[beta] 1.46 beta

[Mark-Simulacrum]

No description provided.

[Mark-Simulacrum]

@bors r+ rollup=never p=20

[bors]

📌 Commit 2af7940 has been approved by Mark-Simulacrum

[bors]

⌛ Testing commit 2af7940 with merge 07b17aaa944283e075e8e965d6954d26076c5823...

[Manishearth]

@bors treeclosed=5

will continue to manually shepherd today

[bors]

💔 Test failed - checks-actions

[Manishearth]

2020-07-14T14:05:10.4562495Z [RUSTC-TIMING] cargo_metadata test:false 8.008
2020-07-14T14:05:11.0796179Z �[0m�[1m�[38;5;9merror[E0635]�[0m�[0m�[1m: unknown feature `const_transmute`�[0m
2020-07-14T14:05:11.0796697Z �[0m  �[0m�[0m�[1m�[38;5;12m--> �[0m�[0m/cargo/registry/src/github.com-1285ae84e5963aae/rustc-ap-rustc_ast-664.0.0/lib.rs:13:12�[0m
2020-07-14T14:05:11.0797080Z �[0m   �[0m�[0m�[1m�[38;5;12m|�[0m
2020-07-14T14:05:11.0797433Z �[0m�[1m�[38;5;12m13�[0m�[0m �[0m�[0m�[1m�[38;5;12m| �[0m�[0m#![feature(const_transmute)]�[0m
2020-07-14T14:05:11.0798073Z �[0m   �[0m�[0m�[1m�[38;5;12m| �[0m�[0m           �[0m�[0m�[1m�[38;5;9m^^^^^^^^^^^^^^^�[0m
2020-07-14T14:05:11.0932281Z 

[Manishearth]

I think we're going to have to wait for the rust ap stuff

[Manishearth]

Once the ap stuff goes through would you mind updating this PR to include the clippy update? (currently in CI)

[Mark-Simulacrum]

Oh right, I forgot we never fixed toolstate this cycle. Indeed this will need to wait for that and we'll just re-promote at that point (basically rebasing atop master) I imagine.

[bors]

⌛ Testing commit 2af7940 with merge 8309d7fda7313002b8b50cbbcbd3efe0827bf789...

[Manishearth]

@bors r-

This had failed, no?

[Manishearth]

homu seems to have gotten confused

[Manishearth]

@bors yield

[Mark-Simulacrum]

#72920 landed in the same rollup as the fix for the toolstate failures on rustfmt/RLS so we didn't note that the fix didn't actually fix things (since that PR introduced a regression, technically, in beta-week). I don't know that tooling could've detected it though -- I guess we should just not rollup toolstate-fixing PRs in beta week?

Anyway, I pushed up a dummy commit that just re-adds that const_transmute feature on a dummy function and seems to fix the build for RLS and rustfmt locally -- hopefully that's also the case on CI.

@bors r+ rollup=never p=20

[bors]

📌 Commit 83b6d84 has been approved by Mark-Simulacrum

[bors]

💡 This pull request was already approved, no need to approve it again.

• There's another pull request that is currently being tested, blocking this pull request: More static symbols #74175

[bors]

📌 Commit 83b6d84 has been approved by Mark-Simulacrum

[bors]

⌛ Testing commit 83b6d84 with merge c20a0a4ad0888cb28993511e6f1e87ce622bd87c...

[Manishearth]

     Running `build/x86_64-unknown-linux-gnu/stage2-tools/x86_64-unknown-linux-gnu/release/cargo-miri miri setup`
ERROR: the sysroot can't be built for the Beta channel. Switch to nightly.
fatal error: Failed to run xargo


command did not execute successfully: "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "run" "--target" "x86_64-unknown-linux-gnu" "-Zbinary-dep-depinfo" "-j" "16" "--release" "--locked" "--color" "always" "--manifest-path" "/checkout/src/tools/miri/cargo-miri/Cargo.toml" "--features" "rustc-workspace-hack/all-static" "--" "miri" "setup"
expected success, got: exit code: 1


[TIMING] Miri { stage: 2, host: "x86_64-unknown-linux-gnu" } -- 33.283

2 command(s) did not execute successfully:

  - "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "test" "--target" "x86_64-unknown-linux-gnu" "-Zbinary-dep-depinfo" "-j" "16" "--release" "--locked" "--color" "always" "--manifest-path" "/checkout/src/tools/rustfmt/Cargo.toml" "--features" "rustc-workspace-hack/all-static"

  - "/checkout/obj/build/x86_64-unknown-linux-gnu/stage0/bin/cargo" "run" "--target" "x86_64-unknown-linux-gnu" "-Zbinary-dep-depinfo" "-j" "16" "--release" "--locked" "--color" "always" "--manifest-path" "/checkout/src/tools/miri/cargo-miri/Cargo.toml" "--features" "rustc-workspace-hack/all-static" "--" "miri" "setup"

failed to run: /checkout/obj/build/bootstrap/debug/bootstrap test --no-fail-fast src/doc/book src/doc/nomicon src/doc/reference src/doc/rust-by-example src/doc/embedded-book src/doc/edition-guide src/tools/rls src/tools/rustfmt src/tools/miri
Build completed unsuccessfully in 0:33:37

[Mark-Simulacrum]

@bors r+ rollup=never p=20

Tools builder is passing on the PR at least.

[bors]

📌 Commit dba515e has been approved by Mark-Simulacrum

[mati865]

Miri should not be built on beta and stable.

[Mark-Simulacrum]

Miri failures are ignored -- we do currently try to build it, though. I'd be happy to take a patch that doesn't do so.

[mati865]

That's what #73232 was supposed to fix 🤔

[Mark-Simulacrum]

That's what #73232 was supposed to fix thinking

You're not wrong, but indeed the tools builder explicitly lists miri (

rust/src/ci/docker/host-x86_64/x86_64-gnu-tools/checktools.sh

Line 19 in c724b67

src/tools/miri \

) so the defaults don't actually matter still -- only for x.py build for example will they affect things.

My guess is we can add some logic around (

rust/src/bootstrap/test.rs

Line 358 in c724b67

let stage = self.stage;

that'll just refuse to test miri on stable/beta.

[bors]

⌛ Testing commit dba515e with merge 139f54761d07ff14bbb3794bb687fa6de4227a19...

[mati865]

I'll cc @RalfJung in case they want to continue work on #74323 (comment)

[RalfJung]

I am not very familiar with that glue code. Wouldn't it make more sense to adjust

rust/src/ci/docker/host-x86_64/x86_64-gnu-tools/checktools.sh

Line 19 in c724b67

src/tools/miri \

to not build nightly-only tools on non-nightly branches?

[Mark-Simulacrum]

That would work too, I'm just less familiar with bash so would need to probably put in more effort (and we'd have a duplicate list of stable/unstable tools).

[RalfJung]

OTOH it seems odd that an explicit request to build & test Miri would be ignored, even if the channel is stable.

Ideally the Python script that has the stable/unstable list would also drive x.py here, but I don't know how to arrange for that.

[Mark-Simulacrum]

@bors retry - this failed forever ago https://github.com/rust-lang-ci/rust/runs/874170463

[bors]

⌛ Testing commit dba515e with merge e51b714...

[bors]

☀️ Test successful - checks-actions, checks-azure
Approved by: Mark-Simulacrum
Pushing e51b714 to beta...

[RalfJung]

How did this land now if Miri is failing?

[mati865]

Miri fail is non fatal. It doesn't stop the build.

[RalfJung]

Oh so the problem that came up here is that the build happened at all, not that the failure was a problem?

[mati865]

I haven't checked the logs but something else must have failed.

[RalfJung]

Okay I thought there was an actual problem here not just "we are running more tests than we should". Oh well.

Here's my partial attempt at fixing this: #74389. But I give up here, rustbuild is too hard to work on I am afraid and my Miri TODO list is already overflowing with other stuff. ;) Hopefully this is a start that someone else can build on, if anyone cares.