2.4.0

Requirements

• Require Node.js 22 or newer (previously 20.17) and @teqfw/di 2.11 or newer within the 2.x line (previously 2.7).

Fixed in 2.4.0

• Describe connections using the selected dialect: SQLite logs its filename; server databases log database, host, and user without treating leftover filename settings as SQLite configuration.
• Omit raw connection setup errors from logs to avoid disclosing credentials while preserving the original thrown error.
• Resolve checked-JavaScript contract errors and verify the published declarations against a consumer of the actual npm artifact.

Documentation

• Clarify when independent applications sharing database space need an application-map prefix; packages contributing to one application do not require an extra prefix.
• Guide concise domain-based DEM names and explicit migration planning for logical path or physical prefix changes.
• Keep compilation access internal to Schema; hosts retain the successful loader result for preparation and history operations.

Development Snapshot Compatibility

• Table-backed identity allocation and the public Schema.getCompilation() getter existed only in intermediate development commits after 2.3.0 and are not part of this release.
• Consumers of those snapshots must retain the loader result instead of calling the getter. Databases created with allocated identities require an explicit migration/rebuild to native generation, preserving IDs, foreign keys, and native sequence state; changing the map alone does not convert existing storage.

Security

• Bind legacy PostgreSQL sequence names and values instead of interpolating SQL, validate sequence values before execution, and restore imported sequences through the active transaction.
• Preserve special JSON keys during fingerprinting and prevent inherited-property merges and prototype mutation.
• Replace the Markdown CLI dependency tree with a direct library runner and an override for the patched KaTeX line.

Development and Publication

• Bound @teqfw/log to 2.x and align Node.js declaration dependencies with Node.js 22.
• Add a verification command, publication checks, Node.js 22/24 CI, and a private vulnerability reporting policy.
• Document agent-operated local npm publication with lifecycle verification; GitHub Actions only verifies changes.
• Inspect a real npm tarball and verify its source syntax, consumer types, closed exports, and runtime DI resolution.
• Remove obsolete scripts that deleted the dependency lockfile and development assets.

2.3.0

2.3.0 - 2026-09-03

Added in 2.3.0

• Added optional lowercase dot-delimited root namespaces for concise DEM fragments.
• Published teqfw.db.schema as an ordinary package-owned DEM fragment for immutable schema snapshots and application history.

Changed

• Expanded fragment roots before composition, resolved local relation paths against the expanded root, and preserved source provenance.
• Simplified the package-owned declaration by omitting empty optional nodes; the compiler handles omitted nodes as empty structures.
• Ensured etc/teqfw.schema.json is included in the npm package and clarified the product positioning and integration boundaries in the README.

Removed in 2.3.0

• Removed the duplicate RELEASE.md; CHANGELOG.md is now the single release history.

@teqfw/db 2.2.0

Breaking changes

• DEM declarations and application maps must explicitly set version: 2.
• DEM v1 decoding, compatibility facades, legacy physical projections, legacy CRUD, legacy selection DTOs, and the logger facade were removed.
• The connection accessor is now getClient(); getKnex() is removed.

Highlights

• Verification is organized into unit, integration, acceptance, opt-in, package, manual, and fixture layers.
• The retained v1 branch remains the historical reference for consumer-specific migration guidance.

@teqfw/db 2.1.1

Changed

• Reworked the README with the current package overview, boundaries, installation guidance, and agent-skill usage.

See CHANGELOG.md for details.

@teqfw/db 2.1.0

Highlights

• Added validated DEM compilation, diagnostics, schema planning, and rebuild execution.
• Added PostgreSQL, MySQL, SQLite, and PostgreSQL pgvector dialect support.
• Migrated runtime composition to @teqfw/di 2.x and the current TeqFW configuration contracts.
• Added public type declarations and package-owned consumer guidance.

Breaking changes

• Consumers must use the DI 2.x composition model and the current configuration contract.
• The previous implementation remains available on the v1 branch.

See CHANGELOG.md for details.

Enhanced selection DTO handling and order-by consistency

• Refactored TeqFw_Db_Shared_Dto_List_Selection: separated class definition from metadata and improved DTO
  initialization via .create().
• Added TeqFw_Db_Shared_Util_Select utility for composing selection DTOs from basic JS objects with support for
  filtering, sorting, and pagination.
• Aligned all method signatures in Crud, Repository, and QueryBuilder to accept unified selection objects
  instead of conditions, sorting, and pagination separately.
• Deprecated createDto() method in selection DTO, use create() instead for consistency.
• Updated docblocks and type annotations across repository and query interfaces for clarity and correctness.
• Minor refactorings to improve internal consistency and readability in database-related modules.

Support for `rowsLimit` and `rowsOffset` in selection

• Added support for rowsLimit and rowsOffset parameters in selection queries.

Added `onCommit` and `onRollback` callbacks in TrxWrapper

• Added onCommit and onRollback callbacks in TrxWrapper for handling post-transaction results.
• Implemented fetchTablesByDependencyOrder in RdbTables and Schema to list tables by dependency order, enhanced date
  item formatting for exports.
• Corrected export mechanisms in Export to properly handle table dependencies and data types.
• Enhanced data import processing in Import, including proper sequence restoration and error handling during data
  insertion.
• Refactored readMany and related methods in Crud and Schema for improved readability and extensibility, added
  selection parameters for data reading operations.

Use Selection in CRUD Base for `readMany`

• Implemented selection functionality in the CRUD base for the readMany operation.

Use operations in `where` clause

• Enhanced the query builder to handle array values in where clauses (e.g., ('id', ['<=', 4]) becomes
  ('id', '<=', 4)).
• Improve JSDoc in TeqFw_Db_Back_App_Crud.updateOne to prevent IDEA highlights.