charlesnicholson
Promote %f to long double option
Currently, nanoprintf casts long double or %Lf arguments to double. It would be good to have an option so float and double could be promoted to long double instead. Especially where long double is 64 bits or less on a given platform.
This may be tricky to test. Since promoting to long double only works if (LDBL_MANT_DIG <= 53) && (LDBL_MAX_EXP <= 1024), which might require additional #ifdef's and conditional compilation.
It could go under the NANOPRINTF_USE_* defines as NANOPRINT_USE_LONG_DOUBLE_PROMOTION like I've done here. Although it might be better to have it implemented similarly to NANOPRINTF_CONVERSION_FLOAT_TYPE as NANOPRINTF_PROMOTE_TO_LONG_DOUBLE 0/1 like I've done here.