Problem with <limit>

Code:

---

#include <iostream>
using namespace std;

#include <limits>

int main() {

  cout << "smallest char ==" << numeric_limits<char>::min() << '\n'
  << "largest char ==" << numeric_limits<char>::max() << '\n'
  << "smallest short ==" << numeric_limits<short>::min() << '\n'
  << "largest short ==" << numeric_limits<short>::max() << '\n';

}

---

This code compiles (and doesn't generate warnings with -Wall), but the output is:

Code:

---

smallest char ==
largest char =
smallest short ==-32768
largest short ==32767

---

The point of this exercise is to print out the implementation-specific lower and upper bounds of each (primitive) type. What's going on with `char'?

These are the same:

Code:

---

    cout << numeric_limits<char>::min();
    cout << '\0';

---

What would you expect to see?

If you want to see an integer value, cast to an int.

gg

Quote:

---

These are the same:

Code:
cout << numeric_limits<char>::min();
cout << '\0';What would you expect to see?

If you want to see an integer value, cast to an int.

gg

---

Not sure if we're on same page

Idea is to extract lower and upper bounds on `char' for this machine in a representation-independent way.

I was expecting to see

smallest char == 0
largest char == 255

for an ANSI-compliant implementation.

Here's the rest:

Code:

---

#include <iostream>
using namespace std;

#include <limits>

int main() {

  cout << "smallest char ==" << numeric_limits<char>::min() << '\n' // Not behaving like others
  << "largest char ==" << numeric_limits<char>::max() << '\n' // Not behaving like others
  << "smallest short ==" << numeric_limits<short>::min() << '\n'
  << "largest short ==" << numeric_limits<short>::max() << '\n'
  << "smallest int ==" << numeric_limits<int>::min() << '\n'
  << "largest int ==" << numeric_limits<int>::max() << '\n'
  << "smallest long ==" << numeric_limits<long>::min() << '\n'
  << "largest long ==" << numeric_limits<long>::max() << '\n'
  << "smallest float ==" << numeric_limits<float>::min() << '\n'
  << "largest float ==" << numeric_limits<float>::max() << '\n'
  << "smallest double ==" << numeric_limits<double>::min() << '\n'
  << "largest double ==" << numeric_limits<double>::max() << '\n'
  << "smallest long double ==" << numeric_limits<long double>::min() << '\n'
  << "largest long double ==" << numeric_limits<long double>::max() << '\n'
  << "smallest unsigned ==" << numeric_limits<unsigned>::min() << '\n'
      << "largest unsigned ==" << numeric_limits<unsigned>::max() << '\n';


}

---

Output:

Code:

---

smallest char ==
largest char =
smallest short ==-32768
largest short ==32767
smallest int ==-2147483648
largest int ==2147483647
smallest long ==-2147483648
largest long ==2147483647
smallest float ==1.17549e-38
largest float ==3.40282e+38
smallest double ==2.22507e-308
largest double ==1.79769e+308
smallest long double ==3.3621e-4932
largest long double ==1.18973e+4932
smallest unsigned ==0
largest unsigned ==4294967295

---

No, you're not on the same page, slippy. Read his post again. He's given you the answer.

Yeah, but not in a very clear way. (And besides, it's wrong in a detail.)

The C++ I/O streams interpret every value of type char as a character. You want the stream to interpret it as an integer. You'll have to cast it explicitly.

Code:

---

std::cout << static_cast<int>(std::numeric_limits<char>::min()) << '\n';

---

You'll also find that the limit is not necessarily 0. The signed-ness of char is implementation-defined. (In fact, those I know use a signed 8-bit char, so the min will actually be -128.) That's why char, signed char and unsigned char are three distinct types.

Quote:

---

Originally Posted by CornedBee

Yeah, but not in a very clear way. (And besides, it's wrong in a detail.)

The C++ I/O streams interpret every value of type char as a character. You want the stream to interpret it as an integer. You'll have to cast it explicitly.

Code:

---

std::cout << static_cast<int>(std::numeric_limits<char>::min()) << '\n';

---

You'll also find that the limit is not necessarily 0. The signed-ness of char is implementation-defined. (In fact, those I know use a signed 8-bit char, so the min will actually be -128.) That's why char, signed char and unsigned char are three distinct types.

---

But of course, all of zero, 255, 127 and -128 may are likely to be invisible characters [127 is "delete" for example]. I know that 128 is not invisible in Windows [it's C-cedilla in my codepage], but the others are in one way or another.

--
Mats

Quote:

---

Originally Posted by CornedBee

Yeah, but not in a very clear way. (And besides, it's wrong in a detail.)

---

The cast is what is important.

Quote:

---

Code:
std::cout << static_cast<int>(std::numeric_limits<char>::min()) << '\n';

---

Thanks!