Fix unsoundness for VecDeque

[MaloJaffre]

See individual commit for more details.

r? @RalfJung.

Fixes#53566, fixes#53529

[RalfJung]

I fully agree with the Debug part.

For the revert, maybe we want to keep the tests and benches?

[varkor]

Shouldn't there be a test to make sure the unsoundness couldn't be accidentally reintroduced in the future?

[RalfJung]

We can't really test for UB...

... except miri can. I guess someone could add a test to the miri test suite. Ideally doing more than just a debug print; essentially running much of the rustc test suite covering VecDeque in miri. (At some point we want to run the rustc test suite proper on miri but that'll take a while.)

[MaloJaffre]

@RalfJung I've now restored the tests and the benches.
Apart from a test, I think this is ready.

[RalfJung]

Apart from a test, I think this is ready.

What kind of test do you plan to add?

[MaloJaffre]

I've just added a compile-pass test for #53529 by @Pazzaz.

[RalfJung]

Hm, I think this would make more sense as a unit test in liballoc/tests/vec_deque.rs. Could you make it one?

[MaloJaffre]

@RalfJung Done!

[RalfJung]

Awesome! r=me once traivs is happy.

@bors delegate+

[bors]

✌️ @MaloJaffre can now approve this pull request

[rust-highfive]

The job x86_64-gnu-llvm-5.0 of your PR failed on Travis (raw log). Through arcane magic we have determined that the following fragments from the build log may contain information about the problem.

Click to expand the log.

[00:48:03] ....................................................................................................
[00:48:06] ....................................................................................................
[00:48:09] ..........i.........................................................................................
[00:48:12] ....................................................................................................
[00:48:14] ...........................................................iiiiiiiii................................
[00:48:20] ....................................................................................................
[00:48:24] ....................................................................................................
[00:48:26] ........................................i...........................................................
[00:48:29] ..........................................................................................i.i..ii...
---
[01:01:28] Compiling alloc v0.0.0 (file:///checkout/src/liballoc)
[01:01:29] error[E0433]: failed to resolve. Use of undeclared type or module `Box`
[01:01:29] --> liballoc/collections/vec_deque.rs:2973:24
[01:01:29] |
[01:01:29] 2973 | dst.push_front(Box::new(1));
[01:01:29] | ^^^ Use of undeclared type or module `Box`
[01:01:29] error[E0433]: failed to resolve. Use of undeclared type or module `Box`
[01:01:29] --> liballoc/collections/vec_deque.rs:2974:24
[01:01:29] |
[01:01:29] |
[01:01:29] 2974 | dst.push_front(Box::new(2));
[01:01:29] | ^^^ Use of undeclared type or module `Box`
[01:01:29] error[E0433]: failed to resolve. Use of undeclared type or module `Box`
[01:01:29] --> liballoc/collections/vec_deque.rs:2978:24
[01:01:29] |
[01:01:29] |
[01:01:29] 2978 | src.push_front(Box::new(0));
[01:01:29] | ^^^ Use of undeclared type or module `Box`
[01:01:32] error: aborting due to 3 previous errors
[01:01:32] [01:01:32] For more information about this error, try `rustc --explain E0433`.
[01:01:32] error: Could not compile `alloc`.
[01:01:32] error: Could not compile `alloc`.
[01:01:32] warning: build failed, waiting for other jobs to finish...
[01:02:06] error: build failed
[01:02:06] [01:02:06] [01:02:06] command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "test" "--target" "x86_64-unknown-linux-gnu" "-j" "4" "--release" "--locked" "--color" "always" "--features" "panic-unwind jemalloc backtrace" "--manifest-path" "/checkout/src/libstd/Cargo.toml" "-p" "alloc" "--" "--quiet"
[01:02:06] [01:02:06] 4776388 .
2940264 ./obj

I'm a bot! I can only do what humans tell me to, so if this was not helpful or you have suggestions for improvements, please ping or otherwise contact @TimNN. (Feature Requests)

[RalfJung]

Oh while you are fixing this anyway, maybe you could add some assertions checking that we are getting back the values we expect (in pop_back and in the iteration loop)?

[MaloJaffre]

Sure!

[MaloJaffre]

@bors r=RalfJung

[bors]

📌 Commit f8d5ed4 has been approved by RalfJung

[bors]

⌛ Testing commit f8d5ed4 with merge 54d82d0...

[bors]

☀️ Test successful - status-appveyor, status-travis
Approved by: RalfJung
Pushing 54d82d0 to master...

[Shnatsel]

Regarding a regression test for this: Memory Sanitizer would also have discovered this issue, and it introduces tolerable (~3x) slowdown, which is much less extreme than MIRI.