Testing __float128 [Cygwin]

[FX]

> Can't you add a configure check and use sqrt/cbrt only if the long double variants are not available?

This is what I plan to do, yes.

> If one looks at Andy's implementation of sqrt, one sees that he only does basic arithmetic without using any library call - thus, cbrtl and sqrtl are never used and thus you do not have those link problems.

I have opted for the simplest (and less error-prone) solution. Once everything is working there will be time for improvements (though I probably won't do it myself, as I am no expert on this).


> I do not know how this compares with gfortran's libquad, but I had the impression that it supports all intrinsics. (So far, I have not yet tried gfortran's quad support.)

I don't know what your "it" is, but gfortran+libquad do support all intrinsics (though some array intrinsics do not work quite yet, as discussed previously)

FX
>