rust: build failure for monterey intel: bootstrap compilation appears to fail

[mascguy (Christopher Nielsen)]

This port was rev-bumped earlier today, to rebuild with the latest version of libgit2. (See: issue:65733)

And despite no apparent changes to rust since it last built successfully, it's now failing for Monterey Intel:

Note: Absolute paths changed to relative, for brevity.

Building stage1 std artifacts (x86_64-apple-darwin -> x86_64-apple-darwin)

running: "rust/work/cargo-1.60.0-x86_64-apple-darwin/cargo/bin/cargo" "build" "--target" "x86_64-apple-darwin" "-Zbinary-dep-depinfo" "-j" "8"
"-v" "-v" "--release" "--features" "panic-unwind backtrace compiler-builtins-c" "--manifest-path" "rust/work/rustc-1.61.0-src/library/test/Cargo.toml"
"--message-format" "json-render-diagnostics"

error: process didn't exit successfully: `rust/work/rustc-1.61.0-src/build/bootstrap/debug/rustc -vV` (exit status: 254)
--- stdout

Did not run successfully: signal: 9
"rust/work/rustc-1.61.0-src/build/x86_64-apple-darwin/stage1/bin/rustc" "-vV" "-Wrust_2018_idioms" "-Wunused_lifetimes"
"-Wsemicolon_in_expressions_from_macros" "-Dwarnings" "-Clinker=rust/work/compwrap/ld/usr/bin/clang" "-Z" "force-unstable-if-unmarked"
-------------

--- stderr
rustc exited with signal: 9

command did not execute successfully: "rust/work/cargo-1.60.0-x86_64-apple-darwin/cargo/bin/cargo" "build" "--target" "x86_64-apple-darwin"
"-Zbinary-dep-depinfo" "-j" "8" "-v" "-v" "--release" "--features" "panic-unwind backtrace compiler-builtins-c"
"--manifest-path" "rust/work/rustc-1.61.0-src/library/test/Cargo.toml" "--message-format" "json-render-diagnostics"

expected success, got: exit status: 101

​https://build.macports.org/builders/ports-12_x86_64-builder/builds/40294/steps/install-port/logs/stdio

While most of the buildbots are still catching up, there have been no other failures thus far. And successful builds include:

• Big Sur, both ARM and Intel
• 10.15
• 10.14
• 10.12
• 10.9

So, the question is: Since the port hasn't changed, have there been any potential changes in our portgroups, etc, that might have caused this? Or...?

[mascguy (Christopher Nielsen)]

Build log also attached, for tracking/historical purposes.

[catap (Kirill A. Korinsky)]

Christopher just a guess: maybe builder hasn't got enough ram?

[mascguy (Christopher Nielsen)]

Replying to catap:

Christopher just a guess: maybe builder hasn't got enough ram?

Well, the Monterey Intel builder has 9gb, which is 1gb more than others. But perhaps there's some other process(es) running - like Software Update, and/or XProtect - which are competing for resources. (Though I think @ryandesign set up the VMs with virtual swap as well, to help avoid running out of memory.)

Still, it's a great point, as perhaps it was simply a fluke. So I've queued up a rebuild, and let's hope it works on the 2nd try.

In the interim, if anyone thinks there could be some other issue, feel free to chime in!

[catap (Kirill A. Korinsky)]

Well, the Monterey Intel builder has 9gb, which is 1gb more than others.

or maybe it has more CPU and rust just consumes all of them?

But it is a guess.

[mascguy (Christopher Nielsen)]

Replying to catap:

Well, the Monterey Intel builder has 9gb, which is 1gb more than others.

or maybe it has more CPU and rust just consumes all of them?

But it is a guess.

Anything's possible!

There were also two unusual failures on the watcher since then, though they could be related to a hosting/network blip:

• ​https://build.macports.org/builders/ports-12_x86_64-watcher/builds/11653
• ​https://build.macports.org/builders/ports-12_x86_64-watcher/builds/11652

Adding @ryandesign and @jmroot, for thoughts/guidance...

[mascguy (Christopher Nielsen)]

The Monterey Intel build succeeded on the 2nd attempt, so perhaps memory and/or disk space was exhausted during the first run.

Builds succeeded everywhere else too, so crisis averted.