m2-exp.y

早期freebsd实现

				{
				  write_exp_elt_opcode (OP_LONG);
				  write_exp_elt_type (builtin_type_int);
				  write_exp_elt_longcst ((LONGEST) msymbol -> address);
				  write_exp_elt_opcode (OP_LONG);
				  write_exp_elt_opcode (UNOP_MEMVAL);
				  if (msymbol -> type == mst_data ||
				      msymbol -> type == mst_bss)
				    write_exp_elt_type (builtin_type_int);
				  else if (msymbol -> type == mst_text)
				    write_exp_elt_type (lookup_function_type (builtin_type_int));
				  else
				    write_exp_elt_type (builtin_type_char);
				  write_exp_elt_opcode (UNOP_MEMVAL);
				}
			      else if (!have_full_symbols () && !have_partial_symbols ())
				error ("No symbol table is loaded.  Use the \"symbol-file\" command.");
			      else
				error ("No symbol \"%s\" in current context.",
				       copy_name ($1));
			    }
			}
	;

type
	:	TYPENAME
			{ $$ = lookup_typename (copy_name ($1),
						expression_context_block, 0); }

	;

%%

#if 0  /* FIXME! */
int
overflow(a,b)
   long a,b;
{
   return (MAX_OF_TYPE(builtin_type_m2_int) - b) < a;
}

int
uoverflow(a,b)
   unsigned long a,b;
{
   return (MAX_OF_TYPE(builtin_type_m2_card) - b) < a;
}
#endif /* FIXME */

/* 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.  */

/*** Needs some error checking for the float case ***/

static int
parse_number (olen)
     int olen;
{
  register char *p = lexptr;
  register LONGEST n = 0;
  register LONGEST prevn = 0;
  register int c,i,ischar=0;
  register int base = input_radix;
  register int len = olen;
  int unsigned_p = number_sign == 1 ? 1 : 0;

  if(p[len-1] == 'H')
  {
     base = 16;
     len--;
  }
  else if(p[len-1] == 'C' || p[len-1] == 'B')
  {
     base = 8;
     ischar = p[len-1] == 'C';
     len--;
  }

  /* Scan the number */
  for (c = 0; c < len; c++)
  {
    if (p[c] == '.' && base == 10)
      {
	/* It's a float since it contains a point.  */
	yylval.dval = atof (p);
	lexptr += len;
	return FLOAT;
      }
    if (p[c] == '.' && base != 10)
       error("Floating point numbers must be base 10.");
    if (base == 10 && (p[c] < '0' || p[c] > '9'))
       error("Invalid digit \'%c\' in number.",p[c]);
 }

  while (len-- > 0)
    {
      c = *p++;
      n *= base;
      if( base == 8 && (c == '8' || c == '9'))
	 error("Invalid digit \'%c\' in octal number.",c);
      if (c >= '0' && c <= '9')
	i = c - '0';
      else
	{
	  if (base == 16 && c >= 'A' && c <= 'F')
	    i = c - 'A' + 10;
	  else
	     return ERROR;
	}
      n+=i;
      if(i >= base)
	 return ERROR;
      if(!unsigned_p && number_sign == 1 && (prevn >= n))
	 unsigned_p=1;		/* Try something unsigned */
      /* Don't do the range check if n==i and i==0, since that special
	 case will give an overflow error. */
      if(RANGE_CHECK && n!=i && i)
      {
	 if((unsigned_p && (unsigned)prevn >= (unsigned)n) ||
	    ((!unsigned_p && number_sign==-1) && -prevn <= -n))
	    range_error("Overflow on numeric constant.");
      }
	 prevn=n;
    }

  lexptr = p;
  if(*p == 'B' || *p == 'C' || *p == 'H')
     lexptr++;			/* Advance past B,C or H */

  if (ischar)
  {
     yylval.ulval = n;
     return CHAR;
  }
  else if ( unsigned_p && number_sign == 1)
  {
     yylval.ulval = n;
     return UINT;
  }
  else if((unsigned_p && (n<0))) {
     range_error("Overflow on numeric constant -- number too large.");
     /* But, this can return if range_check == range_warn.  */
  }
  yylval.lval = n;
  return INT;
}


/* Some tokens */

static struct
{
   char name[4];
   int token;
} tokentab2[] =
{
    { {'<', '>'},    NOTEQUAL 	},
    { {':', '='},    ASSIGN	},
    { {'<', '='},    LEQ	},
    { {'>', '='},    GEQ	},
    { {':', ':'},    COLONCOLON },

};

/* Some specific keywords */

struct keyword {
   char keyw[10];
   int token;
};

static struct keyword keytab[] =
{
    {"OR" ,   OROR	 },
    {"IN",    IN         },/* Note space after IN */
    {"AND",   ANDAND     },
    {"ABS",   ABS	 },
    {"CHR",   CHR	 },
    {"DEC",   DEC	 },
    {"NOT",   NOT	 },
    {"DIV",   DIV    	 },
    {"INC",   INC	 },
    {"MAX",   MAX_FUNC	 },
    {"MIN",   MIN_FUNC	 },
    {"MOD",   MOD	 },
    {"ODD",   ODD	 },
    {"CAP",   CAP	 },
    {"ORD",   ORD	 },
    {"VAL",   VAL	 },
    {"EXCL",  EXCL	 },
    {"HIGH",  HIGH       },
    {"INCL",  INCL	 },
    {"SIZE",  SIZE       },
    {"FLOAT", FLOAT_FUNC },
    {"TRUNC", TRUNC	 },
};


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

/* This is where we will check to make sure that the language and the operators used are
   compatible  */

static int
yylex ()
{
  register int c;
  register int namelen;
  register int i;
  register char *tokstart;
  register char quote;

 retry:

  tokstart = lexptr;


  /* See if it is a special token of length 2 */
  for( i = 0 ; i < sizeof tokentab2 / sizeof tokentab2[0] ; i++)
     if(!strncmp(tokentab2[i].name, tokstart, 2))
     {
	lexptr += 2;
	return tokentab2[i].token;
     }

  switch (c = *tokstart)
    {
    case 0:
      return 0;

    case ' ':
    case '\t':
    case '\n':
      lexptr++;
      goto retry;

    case '(':
      paren_depth++;
      lexptr++;
      return c;

    case ')':
      if (paren_depth == 0)
	return 0;
      paren_depth--;
      lexptr++;
      return c;

    case ',':
      if (comma_terminates && paren_depth == 0)
	return 0;
      lexptr++;
      return c;

    case '.':
      /* Might be a floating point number.  */
      if (lexptr[1] >= '0' && lexptr[1] <= '9')
	break;			/* Falls into number code.  */
      else
      {
	 lexptr++;
	 return DOT;
      }

/* These are character tokens that appear as-is in the YACC grammar */
    case '+':
    case '-':
    case '*':
    case '/':
    case '^':
    case '<':
    case '>':
    case '[':
    case ']':
    case '=':
    case '{':
    case '}':
    case '#':
    case '@':
    case '~':
    case '&':
      lexptr++;
      return c;

    case '\'' :
    case '"':
      quote = c;
      for (namelen = 1; (c = tokstart[namelen]) != quote && c != '\0'; namelen++)
	if (c == '\\')
	  {
	    c = tokstart[++namelen];
	    if (c >= '0' && c <= '9')
	      {
		c = tokstart[++namelen];
		if (c >= '0' && c <= '9')
		  c = tokstart[++namelen];
	      }
	  }
      if(c != quote)
	 error("Unterminated string or character constant.");
      yylval.sval.ptr = tokstart + 1;
      yylval.sval.length = namelen - 1;
      lexptr += namelen + 1;

      if(namelen == 2)  	/* Single character */
      {
	   yylval.ulval = tokstart[1];
	   return CHAR;
      }
      else
	 return STRING;
    }

  /* Is it a number?  */
  /* Note:  We have already dealt with the case of the token '.'.
     See case '.' above.  */
  if ((c >= '0' && c <= '9'))
    {
      /* It's a number.  */
      int got_dot = 0, got_e = 0;
      register char *p = tokstart;
      int toktype;

      for (++p ;; ++p)
	{
	  if (!got_e && (*p == 'e' || *p == 'E'))
	    got_dot = got_e = 1;
	  else if (!got_dot && *p == '.')
	    got_dot = 1;
	  else if (got_e && (p[-1] == 'e' || p[-1] == 'E')
		   && (*p == '-' || *p == '+'))
	    /* This is the sign of the exponent, not the end of the
	       number.  */
	    continue;
	  else if ((*p < '0' || *p > '9') &&
		   (*p < 'A' || *p > 'F') &&
		   (*p != 'H'))  /* Modula-2 hexadecimal number */
	    break;
	}
	toktype = parse_number (p - tokstart);
        if (toktype == ERROR)
	  {
	    char *err_copy = (char *) alloca (p - tokstart + 1);

	    memcpy (err_copy, tokstart, p - tokstart);
	    err_copy[p - tokstart] = 0;
	    error ("Invalid number \"%s\".", err_copy);
	  }
	lexptr = p;
	return toktype;
    }

  if (!(c == '_' || c == '$'
	|| (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z')))
    /* We must have come across a bad character (e.g. ';').  */
    error ("Invalid character '%c' in expression.", c);

  /* It's a name.  See how long it is.  */
  namelen = 0;
  for (c = tokstart[namelen];
       (c == '_' || c == '$' || (c >= '0' && c <= '9')
	|| (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z'));
       c = tokstart[++namelen])
    ;

  /* The token "if" terminates the expression and is NOT
     removed from the input stream.  */
  if (namelen == 2 && tokstart[0] == 'i' && tokstart[1] == 'f')
    {
      return 0;
    }

  lexptr += namelen;

  /* Handle the tokens $digits; also $ (short for $0) and $$ (short for $$1)
     and $$digits (equivalent to $<-digits> if you could type that).
     Make token type LAST, and put the number (the digits) in yylval.  */

  if (*tokstart == '$')
    {
      register int negate = 0;
      c = 1;
      /* Double dollar means negate the number and add -1 as well.
	 Thus $$ alone means -1.  */
      if (namelen >= 2 && tokstart[1] == '$')
	{
	  negate = 1;
	  c = 2;
	}
      if (c == namelen)
	{
	  /* Just dollars (one or two) */
	  yylval.lval = - negate;
	  return LAST;
	}
      /* Is the rest of the token digits?  */
      for (; c < namelen; c++)
	if (!(tokstart[c] >= '0' && tokstart[c] <= '9'))
	  break;
      if (c == namelen)
	{
	  yylval.lval = atoi (tokstart + 1 + negate);
	  if (negate)
	    yylval.lval = - yylval.lval;
	  return LAST;
	}
    }

  /* Handle tokens that refer to machine registers:
     $ followed by a register name.  */

  if (*tokstart == '$') {
    for (c = 0; c < NUM_REGS; c++)
      if (namelen - 1 == strlen (reg_names[c])
	  && !strncmp (tokstart + 1, reg_names[c], namelen - 1))
	{
	  yylval.lval = c;
	  return REGNAME;
	}
    for (c = 0; c < num_std_regs; c++)
     if (namelen - 1 == strlen (std_regs[c].name)
	 && !strncmp (tokstart + 1, std_regs[c].name, namelen - 1))
       {
	 yylval.lval = std_regs[c].regnum;
	 return REGNAME;
       }
  }


  /*  Lookup special keywords */
  for(i = 0 ; i < sizeof(keytab) / sizeof(keytab[0]) ; i++)
     if(namelen == strlen(keytab[i].keyw) && !strncmp(tokstart,keytab[i].keyw,namelen))
	   return keytab[i].token;

  yylval.sval.ptr = tokstart;
  yylval.sval.length = namelen;

  /* Any other names starting in $ are debugger internal variables.  */

  if (*tokstart == '$')
    {
      yylval.ivar = (struct internalvar *) lookup_internalvar (copy_name (yylval.sval) + 1);
      return INTERNAL_VAR;
    }


  /* Use token-type BLOCKNAME for symbols that happen to be defined as
     functions.  If this is not so, then ...
     Use token-type TYPENAME for symbols that happen to be defined
     currently as names of types; NAME for other symbols.
     The caller is not constrained to care about the distinction.  */
 {


    char *tmp = copy_name (yylval.sval);
    struct symbol *sym;

    if (lookup_partial_symtab (tmp))
      return BLOCKNAME;
    sym = lookup_symbol (tmp, expression_context_block,
			 VAR_NAMESPACE, 0, NULL);
    if (sym && SYMBOL_CLASS (sym) == LOC_BLOCK)
      return BLOCKNAME;
    if (lookup_typename (copy_name (yylval.sval), expression_context_block, 1))
      return TYPENAME;

    if(sym)
    {
       switch(sym->class)
       {
       case LOC_STATIC:
       case LOC_REGISTER:
       case LOC_ARG:
       case LOC_REF_ARG:
       case LOC_REGPARM:
       case LOC_LOCAL:
       case LOC_LOCAL_ARG:
       case LOC_CONST:
       case LOC_CONST_BYTES:
	  return NAME;

       case LOC_TYPEDEF:
	  return TYPENAME;

       case LOC_BLOCK:
	  return BLOCKNAME;

       case LOC_UNDEF:
	  error("internal:  Undefined class in m2lex()");

       case LOC_LABEL:
	  error("internal:  Unforseen case in m2lex()");
       }
    }
    else
    {
       /* Built-in BOOLEAN type.  This is sort of a hack. */
       if(!strncmp(tokstart,"TRUE",4))
       {
	  yylval.ulval = 1;
	  return M2_TRUE;
       }
       else if(!strncmp(tokstart,"FALSE",5))
       {
	  yylval.ulval = 0;
	  return M2_FALSE;
       }
    }

    /* Must be another type of name... */
    return NAME;
 }
}

#if 0		/* Unused */
static char *
make_qualname(mod,ident)
   char *mod, *ident;
{
   char *new = xmalloc(strlen(mod)+strlen(ident)+2);

   strcpy(new,mod);
   strcat(new,".");
   strcat(new,ident);
   return new;
}
#endif  /* 0 */

static void
yyerror(msg)
     char *msg;	/* unused */
{
   printf("Parsing:  %s\n",lexptr);
   if (yychar < 256)
     error("Invalid syntax in expression near character '%c'.",yychar);
   else
     error("Invalid syntax in expression");
}

/* Table of operators and their precedences for printing expressions.  */

const static struct op_print m2_op_print_tab[] = {
    {"+",   BINOP_ADD, PREC_ADD, 0},
    {"+",   UNOP_PLUS, PREC_PREFIX, 0},
    {"-",   BINOP_SUB, PREC_ADD, 0},
    {"-",   UNOP_NEG, PREC_PREFIX, 0},
    {"*",   BINOP_MUL, PREC_MUL, 0},
    {"/",   BINOP_DIV, PREC_MUL, 0},
    {"DIV", BINOP_INTDIV, PREC_MUL, 0},
    {"MOD", BINOP_REM, PREC_MUL, 0},
    {":=",  BINOP_ASSIGN, PREC_ASSIGN, 1},
    {"OR",  BINOP_OR, PREC_OR, 0},
    {"AND", BINOP_AND, PREC_AND, 0},
    {"NOT", UNOP_ZEROP, PREC_PREFIX, 0},
    {"=",   BINOP_EQUAL, PREC_EQUAL, 0},
    {"<>",  BINOP_NOTEQUAL, PREC_EQUAL, 0},
    {"<=",  BINOP_LEQ, PREC_ORDER, 0},
    {">=",  BINOP_GEQ, PREC_ORDER, 0},
    {">",   BINOP_GTR, PREC_ORDER, 0},
    {"<",   BINOP_LESS, PREC_ORDER, 0},
    {"^",   UNOP_IND, PREC_PREFIX, 0},
    {"@",   BINOP_REPEAT, PREC_REPEAT, 0},
};

/* The built-in types of Modula-2.  */

struct type *builtin_type_m2_char;
struct type *builtin_type_m2_int;
struct type *builtin_type_m2_card;
struct type *builtin_type_m2_real;
struct type *builtin_type_m2_bool;

struct type ** const (m2_builtin_types[]) = 
{
  &builtin_type_m2_char,
  &builtin_type_m2_int,
  &builtin_type_m2_card,
  &builtin_type_m2_real,
  &builtin_type_m2_bool,
  0
};

const struct language_defn m2_language_defn = {
  "modula-2",
  language_m2,
  m2_builtin_types,
  range_check_on,
  type_check_on,
  m2_parse,			/* parser */
  m2_error,			/* parser error function */
  &builtin_type_m2_int,		/* longest signed   integral type */
  &builtin_type_m2_card,		/* longest unsigned integral type */
  &builtin_type_m2_real,		/* longest floating point type */
  "0%XH", "0%", "XH",		/* Hex   format string, prefix, suffix */
  "%oB",  "%",  "oB",		/* Octal format string, prefix, suffix */
  m2_op_print_tab,		/* expression operators for printing */
  LANG_MAGIC
};

/* Initialization for Modula-2 */

void
_initialize_m2_exp ()
{
  /* Modula-2 "pervasive" types.  NOTE:  these can be redefined!!! */
  builtin_type_m2_int =
    init_type (TYPE_CODE_INT, TARGET_INT_BIT / TARGET_CHAR_BIT,
	       0,
	       "INTEGER", (struct objfile *) NULL);
  builtin_type_m2_card =
    init_type (TYPE_CODE_INT, TARGET_INT_BIT / TARGET_CHAR_BIT,
	       TYPE_FLAG_UNSIGNED,
	       "CARDINAL", (struct objfile *) NULL);
  builtin_type_m2_real =
    init_type (TYPE_CODE_FLT, TARGET_FLOAT_BIT / TARGET_CHAR_BIT,
	       0,
	       "REAL", (struct objfile *) NULL);
  builtin_type_m2_char =
    init_type (TYPE_CODE_CHAR, TARGET_CHAR_BIT / TARGET_CHAR_BIT,
	       TYPE_FLAG_UNSIGNED,
	       "CHAR", (struct objfile *) NULL);
  builtin_type_m2_bool =
    init_type (TYPE_CODE_BOOL, TARGET_INT_BIT / TARGET_CHAR_BIT,
	       TYPE_FLAG_UNSIGNED,
	       "BOOLEAN", (struct objfile *) NULL);

  TYPE_NFIELDS(builtin_type_m2_bool) = 2;
  TYPE_FIELDS(builtin_type_m2_bool) = 
     (struct field *) malloc (sizeof (struct field) * 2);
  TYPE_FIELD_BITPOS(builtin_type_m2_bool,0) = 0;
  TYPE_FIELD_NAME(builtin_type_m2_bool,0) = (char *)malloc(6);
  strcpy(TYPE_FIELD_NAME(builtin_type_m2_bool,0),"FALSE");
  TYPE_FIELD_BITPOS(builtin_type_m2_bool,1) = 1;
  TYPE_FIELD_NAME(builtin_type_m2_bool,1) = (char *)malloc(5);
  strcpy(TYPE_FIELD_NAME(builtin_type_m2_bool,1),"TRUE");

  add_language (&m2_language_defn);
}