Skip to content

CHANGELOG.md

Latest commit

 

History

History
70 lines (51 loc) · 3.12 KB

CHANGELOG.md

File metadata and controls

70 lines (51 loc) · 3.12 KB

Changelog

All notable changes to this project will be documented in this file.

[Unreleased]

Optimized

  • Regex Performance:
    • Extracted inline regex compilation in read_python_file to a module-level constant (PYTHON_STRING_REGEX) to prevent recompilation in loops.
    • Updated apply_rule to use compiled regex objects directly (regex.sub) instead of re.sub for improved performance.

[1.1.0] - 2026-01-22

Changed

  • Switched to AST-based Python parsing: Replaced regex-based string extraction with Python's ast module.
    • Improved Robustness: Eliminated edge cases where comments or non-assignment strings were incorrectly formatted.
    • Enhanced F-string support: More reliable parsing of SQL within f-strings.
    • False Positive Prevention: Correctly ignores SQL-like text in docstrings and comments, which were previously matched by the regex.
    • Strict Targeting: Only formats strings that are actual variable assignments.
  • Modernized File Handling: Adopted pathlib for all file system operations.
  • Type Safety: Added comprehensive type hints to the codebase.

[1.0.0] - 2026-01-21

Changed

  • Migrated from Poetry to uv for dependency management and builds

    • Converted pyproject.toml to PEP 621 format
    • Switched build backend from poetry.masonry.api to hatchling
    • Replaced poetry.lock with uv.lock

  • Simplified codebase structure

    • Merged readsql/parse_args.py into readsql/__main__.py (single module)
    • Renamed entry point from command_line() to main()
    • Standardized parameter naming: file_name → path throughout
    • Replaced read_all_files_into() with generator-based collect_files()
    • Replaced format_files() with simpler run_on_files()
    • Replaced replace() and replace_part_of_string() with apply_rule()
    • Renamed read_regexes() to get_regexes()
    • Removed write_file(), get_message(), get_all_files(), validate() helper functions

  • Bumped version from 0.1.1 to 1.0.0

  • Updated minimum Python version from 3.6 to 3.8

  • Consolidated linting tools: Replaced zimports with ruff for import sorting and updated CI to show diffs on failure.

Added

  • GitHub Actions workflows:
    • ci.yml for testing and linting
    • publish-pypi.yml for publishing releases
  • AGENTS.md with development commands and architecture documentation
  • CHANGELOG.md to track changes

Removed

  • Removed argparse from dependencies (it's part of Python stdlib since 2.7)
  • Deleted setup.py (replaced by PEP 621 metadata in pyproject.toml)
  • Deleted readsql/parse_args.py (merged into __main__.py)
  • Deleted poetry.lock (replaced by uv.lock)

Fixed

  • Fixed SQL string detection in Python files:
    • Regex now correctly identifies SQL strings wrapped in single quotes (''') and triple double quotes (""").
    • Fixed a bug where replacing content with a different length caused index drift, corrupting the file.
  • Updated tests/check.py to use current API (read_sql_file, read_python_file with positional args)
  • Updated tests/test_files.py to use path= instead of file_name= keyword argument