PythonVirtualenvOperator hangs indefinitely on Variable.get() (Task SDK) under airflow dags test and dag.test() — no error, no timeout

### Under which category would you file this issue? Airflow Core ### Apache Airflow version 3.2.0 ### What happened and how to reproduce it? When running a DAG that uses `PythonVirtualenvOperator` whose callable calls `from airflow.sdk import Variable; Variable.get(...)`, the venv subprocess hangs forever as soon as `Variable.get()` is called. There is no error, no timeout, no traceback — the `airflow dags test / dag.test()` invocation simply never returns. The hang reproduces: - under `dag.test()` (in-process execution) - under `airflow dags test <dag_id>` with a separately running airflow api-server reachable via `AIRFLOW__CORE__EXECUTION_API_SERVER_URL` In both cases the last log line from the venv is the user's own `print("Python inside virtualenv: 3.10.17 ...")` — i.e. the subprocess is alive, the callable has started, and execution blocks on the very next line which is Variable.get(...). Versions: - apache-airflow-core==3.2.0 - apache-airflow-providers-standard==1.12.3 - apache-airflow-task-sdk==1.2.0 - Python 3.10.17 (host interpreter and venv interpreter) - SQLite metadata DB - OS: Linux **How to reproduce** Minimal script (single file, runs end-to-end): ``` import json import os import shutil import subprocess import sys from datetime import datetime, timezone from pathlib import Path AIRFLOW_HOME = Path("/tmp/airflow32_venv_hang") AIRFLOW_DB = AIRFLOW_HOME / "airflow.db" DAG_BUNDLE_DIR = AIRFLOW_HOME / "dags_bundle" AIRFLOW_HOME.mkdir(parents=True, exist_ok=True) DAG_BUNDLE_DIR.mkdir(parents=True, exist_ok=True) _THIS_FILE = Path(__file__).resolve() _BUNDLE_FILE = (DAG_BUNDLE_DIR / "venv_hang_repro.py").resolve() if _THIS_FILE != _BUNDLE_FILE: shutil.copy(_THIS_FILE, _BUNDLE_FILE) os.environ["AIRFLOW_HOME"] = str(AIRFLOW_HOME) os.environ["AIRFLOW__DATABASE__SQL_ALCHEMY_CONN"] = f"sqlite:///{AIRFLOW_DB}" os.environ["AIRFLOW__CORE__LOAD_EXAMPLES"] = "False" os.environ["AIRFLOW__DAG_PROCESSOR__DAG_BUNDLE_CONFIG_LIST"] = json.dumps([{ "name": "dags-fol…