Nightly OSS: Incompatibility of assistant with tool_calls and tool role in the /v1/responses endpoint#3174
Conversation
…alls and tool role in the /v1/responses endpoint Nightly Codex produced a focused contribution for openai#2677. Constraint: Automated nightly run; keep changes small and reviewable. Confidence: medium Scope-risk: narrow Tested: See uploaded nightly artifacts and workflow logs. Not-tested: Maintainer CI beyond this workflow.
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 2076f95282
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "type": "function_call", | ||
| "call_id": "call_123", | ||
| "name": "get_order_status", | ||
| "arguments": '{"order_id":"order_123"}', | ||
| }, | ||
| { | ||
| "type": "function_call_output", |
There was a problem hiding this comment.
Cover assistant/tool-role regression in the new test
This test does not exercise the incompatibility described in the commit title (chat-style assistant messages with tool_calls plus tool role follow-ups), because it only sends Responses-native function_call/function_call_output items. As a result, the reported failing migration path can still regress (or remain broken) while this test stays green, so the added coverage does not protect the behavior the change claims to validate.
Useful? React with 👍 / 👎.
Automated nightly Codex contribution for #2677.
Verification:
This PR is generated by xodn348/oss-nightly-control and is opened ready for maintainer review when the local nightly verification step succeeds.