register int c;
  register char *p;
  register int value;
  int wide_flag = 0;
  int objc_flag = 0;

  if (nextchar >= 0)
    c = nextchar, nextchar = -1;
  else
    c = getc (finput);

  /* Effectively do c = skip_white_space (c)
     but do it faster in the usual cases.  */
  while (1)
    switch (c)
      {
      case ' ':
      case '\t':
      case '\f':
      case '\v':
      case '\b':
	c = getc (finput);
	break;

      case '\r':
	/* Call skip_white_space so we can warn if appropriate.  */

      case '\n':
      case '/':
      case '\\':
	c = skip_white_space (c);
      default:
	goto found_nonwhite;
      }
 found_nonwhite:

  token_buffer[0] = c;
  token_buffer[1] = 0;

/*  yylloc.first_line = lineno; */

  switch (c)
    {
    case EOF:
      end_of_file = 1;
      token_buffer[0] = 0;
      value = ENDFILE;
      break;

    case '$':
      if (dollars_in_ident)
	goto letter;
      return '$';

    case 'L':
      /* Capital L may start a wide-string or wide-character constant.  */
      {
	register int c = getc (finput);
	if (c == '\'')
	  {
	    wide_flag = 1;
	    goto char_constant;
	  }
	if (c == '"')
	  {
	    wide_flag = 1;
	    goto string_constant;
	  }
	ungetc (c, finput);
      }
      goto letter;

    case '@':
      if (!doing_objc_thang)
	{
	  value = c;
	  break;
	}
      else
	{
	  /* '@' may start a constant string object.  */
	  register int c = getc(finput);
	  if (c == '"')
	    {
	      objc_flag = 1;
	      goto string_constant;
	    }
	  ungetc(c, finput);
	  /* Fall through to treat '@' as the start of an identifier.  */
	}

    case 'A':  case 'B':  case 'C':  case 'D':  case 'E':
    case 'F':  case 'G':  case 'H':  case 'I':  case 'J':
    case 'K':		  case 'M':  case 'N':  case 'O':
    case 'P':  case 'Q':  case 'R':  case 'S':  case 'T':
    case 'U':  case 'V':  case 'W':  case 'X':  case 'Y':
    case 'Z':
    case 'a':  case 'b':  case 'c':  case 'd':  case 'e':
    case 'f':  case 'g':  case 'h':  case 'i':  case 'j':
    case 'k':  case 'l':  case 'm':  case 'n':  case 'o':
    case 'p':  case 'q':  case 'r':  case 's':  case 't':
    case 'u':  case 'v':  case 'w':  case 'x':  case 'y':
    case 'z':