Skip to content

consensus: CLI renderer is untestable inline print — extract shared render seam + parity test with MCP renderer #62

Description

@kiki830621

Problem

CLI 的 consensus renderer 是 ConsensusCommand.run() 裡的裸 print 序列,無可測 seam;MCP 端的 Server.renderConsensusSummary 是純函式。兩邊的 #38/#39 輸出邏輯是手工複製的雙份——#39 verify 能抓到 overwrote note 在兩個 renderer 都缺席,靠的是人工比對,不在測試裡(#38 verify R1 regression R1 的結構性附帶發現)。

#38 fix round 已在 MCP renderer 補契約測試((n=X) / prior 措辭 / solo 行 / refused 細節),CLI 仍是手工同步、零覆蓋——今天把 CLI 端整段 n 邏輯刪掉,全套件仍綠。

Expected

把 CLI 輸出組裝抽成純函式(或讓 CLI 直接消費同一份 render 函式),加「兩 surface 對同一 summary 產出等價內容」的對稱測試。

Source

surfaced during /idd-verify #38 R1(regression R1 附帶結構發現;verify report: https://github.com/PsychQuant/bestOCR/issues/38#issuecomment-5231307896)

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions