roundf function gives error in clang but runs smoothly in g++

Question

while(a<0)
{
    printf("\n O hai! How much change is owed?");
    scanf("%f",&a);
}

b=(int)roundf(float(a*100));

The above is a part of my code. Clang gives an error 'expected expression' even though I've followed the syntax clearly by referring to the manual(man round). However, when I compile with g++, it doesn't give me an error at all.

Please help!

Answer 1

Regarding the first error b=(int)roundf(float(a*100)); should be following, b=(int)roundf((float)(a*100)); However, typecasting to float is not needed as 'a' is already in float.

Roundf is from math.h library, you need to link that library while compiling. Try following:

clang -o greedy greedy.c -lm

'-lm' tells the linker to use math library. Try this two things and post what happens.

Answer 2

I believe you're trying to do one of these two things (or probably both).

1. To cast the result of the expression a * 100 to float.

2. To pass the expression a * 100 to roundf() assuming that it expects a float as an argument.

If you're trying to cast a value to float, then the correct syntax for doing this is

(float)value;

For example

int x = 10;
float f = (float)x; // cast the value in x to float

If rather you're passing the expression to roundf(), then you don't really need to include the datatype. Just the variable name/the value/the expression.

For example

float f = 10.5;
float r0 = roundf(f); // passing a variable
float r1 = roundf(9.3); // passing a value
float r2 = roundf(5.2 * 100); //passing an expression

That's a thing! Another thing is that you should be printing a newline and a white space before "O hai! How much change is owed?". And probably you should use GetFloat() from the CS50 library since it guarantees that you'll get a float as input from the user.