Avoid alignment mismatch between ABI and layout for unions.

[luqmana]

Fixes #104802
Fixes #103634

r? @eddyb cc @RalfJung

[eddyb]

I would prefer if this was fixed by #103926 - specifically, this comment: #103926 (comment).

If you can arrange with @oli-obk, you could take over in this PR - I don't care as long as both alignment sources are fixed at once without special-casing them.

Also, we need a test for #[repr(packed)] union Foo { zst: [u16; 0], byte: u8 }, which should, by virtue of #[repr(packed)] reducing alignment back, be able to be Abi::Scalar again!

[luqmana]

I would prefer if this was fixed by #103926 - specifically, this comment: #103926 (comment).

If you can arrange with @oli-obk, you could take over in this PR - I don't care as long as both alignment sources are fixed at once without special-casing them.

Ah, admittedly I did see that issue but didn't dig too deep into it. I'm ok with taking a stab at it or leaving it to @oli-obk

[oli-obk]

I'm ok with taking a stab at it or leaving it to @oli-obk

I won't get to it this week or next week, so feel free to take over!

[bors]

☔ The latest upstream changes (presumably #105554) made this pull request unmergeable. Please resolve the merge conflicts.

[apiraino]

Switching to waiting on author to incorporate changes. Feel free to request a review with @rustbot ready, thanks!

@rustbot author

[luqmana]

@rustbot ready

[wesleywiser]

Visited during the compiler team triage meeting. It looks like @eddyb's feedback has been taken into account and resolved. Reassigning to @oli-obk for review. Feel free to reroll if you would prefer!

r? @oli-obk

[oli-obk]

just a nit, then this lgtm

[oli-obk]

This relies on the !is_err check above. Maybe replace that check with a let Ok(common_non_zst_abi_and_align) = &mut common_non_zst_abi_and_align to make sure these two things stay in sync

Suggested change

	if let Ok(Some((common_abi, common_align))) = &mut common_non_zst_abi_and_align {
	if let Some((common_abi, common_align)) = &mut common_non_zst_abi_and_align {

[luqmana]

Sure, reordered it a bit to remove the is_err check.

[bors]

☔ The latest upstream changes (presumably #109626) made this pull request unmergeable. Please resolve the merge conflicts.

[pnkfelix]

@rustbot label: +S-waiting-on-author -S-waiting-on-review

just needs a rebase according to @oli-obk

[luqmana]

@rustbot ready

[oli-obk]

@bors r+

[bors]

📌 Commit 8e7714d has been approved by oli-obk

It is now in the queue for this repository.

[bors]

⌛ Testing commit 8e7714d with merge 151a070...

[bors]

☀️ Test successful - checks-actions
Approved by: oli-obk
Pushing 151a070 to master...

[rust-timer]

Finished benchmarking commit (151a070): comparison URL.

Overall result: no relevant changes - no action needed

@rustbot label: -perf-regression

Instruction count

This benchmark run did not return any relevant results for this metric.

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
Regressions ❌ (secondary)	-	-	0
Improvements ✅ (primary)	-	-	0
Improvements ✅ (secondary)	-2.5%	[-2.7%, -2.2%]	2
All ❌✅ (primary)	-	-	0

Cycles

This benchmark run did not return any relevant results for this metric.

Binary size

This benchmark run did not return any relevant results for this metric.

Bootstrap: 656.275s -> 653.613s (-0.41%)