A high-throughput and memory-efficient inference and serving engine for LLMs
Mirrored from https://github.com/tduncan-lum/Nvidia-Optimization-Stack---VLLM.git
Synced Nov 10, 10:15 UTC
Fix type error (#31999)
Signed-off-by: Adolfo Victoria <adolfokarim@gmail.com> Co-authored-by: Adolfo Victoria <adovi@meta.com>
Adolfo Victoria committed
Jan 9, 2026 at 06:03 UTC
bbf80ede4325ae8bed35dfb1c04be63e6b0ce372
1 file changed
+3
-1
vllm/entrypoints/responses_utils.py
+3
-1
@@ -78,7 +78,9 @@ def _maybe_combine_reasoning_and_tool_call(
78
This function checks if the last message is a reasoning message and
79
the current message is a tool call"""
80
if not (
81
- isinstance(item, ResponseFunctionToolCall) and item.id.startswith(MCP_PREFIX)
81
+ isinstance(item, ResponseFunctionToolCall)
82
+ and item.id
83
+ and item.id.startswith(MCP_PREFIX)
84
):
85
return None
86
if len(messages) == 0: