Story: Hotfix: the ER diagram no longer matches the SQL tree
Table of Contents
+#+filetags: :hotfix:sql:codegen:diagram:sprint_26:v0:
+#+updated: 2026-10-01 +#+environment: clever_dijkstra +#+todo: BACKLOG STARTED | DONE ABANDONED
This page documents a story in Sprint 26. It captures the goal, current status, acceptance criteria, and the tasks that compose it.
1. Goal
The drift job passes its first step and fails its second:
Check generated output matches the models: No drift: regenerated output matches the checked-in tree. Check the ER diagram matches the SQL schema: ##[error]Process completed with exit code 1.
The cross-component diagram is generated from the SQL tree and nothing regenerated it, so it fell behind the SQL that the reporting analytics, the asset class catalogue and the curve configuration merges added. Every pull request is red for it, including ones that touch no SQL.
2. Status
| Field | Value |
|---|---|
| State | DONE |
| Parent sprint | Sprint 26 |
| Now | Nothing. |
| Waiting on | Nothing. |
| Next | Nothing. |
| Last touched | 2026-10-01 |
3. Acceptance
ores_schema.pumlmatches a fresh render, soplantuml_er_generate.py --checkpasses.- The diagram tool is not needed: the render is text from the model.
4. Tasks
| Task | State | Start | End | Description |
|---|---|---|---|---|
| Implement Hotfix: the ER diagram no longer matches the SQL tree | DONE | 2026-10-01 | 2026-10-01 | Parse the SQL tree, regenerate the diagram, and prove the check clean. |
5. Decisions
- The diagram is regenerated rather than the check relaxed. The step exists because the diagram fell a month behind its own generator once before, and the answer then was the same.
6. Out of scope
- Why the merges that drifted it did not regenerate it: their branches do not fire the ER step, because the job's path filter watches the SQL tree and the generator, and the puml is neither.