Description
Motivation
Parameterized tests are currently represented as a single <testcase>
in xUnit output. This makes it difficult to distinguish how many individual runs of the test with distinct parameters passed or failed, and to correlate them back to which input they passed/failed on.
We use parameterized tests extensively as "table driven tests", where common logic is exercised against various input/output expectations represented as an array of test cases.
The current xunit output is not well adapted to this pattern, and we'd love to look for ways of representing the individual [test, arguments] tuples as distinct <testcase>
elements in the report.
Proposed solution
It might make senes to model parameterized tests as their own nested test suites, with their arguments as individual cases. The test cases could be named after the testDescription
, assuming that argument adopts CustomTestStringConvertible
.
Alternatives considered
No response
Additional information
No response