You need to modify the print function in print.h and print.cpp in /hardware/cores/arduino/ to accept an extra parameter for the precision and pass it to printFloat.
As a hack I just modified these functions.
It means that wherever you have serial.print you also have to pass the precision.
When I get round to it ill do it properly. But I believe something like this will be added to ver14 of the ide.
in print.cpp
void Print::print(double n)
{
printFloat(n, 2);
}
To:
void Print::print(double n,uint8_t digits)
{
printFloat(n, digits);
}
and print.h
void print(double);
To:
void print(double,uint8_t);
Gordon