Rust 1.19 regression, euclid 0.10.5, SIGSEGV

[brson]

https://github.com/servo/euclid

commit 8947756ff1436df5cf2cb291e6502fd902ad5fce
Merge: 7097fcb fda90dc
Author: bors-servo <lbergstrom+bors@mozilla.com>
Date:   Fri Jan 27 02:54:00 2017 -0800

    Auto merge of #174 - nox:bump, r=SimonSapin

    Bump version to 0.10.5

    Some .swp files were included in the package on Cargo.

    <!-- Reviewable:start -->
    ---
    This change is [<img src="https://reviewable.io/review_button.svg" height="34" align="absmiddle" alt="Reviewable"/>](https://reviewable.io/reviews/servo/euclid/174)
    <!-- Reviewable:end -->

brian@ip-10-145-43-250:/mnt2/dev⟫ rustc +nightly -Vv
rustc 1.19.0-nightly (0418fa9d3 2017-06-04)
binary: rustc
commit-hash: 0418fa9d382a47d782cc1e195c14573be9c32095
commit-date: 2017-06-04
host: x86_64-unknown-linux-gnu
release: 1.19.0-nightly
LLVM version: 4.0

141 brian@ip-10-145-43-250:~/dev/euclid⟫ cargo +nightly test
    Finished dev [unoptimized + debuginfo] target(s) in 0.0 secs
     Running target/debug/deps/euclid-be4e9b9314cd979d

running 77 tests
test length::tests::test_add ... ok
test length::tests::test_cast ... ok
test length::tests::test_addassign ... ok
test length::tests::test_clone ... ok
test length::tests::test_division_by_length ... ok
test length::tests::test_fmt_debug ... ok
test length::tests::test_division_by_scalefactor ... ok
test length::tests::test_get_clones_length_value ... ok
test length::tests::test_fmt_display ... ok
test length::tests::test_equality ... ok
error: process didn't exit successfully: `/mnt2/dev/euclid/target/debug/deps/euclid-be4e9b9314cd979d` (signal: 11, SIGSEGV: invalid memory reference)

To learn more, run the command again with --verbose.

cc @pcwalton @SimonSapin

[brson]

I also see a sigsegv in heapsize 0.3.9.

[brson]

Oh this is repr(C) bugs.

[brson]

Actually, no, I don't know what this is.

[brson]

Heapsize has a PR to fix this re a change in the empty vector representation: servo/heapsize#81

[brson]

What PR caused this change to the size of Vec?

[Mark-Simulacrum]

42a4f37 - #41773; though I don't see anything obvious in there... not sure.

[Mark-Simulacrum]

Looks like this will indeed be fixed by the heapsize PR.

#0  malloc_usable_size (ptr=<optimized out>) at /checkout/src/liballoc_jemalloc/../jemalloc/include/jemalloc/internal/arena.h:743
#1  0x00005614c6ee2bdd in heapsize::heap_size_of_impl (ptr=0x4) at /root/.cargo/registry/src/github.com-1ecc6299db9ec823/heapsize-0.3.9/src/lib.rs:50
#2  0x00005614c6ee2ba4 in heapsize::heap_size_of (ptr=0x4) at /root/.cargo/registry/src/github.com-1ecc6299db9ec823/heapsize-0.3.9/src/lib.rs:36
#3  0x00005614c6ebd0e4 in heapsize::{{impl}}::heap_size_of_children<f32> (self=0x7f2bae3fe4d0) at /root/.cargo/registry/src/github.com-1ecc6299db9ec823/heapsize-0.3.9/src/lib.rs:234
#4  0x00005614c6ec30fd in euclid::length::{{impl}}::heap_size_of_children<euclid::length::tests::Inch,collections::vec::Vec<f32>> (self=0x7f2bae3fe4d0) at src/length.rs:49
#5  0x00005614c6ec4bd4 in euclid::length::tests::test_heapsizeof_length_vector () at src/length.rs:244
#6  0x00005614c6ef81dc in test::run_test::{{closure}} () at /checkout/src/libtest/lib.rs:1430
#7  core::ops::FnOnce::call_once<closure,(())> () at /checkout/src/libcore/ops.rs:2626
#8  test::{{impl}}::call_box<(),closure> () at /checkout/src/libtest/lib.rs:140
#9  0x00005614c6f3965b in panic_unwind::__rust_maybe_catch_panic () at /checkout/src/libpanic_unwind/lib.rs:98
#10 0x00005614c6eea684 in std::panicking::try<(),std::panic::AssertUnwindSafe<closure>> () at /checkout/src/libstd/panicking.rs:433
#11 std::panic::catch_unwind<std::panic::AssertUnwindSafe<closure>,()> () at /checkout/src/libstd/panic.rs:361
#12 test::run_test::run_test_inner::{{closure}} () at /checkout/src/libtest/lib.rs:1375
#13 std::panic::{{impl}}::call_once<(),closure> () at /checkout/src/libstd/panic.rs:296
#14 std::panicking::try::do_call<std::panic::AssertUnwindSafe<closure>,()> () at /checkout/src/libstd/panicking.rs:454
#15 0x00005614c6f3965b in panic_unwind::__rust_maybe_catch_panic () at /checkout/src/libpanic_unwind/lib.rs:98
#16 0x00005614c6ef258d in std::panicking::try<(),std::panic::AssertUnwindSafe<closure>> () at /checkout/src/libstd/panicking.rs:433
#17 std::panic::catch_unwind<std::panic::AssertUnwindSafe<closure>,()> () at /checkout/src/libstd/panic.rs:361
#18 std::thread::{{impl}}::spawn::{{closure}}<closure,()> () at /checkout/src/libstd/thread/mod.rs:361
#19 alloc::boxed::{{impl}}::call_box<(),closure> () at /checkout/src/liballoc/boxed.rs:648
#20 0x00005614c6f31386 in alloc::boxed::{{impl}}::call_once<(),()> () at /checkout/src/liballoc/boxed.rs:658
#21 std::sys_common::thread::start_thread () at /checkout/src/libstd/sys_common/thread.rs:21
#22 std::sys::imp::thread::{{impl}}::new::thread_start () at /checkout/src/libstd/sys/unix/thread.rs:84
#23 0x00007f2bb148d6ba in start_thread (arg=0x7f2bae3ff700) at pthread_create.c:333
#24 0x00007f2bb0fad82d in clone () at ../sysdeps/unix/sysv/linux/x86_64/clone.S:109

[SimonSapin]

#41064 in src/liballoc/heap.rs changed pointer used for zero-size allocations. heapsize could not use alloc::heap::EMPTY directly since alloc is #[unstable], so instead it used ptr == 1 as *mut ().

[brson]

Expected breakage. Closing.