audit-cache

To detect stale cache data, you audit invalidation signals like mtime, size, and hashes across JSON and SQLite artifacts to verify coherence and prevent silent data corruption. This requires inspecting actual code paths for invalidation logic.

JSON cache corruption occurs when decoding exceptions happen during read operations. Proper corruption handling assesses whether the system catches JSON decode exceptions and automatically cleans up corrupted cache files to prevent stale data persistence.

Validating SQLite cache coherence involves auditing hacienda_cache read and write behavior constraints. You check access patterns to confirm the database is treated consistently and read-only where intended for external fiscal status data.

Relying solely on file mtime for cache invalidation of parsed PDFs can be fragile. A thorough cache invalidation strategy review checks whether mtime, file size, or hashes are used correctly rather than relying on assumptions.

The best way to audit cache invalidation logic is to inspect actual current code paths for invalidation signals and corruption-handling logic. This verifies that pdf_cache and xml_cache are invalidated correctly using mtime, size, or hashes.

Hacienda_cache returns inconsistent fiscal status data when write and read behavior constraints are not treated consistently. A coherence check audits access patterns to confirm the SQLite database is treated read-only where intended.