Can you help me understand the purpose and reasoning behind FASTLED_NO_PINMAP?

I'm walking through the steps to port FastLED. I'm normally able to divine(*) the purpose of the #define options. However, I'm stumped as to the purpose of FASTLED_NO_PINMAP.

The only platforms that #define FASTLED_NO_PINMAP are STM32 and NRF51.

There are only two conditional compilations, both in fastpin.h.

The first wraps a NEARLY identical definition of the Pin class. When FASTLED_NO_PINMAP is defined, the class has the following differences:

1. private _init() function sets mPinMask, mPort, and mInPort to NULL instead of using digitalPinToBitMask(), portOutputRegister(digitalPinToPort(mPin)), and portInputRegister(digitalPinToPort(mPin)) (respectively)
2. public functions setOutput() and setInput() are given empty function bodies
3. additional public function added: setPin(int pin) (sets member and calls _init())
4. removes public functions strobe() and toggle()
5. removes public functions void hi(register port_ptr_t port) and void lo(register port_ptr_t port)

Otherwise, the classes are identical. Based on the above, one guess I came up with is that, when porting to a new board, one would sub-class the above Pin class, and override the functions _init(), setOutput(), and setInput(). But, I can't find where the Pin class would actually be instantiated in FastLED. Thus, it seems unsatisfactory.

My other guess is that this is to enable support for non-Arduino environments, by providing the digitalPinToBitMask(), digitalPinToPort(), portOutputRegister(), and portInputRegister() capabilities that Arduino would normally provide. But, this also seems somewhat unsatisfactory.

Your help understanding this will be appreciated...

(*) the brute-force way, reviewing the code where they are checked. A slow way to learn the code base, but it's a start.