With @Test(arguments:), how are several argument collections combined, and how is a failure reported?

One argument collection runs the body once per element. Two collections give the Cartesian product: arguments: [1, 2], ["a", "b", "c"] runs six cases, which the check confirms (“with 6 test cases passed”). To pair elements instead, pass zip(a, b), which gives two cases from two-element collections. Each case is its own test case with its arguments in the name, so a failure points at the exact input: in the check, only n → 3 of four cases records an issue, the test as a whole is reported failed, and the other three cases still run and pass. Arguments must be Sendable, and cases run in parallel, so the body must not share mutable state between cases.

Prefer this to a for loop inside one test: a loop stops being informative after the first failure (or hides which element failed), cannot be rerun for one input, and gives one result. Prefer a small named type for the inputs when there are several fields (a struct with an expected value and a description), and keep the product small, because it grows multiplicatively: three collections of ten are a thousand cases. Parameterised tests are also the right tool for table-driven tests such as parsing and formatting edge cases. XCTest has no equivalent; people write loops, or generate one method per input.