todo2code

Ticket 042: Bound aggregate LLM deadline for workspace comparison

ID: ticket-042
Owner: human:founder
Status: DONE
Workflow state: DONE
Created: 2026-08-05

Goal and scope

Bound the total wall-clock budget of compare-workspace semantic analysis. The existing per-request adaptive timeout is correct, but a comparison runs two pipelines and may issue many document chunks; therefore sequential waves can outlive the intended task budget even though every request is bounded.

Authorization is the Founder’s explicit instruction in the active Codex session to scale timeout to task/input complexity, use 2x growth when the default is insufficient, continue execution, and use z-ai/glm-5.2 instead of Gemini 3.1 Pro Preview. No human-owned participant file is synthesized.

Acceptance criteria

AC-01: The active Founder instruction authorizes adaptive task/input timeout handling and continued implementation.
AC-02: The aggregate deadline accounts for both pipelines, document chunks and semantic stages using bounded 2x steps.
AC-03: Expiry aborts in-flight OpenRouter requests and leaves no detached comparison process or partial success claim.
AC-04: Focused, full, governance and Docker validation pass.

Participants

Human participant: human:founder, evidenced by the active conversation; no user-* file was generated or edited.
Agent participant: ai-codex.md

Protected completion evidence

Koru run 30987747726 passed exact head 5b51880cd448c295fbfdcb7bcc2e892864ac910f.
Validator run 30987864973 approved that same head for ticket-042 and correlation ID todo2code-pr52-5b51880cd448 using openrouter/z-ai/glm-5.2. The semantic verdict remained advisory rather than the approval trust root.
Review-triggered CI run 30987963474 passed governance, full verification, Docker smoke and the required Java fixture with current-head Validator evidence.
Protected PR #52 merged exact head 5b51880 as main@2c164492e7aa751b0fd159b80e07ee0fcc22384f without an administrative bypass.

This site is open source. Improve this page.