🚀 go-pugleaf

The following program generates the errors:

/tmp/K.C: In function `double  hypot (double, double)':
/tmp/K.C:21: warning: float or double assigned to integer data type
/tmp/K.C:21: warning: float or double assigned to integer data type

However, these conversions should not have been selected at all, and
cfront-3.0.1 correctly calls the double sqrt (double x) routine. The
problem comes to light on a sun-3 when you make libg++, because the
math-68881.h header contains equivalent code, but it would appear that
all recent g++ compilers actually have the bug. I compiled this
particular example on a sun-4 with gcc-2.3.3. Any ideas anyone...?
 --------------------------------------------------------------------------



class Integer
{
public:
                  Integer(long);
                  Integer(const Integer&);
                  operator long() const;
                  operator double() const;
};

extern Integer  sqrt(const Integer&);  

extern "C" {			 
   extern double sqrt (double x);

}

/*extern "C" { */
    double hypot ( double x,  double y)
      {
	double temp= x*x + y*y;
	temp = sqrt(temp);
	return temp;
      }

/*}*/



--
Jeff Smith, Computer Science, Warwick University, Coventry, CV4 7AL, England
jeff@dcs.warwick.ac.uk	phone: +44 203 523485	fax: +44 203 525714