CI flake: Test Suite wasm-opt segfaults while linking C++ sdk-test module

[clockwork-labs-bot]

Created by CI Gremlin from a manually supplied failed GitHub Actions job.

Job:

• Workflow run: https://github.com/clockworklabs/SpacetimeDB/actions/runs/30484985781
• Exact job: https://github.com/clockworklabs/SpacetimeDB/actions/runs/30484985781/job/90689815850?pr=5618
• Run attempt: 1
• Job name: Test Suite
• PR: Smoketests for default values #5618, bfops/consolidate-default-tests into bfops/test-suite-fixes
• Run head SHA: ba860143df4d80e7a33d512dfa8e66a4a080ab26

Failure signature:

• Step Run tests failed while running cargo test -p spacetimedb-sdk --features allow_loopback_http_for_tests -- --test-threads=2 --skip unreal.
• The first real failure was during C++ SDK test module compilation in modules/sdk-test-cpp:
  • emcmake: cmake -S . -B build ...
  • [100%] Linking CXX executable lib.wasm
  • em++: error: '/home/runner/emsdk/upstream/bin/wasm-opt ... lib.wasm -o lib.wasm ...' failed (received SIGSEGV (-11))
  • gmake[2]: *** [CMakeFiles/lib.dir/build.make:101: lib.wasm] Error 1
• After that, many SDK tests failed because the shared module build/cache path was poisoned:
  • called Result::unwrap() on an Err value: PoisonError { .. }
  • test result: FAILED. 25 passed; 261 failed; 4 ignored; 0 measured; 0 filtered out; finished in 60.06s

Root cause hypothesis:

• Toolchain / Binaryen wasm-opt segfault during Emscripten C++ module link. The 261 test failures are fallout from that first build failure, not independent SDK behavior failures.

Duplicate analysis:

• Top duplicate candidate: CI flake: segfault in wasm-opt? #5369, open, same broad wasm-opt segfault during C++ SDK module build with downstream shared-cache fallout.
• Related closed history: CI flake: Test Suite / typescript::insert_call_uuid_v4, typescript::indexed_simple_enum,... #5499 and CI flake: Test Suite / typescript::delete_uuid, typescript::indexed_simple_enum, typescript::insert_call_uuid_v4,... #5412 show similar cascading SDK failures, but CI flake: segfault in wasm-opt? #5369 is the open tracker.
• Per current CI Gremlin duplicate rules, I am not auto-closing this ticket because CI flake: segfault in wasm-opt? #5369 is not in the proactive duplicate-close allowlist.