Show Lecture.Limits as a slide show.

CS253 Limits

C integral types

<limits.h>, alias <climits>

Symbol	Meaning
CHAR_BIT	char: number of bits
SCHAR_MIN…SCHAR_MAX	signed char: range
UCHAR_MAX	unsigned char: max value
CHAR_MIN…CHAR_MAX	char: range
SHRT_MIN…SHRT_MAX	short int: range
USHRT_MAX	unsigned short int: max value
INT_MIN…INT_MAX	int: range
UINT_MAX	unsigned int: max value
LONG_MIN…LONG_MAX	long int: range
ULONG_MAX	unsigned long int: max value
LLONG_MIN…LLONG_MAX	long long int: range
ULLONG_MAX	unsigned long long int: max value

C floating-point types

<float.h>, alias <cfloat>

Symbol	Meaning
FLT_DIG	float: mantissa decimal digits
DBL_DIG	double: mantissa decimal digits
LDBL_DIG	long double: mantissa decimal digits
FLT_MIN_10_EXP…FLT_MAX_10_EXP	float: exponent range
DBL_MIN_10_EXP…DBL_MAX_10_EXP	double exponent range
LDBL_MIN_10_EXP…LDBL_MAX_10_EXP	long double: exponent range
FLT_MIN…FLT_MAX	float: value range
DBL_MIN…DBL_MAX	double: value range
LDBL_MIN…LDBL_MAX	long double: value range

Pop Quiz

Quick, now:

• What symbol says how many decimal digits are in the exponent of a long double?
• What symbol gives the maximum value of a short?
• What symbol gives the maximum value of a size_t?

C++ way

• <limits> defines the template class numeric_limits, which is specialized for all built-in types.
• It has many static methods and constants, including:
  • .min()
  • .max()
  • .digits10
  • .min_exponent10
  • .max_exponent10
• It’s unclear why .min() & .max() are methods, not variables like .min_exponent10.
• Surely, these values are known at compile-time.
• In fact, they’re all constexpr methods & constants. Fine.

Simple Example

What is the range of values for a char?

constexpr int smallest = numeric_limits<char>::min();
constexpr int largest = numeric_limits<char>::max();
cout << "A plain char value ranges from "
     << smallest << " to " << largest << ".\n";

A plain char value ranges from -128 to 127.

• Sure, on today's computers, a char is almost always eight bits, but not always!
• Also, is it signed or unsigned?
• Guessing is a poor way to know--use numeric_limits!

Integral Example

cout << numeric_limits<int>::digits10 << '\n'
     << numeric_limits<int>::min()    << '\n'
     << numeric_limits<int>::max()    << '\n';

9
-2147483648
2147483647

• Note the ::. We’d use . if we had an instance of this class.
• digits10 is a constant, not a method.

Floating-Point Examples

typedef numeric_limits<float> nl; // alias
cout << nl::digits10 << '\n'
     << nl::min()    << '\n'
     << nl::max()    << '\n';

6
1.17549e-38
3.40282e+38

using nl = numeric_limits<double>; // alias
cout << nl::digits10 << '\n'
     << nl::min()    << '\n'
     << nl::max()    << '\n';

15
2.22507e-308
1.79769e+308

numeric_limits<long double> nl; // instance
cout << nl.digits10 << '\n'
     << nl.min()    << '\n'
     << nl.max()    << '\n';

18
3.3621e-4932
1.18973e+4932

A Mystery

Why do these both compile? One must be wrong!

cout << numeric_limits<size_t>::min   << '\n' 
     << numeric_limits<size_t>::min() << '\n';

1
0