Created by: sarfata
This helps when the compiler option -fsingle-precision-constant is
set (teensy 3.x).
There is no suffix to make a constant a double. The L suffix on float constants makes them long double but for most platforms that should be the same as a double so this should not impact any other platform.
See https://stackoverflow.com/questions/12205141/is-there-a-floating-point-literal-suffix-in-c-to-make-a-number-double-precisio for some discussion on this issue.
Another option would be to set the ARDUINOJSON_DOUBLE_IS_64BITS to 0 on Teensy but I think this approach is favorable.