fix: align Undici headersTimeout with apiRequestTimeout for OpenAI-compatible providers

[roomote-v0]

Related GitHub Issue

Closes: #12244

Description

This PR attempts to address Issue #12244. Feedback and guidance are welcome.

The OpenAI SDK's timeout option only controls the SDK-level request timeout but does not override Undici's default headersTimeout of 300 seconds (5 minutes). When time-to-first-byte exceeds 300s (common with large local models via LM Studio, Ollama, or other OpenAI-compatible providers), Undici throws a HeadersTimeoutError before the SDK timeout fires.

How it works:

• Created a shared createFetchWithUndiciTimeout() utility in src/api/providers/utils/undici-fetch.ts that builds a custom fetch function backed by an Undici Agent with headersTimeout and bodyTimeout aligned to the configured apiRequestTimeout.
• Applied the custom fetch to all three provider paths that create OpenAI SDK clients:
  • lm-studio.ts (LM Studio provider)
  • openai.ts (generic OpenAI-compatible provider, including Azure variants)
  • base-openai-compatible-provider.ts (base class for providers like DeepSeek, xAI, etc.)
• No new dependencies needed -- undici is already in src/package.json.

As noted by @yangzetao in the issue, this fix covers not just LM Studio but the generic OpenAI-compatible provider path as well, ensuring consistency across all providers using the OpenAI SDK transport layer.

Test Procedure

• Added unit tests for the new createFetchWithUndiciTimeout() utility (undici-fetch.spec.ts)
• Updated existing timeout tests for all three providers to verify the custom fetch option is passed to the OpenAI SDK constructor
• All 18 tests pass across 4 test files

Run tests:

cd src && npx vitest run api/providers/utils/__tests__/undici-fetch.spec.ts api/providers/__tests__/lm-studio-timeout.spec.ts api/providers/__tests__/openai-timeout.spec.ts api/providers/__tests__/base-openai-compatible-provider-timeout.spec.ts

Pre-Submission Checklist

• Issue Linked: This PR is linked to an approved GitHub Issue (see "Related GitHub Issue" above).
• Scope: My changes are focused on the linked issue (one major feature/fix per PR).
• Self-Review: I have performed a thorough self-review of my code.
• Testing: New and/or updated tests have been added to cover my changes.
• Documentation Impact: No documentation changes needed -- this is an internal transport-layer fix.
• Contribution Guidelines: I have read and agree to the Contributor Guidelines.

Screenshots / Videos

N/A -- no UI changes.

Documentation Updates

No documentation updates needed. This is an internal fix to the Undici transport layer timeout configuration.

Interactively review PR in Roo Code Cloud