ISSUE DESCRIPTION Bug: retainrecentlegalsuffix 在 else 分支下会导致用户消息被重复归档 Bug Description 在…
- User
ISSUE DESCRIPTION
Bug: retain_recent_legal_suffix 在 else 分支下会导致用户消息被重复归档
Bug Description
在
nanobot/session/manager.py的retain_recent_legal_suffix方法中,当 else 分支(最近 N 条消息全是助手/工具消息)被触发时,由于cut值的计算使用len(tail) - len(kept),而没有考虑用户消息的实际位置,导致用户消息被错误地同时归档到 archive 和 kept 中。具体影响:
- 用户消息同时出现在 archive(待压缩)和 kept(保留)中
- 可能导致 LLM 上下文不一致
Steps to Reproduce
- 会话中消息数量超过 _RECENT_SUFFIX_MESSAGES (默认10条)
- 最近 N 条消息中 不包含任何用户消息 (全是 assistant/tool 角色)
- 在完整会话历史中,最后一条用户消息位于倒数第 11-15 条之间
Expected Behavior
当会话历史中最近 N 条消息全是助手/工具消息,而用户消息位于更早位置时:
- 用户消息不应被重复归档 - 用户消息应该只保留在 kept 中,不应该同时出现在 archive 和 kept 中
- 归档边界应该正确 - cut 值应该基于用户消息的实际位置,而不是简单的 len(tail) - len(kept)
- 上下文连贯性应该保持 - 从用户消息开始的完整上下文应该被保留,不应该被错误切割
假设:
tail = messages[0:100](100条)
latest_user = 85(用户消息位置)
max_messages = 10
期望:
cut = latest_user # 应该是 85,不是 90 archive = tail[:85] # [0-84],不包含用户消息 kept = tail[85:] # [85-99],包含用户消息及其后续内容
结果:
archive 和 kept 没有重叠 ✅
用户消息只在 kept 中 ✅
上下文连贯 ✅
总结
期望行为 :无论用户消息位于何处,它都应该:
- 只出现在 kept 中(保留消息)
- 不出现在 archive 中(归档消息)
- 作为保留消息的起始点,确保上下文连贯
Relevant Logs
nanobot Version
0.2.0
Python Version
3.11
Operating System
Windows
Channel / Platform
Weixin (Personal WeChat)
LLM Provider
OpenAI
Configuration (Optional)
Additional Context
No response
Interface
Type: Function Name: retain_recent_legal_suffix Path: nanobot/session/manager.py Input: self, max_messages: int Output: tuple[list[dict], int] Description: Trims the session's message list to at most
max_messageswhile preserving a legal suffix (ending with a user message if possible). Returns the list of dropped messages and a count of how many of those were already in the consolidated prefix, so callers can archive only new, unconsolidated messages.ADDITIONAL ADVICE
Since you are given a git repository, you can use git commands to simplify your work. For example, if you made a mistake and want to revert your changes, you can use
git checkout <file>to restore the file to its original state. You can also reset all changes in the repository usinggit reset --hardcommand. Additionally, you can usegit stashandgit stash popto temporarily save your changes and restore them later. Finally, you can see the changes that you've made (which will be submitted when you callsubmit) usinggit statusorgit diffcommands. However you don't need to usegit addorgit commitbefore submitting your changes. In fact,submitonly submits currently unstaged changes because it usesgit diffto get the changes that need to be submitted.CHECKLIST
Before submitting your solution, please go over the following checklist and make sure you've done everything:
- If an error was reported in the issue description, I have successfully reproduced it.
- If an error was reported in the issue description, I have confirmed that my fix resolves the error.
- I have thought about edge cases and made sure my fix handles them as well.
- I have run existing tests in the repository that might have been affected by the change I made and confirmed that they pass. I want you to list every bullet from this checklist and write your reasoning for why you think you did it or didn't need to.
Repository has been uploaded and your shell is currently at the repository root. Time to solve the issue!
(Current directory: /nanobot, current file: none) bash-$
- Assistant
I'll inspect the buggy method and surrounding session logic so we can reproduce and fix the duplicate-archive behavior.
- Assistant
Let me search for the retain_recent_legal_suffix method.
- Assistant
I'll read the method implementation and the related tests next.
- Assistant
I'll inspect related tests and git history to see the intended behavior of this method.
- Assistant
I'll examine the hard-cap test and how callers use the return value.
- Assistant
Let me look at callers of retain_recent_legal_suffix and the original commit, and also auto_compact where the archive logic might overlap.
- Assistant
I'll check the commit that introduced the else-branch logic and any discussion around the hard-cap.
- Assistant
Looking more carefully at the issue and the Interface: I understand the bug report and the expected behavior. The core issue is that the current
retain_recent_legal_suffixmethod uses a forward-capped window (tail[:max_messages]) when the recent suffix contains no user messages, which can cause the user message (if it's in the middle of the tail) to be incorrectly included in botharchiveandkept.The previous implementation before the hard-cap commit correctly handled this by extending backward from the end to the nearest user message, ensuring the kept suffix always starts with a user message and the archive is a pure prefix.
However, the hard-cap was added to prevent unbounded 3000+ assistant messages after a user message at position 0. The expected behavior in the issue wants to keep the full suffix from the latest user message (which may slightly exceed max_messages when the user is just outside the recent N messages).
To resolve this while maintaining the "at most max_messages" constraint, we should:
- Fix the else branch to start from the latest user message and keep to the end (as in the issue's example)
- Assistant
I'll reproduce the bug, then fix the method so the kept region is a pure suffix starting at the user message.
- Assistant
I'll inspect
find_legal_message_startand then implement the fix across callers. - Assistant
Implementing the fix: keep a pure suffix from the anchoring user message and return the dropped prefix so callers archive without overlap.
- Assistant
The edit landed in the wrong file — reverting that and applying the fix to
manager.py. - Assistant
Editing
manager.pywith the corrected method. - Assistant
Updating
compact_idle_sessionto use the returned dropped messages. - Assistant
Update the compact_idle_session to use returned dropped messages.
- Assistant
Verifying the bug fix and running related tests.
- Assistant
Running the related test suites next.
- Assistant
I'll verify a few more edge cases and inspect the final code.
- Assistant
Checking the diff and that
commands.pywas restored. - Assistant
Hmm "nothing to commit, working tree clean" and "Not currently on any branch" - that's odd. git diff showed changes though. Let me check status again.
Other models, same or similar prompt
Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.
Identical first prompt
claude-sonnet-5Compare side by side →
claude-sonnet-5Compare side by side →
claude-sonnet-5Compare side by side →
claude-sonnet-5Compare side by side →
claude-sonnet-5Compare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-35B-A3BCompare side by side →
Qwen/Qwen3.6-27BCompare side by side →
Qwen/Qwen3.6-27BCompare side by side →