root/Jyotisha
1
0
Fork 0
Code Issues 1 Pull Requests 1 Actions Packages Projects Releases Wiki Activity

docs(testing): consult plain answer — real-model old/new comparison and acceptance

Browse Source 
Co-Authored-By: Claude Opus 5.5 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01N4f2nya58RoRu4yEmJgRGE
This commit is contained in:
Jesse_Chen
2026-10-01 10:36:32 +08:00
co-authored by Claude Opus 5.5
parent a6f6f576bb
commit b9215d21df
2 changed files with 249 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
docs/tasks/PROGRESS-consult-plain-answer-20261001.md
+6
View File
@@ -84,3 +84,9 @@
## 变基复跑
推分支前 `origin/staging` 前进到 `c304cf3a`（BUG-1140 构建修复 + 新增 `runtime-sources-no-test-imports` 测试 2 条）。本分支变基到其上后复跑：tsc 0；lint 0 error / 126 warning；`npm test` 4,815（pass 4,750 / fail 24 / skip 41），失败名单与基线逐条一致；`next build` `/` ○ Static。`docs/BUG_HISTORY.md` 冲突按编号顺序合并（1132～1134 在 1140 之前）。
## Claude 验收（2026-10-01）
- 独立复跑：tsc 0；lint 0 error / 126 warning；`npm test` 4,815 条 fail 24，与基线 `c304cf3a` 逐条同名；`next build` 后 `/` Static。推 staging `44cdd9c9`（快进）。
- 真实模型对比（产品提供临时 DeepSeek key）：见 `docs/testing/consult-plain-answer-20261001-model-runs.md`。旧提示词 3/3 复现事故形状，新提示词 5/5 无格局名 / 推修 / 扮演象，父母题均按妈妈、爸爸、他们怎么对你分段。
- 待产品决定：`deepseek-v4-pro` 一条 77 s 超过 70 s 答题时钟（会截断最后一节）；flash 30–41 s 在钟内。