Disable or enable quadruple precision support by default?

[awvwgk]

trafficstars

Currently test-drive will automatically ~~enable~~ disable support for quadruple precision if the preprocessor is not defined:

https://github.com/fortran-lang/test-drive/blob/00bb499f0c46bc376c3aacbbe93f6b38abbe824b/src/testdrive.F90#L14-L17

In case of the meson and CMake build this never happens, since we can easily detect this feature in the configure stage and set the preprocessor defines for the compiler. For fpm we will always fall back to the default unless the user explicitly ~~disables~~ enables the support for quadruple precision by providing --flag "-DWITH_QP=1".

The default has been changed from enabled by default to disabled by default in #8.

This issue will be relevant for compilers not supporting quadruple precision:

• GFortran on MacOS/Arm64
• NVFortran