Home/pyrite/Bug: priority field type mismatch between entry types and DB column

Bug: priority field type mismatch between entry types and DB columnbacklog_item

bugstorageprotocolstype-system
1 min read · Edit on Pyrite

Bug: priority field type mismatch between entry types and DB column

Problem

The `priority` DB column is defined as `Integer` in `pyrite/storage/models.py`, but entry types like `BacklogItemEntry` use string values ("high", "medium", "low"). The Prioritizable protocol in ADR-0017 defines priority as an integer (1-5 scale).

This mismatch means:

  • String priorities from backlog items are stored in an Integer column (SQLite is lenient, Postgres will reject)
  • The dual-source workaround in `sw backlog` checks `r.get("priority") or r["_meta"].get("priority", "medium")` — mixing column (int) and metadata (string)
  • Cross-type priority queries ("all high-priority items") won't work consistently

    • Options

    1. Change DB column to Text — simple, but loses ability to do numeric comparison 2. Map strings to integers during indexing: high=1, medium=3, low=5 — preserves schema intent 3. Change protocol to use string priorities — matches how humans think about priority

    Files

  • `pyrite/storage/models.py` — `priority` column type
  • `pyrite/storage/index.py` — `_entry_to_dict` priority handling
  • `pyrite/models/protocols.py` — Prioritizable protocol definition

    • Linked from

    Pyrite Prioritized Backlog Schema validate CLI command and pre-commit hook