QVariant doesn't properly convert floats and doubles stored within it to QString.

QVariant doesn't properly convert floats and doubles stored within it to QString.

For example for this float stored within QVariant: 0.035003945
After conversion to QString the string looks like: 0.0350039
After storing string to QVariant and getting float you get: 0.035003901

This is obviously not what is expected and unneeded and unnecessary lose of precision happens.

The Qt implementation internally incorrectly uses: FLT_DIG and DBL_DIG which is only storing number of digits which is always representable by the number (minimum guaranteed).
However this is not maximum guaranteed digits. For C+11 correct solution is to use std::numeric_limits<float>::max_digits10(). If you don't have C+11 support all the time, you need to specify the number manually!