Flaky test not reproduced by RSpec --bisect: diagnosing non‑order‑dependent failures
When a test suite exhibits intermittent failures, developers often turn to RSpec’s --bisect flag to find the minimal set of examples that reproduces the problem under a fixed random seed. The goal is to decide whether the observed flakiness stems from order‑dependent state leakage or from other sources such as timing, concurrency, or external service calls,