Sandbox issue with Bodo: blocking syscalls

What version of Codex is running?

0.36

Which model were you using?

gpt-5-codex high

What platform is your computer?

Darwin 24.6.0 arm64 arm

What steps can reproduce the bug?

If you run
└ python - <<'PY'
import platform, sys
print('Python:', sys.version)
print('Platform:', platform.platform())

try:
import bodo
print('Bodo version:', getattr(bodo, '__version__', 'unknown'))
except Exception as exc:
import traceback
traceback.print_exc()
PY
Abort(136435343): Fatal error in internal_Init_thread: Other MPI error, error stack:
internal_Init_thread(67).: MPI_Init_thread(argc=0x0, argv=0x0, required=3, provided=0x16fd82694) failed
MPII_Init_thread(234)....:
MPID_Init(513)...........:
MPIDI_OFI_init_local(604):
create_vni_context(987)..: OFI EP enable failed (ofi_init.c:987:create_vni_context:Operation not permitted)

It immediately aborts during MPI_Init_thread, failing at create_vni_context with “Operation not permitted”.

I think the sandbox blocks socket syscalls, so libfabric fails to open an endpoint?

What is the expected behavior?

(global_env) rohit@Strange-Loop-Mac jax_for_data % python <<'PY'
import datajax
from datajax.runtime.executor import active_backend_name
print("backend:", active_backend_name())
PY
backend: bodo

What do you see instead?

It doesn't run inside the sandbox. Maybe the sandbox blocks socket syscalls, so libfabric fails to open an endpoint?

Additional information

_No response_