`toml::format` returns the wrong floating value about toml11

[ToruNiina]

This may be unexpected behavior, but the value is actually correct.

In binary numbers, 3.14 cannot be expressed in a finite number of digits.
Computers use binary numbers, so we get 3.1400000000000001 when you write 3.14 as a double-precision floating point number. This value is the closest number to 3.14 of all the values that can be expressed in double.
cf. https://wandbox.org/permlink/RXlDyxSXM3II6q53

By default, the toml::format outputs the most accurate value possible. In other words, it uses std::numeric_limits<double>::max_digits10 as its precision.
To change this setting and perform rounding on a specific digit, do the following.

const std::string fmt = toml::format(v, /*line width = */80, /*precision = */16);

To print 3.1400000000000001 as 3.14, 16 is small enough.

[ups-ishii]

Understood. I learned a lot.
Thank you very much for your very detailed explanation!