Context
Right now, markdown files are the source of truth for skip decisions during import/refresh. This was formalized in #35 — SQLite is no longer consulted for skip logic to avoid stale state (e.g., deleting a page folder then re-importing would silently skip everything).
Friction area
There's no sync mechanism between SQLite and the markdown files. They can drift:
- Delete markdown → SQLite still thinks pages exist
- Edit markdown frontmatter → SQLite is stale
- SQLite-only output mode has no skip optimization now
Future considerations
- Should SQLite mirror markdown state on startup (reconciliation pass)?
- UX patterns for detecting and resolving drift between the two stores
- Whether SQLite should eventually become the source of truth (with markdown as an export artifact)
- Unit tests for
FreezePage skip logic edge cases (currently only covered by system tests)
Low priority — current behavior is correct. This is about polishing the dual-store UX over time.
Context
Right now, markdown files are the source of truth for skip decisions during import/refresh. This was formalized in #35 — SQLite is no longer consulted for skip logic to avoid stale state (e.g., deleting a page folder then re-importing would silently skip everything).
Friction area
There's no sync mechanism between SQLite and the markdown files. They can drift:
Future considerations
FreezePageskip logic edge cases (currently only covered by system tests)Low priority — current behavior is correct. This is about polishing the dual-store UX over time.