Release v0.14.0

Added

• Enhanced association safety with bounds checking to prevent IndexError (#133, #195)
• Association foreign key validation in changesets (#249, #195)
• Comprehensive test suite for association safety
• Transaction support with savepoints for nested operations (#249)
• Enhanced changeset validation with association support

Fixed

• IndexError in belongs_to associations (#133, #195)
• IndexError in has_many associations (#133, #195)
• IndexError in has_one associations (#133, #195)
• Query string interpolation corruption (#217)
• Multi-transaction association operations (#249)
• Association foreign key validation (#249, #195)

Improved

• Query and transaction testing coverage
• Live transaction enhancements for better consistency
• Safe array access patterns to all association macros

v0.13.0 Release

Summary

This release brings significant improvements to Crecto's infrastructure, including comprehensive test coverage, new error types, UUID support, and performance optimizations. It establishes a solid foundation for future user story implementations.

Features

Test Infrastructure

• Comprehensive Test Coverage: Added extensive test coverage for core functionality including:
  • Bulk configuration handling (spec/bulk_config_spec.cr)
  • Error type handling and validation (spec/error_types_spec.cr)
  • Prepared statement cache behavior (spec/prepared_statement_cache_spec.cr)
  • Query iteration and result processing (spec/query_iterator_spec.cr)

Quality Infrastructure

• UUID Type Support: Added UUID type support to schema field types with UUID_FORMAT constant
• Enhanced Error Types: Implemented comprehensive error handling with:
  • BulkError for bulk operation errors
  • AssociationError for association-related errors
  • IteratorError for query iteration errors
• Bulk Configuration: Created BulkConfig singleton for managing operation thresholds
• Performance Optimizations: Added PreparedStatementCache infrastructure for improved query performance
• Query Iterator: Implemented QueryIterator(T) base class for memory-efficient large dataset processing

Type System Improvements

• Enhanced Type Checking: Improved overall type checking system robustness
• Type Cleanup: Removed duplicate Int64 type definition

Development Experience

• Multi-Database Testing: Added comprehensive test database configurations
• Example Repository: Created example spec repository for easier testing
• Flexible Updates: Enhanced update_from_hash method for greater flexibility

Bug Fixes

• Type System: Fixed duplicate Int64 type definition in types.cr

Documentation

• Agent Documentation: Added AGENTS.md with comprehensive agent workflow documentation
• Project Instructions: Created CLAUDE.md with project-specific development guidelines
• Enhanced README: Updated README with improved documentation and structure
• Test Configuration: Added comprehensive multi-database testing setup in spec/test_repos.cr

Breaking Changes

None - this is a minor release with backward-compatible improvements.

Upcoming

This release establishes the foundation for Phase 3 user story implementations including:

• Bulk operations with configurable thresholds
• Prepared statement caching for performance optimization
• Memory-efficient query iteration for large datasets
• Enhanced error handling and reporting

Note: All tests pass across all supported databases (PostgreSQL, MySQL, SQLite3).

• Fixed multi error checking, thanks @hutou
• Fixed installing via shards command

• Update for crystal 1.0

Update for crystal 0.34.0

• Fix shard version/git tag version mismatch

• Updated for crystal 0.31.0

• Fixed breaking bug
• Updated all dependencies to latest versions
• Updated to Crystal 0.30.1

• Update to crystal 0.27.0 - [@drujensen]https://github.com/drujensen

• Updated to crystal 0.26.1
• Allow defining models without any fields - @jreinert
• Queries in preloads
• Live transactions with Repo#transaction! - @jreinert
• Changed model to use JSON::Serializable - @jreinert
• Fields now correctly compare nil - @shiba-hiro
• Added Repo#join method override with string argument - @shiba-hiro
• Added more where and or_where method signatures - @jreinert
• Added int8 and int16 suppoert - @jianghengle
• Improved performance using String::Builder to build query strings - @jreinert
• Added Repo#get_by method overloads to use Query objects - @jreinert
• Compiler improvements by only including associations when used - @jianghengle
• Updated to use new crystal time / timezone format - @wontruefree
• Handle single-column updates for postgres 10.x - @faultyserver
• Added Schema#cast and Schema#cast! mass assignment methods - @jreinert
• Fixed has_one preloads - @jreinert