Fixes float-conversion warning in ostream::putDouble.

[martinbudden]

Fixes warning that occurs when compiled with -Wfloat-conversion.

[greiman]

I will handle this fix like the fix for istream. Please try the following at line 92 of ostream,cpp:

  n += roundDbl;
  uint32_t intPart = static_cast<uint32_t>(n);  // << here
  double fractionPart = n - intPart;

I will apply it to SdFat-beta and later to the release version of SdFat.

[greiman]

Few people use the iostream part of SdFat and printf is not in many board packages.

I don't use iostream much so I am playing with a variadic wrapper for anything derived from Print for my use.

Here is an early test:

VaPrint vp(Serial);

void test() {
  bool b = true;
  uint32_t x = 123456;
  vp.vprintln("x: ", x, Base(16, 'x'), ", " ,x ,Base(), ", ",  x);
  vp.vprintln('|', Width(20), 1234, '|');
  vp.vprintln('|', Width(-20), 1234, '|');
  vp.vprintln("PI: ", PI,  ", ", Prec(7), PI);
  vp.vprintln("bool b: ",  b);
} 

Width() is like iostreams std::setw(), it only applies to the next field. Base() is like iosteram std::hex and std::dec, and Prec() is like std::setprecision().

Here is the output:

x: 123456, 0x1E240, 123456
|                1234|
|1234                |
PI: 3.14, 3.1415927
bool b: true