fix(test): update session integ test for sliding window conversation manager

[mkmeral]

Description

The integration test test_agent_with_file_session_and_conversation_manager fails after #2087 which enforces that the first remaining message after sliding window trimming must be a user message.

The test used window_size=1, which can no longer trim a simple (no-tool) conversation — after one turn the messages are [user, assistant], and trimming to 1 would leave only [assistant], which violates the new user-first rule. The conversation manager now correctly skips trimming in this case, leaving 2 messages instead of the expected 1.

This PR updates the test to use window_size=2 and adds a second agent call before the session reload, which still exercises all the same behaviors:

• Session manager persists all messages (including ones trimmed from agent memory)
• Conversation manager trims in-memory messages when the window is exceeded
• Trimming happens before the session reload (not just after)
• Restored agent loads the trimmed state correctly
• removed_message_count is tracked and restored across sessions
• Full message history is preserved in the session across multiple turns

With window_size=2, two agent calls produce 4 messages which get trimmed to 2 [user, assistant] — a valid window starting with a user message. After restoring and calling again, a third trim verifies the state was correctly persisted and restored.

Related Issues

Follow-up to #2087

Type of Change

Bug fix

Testing

Ran the integration test suite locally. The updated test passes with the new sliding window behavior.

• I ran hatch run prepare

Checklist

• I have read the CONTRIBUTING document
• I have added any necessary tests that prove my fix is effective or my feature works
• I have updated the documentation accordingly
• I have added an appropriate example to the documentation to outline the feature, or no new docs are needed
• My changes generate no new warnings
• Any dependent changes have been merged and published

---

By submitting this pull request, I confirm that you can use, modify, copy, and redistribute this contribution, under the terms of your choice.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!