Skip to content

Releases: aatuh/randutil

v2.1.3

21 May 17:03

Choose a tag to compare

Summary

Patch release focused on documentation quality, executable examples, repository hygiene, and regression-test coverage. No public API signatures changed.

Added

  • Executable examples for root-package derivation and workspace usage, NanoID generation, and ULID generation.
  • Regression tests for workspace usage snapshots, disabled-cache behavior, cached stream eviction, derived root contracts, and byte-token ownership.

Changed

  • Tightened README navigation, quick-start structure, and deterministic testing guidance.
  • Consolidated benchmark evidence guidance in docs/benchmarks.md to reduce release-doc drift.
  • Updated repository ignore rules for local hidden-file artifacts.

Documentation

  • Clarified vulnerability-reporting expectations in SECURITY.md.
  • Removed the temporary production-readiness tracking file now that readiness evidence lives in permanent project docs and GitHub settings.

Validation

v2.1.2

21 May 04:46

Choose a tag to compare

Summary

  • Hardened invalid-input handling: collection probabilities now reject NaN and infinities, and Bernoulli has regression coverage for non-finite probabilities.
  • Fixed signed full-range integer generation so ranges spanning MinInt64 or MinInt can return the minimum value without ErrResultOutOfRange.
  • Rejected UUID v7 timestamps that exceed the 48-bit timestamp field.
  • Updated Go/tooling baseline and dependencies: Go 1.25.0, golang.org/x/crypto v0.51.0, govulncheck v1.3.0, gosec v2.26.1, golangci-lint v2.12.2.
  • Refreshed GitHub Actions versions and normalized SARIF rule tags before upload.

Documentation

  • Clarified production security boundaries: not FIPS/audited crypto; seed quality matters; callers needing strict OS RNG or FIPS paths should use crypto/rand.Reader directly.
  • Clarified default RNG state wording and UUID v7 / ULID ordering semantics.

Validation

Benchmarks

Benchmark evidence is recorded in docs/benchmarks-v2.1.2.md. The v2.1.1 fallback benchmark run found no benchmark functions, so this release does not claim a measured performance improvement.

v2.1.1

02 Oct 13:11

Choose a tag to compare

Add facade for convenience.

v2.0.0

20 Sep 13:27

Choose a tag to compare

Major refactoring and tests.

v1.0.2

16 Sep 17:42

Choose a tag to compare

Missing package updated.

v1.0.1

16 Sep 17:39

Choose a tag to compare

Change old package name. Update README.

v1.0.0

16 Sep 17:34

Choose a tag to compare

Random utility functionality.