On Mon, Feb 17 2020, Christopher Zimmermann <[hidden email]> wrote:
> the diff including the PLIST changes look correct to me. Nevertheless
> I would recommend testing on at least one bytecode-only architecture.
Sadly it doesn't build any more on sparc64. Error #1 is due to the lack
of -fexcess-precision=standard in base-gcc.
When patching this away or using COMPILER=ports-gcc, I hit error #2:
Error: Error while linking kernel/uint63.cmo:
The external function `coq_uint63_to_float_byte' is not available