cexp.y

GCC编译器源代码

/* Parse C expressions for CCCP.
   Copyright (C) 1987, 1992, 1994, 1995, 1996, 1997 Free Software Foundation.

This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2, or (at your option) any
later version.

This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with this program; if not, write to the Free Software
Foundation, 59 Temple Place - Suite 330,
Boston, MA 02111-1307, USA.

 In other words, you are welcome to use, share and improve this program.
 You are forbidden to forbid anyone else to use, share and improve
 what you give them.   Help stamp out software-hoarding!

 Adapted from expread.y of GDB by Paul Rubin, July 1986.  */

/* Parse a C expression from text in a string  */
   
%{
#include "config.h"
#include <setjmp.h>
/* #define YYDEBUG 1 */


#ifdef HAVE_STRING_H
# include <string.h>
#endif

#ifdef HAVE_STDLIB_H
# include <stdlib.h>
#endif

#ifdef HAVE_LIMITS_H
# include <limits.h>
#endif

#ifdef MULTIBYTE_CHARS
#include <locale.h>
#endif

#include <stdio.h>

typedef unsigned char U_CHAR;

/* This is used for communicating lists of keywords with cccp.c.  */
struct arglist {
  struct arglist *next;
  U_CHAR *name;
  int length;
  int argno;
};

/* Define a generic NULL if one hasn't already been defined.  */

#ifndef NULL
#define NULL 0
#endif

#ifndef GENERIC_PTR
#if defined (USE_PROTOTYPES) ? USE_PROTOTYPES : defined (__STDC__)
#define GENERIC_PTR void *
#else
#define GENERIC_PTR char *
#endif
#endif

#ifndef NULL_PTR
#define NULL_PTR ((GENERIC_PTR) 0)
#endif

/* Find the largest host integer type and set its size and type.
   Watch out: on some crazy hosts `long' is shorter than `int'.  */

#ifndef HOST_WIDE_INT
# if HAVE_INTTYPES_H
#  include <inttypes.h>
#  define HOST_WIDE_INT intmax_t
#  define unsigned_HOST_WIDE_INT uintmax_t
# else
#  if (HOST_BITS_PER_LONG <= HOST_BITS_PER_INT && HOST_BITS_PER_LONGLONG <= HOST_BITS_PER_INT)
#   define HOST_WIDE_INT int
#  else
#  if (HOST_BITS_PER_LONGLONG <= HOST_BITS_PER_LONG || ! (defined LONG_LONG_MAX || defined LLONG_MAX))
#   define HOST_WIDE_INT long
#  else
#   define HOST_WIDE_INT long long
#  endif
#  endif
# endif
#endif

#ifndef unsigned_HOST_WIDE_INT
#define unsigned_HOST_WIDE_INT unsigned HOST_WIDE_INT
#endif

#ifndef CHAR_BIT
#define CHAR_BIT 8
#endif

#ifndef HOST_BITS_PER_WIDE_INT
#define HOST_BITS_PER_WIDE_INT (CHAR_BIT * sizeof (HOST_WIDE_INT))
#endif

#if __GNUC__ < 2 || (__GNUC__ == 2 && __GNUC_MINOR__ < 7)
# define __attribute__(x)
#endif

#ifndef PROTO
# if defined (USE_PROTOTYPES) ? USE_PROTOTYPES : defined (__STDC__)
#  define PROTO(ARGS) ARGS
# else
#  define PROTO(ARGS) ()
# endif
#endif

#if defined (__STDC__) && defined (HAVE_VPRINTF)
# include <stdarg.h>
# define VA_START(va_list, var) va_start (va_list, var)
# define PRINTF_ALIST(msg) char *msg, ...
# define PRINTF_DCL(msg)
# define PRINTF_PROTO(ARGS, m, n) PROTO (ARGS) __attribute__ ((format (__printf__, m, n)))
#else
# include <varargs.h>
# define VA_START(va_list, var) va_start (va_list)
# define PRINTF_ALIST(msg) msg, va_alist
# define PRINTF_DCL(msg) char *msg; va_dcl
# define PRINTF_PROTO(ARGS, m, n) () __attribute__ ((format (__printf__, m, n)))
# define vfprintf(file, msg, args) \
    { \
      char *a0 = va_arg(args, char *); \
      char *a1 = va_arg(args, char *); \
      char *a2 = va_arg(args, char *); \
      char *a3 = va_arg(args, char *); \
      fprintf (file, msg, a0, a1, a2, a3); \
    }
#endif

#define PRINTF_PROTO_1(ARGS) PRINTF_PROTO(ARGS, 1, 2)

HOST_WIDE_INT parse_c_expression PROTO((char *, int));

static int yylex PROTO((void));
static void yyerror PROTO((char *)) __attribute__ ((noreturn));
static HOST_WIDE_INT expression_value;
#ifdef TEST_EXP_READER
static int expression_signedp;
#endif

static jmp_buf parse_return_error;

/* Nonzero means count most punctuation as part of a name.  */
static int keyword_parsing = 0;

/* Nonzero means do not evaluate this expression.
   This is a count, since unevaluated expressions can nest.  */
static int skip_evaluation;

/* Nonzero means warn if undefined identifiers are evaluated.  */
static int warn_undef;

/* some external tables of character types */
extern unsigned char is_idstart[], is_idchar[], is_space[];

/* Flag for -pedantic.  */
extern int pedantic;

/* Flag for -traditional.  */
extern int traditional;

/* Flag for -lang-c89.  */
extern int c89;

#ifndef CHAR_TYPE_SIZE
#define CHAR_TYPE_SIZE BITS_PER_UNIT
#endif

#ifndef INT_TYPE_SIZE
#define INT_TYPE_SIZE BITS_PER_WORD
#endif

#ifndef LONG_TYPE_SIZE
#define LONG_TYPE_SIZE BITS_PER_WORD
#endif

#ifndef WCHAR_TYPE_SIZE
#define WCHAR_TYPE_SIZE INT_TYPE_SIZE
#endif

#ifndef MAX_CHAR_TYPE_SIZE
#define MAX_CHAR_TYPE_SIZE CHAR_TYPE_SIZE
#endif

#ifndef MAX_INT_TYPE_SIZE
#define MAX_INT_TYPE_SIZE INT_TYPE_SIZE
#endif

#ifndef MAX_LONG_TYPE_SIZE
#define MAX_LONG_TYPE_SIZE LONG_TYPE_SIZE
#endif

#ifndef MAX_WCHAR_TYPE_SIZE
#define MAX_WCHAR_TYPE_SIZE WCHAR_TYPE_SIZE
#endif

#define MAX_CHAR_TYPE_MASK (MAX_CHAR_TYPE_SIZE < HOST_BITS_PER_WIDE_INT \
			    ? (~ (~ (HOST_WIDE_INT) 0 << MAX_CHAR_TYPE_SIZE)) \
			    : ~ (HOST_WIDE_INT) 0)

#define MAX_WCHAR_TYPE_MASK (MAX_WCHAR_TYPE_SIZE < HOST_BITS_PER_WIDE_INT \
			     ? ~ (~ (HOST_WIDE_INT) 0 << MAX_WCHAR_TYPE_SIZE) \
			     : ~ (HOST_WIDE_INT) 0)

/* Suppose A1 + B1 = SUM1, using 2's complement arithmetic ignoring overflow.
   Suppose A, B and SUM have the same respective signs as A1, B1, and SUM1.
   Suppose SIGNEDP is negative if the result is signed, zero if unsigned.
   Then this yields nonzero if overflow occurred during the addition.
   Overflow occurs if A and B have the same sign, but A and SUM differ in sign,
   and SIGNEDP is negative.
   Use `^' to test whether signs differ, and `< 0' to isolate the sign.  */
#define overflow_sum_sign(a, b, sum, signedp) \
	((~((a) ^ (b)) & ((a) ^ (sum)) & (signedp)) < 0)

struct constant;

GENERIC_PTR xmalloc PROTO((size_t));
HOST_WIDE_INT parse_escape PROTO((char **, HOST_WIDE_INT));
int check_assertion PROTO((U_CHAR *, int, int, struct arglist *));
struct hashnode *lookup PROTO((U_CHAR *, int, int));
void error PRINTF_PROTO_1((char *, ...));
void pedwarn PRINTF_PROTO_1((char *, ...));
void warning PRINTF_PROTO_1((char *, ...));

static int parse_number PROTO((int));
static HOST_WIDE_INT left_shift PROTO((struct constant *, unsigned_HOST_WIDE_INT));
static HOST_WIDE_INT right_shift PROTO((struct constant *, unsigned_HOST_WIDE_INT));
static void integer_overflow PROTO((void));

/* `signedp' values */
#define SIGNED (~0)
#define UNSIGNED 0
%}

%union {
  struct constant {HOST_WIDE_INT value; int signedp;} integer;
  struct name {U_CHAR *address; int length;} name;
  struct arglist *keywords;
}

%type <integer> exp exp1 start
%type <keywords> keywords
%token <integer> INT CHAR
%token <name> NAME
%token <integer> ERROR

%right '?' ':'
%left ','
%left OR
%left AND
%left '|'
%left '^'
%left '&'
%left EQUAL NOTEQUAL
%left '<' '>' LEQ GEQ
%left LSH RSH
%left '+' '-'
%left '*' '/' '%'
%right UNARY

/* %expect 40 */

%%

start   :	exp1
		{
		  expression_value = $1.value;
#ifdef TEST_EXP_READER
		  expression_signedp = $1.signedp;
#endif
		}
	;

/* Expressions, including the comma operator.  */
exp1	:	exp
	|	exp1 ',' exp
			{ if (pedantic)
			    pedwarn ("comma operator in operand of `#if'");
			  $$ = $3; }
	;

/* Expressions, not including the comma operator.  */
exp	:	'-' exp    %prec UNARY
			{ $$.value = - $2.value;
			  $$.signedp = $2.signedp;
			  if (($$.value & $2.value & $$.signedp) < 0)
			    integer_overflow (); }
	|	'!' exp    %prec UNARY
			{ $$.value = ! $2.value;
			  $$.signedp = SIGNED; }
	|	'+' exp    %prec UNARY
			{ $$ = $2; }
	|	'~' exp    %prec UNARY
			{ $$.value = ~ $2.value;
			  $$.signedp = $2.signedp; }
	|	'#' NAME
  			{ $$.value = check_assertion ($2.address, $2.length,
						      0, NULL_PTR);
			  $$.signedp = SIGNED; }
	|	'#' NAME
			{ keyword_parsing = 1; }
		'(' keywords ')'
  			{ $$.value = check_assertion ($2.address, $2.length,
						      1, $5);
			  keyword_parsing = 0;
			  $$.signedp = SIGNED; }
	|	'(' exp1 ')'
			{ $$ = $2; }
	;

/* Binary operators in order of decreasing precedence.  */
exp	:	exp '*' exp
			{ $$.signedp = $1.signedp & $3.signedp;
			  if ($$.signedp)
			    {
			      $$.value = $1.value * $3.value;
			      if ($1.value
				  && ($$.value / $1.value != $3.value
				      || ($$.value & $1.value & $3.value) < 0))
				integer_overflow ();
			    }
			  else
			    $$.value = ((unsigned_HOST_WIDE_INT) $1.value
					* $3.value); }
	|	exp '/' exp
			{ if ($3.value == 0)
			    {
			      if (!skip_evaluation)
				error ("division by zero in #if");
			      $3.value = 1;
			    }
			  $$.signedp = $1.signedp & $3.signedp;
			  if ($$.signedp)
			    {
			      $$.value = $1.value / $3.value;
			      if (($$.value & $1.value & $3.value) < 0)
				integer_overflow ();
			    }
			  else
			    $$.value = ((unsigned_HOST_WIDE_INT) $1.value
					/ $3.value); }
	|	exp '%' exp
			{ if ($3.value == 0)
			    {
			      if (!skip_evaluation)
				error ("division by zero in #if");
			      $3.value = 1;
			    }
			  $$.signedp = $1.signedp & $3.signedp;
			  if ($$.signedp)
			    $$.value = $1.value % $3.value;
			  else
			    $$.value = ((unsigned_HOST_WIDE_INT) $1.value
					% $3.value); }
	|	exp '+' exp
			{ $$.value = $1.value + $3.value;
			  $$.signedp = $1.signedp & $3.signedp;
			  if (overflow_sum_sign ($1.value, $3.value,
						 $$.value, $$.signedp))
			    integer_overflow (); }
	|	exp '-' exp
			{ $$.value = $1.value - $3.value;
			  $$.signedp = $1.signedp & $3.signedp;
			  if (overflow_sum_sign ($$.value, $3.value,
						 $1.value, $$.signedp))
			    integer_overflow (); }
	|	exp LSH exp
			{ $$.signedp = $1.signedp;
			  if (($3.value & $3.signedp) < 0)
			    $$.value = right_shift (&$1, -$3.value);
			  else
			    $$.value = left_shift (&$1, $3.value); }
	|	exp RSH exp
			{ $$.signedp = $1.signedp;
			  if (($3.value & $3.signedp) < 0)
			    $$.value = left_shift (&$1, -$3.value);
			  else
			    $$.value = right_shift (&$1, $3.value); }
	|	exp EQUAL exp
			{ $$.value = ($1.value == $3.value);
			  $$.signedp = SIGNED; }
	|	exp NOTEQUAL exp
			{ $$.value = ($1.value != $3.value);
			  $$.signedp = SIGNED; }
	|	exp LEQ exp
			{ $$.signedp = SIGNED;
			  if ($1.signedp & $3.signedp)
			    $$.value = $1.value <= $3.value;
			  else
			    $$.value = ((unsigned_HOST_WIDE_INT) $1.value
					<= $3.value); }
	|	exp GEQ exp
			{ $$.signedp = SIGNED;
			  if ($1.signedp & $3.signedp)
			    $$.value = $1.value >= $3.value;
			  else
			    $$.value = ((unsigned_HOST_WIDE_INT) $1.value
					>= $3.value); }
	|	exp '<' exp
			{ $$.signedp = SIGNED;
			  if ($1.signedp & $3.signedp)
			    $$.value = $1.value < $3.value;
			  else
			    $$.value = ((unsigned_HOST_WIDE_INT) $1.value
					< $3.value); }
	|	exp '>' exp
			{ $$.signedp = SIGNED;
			  if ($1.signedp & $3.signedp)
			    $$.value = $1.value > $3.value;
			  else
			    $$.value = ((unsigned_HOST_WIDE_INT) $1.value
					> $3.value); }
	|	exp '&' exp
			{ $$.value = $1.value & $3.value;
			  $$.signedp = $1.signedp & $3.signedp; }
	|	exp '^' exp
			{ $$.value = $1.value ^ $3.value;
			  $$.signedp = $1.signedp & $3.signedp; }
	|	exp '|' exp
			{ $$.value = $1.value | $3.value;
			  $$.signedp = $1.signedp & $3.signedp; }
	|	exp AND
			{ skip_evaluation += !$1.value; }
		exp
			{ skip_evaluation -= !$1.value;
			  $$.value = ($1.value && $4.value);
			  $$.signedp = SIGNED; }
	|	exp OR
			{ skip_evaluation += !!$1.value; }
		exp
			{ skip_evaluation -= !!$1.value;
			  $$.value = ($1.value || $4.value);
			  $$.signedp = SIGNED; }
	|	exp '?'
			{ skip_evaluation += !$1.value; }
	        exp ':'
			{ skip_evaluation += !!$1.value - !$1.value; }
		exp
			{ skip_evaluation -= !!$1.value;
			  $$.value = $1.value ? $4.value : $7.value;
			  $$.signedp = $4.signedp & $7.signedp; }
	|	INT
			{ $$ = yylval.integer; }
	|	CHAR
			{ $$ = yylval.integer; }
	|	NAME
			{ if (warn_undef && !skip_evaluation)
			    warning ("`%.*s' is not defined",
				     $1.length, $1.address);
			  $$.value = 0;
			  $$.signedp = SIGNED; }
	;

keywords :
			{ $$ = 0; } 
	|	'(' keywords ')' keywords
			{ struct arglist *temp;
			  $$ = (struct arglist *) xmalloc (sizeof (struct arglist));
			  $$->next = $2;
			  $$->name = (U_CHAR *) "(";
			  $$->length = 1;
			  temp = $$;
			  while (temp != 0 && temp->next != 0)
			    temp = temp->next;
			  temp->next = (struct arglist *) xmalloc (sizeof (struct arglist));
			  temp->next->next = $4;
			  temp->next->name = (U_CHAR *) ")";
			  temp->next->length = 1; }
	|	NAME keywords
			{ $$ = (struct arglist *) xmalloc (sizeof (struct arglist));
			  $$->name = $1.address;
			  $$->length = $1.length;
			  $$->next = $2; } 
	;
%%

/* During parsing of a C expression, the pointer to the next character
   is in this variable.  */

static char *lexptr;

/* Take care of parsing a number (anything that starts with a digit).
   Set yylval and return the token type; update lexptr.
   LEN is the number of characters in it.  */

/* maybe needs to actually deal with floating point numbers */

static int
parse_number (olen)
     int olen;
{
  register char *p = lexptr;
  register int c;
  register unsigned_HOST_WIDE_INT n = 0, nd, max_over_base;
  register int base = 10;
  register int len = olen;
  register int overflow = 0;
  register int digit, largest_digit = 0;
  int spec_long = 0;

  yylval.integer.signedp = SIGNED;

  if (*p == '0') {
    base = 8;
    if (len >= 3 && (p[1] == 'x' || p[1] == 'X')) {
      p += 2;
      base = 16;
      len -= 2;
    }
  }

  max_over_base = (unsigned_HOST_WIDE_INT) -1 / base;

  for (; len > 0; len--) {
    c = *p++;

    if (c >= '0' && c <= '9')
      digit = c - '0';
    else if (base == 16 && c >= 'a' && c <= 'f')
      digit = c - 'a' + 10;
    else if (base == 16 && c >= 'A' && c <= 'F')
      digit = c - 'A' + 10;
    else {
      /* `l' means long, and `u' means unsigned.  */
      while (1) {
	if (c == 'l' || c == 'L')
	  {
	    if (!pedantic < spec_long)
	      yyerror ("too many `l's in integer constant");
	    spec_long++;
	  }
	else if (c == 'u' || c == 'U')
	  {
	    if (! yylval.integer.signedp)
	      yyerror ("two `u's in integer constant");
	    yylval.integer.signedp = UNSIGNED;
	  }
	else {
	  if (c == '.' || c == 'e' || c == 'E' || c == 'p' || c == 'P')
	    yyerror ("Floating point numbers not allowed in #if expressions");
	  else {
	    char *buf = (char *) alloca (p - lexptr + 40);
	    sprintf (buf, "missing white space after number `%.*s'",
		     (int) (p - lexptr - 1), lexptr);
	    yyerror (buf);
	  }
	}

	if (--len == 0)
	  break;
	c = *p++;
      }
      /* Don't look for any more digits after the suffixes.  */
      break;
    }
    if (largest_digit < digit)
      largest_digit = digit;
    nd = n * base + digit;
    overflow |= (max_over_base < n) | (nd < n);
    n = nd;
  }

  if (base <= largest_digit)
    pedwarn ("integer constant contains digits beyond the radix");

  if (overflow)
    pedwarn ("integer constant out of range");

  /* If too big to be signed, consider it unsigned.  */
  if (((HOST_WIDE_INT) n & yylval.integer.signedp) < 0)
    {
      if (base == 10)
	warning ("integer constant is so large that it is unsigned");
      yylval.integer.signedp = UNSIGNED;
    }

  lexptr = p;
  yylval.integer.value = n;
  return INT;
}

struct token {
  char *operator;
  int token;
};

static struct token tokentab2[] = {
  {"&&", AND},
  {"||", OR},
  {"<<", LSH},
  {">>", RSH},
  {"==", EQUAL},
  {"!=", NOTEQUAL},
  {"<=", LEQ},
  {">=", GEQ},
  {"++", ERROR},
  {"--", ERROR},
  {NULL, ERROR}
};

/* Read one token, getting characters through lexptr.  */

static int
yylex ()
{
  register int c;
  register int namelen;
  register unsigned char *tokstart;
  register struct token *toktab;
  int wide_flag;
  HOST_WIDE_INT mask;

 retry: