Wasmtime tests hanging locally

[bjorn3]

$ cargo test
[snip]
test cli_tests::test_programs::cli_serve_trap_before_set has been running for over 60 seconds
test gc::drc_traces_the_correct_number_of_gc_refs_in_arrays has been running for over 60 seconds
test linker::module_interposition has been running for over 60 seconds
test memory::craneliftnative_offsets_static_dynamic_oh_my has been running for over 60 seconds
test memory::craneliftnative_tiny_static_heap has been running for over 60 seconds
test memory::craneliftpulley_offsets_static_dynamic_oh_my has been running for over 60 seconds
test memory::winch_offsets_static_dynamic_oh_my has been running for over 60 seconds
test memory_creator::not_for_windows::host_memory_grow has been running for over 60 seconds
test stack_overflow::craneliftpulley_big_stack_works_ok ... ok
test memory::winch_offsets_static_dynamic_oh_my ... ok
test module::craneliftpulley_concurrent_type_registry_modifications ... ok
test memory_creator::not_for_windows::host_memory_grow ... ok
test gc::drc_traces_the_correct_number_of_gc_refs_in_arrays ... ok
test stack_overflow::craneliftnative_big_stack_works_ok ... ok
test relocs::mixed_forced ... ok
test relocs::mixed ... ok
test memory::craneliftpulley_offsets_static_dynamic_oh_my ... ok
test memory::craneliftnative_tiny_static_heap ... ok
test linker::module_interposition ... ok
test memory::craneliftnative_offsets_static_dynamic_oh_my ... ok

The test binary was using 200% cpu when the warnings about slow execution were printed and dropped to 0% when the couple test passes happened with a wasmtime instance also using 0%. This wasmtime instance is wasmtime serve --addr=127.0.0.1:0 /path/to/wasmtime/target/debug/build/test-programs-artifacts-51f05083d9eb061b/out/wasm32-wasip1/debug/cli_serve_trap_before_set.component.wasm -Scli --shutdown-addr=127.0.0.1:0.

This is on current main (9362e47) on Linux and consistently reproduces for me.

[alexcrichton]

Do you have RUST_BACKTRACE=1 set in your environment? I've seen this befor with that on this specific test and I keep forgetting to either file an issue for it or fix it. Basically the test generates too much output and the server integration we have doesn't work with that.

[bjorn3]

Yes, I do have RUST_BACKTRACE=1.