rvben/rumdl v0.0.164

on GitHub

Added

• File-Level Caching (Ruff-inspired): Dramatic performance improvements for repeat runs

  • Blake3-based content hashing for fast cache lookups
  • Automatic cache invalidation on content, config, or version changes
  • Cache stored in .rumdl-cache/{version}/{hash}.json
  • CLI flags: --no-cache to disable, --cache-dir to customize location
  • Enabled by default for instant subsequent runs

• Thread-Safe Parallel Caching: Best of both worlds - parallelization AND caching

  • Implemented Arc<Mutex> for safe cache sharing across threads
  • Mutex locked ONLY for brief cache get/set operations
  • Full parallelization during expensive linting operations
  • Matches Ruff's architecture for optimal performance

• Convergence Detection: Added hash-based detection to identify when fixes have stabilized

  • Stops iteration when content hash remains unchanged
  • More efficient than counting rule applications
  • Returns convergence status in fix results

• Convergence Failure Warnings: Report when auto-fix doesn't converge (Ruff-style)

  • Warns if 100 iteration limit reached without convergence
  • Shows rule codes involved in potential infinite loop
  • Encourages bug reports for convergence failures
  • Available via RUMDL_DEBUG_FIX_PERF environment variable

Changed

• Auto-fix Iteration: Automatic iteration until convergence (fixes #88)

  • --fix now automatically iterates up to 100 passes until content stabilizes (same as Ruff)
  • No need to manually re-run rumdl check --fix multiple times
  • Hash-based convergence detection prevents unnecessary iterations
  • Significantly improves user experience for multi-pass fix scenarios

• Unified Linting Architecture: Removed ~60 lines of duplicate linting logic

  • Refactored process_file_collect_warnings to use process_file_inner
  • Single code path for all file processing
  • Cache works for ALL output formats (text, JSON, GitLab, SARIF, JUnit)

• Parallel File Processing for Fix Mode: 4.8x speedup on multi-file fixes

  • Previously fix mode was always sequential
  • Now uses parallel processing when safe (multiple independent files)
  • Each file processes all its fix iterations independently

Fixed

• Multi-pass Fixes: No longer require manual re-runs to apply all possible fixes

  • Previously users had to run rumdl check --fix multiple times
  • Now automatically handles dependent rule fixes in single command
  • Examples: MD010 (tabs) before MD007 (list indent), MD013 (line length) before MD009 (trailing spaces)

• Cache Correctness: Include enabled rules in cache key (Ruff-style)

  • Cache now respects --enable/--disable CLI flags
  • Different rule configurations create separate cache entries
  • Prevents incorrect cached results when switching rule sets
  • Changed LintWarning.rule_name from Option<&'static str> to Option<String> for proper serialization

• Cache Parallelization: Cache now works correctly with parallel processing

  • No mutex contention during parallel file processing
  • All output formats benefit from caching (previously only JSON/GitLab/SARIF/JUnit)

Performance

• Single file with cache: 943ms → 7ms (135x faster)
• Multi-file (21 files) cold cache: 14.4s → 4s (parallel processing)
• Multi-file (21 files) warm cache: 14.4s → 0.019s (757x faster!)
• JSON format (17 files) with cache: 13.9s → 60ms (231x faster)

Downloads

Installation

Using uv (Recommended)

uv tool install rumdl

Using pip

pip install rumdl

Using pipx

pipx install rumdl

Direct Download

Download the appropriate binary for your platform from the table above, extract it, and add it to your PATH.