Handle inactive enum variants in MaybeUninitializedPlaces

[ecstatic-morse]

Resolves the first part of #69715.

This is the equivalent of #68528 but for MaybeUninitializedPlaces. Because we now notify drop elaboration that inactive enum variants might be uninitialized, some drops get marked as "open" that were previously "static". Unlike in #69715, this isn't strictly better: An "open" drop expands to more MIR than a simple call to the drop shim. However, because drop elaboration considers each field of an "open" drop separately, it can sometimes eliminate unnecessary drops of moved-from or unit-like enum variants. This is the case for Option::unwrap, which is reflected in the mir-opt test.

cc @eddyb
r? @oli-obk

[ecstatic-morse]

@bors try
@rust-timer queue

[rust-timer]

Awaiting bors try build completion

[bors]

⌛ Trying commit 7f8b034f63e2914164a1a2ef52b077272363f3fb with merge 7eb15f79d388f9a78cd1fb2e8eeb7267cf56f4a4...

[Aaron1011]

@ecstatic-morse: How does interact with generators?

[bors]

☀️ Try build successful - checks-azure
Build commit: 7eb15f79d388f9a78cd1fb2e8eeb7267cf56f4a4 (7eb15f79d388f9a78cd1fb2e8eeb7267cf56f4a4)

[rust-timer]

Queued 7eb15f79d388f9a78cd1fb2e8eeb7267cf56f4a4 with parent 0ca7f74, future comparison URL.

[rust-timer]

Finished benchmarking try commit (7eb15f79d388f9a78cd1fb2e8eeb7267cf56f4a4): comparison url.

[jonas-schievink]

Neat!

[ecstatic-morse]

How does interact with generators?

Hey, you're not a ghost! 👻

The generator transform doesn't use MaybeUninitializedLocals to compute generator saved locals, and it does drop elaboration with DropShimElaborator , which also doesn't use MaybeUninitializedLocals. Only Elaborator, which seems to be the "general purpose" drop elaborator uses it. Furthermore, apply_discriminant_switch_effect doesn't get called for TyKind::Generators, so I don't believe that this affects generators any more than regular functions. I don't know what the benefits may be from changing that last part, but it might be worth investigating. I'm pretty much terrified of touching the generator transform, though.

[ecstatic-morse]

This is ready for review now.

r? @oli-obk
cc @eddyb

[oli-obk]

Oh god diffs of diffs XD. I love how this basically says we now have an additional variable, but it's optimized out here.

[oli-obk]

@bors r+

[bors]

📌 Commit eb4d28b has been approved by oli-obk

[oli-obk]

You're really raising the bar on what a high quality PR means... Readable PR diff, modular design, documentation and it all causes a perf improvement along the way.

[ecstatic-morse]

@bors rollup=never (perf)

[Manishearth]

@bors p=1

[bors]

⌛ Testing commit eb4d28b with merge 04c24cf3654f0427ebb9bdb5a8aaf879d169aeb7...

[Manishearth]

@bors yield retry

[Manishearth]

@bors retry yield

[bors]

⌛ Testing commit eb4d28b with merge 0202cee4aabe48154df4a45a442656e0598f3d9b...

[Manishearth]

@bors retry yield

[bors]

⌛ Testing commit eb4d28b with merge b5c3197d9d1e60bdd5ca98e8a8533a5742164739...

[Manishearth]

@bors retry yield

[bors]

⌛ Testing commit eb4d28b with merge aafc3c614dfaf82515c3c9c8f5eccd812823603c...

[Manishearth]

@bors retry yield

[bors]

⌛ Testing commit eb4d28b with merge 4a0a37a3373d93f4332f326546224e43d020d36c...

[bors]

💔 Test failed - checks-actions

[JohnTitor]

Another case of #73992, @bors retry

[bors]

⌛ Testing commit eb4d28b with merge 2753fab...

[bors]

☀️ Test successful - checks-actions, checks-azure
Approved by: oli-obk
Pushing 2753fab to master...