Rantly

What Rantly is

Rantly is a Ruby library for producing random test data and checking properties across generated values. Its recursive-descent interpreter returns values such as strings, integers, and floats, with generators for integers, ranges, floats, booleans, choices, weighted frequencies, arrays, strings, and hashes. The project provides a Test::Unit extension and documents extensions for Minitest and RSpec, so property checks can be used with those test runners. Checks run 100 times by default; you can change the count with an argument to `check` or through the `RANTLY_COUNT` environment variable. Optional shrinking can reduce failing values, but its search is capped at 1,024 operations, does not explore the entire failure tree, and leaves ordinary arrays and hashes unshrunk. Rantly is free, self-hosted software for Ruby, published under the MIT License. Install it with `gem install rantly` or add it to a Gemfile. The README states a Ruby requirement of 3.2 or higher; RubyGems lists version 3.0.0, dated October 21, 2024, as requiring Ruby 3.3.0 or higher. The project welcomes new contributors.

Who it suits

Rantly suits Ruby developers who want to generate varied test inputs and check properties using Test::Unit, Minitest, or RSpec. It is a fit for projects that can use a Ruby library and want optional shrinking to help narrow failing values.

Limits

Property checks run 100 times by default. Optional shrinking is limited to at most 1,024 operations, does not fully traverse the failure tree, and does not shrink ordinary arrays or hashes.

Getting started

1. Open the Rantly GitHub repository.
2. Use Ruby 3.3.0 or higher for the listed RubyGems version 3.0.0.
3. Install with `gem install rantly` or add Rantly to a Gemfile.
4. Use its documented extension for Test::Unit, Minitest, or RSpec.
5. Adjust the default check count with an argument to `check` or `RANTLY_COUNT` if needed.

Verdict

Choose Rantly if you need free Ruby generators and property checks with the listed test runners. Look elsewhere if you need ordinary arrays and hashes to shrink, or a shrinking search that explores the full failure tree.