Intermittent failures in macOS CI jobs

[adpaco-aws]

Proposed change: Investigate the spurious failures in macOS CI jobs and attempt to fix if possible.

Motivation: This week some CI jobs running on macOS runners have been failing without a clear reason to do so. Some examples:

• This CI run for Automatic cargo update to 2024-01-08 #2968
• This CI run for Rethink should_panic and fail_uncoverable options as global conditions #2967 (no code changes)
• This CI run in main after we merged Rethink should_panic and fail_uncoverable options as global conditions #2967 (again, no code changes)

In both cases, the processes are compiling a crate when they receive an abort signal. It's possible that we're hitting some resource limits in the runners themselves. But re-running the CI jobs a few times does the trick.

[tautschnig]

Adding another such failure: https://github.com/model-checking/kani/actions/runs/7568332070/job/20609324630?pr=2977. This looks a lot like what they saw in Miri: rust-lang/miri#3255. Perhaps we should try to adapt rust-lang/miri@0b79240 for Kani.

[tautschnig]

Resolving with the assumption that #2987 has fixed the problem, which is in line with the claims made in rust-lang/rust#92173 (comment) (that the problem is an XCode issue). We may have to re-open this issue if the problem isn't actually fixed by the OS upgrade.