Python: fix(core): correct FunctionResultContent ordering in WorkflowAgent.merge_updates

[moonbox3]

Motivation and Context

When using a thread with WorkflowAgent, the second call to run_stream fails with an OpenAI error:

Error: service failed to complete the prompt: Error code: 400 - {'error': {'message': "An assistant message with 'tool_calls' must be followed by tool messages responding to each 'tool_call_id'. The following tool_call_ids did not have response messages: call_F09je20iUue6DlFRDLLh3dGK", 'type': 'invalid_request_error', 'param': 'messages.[3].role', 'code': None}}

The root cause is in WorkflowAgent.merge_updates(). FunctionResultContent updates lack a response_id, so they were being added to global_dangling and appended at the end of messages instead of immediately after their corresponding FunctionCallContent.

Incorrect ordering (before fix):

- User Question
- FunctionCallContent
- Assistant Answer
- FunctionResultContent  <-- Wrong position

Correct ordering (after fix):

- User Question
- FunctionCallContent
- FunctionResultContent
- Assistant Answer

Description

• Fixes Python: WorkflowAgent Thread FunctionResultContent out of order #2977
• Add unit tests

Contribution Checklist

• The code builds clean without any errors or warnings
• The PR follows the Contribution Guidelines
• All unit tests pass, and I have added new tests where possible
• Is this a breaking change? If yes, add "[BREAKING]" prefix to the title of the PR.

[markwallace-microsoft]

Python Test Coverage Report •

File	Stmts	Miss	Cover	Missing
packages/core/agent_framework/_workflows
_agent.py	283	43	84%	61, 69–75, 103–104, 260, 316, 330, 343, 392–395, 401, 407, 411–412, 415–421, 425–426, 495, 502, 508–509, 515, 527, 559, 566, 587, 596, 600, 611
TOTAL	17658	2896	83%

Python Unit Test Overview

Tests	Skipped	Failures	Errors	Time
2620	154 💤	0 ❌	0 🔥	1m 0s ⏱️

[Copilot]

Pull request overview

This PR fixes a critical bug in WorkflowAgent.merge_updates() where FunctionResultContent updates without a response_id were incorrectly placed at the end of messages, causing OpenAI API validation errors.

Changes:

• Modified the merge logic to map FunctionResultContent to their corresponding FunctionCallContent via call_id matching
• Added comprehensive unit tests covering single and multiple function call/result scenarios, plus edge cases

Reviewed changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated 2 comments.

File	Description
python/packages/core/agent_framework/_workflows/_agent.py	Implemented two-pass algorithm: first pass builds call_id→response_id mapping from FunctionCallContent, second pass associates FunctionResultContent with matching calls
python/packages/core/tests/workflow/test_workflow_agent.py	Added three comprehensive tests validating correct ordering, multiple function calls, and orphan result handling