Skip to content

v0.1.17

Choose a tag to compare

@danielsreichenbach danielsreichenbach released this 23 Feb 05:39
· 82 commits to main since this release
04278a6

Added

  • add per-test benchmark script for PUC-Rio vs rilua comparison

Documentation

  • update benchmarks with bench-all.lua results after dead-key fix
  • update CHANGELOG with dead-key fix and benchmark additions
  • add per-test benchmark results and optimization priorities

Fixed

  • reuse dead hash slots in new_key to prevent chain corruption

Fixed

  • Fix hash chain corruption when inserting keys into tables with GC-swept
    nil-valued entries. After GC sweeps key strings, stale GcRefs caused
    main_position to return incorrect buckets, corrupting hash chains via
    Brent's Case A relocation. Detected as globals (e.g. assert) becoming
    nil when running multiple test files sequentially.

Added

  • Per-test benchmark script (scripts/benchmark-tests.sh) for comparing
    PUC-Rio and rilua performance on each test file individually

Documentation

  • Update per-test benchmark results with bench-all.lua: rilua is 1.75x
    slower individual (1.93x combined runner) vs PUC-Rio
  • Add optimization priority analysis and profiling workflow