rvben/rumdl v0.0.176

on GitHub

Added

• Git-style intelligent configuration merging
  • User and project configs can now intelligently combine instead of simply overriding
  • disable array uses union semantics - user can add to project disables
  • enable array uses replace semantics - project can enforce specific rules
  • When a rule appears in both enable and disable, enable wins (project can override user)
  • Configuration hierarchy is now explicit and follows Git's pattern:
    • Default (0) < UserConfig (1) < PyprojectToml (2) < ProjectConfig (3) < CLI (4)
  • Renamed config sources for clarity:
    • ConfigSource::RumdlToml → UserConfig (global user config)
    • ConfigSource::RumdlToml → ProjectConfig (project-level config files)
    • Removed Markdownlint source (now uses ProjectConfig)
  • New API methods:
    • SourcedValue::merge_union() - additive merging for disable arrays
    • SourcedValue::merge_override() - replacement merging for enable arrays
  • Example use cases:
    • User disables MD013 globally → Project enables MD013 → MD013 is enabled ✓
    • Project disables MD001, MD003 → User disables MD013 → All three are disabled ✓
  • 17 comprehensive tests covering merge semantics, precedence, and conflict resolution
  • Matches configuration pattern used by git, eslint, prettier, and other modern tools

Performance

• Major memory optimizations eliminating string allocations

  • Converted LineInfo to use byte ranges instead of owned strings (50-80% memory reduction)
  • Eliminates N string allocations where N = number of lines in document
  • Added zero-copy content(&self, source: &str) -> &str method for on-demand access
  • Converted LineIndex to borrow &'a str instead of cloning (eliminates one full document copy per lint operation)
  • Converted ParsedLink/ParsedImage to use Cow<'a, str> (60-80% reduction in heap allocations for link/image parsing)
  • Zero-cost when borrowing, minimal overhead when owning - most data borrowed directly from source

• Eliminated O(n²) complexity in multiple rules

  • Fixed quadratic bottlenecks in MD027 (Multiple spaces after blockquote symbol)
  • Fixed quadratic bottlenecks in MD020 (No space inside hashes on closed atx style heading)
  • Fixed quadratic bottlenecks in MD046 (Code block style)
  • All three rules now use pre-computed context data with O(1) lookups

• Optimized CLI commands

  • rumdl config now executes in ~18ms (eliminated duplicate rule instantiation)
  • Removed duplicate instantiation of ~50 rule objects between main.rs and formatter.rs
  • Cleaner architecture with single source of truth for rule instances

Fixed

• Config loading: User config now always loaded as base layer (#131)

  • Fixed LSP server ignoring user config when finding project-level config files
  • User configuration is now always loaded first (unless --no-config is used)
  • Project configs merge on top of user config, CLI flags have highest priority
  • Configuration hierarchy now consistent between CLI and LSP:
    1. User/global config (~/.config/rumdl/rumdl.toml) - base layer
    2. Project config (discovered or explicit) - overrides user config
    3. CLI flags - highest priority
  • Matches pattern used by git, eslint, prettier, and other tools
  • Added regression test verifying user config preserved with explicit project config

• MD035: Frontmatter delimiter false positives (#40)

  • Fixed incorrect flagging of YAML/TOML frontmatter delimiters (---/+++) as horizontal rules
  • Rule now correctly skips frontmatter in three places:
    • most_prevalent_hr_style() - don't count frontmatter HRs for prevalence
    • check() - don't flag frontmatter delimiters as violations
    • fix() - don't replace frontmatter delimiters when fixing
  • Uses pre-computed LineInfo.in_front_matter field for efficient detection

Changed

• Architecture: Consistent use of pre-computed context data

  • MD035 now uses pre-computed LineInfo.in_front_matter instead of function calls
  • All rules now consistently use ctx.line_index for line-based operations
  • Eliminates redundant function calls and O(n) scans
  • More efficient with O(1) field access patterns

• Test infrastructure improvements

  • Implemented dynamic fixture downloading for performance tests
  • Test downloads now happen on-demand instead of being checked into repository
  • Better handling of large test files for performance benchmarking

• Test quality improvements

  • Corrected ESM block test assertions to match actual MDX behavior
  • ESM blocks only exist at TOP of MDX files and end at first non-ESM line
  • Tests now assert correct behavior rather than documenting implementation quirks
  • Fixed MD033 test assertions to properly validate HTML inline rules

Breaking Changes

For library users only (CLI users unaffected):

• LineInfo.content field is no longer public - use new content(source: &str) method instead
• Since LineInfo was primarily internal API, impact should be minimal

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.