Mike Myers

I added this semantic in https://github.com/trailofbits/remill/commit/03fdb2d8c9363f36abd30440f430e860c279cdc3#diff-ad103422aab8c2d7d4f4aa2999e5f4f2 I also didn't keep notes to remember my rationale for this, but I think I based it on the semantic for `FNSTCW` https://github.com/trailofbits/remill/blob/8e0a28ebae9460bc71302ab9dc4b8fe4fbbf32a0/remill/Arch/X86/Semantics/X87.cpp#L1276 I...

Well that's good to hear, and I can revisit this. I think Peter added something to the test runner that ignores small floating point rounding errors though, could this be...

For not completely understood reasons, this bug went away on its own (perhaps our migration to a newer LLVM, perhaps our improvements to state modeling). Thanks for noticing @ranweiler !

Yea, so the reproducibility of the failing test case was temporarily hidden by issue #206 but it looks like this when the test case runs as it should: ``` gdb...

Ah! I will read the paper you linked from the year 2099 (?), and if you've already implemented something then I'm interested in that too. Briefly looking at the paper...

- [ ] `(X86 f115d018 5 (BYTES 0f ae 44 24 30) FXSAVE_MEMmfpxenv (WRITE_OP (DWORD_PTR (ADD (REG_32 SS_BASE) (REG_32 ESP) (SIGNED_IMM_32 0x30)))))` - [ ] FXRSTOR

Ok, moving to next release milestone since it seems like there just a couple more changes to address 👍🏻

Ref: https://github.com/actions/virtual-environments/blob/main/images/macos/macos-12-Readme.md

- [ ] The Packaging job already uses macOS 11 Big Sur. It could use macOS 12 Monterey. - [x] Deprecate Catalina-based build in CI - [x] Move the build...

[As of July 28th 2022](https://docs.github.com/en/actions/using-github-hosted-runners/about-github-hosted-runners#supported-runners-and-hardware-resources), all GitHub Actions hosted runners are still x86-64. We would still need a self-hosted runner for ARM Mac. [We've been _building_ osquery as Universal binaries](https://github.com/osquery/osquery/blob/master/.github/workflows/hosted_runners.yml#L797)...