You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
miri no longer builds after rust-lang/rust#101333 #101344
Hello, this is your friendly neighborhood mergebot.
After merging PR #101333, I observed that the tool miri has failing tests.
A follow-up PR to the repository https://github.com/rust-lang/miri is needed to fix the fallout.
cc @matthiaskrgr, do you think you would have time to do the follow-up work?
If so, that would be great!
thread 'rustc' panicked at 'range start index 1 out of range for slice of length 0', library\core\src\slice\index.rs:52:5
stack backtrace:
note: Some details are omitted, run with `RUST_BACKTRACE=full` for a verbose backtrace.
error: internal compiler error: unexpected panic
note: the compiler unexpectedly panicked. this is a bug.
note: we would appreciate a bug report: https://github.com/rust-lang/rust/issues/new?labels=C-bug%2C+I-ICE%2C+T-compiler&template=ice.md
note: rustc 1.65.0-nightly (9ba169a73 2022-09-02) running on x86_64-pc-windows-msvc
note: compiler flags: --crate-type dylib --crate-type rlib -C opt-level=3 -C embed-bitcode=no -C debuginfo=0 -C debug-assertions=on -C debug-assertions=off -C overflow-checks=on -Z force-unstable-if-unmarked -Z tls-model=initial-exec
note: some of the compiler flags provided by cargo are hidden
query stack during panic:
#0 [mir_drops_elaborated_and_const_checked] elaborating drops for `sys::windows::fs::<impl at D:\a\rust\rust\library\std\src\sys\windows\fs.rs:278:1: 278:10>::file_attr`
#1 [optimized_mir] optimizing MIR for `sys::windows::fs::<impl at D:\a\rust\rust\library\std\src\sys\windows\fs.rs:278:1: 278:10>::file_attr`
end of query stack
[RUSTC-TIMING] std test:false 5.705
error: could not compile `std`
The same target works fine on my system, so it is host-specific. Sounds like one of the PRs in that rollup is buggy...
We're seeing a related-looking issue on Fuchsia on the same commit:
Not sure how related they really are, the original ICE is some kind of out-of-bounds slice access within rustc itself...
Could someone try upgrading Miri to current git master? Miri CI passes with latest rustc, after all... even if it fails again, we'll at least get better line numbers since #100696 was fixed. Maybe there is a way to run the windows tools runner without using up a full bors slot?
Activity
RalfJung commentedon Sep 2, 2022
Ah damn it, we keep getting broken so there's still no nightly with Miri. :(
RalfJung commentedon Sep 2, 2022
Building std now ICEs on windows...?
The same target works fine on my system, so it is host-specific. Sounds like one of the PRs in that rollup is buggy...
compiler-errors commentedon Sep 2, 2022
@RalfJung do you have a stack trace for that ICE?
RalfJung commentedon Sep 2, 2022
Nope, I cannot reproduce so all we got is the log from CI. The ICE is here.
RalfJung commentedon Sep 2, 2022
Miri CI cannot reproduce either, not even in a Windows host. It's a 32bit host though.
Would be great if someone with an x86_64 windows host could try to clone the Miri repo and do
./rustup-toolchain HEAD && ./miri test.tmandry commentedon Sep 2, 2022
We're seeing a related-looking issue on Fuchsia on the same commit:
https://logs.chromium.org/logs/fuchsia/buildbucket/cr-buildbucket/8804097715115352817/+/u/build/failure_summary
saethlin commentedon Sep 3, 2022
I'm unable to reproduce the ICE with
./miri testonBtw the advice of
./rustup-toolchain HEADresults in a build error, because Miri was broken again by another PR.Noratrieb commentedon Sep 3, 2022
This does look like my PR causes this.
RalfJung commentedon Sep 3, 2022
Yeah when I wrote that, the other PR had not landed yet. ;)
RalfJung commentedon Sep 3, 2022
Hm, that is odd. Not sure what the tools builder does differently, maybe it has debug assertions enabled?
The line number in the ICE is so bad due to #100696.
RalfJung commentedon Sep 6, 2022
Not sure how related they really are, the original ICE is some kind of out-of-bounds slice access within rustc itself...
Could someone try upgrading Miri to current git master? Miri CI passes with latest rustc, after all... even if it fails again, we'll at least get better line numbers since #100696 was fixed. Maybe there is a way to run the windows tools runner without using up a full bors slot?
oli-obk commentedon Sep 7, 2022
on it
RalfJung commentedon Sep 7, 2022
Turns out the strange Windows-only ICE was caused by a buggy debug assertion:
https://github.com/rust-lang/rust/blob/58286b8f310c4884e250ed622e8678c5f2dff666/compiler/rustc_middle/src/mir/mod.rs#L1481
Auto merge of rust-lang#101437 - compiler-errors:erase-normalize-orde…
Auto merge of #101437 - compiler-errors:erase-normalize-ordering, r=t…
Auto merge of #101437 - compiler-errors:erase-normalize-ordering, r=t…
Auto merge of #101437 - compiler-errors:erase-normalize-ordering, r=t…