Kirk Klobe

> Yeah me and @kcgen knew this would come bite our asses hence we've postponed upgrading from macOS 11. Already planned to moving to MacPorts to avoid this but lets...

Thanks for testing @Burrito78 - this build (https://github.com/dosbox-staging/dosbox-staging/actions/runs/7566094859) works, with duct tape and baling wire, but should hopefully let us get the release out the door.

Update: M1 runners now available for open source projects! https://github.blog/changelog/2024-01-30-github-actions-introducing-the-new-m1-macos-runner-available-to-open-source/

I am also seeing the "rejected". Although, according to [Apple](https://forums.developer.apple.com/forums/thread/696235), "First up, spctl is not a great techniques for evaluating whether something will pass Gatekeeper.", but if that's what we...

OK, got it figured out - the `sed` command in the packaging script is creating some backup files that don't belong in the bundle, most notably `Info.plist-e`: ![CleanShot 2024-02-14 at...

ok @Burrito78 , @johnnovak posted the new DMG and hash on the website, maybe give it a re-submit?

> > implementing the 80-bit operations accurately in software is going to be a lot of work. > > Good news is there a lots of multi and extended precision...

> But I don't think that's true. Single precision is useless, you could use integer maths for that and it would be much faster, as I explained. If something _needs_...

The MPFR basic arithmetic operations are much faster than the math functions. Quick benchmark: https://gist.github.com/kklobe/2df6ac7d485dce90a51c430c6fbdf978 ```log Native (double) arithmetic: 5768 microseconds Native (double) math functions: 4142 microseconds MPFR (80-bit) arithmetic:...

OK, did a quick test replacing `FPU_FADD`,`FPU_FSUB`,`FPU_FMUL`, and `FPU_FDIV` with MPFR 80-bit, and it dropped my QTD scores right about in half (~64 vs. ~120). Definitely still playable.