Architecture Changes:
- Split dialect implementations into separate files for better organization
- Move SQLite dialect from dialect.go to sqlite.go
- Keep only Dialect interface in dialect.go with comprehensive documentation
- Each dialect now in its own file following single responsibility principle
New Features:
- Add PostgreSQL dialect support (Postgres() function)
- PostgreSQL uses SERIAL PRIMARY KEY (auto-incrementing integer)
- PostgreSQL uses $1, $2 placeholders instead of ? for parameters
- PostgreSQL uses SELECT 1 for table existence check (more efficient)
- Both dialects implement proper SQL identifier quoting for security
Testing:
- Add comprehensive dialect-specific tests in sqlite_test.go
- Add comprehensive dialect-specific tests in postgres_test.go
- Test SQL generation for all dialect methods
- Test SQL injection protection via identifier escaping
- All tests pass (8 test functions, 10 subtests)
Documentation:
- Update README with PostgreSQL usage example
- Add "Supported Databases" section listing SQLite and PostgreSQL
- Improve code examples with proper imports and error handling
- Document how to implement Dialect interface for other databases
File Structure:
- dialect.go: Interface definition only (18 lines)
- sqlite.go: SQLite dialect implementation (39 lines)
- postgres.go: PostgreSQL dialect implementation (42 lines)
- sqlite_test.go: SQLite dialect tests (67 lines)
- postgres_test.go: PostgreSQL dialect tests (67 lines)
Security:
- Both dialects use quoteIdentifier() for SQL injection protection
- Identifiers are quoted and internal quotes are escaped
- Follows SQL standard quoting mechanism (double quotes for escaping)
This change maintains backward compatibility while adding PostgreSQL support
and improving code organization for future dialect additions.
