Unfortunately, portlint didn't complain about this.

Address comments that portlint didn't complain about.

Addressed comments

Address review comments.

use IGNORE_FreeBSD_10 and extend pkg-descr

Comments addressed

Done

Add check for older than 12.0

LGTM, thanks!

Order variables

Pet portlint.

Version using weak_reference instead of imprecise. Does not build world on power* platforms.

Results for using weak_reference as opposed to imprecise.c

Sorted cpowl into the list and re-instantiated the tgammal comment.

Thanks for the comments!

I've learned a lot already.

universe on riscv reveals that we generally didn't have a powl but instead re-used regular pow. Only gcc on risvc complains about this though. Hence, import powl for ld80 and ld128 from openbsd (and some helper routines) and restrict the workaround to mantissa 53 architectures (power). with this, universe passes on amd64, i386, riscv (except sf but that seems to be expected), arm64, and powerpc.

Add patch to complex.h that was missing.

Correct file locations and alphabetic ordering.

Change to ONLY_FOR_ARCHS_REASON from the amd64 specific one.

Address comments. Unmute INSTALL_DATA, portlint check

Address @mat 's comments and enable i386 compilation.

On a AMD Threadripper 1900X I see "microcode updating..." and output of the CPU information. On a Carrizo APU, I see "microcode updating..." and "done", nothing else. How to check it DTRT?

Thanks for letting me know it was marked broken already! Updated diff accordingly!

necessary lkpi changes were MFC'd to STABLE.

After discussion with @swills , use the Linux release ID as the PORTVERSION (switched between 11-STABLE and CURRENT) to avoid confusion "port is g20180224". This solution is still ugly but it allows users of the port to seamlessly upgrade, without needing to know which DRM their release supports (and get a new DRM automagically if the requirements are MFC'd).