Aarch64 floating point operands support

[psi-func]

Implemented initial support for Aarch64 float-point operands in instructions
tested on FMOV, other instructions with this feature are not implemented in triton (float point nature)

[JonathanSalwan]

Thx!

What about moving encode_fp_imm into the Immediate class as a new constructor so that the other architectures can benefit of this feature?

Something like this:

Immediate::Immediate(double value, triton::uint32 size) {
    this->setValue(this->encode_fp_imm(value), size);
}

Then, at this line you can just call the new constructor as follow: Immediate imm{op->fp, size};.

[psi-func]

I agree, move feature to Immediate ctor + add endianness check