Windows ARM build

[qurious-pixel]

Windows ARM build

[schm1dtmac]

Quick note: I’ve removed googletest in PR #17735 from arm64 CI too to bypass arm64 build lockups but tbh I think in both of our PRs it shouldn't be removed, it’s a workaround for what really is a unit test failure for changes that actually appear to lead to crashes in game on builds that bypass such tests from my tests (see #17746). (That and it’s not uber relevant to Windows ARM CI I guess)

[qurious-pixel]

Quick note: I’ve removed googletest in PR #17735 from arm64 CI too to bypass arm64 build lockups but tbh I think in both of our PRs it shouldn't be removed, it’s a workaround for what really is a unit test failure for changes that actually appear to lead to crashes in game on builds that bypass such tests from my tests (see #17746). (That and it’s not uber relevant to Windows ARM CI I guess)

Thanks @schm1dtmac. I am currently bisecting the googletest issue. test_pair test_simple_array is causing the lockup. I will revert the change, if I can resolve the issue.

[hcorion]

Mac tests should be addressed as part of a different PR, but the Windows changes look good

[schm1dtmac]

Mac tests should be addressed as part of a different PR, but the Windows changes look good

Yeah the Mac stuff’s long since been addressed properly and all unit tests succeed now (#17800), so any of that stuff here just needs complete removal.

[kd-11]

I'll be able to test this in a few days when I go on break from work.

[Megamouse]

Why this?

[qurious-pixel]

Fixes the following error:

void vm::init() (error=0x8)

vm.cpp error

[Megamouse]

Why remove the default?
Why exclude all clang?

[qurious-pixel]

This was removed entirely, then hacked back in. not needed anymore after #17899, but should have been a separate PR anyways.

FreeBSD clang behaves differently; it was the only runner to fail after the changes from #17736.

Now all jobs build with L396-398 removed as was inteneded in that PR.