Skip to content

Fix tool_choice stall on multi-call prompts (#17998)#19848

Open
AjAnubolu wants to merge 1 commit intosgl-project:mainfrom
AjAnubolu:fix/tool-choice-multi-call-stall
Open

Fix tool_choice stall on multi-call prompts (#17998)#19848
AjAnubolu wants to merge 1 commit intosgl-project:mainfrom
AjAnubolu:fix/tool-choice-multi-call-stall

Conversation

@AjAnubolu
Copy link
Copy Markdown
Contributor

Summary

When tool_choice specifies a function name, the JSON schema included maxItems: 1 which masked the comma token after the first tool call, causing the model to loop on whitespace until max_tokens. Removed the constraint so multiple calls to the same function work correctly.

Closes #17998

@gemini-code-assist
Copy link
Copy Markdown
Contributor

Summary of Changes

Hello, I'm Gemini Code Assist1! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request addresses a critical bug that caused the model to stall when processing multi-call prompts with a specified tool_choice. By removing an erroneous "maxItems" constraint from the generated JSON schema, the system can now correctly handle multiple tool calls, preventing infinite loops and improving the reliability of function calling.

Highlights

  • JSON Schema Correction: Removed the "maxItems: 1" constraint from the JSON schema generated for tool_choice when a specific function name is provided.
  • Multi-Call Prompt Fix: Resolved an issue where the model would stall on multi-call prompts due to the "maxItems" constraint masking the comma token after the first tool call.
  • Test Updates: Modified existing tests to assert the absence of the "maxItems" property in the generated schema, reflecting the change.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Changelog
  • python/sglang/srt/function_call/utils.py
    • Removed the "maxItems: 1" property from the JSON schema generated for tool call arrays.
  • test/registered/function_call/test_json_schema_constraint.py
    • Updated test_specific_tool_choice_schema to assert that "maxItems" is no longer present in the schema.
    • Updated test_specific_tool_choice_dict_schema to assert that "maxItems" is no longer present in the schema.
Activity
  • No human activity has been recorded on this pull request yet.
Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature Command Description
Code Review /gemini review Performs a code review for the current pull request in its current state.
Pull Request Summary /gemini summary Provides a summary of the current pull request in its current state.
Comment @gemini-code-assist Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help /gemini help Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for Github and other Google products, sign up here.

You can also get AI-powered code generation, chat, as well as code reviews directly in the IDE at no cost with the Gemini Code Assist IDE Extension.

Footnotes

  1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution.

Copy link
Copy Markdown
Contributor

@gemini-code-assist gemini-code-assist Bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Code Review

This pull request addresses a bug that caused stalls on multi-call prompts when tool_choice specified a particular function. The fix involves removing the maxItems: 1 constraint from the generated JSON schema, which correctly allows for multiple calls to the same function. The associated tests have been updated to verify that the maxItems property is no longer part of the schema. The changes are correct and effectively resolve the issue.

@b8zhong b8zhong closed this Mar 5, 2026
@b8zhong b8zhong reopened this Mar 5, 2026
@JustinTong0323 JustinTong0323 self-assigned this Mar 6, 2026
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

[Bug] tool_choice with specific function causes model to stall on multi-call prompts

3 participants