CI: Enable overflow checks for test (non-dist) builds

[hkratz]

They stay disabled for Apple builds though, which take the most time already due to running on slow hw.

[rust-highfive]

r? @kennytm

(rust-highfive has picked a reviewer for you, use r? to override)

[hkratz]

@rustbot label -S-waiting-on-review +S-waiting-on-author

[hkratz]

Currently fails due to #89639.

[hkratz]

@rustbot label -S-waiting-on-author +S-blocked

[alex]

FWIW I think you're going to find that #72549 fails as well.

…

On Mon, Oct 11, 2021 at 2:28 PM Hans Kratz ***@***.***> wrote: rustbot label -S-waiting-on-author +S-blocked — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#89776 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAAAGBAKSOGSWVC5HG4DGK3UGMUDBANCNFSM5FYXUERA> . Triage notifications on the go with GitHub Mobile for iOS <https://apps.apple.com/app/apple-store/id1477376905?ct=notification-email&mt=8&pt=524675> or Android <https://play.google.com/store/apps/details?id=com.github.android&referrer=utm_campaign%3Dnotification-email%26utm_medium%3Demail%26utm_source%3Dgithub>.

-- All that is necessary for evil to succeed is for good people to do nothing.

[Mark-Simulacrum]

Do we have a recent-ish measurement of how much this would slow down the compiler if enabled by default? I feel like there was a measurement a few months (years?) back...

Even if we don't enable the overflow checks within std, it seems reasonable to enable them for the compiler only, if we can afford it.

[hkratz]

@Mark-Simulacrum We have measured this as part of #87784 in August. Perf results:

• Overflow checks enabled for rustc and std
• Overflow checks enabled just for rustc

The slowdown is significant so to distributing rustc/std with overflow checks enabled is not a good idea. But I think it makes sense to enable the overflow checks for non-Apple non-dist CI builds. Debug assertions are enabled for those builds already and the overall impact on the CI (at least on x86_64-gnu-llvm-10 in this PR) is not that big. It avoids developer confusion encountering unexpected errors when using debug builds locally like #89639.

The slowest CI builds are the Apple ones anyway and for those we would keep it disabled (just as with debug assertions).

[Mark-Simulacrum]

Thanks. Yeah, I recall those measurements now. Agreed we likely can't accept that hit just yet.

I'm fine with us merging this and we can evaluate the CI time impact if it becomes a problem, it seems reasonable to have our test builders run with more assertions where we can afford it.

[oli-obk]

will also fail the tools builder due to #89280 which breaks clippy tests

[hkratz]

#89280 has been fixed. I have tested locally that x.py test --stage 2 and ./src/ci/docker/run.sh x86_64-gnu-tools pass with overflow checks enabled.

@rustbot label -S-blocked +S-waiting-on-review

r? @Mark-Simulacrum

[kennytm]

@bors r=Mark-Simulacrum rollup=iffy

[bors]

📌 Commit 5c8fca5 has been approved by Mark-Simulacrum

[bors]

⌛ Testing commit 5c8fca5 with merge 8ed99dd012f1305de11e33ffcb82e7f193b2feee...

[bors]

💔 Test failed - checks-actions

[rust-log-analyzer]

The job i686-msvc-1 failed! Check out the build log: (web)(plain)

Click to see the possible cause of the failure (guessed by this bot)

---- [js-doc-test] rustdoc-js\struct-like-variant.rs stdout ----
error: rustdoc-js test failed!
status: exit code: 1
command: "C:\\Program Files\\nodejs\\node" "D:\\a\\rust\\rust\\src/tools/rustdoc-js/tester.js" "--doc-folder" "D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\test\\rustdoc-js\\struct-like-variant" "--crate-name" "struct_like_variant" "--test-file" "D:\\a\\rust\\rust\\src/test\\rustdoc-js\\struct-like-variant.js"
------------------------------------------
------------------------------------------
stderr:
stderr:
------------------------------------------
internal/fs/utils.js:332
throw err;
Error: ENOENT: no such file or directory, open 'D:\a\rust\rust\build\i686-pc-windows-msvc\test\rustdoc-js\struct-like-variant\search.js'
at Object.openSync (fs.js:497:3)
at Object.readFileSync (fs.js:393:35)
at readFile (D:\a\rust\rust\src\tools\rustdoc-js\tester.js:181:15)
at load_files (D:\a\rust\rust\src\tools\rustdoc-js\tester.js:391:20)
at main (D:\a\rust\rust\src\tools\rustdoc-js\tester.js:471:27)
at Object.<anonymous> (D:\a\rust\rust\src\tools\rustdoc-js\tester.js:491:14)
at Module._compile (internal/modules/cjs/loader.js:1085:14)
at Object.Module._extensions..js (internal/modules/cjs/loader.js:1114:10)
at Module.load (internal/modules/cjs/loader.js:950:32)
at Function.Module._load (internal/modules/cjs/loader.js:790:12) {
errno: -4058,
syscall: 'open',
code: 'ENOENT',
path: 'D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\test\\rustdoc-js\\struct-like-variant\\search.js'
------------------------------------------
---
test result: FAILED. 13 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 2.54s
command did not execute successfully: "D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\stage0-tools-bin\\compiletest.exe" "--compile-lib-path" "D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\stage2\\bin" "--run-lib-path" "D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\stage2\\lib\\rustlib\\i686-pc-windows-msvc\\lib" "--rustc-path" "D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\stage2\\bin\\rustc.exe" "--rustdoc-path" "D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\stage2\\bin\\rustdoc.exe" "--src-base" "D:\\a\\rust\\rust\\src/test\\rustdoc-js" "--build-base" "D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\test\\rustdoc-js" "--stage-id" "stage2-i686-pc-windows-msvc" "--suite" "rustdoc-js" "--mode" "js-doc-test" "--target" "i686-pc-windows-msvc" "--host" "i686-pc-windows-msvc" "--llvm-filecheck" "D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\llvm\\build\\bin\\FileCheck.exe" "--nodejs" "C:\\Program Files\\nodejs\\node" "--npm" "C:\\Program Files\\nodejs\\npm" "--host-rustcflags" "-Cdebuginfo=0 -Lnative=D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\native\\rust-test-helpers" "--target-rustcflags" "-Cdebuginfo=0 -Lnative=D:\\a\\rust\\rust\\build\\i686-pc-windows-msvc\\native\\rust-test-helpers" "--docck-python" "C:\\hostedtoolcache\\windows\\Python\\3.10.0\\x64\\python3.exe" "--lldb-python" "C:\\hostedtoolcache\\windows\\Python\\3.10.0\\x64\\python3.exe" "--gdb" "C:\\msys64\\usr\\bin\\gdb" "--llvm-version" "13.0.0-rust-1.58.0-nightly" "--llvm-components" "aarch64 aarch64asmparser aarch64codegen aarch64desc aarch64disassembler aarch64info aarch64utils aggressiveinstcombine all all-targets analysis arm armasmparser armcodegen armdesc armdisassembler arminfo armutils asmparser asmprinter avr avrasmparser avrcodegen avrdesc avrdisassembler avrinfo binaryformat bitreader bitstreamreader bitwriter bpf bpfasmparser bpfcodegen bpfdesc bpfdisassembler bpfinfo cfguard codegen core coroutines coverage debuginfocodeview debuginfodwarf debuginfogsym debuginfomsf debuginfopdb demangle dlltooldriver dwarflinker dwp engine executionengine extensions filecheck frontendopenacc frontendopenmp fuzzmutate globalisel hexagon hexagonasmparser hexagoncodegen hexagondesc hexagondisassembler hexagoninfo instcombine instrumentation interfacestub interpreter ipo irreader jitlink libdriver lineeditor linker lto m68k m68kasmparser m68kcodegen m68kdesc m68kdisassembler m68kinfo mc mca mcdisassembler mcjit mcparser mips mipsasmparser mipscodegen mipsdesc mipsdisassembler mipsinfo mirparser msp430 msp430asmparser msp430codegen msp430desc msp430disassembler msp430info native nativecodegen nvptx nvptxcodegen nvptxdesc nvptxinfo objcarcopts object objectyaml option orcjit orcshared orctargetprocess passes powerpc powerpcasmparser powerpccodegen powerpcdesc powerpcdisassembler powerpcinfo profiledata remarks riscv riscvasmparser riscvcodegen riscvdesc riscvdisassembler riscvinfo runtimedyld scalaropts selectiondag sparc sparcasmparser sparccodegen sparcdesc sparcdisassembler sparcinfo support symbolize systemz systemzasmparser systemzcodegen systemzdesc systemzdisassembler systemzinfo tablegen target textapi transformutils vectorize webassembly webassemblyasmparser webassemblycodegen webassemblydesc webassemblydisassembler webassemblyinfo webassemblyutils windowsmanifest x86 x86asmparser x86codegen x86desc x86disassembler x86info xray" "--cc" "" "--cxx" "" "--cflags" "" "--adb-path" "adb" "--adb-test-dir" "/data/tmp/work" "--android-cross-path" "" "--channel" "nightly" "--color" "always"
Build completed unsuccessfully in 0:59:23
Build completed unsuccessfully in 0:59:23
make: *** [Makefile:72: ci-subset-1] Error 1

[Mark-Simulacrum]

@bors retry - missing search.js in rustdoc-js (?) tests

[bors]

⌛ Testing commit 5c8fca5 with merge a99c9d6...

[bors]

☀️ Test successful - checks-actions
Approved by: Mark-Simulacrum
Pushing a99c9d6 to master...

[rust-timer]

Finished benchmarking commit (a99c9d6): comparison url.

Summary: This benchmark run did not return any relevant changes.

If you disagree with this performance assessment, please file an issue in rust-lang/rustc-perf.

@rustbot label: -perf-regression

[hkratz]

It seems that enabling this uncovers non-deterministic overflows in rustc-rayon-core causing the build failures of #90042 and #90222:

• https://github.com/rust-lang-ci/rust/runs/3987259810?check_suite_focus=true#step:25:26550
• https://github.com/rust-lang-ci/rust/runs/3987471878?check_suite_focus=true#step:25:21315

Overflow location:
https://github.com/rust-lang/rustc-rayon/blob/c8ec88d8a2236d1fe19d65a4ab38834f76d256b7/rayon-core/src/sleep/mod.rs#L330

This code is not in upstream rayon but just in our fork, introduced here:
rust-lang/rustc-rayon@27911f7#diff-cde9d726ca2f32c319420be6c61d2e57ad9daff7f85a5f8f7be25474a15f9c4dR328