Significant decimal digits of double?

Question

What does the C/C++-Standard guarantee about the minimum significant decimal digits of a double?

The C++ standard actually says in a footnote of § 18.3.2.4/10 this is equivalent to DBL_DIG as defined in the C standard. But I can't find anything about a minimum value in the C standard.

So, what is the minimum value of one of the following?

• std::numeric_limits<double>::digits10
• DBL_MANT_DIG edit: No, DBL_DIG

Answer 1

You won't find it in the C++ standard, because C++ defers to the C standard on this one. For decimal digits, the minimum values are:

FLT_DIG   6
DBL_DIG  10
LDBL_DIG 10

So, 6 digits for float, and 10 for double and long double. (This is the number of digits for which it is guaranteed that a conversion from text to the type and back will result in the same value.)

Note that DBL_MANT_DIG corresponds to the number of digits in the base, thus usually the number of binary digits, and not the number of decimal digits.