Repository navigation
fix: 修复 v2.8.5 工具结果历史、文本动作解析、空预留输出上限及北向 OTel - #4328
Merged
Merged
Conversation
Codecov Report❌ Patch coverage is
📢 Thoughts on this report? Let us know! |
JasonW404
marked this pull request as ready for review
November 13, 2026 03:14
JasonW404
requested review from
Dallas98,
WMC001,
YehongPan,
hhhhsc701 and
jeffwu-1999
as code owners
November 13, 2026 03:14
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
本 PR 修复 v2.8.5 中的工具结果历史渲染、CodeAgent 文本动作解析、空输出预留导致的隐式生成上限,以及北向服务 OTel 链路问题。
VERSION为v2.8.5,源分支为hotfix/v2.8.5,目标分支为main。修改范围:
default_output_reserve_tokens为空且未显式指定输出预算时,不再将内部兜底预留作为max_tokens=4096发给模型。内部仍保留用于输入容量、压缩阈值及上下文检查的规划预算;显式模型/Agent/请求预算、快照校验和有界溢出恢复保持生效。诊断信息区分内部规划预留与实际发送的输出上限。此项产品代码仅修改openai_llm.py(新增 38 行、删除 13 行);含测试和资产说明共 7 个文件(新增 429 行、删除 18 行)。/nb/覆盖原有/nb/v1/generate_title路由,/agent/internal/northbound/覆盖 Runtime 内部调用;显式 URL 排除配置仍优先。验证结果:
43b6a2bd9):1303 项产品单元测试、25 项测试基础设施单元测试通过;TPC-D3-001 的 35 项运行时检查、NB-OTEL-001 的 3 项真实模型/Phoenix 检查,以及新旧解析路径的 8 组容器历史验证通过。覆盖工具执行一次、真实结果进入后续模型请求、普通流式行为、跨服务 trace 父子关系、未授权请求及断连后的持久化运行继续完成。72a308497):24 个独立模块共 901 项产品单元测试通过;8 个正式 D1 组件用例共 11 项检查通过,分别统计。19 项针对性回归覆盖空预留省略参数、显式预算、采样参数回退、实际流式响应、历史压缩、最终上下文容量检查、快照一致性及有界溢出恢复。git diff --check通过。更新描述时,最新提交72a308497的 12 项 GitHub 检查全部通过,包括自动单元测试、已配置的 AMD64/ARM64 镜像构建、CodeQL 及变更覆盖率检查。本机部署验证:
已用最新提交构建 AMD64
nexent/nexent:v2.8.5,更新 Config、Runtime、Northbound、MCP 四个服务。持久化挂载及原有网络连接保持一致,其余受检容器未重建;Config/Runtime/Northbound/Web 的 HTTP 检查均返回 200。镜像内 SDK 代码哈希与源码一致,部署后的 Runtime 内 19 项针对性回归全部通过。另通过真实模型执行一次不保存会话的发票技能创建,约 15 秒完成,正常生成技能正文和摘要且无错误事件;运行日志确认该模型请求未发送max_tokens。已知限制及范围边界:
<code>误判为未完成动作,造成不必要的重试;前端复用答案位置时,后续答案可能显示在先前答案的位置。这些问题暂不处理。<||DSML|| calls>等调用控制标记,偏离 Nexent 的<code>...</code>动作协议。关闭协议自动修复的旧版兜底路径可能将这类文本当作最终答案直接结束运行,其中的代码不会执行。该问题已在本机工作台会话中确认,本 PR 暂不修复,也不声明支持 DSML 调用格式;前述“合法代码正常执行”的验证不覆盖此类不受支持的格式。max_tokens后,实际输出上限由供应商决定;内部规划预留不保证实际生成长度。真实模型冒烟及受控组件测试不代表所有供应商兼容性验收。