Question

I have a class called PointF and it has a constructor that takes a Point, and I keep getting “possible loss of data warnings”. How can I show that my intention is in fact to make a float value into an int, among other tings? I tried static_cast and (float) but they did not fix the warning.

For example:

int curPos = ceil(float(newMouseY / font.getLineHeight())) ; //float to int

And

outputChars[0] = uc; //converting from size_t to char

Answer 1

A cast should do the trick; that says “explicitly make this type into that type”, which is generally pretty silly for a compiler to warn for:

int curPos = static_cast<int>(ceil(float(newMouseY / font.getLineHeight())));

Or:

outputChars[0] = static_cast<char>(uc);

Make sure the casts you tried were akin to that. You say “I tried …(float)” which leads me to believe you tried something like this:

int curPos = (float)(ceil(float(newMouseY / font.getLineHeight())));

Which does nothing. The type of the expression is already a float, what would casting it to the same type do? You need to cast it to the destination type.

---

Keep in mind casts are generally to be avoided. In your first snippet, the cast is sensible because when you quantize something you necessarily need to drop information.

But your second cast is not. Why is uc a size_t in the first place? What happens when the cast does drop information? Is that really a good idea?