Add 'long double' template specializations

**Description**
It would be interesting to have `long double` templated functions' specializations. `long double` is most notably used with user-defined literals, and this forces to "downscale" the type to double.

As an example, having a custom templated matcher on floating point types, I can't directly set precision when used with `long double` parameters:

```cpp
Matcher() { Catch::StringMaker::precision = std::numeric_limits::digits10 + 1; }

/*
error: 'precision' is not a member of 'Catch::StringMaker'
Matcher() { Catch::StringMaker::precision = std::numeric_limits::digits10 + 1; }
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
*/
```