zai-org / zai-org/feedback

[Bug] 对话长文本生成时陷入死循环且无法恢复,重启后报错 "Partial assistant output was discarded before a streaming retry"

Open
#514 1 comment 0 reactions 0 assignees View on GitHub

Nobody has claimed this yet.

priority: P2 status: 待评估 type: Bug
Dominant language
No language data
Stars
22
Forks
1
PR merge metrics
No merged PRs in 30d

Description

提交前确认 · Pre-submission checklist
  • 我已搜索过现有 issue,确认这不是重复 / I searched existing issues and confirmed this isn't a duplicate.
  • 我已阅读 CONTRIBUTING.md / I've read CONTRIBUTING.md.
问题类别 · Category

对话 / Agent 交互 · Agent chat

涉及的 Agent 框架 · Agent framework

ZCode Agent(自研)

严重程度 · Severity

阻塞使用 · Blocking (无法使用核心功能 / core function unusable)

复现频率 · Reproducibility

偶现 · Sometimes

问题描述 · Description

在进行长代码/长文本(如全量重写某个核心代码文件)生成时,AI 陷入思考与流式输出的死循环,界面一直显示正在发送/生成请求且完全不会恢复。
当强制重启客户端后,由于本地对话状态加载异常,直接抛出流式重试报错:
"Partial assistant output was discarded before a streaming retry. TraceID: hydrate-trace"

复现步骤 · Steps to reproduce
  1. 打开 ZCode,进入包含较长上下文或大文件的项目。
  2. 输入提示词要求对大文件进行全量生成/重构(例如:全量重写 cfg_store.py)。
  3. AI 开始思考(Thinking)并进入多轮流式生成与“继续”的递归状态。
  4. 客户端陷入无限请求/等待状态,停止响应新操作且完全无法恢复。
  5. 强制关闭并重启 ZCode,重新打开该会话。
  6. 界面提示:Partial assistant output was discarded before a streaming retry. TraceID: hydrate-trace
期望表现 · Expected behavior
  1. 长文本或深度思考遇到超时/Token上限时,应正常捕获并终止,给出明确的截断或超时提示,而不是陷入持续发送请求的假死状态。
实际表现 · Actual behavior
  1. 遇到复杂长任务时持续处于发送状态,前端无法自行恢复。中止后下次对话仍有问题
  2. 重启应用后,状态水合失败,报错:
    报错摘要: Partial assistant output was discarded before a streaming retry.
    TraceID: hydrate-trace
ZCode 版本 · ZCode version

v3.11.2

设备 / 系统 / 浏览器 · Device / OS / Browser

Windows11 25H2 26200.9168

截图 / 录屏 / 日志 · Screenshots / Recordings / Logs
Image

Contributor guide

Open the contributing guide

First steps

  1. Read the whole issue, then the project's contributing guide.
  2. Comment on the issue to say you are picking it up — it saves two people doing the same work.
  3. Fork the repository and make your change on a branch.
  4. Open a pull request that references the issue number.

Research direction

No source files or tests are named. Start by reproducing the long-text generation loop, then trace the streaming retry flow and local conversation-state hydration after restart. Done means long requests terminate with a clear timeout or truncation message and the conversation reloads without the reported error.

Written by the indexing model from the issue text.

Assessment

Domain
desktop
Issue type
Bug
Difficulty
4/5
Estimated time
3-5 days
Activity status
Active
Clarity
Mostly clear
Newbie friendliness
45/100

Get new issues in your inbox

A short digest of beginner-friendly GitHub issues.