fix and (re-)enable Miri cross-target checks on macOS and Windows

[RalfJung]

Fixes #103519
r? @Mark-Simulacrum

[Noratrieb]

Wouldn't it make more sense to relax the sanity checks depending on the target instead of removing them entirely?

[RalfJung]

These sanity checks make no sense for test src/tools/miri --target $TARGET, no matter the target. The sanity checks ensure that we can actually build code for that target with a C compiler, but when testing a target in Miri we don't build any code for the target, we just interpret it.

[Mark-Simulacrum]

#103731 disables the cross-test to Windows -- I suspect this PR may actually fix the problem in any case for both macOS and Windows, and it's just coincidence that Windows appeared to work before.

[Mark-Simulacrum]

There's other checks similar to this (e.g., the cmake check for msvc below). I think my recommendation would be to see if we can refactor the sanity checking as a whole so there's a x.py flag that disables it. I think skipping it entirely won't work today, but if we move some of the detection logic to config.rs that might fix it.

[RalfJung]

That will probably require more digging into x.py than I will have time for, unfortunately.

[Mark-Simulacrum]

Alright. I can probably make an attempt to do that.

[RalfJung]

I added another check of BOOTSTRAP_SKIP_TARGET_SANITY to the second loop that iterates all targets, which could help in the interim. (When writing this PR I did not notice there are two such loops.)

[bors]

☔ The latest upstream changes (presumably #103731) made this pull request unmergeable. Please resolve the merge conflicts.

[RalfJung]

@Mark-Simulacrum now that the release has gone through, do you think it'd make sense to merge this PR until someone has time to do the sanity check refactor you mentioned?

[Mark-Simulacrum]

r=me with a short comment on the env variable in bootstrap noting the desire for a refactor here (feel free to just point at this PR in terms of context/explanation)

[RalfJung]

@bors r=Mark-Simulacrum rollup=iffy

[bors]

📌 Commit 397e5bb has been approved by Mark-Simulacrum

It is now in the queue for this repository.

[bors]

⌛ Testing commit 397e5bb with merge d69c33a...

[bors]

☀️ Test successful - checks-actions
Approved by: Mark-Simulacrum
Pushing d69c33a to master...

[rust-timer]

Finished benchmarking commit (d69c33a): comparison URL.

Overall result: ✅ improvements - no action needed

@rustbot label: -perf-regression

Instruction count

This is a highly reliable metric that was used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-0.6%	[-0.7%, -0.6%]	2
Improvements ✅ (secondary)	-1.4%	[-1.6%, -1.3%]	6
All ❌✅ (primary)	-0.6%	[-0.7%, -0.6%]	2

Max RSS (memory usage)

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	0.9%	[0.5%, 1.3%]	2
Regressions ❌ (secondary)	6.6%	[6.6%, 6.6%]	1
Improvements ✅ (primary)	-0.6%	[-0.7%, -0.6%]	2
Improvements ✅ (secondary)	-3.6%	[-3.6%, -3.6%]	1
All ❌✅ (primary)	0.1%	[-0.7%, 1.3%]	4

Cycles

Results

This is a less reliable metric that may be of interest but was not used to determine the overall result at the top of this comment.

	mean	range	count
Regressions ❌ (primary)	-	-	0
Regressions ❌ (secondary)	3.5%	[3.5%, 3.5%]	1
Improvements ✅ (primary)	-0.5%	[-0.5%, -0.5%]	2
Improvements ✅ (secondary)	-	-	0
All ❌✅ (primary)	-0.5%	[-0.5%, -0.5%]	2

[Mark-Simulacrum]

Delta here is pretty certainly noise, though it's also an improvement so either way fine.