Release 0.2.0

Release version 0.2.0

Release 0.1.0

Added

• check-chars: Character validation for source files
  • ASCII + emoji + technical symbols (arrows, box-drawing, math operators)
  • Dangerous invisible/bidi character detection (Trojan Source - CVE-2021-42574)
  • Comment-aware checking for .rs, .py, .js, .ts, .tsx, .jsx files
  • Built-in charset presets: chinese, japanese, korean
  • Custom charset JSON file support
  • APDEV_EXTRA_CHARS environment variable fallback
  • Automatic directory walking with configurable skip lists
  • Default project paths: src/, tests/, examples/, and config files
• release: Interactive release automation for crates.io
  • Auto-detect project name from Cargo.toml and GitHub repo from git remote
  • Step-by-step menu: version verify, test, build, tag, GitHub release, publish
  • Silent mode (--yes) for CI/CD
• Pre-commit hook integration via language: system

Removed

• check-imports command (Python-specific, not applicable to Rust projects)
• config.rs / pyproject.toml config loader (no longer needed)
• toml dependency

Release 0.2.2

Release version 0.2.2

Release 0.2.1

Release version 0.2.1

Release 0.2.3

Release version 0.2.3

See CHANGELOG.md for details.

Release 0.2.2

Release version 0.2.2

See CHANGELOG.md for details.

Release 0.2.1

Fixed

• Release script version check now supports dynamic versions via importlib.metadata (not just static __version__ = "x.x.x")
• Added hint message suggesting pip install -e . when version mismatch is detected

Release 0.2.0

Added

• Custom charset support via shared JSON definition files (shared/charsets/)
• loadCharset(nameOrPath) — load a bundled preset (e.g. chinese) or custom JSON file
• resolveCharsets(charsetNames, charsetFiles) — merge base + extra charsets with deduplication
• Bundled charset presets: base, chinese, japanese, korean
• CLI --charset flag (repeatable) to enable preset charsets
• CLI --charset-file flag (repeatable) to load custom charset JSON files
• APDEV_EXTRA_CHARS environment variable — comma-separated charset names or file paths, used as fallback when no CLI args provided
• check-chars: Block Elements (U+2580-U+259F) and Braille Patterns (U+2800-U+28FF) to default allowed ranges
• Exported isDangerousChar from public API (index.ts)
• Exported RangeEntry, DangerousEntry, CharsetData interfaces

Changed

• Character ranges and dangerous codepoints now loaded from charsets/base.json instead of hardcoded constants
• checkFile() and checkPaths() accept optional extraRanges and dangerousMap params for custom charset support
• isAllowedChar() now excludes dangerous codepoints (Trojan Source vectors) even though they fall within the General Punctuation allowed range
• typescript moved from dependencies to optional peerDependencies — build output reduced from 9.50 MB to 17.88 KB
• Commander.js --yes option syntax fixed to -y, --yes

Fixed

• isAllowedChar() previously returned true for dangerous codepoints (U+200B, U+202E, etc.) because they fall within the General Punctuation range (U+2000-U+206F)
• checkFile() now uses lazy-loaded cache instead of re-parsing base.json on every call
• Error details now included in file read failure messages

Release 0.2.0

Added

• Custom charset support via shared JSON definition files (shared/charsets/)
• load_charset(name_or_path) — load a bundled preset (e.g. chinese) or custom JSON file
• resolve_charsets(charset_names, charset_files) — merge base + extra charsets with deduplication
• Bundled charset presets: base, chinese, japanese, korean
• CLI --charset flag (repeatable) to enable preset charsets
• CLI --charset-file flag (repeatable) to load custom charset JSON files
• APDEV_EXTRA_CHARS environment variable — comma-separated charset names or file paths, used as fallback when no CLI args provided
• check-chars: Block Elements (U+2580-U+259F) and Braille Patterns (U+2800-U+28FF) to default allowed ranges

Changed

• Character ranges and dangerous codepoints now loaded from charsets/base.json instead of hardcoded constants
• check_file() and check_paths() accept optional extra_ranges and dangerous kwargs for custom charset support
• is_allowed_char() now excludes dangerous codepoints (Trojan Source vectors) even though they fall within the General Punctuation allowed range

Fixed

• is_allowed_char() previously returned True for dangerous codepoints (U+200B, U+202E, etc.) because they fall within the General Punctuation range (U+2000-U+206F)
• load_charset() now catches all exceptions from importlib.resources (not just FileNotFoundError)
• check_file() now uses lazy-loaded cache instead of re-parsing base.json on every call

Release 0.1.6

Added

• check-chars: Dangerous character blacklist (bidi controls + zero-width chars, CVE-2021-42574) with comment-aware detection — flags in code, allows in comments
• check-chars: is_dangerous_char() public function and DANGEROUS_CODEPOINTS constant
• check-chars: Comment region detection state machine for Python (#) and JS/TS (//, /* */) with string literal tracking
• check-imports: Relative import support — from .foo import x and from . import foo now resolve correctly for cycle detection

Fixed

• check-imports: ImportAnalyzer now handles node.level > 0 (relative imports) by resolving against current_module with correct package vs module distinction