On Thu, Aug 18, 2022 at 9:20 AM Michał Winiarski
<michal.winiarski@xxxxxxxxx> wrote:
> Already did that - s/passing negative cmdline/passing invalid cmdline.
> The tests are still "negative tests" - in other words, tests that pass invalid
> data, and expect specific error condition to happen. We can't use "invalid
> tests" here, as that has different meaning (broken test).
>
> We could expand it into:
> "Tests that pass invalid data can be expressed as a single parameterized test
> case (...)"
>
> Would that work? Or should we keep it as "negative tests"?
Just my 2c, "negative tests" was easier for me to immediately grok this patch.
I'd prefer this patch as-is ("invalid" in the test names, "negative"
in the commit desc).
But it is a term that some people might not be familiar with, so I
don't feel too strongly either way.
Daniel
