• Nick (19/19) Aug 01 2004 Should this really be allowed?
• Matthew (2/21) Aug 01 2004

Nick <Nick_member pathlink.com> writes:

Should this really be allowed?

creal c = 3 + 5i;
double d = c;    // d is now 3

Besides being a lossy conversion, it can make it difficult to write a complex
math library, and lead to some silent bugs if there ever comes to BE a library.

For example:
creal exp(creal);
real exp(real);
..
cdouble d = 3 + 4i;
exp(d); // The compiler chokes on this one, so we can't overload names

Another example: If we rename the complex functions (as done in C99) we get
real exp(real);
creal cexp(creal);
..
creal d = 2i;
exp(d); // Typo, should be cexp, but the compiler accepts and exp returns the
wrong answer.

Nick

"Matthew" <admin stlsoft.dot.dot.dot.dot.org> writes:

"Nick" <Nick_member pathlink.com> wrote in message
news:cejsno$2g2s$1 digitaldaemon.com...
 Should this really be allowed?

 creal c = 3 + 5i;
 double d = c;    // d is now 3

 Besides being a lossy conversion, it can make it difficult to write a complex
 math library, and lead to some silent bugs if there ever comes to BE a library.

Should require explicit cast  IMO

 For example:
 creal exp(creal);
 real exp(real);
 ..
 cdouble d = 3 + 4i;
 exp(d); // The compiler chokes on this one, so we can't overload names

 Another example: If we rename the complex functions (as done in C99) we get
 real exp(real);
 creal cexp(creal);
 ..
 creal d = 2i;
 exp(d); // Typo, should be cexp, but the compiler accepts and exp returns the
 wrong answer.

 Nick