Changelog

1.2.6

Fixed

• SQLite runtime schema sync now preserves explicit raw SQL defaults such as strftime(...) through sqlDefault(...), instead of serializing them as quoted string literals or treating them as rebuild-causing mismatches.
• SQLite schema inspection and entity generation now preserve expressive default SQL by emitting sqlDefault(...) for non-literal defaults instead of dropping them from generated models.

Changed

• The public column metadata surface now supports explicit raw SQL defaults through sqlDefault(...), keeping runtime/query adoption compatible with bootstrap DDL scenarios that depend on engine expressions.
• PostgreSQL and MySQL default serializers now accept the same raw-default wrapper so model metadata stays consistent across engines.

Validation

• focused SQLite regressions passed for schema sync and inspector generation with strftime(...) defaults
• TypeScript typecheck passed

1.2.5

Added

• The published package now exposes both ESM and CommonJS entry points so consumers can use either import ... from "sqlqueryhelperjs" or require("sqlqueryhelperjs").

Changed

• The build now emits both ESM output and a dedicated CommonJS bundle under dist/cjs, with conditional exports for each consumer type.
• The README now documents CommonJS usage alongside the existing ESM examples.

Validation

• npm run build passed
• npm pack --dry-run passed for sqlqueryhelperjs@1.2.5
• package entry smoke tests passed for both require("sqlqueryhelperjs") and import("sqlqueryhelperjs")

1.2.4

Added

• SQLite, PostgreSQL, and MySQL schema reflectors now support explicit data-preserving column renames through approvedColumnRenames: [{ table, from, to }].
• Schema sync previews now expose approved rename mappings through columnRenames, making destructive rebuild intent reviewable before apply.

Changed

• SQLite rebuild planning now combines general approved rename mappings with the existing case-only rename approval flow.
• PostgreSQL and MySQL rebuild copy steps now preserve values for approved renames with explicit source-to-target projections instead of treating them as dropped and re-added columns.
• Runtime guides now document the cross-engine rename approval workflow and semantic change detection guidance.

Validation

• focused schema sync tests passed for SQLite, PostgreSQL, and MySQL explicit rename preservation
• TypeScript typecheck passed
• documentation site regenerated successfully

1.2.3

Fixed

• SQLite inspector and reflector now preserve advanced indexes with expressions and where, instead of degrading them during inspection or regeneration.
• PostgreSQL inspector and reflector now preserve expression indexes and partial index predicates in generated entity metadata and reflect(...) index recreation.
• MySQL inspector and reflector now preserve functional indexes on servers that expose information_schema.statistics.expression, with a fallback path for older servers that only expose column metadata.
• CLI PostgreSQL inspection output now matches the runtime inspector behavior for expression and partial indexes.

Validation

• focused regression tests added for SQLite, PostgreSQL, and MySQL advanced index inspection and sync
• PostgreSQL and MySQL live integration tests passed against real database containers
• TypeScript typecheck passed
• documentation site regenerated

Operational Notes

• Existing generated entity files should be regenerated if they were produced before this fix and contain degraded advanced index output such as index([]) without the original expressions or where metadata.
• MySQL functional index fidelity depends on server metadata support for information_schema.statistics.expression; older servers still need manual review for version-specific index details.