Add support to disable quadmath on aarch64 platforms? #33
Description
From what I understand about quadmath, the main benefit it gives is the definition of __float128 types. However, aarch64 already has a native __float128 type (c.f. conda-forge/ctng-compilers-feedstock#160 (comment)) and so quadmath is not supported on aarch64 Linux platforms.
Would it be possible to add in a build time guard that disables quadmath if the platform is detected to be aarch64 and use native platform types instead? Or is there a technical blocker to this given how the rest of quadmath works?